From b77bd3ba762f34e5eb731134cf50e233d1060053 Mon Sep 17 00:00:00 2001 From: Fabio Estevam Date: Wed, 27 Nov 2024 16:06:05 -0300 Subject: [PATCH 001/196] ARM: imx: Re-introduce the PINCTRL selection Since commit 17d210018914 ("ARM: imx: Allow user to disable pinctrl"), the CONFIG_PINCTRL option is no longer implicitly selected, causing several i.MX SoC pinctrl drivers no longer getting selected by default. This causes boot regressions on the ARMv4, ARMv5, ARMv6 and ARMv7 i.MX SoCs. Fix it by selecting CONFIG_PINCTRL as before. This defeats the purpose of 7d210018914 ("ARM: imx: Allow user to disable pinctrl"), but it is the less invasive fix for the boot regressions. The attempt to build Layerscape without pinctrl can still be explored later as suggested by Arnd: "Overall, my best advice here is still to not change the way i.MX pinctrl works at all, but just fix Layerscape to not depend on i.MX. The reason for the 'select' here is clearly that the i.MX machines would fail to boot without pinctrl, and changing that because of Layerscape seems backwards." Fixes: 17d210018914 ("ARM: imx: Allow user to disable pinctrl") Reported-by: Guenter Roeck Closes: https://lore.kernel.org/linux-arm-kernel/49ff070a-ce67-42d7-84ec-8b54fd7e9742@roeck-us.net/ Signed-off-by: Fabio Estevam Acked-by: Arnd Bergmann Reviewed-by: Linus Walleij Tested-by: Guenter Roeck Link: https://lore.kernel.org/20241127190605.1367157-1-festevam@gmail.com Signed-off-by: Linus Walleij --- arch/arm/mach-imx/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/arm/mach-imx/Kconfig b/arch/arm/mach-imx/Kconfig index e4fe059cd861..dc47b2312127 100644 --- a/arch/arm/mach-imx/Kconfig +++ b/arch/arm/mach-imx/Kconfig @@ -6,6 +6,7 @@ menuconfig ARCH_MXC select CLKSRC_IMX_GPT select GENERIC_IRQ_CHIP select GPIOLIB + select PINCTRL select PM_OPP if PM select SOC_BUS select SRAM From 793baff3f24f16dab9061045e23eea67724feae6 Mon Sep 17 00:00:00 2001 From: Honglei Wang Date: Fri, 29 Nov 2024 17:10:03 +0800 Subject: [PATCH 002/196] sched_ext: Add __weak to fix the build errors commit 5cbb302880f5 ("sched_ext: Rename scx_bpf_dispatch[_vtime]_from_dsq*() -> scx_bpf_dsq_move[_vtime]*()") introduced several new functions which caused compilation errors when compiled with clang. Let's fix this by adding __weak markers. Signed-off-by: Honglei Wang Signed-off-by: Tejun Heo Fixes: 5cbb302880f5 ("sched_ext: Rename scx_bpf_dispatch[_vtime]_from_dsq*() -> scx_bpf_dsq_move[_vtime]*()") Acked-by: Andrii Nakryiko --- tools/sched_ext/include/scx/common.bpf.h | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/tools/sched_ext/include/scx/common.bpf.h b/tools/sched_ext/include/scx/common.bpf.h index 2f36b7b6418d..625f5b046776 100644 --- a/tools/sched_ext/include/scx/common.bpf.h +++ b/tools/sched_ext/include/scx/common.bpf.h @@ -40,9 +40,9 @@ void scx_bpf_dsq_insert(struct task_struct *p, u64 dsq_id, u64 slice, u64 enq_fl void scx_bpf_dsq_insert_vtime(struct task_struct *p, u64 dsq_id, u64 slice, u64 vtime, u64 enq_flags) __ksym __weak; u32 scx_bpf_dispatch_nr_slots(void) __ksym; void scx_bpf_dispatch_cancel(void) __ksym; -bool scx_bpf_dsq_move_to_local(u64 dsq_id) __ksym; -void scx_bpf_dsq_move_set_slice(struct bpf_iter_scx_dsq *it__iter, u64 slice) __ksym; -void scx_bpf_dsq_move_set_vtime(struct bpf_iter_scx_dsq *it__iter, u64 vtime) __ksym; +bool scx_bpf_dsq_move_to_local(u64 dsq_id) __ksym __weak; +void scx_bpf_dsq_move_set_slice(struct bpf_iter_scx_dsq *it__iter, u64 slice) __ksym __weak; +void scx_bpf_dsq_move_set_vtime(struct bpf_iter_scx_dsq *it__iter, u64 vtime) __ksym __weak; bool scx_bpf_dsq_move(struct bpf_iter_scx_dsq *it__iter, struct task_struct *p, u64 dsq_id, u64 enq_flags) __ksym __weak; bool scx_bpf_dsq_move_vtime(struct bpf_iter_scx_dsq *it__iter, struct task_struct *p, u64 dsq_id, u64 enq_flags) __ksym __weak; u32 scx_bpf_reenqueue_local(void) __ksym; From b03917e02bf9861be887a7e67c399b3b014f88be Mon Sep 17 00:00:00 2001 From: Konstantin Andrikopoulos Date: Wed, 27 Nov 2024 15:07:38 +0000 Subject: [PATCH 003/196] rust: add safety comment in workqueue traits Add missing safety comments for the implementation of the unsafe traits WorkItemPointer and RawWorkItem for Arc in workqueue.rs Link: https://github.com/Rust-for-Linux/linux/issues/351. Co-developed-by: Vangelis Mamalakis Signed-off-by: Vangelis Mamalakis Suggested-by: Miguel Ojeda Reviewed-by: Alice Ryhl Signed-off-by: Konstantin Andrikopoulos Signed-off-by: Tejun Heo --- rust/kernel/workqueue.rs | 18 ++++++++++++++++-- 1 file changed, 16 insertions(+), 2 deletions(-) diff --git a/rust/kernel/workqueue.rs b/rust/kernel/workqueue.rs index 4d1d2062f6eb..fd3e97192ed8 100644 --- a/rust/kernel/workqueue.rs +++ b/rust/kernel/workqueue.rs @@ -519,7 +519,15 @@ unsafe fn raw_get_work(ptr: *mut Self) -> *mut $crate::workqueue::Work<$work_typ impl{T} HasWork for ClosureWork { self.work } } -// SAFETY: TODO. +// SAFETY: The `__enqueue` implementation in RawWorkItem uses a `work_struct` initialized with the +// `run` method of this trait as the function pointer because: +// - `__enqueue` gets the `work_struct` from the `Work` field, using `T::raw_get_work`. +// - The only safe way to create a `Work` object is through `Work::new`. +// - `Work::new` makes sure that `T::Pointer::run` is passed to `init_work_with_key`. +// - Finally `Work` and `RawWorkItem` guarantee that the correct `Work` field +// will be used because of the ID const generic bound. This makes sure that `T::raw_get_work` +// uses the correct offset for the `Work` field, and `Work::new` picks the correct +// implementation of `WorkItemPointer` for `Arc`. unsafe impl WorkItemPointer for Arc where T: WorkItem, @@ -537,7 +545,13 @@ unsafe impl WorkItemPointer for Arc } } -// SAFETY: TODO. +// SAFETY: The `work_struct` raw pointer is guaranteed to be valid for the duration of the call to +// the closure because we get it from an `Arc`, which means that the ref count will be at least 1, +// and we don't drop the `Arc` ourselves. If `queue_work_on` returns true, it is further guaranteed +// to be valid until a call to the function pointer in `work_struct` because we leak the memory it +// points to, and only reclaim it if the closure returns false, or in `WorkItemPointer::run`, which +// is what the function pointer in the `work_struct` must be pointing to, according to the safety +// requirements of `WorkItemPointer`. unsafe impl RawWorkItem for Arc where T: WorkItem, From eb867d797d294a00a092b5027d08439da68940b2 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Tue, 26 Nov 2024 15:10:31 +0200 Subject: [PATCH 004/196] RDMA/bnxt_re: Remove always true dattr validity check res->dattr is always valid at this point as it was initialized during device addition in bnxt_re_add_device(). This change is fixing the following smatch error: drivers/infiniband/hw/bnxt_re/qplib_fp.c:1090 bnxt_qplib_create_qp() error: we previously assumed 'res->dattr' could be null (see line 985) Fixes: 07f830ae4913 ("RDMA/bnxt_re: Adds MSN table capability for Gen P7 adapters") Reported-by: kernel test robot Reported-by: Dan Carpenter Closes: https://lore.kernel.org/r/202411222329.YTrwonWi-lkp@intel.com/ Link: https://patch.msgid.link/be0d8836b64cba3e479fbcbca717acad04aae02e.1732626579.git.leonro@nvidia.com Acked-by: Selvin Xavier Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_fp.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index e42abf5be6c0..9af8aaadc99a 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -1000,9 +1000,7 @@ int bnxt_qplib_create_qp(struct bnxt_qplib_res *res, struct bnxt_qplib_qp *qp) u32 tbl_indx; u16 nsge; - if (res->dattr) - qp->is_host_msn_tbl = _is_host_msn_table(res->dattr->dev_cap_flags2); - + qp->is_host_msn_tbl = _is_host_msn_table(res->dattr->dev_cap_flags2); sq->dbinfo.flags = 0; bnxt_qplib_rcfw_cmd_prep((struct cmdq_base *)&req, CMDQ_BASE_OPCODE_CREATE_QP, From d0257e089d1bbd35c69b6c97ff73e3690ab149a9 Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Sat, 30 Nov 2024 13:06:41 +0300 Subject: [PATCH 005/196] RDMA/uverbs: Prevent integer overflow issue In the expression "cmd.wqe_size * cmd.wr_count", both variables are u32 values that come from the user so the multiplication can lead to integer wrapping. Then we pass the result to uverbs_request_next_ptr() which also could potentially wrap. The "cmd.sge_count * sizeof(struct ib_uverbs_sge)" multiplication can also overflow on 32bit systems although it's fine on 64bit systems. This patch does two things. First, I've re-arranged the condition in uverbs_request_next_ptr() so that the use controlled variable "len" is on one side of the comparison by itself without any math. Then I've modified all the callers to use size_mul() for the multiplications. Fixes: 67cdb40ca444 ("[IB] uverbs: Implement more commands") Cc: stable@vger.kernel.org Signed-off-by: Dan Carpenter Link: https://patch.msgid.link/b8765ab3-c2da-4611-aae0-ddd6ba173d23@stanley.mountain Signed-off-by: Leon Romanovsky --- drivers/infiniband/core/uverbs_cmd.c | 16 +++++++++------- 1 file changed, 9 insertions(+), 7 deletions(-) diff --git a/drivers/infiniband/core/uverbs_cmd.c b/drivers/infiniband/core/uverbs_cmd.c index 66b02fbf077a..5ad14c39d48c 100644 --- a/drivers/infiniband/core/uverbs_cmd.c +++ b/drivers/infiniband/core/uverbs_cmd.c @@ -161,7 +161,7 @@ static const void __user *uverbs_request_next_ptr(struct uverbs_req_iter *iter, { const void __user *res = iter->cur; - if (iter->cur + len > iter->end) + if (len > iter->end - iter->cur) return (void __force __user *)ERR_PTR(-ENOSPC); iter->cur += len; return res; @@ -2008,11 +2008,13 @@ static int ib_uverbs_post_send(struct uverbs_attr_bundle *attrs) ret = uverbs_request_start(attrs, &iter, &cmd, sizeof(cmd)); if (ret) return ret; - wqes = uverbs_request_next_ptr(&iter, cmd.wqe_size * cmd.wr_count); + wqes = uverbs_request_next_ptr(&iter, size_mul(cmd.wqe_size, + cmd.wr_count)); if (IS_ERR(wqes)) return PTR_ERR(wqes); - sgls = uverbs_request_next_ptr( - &iter, cmd.sge_count * sizeof(struct ib_uverbs_sge)); + sgls = uverbs_request_next_ptr(&iter, + size_mul(cmd.sge_count, + sizeof(struct ib_uverbs_sge))); if (IS_ERR(sgls)) return PTR_ERR(sgls); ret = uverbs_request_finish(&iter); @@ -2198,11 +2200,11 @@ ib_uverbs_unmarshall_recv(struct uverbs_req_iter *iter, u32 wr_count, if (wqe_size < sizeof(struct ib_uverbs_recv_wr)) return ERR_PTR(-EINVAL); - wqes = uverbs_request_next_ptr(iter, wqe_size * wr_count); + wqes = uverbs_request_next_ptr(iter, size_mul(wqe_size, wr_count)); if (IS_ERR(wqes)) return ERR_CAST(wqes); - sgls = uverbs_request_next_ptr( - iter, sge_count * sizeof(struct ib_uverbs_sge)); + sgls = uverbs_request_next_ptr(iter, size_mul(sge_count, + sizeof(struct ib_uverbs_sge))); if (IS_ERR(sgls)) return ERR_CAST(sgls); ret = uverbs_request_finish(iter); From ef7009decc30eb2515a64253791d61b72229c119 Mon Sep 17 00:00:00 2001 From: Ihor Solodrai Date: Thu, 21 Nov 2024 21:40:17 +0000 Subject: [PATCH 006/196] selftests/sched_ext: fix build after renames in sched_ext API The selftests are falining to build on current tip of bpf-next and sched_ext [1]. This has broken BPF CI [2] after merge from upstream. Use appropriate function names in the selftests according to the recent changes in the sched_ext API [3]. [1] https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git/commit/?id=fc39fb56917bb3cb53e99560ca3612a84456ada2 [2] https://github.com/kernel-patches/bpf/actions/runs/11959327258/job/33340923745 [3] https://lore.kernel.org/all/20241109194853.580310-1-tj@kernel.org/ Signed-off-by: Ihor Solodrai Acked-by: Andrea Righi Acked-by: David Vernet Signed-off-by: Tejun Heo --- .../testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c | 2 +- .../selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c | 4 ++-- tools/testing/selftests/sched_ext/dsp_local_on.bpf.c | 2 +- .../selftests/sched_ext/enq_select_cpu_fails.bpf.c | 2 +- tools/testing/selftests/sched_ext/exit.bpf.c | 4 ++-- tools/testing/selftests/sched_ext/maximal.bpf.c | 4 ++-- tools/testing/selftests/sched_ext/select_cpu_dfl.bpf.c | 2 +- .../selftests/sched_ext/select_cpu_dfl_nodispatch.bpf.c | 2 +- .../testing/selftests/sched_ext/select_cpu_dispatch.bpf.c | 2 +- .../selftests/sched_ext/select_cpu_dispatch_bad_dsq.bpf.c | 2 +- .../selftests/sched_ext/select_cpu_dispatch_dbl_dsp.bpf.c | 4 ++-- tools/testing/selftests/sched_ext/select_cpu_vtime.bpf.c | 8 ++++---- 12 files changed, 19 insertions(+), 19 deletions(-) diff --git a/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c b/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c index 37d9bf6fb745..6f4c3f5a1c5d 100644 --- a/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c +++ b/tools/testing/selftests/sched_ext/ddsp_bogus_dsq_fail.bpf.c @@ -20,7 +20,7 @@ s32 BPF_STRUCT_OPS(ddsp_bogus_dsq_fail_select_cpu, struct task_struct *p, * If we dispatch to a bogus DSQ that will fall back to the * builtin global DSQ, we fail gracefully. */ - scx_bpf_dispatch_vtime(p, 0xcafef00d, SCX_SLICE_DFL, + scx_bpf_dsq_insert_vtime(p, 0xcafef00d, SCX_SLICE_DFL, p->scx.dsq_vtime, 0); return cpu; } diff --git a/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c b/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c index dffc97d9cdf1..e4a55027778f 100644 --- a/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c +++ b/tools/testing/selftests/sched_ext/ddsp_vtimelocal_fail.bpf.c @@ -17,8 +17,8 @@ s32 BPF_STRUCT_OPS(ddsp_vtimelocal_fail_select_cpu, struct task_struct *p, if (cpu >= 0) { /* Shouldn't be allowed to vtime dispatch to a builtin DSQ. */ - scx_bpf_dispatch_vtime(p, SCX_DSQ_LOCAL, SCX_SLICE_DFL, - p->scx.dsq_vtime, 0); + scx_bpf_dsq_insert_vtime(p, SCX_DSQ_LOCAL, SCX_SLICE_DFL, + p->scx.dsq_vtime, 0); return cpu; } diff --git a/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c b/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c index 6a7db1502c29..6325bf76f47e 100644 --- a/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c +++ b/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c @@ -45,7 +45,7 @@ void BPF_STRUCT_OPS(dsp_local_on_dispatch, s32 cpu, struct task_struct *prev) target = bpf_get_prandom_u32() % nr_cpus; - scx_bpf_dispatch(p, SCX_DSQ_LOCAL_ON | target, SCX_SLICE_DFL, 0); + scx_bpf_dsq_insert(p, SCX_DSQ_LOCAL_ON | target, SCX_SLICE_DFL, 0); bpf_task_release(p); } diff --git a/tools/testing/selftests/sched_ext/enq_select_cpu_fails.bpf.c b/tools/testing/selftests/sched_ext/enq_select_cpu_fails.bpf.c index 1efb50d61040..a7cf868d5e31 100644 --- a/tools/testing/selftests/sched_ext/enq_select_cpu_fails.bpf.c +++ b/tools/testing/selftests/sched_ext/enq_select_cpu_fails.bpf.c @@ -31,7 +31,7 @@ void BPF_STRUCT_OPS(enq_select_cpu_fails_enqueue, struct task_struct *p, /* Can only call from ops.select_cpu() */ scx_bpf_select_cpu_dfl(p, 0, 0, &found); - scx_bpf_dispatch(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, enq_flags); + scx_bpf_dsq_insert(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, enq_flags); } SEC(".struct_ops.link") diff --git a/tools/testing/selftests/sched_ext/exit.bpf.c b/tools/testing/selftests/sched_ext/exit.bpf.c index d75d4faf07f6..4bc36182d3ff 100644 --- a/tools/testing/selftests/sched_ext/exit.bpf.c +++ b/tools/testing/selftests/sched_ext/exit.bpf.c @@ -33,7 +33,7 @@ void BPF_STRUCT_OPS(exit_enqueue, struct task_struct *p, u64 enq_flags) if (exit_point == EXIT_ENQUEUE) EXIT_CLEANLY(); - scx_bpf_dispatch(p, DSQ_ID, SCX_SLICE_DFL, enq_flags); + scx_bpf_dsq_insert(p, DSQ_ID, SCX_SLICE_DFL, enq_flags); } void BPF_STRUCT_OPS(exit_dispatch, s32 cpu, struct task_struct *p) @@ -41,7 +41,7 @@ void BPF_STRUCT_OPS(exit_dispatch, s32 cpu, struct task_struct *p) if (exit_point == EXIT_DISPATCH) EXIT_CLEANLY(); - scx_bpf_consume(DSQ_ID); + scx_bpf_dsq_move_to_local(DSQ_ID); } void BPF_STRUCT_OPS(exit_enable, struct task_struct *p) diff --git a/tools/testing/selftests/sched_ext/maximal.bpf.c b/tools/testing/selftests/sched_ext/maximal.bpf.c index 4d4cd8d966db..4c005fa71810 100644 --- a/tools/testing/selftests/sched_ext/maximal.bpf.c +++ b/tools/testing/selftests/sched_ext/maximal.bpf.c @@ -20,7 +20,7 @@ s32 BPF_STRUCT_OPS(maximal_select_cpu, struct task_struct *p, s32 prev_cpu, void BPF_STRUCT_OPS(maximal_enqueue, struct task_struct *p, u64 enq_flags) { - scx_bpf_dispatch(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, enq_flags); + scx_bpf_dsq_insert(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, enq_flags); } void BPF_STRUCT_OPS(maximal_dequeue, struct task_struct *p, u64 deq_flags) @@ -28,7 +28,7 @@ void BPF_STRUCT_OPS(maximal_dequeue, struct task_struct *p, u64 deq_flags) void BPF_STRUCT_OPS(maximal_dispatch, s32 cpu, struct task_struct *prev) { - scx_bpf_consume(SCX_DSQ_GLOBAL); + scx_bpf_dsq_move_to_local(SCX_DSQ_GLOBAL); } void BPF_STRUCT_OPS(maximal_runnable, struct task_struct *p, u64 enq_flags) diff --git a/tools/testing/selftests/sched_ext/select_cpu_dfl.bpf.c b/tools/testing/selftests/sched_ext/select_cpu_dfl.bpf.c index f171ac470970..13d0f5be788d 100644 --- a/tools/testing/selftests/sched_ext/select_cpu_dfl.bpf.c +++ b/tools/testing/selftests/sched_ext/select_cpu_dfl.bpf.c @@ -30,7 +30,7 @@ void BPF_STRUCT_OPS(select_cpu_dfl_enqueue, struct task_struct *p, } scx_bpf_put_idle_cpumask(idle_mask); - scx_bpf_dispatch(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, enq_flags); + scx_bpf_dsq_insert(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, enq_flags); } SEC(".struct_ops.link") diff --git a/tools/testing/selftests/sched_ext/select_cpu_dfl_nodispatch.bpf.c b/tools/testing/selftests/sched_ext/select_cpu_dfl_nodispatch.bpf.c index 9efdbb7da928..815f1d5d61ac 100644 --- a/tools/testing/selftests/sched_ext/select_cpu_dfl_nodispatch.bpf.c +++ b/tools/testing/selftests/sched_ext/select_cpu_dfl_nodispatch.bpf.c @@ -67,7 +67,7 @@ void BPF_STRUCT_OPS(select_cpu_dfl_nodispatch_enqueue, struct task_struct *p, saw_local = true; } - scx_bpf_dispatch(p, dsq_id, SCX_SLICE_DFL, enq_flags); + scx_bpf_dsq_insert(p, dsq_id, SCX_SLICE_DFL, enq_flags); } s32 BPF_STRUCT_OPS(select_cpu_dfl_nodispatch_init_task, diff --git a/tools/testing/selftests/sched_ext/select_cpu_dispatch.bpf.c b/tools/testing/selftests/sched_ext/select_cpu_dispatch.bpf.c index 59bfc4f36167..4bb99699e920 100644 --- a/tools/testing/selftests/sched_ext/select_cpu_dispatch.bpf.c +++ b/tools/testing/selftests/sched_ext/select_cpu_dispatch.bpf.c @@ -29,7 +29,7 @@ s32 BPF_STRUCT_OPS(select_cpu_dispatch_select_cpu, struct task_struct *p, cpu = prev_cpu; dispatch: - scx_bpf_dispatch(p, dsq_id, SCX_SLICE_DFL, 0); + scx_bpf_dsq_insert(p, dsq_id, SCX_SLICE_DFL, 0); return cpu; } diff --git a/tools/testing/selftests/sched_ext/select_cpu_dispatch_bad_dsq.bpf.c b/tools/testing/selftests/sched_ext/select_cpu_dispatch_bad_dsq.bpf.c index 3bbd5fcdfb18..2a75de11b2cf 100644 --- a/tools/testing/selftests/sched_ext/select_cpu_dispatch_bad_dsq.bpf.c +++ b/tools/testing/selftests/sched_ext/select_cpu_dispatch_bad_dsq.bpf.c @@ -18,7 +18,7 @@ s32 BPF_STRUCT_OPS(select_cpu_dispatch_bad_dsq_select_cpu, struct task_struct *p s32 prev_cpu, u64 wake_flags) { /* Dispatching to a random DSQ should fail. */ - scx_bpf_dispatch(p, 0xcafef00d, SCX_SLICE_DFL, 0); + scx_bpf_dsq_insert(p, 0xcafef00d, SCX_SLICE_DFL, 0); return prev_cpu; } diff --git a/tools/testing/selftests/sched_ext/select_cpu_dispatch_dbl_dsp.bpf.c b/tools/testing/selftests/sched_ext/select_cpu_dispatch_dbl_dsp.bpf.c index 0fda57fe0ecf..99d075695c97 100644 --- a/tools/testing/selftests/sched_ext/select_cpu_dispatch_dbl_dsp.bpf.c +++ b/tools/testing/selftests/sched_ext/select_cpu_dispatch_dbl_dsp.bpf.c @@ -18,8 +18,8 @@ s32 BPF_STRUCT_OPS(select_cpu_dispatch_dbl_dsp_select_cpu, struct task_struct *p s32 prev_cpu, u64 wake_flags) { /* Dispatching twice in a row is disallowed. */ - scx_bpf_dispatch(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, 0); - scx_bpf_dispatch(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, 0); + scx_bpf_dsq_insert(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, 0); + scx_bpf_dsq_insert(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, 0); return prev_cpu; } diff --git a/tools/testing/selftests/sched_ext/select_cpu_vtime.bpf.c b/tools/testing/selftests/sched_ext/select_cpu_vtime.bpf.c index e6c67bcf5e6e..bfcb96cd4954 100644 --- a/tools/testing/selftests/sched_ext/select_cpu_vtime.bpf.c +++ b/tools/testing/selftests/sched_ext/select_cpu_vtime.bpf.c @@ -2,8 +2,8 @@ /* * A scheduler that validates that enqueue flags are properly stored and * applied at dispatch time when a task is directly dispatched from - * ops.select_cpu(). We validate this by using scx_bpf_dispatch_vtime(), and - * making the test a very basic vtime scheduler. + * ops.select_cpu(). We validate this by using scx_bpf_dsq_insert_vtime(), + * and making the test a very basic vtime scheduler. * * Copyright (c) 2024 Meta Platforms, Inc. and affiliates. * Copyright (c) 2024 David Vernet @@ -47,13 +47,13 @@ s32 BPF_STRUCT_OPS(select_cpu_vtime_select_cpu, struct task_struct *p, cpu = prev_cpu; scx_bpf_test_and_clear_cpu_idle(cpu); ddsp: - scx_bpf_dispatch_vtime(p, VTIME_DSQ, SCX_SLICE_DFL, task_vtime(p), 0); + scx_bpf_dsq_insert_vtime(p, VTIME_DSQ, SCX_SLICE_DFL, task_vtime(p), 0); return cpu; } void BPF_STRUCT_OPS(select_cpu_vtime_dispatch, s32 cpu, struct task_struct *p) { - if (scx_bpf_consume(VTIME_DSQ)) + if (scx_bpf_dsq_move_to_local(VTIME_DSQ)) consumed = true; } From f24d192985cbd6782850fdbb3839039da2f0ee76 Mon Sep 17 00:00:00 2001 From: guanjing Date: Sun, 17 Nov 2024 10:51:29 +0800 Subject: [PATCH 007/196] sched_ext: fix application of sizeof to pointer sizeof when applied to a pointer typed expression gives the size of the pointer. The proper fix in this particular case is to code sizeof(*cpuset) instead of sizeof(cpuset). This issue was detected with the help of Coccinelle. Fixes: 22a920209ab6 ("sched_ext: Implement tickless support") Signed-off-by: guanjing Acked-by: Andrea Righi Signed-off-by: Tejun Heo --- tools/sched_ext/scx_central.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/tools/sched_ext/scx_central.c b/tools/sched_ext/scx_central.c index 21deea320bd7..e938156ed0a0 100644 --- a/tools/sched_ext/scx_central.c +++ b/tools/sched_ext/scx_central.c @@ -97,7 +97,7 @@ int main(int argc, char **argv) SCX_BUG_ON(!cpuset, "Failed to allocate cpuset"); CPU_ZERO(cpuset); CPU_SET(skel->rodata->central_cpu, cpuset); - SCX_BUG_ON(sched_setaffinity(0, sizeof(cpuset), cpuset), + SCX_BUG_ON(sched_setaffinity(0, sizeof(*cpuset), cpuset), "Failed to affinitize to central CPU %d (max %d)", skel->rodata->central_cpu, skel->rodata->nr_cpu_ids - 1); CPU_FREE(cpuset); From e05feab22fd7dabcd6d272c4e2401ec1acdfdb9b Mon Sep 17 00:00:00 2001 From: Patrisious Haddad Date: Tue, 3 Dec 2024 15:45:37 +0200 Subject: [PATCH 008/196] RDMA/mlx5: Enforce same type port association for multiport RoCE Different core device types such as PFs and VFs shouldn't be affiliated together since they have different capabilities, fix that by enforcing type check before doing the affiliation. Fixes: 32f69e4be269 ("{net, IB}/mlx5: Manage port association for multiport RoCE") Reviewed-by: Mark Bloch Signed-off-by: Patrisious Haddad Link: https://patch.msgid.link/88699500f690dff1c1852c1ddb71f8a1cc8b956e.1733233480.git.leonro@nvidia.com Reviewed-by: Mateusz Polchlopek Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/mlx5/main.c | 6 ++++-- include/linux/mlx5/driver.h | 6 ++++++ 2 files changed, 10 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index bc7930d0c564..c2314797afc9 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -3639,7 +3639,8 @@ static int mlx5_ib_init_multiport_master(struct mlx5_ib_dev *dev) list_for_each_entry(mpi, &mlx5_ib_unaffiliated_port_list, list) { if (dev->sys_image_guid == mpi->sys_image_guid && - (mlx5_core_native_port_num(mpi->mdev) - 1) == i) { + (mlx5_core_native_port_num(mpi->mdev) - 1) == i && + mlx5_core_same_coredev_type(dev->mdev, mpi->mdev)) { bound = mlx5_ib_bind_slave_port(dev, mpi); } @@ -4785,7 +4786,8 @@ static int mlx5r_mp_probe(struct auxiliary_device *adev, mutex_lock(&mlx5_ib_multiport_mutex); list_for_each_entry(dev, &mlx5_ib_dev_list, ib_dev_list) { - if (dev->sys_image_guid == mpi->sys_image_guid) + if (dev->sys_image_guid == mpi->sys_image_guid && + mlx5_core_same_coredev_type(dev->mdev, mpi->mdev)) bound = mlx5_ib_bind_slave_port(dev, mpi); if (bound) { diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index fc7e6153b73d..4f9e6f6dbaab 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -1202,6 +1202,12 @@ static inline bool mlx5_core_is_vf(const struct mlx5_core_dev *dev) return dev->coredev_type == MLX5_COREDEV_VF; } +static inline bool mlx5_core_same_coredev_type(const struct mlx5_core_dev *dev1, + const struct mlx5_core_dev *dev2) +{ + return dev1->coredev_type == dev2->coredev_type; +} + static inline bool mlx5_core_is_ecpf(const struct mlx5_core_dev *dev) { return dev->caps.embedded_cpu; From 79d330fbdffd8cee06d8bdf38d82cb62d8363a27 Mon Sep 17 00:00:00 2001 From: Kashyap Desai Date: Wed, 4 Dec 2024 13:24:12 +0530 Subject: [PATCH 009/196] RDMA/bnxt_re: Fix max SGEs for the Work Request Gen P7 supports up to 13 SGEs for now. WQE software structure can hold only 6 now. Since the max send sge is reported as 13, the stack can give requests up to 13 SGEs. This is causing traffic failures and system crashes. Use the define for max SGE supported for variable size. This will work for both static and variable WQEs. Fixes: 227f51743b61 ("RDMA/bnxt_re: Fix the max WQE size for static WQE support") Signed-off-by: Kashyap Desai Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241204075416.478431-2-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_fp.h | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.h b/drivers/infiniband/hw/bnxt_re/qplib_fp.h index ef3424c81345..19e279871f10 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.h @@ -114,7 +114,6 @@ struct bnxt_qplib_sge { u32 size; }; -#define BNXT_QPLIB_QP_MAX_SGL 6 struct bnxt_qplib_swq { u64 wr_id; int next_idx; @@ -154,7 +153,7 @@ struct bnxt_qplib_swqe { #define BNXT_QPLIB_SWQE_FLAGS_UC_FENCE BIT(2) #define BNXT_QPLIB_SWQE_FLAGS_SOLICIT_EVENT BIT(3) #define BNXT_QPLIB_SWQE_FLAGS_INLINE BIT(4) - struct bnxt_qplib_sge sg_list[BNXT_QPLIB_QP_MAX_SGL]; + struct bnxt_qplib_sge sg_list[BNXT_VAR_MAX_SGE]; int num_sge; /* Max inline data is 96 bytes */ u32 inline_len; From 5effcacc8a8f3eb2a9f069d7e81a9ac793598dfb Mon Sep 17 00:00:00 2001 From: Selvin Xavier Date: Wed, 4 Dec 2024 13:24:13 +0530 Subject: [PATCH 010/196] RDMA/bnxt_re: Avoid initializing the software queue for user queues Software Queues to hold the WRs needs to be created for only kernel queues. Avoid allocating the unnecessary memory for user Queues. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Fixes: 159fb4ceacd7 ("RDMA/bnxt_re: introduce a function to allocate swq") Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241204075416.478431-3-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_fp.c | 42 +++++++++++++----------- 1 file changed, 23 insertions(+), 19 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index 9af8aaadc99a..72f35070f671 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -659,13 +659,6 @@ int bnxt_qplib_create_srq(struct bnxt_qplib_res *res, rc = bnxt_qplib_alloc_init_hwq(&srq->hwq, &hwq_attr); if (rc) return rc; - - srq->swq = kcalloc(srq->hwq.max_elements, sizeof(*srq->swq), - GFP_KERNEL); - if (!srq->swq) { - rc = -ENOMEM; - goto fail; - } srq->dbinfo.flags = 0; bnxt_qplib_rcfw_cmd_prep((struct cmdq_base *)&req, CMDQ_BASE_OPCODE_CREATE_SRQ, @@ -694,9 +687,17 @@ int bnxt_qplib_create_srq(struct bnxt_qplib_res *res, spin_lock_init(&srq->lock); srq->start_idx = 0; srq->last_idx = srq->hwq.max_elements - 1; - for (idx = 0; idx < srq->hwq.max_elements; idx++) - srq->swq[idx].next_idx = idx + 1; - srq->swq[srq->last_idx].next_idx = -1; + if (!srq->hwq.is_user) { + srq->swq = kcalloc(srq->hwq.max_elements, sizeof(*srq->swq), + GFP_KERNEL); + if (!srq->swq) { + rc = -ENOMEM; + goto fail; + } + for (idx = 0; idx < srq->hwq.max_elements; idx++) + srq->swq[idx].next_idx = idx + 1; + srq->swq[srq->last_idx].next_idx = -1; + } srq->id = le32_to_cpu(resp.xid); srq->dbinfo.hwq = &srq->hwq; @@ -1042,13 +1043,14 @@ int bnxt_qplib_create_qp(struct bnxt_qplib_res *res, struct bnxt_qplib_qp *qp) if (rc) return rc; - rc = bnxt_qplib_alloc_init_swq(sq); - if (rc) - goto fail_sq; - - if (psn_sz) - bnxt_qplib_init_psn_ptr(qp, psn_sz); + if (!sq->hwq.is_user) { + rc = bnxt_qplib_alloc_init_swq(sq); + if (rc) + goto fail_sq; + if (psn_sz) + bnxt_qplib_init_psn_ptr(qp, psn_sz); + } req.sq_size = cpu_to_le32(bnxt_qplib_set_sq_size(sq, qp->wqe_mode)); pbl = &sq->hwq.pbl[PBL_LVL_0]; req.sq_pbl = cpu_to_le64(pbl->pg_map_arr[0]); @@ -1074,9 +1076,11 @@ int bnxt_qplib_create_qp(struct bnxt_qplib_res *res, struct bnxt_qplib_qp *qp) rc = bnxt_qplib_alloc_init_hwq(&rq->hwq, &hwq_attr); if (rc) goto sq_swq; - rc = bnxt_qplib_alloc_init_swq(rq); - if (rc) - goto fail_rq; + if (!rq->hwq.is_user) { + rc = bnxt_qplib_alloc_init_swq(rq); + if (rc) + goto fail_rq; + } req.rq_size = cpu_to_le32(rq->max_wqe); pbl = &rq->hwq.pbl[PBL_LVL_0]; From 064c22408a73b9e945139b64614c534cbbefb591 Mon Sep 17 00:00:00 2001 From: Kashyap Desai Date: Wed, 4 Dec 2024 13:24:14 +0530 Subject: [PATCH 011/196] RDMA/bnxt_re: Avoid sending the modify QP workaround for latest adapters The workaround to modify the UD QP from RTS to RTS is required only for older adapters. Issuing this for latest adapters can caus some unexpected behavior. Fix it Fixes: 1801d87b3598 ("RDMA/bnxt_re: Support new 5760X P7 devices") Signed-off-by: Kashyap Desai Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241204075416.478431-4-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/ib_verbs.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index 82023394e330..5428a1408cee 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -2824,7 +2824,8 @@ static int bnxt_re_post_send_shadow_qp(struct bnxt_re_dev *rdev, wr = wr->next; } bnxt_qplib_post_send_db(&qp->qplib_qp); - bnxt_ud_qp_hw_stall_workaround(qp); + if (!bnxt_qplib_is_chip_gen_p5_p7(qp->rdev->chip_ctx)) + bnxt_ud_qp_hw_stall_workaround(qp); spin_unlock_irqrestore(&qp->sq_lock, flags); return rc; } @@ -2936,7 +2937,8 @@ int bnxt_re_post_send(struct ib_qp *ib_qp, const struct ib_send_wr *wr, wr = wr->next; } bnxt_qplib_post_send_db(&qp->qplib_qp); - bnxt_ud_qp_hw_stall_workaround(qp); + if (!bnxt_qplib_is_chip_gen_p5_p7(qp->rdev->chip_ctx)) + bnxt_ud_qp_hw_stall_workaround(qp); spin_unlock_irqrestore(&qp->sq_lock, flags); return rc; From d507d29bfde3fee6a74d098a9ac640b8fc1a549b Mon Sep 17 00:00:00 2001 From: Kalesh AP Date: Wed, 4 Dec 2024 13:24:16 +0530 Subject: [PATCH 012/196] RDMA/bnxt_re: Don't fail destroy QP and cleanup debugfs earlier Change bnxt_re_destroy_qp to always return 0 and don't fail in case of error during destroy. In addition, delete debugfs QP to earlier stage. Fixes: d7d54769c042 ("RDMA/bnxt_re: Add debugfs hook in the driver") Reviewed-by: Kashyap Desai Signed-off-by: Kalesh AP Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241204075416.478431-6-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/ib_verbs.c | 15 +++++---------- 1 file changed, 5 insertions(+), 10 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index 5428a1408cee..215074c0860b 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -967,13 +967,13 @@ int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata) unsigned int flags; int rc; + bnxt_re_debug_rem_qpinfo(rdev, qp); + bnxt_qplib_flush_cqn_wq(&qp->qplib_qp); rc = bnxt_qplib_destroy_qp(&rdev->qplib_res, &qp->qplib_qp); - if (rc) { + if (rc) ibdev_err(&rdev->ibdev, "Failed to destroy HW QP"); - return rc; - } if (rdma_is_kernel_res(&qp->ib_qp.res)) { flags = bnxt_re_lock_cqs(qp); @@ -983,11 +983,8 @@ int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata) bnxt_qplib_free_qp_res(&rdev->qplib_res, &qp->qplib_qp); - if (ib_qp->qp_type == IB_QPT_GSI && rdev->gsi_ctx.gsi_sqp) { - rc = bnxt_re_destroy_gsi_sqp(qp); - if (rc) - return rc; - } + if (ib_qp->qp_type == IB_QPT_GSI && rdev->gsi_ctx.gsi_sqp) + bnxt_re_destroy_gsi_sqp(qp); mutex_lock(&rdev->qp_lock); list_del(&qp->list); @@ -998,8 +995,6 @@ int bnxt_re_destroy_qp(struct ib_qp *ib_qp, struct ib_udata *udata) else if (qp->qplib_qp.type == CMDQ_CREATE_QP_TYPE_UD) atomic_dec(&rdev->stats.res.ud_qp_count); - bnxt_re_debug_rem_qpinfo(rdev, qp); - ib_umem_release(qp->rumem); ib_umem_release(qp->sumem); From a4048c83fd87c65657a4acb17d639092d4b6133d Mon Sep 17 00:00:00 2001 From: Anumula Murali Mohan Reddy Date: Tue, 3 Dec 2024 19:30:53 +0530 Subject: [PATCH 013/196] RDMA/core: Fix ENODEV error for iWARP test over vlan If traffic is over vlan, cma_validate_port() fails to match net_device ifindex with bound_if_index and results in ENODEV error. As iWARP gid table is static, it contains entry corresponding to only one net device which is either real netdev or vlan netdev for cases like siw attached to a vlan interface. This patch fixes the issue by assigning bound_if_index with net device index, if real net device obtained from bound if index matches with net device retrieved from gid table Fixes: f8ef1be816bf ("RDMA/cma: Avoid GID lookups on iWARP devices") Link: https://lore.kernel.org/all/ZzNgdrjo1kSCGbRz@chelsio.com/ Signed-off-by: Anumula Murali Mohan Reddy Signed-off-by: Potnuri Bharat Teja Link: https://patch.msgid.link/20241203140052.3985-1-anumula@chelsio.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/core/cma.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 64ace0b968f0..91db10515d74 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -690,6 +690,7 @@ cma_validate_port(struct ib_device *device, u32 port, int bound_if_index = dev_addr->bound_dev_if; int dev_type = dev_addr->dev_type; struct net_device *ndev = NULL; + struct net_device *pdev = NULL; if (!rdma_dev_access_netns(device, id_priv->id.route.addr.dev_addr.net)) goto out; @@ -714,6 +715,21 @@ cma_validate_port(struct ib_device *device, u32 port, rcu_read_lock(); ndev = rcu_dereference(sgid_attr->ndev); + if (ndev->ifindex != bound_if_index) { + pdev = dev_get_by_index_rcu(dev_addr->net, bound_if_index); + if (pdev) { + if (is_vlan_dev(pdev)) { + pdev = vlan_dev_real_dev(pdev); + if (ndev->ifindex == pdev->ifindex) + bound_if_index = pdev->ifindex; + } + if (is_vlan_dev(ndev)) { + pdev = vlan_dev_real_dev(ndev); + if (bound_if_index == pdev->ifindex) + bound_if_index = ndev->ifindex; + } + } + } if (!net_eq(dev_net(ndev), dev_addr->net) || ndev->ifindex != bound_if_index) { rdma_put_gid_attr(sgid_attr); From b8f614207b0d5e4abd6df8d5cb3cc11f009d1d93 Mon Sep 17 00:00:00 2001 From: David Vernet Date: Mon, 9 Dec 2024 09:29:24 -0600 Subject: [PATCH 014/196] scx: Fix maximal BPF selftest prog maximal.bpf.c is still dispatching to and consuming from SCX_DSQ_GLOBAL. Let's have it use its own DSQ to avoid any runtime errors. Signed-off-by: David Vernet Tested-by: Andrea Righi Signed-off-by: Tejun Heo --- tools/testing/selftests/sched_ext/maximal.bpf.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/sched_ext/maximal.bpf.c b/tools/testing/selftests/sched_ext/maximal.bpf.c index 4c005fa71810..430f5e13bf55 100644 --- a/tools/testing/selftests/sched_ext/maximal.bpf.c +++ b/tools/testing/selftests/sched_ext/maximal.bpf.c @@ -12,6 +12,8 @@ char _license[] SEC("license") = "GPL"; +#define DSQ_ID 0 + s32 BPF_STRUCT_OPS(maximal_select_cpu, struct task_struct *p, s32 prev_cpu, u64 wake_flags) { @@ -20,7 +22,7 @@ s32 BPF_STRUCT_OPS(maximal_select_cpu, struct task_struct *p, s32 prev_cpu, void BPF_STRUCT_OPS(maximal_enqueue, struct task_struct *p, u64 enq_flags) { - scx_bpf_dsq_insert(p, SCX_DSQ_GLOBAL, SCX_SLICE_DFL, enq_flags); + scx_bpf_dsq_insert(p, DSQ_ID, SCX_SLICE_DFL, enq_flags); } void BPF_STRUCT_OPS(maximal_dequeue, struct task_struct *p, u64 deq_flags) @@ -28,7 +30,7 @@ void BPF_STRUCT_OPS(maximal_dequeue, struct task_struct *p, u64 deq_flags) void BPF_STRUCT_OPS(maximal_dispatch, s32 cpu, struct task_struct *prev) { - scx_bpf_dsq_move_to_local(SCX_DSQ_GLOBAL); + scx_bpf_dsq_move_to_local(DSQ_ID); } void BPF_STRUCT_OPS(maximal_runnable, struct task_struct *p, u64 enq_flags) @@ -123,7 +125,7 @@ void BPF_STRUCT_OPS(maximal_cgroup_set_weight, struct cgroup *cgrp, u32 weight) s32 BPF_STRUCT_OPS_SLEEPABLE(maximal_init) { - return 0; + return scx_bpf_create_dsq(DSQ_ID, -1); } void BPF_STRUCT_OPS(maximal_exit, struct scx_exit_info *info) From 2d2f25405a87cfa270ea7b5bb03a612c1a16020a Mon Sep 17 00:00:00 2001 From: Andrea Righi Date: Mon, 9 Dec 2024 18:45:01 +0100 Subject: [PATCH 015/196] MAINTAINERS: add self as reviewer for sched_ext Add myself as a reviewer for sched_ext, as I am actively working on this project and would like to help review relevant patches and address any related kernel issues. Signed-off-by: Andrea Righi Signed-off-by: Tejun Heo --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 1e930c7a58b1..41bae8792a77 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -20902,6 +20902,7 @@ F: kernel/sched/ SCHEDULER - SCHED_EXT R: Tejun Heo R: David Vernet +R: Andrea Righi L: linux-kernel@vger.kernel.org S: Maintained W: https://github.com/sched-ext/scx From 48ca421268735c60ea0d4c2e19610b224d5c8656 Mon Sep 17 00:00:00 2001 From: Changwoo Min Date: Wed, 11 Dec 2024 17:39:45 +0900 Subject: [PATCH 016/196] MAINTAINERS: add me as reviewer for sched_ext Add me as a reviewer for sched_ext. I have been actively working on the project and would like to help review patches and address related kernel issues/features. Signed-off-by: Changwoo Min Signed-off-by: Tejun Heo --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 41bae8792a77..cbfa4e97a3ff 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -20903,6 +20903,7 @@ SCHEDULER - SCHED_EXT R: Tejun Heo R: David Vernet R: Andrea Righi +R: Changwoo Min L: linux-kernel@vger.kernel.org S: Maintained W: https://github.com/sched-ext/scx From 18b2093f4598d8ee67a8153badc93f0fa7686b8a Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Wed, 11 Dec 2024 11:01:51 -1000 Subject: [PATCH 017/196] sched_ext: Fix invalid irq restore in scx_ops_bypass() While adding outer irqsave/restore locking, 0e7ffff1b811 ("scx: Fix raciness in scx_ops_bypass()") forgot to convert an inner rq_unlock_irqrestore() to rq_unlock() which could re-enable IRQ prematurely leading to the following warning: raw_local_irq_restore() called with IRQs enabled WARNING: CPU: 1 PID: 96 at kernel/locking/irqflag-debug.c:10 warn_bogus_irq_restore+0x30/0x40 ... Sched_ext: create_dsq (enabling) pstate: 60400005 (nZCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) pc : warn_bogus_irq_restore+0x30/0x40 lr : warn_bogus_irq_restore+0x30/0x40 ... Call trace: warn_bogus_irq_restore+0x30/0x40 (P) warn_bogus_irq_restore+0x30/0x40 (L) scx_ops_bypass+0x224/0x3b8 scx_ops_enable.isra.0+0x2c8/0xaa8 bpf_scx_reg+0x18/0x30 ... irq event stamp: 33739 hardirqs last enabled at (33739): [] scx_ops_bypass+0x174/0x3b8 hardirqs last disabled at (33738): [] _raw_spin_lock_irqsave+0xb4/0xd8 Drop the stray _irqrestore(). Signed-off-by: Tejun Heo Reported-by: Ihor Solodrai Link: http://lkml.kernel.org/r/qC39k3UsonrBYD_SmuxHnZIQLsuuccoCrkiqb_BT7DvH945A1_LZwE4g-5Pu9FcCtqZt4lY1HhIPi0homRuNWxkgo1rgP3bkxa0donw8kV4=@pm.me Fixes: 0e7ffff1b811 ("scx: Fix raciness in scx_ops_bypass()") Cc: stable@vger.kernel.org # v6.12 --- kernel/sched/ext.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 7fff1d045477..98519e6d0dcd 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -4763,7 +4763,7 @@ static void scx_ops_bypass(bool bypass) * sees scx_rq_bypassing() before moving tasks to SCX. */ if (!scx_enabled()) { - rq_unlock_irqrestore(rq, &rf); + rq_unlock(rq, &rf); continue; } From ebefac5647968679f6ef5803e5d35a71997d20fa Mon Sep 17 00:00:00 2001 From: Robert Beckett Date: Tue, 12 Nov 2024 19:50:00 +0000 Subject: [PATCH 018/196] nvme-pci: 512 byte aligned dma pool segment quirk MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We initially introduced a quick fix limiting the queue depth to 1 as experimentation showed that it fixed data corruption on 64GB steamdecks. Further experimentation revealed corruption only happens when the last PRP data element aligns to the end of the page boundary. The device appears to treat this as a PRP chain to a new list instead of the data element that it actually is. This implementation is in violation of the spec. Encountering this errata with the Linux driver requires the host request a 128k transfer and coincidently be handed the last small pool dma buffer within a page. The QD1 quirk effectly works around this because the last data PRP always was at a 248 byte offset from the page start, so it never appeared at the end of the page, but comes at the expense of throttling IO and wasting the remainder of the PRP page beyond 256 bytes. Also to note, the MDTS on these devices is small enough that the "large" prp pool can hold enough PRP elements to never reach the end, so that pool is not a problem either. Introduce a new quirk to ensure the small pool is always aligned such that the last PRP element can't appear a the end of the page. This comes at the expense of wasting 256 bytes per small pool page allocated. Link: https://lore.kernel.org/linux-nvme/20241113043151.GA20077@lst.de/T/#u Fixes: 83bdfcbdbe5d ("nvme-pci: qdepth 1 quirk") Cc: Paweł Anikiel Signed-off-by: Robert Beckett Signed-off-by: Keith Busch --- drivers/nvme/host/nvme.h | 5 +++++ drivers/nvme/host/pci.c | 9 +++++++-- 2 files changed, 12 insertions(+), 2 deletions(-) diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h index 611b02c8a8b3..c4bb8dfe1a45 100644 --- a/drivers/nvme/host/nvme.h +++ b/drivers/nvme/host/nvme.h @@ -173,6 +173,11 @@ enum nvme_quirks { * MSI (but not MSI-X) interrupts are broken and never fire. */ NVME_QUIRK_BROKEN_MSI = (1 << 21), + + /* + * Align dma pool segment size to 512 bytes + */ + NVME_QUIRK_DMAPOOL_ALIGN_512 = (1 << 22), }; /* diff --git a/drivers/nvme/host/pci.c b/drivers/nvme/host/pci.c index 1a5ba80f1811..e2634f437f33 100644 --- a/drivers/nvme/host/pci.c +++ b/drivers/nvme/host/pci.c @@ -2834,15 +2834,20 @@ static int nvme_disable_prepare_reset(struct nvme_dev *dev, bool shutdown) static int nvme_setup_prp_pools(struct nvme_dev *dev) { + size_t small_align = 256; + dev->prp_page_pool = dma_pool_create("prp list page", dev->dev, NVME_CTRL_PAGE_SIZE, NVME_CTRL_PAGE_SIZE, 0); if (!dev->prp_page_pool) return -ENOMEM; + if (dev->ctrl.quirks & NVME_QUIRK_DMAPOOL_ALIGN_512) + small_align = 512; + /* Optimisation for I/Os between 4k and 128k */ dev->prp_small_pool = dma_pool_create("prp list 256", dev->dev, - 256, 256, 0); + 256, small_align, 0); if (!dev->prp_small_pool) { dma_pool_destroy(dev->prp_page_pool); return -ENOMEM; @@ -3607,7 +3612,7 @@ static const struct pci_device_id nvme_id_table[] = { { PCI_VDEVICE(REDHAT, 0x0010), /* Qemu emulated controller */ .driver_data = NVME_QUIRK_BOGUS_NID, }, { PCI_DEVICE(0x1217, 0x8760), /* O2 Micro 64GB Steam Deck */ - .driver_data = NVME_QUIRK_QDEPTH_ONE }, + .driver_data = NVME_QUIRK_DMAPOOL_ALIGN_512, }, { PCI_DEVICE(0x126f, 0x2262), /* Silicon Motion generic */ .driver_data = NVME_QUIRK_NO_DEEPEST_PS | NVME_QUIRK_BOGUS_NID, }, From 0e8c52091633b354b12d0c29a27a22077584c111 Mon Sep 17 00:00:00 2001 From: Emmanuel Grumbach Date: Thu, 12 Dec 2024 13:29:42 +0200 Subject: [PATCH 019/196] wifi: iwlwifi: fix CRF name for Bz We had BE201 hard coded. Look at the RF_ID and decide based on its value. Fixes: 6795a37161fb ("wifi: iwlwifi: Print a specific device name.") Signed-off-by: Emmanuel Grumbach Signed-off-by: Miri Korenblit Link: https://patch.msgid.link/20241212132940.b9eebda1ca60.I36791a134ed5e538e059418eb6520761da97b44c@changeid Signed-off-by: Johannes Berg --- drivers/net/wireless/intel/iwlwifi/cfg/bz.c | 1 + .../net/wireless/intel/iwlwifi/iwl-config.h | 1 + drivers/net/wireless/intel/iwlwifi/pcie/drv.c | 41 +++++++++++++++++-- 3 files changed, 40 insertions(+), 3 deletions(-) diff --git a/drivers/net/wireless/intel/iwlwifi/cfg/bz.c b/drivers/net/wireless/intel/iwlwifi/cfg/bz.c index cd1fe8490ae5..1c43f283ac4a 100644 --- a/drivers/net/wireless/intel/iwlwifi/cfg/bz.c +++ b/drivers/net/wireless/intel/iwlwifi/cfg/bz.c @@ -161,6 +161,7 @@ const struct iwl_cfg_trans_params iwl_gl_trans_cfg = { const char iwl_bz_name[] = "Intel(R) TBD Bz device"; const char iwl_fm_name[] = "Intel(R) Wi-Fi 7 BE201 320MHz"; +const char iwl_wh_name[] = "Intel(R) Wi-Fi 7 BE211 320MHz"; const char iwl_gl_name[] = "Intel(R) Wi-Fi 7 BE200 320MHz"; const char iwl_mtp_name[] = "Intel(R) Wi-Fi 7 BE202 160MHz"; diff --git a/drivers/net/wireless/intel/iwlwifi/iwl-config.h b/drivers/net/wireless/intel/iwlwifi/iwl-config.h index 34c91deca57b..17721bb47e25 100644 --- a/drivers/net/wireless/intel/iwlwifi/iwl-config.h +++ b/drivers/net/wireless/intel/iwlwifi/iwl-config.h @@ -545,6 +545,7 @@ extern const char iwl_ax231_name[]; extern const char iwl_ax411_name[]; extern const char iwl_bz_name[]; extern const char iwl_fm_name[]; +extern const char iwl_wh_name[]; extern const char iwl_gl_name[]; extern const char iwl_mtp_name[]; extern const char iwl_sc_name[]; diff --git a/drivers/net/wireless/intel/iwlwifi/pcie/drv.c b/drivers/net/wireless/intel/iwlwifi/pcie/drv.c index 805fb249a0c6..8fb2aa282242 100644 --- a/drivers/net/wireless/intel/iwlwifi/pcie/drv.c +++ b/drivers/net/wireless/intel/iwlwifi/pcie/drv.c @@ -1106,18 +1106,53 @@ VISIBLE_IF_IWLWIFI_KUNIT const struct iwl_dev_info iwl_dev_info_table[] = { iwlax210_2ax_cfg_so_jf_b0, iwl9462_name), /* Bz */ -/* FIXME: need to change the naming according to the actual CRF */ _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_MAC_TYPE_BZ, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_HR2, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, + iwl_cfg_bz, iwl_ax201_name), + + _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_MAC_TYPE_BZ, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_GF, IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, + iwl_cfg_bz, iwl_ax211_name), + + _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_MAC_TYPE_BZ, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_FM, IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, + iwl_cfg_bz, iwl_fm_name), + + _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_MAC_TYPE_BZ, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_WH, IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, + iwl_cfg_bz, iwl_wh_name), + + _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_MAC_TYPE_BZ_W, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_HR2, IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, + iwl_cfg_bz, iwl_ax201_name), + + _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_MAC_TYPE_BZ_W, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_GF, IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, + iwl_cfg_bz, iwl_ax211_name), + + _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, + IWL_CFG_MAC_TYPE_BZ_W, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_FM, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, iwl_cfg_bz, iwl_fm_name), _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_MAC_TYPE_BZ_W, IWL_CFG_ANY, + IWL_CFG_RF_TYPE_WH, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, - IWL_CFG_ANY, IWL_CFG_ANY, IWL_CFG_ANY, - iwl_cfg_bz, iwl_fm_name), + iwl_cfg_bz, iwl_wh_name), /* Ga (Gl) */ _IWL_DEV_INFO(IWL_CFG_ANY, IWL_CFG_ANY, From b83accfec0811421df065f820e73ca8df7f6439a Mon Sep 17 00:00:00 2001 From: Jeff Johnson Date: Thu, 12 Dec 2024 08:27:05 -0800 Subject: [PATCH 020/196] MAINTAINERS: wifi: ath: add Jeff Johnson as maintainer The "ATHEROS ATH GENERIC UTILITIES" entry shares the same git tree as the ATH10K, ATH11K, and ATH12K entries which I already maintain, so add me to that entry as well. Signed-off-by: Jeff Johnson Signed-off-by: Kalle Valo Link: https://patch.msgid.link/20241212-ath-maintainer-v1-1-7ea5e86780a8@kernel.org --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index e6e71b05710b..158740a571b3 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3608,6 +3608,7 @@ F: drivers/phy/qualcomm/phy-ath79-usb.c ATHEROS ATH GENERIC UTILITIES M: Kalle Valo +M: Jeff Johnson L: linux-wireless@vger.kernel.org S: Supported F: drivers/net/wireless/ath/* From 38651476e46e088598354510502c383e932e2297 Mon Sep 17 00:00:00 2001 From: Kalesh AP Date: Wed, 11 Dec 2024 14:09:27 +0530 Subject: [PATCH 021/196] RDMA/bnxt_re: Fix the check for 9060 condition The check for 9060 condition should only be made for legacy chips. Fixes: 9152e0b722b2 ("RDMA/bnxt_re: HW workarounds for handling specific conditions") Reviewed-by: Kashyap Desai Signed-off-by: Kalesh AP Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241211083931.968831-2-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_fp.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index 72f35070f671..093bfb748cdf 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -2669,10 +2669,12 @@ static int bnxt_qplib_cq_process_req(struct bnxt_qplib_cq *cq, bnxt_qplib_add_flush_qp(qp); } else { /* Before we complete, do WA 9060 */ - if (do_wa9060(qp, cq, cq_cons, sq->swq_last, - cqe_sq_cons)) { - *lib_qp = qp; - goto out; + if (!bnxt_qplib_is_chip_gen_p5_p7(qp->cctx)) { + if (do_wa9060(qp, cq, cq_cons, sq->swq_last, + cqe_sq_cons)) { + *lib_qp = qp; + goto out; + } } if (swq->flags & SQ_SEND_FLAGS_SIGNAL_COMP) { cqe->status = CQ_REQ_STATUS_OK; From 798653a0ee30d3cd495099282751c0f248614ae7 Mon Sep 17 00:00:00 2001 From: Saravanan Vajravel Date: Wed, 11 Dec 2024 14:09:28 +0530 Subject: [PATCH 022/196] RDMA/bnxt_re: Add check for path mtu in modify_qp When RDMA app configures path MTU, add a check in modify_qp verb to make sure that it doesn't go beyond interface MTU. If this check fails, driver will fail the modify_qp verb. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Reviewed-by: Kalesh AP Signed-off-by: Saravanan Vajravel Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241211083931.968831-3-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/ib_verbs.c | 26 +++++++++++++----------- 1 file changed, 14 insertions(+), 12 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index 215074c0860b..a609e1635a3d 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -2162,18 +2162,20 @@ int bnxt_re_modify_qp(struct ib_qp *ib_qp, struct ib_qp_attr *qp_attr, } } - if (qp_attr_mask & IB_QP_PATH_MTU) { - qp->qplib_qp.modify_flags |= - CMDQ_MODIFY_QP_MODIFY_MASK_PATH_MTU; - qp->qplib_qp.path_mtu = __from_ib_mtu(qp_attr->path_mtu); - qp->qplib_qp.mtu = ib_mtu_enum_to_int(qp_attr->path_mtu); - } else if (qp_attr->qp_state == IB_QPS_RTR) { - qp->qplib_qp.modify_flags |= - CMDQ_MODIFY_QP_MODIFY_MASK_PATH_MTU; - qp->qplib_qp.path_mtu = - __from_ib_mtu(iboe_get_mtu(rdev->netdev->mtu)); - qp->qplib_qp.mtu = - ib_mtu_enum_to_int(iboe_get_mtu(rdev->netdev->mtu)); + if (qp_attr->qp_state == IB_QPS_RTR) { + enum ib_mtu qpmtu; + + qpmtu = iboe_get_mtu(rdev->netdev->mtu); + if (qp_attr_mask & IB_QP_PATH_MTU) { + if (ib_mtu_enum_to_int(qp_attr->path_mtu) > + ib_mtu_enum_to_int(qpmtu)) + return -EINVAL; + qpmtu = qp_attr->path_mtu; + } + + qp->qplib_qp.modify_flags |= CMDQ_MODIFY_QP_MODIFY_MASK_PATH_MTU; + qp->qplib_qp.path_mtu = __from_ib_mtu(qpmtu); + qp->qplib_qp.mtu = ib_mtu_enum_to_int(qpmtu); } if (qp_attr_mask & IB_QP_TIMEOUT) { From da2132e683954e7ddda3cd674e866a847b7389eb Mon Sep 17 00:00:00 2001 From: Damodharam Ammepalli Date: Wed, 11 Dec 2024 14:09:29 +0530 Subject: [PATCH 023/196] RDMA/bnxt_re: Fix setting mandatory attributes for modify_qp Firmware expects "min_rnr_timer" as a mandatory attribute in MODIFY_QP command during the RTR-RTS transition. This needs to be enforced by the driver which is missing while setting bnxt_set_mandatory_attributes that sends these flags as part of modify_qp optimization. Fixes: 82c32d219272 ("RDMA/bnxt_re: Add support for optimized modify QP") Reviewed-by: Rukhsana Ansari Reviewed-by: Kalesh AP Signed-off-by: Damodharam Ammepalli Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241211083931.968831-4-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_fp.c | 13 +++++++++++-- drivers/infiniband/hw/bnxt_re/qplib_res.h | 5 +++++ drivers/infiniband/hw/bnxt_re/roce_hsi.h | 1 + 3 files changed, 17 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index 093bfb748cdf..5169804e6f12 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -1285,7 +1285,8 @@ static void __filter_modify_flags(struct bnxt_qplib_qp *qp) } } -static void bnxt_set_mandatory_attributes(struct bnxt_qplib_qp *qp, +static void bnxt_set_mandatory_attributes(struct bnxt_qplib_res *res, + struct bnxt_qplib_qp *qp, struct cmdq_modify_qp *req) { u32 mandatory_flags = 0; @@ -1300,6 +1301,14 @@ static void bnxt_set_mandatory_attributes(struct bnxt_qplib_qp *qp, mandatory_flags |= CMDQ_MODIFY_QP_MODIFY_MASK_PKEY; } + if (_is_min_rnr_in_rtr_rts_mandatory(res->dattr->dev_cap_flags2) && + (qp->cur_qp_state == CMDQ_MODIFY_QP_NEW_STATE_RTR && + qp->state == CMDQ_MODIFY_QP_NEW_STATE_RTS)) { + if (qp->type == CMDQ_MODIFY_QP_QP_TYPE_RC) + mandatory_flags |= + CMDQ_MODIFY_QP_MODIFY_MASK_MIN_RNR_TIMER; + } + if (qp->type == CMDQ_MODIFY_QP_QP_TYPE_UD || qp->type == CMDQ_MODIFY_QP_QP_TYPE_GSI) mandatory_flags |= CMDQ_MODIFY_QP_MODIFY_MASK_QKEY; @@ -1340,7 +1349,7 @@ int bnxt_qplib_modify_qp(struct bnxt_qplib_res *res, struct bnxt_qplib_qp *qp) /* Set mandatory attributes for INIT -> RTR and RTR -> RTS transition */ if (_is_optimize_modify_qp_supported(res->dattr->dev_cap_flags2) && is_optimized_state_transition(qp)) - bnxt_set_mandatory_attributes(qp, &req); + bnxt_set_mandatory_attributes(res, qp, &req); } bmask = qp->modify_flags; req.modify_mask = cpu_to_le32(qp->modify_flags); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_res.h b/drivers/infiniband/hw/bnxt_re/qplib_res.h index 21fb148713a6..cbfc49a1a56d 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_res.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_res.h @@ -584,6 +584,11 @@ static inline bool _is_optimize_modify_qp_supported(u16 dev_cap_ext_flags2) return dev_cap_ext_flags2 & CREQ_QUERY_FUNC_RESP_SB_OPTIMIZE_MODIFY_QP_SUPPORTED; } +static inline bool _is_min_rnr_in_rtr_rts_mandatory(u16 dev_cap_ext_flags2) +{ + return !!(dev_cap_ext_flags2 & CREQ_QUERY_FUNC_RESP_SB_MIN_RNR_RTR_RTS_OPT_SUPPORTED); +} + static inline bool _is_cq_coalescing_supported(u16 dev_cap_ext_flags2) { return dev_cap_ext_flags2 & CREQ_QUERY_FUNC_RESP_SB_CQ_COALESCING_SUPPORTED; diff --git a/drivers/infiniband/hw/bnxt_re/roce_hsi.h b/drivers/infiniband/hw/bnxt_re/roce_hsi.h index a98fc9c2313e..0ee60fdc18b3 100644 --- a/drivers/infiniband/hw/bnxt_re/roce_hsi.h +++ b/drivers/infiniband/hw/bnxt_re/roce_hsi.h @@ -2215,6 +2215,7 @@ struct creq_query_func_resp_sb { #define CREQ_QUERY_FUNC_RESP_SB_REQ_RETRANSMISSION_SUPPORT_IQM_MSN_TABLE (0x2UL << 4) #define CREQ_QUERY_FUNC_RESP_SB_REQ_RETRANSMISSION_SUPPORT_LAST \ CREQ_QUERY_FUNC_RESP_SB_REQ_RETRANSMISSION_SUPPORT_IQM_MSN_TABLE + #define CREQ_QUERY_FUNC_RESP_SB_MIN_RNR_RTR_RTS_OPT_SUPPORTED 0x1000UL __le16 max_xp_qp_size; __le16 create_qp_batch_size; __le16 destroy_qp_batch_size; From 34db8ec931b84d1426423f263b1927539e73b397 Mon Sep 17 00:00:00 2001 From: Hongguang Gao Date: Wed, 11 Dec 2024 14:09:30 +0530 Subject: [PATCH 024/196] RDMA/bnxt_re: Fix to export port num to ib_query_qp Current driver implementation doesn't populate the port_num field in query_qp. Adding the code to convert internal firmware port id to ibv defined port number and export it. Reviewed-by: Saravanan Vajravel Reviewed-by: Kalesh AP Signed-off-by: Hongguang Gao Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241211083931.968831-5-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/ib_verbs.c | 1 + drivers/infiniband/hw/bnxt_re/ib_verbs.h | 4 ++++ drivers/infiniband/hw/bnxt_re/qplib_fp.c | 1 + drivers/infiniband/hw/bnxt_re/qplib_fp.h | 1 + 4 files changed, 7 insertions(+) diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index a609e1635a3d..bcb7cfc63d09 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -2325,6 +2325,7 @@ int bnxt_re_query_qp(struct ib_qp *ib_qp, struct ib_qp_attr *qp_attr, qp_attr->retry_cnt = qplib_qp->retry_cnt; qp_attr->rnr_retry = qplib_qp->rnr_retry; qp_attr->min_rnr_timer = qplib_qp->min_rnr_timer; + qp_attr->port_num = __to_ib_port_num(qplib_qp->port_id); qp_attr->rq_psn = qplib_qp->rq.psn; qp_attr->max_rd_atomic = qplib_qp->max_rd_atomic; qp_attr->sq_psn = qplib_qp->sq.psn; diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.h b/drivers/infiniband/hw/bnxt_re/ib_verbs.h index ac59f1d73b15..fbb16a411d6a 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.h +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.h @@ -268,6 +268,10 @@ void bnxt_re_dealloc_ucontext(struct ib_ucontext *context); int bnxt_re_mmap(struct ib_ucontext *context, struct vm_area_struct *vma); void bnxt_re_mmap_free(struct rdma_user_mmap_entry *rdma_entry); +static inline u32 __to_ib_port_num(u16 port_id) +{ + return (u32)port_id + 1; +} unsigned long bnxt_re_lock_cqs(struct bnxt_re_qp *qp); void bnxt_re_unlock_cqs(struct bnxt_re_qp *qp, unsigned long flags); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index 5169804e6f12..d8a2a929bbe3 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -1532,6 +1532,7 @@ int bnxt_qplib_query_qp(struct bnxt_qplib_res *res, struct bnxt_qplib_qp *qp) qp->dest_qpn = le32_to_cpu(sb->dest_qp_id); memcpy(qp->smac, sb->src_mac, 6); qp->vlan_id = le16_to_cpu(sb->vlan_pcp_vlan_dei_vlan_id); + qp->port_id = le16_to_cpu(sb->port_id); bail: dma_free_coherent(&rcfw->pdev->dev, sbuf.size, sbuf.sb, sbuf.dma_addr); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.h b/drivers/infiniband/hw/bnxt_re/qplib_fp.h index 19e279871f10..0660101b5310 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.h +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.h @@ -298,6 +298,7 @@ struct bnxt_qplib_qp { u32 dest_qpn; u8 smac[6]; u16 vlan_id; + u16 port_id; u8 nw_type; struct bnxt_qplib_ah ah; From 7179fe0074a3c962e43a9e51169304c4911989ed Mon Sep 17 00:00:00 2001 From: Kalesh AP Date: Wed, 11 Dec 2024 14:09:31 +0530 Subject: [PATCH 025/196] RDMA/bnxt_re: Fix reporting hw_ver in query_device Driver currently populates subsystem_device id in the "hw_ver" field of ib_attr structure in query_device. Updated to populate PCI revision ID. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Reviewed-by: Preethi G Signed-off-by: Kalesh AP Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241211083931.968831-6-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/ib_verbs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index bcb7cfc63d09..e3d26bd6de05 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -199,7 +199,7 @@ int bnxt_re_query_device(struct ib_device *ibdev, ib_attr->vendor_id = rdev->en_dev->pdev->vendor; ib_attr->vendor_part_id = rdev->en_dev->pdev->device; - ib_attr->hw_ver = rdev->en_dev->pdev->subsystem_device; + ib_attr->hw_ver = rdev->en_dev->pdev->revision; ib_attr->max_qp = dev_attr->max_qp; ib_attr->max_qp_wr = dev_attr->max_qp_wqes; ib_attr->device_cap_flags = From 0efbca0fec7d1665e19fa008ba95daab71f76f4d Mon Sep 17 00:00:00 2001 From: Thorsten Blum Date: Mon, 9 Dec 2024 12:06:39 +0100 Subject: [PATCH 026/196] nios2: Use str_yes_no() helper in show_cpuinfo() Remove hard-coded strings by using the str_yes_no() helper function. Signed-off-by: Thorsten Blum Signed-off-by: Dinh Nguyen --- arch/nios2/kernel/cpuinfo.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/arch/nios2/kernel/cpuinfo.c b/arch/nios2/kernel/cpuinfo.c index 338849c430a5..7b1e8f9128e9 100644 --- a/arch/nios2/kernel/cpuinfo.c +++ b/arch/nios2/kernel/cpuinfo.c @@ -143,11 +143,11 @@ static int show_cpuinfo(struct seq_file *m, void *v) " DIV:\t\t%s\n" " BMX:\t\t%s\n" " CDX:\t\t%s\n", - cpuinfo.has_mul ? "yes" : "no", - cpuinfo.has_mulx ? "yes" : "no", - cpuinfo.has_div ? "yes" : "no", - cpuinfo.has_bmx ? "yes" : "no", - cpuinfo.has_cdx ? "yes" : "no"); + str_yes_no(cpuinfo.has_mul), + str_yes_no(cpuinfo.has_mulx), + str_yes_no(cpuinfo.has_div), + str_yes_no(cpuinfo.has_bmx), + str_yes_no(cpuinfo.has_cdx)); seq_printf(m, "Icache:\t\t%ukB, line length: %u\n", From a37eecb705f33726f1fb7cd2a67e514a15dfe693 Mon Sep 17 00:00:00 2001 From: Evgenii Shatokhin Date: Mon, 9 Dec 2024 10:46:59 +0300 Subject: [PATCH 027/196] pinctrl: mcp23s08: Fix sleeping in atomic context due to regmap locking If a device uses MCP23xxx IO expander to receive IRQs, the following bug can happen: BUG: sleeping function called from invalid context at kernel/locking/mutex.c:283 in_atomic(): 1, irqs_disabled(): 1, non_block: 0, ... preempt_count: 1, expected: 0 ... Call Trace: ... __might_resched+0x104/0x10e __might_sleep+0x3e/0x62 mutex_lock+0x20/0x4c regmap_lock_mutex+0x10/0x18 regmap_update_bits_base+0x2c/0x66 mcp23s08_irq_set_type+0x1ae/0x1d6 __irq_set_trigger+0x56/0x172 __setup_irq+0x1e6/0x646 request_threaded_irq+0xb6/0x160 ... We observed the problem while experimenting with a touchscreen driver which used MCP23017 IO expander (I2C). The regmap in the pinctrl-mcp23s08 driver uses a mutex for protection from concurrent accesses, which is the default for regmaps without .fast_io, .disable_locking, etc. mcp23s08_irq_set_type() calls regmap_update_bits_base(), and the latter locks the mutex. However, __setup_irq() locks desc->lock spinlock before calling these functions. As a result, the system tries to lock the mutex whole holding the spinlock. It seems, the internal regmap locks are not needed in this driver at all. mcp->lock seems to protect the regmap from concurrent accesses already, except, probably, in mcp_pinconf_get/set. mcp23s08_irq_set_type() and mcp23s08_irq_mask/unmask() are called under chip_bus_lock(), which calls mcp23s08_irq_bus_lock(). The latter takes mcp->lock and enables regmap caching, so that the potentially slow I2C accesses are deferred until chip_bus_unlock(). The accesses to the regmap from mcp23s08_probe_one() do not need additional locking. In all remaining places where the regmap is accessed, except mcp_pinconf_get/set(), the driver already takes mcp->lock. This patch adds locking in mcp_pinconf_get/set() and disables internal locking in the regmap config. Among other things, it fixes the sleeping in atomic context described above. Fixes: 8f38910ba4f6 ("pinctrl: mcp23s08: switch to regmap caching") Cc: stable@vger.kernel.org Signed-off-by: Evgenii Shatokhin Link: https://lore.kernel.org/20241209074659.1442898-1-e.shatokhin@yadro.com Signed-off-by: Linus Walleij --- drivers/pinctrl/pinctrl-mcp23s08.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/pinctrl/pinctrl-mcp23s08.c b/drivers/pinctrl/pinctrl-mcp23s08.c index d66c3a3e8429..b96e6368a956 100644 --- a/drivers/pinctrl/pinctrl-mcp23s08.c +++ b/drivers/pinctrl/pinctrl-mcp23s08.c @@ -86,6 +86,7 @@ const struct regmap_config mcp23x08_regmap = { .num_reg_defaults = ARRAY_SIZE(mcp23x08_defaults), .cache_type = REGCACHE_FLAT, .max_register = MCP_OLAT, + .disable_locking = true, /* mcp->lock protects the regmap */ }; EXPORT_SYMBOL_GPL(mcp23x08_regmap); @@ -132,6 +133,7 @@ const struct regmap_config mcp23x17_regmap = { .num_reg_defaults = ARRAY_SIZE(mcp23x17_defaults), .cache_type = REGCACHE_FLAT, .val_format_endian = REGMAP_ENDIAN_LITTLE, + .disable_locking = true, /* mcp->lock protects the regmap */ }; EXPORT_SYMBOL_GPL(mcp23x17_regmap); @@ -228,7 +230,9 @@ static int mcp_pinconf_get(struct pinctrl_dev *pctldev, unsigned int pin, switch (param) { case PIN_CONFIG_BIAS_PULL_UP: + mutex_lock(&mcp->lock); ret = mcp_read(mcp, MCP_GPPU, &data); + mutex_unlock(&mcp->lock); if (ret < 0) return ret; status = (data & BIT(pin)) ? 1 : 0; @@ -257,7 +261,9 @@ static int mcp_pinconf_set(struct pinctrl_dev *pctldev, unsigned int pin, switch (param) { case PIN_CONFIG_BIAS_PULL_UP: + mutex_lock(&mcp->lock); ret = mcp_set_bit(mcp, MCP_GPPU, pin, arg); + mutex_unlock(&mcp->lock); break; default: dev_dbg(mcp->dev, "Invalid config param %04x\n", param); From 185e1b1d91e419445d3fd99c1c0376a970438acf Mon Sep 17 00:00:00 2001 From: Joe Hattori Date: Mon, 16 Dec 2024 11:25:38 +0900 Subject: [PATCH 028/196] platform/x86: mlx-platform: call pci_dev_put() to balance the refcount MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit mlxplat_pci_fpga_device_init() calls pci_get_device() but does not release the refcount on error path. Call pci_dev_put() on the error path and in mlxplat_pci_fpga_device_exit() to fix this. This bug was found by an experimental static analysis tool that I am developing. Fixes: 02daa222fbdd ("platform: mellanox: Add initial support for PCIe based programming logic device") Signed-off-by: Joe Hattori Reviewed-by: Vadim Pasternak Link: https://lore.kernel.org/r/20241216022538.381209-1-joe@pf.is.s.u-tokyo.ac.jp Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen --- drivers/platform/x86/mlx-platform.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/platform/x86/mlx-platform.c b/drivers/platform/x86/mlx-platform.c index 671021cd1f59..9c7f30a47f1f 100644 --- a/drivers/platform/x86/mlx-platform.c +++ b/drivers/platform/x86/mlx-platform.c @@ -6237,6 +6237,7 @@ mlxplat_pci_fpga_device_init(unsigned int device, const char *res_name, struct p fail_pci_request_regions: pci_disable_device(pci_dev); fail_pci_enable_device: + pci_dev_put(pci_dev); return err; } @@ -6247,6 +6248,7 @@ mlxplat_pci_fpga_device_exit(struct pci_dev *pci_bridge, iounmap(pci_bridge_addr); pci_release_regions(pci_bridge); pci_disable_device(pci_bridge); + pci_dev_put(pci_bridge); } static int From f4bf0b909a6bf64a2220a42a7c8b8c2ee1b77b89 Mon Sep 17 00:00:00 2001 From: Maksim Kiselev Date: Tue, 10 Dec 2024 11:30:27 +0300 Subject: [PATCH 029/196] clk: thead: Fix TH1520 emmc and shdci clock rate In accordance with LicheePi 4A BSP the clock that comes to emmc/sdhci is 198Mhz which is got through frequency division of source clock VIDEO PLL by 4 [1]. But now the AP_SUBSYS driver sets the CLK EMMC SDIO to the same frequency as the VIDEO PLL, equal to 792 MHz. This causes emmc/sdhci to work 4 times slower. Let's fix this issue by adding fixed factor clock that divides VIDEO PLL by 4 for emmc/sdhci. Link: https://github.com/revyos/thead-kernel/blob/7563179071a314f41cdcdbfd8cf6e101e73707f3/drivers/clk/thead/clk-light-fm.c#L454 Fixes: ae81b69fd2b1 ("clk: thead: Add support for T-Head TH1520 AP_SUBSYS clocks") Signed-off-by: Maksim Kiselev Link: https://lore.kernel.org/r/20241210083029.92620-1-bigunclemax@gmail.com Tested-by: Xi Ruoyao Reviewed-by: Drew Fustini Signed-off-by: Stephen Boyd --- drivers/clk/thead/clk-th1520-ap.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/drivers/clk/thead/clk-th1520-ap.c b/drivers/clk/thead/clk-th1520-ap.c index 17e32ae08720..1015fab95251 100644 --- a/drivers/clk/thead/clk-th1520-ap.c +++ b/drivers/clk/thead/clk-th1520-ap.c @@ -779,6 +779,13 @@ static struct ccu_div dpu1_clk = { }, }; +static CLK_FIXED_FACTOR_HW(emmc_sdio_ref_clk, "emmc-sdio-ref", + &video_pll_clk.common.hw, 4, 1, 0); + +static const struct clk_parent_data emmc_sdio_ref_clk_pd[] = { + { .hw = &emmc_sdio_ref_clk.hw }, +}; + static CCU_GATE(CLK_BROM, brom_clk, "brom", ahb2_cpusys_hclk_pd, 0x100, BIT(4), 0); static CCU_GATE(CLK_BMU, bmu_clk, "bmu", axi4_cpusys2_aclk_pd, 0x100, BIT(5), 0); static CCU_GATE(CLK_AON2CPU_A2X, aon2cpu_a2x_clk, "aon2cpu-a2x", axi4_cpusys2_aclk_pd, @@ -798,7 +805,7 @@ static CCU_GATE(CLK_PERISYS_APB4_HCLK, perisys_apb4_hclk, "perisys-apb4-hclk", p 0x150, BIT(12), 0); static CCU_GATE(CLK_NPU_AXI, npu_axi_clk, "npu-axi", axi_aclk_pd, 0x1c8, BIT(5), 0); static CCU_GATE(CLK_CPU2VP, cpu2vp_clk, "cpu2vp", axi_aclk_pd, 0x1e0, BIT(13), 0); -static CCU_GATE(CLK_EMMC_SDIO, emmc_sdio_clk, "emmc-sdio", video_pll_clk_pd, 0x204, BIT(30), 0); +static CCU_GATE(CLK_EMMC_SDIO, emmc_sdio_clk, "emmc-sdio", emmc_sdio_ref_clk_pd, 0x204, BIT(30), 0); static CCU_GATE(CLK_GMAC1, gmac1_clk, "gmac1", gmac_pll_clk_pd, 0x204, BIT(26), 0); static CCU_GATE(CLK_PADCTRL1, padctrl1_clk, "padctrl1", perisys_apb_pclk_pd, 0x204, BIT(24), 0); static CCU_GATE(CLK_DSMART, dsmart_clk, "dsmart", perisys_apb_pclk_pd, 0x204, BIT(23), 0); @@ -1059,6 +1066,10 @@ static int th1520_clk_probe(struct platform_device *pdev) return ret; priv->hws[CLK_PLL_GMAC_100M] = &gmac_pll_clk_100m.hw; + ret = devm_clk_hw_register(dev, &emmc_sdio_ref_clk.hw); + if (ret) + return ret; + ret = devm_of_clk_add_hw_provider(dev, of_clk_hw_onecell_get, priv); if (ret) return ret; From cc0c53f4fac562efb3aca2bc493515e77642ae33 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Wed, 19 Jun 2024 14:12:45 -0700 Subject: [PATCH 030/196] wifi: iwlwifi: mvm: Fix __counted_by usage in cfg80211_wowlan_nd_* Both struct cfg80211_wowlan_nd_match and struct cfg80211_wowlan_nd_info pre-allocate space for channels and matches, but then may end up using fewer that the full allocation. Shrink the associated counter (n_channels and n_matches) after counting the results. This avoids compile-time (and run-time) warnings from __counted_by. (The counter member needs to be updated _before_ accessing the array index.) Seen with coming GCC 15: drivers/net/wireless/intel/iwlwifi/mvm/d3.c: In function 'iwl_mvm_query_set_freqs': drivers/net/wireless/intel/iwlwifi/mvm/d3.c:2877:66: warning: operation on 'match->n_channels' may be undefined [-Wsequence-point] 2877 | match->channels[match->n_channels++] = | ~~~~~~~~~~~~~~~~~^~ drivers/net/wireless/intel/iwlwifi/mvm/d3.c:2885:66: warning: operation on 'match->n_channels' may be undefined [-Wsequence-point] 2885 | match->channels[match->n_channels++] = | ~~~~~~~~~~~~~~~~~^~ drivers/net/wireless/intel/iwlwifi/mvm/d3.c: In function 'iwl_mvm_query_netdetect_reasons': drivers/net/wireless/intel/iwlwifi/mvm/d3.c:2982:58: warning: operation on 'net_detect->n_matches' may be undefined [-Wsequence-point] 2982 | net_detect->matches[net_detect->n_matches++] = match; | ~~~~~~~~~~~~~~~~~~~~~^~ Cc: stable@vger.kernel.org Fixes: aa4ec06c455d ("wifi: cfg80211: use __counted_by where appropriate") Signed-off-by: Kees Cook Reviewed-by: Gustavo A. R. Silva Link: https://patch.msgid.link/20240619211233.work.355-kees@kernel.org Signed-off-by: Johannes Berg --- drivers/net/wireless/intel/iwlwifi/mvm/d3.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/drivers/net/wireless/intel/iwlwifi/mvm/d3.c b/drivers/net/wireless/intel/iwlwifi/mvm/d3.c index f85c01e04ebf..7d973546c9fb 100644 --- a/drivers/net/wireless/intel/iwlwifi/mvm/d3.c +++ b/drivers/net/wireless/intel/iwlwifi/mvm/d3.c @@ -2954,6 +2954,7 @@ static void iwl_mvm_query_set_freqs(struct iwl_mvm *mvm, int idx) { int i; + int n_channels = 0; if (fw_has_api(&mvm->fw->ucode_capa, IWL_UCODE_TLV_API_SCAN_OFFLOAD_CHANS)) { @@ -2962,7 +2963,7 @@ static void iwl_mvm_query_set_freqs(struct iwl_mvm *mvm, for (i = 0; i < SCAN_OFFLOAD_MATCHING_CHANNELS_LEN * 8; i++) if (matches[idx].matching_channels[i / 8] & (BIT(i % 8))) - match->channels[match->n_channels++] = + match->channels[n_channels++] = mvm->nd_channels[i]->center_freq; } else { struct iwl_scan_offload_profile_match_v1 *matches = @@ -2970,9 +2971,11 @@ static void iwl_mvm_query_set_freqs(struct iwl_mvm *mvm, for (i = 0; i < SCAN_OFFLOAD_MATCHING_CHANNELS_LEN_V1 * 8; i++) if (matches[idx].matching_channels[i / 8] & (BIT(i % 8))) - match->channels[match->n_channels++] = + match->channels[n_channels++] = mvm->nd_channels[i]->center_freq; } + /* We may have ended up with fewer channels than we allocated. */ + match->n_channels = n_channels; } /** @@ -3053,6 +3056,8 @@ static void iwl_mvm_query_netdetect_reasons(struct iwl_mvm *mvm, GFP_KERNEL); if (!net_detect || !n_matches) goto out_report_nd; + net_detect->n_matches = n_matches; + n_matches = 0; for_each_set_bit(i, &matched_profiles, mvm->n_nd_match_sets) { struct cfg80211_wowlan_nd_match *match; @@ -3066,8 +3071,9 @@ static void iwl_mvm_query_netdetect_reasons(struct iwl_mvm *mvm, GFP_KERNEL); if (!match) goto out_report_nd; + match->n_channels = n_channels; - net_detect->matches[net_detect->n_matches++] = match; + net_detect->matches[n_matches++] = match; /* We inverted the order of the SSIDs in the scan * request, so invert the index here. @@ -3082,6 +3088,8 @@ static void iwl_mvm_query_netdetect_reasons(struct iwl_mvm *mvm, iwl_mvm_query_set_freqs(mvm, d3_data->nd_results, match, i); } + /* We may have fewer matches than we allocated. */ + net_detect->n_matches = n_matches; out_report_nd: wakeup.net_detect = net_detect; From 146b6057e1fd28fb1a38d300bf76a38dfba7f9fb Mon Sep 17 00:00:00 2001 From: Linus Walleij Date: Tue, 17 Dec 2024 13:55:48 +0100 Subject: [PATCH 031/196] wifi: cw1200: Fix potential NULL dereference A recent refactoring was identified by smatch to cause another potential NULL dereference: drivers/net/wireless/st/cw1200/cw1200_spi.c:440 cw1200_spi_disconnect() error: we previously assumed 'self' could be null (see line 433) Reported-by: kernel test robot Reported-by: Dan Carpenter Closes: https://lore.kernel.org/r/202411271742.Xa7CNVh1-lkp@intel.com/ Fixes: 2719a9e7156c ("wifi: cw1200: Convert to GPIO descriptors") Signed-off-by: Linus Walleij Signed-off-by: Kalle Valo Link: https://patch.msgid.link/20241217-cw1200-fix-v1-1-911e6b5823ec@linaro.org --- drivers/net/wireless/st/cw1200/cw1200_spi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/wireless/st/cw1200/cw1200_spi.c b/drivers/net/wireless/st/cw1200/cw1200_spi.c index 862964a8cc87..52386dfb5f4a 100644 --- a/drivers/net/wireless/st/cw1200/cw1200_spi.c +++ b/drivers/net/wireless/st/cw1200/cw1200_spi.c @@ -442,8 +442,8 @@ static void cw1200_spi_disconnect(struct spi_device *func) cw1200_core_release(self->core); self->core = NULL; } + cw1200_spi_off(self, dev_get_platdata(&func->dev)); } - cw1200_spi_off(self, dev_get_platdata(&func->dev)); } static int __maybe_unused cw1200_spi_suspend(struct device *dev) From 13a6691910cc23ea9ba4066e098603088673d5b0 Mon Sep 17 00:00:00 2001 From: Chiara Meiohas Date: Tue, 10 Dec 2024 09:33:10 +0200 Subject: [PATCH 032/196] RDMA/nldev: Set error code in rdma_nl_notify_event In case of error set the error code before the goto. Fixes: 6ff57a2ea7c2 ("RDMA/nldev: Fix NULL pointer dereferences issue in rdma_nl_notify_event") Reported-by: Dan Carpenter Closes: https://lore.kernel.org/linux-rdma/a84a2fc3-33b6-46da-a1bd-3343fa07eaf9@stanley.mountain/ Signed-off-by: Chiara Meiohas Reviewed-by: Maher Sanalla Link: https://patch.msgid.link/13eb25961923f5de9eb9ecbbc94e26113d6049ef.1733815944.git.leonro@nvidia.com Reviewed-by: Kalesh AP Signed-off-by: Leon Romanovsky --- drivers/infiniband/core/nldev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/core/nldev.c b/drivers/infiniband/core/nldev.c index ff121e59b9c0..cb987ab0177c 100644 --- a/drivers/infiniband/core/nldev.c +++ b/drivers/infiniband/core/nldev.c @@ -2833,8 +2833,8 @@ int rdma_nl_notify_event(struct ib_device *device, u32 port_num, enum rdma_nl_notify_event_type type) { struct sk_buff *skb; + int ret = -EMSGSIZE; struct net *net; - int ret = 0; void *nlh; net = read_pnet(&device->coredev.rdma_net); From 16b87037b48889d21854c8e97aec8a1baf2642b3 Mon Sep 17 00:00:00 2001 From: Bernard Metzler Date: Thu, 12 Dec 2024 16:18:48 +0100 Subject: [PATCH 033/196] RDMA/siw: Remove direct link to net_device Do not manage a per device direct link to net_device. Rely on associated ib_devices net_device management, not doubling the effort locally. A badly managed local link to net_device was causing a 'KASAN: slab-use-after-free' exception during siw_query_port() call. Fixes: bdcf26bf9b3a ("rdma/siw: network and RDMA core interface") Reported-by: syzbot+4b87489410b4efd181bf@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=4b87489410b4efd181bf Signed-off-by: Bernard Metzler Link: https://patch.msgid.link/20241212151848.564872-1-bmt@zurich.ibm.com Reviewed-by: Zhu Yanjun Signed-off-by: Leon Romanovsky --- drivers/infiniband/sw/siw/siw.h | 7 +++--- drivers/infiniband/sw/siw/siw_cm.c | 27 ++++++++++++++++----- drivers/infiniband/sw/siw/siw_main.c | 15 +----------- drivers/infiniband/sw/siw/siw_verbs.c | 35 ++++++++++++++++++--------- 4 files changed, 49 insertions(+), 35 deletions(-) diff --git a/drivers/infiniband/sw/siw/siw.h b/drivers/infiniband/sw/siw/siw.h index 86d4d6a2170e..ea5eee50dc39 100644 --- a/drivers/infiniband/sw/siw/siw.h +++ b/drivers/infiniband/sw/siw/siw.h @@ -46,6 +46,9 @@ */ #define SIW_IRQ_MAXBURST_SQ_ACTIVE 4 +/* There is always only a port 1 per siw device */ +#define SIW_PORT 1 + struct siw_dev_cap { int max_qp; int max_qp_wr; @@ -69,16 +72,12 @@ struct siw_pd { struct siw_device { struct ib_device base_dev; - struct net_device *netdev; struct siw_dev_cap attrs; u32 vendor_part_id; int numa_node; char raw_gid[ETH_ALEN]; - /* physical port state (only one port per device) */ - enum ib_port_state state; - spinlock_t lock; struct xarray qp_xa; diff --git a/drivers/infiniband/sw/siw/siw_cm.c b/drivers/infiniband/sw/siw/siw_cm.c index 86323918a570..708b13993fdf 100644 --- a/drivers/infiniband/sw/siw/siw_cm.c +++ b/drivers/infiniband/sw/siw/siw_cm.c @@ -1759,6 +1759,7 @@ int siw_create_listen(struct iw_cm_id *id, int backlog) { struct socket *s; struct siw_cep *cep = NULL; + struct net_device *ndev = NULL; struct siw_device *sdev = to_siw_dev(id->device); int addr_family = id->local_addr.ss_family; int rv = 0; @@ -1779,9 +1780,15 @@ int siw_create_listen(struct iw_cm_id *id, int backlog) struct sockaddr_in *laddr = &to_sockaddr_in(id->local_addr); /* For wildcard addr, limit binding to current device only */ - if (ipv4_is_zeronet(laddr->sin_addr.s_addr)) - s->sk->sk_bound_dev_if = sdev->netdev->ifindex; - + if (ipv4_is_zeronet(laddr->sin_addr.s_addr)) { + ndev = ib_device_get_netdev(id->device, SIW_PORT); + if (ndev) { + s->sk->sk_bound_dev_if = ndev->ifindex; + } else { + rv = -ENODEV; + goto error; + } + } rv = s->ops->bind(s, (struct sockaddr *)laddr, sizeof(struct sockaddr_in)); } else { @@ -1797,9 +1804,15 @@ int siw_create_listen(struct iw_cm_id *id, int backlog) } /* For wildcard addr, limit binding to current device only */ - if (ipv6_addr_any(&laddr->sin6_addr)) - s->sk->sk_bound_dev_if = sdev->netdev->ifindex; - + if (ipv6_addr_any(&laddr->sin6_addr)) { + ndev = ib_device_get_netdev(id->device, SIW_PORT); + if (ndev) { + s->sk->sk_bound_dev_if = ndev->ifindex; + } else { + rv = -ENODEV; + goto error; + } + } rv = s->ops->bind(s, (struct sockaddr *)laddr, sizeof(struct sockaddr_in6)); } @@ -1860,6 +1873,7 @@ int siw_create_listen(struct iw_cm_id *id, int backlog) } list_add_tail(&cep->listenq, (struct list_head *)id->provider_data); cep->state = SIW_EPSTATE_LISTENING; + dev_put(ndev); siw_dbg(id->device, "Listen at laddr %pISp\n", &id->local_addr); @@ -1879,6 +1893,7 @@ int siw_create_listen(struct iw_cm_id *id, int backlog) siw_cep_set_free_and_put(cep); } sock_release(s); + dev_put(ndev); return rv; } diff --git a/drivers/infiniband/sw/siw/siw_main.c b/drivers/infiniband/sw/siw/siw_main.c index 17abef48abcd..14d3103aee6f 100644 --- a/drivers/infiniband/sw/siw/siw_main.c +++ b/drivers/infiniband/sw/siw/siw_main.c @@ -287,7 +287,6 @@ static struct siw_device *siw_device_create(struct net_device *netdev) return NULL; base_dev = &sdev->base_dev; - sdev->netdev = netdev; if (netdev->addr_len) { memcpy(sdev->raw_gid, netdev->dev_addr, @@ -381,12 +380,10 @@ static int siw_netdev_event(struct notifier_block *nb, unsigned long event, switch (event) { case NETDEV_UP: - sdev->state = IB_PORT_ACTIVE; siw_port_event(sdev, 1, IB_EVENT_PORT_ACTIVE); break; case NETDEV_DOWN: - sdev->state = IB_PORT_DOWN; siw_port_event(sdev, 1, IB_EVENT_PORT_ERR); break; @@ -407,12 +404,8 @@ static int siw_netdev_event(struct notifier_block *nb, unsigned long event, siw_port_event(sdev, 1, IB_EVENT_LID_CHANGE); break; /* - * Todo: Below netdev events are currently not handled. + * All other events are not handled */ - case NETDEV_CHANGEMTU: - case NETDEV_CHANGE: - break; - default: break; } @@ -442,12 +435,6 @@ static int siw_newlink(const char *basedev_name, struct net_device *netdev) sdev = siw_device_create(netdev); if (sdev) { dev_dbg(&netdev->dev, "siw: new device\n"); - - if (netif_running(netdev) && netif_carrier_ok(netdev)) - sdev->state = IB_PORT_ACTIVE; - else - sdev->state = IB_PORT_DOWN; - ib_mark_name_assigned_by_user(&sdev->base_dev); rv = siw_device_register(sdev, basedev_name); if (rv) diff --git a/drivers/infiniband/sw/siw/siw_verbs.c b/drivers/infiniband/sw/siw/siw_verbs.c index 986666c19378..7ca0297d68a4 100644 --- a/drivers/infiniband/sw/siw/siw_verbs.c +++ b/drivers/infiniband/sw/siw/siw_verbs.c @@ -171,21 +171,29 @@ int siw_query_device(struct ib_device *base_dev, struct ib_device_attr *attr, int siw_query_port(struct ib_device *base_dev, u32 port, struct ib_port_attr *attr) { - struct siw_device *sdev = to_siw_dev(base_dev); + struct net_device *ndev; int rv; memset(attr, 0, sizeof(*attr)); rv = ib_get_eth_speed(base_dev, port, &attr->active_speed, &attr->active_width); + if (rv) + return rv; + + ndev = ib_device_get_netdev(base_dev, SIW_PORT); + if (!ndev) + return -ENODEV; + attr->gid_tbl_len = 1; attr->max_msg_sz = -1; - attr->max_mtu = ib_mtu_int_to_enum(sdev->netdev->mtu); - attr->active_mtu = ib_mtu_int_to_enum(sdev->netdev->mtu); - attr->phys_state = sdev->state == IB_PORT_ACTIVE ? + attr->max_mtu = ib_mtu_int_to_enum(ndev->max_mtu); + attr->active_mtu = ib_mtu_int_to_enum(READ_ONCE(ndev->mtu)); + attr->phys_state = (netif_running(ndev) && netif_carrier_ok(ndev)) ? IB_PORT_PHYS_STATE_LINK_UP : IB_PORT_PHYS_STATE_DISABLED; + attr->state = attr->phys_state == IB_PORT_PHYS_STATE_LINK_UP ? + IB_PORT_ACTIVE : IB_PORT_DOWN; attr->port_cap_flags = IB_PORT_CM_SUP | IB_PORT_DEVICE_MGMT_SUP; - attr->state = sdev->state; /* * All zero * @@ -199,6 +207,7 @@ int siw_query_port(struct ib_device *base_dev, u32 port, * attr->subnet_timeout = 0; * attr->init_type_repy = 0; */ + dev_put(ndev); return rv; } @@ -505,21 +514,24 @@ int siw_query_qp(struct ib_qp *base_qp, struct ib_qp_attr *qp_attr, int qp_attr_mask, struct ib_qp_init_attr *qp_init_attr) { struct siw_qp *qp; - struct siw_device *sdev; + struct net_device *ndev; - if (base_qp && qp_attr && qp_init_attr) { + if (base_qp && qp_attr && qp_init_attr) qp = to_siw_qp(base_qp); - sdev = to_siw_dev(base_qp->device); - } else { + else return -EINVAL; - } + + ndev = ib_device_get_netdev(base_qp->device, SIW_PORT); + if (!ndev) + return -ENODEV; + qp_attr->qp_state = siw_qp_state_to_ib_qp_state[qp->attrs.state]; qp_attr->cap.max_inline_data = SIW_MAX_INLINE; qp_attr->cap.max_send_wr = qp->attrs.sq_size; qp_attr->cap.max_send_sge = qp->attrs.sq_max_sges; qp_attr->cap.max_recv_wr = qp->attrs.rq_size; qp_attr->cap.max_recv_sge = qp->attrs.rq_max_sges; - qp_attr->path_mtu = ib_mtu_int_to_enum(sdev->netdev->mtu); + qp_attr->path_mtu = ib_mtu_int_to_enum(READ_ONCE(ndev->mtu)); qp_attr->max_rd_atomic = qp->attrs.irq_size; qp_attr->max_dest_rd_atomic = qp->attrs.orq_size; @@ -534,6 +546,7 @@ int siw_query_qp(struct ib_qp *base_qp, struct ib_qp_attr *qp_attr, qp_init_attr->cap = qp_attr->cap; + dev_put(ndev); return 0; } From 40be32303ec829ea12f9883e499bfd3fe9e52baf Mon Sep 17 00:00:00 2001 From: Selvin Xavier Date: Tue, 17 Dec 2024 15:56:45 +0530 Subject: [PATCH 034/196] RDMA/bnxt_re: Fix max_qp_wrs reported While creating qps, driver adds one extra entry to the sq size passed by the ULPs in order to avoid queue full condition. When ULPs creates QPs with max_qp_wr reported, driver creates QP with 1 more than the max_wqes supported by HW. Create QP fails in this case. To avoid this error, reduce 1 entry in max_qp_wqes and report it to the stack. Fixes: 1ac5a4047975 ("RDMA/bnxt_re: Add bnxt_re RoCE driver") Reviewed-by: Kalesh AP Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241217102649.1377704-2-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_sp.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_sp.c b/drivers/infiniband/hw/bnxt_re/qplib_sp.c index 7e20ae3d2c4f..73c9baaebb4e 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_sp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_sp.c @@ -129,7 +129,7 @@ int bnxt_qplib_get_dev_attr(struct bnxt_qplib_rcfw *rcfw, attr->max_qp_init_rd_atom = sb->max_qp_init_rd_atom > BNXT_QPLIB_MAX_OUT_RD_ATOM ? BNXT_QPLIB_MAX_OUT_RD_ATOM : sb->max_qp_init_rd_atom; - attr->max_qp_wqes = le16_to_cpu(sb->max_qp_wr); + attr->max_qp_wqes = le16_to_cpu(sb->max_qp_wr) - 1; /* * 128 WQEs needs to be reserved for the HW (8916). Prevent * reporting the max number From d5a38bf2f35979537c526acbc56bc435ed40685f Mon Sep 17 00:00:00 2001 From: Kalesh AP Date: Tue, 17 Dec 2024 15:56:46 +0530 Subject: [PATCH 035/196] RDMA/bnxt_re: Disable use of reserved wqes Disabling the reserved wqes logic for Gen P5/P7 devices because this workaround is required only for legacy devices. Fixes: ecb53febfcad ("RDMA/bnxt_en: Enable RDMA driver support for 57500 chip") Signed-off-by: Kalesh AP Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241217102649.1377704-3-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_sp.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_sp.c b/drivers/infiniband/hw/bnxt_re/qplib_sp.c index 73c9baaebb4e..776f8f1f1432 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_sp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_sp.c @@ -130,11 +130,13 @@ int bnxt_qplib_get_dev_attr(struct bnxt_qplib_rcfw *rcfw, sb->max_qp_init_rd_atom > BNXT_QPLIB_MAX_OUT_RD_ATOM ? BNXT_QPLIB_MAX_OUT_RD_ATOM : sb->max_qp_init_rd_atom; attr->max_qp_wqes = le16_to_cpu(sb->max_qp_wr) - 1; - /* - * 128 WQEs needs to be reserved for the HW (8916). Prevent - * reporting the max number - */ - attr->max_qp_wqes -= BNXT_QPLIB_RESERVED_QP_WRS + 1; + if (!bnxt_qplib_is_chip_gen_p5_p7(rcfw->res->cctx)) { + /* + * 128 WQEs needs to be reserved for the HW (8916). Prevent + * reporting the max number on legacy devices + */ + attr->max_qp_wqes -= BNXT_QPLIB_RESERVED_QP_WRS + 1; + } attr->max_qp_sges = cctx->modes.wqe_mode == BNXT_QPLIB_WQE_MODE_VARIABLE ? min_t(u32, sb->max_sge_var_wqe, BNXT_VAR_MAX_SGE) : 6; From d13be54dc18baee7a3e44349b80755a8c8205d3f Mon Sep 17 00:00:00 2001 From: Damodharam Ammepalli Date: Tue, 17 Dec 2024 15:56:47 +0530 Subject: [PATCH 036/196] RDMA/bnxt_re: Add send queue size check for variable wqe For the fixed WQE case, HW supports 0xFFFF WQEs. For variable Size WQEs, HW treats this number as the 16 bytes slots. The maximum supported WQEs needs to be adjusted based on the number of slots. Set a maximum WQE limit for variable WQE scenario. Fixes: de1d364c3815 ("RDMA/bnxt_re: Add support for Variable WQE in Genp7 adapters") Reviewed-by: Kalesh AP Signed-off-by: Damodharam Ammepalli Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241217102649.1377704-4-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_sp.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_sp.c b/drivers/infiniband/hw/bnxt_re/qplib_sp.c index 776f8f1f1432..9df3e3271577 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_sp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_sp.c @@ -138,6 +138,10 @@ int bnxt_qplib_get_dev_attr(struct bnxt_qplib_rcfw *rcfw, attr->max_qp_wqes -= BNXT_QPLIB_RESERVED_QP_WRS + 1; } + /* Adjust for max_qp_wqes for variable wqe */ + if (cctx->modes.wqe_mode == BNXT_QPLIB_WQE_MODE_VARIABLE) + attr->max_qp_wqes = BNXT_VAR_MAX_WQE - 1; + attr->max_qp_sges = cctx->modes.wqe_mode == BNXT_QPLIB_WQE_MODE_VARIABLE ? min_t(u32, sb->max_sge_var_wqe, BNXT_VAR_MAX_SGE) : 6; attr->max_cq = le32_to_cpu(sb->max_cq); From bb839f3ace0fee532a0487b692cc4d868fccb7cf Mon Sep 17 00:00:00 2001 From: Damodharam Ammepalli Date: Tue, 17 Dec 2024 15:56:48 +0530 Subject: [PATCH 037/196] RDMA/bnxt_re: Fix MSN table size for variable wqe mode For variable size wqe mode, the MSN table size should be half the size of the SQ depth. Fixing this to avoid wrap around problems in the retransmission path. Fixes: de1d364c3815 ("RDMA/bnxt_re: Add support for Variable WQE in Genp7 adapters") Reviewed-by: Kashyap Desai Reviewed-by: Kalesh AP Signed-off-by: Damodharam Ammepalli Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241217102649.1377704-5-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_fp.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index d8a2a929bbe3..951ad90f5aa9 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -1033,7 +1033,12 @@ int bnxt_qplib_create_qp(struct bnxt_qplib_res *res, struct bnxt_qplib_qp *qp) : 0; /* Update msn tbl size */ if (qp->is_host_msn_tbl && psn_sz) { - hwq_attr.aux_depth = roundup_pow_of_two(bnxt_qplib_set_sq_size(sq, qp->wqe_mode)); + if (qp->wqe_mode == BNXT_QPLIB_WQE_MODE_STATIC) + hwq_attr.aux_depth = + roundup_pow_of_two(bnxt_qplib_set_sq_size(sq, qp->wqe_mode)); + else + hwq_attr.aux_depth = + roundup_pow_of_two(bnxt_qplib_set_sq_size(sq, qp->wqe_mode)) / 2; qp->msn_tbl_sz = hwq_attr.aux_depth; qp->msn = 0; } From 9272cba0ded71b5a2084da3004ec7806b8cb7fd2 Mon Sep 17 00:00:00 2001 From: Selvin Xavier Date: Tue, 17 Dec 2024 15:56:49 +0530 Subject: [PATCH 038/196] RDMA/bnxt_re: Fix the locking while accessing the QP table QP table handling is synchronized with destroy QP and Async event from the HW. The same needs to be synchronized during create_qp also. Use the same lock in create_qp also. Fixes: 76d3ddff7153 ("RDMA/bnxt_re: synchronize the qp-handle table array") Fixes: f218d67ef004 ("RDMA/bnxt_re: Allow posting when QPs are in error") Fixes: 84cf229f4001 ("RDMA/bnxt_re: Fix the qp table indexing") Signed-off-by: Selvin Xavier Link: https://patch.msgid.link/20241217102649.1377704-6-kalesh-anakkur.purayil@broadcom.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/qplib_fp.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index 951ad90f5aa9..5336f74297f8 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -1181,9 +1181,11 @@ int bnxt_qplib_create_qp(struct bnxt_qplib_res *res, struct bnxt_qplib_qp *qp) rq->dbinfo.db = qp->dpi->dbr; rq->dbinfo.max_slot = bnxt_qplib_set_rq_max_slot(rq->wqe_size); } + spin_lock_bh(&rcfw->tbl_lock); tbl_indx = map_qp_id_to_tbl_indx(qp->id, rcfw); rcfw->qp_tbl[tbl_indx].qp_id = qp->id; rcfw->qp_tbl[tbl_indx].qp_handle = (void *)qp; + spin_unlock_bh(&rcfw->tbl_lock); return 0; fail: From 8d90a86ed053226a297ce062f4d9f4f521e05c4c Mon Sep 17 00:00:00 2001 From: Eric Biggers Date: Thu, 12 Dec 2024 20:19:48 -0800 Subject: [PATCH 039/196] mmc: sdhci-msm: fix crypto key eviction Commit c7eed31e235c ("mmc: sdhci-msm: Switch to the new ICE API") introduced an incorrect check of the algorithm ID into the key eviction path, and thus qcom_ice_evict_key() is no longer ever called. Fix it. Fixes: c7eed31e235c ("mmc: sdhci-msm: Switch to the new ICE API") Cc: stable@vger.kernel.org Cc: Abel Vesa Signed-off-by: Eric Biggers Message-ID: <20241213041958.202565-6-ebiggers@kernel.org> Signed-off-by: Ulf Hansson --- drivers/mmc/host/sdhci-msm.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/drivers/mmc/host/sdhci-msm.c b/drivers/mmc/host/sdhci-msm.c index e00208535bd1..319f0ebbe652 100644 --- a/drivers/mmc/host/sdhci-msm.c +++ b/drivers/mmc/host/sdhci-msm.c @@ -1867,20 +1867,20 @@ static int sdhci_msm_program_key(struct cqhci_host *cq_host, struct sdhci_msm_host *msm_host = sdhci_pltfm_priv(pltfm_host); union cqhci_crypto_cap_entry cap; + if (!(cfg->config_enable & CQHCI_CRYPTO_CONFIGURATION_ENABLE)) + return qcom_ice_evict_key(msm_host->ice, slot); + /* Only AES-256-XTS has been tested so far. */ cap = cq_host->crypto_cap_array[cfg->crypto_cap_idx]; if (cap.algorithm_id != CQHCI_CRYPTO_ALG_AES_XTS || cap.key_size != CQHCI_CRYPTO_KEY_SIZE_256) return -EINVAL; - if (cfg->config_enable & CQHCI_CRYPTO_CONFIGURATION_ENABLE) - return qcom_ice_program_key(msm_host->ice, - QCOM_ICE_CRYPTO_ALG_AES_XTS, - QCOM_ICE_CRYPTO_KEY_SIZE_256, - cfg->crypto_key, - cfg->data_unit_size, slot); - else - return qcom_ice_evict_key(msm_host->ice, slot); + return qcom_ice_program_key(msm_host->ice, + QCOM_ICE_CRYPTO_ALG_AES_XTS, + QCOM_ICE_CRYPTO_KEY_SIZE_256, + cfg->crypto_key, + cfg->data_unit_size, slot); } #else /* CONFIG_MMC_CRYPTO */ From 469c0682e03d67d8dc970ecaa70c2d753057c7c0 Mon Sep 17 00:00:00 2001 From: Joe Hattori Date: Sun, 15 Dec 2024 12:01:59 +0900 Subject: [PATCH 040/196] pmdomain: imx: gpcv2: fix an OF node reference leak in imx_gpcv2_probe() imx_gpcv2_probe() leaks an OF node reference obtained by of_get_child_by_name(). Fix it by declaring the device node with the __free(device_node) cleanup construct. This bug was found by an experimental static analysis tool that I am developing. Fixes: 03aa12629fc4 ("soc: imx: Add GPCv2 power gating driver") Signed-off-by: Joe Hattori Cc: stable@vger.kernel.org Message-ID: <20241215030159.1526624-1-joe@pf.is.s.u-tokyo.ac.jp> Signed-off-by: Ulf Hansson --- drivers/pmdomain/imx/gpcv2.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/pmdomain/imx/gpcv2.c b/drivers/pmdomain/imx/gpcv2.c index e67ecf99ef84..9bdb80fd7210 100644 --- a/drivers/pmdomain/imx/gpcv2.c +++ b/drivers/pmdomain/imx/gpcv2.c @@ -1458,12 +1458,12 @@ static int imx_gpcv2_probe(struct platform_device *pdev) .max_register = SZ_4K, }; struct device *dev = &pdev->dev; - struct device_node *pgc_np; + struct device_node *pgc_np __free(device_node) = + of_get_child_by_name(dev->of_node, "pgc"); struct regmap *regmap; void __iomem *base; int ret; - pgc_np = of_get_child_by_name(dev->of_node, "pgc"); if (!pgc_np) { dev_err(dev, "No power domains specified in DT\n"); return -EINVAL; From f64f610ec6ab59dd0391b03842cea3a4cd8ee34f Mon Sep 17 00:00:00 2001 From: Lucas Stach Date: Wed, 18 Dec 2024 19:44:33 +0100 Subject: [PATCH 041/196] pmdomain: core: add dummy release function to genpd device The genpd device, which is really only used as a handle to lookup OPP, but not even registered to the device core otherwise and thus lifetime linked to the genpd struct it is contained in, is missing a release function. After b8f7bbd1f4ec ("pmdomain: core: Add missing put_device()") the device will be cleaned up going through the driver core device_release() function, which will warn when no release callback is present for the device. Add a dummy release function to shut up the warning. Signed-off-by: Lucas Stach Tested-by: Luca Ceresoli Fixes: b8f7bbd1f4ec ("pmdomain: core: Add missing put_device()") Cc: stable@vger.kernel.org Message-ID: <20241218184433.1930532-1-l.stach@pengutronix.de> Signed-off-by: Ulf Hansson --- drivers/pmdomain/core.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/pmdomain/core.c b/drivers/pmdomain/core.c index bb11f467dc78..20a9efebbcb7 100644 --- a/drivers/pmdomain/core.c +++ b/drivers/pmdomain/core.c @@ -2142,6 +2142,11 @@ static int genpd_set_default_power_state(struct generic_pm_domain *genpd) return 0; } +static void genpd_provider_release(struct device *dev) +{ + /* nothing to be done here */ +} + static int genpd_alloc_data(struct generic_pm_domain *genpd) { struct genpd_governor_data *gd = NULL; @@ -2173,6 +2178,7 @@ static int genpd_alloc_data(struct generic_pm_domain *genpd) genpd->gd = gd; device_initialize(&genpd->dev); + genpd->dev.release = genpd_provider_release; if (!genpd_is_dev_name_fw(genpd)) { dev_set_name(&genpd->dev, "%s", genpd->name); From de35994ecd2dd6148ab5a6c5050a1670a04dec77 Mon Sep 17 00:00:00 2001 From: Tvrtko Ursulin Date: Thu, 19 Dec 2024 09:30:30 +0000 Subject: [PATCH 042/196] workqueue: Do not warn when cancelling WQ_MEM_RECLAIM work from !WQ_MEM_RECLAIM worker MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit After commit 746ae46c1113 ("drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM") amdgpu started seeing the following warning: [ ] workqueue: WQ_MEM_RECLAIM sdma0:drm_sched_run_job_work [gpu_sched] is flushing !WQ_MEM_RECLAIM events:amdgpu_device_delay_enable_gfx_off [amdgpu] ... [ ] Workqueue: sdma0 drm_sched_run_job_work [gpu_sched] ... [ ] Call Trace: [ ] ... [ ] ? check_flush_dependency+0xf5/0x110 ... [ ] cancel_delayed_work_sync+0x6e/0x80 [ ] amdgpu_gfx_off_ctrl+0xab/0x140 [amdgpu] [ ] amdgpu_ring_alloc+0x40/0x50 [amdgpu] [ ] amdgpu_ib_schedule+0xf4/0x810 [amdgpu] [ ] ? drm_sched_run_job_work+0x22c/0x430 [gpu_sched] [ ] amdgpu_job_run+0xaa/0x1f0 [amdgpu] [ ] drm_sched_run_job_work+0x257/0x430 [gpu_sched] [ ] process_one_work+0x217/0x720 ... [ ] The intent of the verifcation done in check_flush_depedency is to ensure forward progress during memory reclaim, by flagging cases when either a memory reclaim process, or a memory reclaim work item is flushed from a context not marked as memory reclaim safe. This is correct when flushing, but when called from the cancel(_delayed)_work_sync() paths it is a false positive because work is either already running, or will not be running at all. Therefore cancelling it is safe and we can relax the warning criteria by letting the helper know of the calling context. Signed-off-by: Tvrtko Ursulin Fixes: fca839c00a12 ("workqueue: warn if memory reclaim tries to flush !WQ_MEM_RECLAIM workqueue") References: 746ae46c1113 ("drm/sched: Mark scheduler work queues with WQ_MEM_RECLAIM") Cc: Tejun Heo Cc: Peter Zijlstra Cc: Lai Jiangshan Cc: Alex Deucher Cc: Christian König Cc: # v4.5+ Signed-off-by: Tejun Heo --- kernel/workqueue.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 8b07576814a5..8336218ec4b8 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -3680,23 +3680,27 @@ void workqueue_softirq_dead(unsigned int cpu) * check_flush_dependency - check for flush dependency sanity * @target_wq: workqueue being flushed * @target_work: work item being flushed (NULL for workqueue flushes) + * @from_cancel: are we called from the work cancel path * * %current is trying to flush the whole @target_wq or @target_work on it. - * If @target_wq doesn't have %WQ_MEM_RECLAIM, verify that %current is not - * reclaiming memory or running on a workqueue which doesn't have - * %WQ_MEM_RECLAIM as that can break forward-progress guarantee leading to - * a deadlock. + * If this is not the cancel path (which implies work being flushed is either + * already running, or will not be at all), check if @target_wq doesn't have + * %WQ_MEM_RECLAIM and verify that %current is not reclaiming memory or running + * on a workqueue which doesn't have %WQ_MEM_RECLAIM as that can break forward- + * progress guarantee leading to a deadlock. */ static void check_flush_dependency(struct workqueue_struct *target_wq, - struct work_struct *target_work) + struct work_struct *target_work, + bool from_cancel) { - work_func_t target_func = target_work ? target_work->func : NULL; + work_func_t target_func; struct worker *worker; - if (target_wq->flags & WQ_MEM_RECLAIM) + if (from_cancel || target_wq->flags & WQ_MEM_RECLAIM) return; worker = current_wq_worker(); + target_func = target_work ? target_work->func : NULL; WARN_ONCE(current->flags & PF_MEMALLOC, "workqueue: PF_MEMALLOC task %d(%s) is flushing !WQ_MEM_RECLAIM %s:%ps", @@ -3980,7 +3984,7 @@ void __flush_workqueue(struct workqueue_struct *wq) list_add_tail(&this_flusher.list, &wq->flusher_overflow); } - check_flush_dependency(wq, NULL); + check_flush_dependency(wq, NULL, false); mutex_unlock(&wq->mutex); @@ -4155,7 +4159,7 @@ static bool start_flush_work(struct work_struct *work, struct wq_barrier *barr, } wq = pwq->wq; - check_flush_dependency(wq, work); + check_flush_dependency(wq, work, from_cancel); insert_wq_barrier(pwq, barr, work, worker); raw_spin_unlock_irq(&pool->lock); From 25c6a5ab151fb9c886552bf5aa7cbf2a5c6e96af Mon Sep 17 00:00:00 2001 From: Wei Fang Date: Tue, 17 Dec 2024 14:35:00 +0800 Subject: [PATCH 043/196] net: phy: micrel: Dynamically control external clock of KSZ PHY On the i.MX6ULL-14x14-EVK board, enet1_ref and enet2_ref are used as the clock sources for two external KSZ PHYs. However, after closing the two FEC ports, the clk_enable_count of the enet1_ref and enet2_ref clocks is not 0. The root cause is that since the commit 985329462723 ("net: phy: micrel: use devm_clk_get_optional_enabled for the rmii-ref clock"), the external clock of KSZ PHY has been enabled when the PHY driver probes, and it can only be disabled when the PHY driver is removed. This causes the clock to continue working when the system is suspended or the network port is down. Although Heiko explained in the commit message that the patch was because some clock suppliers need to enable the clock to get the valid clock rate , it seems that the simple fix is to disable the clock after getting the clock rate to solve the current problem. This is indeed true, but we need to admit that Heiko's patch has been applied for more than a year, and we cannot guarantee whether there are platforms that only enable rmii-ref in the KSZ PHY driver during this period. If this is the case, disabling rmii-ref will cause RMII on these platforms to not work. Secondly, commit 99ac4cbcc2a5 ("net: phy: micrel: allow usage of generic ethernet-phy clock") just simply enables the generic clock permanently, which seems like the generic clock may only be enabled in the PHY driver. If we simply disable the generic clock, RMII may not work. If we keep it as it is, the platform using the generic clock will have the same problem as the i.MX6ULL platform. To solve this problem, the clock is enabled when phy_driver::resume() is called, and the clock is disabled when phy_driver::suspend() is called. Since phy_driver::resume() and phy_driver::suspend() are not called in pairs, an additional clk_enable flag is added. When phy_driver::suspend() is called, the clock is disabled only if clk_enable is true. Conversely, when phy_driver::resume() is called, the clock is enabled if clk_enable is false. The changes that introduced the problem were only a few lines, while the current fix is about a hundred lines, which seems out of proportion, but it is necessary because kszphy_probe() is used by multiple KSZ PHYs and we need to fix all of them. Fixes: 985329462723 ("net: phy: micrel: use devm_clk_get_optional_enabled for the rmii-ref clock") Fixes: 99ac4cbcc2a5 ("net: phy: micrel: allow usage of generic ethernet-phy clock") Signed-off-by: Wei Fang Reviewed-by: Andrew Lunn Reviewed-by: Florian Fainelli Link: https://patch.msgid.link/20241217063500.1424011-1-wei.fang@nxp.com Signed-off-by: Jakub Kicinski --- drivers/net/phy/micrel.c | 114 ++++++++++++++++++++++++++++++++++----- 1 file changed, 101 insertions(+), 13 deletions(-) diff --git a/drivers/net/phy/micrel.c b/drivers/net/phy/micrel.c index 3ef508840674..eeb33eb181ac 100644 --- a/drivers/net/phy/micrel.c +++ b/drivers/net/phy/micrel.c @@ -432,10 +432,12 @@ struct kszphy_ptp_priv { struct kszphy_priv { struct kszphy_ptp_priv ptp_priv; const struct kszphy_type *type; + struct clk *clk; int led_mode; u16 vct_ctrl1000; bool rmii_ref_clk_sel; bool rmii_ref_clk_sel_val; + bool clk_enable; u64 stats[ARRAY_SIZE(kszphy_hw_stats)]; }; @@ -2050,6 +2052,46 @@ static void kszphy_get_stats(struct phy_device *phydev, data[i] = kszphy_get_stat(phydev, i); } +static void kszphy_enable_clk(struct phy_device *phydev) +{ + struct kszphy_priv *priv = phydev->priv; + + if (!priv->clk_enable && priv->clk) { + clk_prepare_enable(priv->clk); + priv->clk_enable = true; + } +} + +static void kszphy_disable_clk(struct phy_device *phydev) +{ + struct kszphy_priv *priv = phydev->priv; + + if (priv->clk_enable && priv->clk) { + clk_disable_unprepare(priv->clk); + priv->clk_enable = false; + } +} + +static int kszphy_generic_resume(struct phy_device *phydev) +{ + kszphy_enable_clk(phydev); + + return genphy_resume(phydev); +} + +static int kszphy_generic_suspend(struct phy_device *phydev) +{ + int ret; + + ret = genphy_suspend(phydev); + if (ret) + return ret; + + kszphy_disable_clk(phydev); + + return 0; +} + static int kszphy_suspend(struct phy_device *phydev) { /* Disable PHY Interrupts */ @@ -2059,7 +2101,7 @@ static int kszphy_suspend(struct phy_device *phydev) phydev->drv->config_intr(phydev); } - return genphy_suspend(phydev); + return kszphy_generic_suspend(phydev); } static void kszphy_parse_led_mode(struct phy_device *phydev) @@ -2090,7 +2132,9 @@ static int kszphy_resume(struct phy_device *phydev) { int ret; - genphy_resume(phydev); + ret = kszphy_generic_resume(phydev); + if (ret) + return ret; /* After switching from power-down to normal mode, an internal global * reset is automatically generated. Wait a minimum of 1 ms before @@ -2112,6 +2156,24 @@ static int kszphy_resume(struct phy_device *phydev) return 0; } +/* Because of errata DS80000700A, receiver error following software + * power down. Suspend and resume callbacks only disable and enable + * external rmii reference clock. + */ +static int ksz8041_resume(struct phy_device *phydev) +{ + kszphy_enable_clk(phydev); + + return 0; +} + +static int ksz8041_suspend(struct phy_device *phydev) +{ + kszphy_disable_clk(phydev); + + return 0; +} + static int ksz9477_resume(struct phy_device *phydev) { int ret; @@ -2159,7 +2221,10 @@ static int ksz8061_resume(struct phy_device *phydev) if (!(ret & BMCR_PDOWN)) return 0; - genphy_resume(phydev); + ret = kszphy_generic_resume(phydev); + if (ret) + return ret; + usleep_range(1000, 2000); /* Re-program the value after chip is reset. */ @@ -2177,6 +2242,11 @@ static int ksz8061_resume(struct phy_device *phydev) return 0; } +static int ksz8061_suspend(struct phy_device *phydev) +{ + return kszphy_suspend(phydev); +} + static int kszphy_probe(struct phy_device *phydev) { const struct kszphy_type *type = phydev->drv->driver_data; @@ -2217,10 +2287,14 @@ static int kszphy_probe(struct phy_device *phydev) } else if (!clk) { /* unnamed clock from the generic ethernet-phy binding */ clk = devm_clk_get_optional_enabled(&phydev->mdio.dev, NULL); - if (IS_ERR(clk)) - return PTR_ERR(clk); } + if (IS_ERR(clk)) + return PTR_ERR(clk); + + clk_disable_unprepare(clk); + priv->clk = clk; + if (ksz8041_fiber_mode(phydev)) phydev->port = PORT_FIBRE; @@ -5290,6 +5364,21 @@ static int lan8841_probe(struct phy_device *phydev) return 0; } +static int lan8804_resume(struct phy_device *phydev) +{ + return kszphy_resume(phydev); +} + +static int lan8804_suspend(struct phy_device *phydev) +{ + return kszphy_generic_suspend(phydev); +} + +static int lan8841_resume(struct phy_device *phydev) +{ + return kszphy_generic_resume(phydev); +} + static int lan8841_suspend(struct phy_device *phydev) { struct kszphy_priv *priv = phydev->priv; @@ -5298,7 +5387,7 @@ static int lan8841_suspend(struct phy_device *phydev) if (ptp_priv->ptp_clock) ptp_cancel_worker_sync(ptp_priv->ptp_clock); - return genphy_suspend(phydev); + return kszphy_generic_suspend(phydev); } static struct phy_driver ksphy_driver[] = { @@ -5358,9 +5447,8 @@ static struct phy_driver ksphy_driver[] = { .get_sset_count = kszphy_get_sset_count, .get_strings = kszphy_get_strings, .get_stats = kszphy_get_stats, - /* No suspend/resume callbacks because of errata DS80000700A, - * receiver error following software power down. - */ + .suspend = ksz8041_suspend, + .resume = ksz8041_resume, }, { .phy_id = PHY_ID_KSZ8041RNLI, .phy_id_mask = MICREL_PHY_ID_MASK, @@ -5436,7 +5524,7 @@ static struct phy_driver ksphy_driver[] = { .soft_reset = genphy_soft_reset, .config_intr = kszphy_config_intr, .handle_interrupt = kszphy_handle_interrupt, - .suspend = kszphy_suspend, + .suspend = ksz8061_suspend, .resume = ksz8061_resume, }, { .phy_id = PHY_ID_KSZ9021, @@ -5507,8 +5595,8 @@ static struct phy_driver ksphy_driver[] = { .get_sset_count = kszphy_get_sset_count, .get_strings = kszphy_get_strings, .get_stats = kszphy_get_stats, - .suspend = genphy_suspend, - .resume = kszphy_resume, + .suspend = lan8804_suspend, + .resume = lan8804_resume, .config_intr = lan8804_config_intr, .handle_interrupt = lan8804_handle_interrupt, }, { @@ -5526,7 +5614,7 @@ static struct phy_driver ksphy_driver[] = { .get_strings = kszphy_get_strings, .get_stats = kszphy_get_stats, .suspend = lan8841_suspend, - .resume = genphy_resume, + .resume = lan8841_resume, .cable_test_start = lan8814_cable_test_start, .cable_test_get_status = ksz886x_cable_test_get_status, }, { From 902806baf3c1e8383c1fe3ff0b6042b8cb5c2707 Mon Sep 17 00:00:00 2001 From: Stefan Ekenberg Date: Tue, 19 Nov 2024 08:40:29 +0100 Subject: [PATCH 044/196] drm/bridge: adv7511_audio: Update Audio InfoFrame properly AUDIO_UPDATE bit (Bit 5 of MAIN register 0x4A) needs to be set to 1 while updating Audio InfoFrame information and then set to 0 when done. Otherwise partially updated Audio InfoFrames could be sent out. Two cases where this rule were not followed are fixed: - In adv7511_hdmi_hw_params() make sure AUDIO_UPDATE bit is updated before/after setting ADV7511_REG_AUDIO_INFOFRAME. - In audio_startup() use the correct register for clearing AUDIO_UPDATE bit. The problem with corrupted audio infoframes were discovered by letting a HDMI logic analyser check the output of ADV7535. Note that this patchs replaces writing REG_GC(1) with REG_INFOFRAME_UPDATE. Bit 5 of REG_GC(1) is positioned within field GC_PP[3:0] and that field doesn't control audio infoframe and is read- only. My conclusion therefore was that the author if this code meant to clear bit 5 of REG_INFOFRAME_UPDATE from the very beginning. Tested-by: Biju Das Fixes: 53c515befe28 ("drm/bridge: adv7511: Add Audio support") Signed-off-by: Stefan Ekenberg Reviewed-by: Dmitry Baryshkov Link: https://patchwork.freedesktop.org/patch/msgid/20241119-adv7511-audio-info-frame-v4-1-4ae68e76c89c@axis.com Signed-off-by: Dmitry Baryshkov --- drivers/gpu/drm/bridge/adv7511/adv7511_audio.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/bridge/adv7511/adv7511_audio.c b/drivers/gpu/drm/bridge/adv7511/adv7511_audio.c index 61f4a38e7d2b..8f786592143b 100644 --- a/drivers/gpu/drm/bridge/adv7511/adv7511_audio.c +++ b/drivers/gpu/drm/bridge/adv7511/adv7511_audio.c @@ -153,7 +153,16 @@ static int adv7511_hdmi_hw_params(struct device *dev, void *data, ADV7511_AUDIO_CFG3_LEN_MASK, len); regmap_update_bits(adv7511->regmap, ADV7511_REG_I2C_FREQ_ID_CFG, ADV7511_I2C_FREQ_ID_CFG_RATE_MASK, rate << 4); - regmap_write(adv7511->regmap, 0x73, 0x1); + + /* send current Audio infoframe values while updating */ + regmap_update_bits(adv7511->regmap, ADV7511_REG_INFOFRAME_UPDATE, + BIT(5), BIT(5)); + + regmap_write(adv7511->regmap, ADV7511_REG_AUDIO_INFOFRAME(0), 0x1); + + /* use Audio infoframe updated info */ + regmap_update_bits(adv7511->regmap, ADV7511_REG_INFOFRAME_UPDATE, + BIT(5), 0); return 0; } @@ -184,8 +193,9 @@ static int audio_startup(struct device *dev, void *data) regmap_update_bits(adv7511->regmap, ADV7511_REG_GC(0), BIT(7) | BIT(6), BIT(7)); /* use Audio infoframe updated info */ - regmap_update_bits(adv7511->regmap, ADV7511_REG_GC(1), + regmap_update_bits(adv7511->regmap, ADV7511_REG_INFOFRAME_UPDATE, BIT(5), 0); + /* enable SPDIF receiver */ if (adv7511->audio_source == ADV7511_AUDIO_SOURCE_SPDIF) regmap_update_bits(adv7511->regmap, ADV7511_REG_AUDIO_CONFIG, From 81adbd3ff21c1182e06aa02c6be0bfd9ea02d8e8 Mon Sep 17 00:00:00 2001 From: Biju Das Date: Tue, 19 Nov 2024 19:20:29 +0000 Subject: [PATCH 045/196] drm: adv7511: Fix use-after-free in adv7533_attach_dsi() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The host_node pointer was assigned and freed in adv7533_parse_dt(), and later, adv7533_attach_dsi() uses the same. Fix this use-after-free issue by dropping of_node_put() in adv7533_parse_dt() and calling of_node_put() in error path of probe() and also in the remove(). Fixes: 1e4d58cd7f88 ("drm/bridge: adv7533: Create a MIPI DSI device") Cc: stable@vger.kernel.org Reviewed-by: Laurent Pinchart Signed-off-by: Biju Das Link: https://patchwork.freedesktop.org/patch/msgid/20241119192040.152657-2-biju.das.jz@bp.renesas.com Signed-off-by: Dmitry Baryshkov --- drivers/gpu/drm/bridge/adv7511/adv7511_drv.c | 10 ++++++++-- drivers/gpu/drm/bridge/adv7511/adv7533.c | 2 -- 2 files changed, 8 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c b/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c index eb5919b38263..a13b3d8ab6ac 100644 --- a/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c +++ b/drivers/gpu/drm/bridge/adv7511/adv7511_drv.c @@ -1241,8 +1241,10 @@ static int adv7511_probe(struct i2c_client *i2c) return ret; ret = adv7511_init_regulators(adv7511); - if (ret) - return dev_err_probe(dev, ret, "failed to init regulators\n"); + if (ret) { + dev_err_probe(dev, ret, "failed to init regulators\n"); + goto err_of_node_put; + } /* * The power down GPIO is optional. If present, toggle it from active to @@ -1363,6 +1365,8 @@ static int adv7511_probe(struct i2c_client *i2c) i2c_unregister_device(adv7511->i2c_edid); uninit_regulators: adv7511_uninit_regulators(adv7511); +err_of_node_put: + of_node_put(adv7511->host_node); return ret; } @@ -1371,6 +1375,8 @@ static void adv7511_remove(struct i2c_client *i2c) { struct adv7511 *adv7511 = i2c_get_clientdata(i2c); + of_node_put(adv7511->host_node); + adv7511_uninit_regulators(adv7511); drm_bridge_remove(&adv7511->bridge); diff --git a/drivers/gpu/drm/bridge/adv7511/adv7533.c b/drivers/gpu/drm/bridge/adv7511/adv7533.c index 4481489aaf5e..5f195e91b3e6 100644 --- a/drivers/gpu/drm/bridge/adv7511/adv7533.c +++ b/drivers/gpu/drm/bridge/adv7511/adv7533.c @@ -181,8 +181,6 @@ int adv7533_parse_dt(struct device_node *np, struct adv7511 *adv) if (!adv->host_node) return -ENODEV; - of_node_put(adv->host_node); - adv->use_timing_gen = !of_property_read_bool(np, "adi,disable-timing-generator"); From ee8f9ed57a397605434caeef351bafa3ec4dfdd4 Mon Sep 17 00:00:00 2001 From: Biju Das Date: Tue, 19 Nov 2024 19:20:30 +0000 Subject: [PATCH 046/196] dt-bindings: display: adi,adv7533: Drop single lane support As per [1] and [2], ADV7535/7533 supports only 2-, 3-, or 4-lane. Drop unsupported 1-lane from bindings. [1] https://www.analog.com/media/en/technical-documentation/data-sheets/ADV7535.pdf [2] https://www.analog.com/media/en/technical-documentation/data-sheets/ADV7533.pdf Fixes: 1e4d58cd7f88 ("drm/bridge: adv7533: Create a MIPI DSI device") Cc: stable@vger.kernel.org Acked-by: Krzysztof Kozlowski Reviewed-by: Geert Uytterhoeven Reviewed-by: Laurent Pinchart Signed-off-by: Biju Das Link: https://patchwork.freedesktop.org/patch/msgid/20241119192040.152657-3-biju.das.jz@bp.renesas.com Signed-off-by: Dmitry Baryshkov --- .../devicetree/bindings/display/bridge/adi,adv7533.yaml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/devicetree/bindings/display/bridge/adi,adv7533.yaml b/Documentation/devicetree/bindings/display/bridge/adi,adv7533.yaml index df20a3c9c744..ec89115c74e4 100644 --- a/Documentation/devicetree/bindings/display/bridge/adi,adv7533.yaml +++ b/Documentation/devicetree/bindings/display/bridge/adi,adv7533.yaml @@ -90,7 +90,7 @@ properties: adi,dsi-lanes: description: Number of DSI data lanes connected to the DSI host. $ref: /schemas/types.yaml#/definitions/uint32 - enum: [ 1, 2, 3, 4 ] + enum: [ 2, 3, 4 ] "#sound-dai-cells": const: 0 From 79d67c499c3f886202a40c5cb27e747e4fa4d738 Mon Sep 17 00:00:00 2001 From: Biju Das Date: Tue, 19 Nov 2024 19:20:31 +0000 Subject: [PATCH 047/196] drm: adv7511: Drop dsi single lane support As per [1] and [2], ADV7535/7533 supports only 2-, 3-, or 4-lane. Drop unsupported 1-lane. [1] https://www.analog.com/media/en/technical-documentation/data-sheets/ADV7535.pdf [2] https://www.analog.com/media/en/technical-documentation/data-sheets/ADV7533.pdf Fixes: 1e4d58cd7f88 ("drm/bridge: adv7533: Create a MIPI DSI device") Reported-by: Hien Huynh Cc: stable@vger.kernel.org Reviewed-by: Laurent Pinchart Reviewed-by: Adam Ford Signed-off-by: Biju Das Link: https://patchwork.freedesktop.org/patch/msgid/20241119192040.152657-4-biju.das.jz@bp.renesas.com Signed-off-by: Dmitry Baryshkov --- drivers/gpu/drm/bridge/adv7511/adv7533.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/bridge/adv7511/adv7533.c b/drivers/gpu/drm/bridge/adv7511/adv7533.c index 5f195e91b3e6..122ad91e8a32 100644 --- a/drivers/gpu/drm/bridge/adv7511/adv7533.c +++ b/drivers/gpu/drm/bridge/adv7511/adv7533.c @@ -172,7 +172,7 @@ int adv7533_parse_dt(struct device_node *np, struct adv7511 *adv) of_property_read_u32(np, "adi,dsi-lanes", &num_lanes); - if (num_lanes < 1 || num_lanes > 4) + if (num_lanes < 2 || num_lanes > 4) return -EINVAL; adv->num_dsi_lanes = num_lanes; From 262bfba8ab820641c8cfbbf03b86d6c00242c078 Mon Sep 17 00:00:00 2001 From: Tristram Ha Date: Tue, 17 Dec 2024 18:02:23 -0800 Subject: [PATCH 048/196] net: dsa: microchip: Fix KSZ9477 set_ageing_time function The aging count is not a simple 11-bit value but comprises a 3-bit multiplier and an 8-bit second count. The code tries to use the original multiplier which is 4 as the second count is still 300 seconds by default. Fixes: 2c119d9982b1 ("net: dsa: microchip: add the support for set_ageing_time") Signed-off-by: Tristram Ha Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20241218020224.70590-2-Tristram.Ha@microchip.com Signed-off-by: Jakub Kicinski --- drivers/net/dsa/microchip/ksz9477.c | 47 +++++++++++++++++++------ drivers/net/dsa/microchip/ksz9477_reg.h | 4 +-- 2 files changed, 37 insertions(+), 14 deletions(-) diff --git a/drivers/net/dsa/microchip/ksz9477.c b/drivers/net/dsa/microchip/ksz9477.c index d16817e0476f..29fe79ea74cd 100644 --- a/drivers/net/dsa/microchip/ksz9477.c +++ b/drivers/net/dsa/microchip/ksz9477.c @@ -2,7 +2,7 @@ /* * Microchip KSZ9477 switch driver main logic * - * Copyright (C) 2017-2019 Microchip Technology Inc. + * Copyright (C) 2017-2024 Microchip Technology Inc. */ #include @@ -983,26 +983,51 @@ void ksz9477_get_caps(struct ksz_device *dev, int port, int ksz9477_set_ageing_time(struct ksz_device *dev, unsigned int msecs) { u32 secs = msecs / 1000; - u8 value; - u8 data; + u8 data, mult, value; + u32 max_val; int ret; - value = FIELD_GET(SW_AGE_PERIOD_7_0_M, secs); +#define MAX_TIMER_VAL ((1 << 8) - 1) - ret = ksz_write8(dev, REG_SW_LUE_CTRL_3, value); - if (ret < 0) - return ret; + /* The aging timer comprises a 3-bit multiplier and an 8-bit second + * value. Either of them cannot be zero. The maximum timer is then + * 7 * 255 = 1785 seconds. + */ + if (!secs) + secs = 1; - data = FIELD_GET(SW_AGE_PERIOD_10_8_M, secs); + /* Return error if too large. */ + else if (secs > 7 * MAX_TIMER_VAL) + return -EINVAL; ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value); if (ret < 0) return ret; - value &= ~SW_AGE_CNT_M; - value |= FIELD_PREP(SW_AGE_CNT_M, data); + /* Check whether there is need to update the multiplier. */ + mult = FIELD_GET(SW_AGE_CNT_M, value); + max_val = MAX_TIMER_VAL; + if (mult > 0) { + /* Try to use the same multiplier already in the register as + * the hardware default uses multiplier 4 and 75 seconds for + * 300 seconds. + */ + max_val = DIV_ROUND_UP(secs, mult); + if (max_val > MAX_TIMER_VAL || max_val * mult != secs) + max_val = MAX_TIMER_VAL; + } - return ksz_write8(dev, REG_SW_LUE_CTRL_0, value); + data = DIV_ROUND_UP(secs, max_val); + if (mult != data) { + value &= ~SW_AGE_CNT_M; + value |= FIELD_PREP(SW_AGE_CNT_M, data); + ret = ksz_write8(dev, REG_SW_LUE_CTRL_0, value); + if (ret < 0) + return ret; + } + + value = DIV_ROUND_UP(secs, data); + return ksz_write8(dev, REG_SW_LUE_CTRL_3, value); } void ksz9477_port_queue_split(struct ksz_device *dev, int port) diff --git a/drivers/net/dsa/microchip/ksz9477_reg.h b/drivers/net/dsa/microchip/ksz9477_reg.h index 04235c22bf40..ff579920078e 100644 --- a/drivers/net/dsa/microchip/ksz9477_reg.h +++ b/drivers/net/dsa/microchip/ksz9477_reg.h @@ -2,7 +2,7 @@ /* * Microchip KSZ9477 register definitions * - * Copyright (C) 2017-2018 Microchip Technology Inc. + * Copyright (C) 2017-2024 Microchip Technology Inc. */ #ifndef __KSZ9477_REGS_H @@ -165,8 +165,6 @@ #define SW_VLAN_ENABLE BIT(7) #define SW_DROP_INVALID_VID BIT(6) #define SW_AGE_CNT_M GENMASK(5, 3) -#define SW_AGE_CNT_S 3 -#define SW_AGE_PERIOD_10_8_M GENMASK(10, 8) #define SW_RESV_MCAST_ENABLE BIT(2) #define SW_HASH_OPTION_M 0x03 #define SW_HASH_OPTION_CRC 1 From bb9869043438af5b94230f94fb4c39206525d758 Mon Sep 17 00:00:00 2001 From: Tristram Ha Date: Tue, 17 Dec 2024 18:02:24 -0800 Subject: [PATCH 049/196] net: dsa: microchip: Fix LAN937X set_ageing_time function The aging count is not a simple 20-bit value but comprises a 3-bit multiplier and a 20-bit second time. The code tries to use the original multiplier which is 4 as the second count is still 300 seconds by default. As the 20-bit number is now too large for practical use there is an option to interpret it as microseconds instead of seconds. Fixes: 2c119d9982b1 ("net: dsa: microchip: add the support for set_ageing_time") Signed-off-by: Tristram Ha Reviewed-by: Andrew Lunn Link: https://patch.msgid.link/20241218020224.70590-3-Tristram.Ha@microchip.com Signed-off-by: Jakub Kicinski --- drivers/net/dsa/microchip/lan937x_main.c | 62 ++++++++++++++++++++++-- drivers/net/dsa/microchip/lan937x_reg.h | 9 ++-- 2 files changed, 65 insertions(+), 6 deletions(-) diff --git a/drivers/net/dsa/microchip/lan937x_main.c b/drivers/net/dsa/microchip/lan937x_main.c index b7652efd632e..b1ae3b9de3d1 100644 --- a/drivers/net/dsa/microchip/lan937x_main.c +++ b/drivers/net/dsa/microchip/lan937x_main.c @@ -1,6 +1,6 @@ // SPDX-License-Identifier: GPL-2.0 /* Microchip LAN937X switch driver main logic - * Copyright (C) 2019-2022 Microchip Technology Inc. + * Copyright (C) 2019-2024 Microchip Technology Inc. */ #include #include @@ -461,10 +461,66 @@ int lan937x_change_mtu(struct ksz_device *dev, int port, int new_mtu) int lan937x_set_ageing_time(struct ksz_device *dev, unsigned int msecs) { - u32 secs = msecs / 1000; - u32 value; + u8 data, mult, value8; + bool in_msec = false; + u32 max_val, value; + u32 secs = msecs; int ret; +#define MAX_TIMER_VAL ((1 << 20) - 1) + + /* The aging timer comprises a 3-bit multiplier and a 20-bit second + * value. Either of them cannot be zero. The maximum timer is then + * 7 * 1048575 = 7340025 seconds. As this value is too large for + * practical use it can be interpreted as microseconds, making the + * maximum timer 7340 seconds with finer control. This allows for + * maximum 122 minutes compared to 29 minutes in KSZ9477 switch. + */ + if (msecs % 1000) + in_msec = true; + else + secs /= 1000; + if (!secs) + secs = 1; + + /* Return error if too large. */ + else if (secs > 7 * MAX_TIMER_VAL) + return -EINVAL; + + /* Configure how to interpret the number value. */ + ret = ksz_rmw8(dev, REG_SW_LUE_CTRL_2, SW_AGE_CNT_IN_MICROSEC, + in_msec ? SW_AGE_CNT_IN_MICROSEC : 0); + if (ret < 0) + return ret; + + ret = ksz_read8(dev, REG_SW_LUE_CTRL_0, &value8); + if (ret < 0) + return ret; + + /* Check whether there is need to update the multiplier. */ + mult = FIELD_GET(SW_AGE_CNT_M, value8); + max_val = MAX_TIMER_VAL; + if (mult > 0) { + /* Try to use the same multiplier already in the register as + * the hardware default uses multiplier 4 and 75 seconds for + * 300 seconds. + */ + max_val = DIV_ROUND_UP(secs, mult); + if (max_val > MAX_TIMER_VAL || max_val * mult != secs) + max_val = MAX_TIMER_VAL; + } + + data = DIV_ROUND_UP(secs, max_val); + if (mult != data) { + value8 &= ~SW_AGE_CNT_M; + value8 |= FIELD_PREP(SW_AGE_CNT_M, data); + ret = ksz_write8(dev, REG_SW_LUE_CTRL_0, value8); + if (ret < 0) + return ret; + } + + secs = DIV_ROUND_UP(secs, data); + value = FIELD_GET(SW_AGE_PERIOD_7_0_M, secs); ret = ksz_write8(dev, REG_SW_AGE_PERIOD__1, value); diff --git a/drivers/net/dsa/microchip/lan937x_reg.h b/drivers/net/dsa/microchip/lan937x_reg.h index 4ec93e421da4..72042fd64e5b 100644 --- a/drivers/net/dsa/microchip/lan937x_reg.h +++ b/drivers/net/dsa/microchip/lan937x_reg.h @@ -1,6 +1,6 @@ /* SPDX-License-Identifier: GPL-2.0 */ /* Microchip LAN937X switch register definitions - * Copyright (C) 2019-2021 Microchip Technology Inc. + * Copyright (C) 2019-2024 Microchip Technology Inc. */ #ifndef __LAN937X_REG_H #define __LAN937X_REG_H @@ -56,8 +56,7 @@ #define SW_VLAN_ENABLE BIT(7) #define SW_DROP_INVALID_VID BIT(6) -#define SW_AGE_CNT_M 0x7 -#define SW_AGE_CNT_S 3 +#define SW_AGE_CNT_M GENMASK(5, 3) #define SW_RESV_MCAST_ENABLE BIT(2) #define REG_SW_LUE_CTRL_1 0x0311 @@ -70,6 +69,10 @@ #define SW_FAST_AGING BIT(1) #define SW_LINK_AUTO_AGING BIT(0) +#define REG_SW_LUE_CTRL_2 0x0312 + +#define SW_AGE_CNT_IN_MICROSEC BIT(7) + #define REG_SW_AGE_PERIOD__1 0x0313 #define SW_AGE_PERIOD_7_0_M GENMASK(7, 0) From 6321f5fb70d502d95de8a212a7b484c297ec9644 Mon Sep 17 00:00:00 2001 From: Joshua Washington Date: Wed, 18 Dec 2024 05:34:11 -0800 Subject: [PATCH 050/196] gve: clean XDP queues in gve_tx_stop_ring_gqi When stopping XDP TX rings, the XDP clean function needs to be called to clean out the entire queue, similar to what happens in the normal TX queue case. Otherwise, the FIFO won't be cleared correctly, and xsk_tx_completed won't be reported. Fixes: 75eaae158b1b ("gve: Add XDP DROP and TX support for GQI-QPL format") Cc: stable@vger.kernel.org Signed-off-by: Joshua Washington Signed-off-by: Praveen Kaligineedi Reviewed-by: Praveen Kaligineedi Reviewed-by: Willem de Bruijn Signed-off-by: David S. Miller --- drivers/net/ethernet/google/gve/gve_tx.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/google/gve/gve_tx.c b/drivers/net/ethernet/google/gve/gve_tx.c index e7fb7d6d283d..83ad278ec91f 100644 --- a/drivers/net/ethernet/google/gve/gve_tx.c +++ b/drivers/net/ethernet/google/gve/gve_tx.c @@ -206,7 +206,10 @@ void gve_tx_stop_ring_gqi(struct gve_priv *priv, int idx) return; gve_remove_napi(priv, ntfy_idx); - gve_clean_tx_done(priv, tx, priv->tx_desc_cnt, false); + if (tx->q_num < priv->tx_cfg.num_queues) + gve_clean_tx_done(priv, tx, priv->tx_desc_cnt, false); + else + gve_clean_xdp_done(priv, tx, priv->tx_desc_cnt); netdev_tx_reset_queue(tx->netdev_txq); gve_tx_remove_from_block(priv, idx); } From ff7c2dea9dd1a436fc79d6273adffdcc4a7ffea3 Mon Sep 17 00:00:00 2001 From: Joshua Washington Date: Wed, 18 Dec 2024 05:34:12 -0800 Subject: [PATCH 051/196] gve: guard XDP xmit NDO on existence of xdp queues In GVE, dedicated XDP queues only exist when an XDP program is installed and the interface is up. As such, the NDO XDP XMIT callback should return early if either of these conditions are false. In the case of no loaded XDP program, priv->num_xdp_queues=0 which can cause a divide-by-zero error, and in the case of interface down, num_xdp_queues remains untouched to persist XDP queue count for the next interface up, but the TX pointer itself would be NULL. The XDP xmit callback also needs to synchronize with a device transitioning from open to close. This synchronization will happen via the GVE_PRIV_FLAGS_NAPI_ENABLED bit along with a synchronize_net() call, which waits for any RCU critical sections at call-time to complete. Fixes: 39a7f4aa3e4a ("gve: Add XDP REDIRECT support for GQI-QPL format") Cc: stable@vger.kernel.org Signed-off-by: Joshua Washington Signed-off-by: Praveen Kaligineedi Reviewed-by: Praveen Kaligineedi Reviewed-by: Shailend Chand Reviewed-by: Willem de Bruijn Signed-off-by: David S. Miller --- drivers/net/ethernet/google/gve/gve_main.c | 3 +++ drivers/net/ethernet/google/gve/gve_tx.c | 5 ++++- 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index e171ca248f9a..5d7b0cc59959 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -1899,6 +1899,9 @@ static void gve_turndown(struct gve_priv *priv) gve_clear_napi_enabled(priv); gve_clear_report_stats(priv); + + /* Make sure that all traffic is finished processing. */ + synchronize_net(); } static void gve_turnup(struct gve_priv *priv) diff --git a/drivers/net/ethernet/google/gve/gve_tx.c b/drivers/net/ethernet/google/gve/gve_tx.c index 83ad278ec91f..852f8c7e39d2 100644 --- a/drivers/net/ethernet/google/gve/gve_tx.c +++ b/drivers/net/ethernet/google/gve/gve_tx.c @@ -837,9 +837,12 @@ int gve_xdp_xmit(struct net_device *dev, int n, struct xdp_frame **frames, struct gve_tx_ring *tx; int i, err = 0, qid; - if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK)) + if (unlikely(flags & ~XDP_XMIT_FLAGS_MASK) || !priv->xdp_prog) return -EINVAL; + if (!gve_get_napi_enabled(priv)) + return -ENETDOWN; + qid = gve_xdp_tx_queue_id(priv, smp_processor_id() % priv->num_xdp_queues); From 40338d7987d810fcaa95c500b1068a52b08eec9b Mon Sep 17 00:00:00 2001 From: Joshua Washington Date: Wed, 18 Dec 2024 05:34:13 -0800 Subject: [PATCH 052/196] gve: guard XSK operations on the existence of queues This patch predicates the enabling and disabling of XSK pools on the existence of queues. As it stands, if the interface is down, disabling or enabling XSK pools would result in a crash, as the RX queue pointer would be NULL. XSK pool registration will occur as part of the next interface up. Similarly, xsk_wakeup needs be guarded against queues disappearing while the function is executing, so a check against the GVE_PRIV_FLAGS_NAPI_ENABLED flag is added to synchronize with the disabling of the bit and the synchronize_net() in gve_turndown. Fixes: fd8e40321a12 ("gve: Add AF_XDP zero-copy support for GQI-QPL format") Cc: stable@vger.kernel.org Signed-off-by: Joshua Washington Signed-off-by: Praveen Kaligineedi Reviewed-by: Praveen Kaligineedi Reviewed-by: Shailend Chand Reviewed-by: Willem de Bruijn Reviewed-by: Larysa Zaremba Signed-off-by: David S. Miller --- drivers/net/ethernet/google/gve/gve_main.c | 22 ++++++++++------------ 1 file changed, 10 insertions(+), 12 deletions(-) diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index 5d7b0cc59959..e4e8ff4f9f80 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -1623,8 +1623,8 @@ static int gve_xsk_pool_enable(struct net_device *dev, if (err) return err; - /* If XDP prog is not installed, return */ - if (!priv->xdp_prog) + /* If XDP prog is not installed or interface is down, return. */ + if (!priv->xdp_prog || !netif_running(dev)) return 0; rx = &priv->rx[qid]; @@ -1669,21 +1669,16 @@ static int gve_xsk_pool_disable(struct net_device *dev, if (qid >= priv->rx_cfg.num_queues) return -EINVAL; - /* If XDP prog is not installed, unmap DMA and return */ - if (!priv->xdp_prog) + /* If XDP prog is not installed or interface is down, unmap DMA and + * return. + */ + if (!priv->xdp_prog || !netif_running(dev)) goto done; - tx_qid = gve_xdp_tx_queue_id(priv, qid); - if (!netif_running(dev)) { - priv->rx[qid].xsk_pool = NULL; - xdp_rxq_info_unreg(&priv->rx[qid].xsk_rxq); - priv->tx[tx_qid].xsk_pool = NULL; - goto done; - } - napi_rx = &priv->ntfy_blocks[priv->rx[qid].ntfy_id].napi; napi_disable(napi_rx); /* make sure current rx poll is done */ + tx_qid = gve_xdp_tx_queue_id(priv, qid); napi_tx = &priv->ntfy_blocks[priv->tx[tx_qid].ntfy_id].napi; napi_disable(napi_tx); /* make sure current tx poll is done */ @@ -1711,6 +1706,9 @@ static int gve_xsk_wakeup(struct net_device *dev, u32 queue_id, u32 flags) struct gve_priv *priv = netdev_priv(dev); int tx_queue_id = gve_xdp_tx_queue_id(priv, queue_id); + if (!gve_get_napi_enabled(priv)) + return -ENETDOWN; + if (queue_id >= priv->rx_cfg.num_queues || !priv->xdp_prog) return -EINVAL; From ba0925c34e0fa6fe02d3d642bc02ab099ab312c7 Mon Sep 17 00:00:00 2001 From: Joshua Washington Date: Wed, 18 Dec 2024 05:34:14 -0800 Subject: [PATCH 053/196] gve: process XSK TX descriptors as part of RX NAPI When busy polling is enabled, xsk_sendmsg for AF_XDP zero copy marks the NAPI ID corresponding to the memory pool allocated for the socket. In GVE, this NAPI ID will never correspond to a NAPI ID of one of the dedicated XDP TX queues registered with the umem because XDP TX is not set up to share a NAPI with a corresponding RX queue. This patch moves XSK TX descriptor processing from the TX NAPI to the RX NAPI, and the gve_xsk_wakeup callback is updated to use the RX NAPI instead of the TX NAPI, accordingly. The branch on if the wakeup is for TX is removed, as the NAPI poll should be invoked whether the wakeup is for TX or for RX. Fixes: fd8e40321a12 ("gve: Add AF_XDP zero-copy support for GQI-QPL format") Cc: stable@vger.kernel.org Signed-off-by: Praveen Kaligineedi Signed-off-by: Joshua Washington Reviewed-by: Willem de Bruijn Signed-off-by: David S. Miller --- drivers/net/ethernet/google/gve/gve.h | 1 + drivers/net/ethernet/google/gve/gve_main.c | 8 +++++ drivers/net/ethernet/google/gve/gve_tx.c | 36 +++++++++++++--------- 3 files changed, 31 insertions(+), 14 deletions(-) diff --git a/drivers/net/ethernet/google/gve/gve.h b/drivers/net/ethernet/google/gve/gve.h index dd92949bb214..8167cc5fb0df 100644 --- a/drivers/net/ethernet/google/gve/gve.h +++ b/drivers/net/ethernet/google/gve/gve.h @@ -1140,6 +1140,7 @@ int gve_xdp_xmit_one(struct gve_priv *priv, struct gve_tx_ring *tx, void gve_xdp_tx_flush(struct gve_priv *priv, u32 xdp_qid); bool gve_tx_poll(struct gve_notify_block *block, int budget); bool gve_xdp_poll(struct gve_notify_block *block, int budget); +int gve_xsk_tx_poll(struct gve_notify_block *block, int budget); int gve_tx_alloc_rings_gqi(struct gve_priv *priv, struct gve_tx_alloc_rings_cfg *cfg); void gve_tx_free_rings_gqi(struct gve_priv *priv, diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index e4e8ff4f9f80..5cab7b88610f 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -333,6 +333,14 @@ int gve_napi_poll(struct napi_struct *napi, int budget) if (block->rx) { work_done = gve_rx_poll(block, budget); + + /* Poll XSK TX as part of RX NAPI. Setup re-poll based on max of + * TX and RX work done. + */ + if (priv->xdp_prog) + work_done = max_t(int, work_done, + gve_xsk_tx_poll(block, budget)); + reschedule |= work_done == budget; } diff --git a/drivers/net/ethernet/google/gve/gve_tx.c b/drivers/net/ethernet/google/gve/gve_tx.c index 852f8c7e39d2..4350ebd9c2bd 100644 --- a/drivers/net/ethernet/google/gve/gve_tx.c +++ b/drivers/net/ethernet/google/gve/gve_tx.c @@ -981,33 +981,41 @@ static int gve_xsk_tx(struct gve_priv *priv, struct gve_tx_ring *tx, return sent; } +int gve_xsk_tx_poll(struct gve_notify_block *rx_block, int budget) +{ + struct gve_rx_ring *rx = rx_block->rx; + struct gve_priv *priv = rx->gve; + struct gve_tx_ring *tx; + int sent = 0; + + tx = &priv->tx[gve_xdp_tx_queue_id(priv, rx->q_num)]; + if (tx->xsk_pool) { + sent = gve_xsk_tx(priv, tx, budget); + + u64_stats_update_begin(&tx->statss); + tx->xdp_xsk_sent += sent; + u64_stats_update_end(&tx->statss); + if (xsk_uses_need_wakeup(tx->xsk_pool)) + xsk_set_tx_need_wakeup(tx->xsk_pool); + } + + return sent; +} + bool gve_xdp_poll(struct gve_notify_block *block, int budget) { struct gve_priv *priv = block->priv; struct gve_tx_ring *tx = block->tx; u32 nic_done; - bool repoll; u32 to_do; /* Find out how much work there is to be done */ nic_done = gve_tx_load_event_counter(priv, tx); to_do = min_t(u32, (nic_done - tx->done), budget); gve_clean_xdp_done(priv, tx, to_do); - repoll = nic_done != tx->done; - - if (tx->xsk_pool) { - int sent = gve_xsk_tx(priv, tx, budget); - - u64_stats_update_begin(&tx->statss); - tx->xdp_xsk_sent += sent; - u64_stats_update_end(&tx->statss); - repoll |= (sent == budget); - if (xsk_uses_need_wakeup(tx->xsk_pool)) - xsk_set_tx_need_wakeup(tx->xsk_pool); - } /* If we still have work we want to repoll */ - return repoll; + return nic_done != tx->done; } bool gve_tx_poll(struct gve_notify_block *block, int budget) From de63ac44a527b2c5067551dbd70d939fe151325a Mon Sep 17 00:00:00 2001 From: Joshua Washington Date: Wed, 18 Dec 2024 05:34:15 -0800 Subject: [PATCH 054/196] gve: fix XDP allocation path in edge cases This patch fixes a number of consistency issues in the queue allocation path related to XDP. As it stands, the number of allocated XDP queues changes in three different scenarios. 1) Adding an XDP program while the interface is up via gve_add_xdp_queues 2) Removing an XDP program while the interface is up via gve_remove_xdp_queues 3) After queues have been allocated and the old queue memory has been removed in gve_queues_start. However, the requirement for the interface to be up for gve_(add|remove)_xdp_queues to be called, in conjunction with the fact that the number of queues stored in priv isn't updated until _after_ XDP queues have been allocated in the normal queue allocation path means that if an XDP program is added while the interface is down, XDP queues won't be added until the _second_ if_up, not the first. Given the expectation that the number of XDP queues is equal to the number of RX queues, scenario (3) has another problematic implication. When changing the number of queues while an XDP program is loaded, the number of XDP queues must be updated as well, as there is logic in the driver (gve_xdp_tx_queue_id()) which relies on every RX queue having a corresponding XDP TX queue. However, the number of XDP queues stored in priv would not be updated until _after_ a close/open leading to a mismatch in the number of XDP queues reported vs the number of XDP queues which actually exist after the queue count update completes. This patch remedies these issues by doing the following: 1) The allocation config getter function is set up to retrieve the _expected_ number of XDP queues to allocate instead of relying on the value stored in `priv` which is only updated once the queues have been allocated. 2) When adjusting queues, XDP queues are adjusted to match the number of RX queues when XDP is enabled. This only works in the case when queues are live, so part (1) of the fix must still be available in the case that queues are adjusted when there is an XDP program and the interface is down. Fixes: 5f08cd3d6423 ("gve: Alloc before freeing when adjusting queues") Cc: stable@vger.kernel.org Signed-off-by: Joshua Washington Signed-off-by: Praveen Kaligineedi Reviewed-by: Praveen Kaligineedi Reviewed-by: Shailend Chand Reviewed-by: Willem de Bruijn Signed-off-by: David S. Miller --- drivers/net/ethernet/google/gve/gve_main.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index 5cab7b88610f..09fb7f16f73e 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -930,11 +930,13 @@ static void gve_init_sync_stats(struct gve_priv *priv) static void gve_tx_get_curr_alloc_cfg(struct gve_priv *priv, struct gve_tx_alloc_rings_cfg *cfg) { + int num_xdp_queues = priv->xdp_prog ? priv->rx_cfg.num_queues : 0; + cfg->qcfg = &priv->tx_cfg; cfg->raw_addressing = !gve_is_qpl(priv); cfg->ring_size = priv->tx_desc_cnt; cfg->start_idx = 0; - cfg->num_rings = gve_num_tx_queues(priv); + cfg->num_rings = priv->tx_cfg.num_queues + num_xdp_queues; cfg->tx = priv->tx; } @@ -1843,6 +1845,7 @@ int gve_adjust_queues(struct gve_priv *priv, { struct gve_tx_alloc_rings_cfg tx_alloc_cfg = {0}; struct gve_rx_alloc_rings_cfg rx_alloc_cfg = {0}; + int num_xdp_queues; int err; gve_get_curr_alloc_cfgs(priv, &tx_alloc_cfg, &rx_alloc_cfg); @@ -1853,6 +1856,10 @@ int gve_adjust_queues(struct gve_priv *priv, rx_alloc_cfg.qcfg = &new_rx_config; tx_alloc_cfg.num_rings = new_tx_config.num_queues; + /* Add dedicated XDP TX queues if enabled. */ + num_xdp_queues = priv->xdp_prog ? new_rx_config.num_queues : 0; + tx_alloc_cfg.num_rings += num_xdp_queues; + if (netif_running(priv->dev)) { err = gve_adjust_config(priv, &tx_alloc_cfg, &rx_alloc_cfg); return err; From 246068b86b1c36e4590388ab8f278e21f1997dc1 Mon Sep 17 00:00:00 2001 From: Vladimir Oltean Date: Thu, 19 Dec 2024 17:54:10 +0200 Subject: [PATCH 055/196] selftests: net: local_termination: require mausezahn Since the blamed commit, we require mausezahn because send_raw() uses it. Remove the "REQUIRE_MZ=no" line, which overwrites the default of requiring it. Fixes: 237979504264 ("selftests: net: local_termination: add PTP frames to the mix") Signed-off-by: Vladimir Oltean Link: https://patch.msgid.link/20241219155410.1856868-1-vladimir.oltean@nxp.com Signed-off-by: Jakub Kicinski --- tools/testing/selftests/net/forwarding/local_termination.sh | 1 - 1 file changed, 1 deletion(-) diff --git a/tools/testing/selftests/net/forwarding/local_termination.sh b/tools/testing/selftests/net/forwarding/local_termination.sh index c35548767756..ecd34f364125 100755 --- a/tools/testing/selftests/net/forwarding/local_termination.sh +++ b/tools/testing/selftests/net/forwarding/local_termination.sh @@ -7,7 +7,6 @@ ALL_TESTS="standalone vlan_unaware_bridge vlan_aware_bridge test_vlan \ NUM_NETIFS=2 PING_COUNT=1 REQUIRE_MTOOLS=yes -REQUIRE_MZ=no source lib.sh From 4a25201aa46ce88e8e31f9ccdec0e4e3dd6bb736 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Wed, 18 Dec 2024 19:28:32 -0800 Subject: [PATCH 056/196] netdev-genl: avoid empty messages in napi get Empty netlink responses from do() are not correct (as opposed to dump() where not dumping anything is perfectly fine). We should return an error if the target object does not exist, in this case if the netdev is down we "hide" the NAPI instances. Fixes: 27f91aaf49b3 ("netdev-genl: Add netlink framework functions for napi") Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/20241219032833.1165433-1-kuba@kernel.org Signed-off-by: Jakub Kicinski --- net/core/netdev-genl.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/net/core/netdev-genl.c b/net/core/netdev-genl.c index 2d3ae0cd3ad2..b0772d135efb 100644 --- a/net/core/netdev-genl.c +++ b/net/core/netdev-genl.c @@ -246,8 +246,12 @@ int netdev_nl_napi_get_doit(struct sk_buff *skb, struct genl_info *info) rcu_read_unlock(); rtnl_unlock(); - if (err) + if (err) { goto err_free_msg; + } else if (!rsp->len) { + err = -ENOENT; + goto err_free_msg; + } return genlmsg_reply(rsp, info); From 30b981796b94b083da8fdded7cb74cb493608760 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Wed, 18 Dec 2024 19:28:33 -0800 Subject: [PATCH 057/196] selftests: drv-net: test empty queue and NAPI responses in netlink Make sure kernel doesn't respond to GETs for queues and NAPIs when link is down. Not with valid data, or with empty message, we want a ENOENT. Link: https://patch.msgid.link/20241219032833.1165433-2-kuba@kernel.org Signed-off-by: Jakub Kicinski --- tools/testing/selftests/drivers/net/queues.py | 28 ++++++++++++++++--- 1 file changed, 24 insertions(+), 4 deletions(-) diff --git a/tools/testing/selftests/drivers/net/queues.py b/tools/testing/selftests/drivers/net/queues.py index 9c5473abbd78..38303da957ee 100755 --- a/tools/testing/selftests/drivers/net/queues.py +++ b/tools/testing/selftests/drivers/net/queues.py @@ -1,10 +1,12 @@ #!/usr/bin/env python3 # SPDX-License-Identifier: GPL-2.0 -from lib.py import ksft_run, ksft_exit, ksft_eq, KsftSkipEx -from lib.py import EthtoolFamily, NetdevFamily +from lib.py import ksft_disruptive, ksft_exit, ksft_run +from lib.py import ksft_eq, ksft_raises, KsftSkipEx +from lib.py import EthtoolFamily, NetdevFamily, NlError from lib.py import NetDrvEnv -from lib.py import cmd +from lib.py import cmd, defer, ip +import errno import glob @@ -59,9 +61,27 @@ def addremove_queues(cfg, nl) -> None: ksft_eq(queues, expected) +@ksft_disruptive +def check_down(cfg, nl) -> None: + # Check the NAPI IDs before interface goes down and hides them + napis = nl.napi_get({'ifindex': cfg.ifindex}, dump=True) + + ip(f"link set dev {cfg.dev['ifname']} down") + defer(ip, f"link set dev {cfg.dev['ifname']} up") + + with ksft_raises(NlError) as cm: + nl.queue_get({'ifindex': cfg.ifindex, 'id': 0, 'type': 'rx'}) + ksft_eq(cm.exception.nl_msg.error, -errno.ENOENT) + + if napis: + with ksft_raises(NlError) as cm: + nl.napi_get({'id': napis[0]['id']}) + ksft_eq(cm.exception.nl_msg.error, -errno.ENOENT) + + def main() -> None: with NetDrvEnv(__file__, queue_count=100) as cfg: - ksft_run([get_queues, addremove_queues], args=(cfg, NetdevFamily())) + ksft_run([get_queues, addremove_queues, check_down], args=(cfg, NetdevFamily())) ksft_exit() From c384481006476ac65478fa3584c7245782e52f34 Mon Sep 17 00:00:00 2001 From: Nikolaus Voss Date: Thu, 19 Dec 2024 11:54:11 +0100 Subject: [PATCH 058/196] clk: clk-imx8mp-audiomix: fix function signature clk_imx8mp_audiomix_reset_controller_register() in the "if !CONFIG_RESET_CONTROLLER" branch had the first argument missing. It is an empty function for this branch so it wasn't immediately apparent. Fixes: 6f0e817175c5 ("clk: imx: clk-audiomix: Add reset controller") Cc: # 6.12.x Signed-off-by: Nikolaus Voss Link: https://lore.kernel.org/r/20241219105447.889CB11FE@mail.steuer-voss.de Reviewed-by: Daniel Baluta Acked-by: Shengjiu Wang Reviewed-by: Peng Fan Signed-off-by: Stephen Boyd --- drivers/clk/imx/clk-imx8mp-audiomix.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c b/drivers/clk/imx/clk-imx8mp-audiomix.c index b2cb157703c5..c409fc7e0618 100644 --- a/drivers/clk/imx/clk-imx8mp-audiomix.c +++ b/drivers/clk/imx/clk-imx8mp-audiomix.c @@ -278,7 +278,8 @@ static int clk_imx8mp_audiomix_reset_controller_register(struct device *dev, #else /* !CONFIG_RESET_CONTROLLER */ -static int clk_imx8mp_audiomix_reset_controller_register(struct clk_imx8mp_audiomix_priv *priv) +static int clk_imx8mp_audiomix_reset_controller_register(struct device *dev, + struct clk_imx8mp_audiomix_priv *priv) { return 0; } From 8673a6c2d9e483dfeeef83a1f06f59e05636f4d1 Mon Sep 17 00:00:00 2001 From: wenglianfa Date: Fri, 20 Dec 2024 13:52:46 +0800 Subject: [PATCH 059/196] RDMA/hns: Fix mapping error of zero-hop WQE buffer Due to HW limitation, the three region of WQE buffer must be mapped and set to HW in a fixed order: SQ buffer, SGE buffer, and RQ buffer. Currently when one region is zero-hop while the other two are not, the zero-hop region will not be mapped. This violate the limitation above and leads to address error. Fixes: 38389eaa4db1 ("RDMA/hns: Add mtr support for mixed multihop addressing") Signed-off-by: wenglianfa Signed-off-by: Junxian Huang Link: https://patch.msgid.link/20241220055249.146943-2-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/hns/hns_roce_hem.c | 43 ++++++++++++++++-------- drivers/infiniband/hw/hns/hns_roce_mr.c | 5 --- 2 files changed, 29 insertions(+), 19 deletions(-) diff --git a/drivers/infiniband/hw/hns/hns_roce_hem.c b/drivers/infiniband/hw/hns/hns_roce_hem.c index f84521be3bea..605562122ecc 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hem.c +++ b/drivers/infiniband/hw/hns/hns_roce_hem.c @@ -931,6 +931,7 @@ struct hns_roce_hem_item { size_t count; /* max ba numbers */ int start; /* start buf offset in this hem */ int end; /* end buf offset in this hem */ + bool exist_bt; }; /* All HEM items are linked in a tree structure */ @@ -959,6 +960,7 @@ hem_list_alloc_item(struct hns_roce_dev *hr_dev, int start, int end, int count, } } + hem->exist_bt = exist_bt; hem->count = count; hem->start = start; hem->end = end; @@ -969,22 +971,22 @@ hem_list_alloc_item(struct hns_roce_dev *hr_dev, int start, int end, int count, } static void hem_list_free_item(struct hns_roce_dev *hr_dev, - struct hns_roce_hem_item *hem, bool exist_bt) + struct hns_roce_hem_item *hem) { - if (exist_bt) + if (hem->exist_bt) dma_free_coherent(hr_dev->dev, hem->count * BA_BYTE_LEN, hem->addr, hem->dma_addr); kfree(hem); } static void hem_list_free_all(struct hns_roce_dev *hr_dev, - struct list_head *head, bool exist_bt) + struct list_head *head) { struct hns_roce_hem_item *hem, *temp_hem; list_for_each_entry_safe(hem, temp_hem, head, list) { list_del(&hem->list); - hem_list_free_item(hr_dev, hem, exist_bt); + hem_list_free_item(hr_dev, hem); } } @@ -1084,6 +1086,10 @@ int hns_roce_hem_list_calc_root_ba(const struct hns_roce_buf_region *regions, for (i = 0; i < region_cnt; i++) { r = (struct hns_roce_buf_region *)®ions[i]; + /* when r->hopnum = 0, the region should not occupy root_ba. */ + if (!r->hopnum) + continue; + if (r->hopnum > 1) { step = hem_list_calc_ba_range(r->hopnum, 1, unit); if (step > 0) @@ -1177,7 +1183,7 @@ static int hem_list_alloc_mid_bt(struct hns_roce_dev *hr_dev, err_exit: for (level = 1; level < hopnum; level++) - hem_list_free_all(hr_dev, &temp_list[level], true); + hem_list_free_all(hr_dev, &temp_list[level]); return ret; } @@ -1218,16 +1224,26 @@ static int alloc_fake_root_bt(struct hns_roce_dev *hr_dev, void *cpu_base, { struct hns_roce_hem_item *hem; + /* This is on the has_mtt branch, if r->hopnum + * is 0, there is no root_ba to reuse for the + * region's fake hem, so a dma_alloc request is + * necessary here. + */ hem = hem_list_alloc_item(hr_dev, r->offset, r->offset + r->count - 1, - r->count, false); + r->count, !r->hopnum); if (!hem) return -ENOMEM; - hem_list_assign_bt(hem, cpu_base, phy_base); + /* The root_ba can be reused only when r->hopnum > 0. */ + if (r->hopnum) + hem_list_assign_bt(hem, cpu_base, phy_base); list_add(&hem->list, branch_head); list_add(&hem->sibling, leaf_head); - return r->count; + /* If r->hopnum == 0, 0 is returned, + * so that the root_bt entry is not occupied. + */ + return r->hopnum ? r->count : 0; } static int setup_middle_bt(struct hns_roce_dev *hr_dev, void *cpu_base, @@ -1271,7 +1287,7 @@ setup_root_hem(struct hns_roce_dev *hr_dev, struct hns_roce_hem_list *hem_list, return -ENOMEM; total = 0; - for (i = 0; i < region_cnt && total < max_ba_num; i++) { + for (i = 0; i < region_cnt && total <= max_ba_num; i++) { r = ®ions[i]; if (!r->count) continue; @@ -1337,9 +1353,9 @@ static int hem_list_alloc_root_bt(struct hns_roce_dev *hr_dev, region_cnt); if (ret) { for (i = 0; i < region_cnt; i++) - hem_list_free_all(hr_dev, &head.branch[i], false); + hem_list_free_all(hr_dev, &head.branch[i]); - hem_list_free_all(hr_dev, &head.root, true); + hem_list_free_all(hr_dev, &head.root); } return ret; @@ -1402,10 +1418,9 @@ void hns_roce_hem_list_release(struct hns_roce_dev *hr_dev, for (i = 0; i < HNS_ROCE_MAX_BT_REGION; i++) for (j = 0; j < HNS_ROCE_MAX_BT_LEVEL; j++) - hem_list_free_all(hr_dev, &hem_list->mid_bt[i][j], - j != 0); + hem_list_free_all(hr_dev, &hem_list->mid_bt[i][j]); - hem_list_free_all(hr_dev, &hem_list->root_bt, true); + hem_list_free_all(hr_dev, &hem_list->root_bt); INIT_LIST_HEAD(&hem_list->btm_bt); hem_list->root_ba = 0; } diff --git a/drivers/infiniband/hw/hns/hns_roce_mr.c b/drivers/infiniband/hw/hns/hns_roce_mr.c index bf30b3a65a9b..55b9283bfc6f 100644 --- a/drivers/infiniband/hw/hns/hns_roce_mr.c +++ b/drivers/infiniband/hw/hns/hns_roce_mr.c @@ -814,11 +814,6 @@ int hns_roce_mtr_map(struct hns_roce_dev *hr_dev, struct hns_roce_mtr *mtr, for (i = 0, mapped_cnt = 0; i < mtr->hem_cfg.region_count && mapped_cnt < page_cnt; i++) { r = &mtr->hem_cfg.region[i]; - /* if hopnum is 0, no need to map pages in this region */ - if (!r->hopnum) { - mapped_cnt += r->count; - continue; - } if (r->offset + r->count > page_cnt) { ret = -EINVAL; From 0572eccf239ce4bd89bd531767ec5ab20e249290 Mon Sep 17 00:00:00 2001 From: Chengchang Tang Date: Fri, 20 Dec 2024 13:52:47 +0800 Subject: [PATCH 060/196] RDMA/hns: Fix accessing invalid dip_ctx during destroying QP If it fails to modify QP to RTR, dip_ctx will not be attached. And during detroying QP, the invalid dip_ctx pointer will be accessed. Fixes: faa62440a577 ("RDMA/hns: Fix different dgids mapping to the same dip_idx") Signed-off-by: Chengchang Tang Signed-off-by: Junxian Huang Link: https://patch.msgid.link/20241220055249.146943-3-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c index 697b17cca02e..6dddadb90e02 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c @@ -5619,6 +5619,9 @@ static void put_dip_ctx_idx(struct hns_roce_dev *hr_dev, { struct hns_roce_dip *hr_dip = hr_qp->dip; + if (!hr_dip) + return; + xa_lock(&hr_dev->qp_table.dip_xa); hr_dip->qp_cnt--; From fa5c4ba8cdbfd2c2d6422e001311c8213283ebbf Mon Sep 17 00:00:00 2001 From: Chengchang Tang Date: Fri, 20 Dec 2024 13:52:48 +0800 Subject: [PATCH 061/196] RDMA/hns: Fix warning storm caused by invalid input in IO path WARN_ON() is called in the IO path. And it could lead to a warning storm. Use WARN_ON_ONCE() instead of WARN_ON(). Fixes: 12542f1de179 ("RDMA/hns: Refactor process about opcode in post_send()") Signed-off-by: Chengchang Tang Signed-off-by: Junxian Huang Link: https://patch.msgid.link/20241220055249.146943-4-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c index 6dddadb90e02..d0469d27c63c 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c @@ -468,7 +468,7 @@ static inline int set_ud_wqe(struct hns_roce_qp *qp, valid_num_sge = calc_wr_sge_num(wr, &msg_len); ret = set_ud_opcode(ud_sq_wqe, wr); - if (WARN_ON(ret)) + if (WARN_ON_ONCE(ret)) return ret; ud_sq_wqe->msg_len = cpu_to_le32(msg_len); @@ -572,7 +572,7 @@ static inline int set_rc_wqe(struct hns_roce_qp *qp, rc_sq_wqe->msg_len = cpu_to_le32(msg_len); ret = set_rc_opcode(hr_dev, rc_sq_wqe, wr); - if (WARN_ON(ret)) + if (WARN_ON_ONCE(ret)) return ret; hr_reg_write(rc_sq_wqe, RC_SEND_WQE_SO, From e3debdd48423d3d75b9d366399228d7225d902cd Mon Sep 17 00:00:00 2001 From: Chengchang Tang Date: Fri, 20 Dec 2024 13:52:49 +0800 Subject: [PATCH 062/196] RDMA/hns: Fix missing flush CQE for DWQE Flush CQE handler has not been called if QP state gets into errored mode in DWQE path. So, the new added outstanding WQEs will never be flushed. It leads to a hung task timeout when using NFS over RDMA: __switch_to+0x7c/0xd0 __schedule+0x350/0x750 schedule+0x50/0xf0 schedule_timeout+0x2c8/0x340 wait_for_common+0xf4/0x2b0 wait_for_completion+0x20/0x40 __ib_drain_sq+0x140/0x1d0 [ib_core] ib_drain_sq+0x98/0xb0 [ib_core] rpcrdma_xprt_disconnect+0x68/0x270 [rpcrdma] xprt_rdma_close+0x20/0x60 [rpcrdma] xprt_autoclose+0x64/0x1cc [sunrpc] process_one_work+0x1d8/0x4e0 worker_thread+0x154/0x420 kthread+0x108/0x150 ret_from_fork+0x10/0x18 Fixes: 01584a5edcc4 ("RDMA/hns: Add support of direct wqe") Signed-off-by: Chengchang Tang Signed-off-by: Junxian Huang Link: https://patch.msgid.link/20241220055249.146943-5-huangjunxian6@hisilicon.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c index d0469d27c63c..0144e7210d05 100644 --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c @@ -670,6 +670,10 @@ static void write_dwqe(struct hns_roce_dev *hr_dev, struct hns_roce_qp *qp, #define HNS_ROCE_SL_SHIFT 2 struct hns_roce_v2_rc_send_wqe *rc_sq_wqe = wqe; + if (unlikely(qp->state == IB_QPS_ERR)) { + flush_cqe(hr_dev, qp); + return; + } /* All kinds of DirectWQE have the same header field layout */ hr_reg_enable(rc_sq_wqe, RC_SEND_WQE_FLAG); hr_reg_write(rc_sq_wqe, RC_SEND_WQE_DB_SL_L, qp->sl); From a53da2fb25a31f4fb8eaeb93c7b1134fc14fd209 Mon Sep 17 00:00:00 2001 From: John Harrison Date: Fri, 13 Dec 2024 09:28:33 -0800 Subject: [PATCH 063/196] drm/xe: Revert some changes that break a mesa debug tool MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There is a mesa debug tool for decoding devcoredump files. Recent changes to improve the devcoredump output broke that tool. So revert the changes until the tool can be extended to support the new fields. Signed-off-by: John Harrison Fixes: c28fd6c358db ("drm/xe/devcoredump: Improve section headings and add tile info") Fixes: ec1455ce7e35 ("drm/xe/devcoredump: Add ASCII85 dump helper function") Cc: John Harrison Cc: Julia Filipchuk Cc: Lucas De Marchi Cc: Thomas Hellström Cc: Rodrigo Vivi Cc: intel-xe@lists.freedesktop.org Reviewed-by: Jonathan Cavitt Reviewed-by: José Roberto de Souza Link: https://patchwork.freedesktop.org/patch/msgid/20241213172833.1733376-1-John.C.Harrison@Intel.com Signed-off-by: Rodrigo Vivi (cherry picked from commit 70fb86a85dc9fd66014d7eb2fe356f50702ceeb6) Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_devcoredump.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_devcoredump.c b/drivers/gpu/drm/xe/xe_devcoredump.c index f8947e7e917e..21a50d539426 100644 --- a/drivers/gpu/drm/xe/xe_devcoredump.c +++ b/drivers/gpu/drm/xe/xe_devcoredump.c @@ -109,7 +109,11 @@ static ssize_t __xe_devcoredump_read(char *buffer, size_t count, drm_puts(&p, "\n**** GuC CT ****\n"); xe_guc_ct_snapshot_print(ss->guc.ct, &p); - drm_puts(&p, "\n**** Contexts ****\n"); + /* + * Don't add a new section header here because the mesa debug decoder + * tool expects the context information to be in the 'GuC CT' section. + */ + /* drm_puts(&p, "\n**** Contexts ****\n"); */ xe_guc_exec_queue_snapshot_print(ss->ge, &p); drm_puts(&p, "\n**** Job ****\n"); @@ -363,6 +367,15 @@ void xe_print_blob_ascii85(struct drm_printer *p, const char *prefix, char buff[ASCII85_BUFSZ], *line_buff; size_t line_pos = 0; + /* + * Splitting blobs across multiple lines is not compatible with the mesa + * debug decoder tool. Note that even dropping the explicit '\n' below + * doesn't help because the GuC log is so big some underlying implementation + * still splits the lines at 512K characters. So just bail completely for + * the moment. + */ + return; + #define DMESG_MAX_LINE_LEN 800 #define MIN_SPACE (ASCII85_BUFSZ + 2) /* 85 + "\n\0" */ From 528cef1b4170f328d28d4e9b437380d8e5a2d18f Mon Sep 17 00:00:00 2001 From: Nirmoy Das Date: Fri, 13 Dec 2024 13:24:14 +0100 Subject: [PATCH 064/196] drm/xe: Use non-interruptible wait when moving BO to system MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Ensure a non-interruptible wait is used when moving a bo to XE_PL_SYSTEM. This prevents dma_mappings from being removed prematurely while a GPU job is still in progress, even if the CPU receives a signal during the operation. Fixes: 75521e8b56e8 ("drm/xe: Perform dma_map when moving system buffer objects to TT") Cc: Thomas Hellström Cc: Matthew Brost Cc: Lucas De Marchi Cc: stable@vger.kernel.org # v6.11+ Suggested-by: Matthew Auld Reviewed-by: Matthew Auld Reviewed-by: Thomas Hellström Link: https://patchwork.freedesktop.org/patch/msgid/20241213122415.3880017-1-nirmoy.das@intel.com Signed-off-by: Nirmoy Das (cherry picked from commit dc5e20ae1f8a7c354dc9833faa2720254e5a5443) Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_bo.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index ae6b337cdc54..1aec4133008e 100644 --- a/drivers/gpu/drm/xe/xe_bo.c +++ b/drivers/gpu/drm/xe/xe_bo.c @@ -724,7 +724,7 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict, new_mem->mem_type == XE_PL_SYSTEM) { long timeout = dma_resv_wait_timeout(ttm_bo->base.resv, DMA_RESV_USAGE_BOOKKEEP, - true, + false, MAX_SCHEDULE_TIMEOUT); if (timeout < 0) { ret = timeout; From 5e0a67fdb894d34c5f109e969320eef9ddae7480 Mon Sep 17 00:00:00 2001 From: Nirmoy Das Date: Fri, 13 Dec 2024 13:24:15 +0100 Subject: [PATCH 065/196] drm/xe: Wait for migration job before unmapping pages MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix a potential GPU page fault during tt -> system moves by waiting for migration jobs to complete before unmapping SG. This ensures that IOMMU mappings are not prematurely torn down while a migration job is still in progress. v2: Use intr=false(Matt A) v3: Update commit message(Matt A) v4: s/DMA_RESV_USAGE_BOOKKEEP/DMA_RESV_USAGE_KERNEL(Thomas) Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/3466 Fixes: 75521e8b56e8 ("drm/xe: Perform dma_map when moving system buffer objects to TT") Cc: Thomas Hellström Cc: Matthew Brost Cc: Lucas De Marchi Cc: stable@vger.kernel.org # v6.11+ Cc: Matthew Auld Reviewed-by: Matthew Auld Reviewed-by: Thomas Hellström Link: https://patchwork.freedesktop.org/patch/msgid/20241213122415.3880017-2-nirmoy.das@intel.com Signed-off-by: Nirmoy Das (cherry picked from commit cda06412c06893a6f07a2fbf89d42a0972ec9e8e) Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_bo.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c index 1aec4133008e..f61a8ef38094 100644 --- a/drivers/gpu/drm/xe/xe_bo.c +++ b/drivers/gpu/drm/xe/xe_bo.c @@ -848,8 +848,16 @@ static int xe_bo_move(struct ttm_buffer_object *ttm_bo, bool evict, out: if ((!ttm_bo->resource || ttm_bo->resource->mem_type == XE_PL_SYSTEM) && - ttm_bo->ttm) + ttm_bo->ttm) { + long timeout = dma_resv_wait_timeout(ttm_bo->base.resv, + DMA_RESV_USAGE_KERNEL, + false, + MAX_SCHEDULE_TIMEOUT); + if (timeout < 0) + ret = timeout; + xe_tt_unmap_sg(ttm_bo->ttm); + } return ret; } From af12ba67d09ebe2b31ab997cea1a930864028562 Mon Sep 17 00:00:00 2001 From: Michal Wajdeczko Date: Mon, 16 Dec 2024 23:32:53 +0100 Subject: [PATCH 066/196] drm/xe/pf: Use correct function to check LMEM provisioning MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There is a typo in function call and instead of VF LMEM we were looking at VF GGTT provisioning. Fix that. Fixes: 234670cea9a2 ("drm/xe/pf: Skip fair VFs provisioning if already provisioned") Signed-off-by: Michal Wajdeczko Cc: Piotr Piórkowski Reviewed-by: Himal Prasad Ghimiray Link: https://patchwork.freedesktop.org/patch/msgid/20241216223253.819-1-michal.wajdeczko@intel.com (cherry picked from commit a8d0aa0e7fcd20c9f1992688c0f0d07a68287403) Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c index 192643d63d22..ca49860168f6 100644 --- a/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c +++ b/drivers/gpu/drm/xe/xe_gt_sriov_pf_config.c @@ -2046,7 +2046,7 @@ static int pf_validate_vf_config(struct xe_gt *gt, unsigned int vfid) valid_any = valid_any || (valid_ggtt && is_primary); if (IS_DGFX(xe)) { - bool valid_lmem = pf_get_vf_config_ggtt(primary_gt, vfid); + bool valid_lmem = pf_get_vf_config_lmem(primary_gt, vfid); valid_any = valid_any || (valid_lmem && is_primary); valid_all = valid_all && valid_lmem; From fe39b222a4139354d32ff9d46b88757f63f71d63 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Tue, 17 Dec 2024 21:31:21 -0800 Subject: [PATCH 067/196] drm/xe: Fix fault on fd close after unbind MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If userspace holds an fd open, unbinds the device and then closes it, the driver shouldn't try to access the hardware. Protect it by using drm_dev_enter()/drm_dev_exit(). This fixes the following page fault: <6> [IGT] xe_wedged: exiting, ret=98 <1> BUG: unable to handle page fault for address: ffffc901bc5e508c <1> #PF: supervisor read access in kernel mode <1> #PF: error_code(0x0000) - not-present page ... <4> xe_lrc_update_timestamp+0x1c/0xd0 [xe] <4> xe_exec_queue_update_run_ticks+0x50/0xb0 [xe] <4> xe_exec_queue_fini+0x16/0xb0 [xe] <4> __guc_exec_queue_fini_async+0xc4/0x190 [xe] <4> guc_exec_queue_fini_async+0xa0/0xe0 [xe] <4> guc_exec_queue_fini+0x23/0x40 [xe] <4> xe_exec_queue_destroy+0xb3/0xf0 [xe] <4> xe_file_close+0xd4/0x1a0 [xe] <4> drm_file_free+0x210/0x280 [drm] <4> drm_close_helper.isra.0+0x6d/0x80 [drm] <4> drm_release_noglobal+0x20/0x90 [drm] Fixes: 514447a12190 ("drm/xe: Stop accumulating LRC timestamp on job_free") Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/3421 Reviewed-by: Umesh Nerlige Ramappa Link: https://patchwork.freedesktop.org/patch/msgid/20241218053122.2730195-1-lucas.demarchi@intel.com Signed-off-by: Lucas De Marchi (cherry picked from commit 4ca1fd418338d4d135428a0eb1e16e3b3ce17ee8) Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_exec_queue.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c index fd0f3b3c9101..268cd3123be9 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue.c +++ b/drivers/gpu/drm/xe/xe_exec_queue.c @@ -8,6 +8,7 @@ #include #include +#include #include #include @@ -762,9 +763,11 @@ bool xe_exec_queue_is_idle(struct xe_exec_queue *q) */ void xe_exec_queue_update_run_ticks(struct xe_exec_queue *q) { + struct xe_device *xe = gt_to_xe(q->gt); struct xe_file *xef; struct xe_lrc *lrc; u32 old_ts, new_ts; + int idx; /* * Jobs that are run during driver load may use an exec_queue, but are @@ -774,6 +777,10 @@ void xe_exec_queue_update_run_ticks(struct xe_exec_queue *q) if (!q->vm || !q->vm->xef) return; + /* Synchronize with unbind while holding the xe file open */ + if (!drm_dev_enter(&xe->drm, &idx)) + return; + xef = q->vm->xef; /* @@ -787,6 +794,8 @@ void xe_exec_queue_update_run_ticks(struct xe_exec_queue *q) lrc = q->lrc[0]; new_ts = xe_lrc_update_timestamp(lrc, &old_ts); xef->run_ticks[q->class] += (new_ts - old_ts) * q->width; + + drm_dev_exit(idx); } /** From a072ffd896efa6a6c8a0334c712fbc98a63c789c Mon Sep 17 00:00:00 2001 From: Mohsin Bashir Date: Wed, 18 Dec 2024 15:25:58 -0800 Subject: [PATCH 068/196] eth: fbnic: fix csr boundary for RPM RAM section MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The CSR dump support leverages the FBNIC_BOUNDS macro, which pads the end condition for each section by adding an offset of 1. However, the RPC RAM section, which is dumped differently from other sections, does not rely on this macro and instead directly uses end boundary address. Hence, subtracting 1 from the end address results in skipping a register. Fixes 3d12862b216d (“eth: fbnic: Add support to dump registers”) Signed-off-by: Mohsin Bashir Reviewed-by: Michal Swiatkowski Link: https://patch.msgid.link/20241218232614.439329-1-mohsin.bashr@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/meta/fbnic/fbnic_csr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/meta/fbnic/fbnic_csr.c b/drivers/net/ethernet/meta/fbnic/fbnic_csr.c index 2118901b25e9..aeb9f333f4c7 100644 --- a/drivers/net/ethernet/meta/fbnic/fbnic_csr.c +++ b/drivers/net/ethernet/meta/fbnic/fbnic_csr.c @@ -64,7 +64,7 @@ static void fbnic_csr_get_regs_rpc_ram(struct fbnic_dev *fbd, u32 **data_p) u32 i, j; *(data++) = start; - *(data++) = end - 1; + *(data++) = end; /* FBNIC_RPC_TCAM_ACT */ for (i = 0; i < FBNIC_RPC_TCAM_ACT_NUM_ENTRIES; i++) { From 2b6ffcd7873b7e8a62c3e15a6f305bfc747c466b Mon Sep 17 00:00:00 2001 From: Joe Hattori Date: Thu, 19 Dec 2024 11:41:19 +0900 Subject: [PATCH 069/196] net: stmmac: restructure the error path of stmmac_probe_config_dt() Current implementation of stmmac_probe_config_dt() does not release the OF node reference obtained by of_parse_phandle() in some error paths. The problem is that some error paths call stmmac_remove_config_dt() to clean up but others use and unwind ladder. These two types of error handling have not kept in sync and have been a recurring source of bugs. Re-write the error handling in stmmac_probe_config_dt() to use an unwind ladder. Consequently, stmmac_remove_config_dt() is not needed anymore, thus remove it. This bug was found by an experimental verification tool that I am developing. Fixes: 4838a5405028 ("net: stmmac: Fix wrapper drivers not detecting PHY") Signed-off-by: Joe Hattori Link: https://patch.msgid.link/20241219024119.2017012-1-joe@pf.is.s.u-tokyo.ac.jp Signed-off-by: Jakub Kicinski --- .../ethernet/stmicro/stmmac/stmmac_platform.c | 43 ++++++++----------- 1 file changed, 17 insertions(+), 26 deletions(-) diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c index 3ac32444e492..dc9884130b91 100644 --- a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c +++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c @@ -405,22 +405,6 @@ static int stmmac_of_get_mac_mode(struct device_node *np) return -ENODEV; } -/** - * stmmac_remove_config_dt - undo the effects of stmmac_probe_config_dt() - * @pdev: platform_device structure - * @plat: driver data platform structure - * - * Release resources claimed by stmmac_probe_config_dt(). - */ -static void stmmac_remove_config_dt(struct platform_device *pdev, - struct plat_stmmacenet_data *plat) -{ - clk_disable_unprepare(plat->stmmac_clk); - clk_disable_unprepare(plat->pclk); - of_node_put(plat->phy_node); - of_node_put(plat->mdio_node); -} - /** * stmmac_probe_config_dt - parse device-tree driver parameters * @pdev: platform_device structure @@ -490,8 +474,10 @@ stmmac_probe_config_dt(struct platform_device *pdev, u8 *mac) dev_warn(&pdev->dev, "snps,phy-addr property is deprecated\n"); rc = stmmac_mdio_setup(plat, np, &pdev->dev); - if (rc) - return ERR_PTR(rc); + if (rc) { + ret = ERR_PTR(rc); + goto error_put_phy; + } of_property_read_u32(np, "tx-fifo-depth", &plat->tx_fifo_size); @@ -581,8 +567,8 @@ stmmac_probe_config_dt(struct platform_device *pdev, u8 *mac) dma_cfg = devm_kzalloc(&pdev->dev, sizeof(*dma_cfg), GFP_KERNEL); if (!dma_cfg) { - stmmac_remove_config_dt(pdev, plat); - return ERR_PTR(-ENOMEM); + ret = ERR_PTR(-ENOMEM); + goto error_put_mdio; } plat->dma_cfg = dma_cfg; @@ -610,8 +596,8 @@ stmmac_probe_config_dt(struct platform_device *pdev, u8 *mac) rc = stmmac_mtl_setup(pdev, plat); if (rc) { - stmmac_remove_config_dt(pdev, plat); - return ERR_PTR(rc); + ret = ERR_PTR(rc); + goto error_put_mdio; } /* clock setup */ @@ -663,6 +649,10 @@ stmmac_probe_config_dt(struct platform_device *pdev, u8 *mac) clk_disable_unprepare(plat->pclk); error_pclk_get: clk_disable_unprepare(plat->stmmac_clk); +error_put_mdio: + of_node_put(plat->mdio_node); +error_put_phy: + of_node_put(plat->phy_node); return ret; } @@ -671,16 +661,17 @@ static void devm_stmmac_remove_config_dt(void *data) { struct plat_stmmacenet_data *plat = data; - /* Platform data argument is unused */ - stmmac_remove_config_dt(NULL, plat); + clk_disable_unprepare(plat->stmmac_clk); + clk_disable_unprepare(plat->pclk); + of_node_put(plat->mdio_node); + of_node_put(plat->phy_node); } /** * devm_stmmac_probe_config_dt * @pdev: platform_device structure * @mac: MAC address to use - * Description: Devres variant of stmmac_probe_config_dt(). Does not require - * the user to call stmmac_remove_config_dt() at driver detach. + * Description: Devres variant of stmmac_probe_config_dt(). */ struct plat_stmmacenet_data * devm_stmmac_probe_config_dt(struct platform_device *pdev, u8 *mac) From 4f4aa4aa28142d53f8b06585c478476cfe325cfc Mon Sep 17 00:00:00 2001 From: Wang Liang Date: Thu, 19 Dec 2024 15:28:59 +0800 Subject: [PATCH 070/196] net: fix memory leak in tcp_conn_request() If inet_csk_reqsk_queue_hash_add() return false, tcp_conn_request() will return without free the dst memory, which allocated in af_ops->route_req. Here is the kmemleak stack: unreferenced object 0xffff8881198631c0 (size 240): comm "softirq", pid 0, jiffies 4299266571 (age 1802.392s) hex dump (first 32 bytes): 00 10 9b 03 81 88 ff ff 80 98 da bc ff ff ff ff ................ 81 55 18 bb ff ff ff ff 00 00 00 00 00 00 00 00 .U.............. backtrace: [] kmem_cache_alloc+0x60c/0xa80 [] dst_alloc+0x55/0x250 [] rt_dst_alloc+0x46/0x1d0 [] __mkroute_output+0x29a/0xa50 [] ip_route_output_key_hash+0x10b/0x240 [] ip_route_output_flow+0x1d/0x90 [] inet_csk_route_req+0x2c5/0x500 [] tcp_conn_request+0x691/0x12c0 [] tcp_rcv_state_process+0x3c8/0x11b0 [] tcp_v4_do_rcv+0x156/0x3b0 [] tcp_v4_rcv+0x1cf8/0x1d80 [] ip_protocol_deliver_rcu+0xf6/0x360 [] ip_local_deliver_finish+0xe6/0x1e0 [] ip_local_deliver+0xee/0x360 [] ip_rcv+0xad/0x2f0 [] __netif_receive_skb_one_core+0x123/0x140 Call dst_release() to free the dst memory when inet_csk_reqsk_queue_hash_add() return false in tcp_conn_request(). Fixes: ff46e3b44219 ("Fix race for duplicate reqsk on identical SYN") Signed-off-by: Wang Liang Link: https://patch.msgid.link/20241219072859.3783576-1-wangliang74@huawei.com Signed-off-by: Jakub Kicinski --- net/ipv4/tcp_input.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 5bdf13ac26ef..4811727b8a02 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -7328,6 +7328,7 @@ int tcp_conn_request(struct request_sock_ops *rsk_ops, if (unlikely(!inet_csk_reqsk_queue_hash_add(sk, req, req->timeout))) { reqsk_free(req); + dst_release(dst); return 0; } From b5a7b661a073727219fedc35f5619f62418ffe72 Mon Sep 17 00:00:00 2001 From: Xiao Liang Date: Thu, 19 Dec 2024 21:03:36 +0800 Subject: [PATCH 071/196] net: Fix netns for ip_tunnel_init_flow() The device denoted by tunnel->parms.link resides in the underlay net namespace. Therefore pass tunnel->net to ip_tunnel_init_flow(). Fixes: db53cd3d88dc ("net: Handle l3mdev in ip_tunnel_init_flow") Signed-off-by: Xiao Liang Reviewed-by: Ido Schimmel Link: https://patch.msgid.link/20241219130336.103839-1-shaw.leon@gmail.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c | 3 +-- net/ipv4/ip_tunnel.c | 6 +++--- 2 files changed, 4 insertions(+), 5 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c index 4b5fd71c897d..32d2e61f2b82 100644 --- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c +++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_span.c @@ -423,8 +423,7 @@ mlxsw_sp_span_gretap4_route(const struct net_device *to_dev, parms = mlxsw_sp_ipip_netdev_parms4(to_dev); ip_tunnel_init_flow(&fl4, parms.iph.protocol, *daddrp, *saddrp, - 0, 0, dev_net(to_dev), parms.link, tun->fwmark, 0, - 0); + 0, 0, tun->net, parms.link, tun->fwmark, 0, 0); rt = ip_route_output_key(tun->net, &fl4); if (IS_ERR(rt)) diff --git a/net/ipv4/ip_tunnel.c b/net/ipv4/ip_tunnel.c index 25505f9b724c..09b73acf037a 100644 --- a/net/ipv4/ip_tunnel.c +++ b/net/ipv4/ip_tunnel.c @@ -294,7 +294,7 @@ static int ip_tunnel_bind_dev(struct net_device *dev) ip_tunnel_init_flow(&fl4, iph->protocol, iph->daddr, iph->saddr, tunnel->parms.o_key, - iph->tos & INET_DSCP_MASK, dev_net(dev), + iph->tos & INET_DSCP_MASK, tunnel->net, tunnel->parms.link, tunnel->fwmark, 0, 0); rt = ip_route_output_key(tunnel->net, &fl4); @@ -611,7 +611,7 @@ void ip_md_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, } ip_tunnel_init_flow(&fl4, proto, key->u.ipv4.dst, key->u.ipv4.src, tunnel_id_to_key32(key->tun_id), - tos & INET_DSCP_MASK, dev_net(dev), 0, skb->mark, + tos & INET_DSCP_MASK, tunnel->net, 0, skb->mark, skb_get_hash(skb), key->flow_flags); if (!tunnel_hlen) @@ -774,7 +774,7 @@ void ip_tunnel_xmit(struct sk_buff *skb, struct net_device *dev, ip_tunnel_init_flow(&fl4, protocol, dst, tnl_params->saddr, tunnel->parms.o_key, tos & INET_DSCP_MASK, - dev_net(dev), READ_ONCE(tunnel->parms.link), + tunnel->net, READ_ONCE(tunnel->parms.link), tunnel->fwmark, skb_get_hash(skb), 0); if (ip_tunnel_encap(skb, &tunnel->encap, &protocol, &fl4) < 0) From a4fd163aed2edd967a244499754dec991d8b4c7d Mon Sep 17 00:00:00 2001 From: Ilya Shchipletsov Date: Thu, 19 Dec 2024 08:23:07 +0000 Subject: [PATCH 072/196] netrom: check buffer length before accessing it Syzkaller reports an uninit value read from ax25cmp when sending raw message through ieee802154 implementation. ===================================================== BUG: KMSAN: uninit-value in ax25cmp+0x3a5/0x460 net/ax25/ax25_addr.c:119 ax25cmp+0x3a5/0x460 net/ax25/ax25_addr.c:119 nr_dev_get+0x20e/0x450 net/netrom/nr_route.c:601 nr_route_frame+0x1a2/0xfc0 net/netrom/nr_route.c:774 nr_xmit+0x5a/0x1c0 net/netrom/nr_dev.c:144 __netdev_start_xmit include/linux/netdevice.h:4940 [inline] netdev_start_xmit include/linux/netdevice.h:4954 [inline] xmit_one net/core/dev.c:3548 [inline] dev_hard_start_xmit+0x247/0xa10 net/core/dev.c:3564 __dev_queue_xmit+0x33b8/0x5130 net/core/dev.c:4349 dev_queue_xmit include/linux/netdevice.h:3134 [inline] raw_sendmsg+0x654/0xc10 net/ieee802154/socket.c:299 ieee802154_sock_sendmsg+0x91/0xc0 net/ieee802154/socket.c:96 sock_sendmsg_nosec net/socket.c:730 [inline] __sock_sendmsg net/socket.c:745 [inline] ____sys_sendmsg+0x9c2/0xd60 net/socket.c:2584 ___sys_sendmsg+0x28d/0x3c0 net/socket.c:2638 __sys_sendmsg net/socket.c:2667 [inline] __do_sys_sendmsg net/socket.c:2676 [inline] __se_sys_sendmsg net/socket.c:2674 [inline] __x64_sys_sendmsg+0x307/0x490 net/socket.c:2674 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0x44/0x110 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x63/0x6b Uninit was created at: slab_post_alloc_hook+0x129/0xa70 mm/slab.h:768 slab_alloc_node mm/slub.c:3478 [inline] kmem_cache_alloc_node+0x5e9/0xb10 mm/slub.c:3523 kmalloc_reserve+0x13d/0x4a0 net/core/skbuff.c:560 __alloc_skb+0x318/0x740 net/core/skbuff.c:651 alloc_skb include/linux/skbuff.h:1286 [inline] alloc_skb_with_frags+0xc8/0xbd0 net/core/skbuff.c:6334 sock_alloc_send_pskb+0xa80/0xbf0 net/core/sock.c:2780 sock_alloc_send_skb include/net/sock.h:1884 [inline] raw_sendmsg+0x36d/0xc10 net/ieee802154/socket.c:282 ieee802154_sock_sendmsg+0x91/0xc0 net/ieee802154/socket.c:96 sock_sendmsg_nosec net/socket.c:730 [inline] __sock_sendmsg net/socket.c:745 [inline] ____sys_sendmsg+0x9c2/0xd60 net/socket.c:2584 ___sys_sendmsg+0x28d/0x3c0 net/socket.c:2638 __sys_sendmsg net/socket.c:2667 [inline] __do_sys_sendmsg net/socket.c:2676 [inline] __se_sys_sendmsg net/socket.c:2674 [inline] __x64_sys_sendmsg+0x307/0x490 net/socket.c:2674 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0x44/0x110 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x63/0x6b CPU: 0 PID: 5037 Comm: syz-executor166 Not tainted 6.7.0-rc7-syzkaller-00003-gfbafc3e621c3 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/17/2023 ===================================================== This issue occurs because the skb buffer is too small, and it's actual allocation is aligned. This hides an actual issue, which is that nr_route_frame does not validate the buffer size before using it. Fix this issue by checking skb->len before accessing any fields in skb->data. Found by Linux Verification Center (linuxtesting.org) with Syzkaller. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Co-developed-by: Nikita Marushkin Signed-off-by: Nikita Marushkin Signed-off-by: Ilya Shchipletsov Link: https://patch.msgid.link/20241219082308.3942-1-rabbelkin@mail.ru Signed-off-by: Jakub Kicinski --- net/netrom/nr_route.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/net/netrom/nr_route.c b/net/netrom/nr_route.c index 2b5e246b8d9a..b94cb2ffbaf8 100644 --- a/net/netrom/nr_route.c +++ b/net/netrom/nr_route.c @@ -754,6 +754,12 @@ int nr_route_frame(struct sk_buff *skb, ax25_cb *ax25) int ret; struct sk_buff *skbn; + /* + * Reject malformed packets early. Check that it contains at least 2 + * addresses and 1 byte more for Time-To-Live + */ + if (skb->len < 2 * sizeof(ax25_address) + 1) + return 0; nr_src = (ax25_address *)(skb->data + 0); nr_dest = (ax25_address *)(skb->data + 7); From 4e86729d1ff329815a6e8a920cb554a1d4cb5b8d Mon Sep 17 00:00:00 2001 From: Nikolay Kuratov Date: Thu, 19 Dec 2024 19:21:14 +0300 Subject: [PATCH 073/196] net/sctp: Prevent autoclose integer overflow in sctp_association_init() While by default max_autoclose equals to INT_MAX / HZ, one may set net.sctp.max_autoclose to UINT_MAX. There is code in sctp_association_init() that can consequently trigger overflow. Cc: stable@vger.kernel.org Fixes: 9f70f46bd4c7 ("sctp: properly latch and use autoclose value from sock to association") Signed-off-by: Nikolay Kuratov Acked-by: Xin Long Link: https://patch.msgid.link/20241219162114.2863827-1-kniv@yandex-team.ru Signed-off-by: Jakub Kicinski --- net/sctp/associola.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/net/sctp/associola.c b/net/sctp/associola.c index c45c192b7878..0b0794f164cf 100644 --- a/net/sctp/associola.c +++ b/net/sctp/associola.c @@ -137,7 +137,8 @@ static struct sctp_association *sctp_association_init( = 5 * asoc->rto_max; asoc->timeouts[SCTP_EVENT_TIMEOUT_SACK] = asoc->sackdelay; - asoc->timeouts[SCTP_EVENT_TIMEOUT_AUTOCLOSE] = sp->autoclose * HZ; + asoc->timeouts[SCTP_EVENT_TIMEOUT_AUTOCLOSE] = + (unsigned long)sp->autoclose * HZ; /* Initializes the timers */ for (i = SCTP_EVENT_TIMEOUT_NONE; i < SCTP_NUM_TIMEOUT_TYPES; ++i) From 4a4d38ace1fb0586bffd2aab03caaa05d6011748 Mon Sep 17 00:00:00 2001 From: Siddharth Vadapalli Date: Fri, 20 Dec 2024 13:26:14 +0530 Subject: [PATCH 074/196] net: ethernet: ti: am65-cpsw: default to round-robin for host port receive The Host Port (i.e. CPU facing port) of CPSW receives traffic from Linux via TX DMA Channels which are Hardware Queues consisting of traffic categorized according to their priority. The Host Port is configured to dequeue traffic from these Hardware Queues on the basis of priority i.e. as long as traffic exists on a Hardware Queue of a higher priority, the traffic on Hardware Queues of lower priority isn't dequeued. An alternate operation is also supported wherein traffic can be dequeued by the Host Port in a Round-Robin manner. Until commit under Fixes, the am65-cpsw driver enabled a single TX DMA Channel, due to which, unless modified by user via "ethtool", all traffic from Linux is transmitted on DMA Channel 0. Therefore, configuring the Host Port for priority based dequeuing or Round-Robin operation is identical since there is a single DMA Channel. Since commit under Fixes, all 8 TX DMA Channels are enabled by default. Additionally, the default "tc mapping" doesn't take into account the possibility of different traffic profiles which various users might have. This results in traffic starvation at the Host Port due to the priority based dequeuing which has been enabled by default since the inception of the driver. The traffic starvation triggers NETDEV WATCHDOG timeout for all TX DMA Channels that haven't been serviced due to the presence of traffic on the higher priority TX DMA Channels. Fix this by defaulting to Round-Robin dequeuing at the Host Port, which shall ensure that traffic is dequeued from all TX DMA Channels irrespective of the traffic profile. This will address the NETDEV WATCHDOG timeouts. At the same time, users can still switch from Round-Robin to Priority based dequeuing at the Host Port with the help of the "p0-rx-ptype-rrobin" private flag of "ethtool". Users are expected to setup an appropriate "tc mapping" that suits their traffic profile when switching to priority based dequeuing at the Host Port. Fixes: be397ea3473d ("net: ethernet: am65-cpsw: Set default TX channels to maximum") Cc: Signed-off-by: Siddharth Vadapalli Link: https://patch.msgid.link/20241220075618.228202-1-s-vadapalli@ti.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/ti/am65-cpsw-nuss.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/ti/am65-cpsw-nuss.c b/drivers/net/ethernet/ti/am65-cpsw-nuss.c index 14e1df721f2e..5465bf872734 100644 --- a/drivers/net/ethernet/ti/am65-cpsw-nuss.c +++ b/drivers/net/ethernet/ti/am65-cpsw-nuss.c @@ -3551,7 +3551,7 @@ static int am65_cpsw_nuss_probe(struct platform_device *pdev) init_completion(&common->tdown_complete); common->tx_ch_num = AM65_CPSW_DEFAULT_TX_CHNS; common->rx_ch_num_flows = AM65_CPSW_DEFAULT_RX_CHN_FLOWS; - common->pf_p0_rx_ptype_rrobin = false; + common->pf_p0_rx_ptype_rrobin = true; common->default_vlan = 1; common->ports = devm_kcalloc(dev, common->port_num, From 75221e96101fa93390d3db5c23e026f5e3565d9b Mon Sep 17 00:00:00 2001 From: Kory Maincent Date: Fri, 20 Dec 2024 18:04:00 +0100 Subject: [PATCH 075/196] net: pse-pd: tps23881: Fix power on/off issue An issue was present in the initial driver implementation. The driver read the power status of all channels before toggling the bit of the desired one. Using the power status register as a base value introduced a problem, because only the bit corresponding to the concerned channel ID should be set in the write-only power enable register. This led to cases where disabling power for one channel also powered off other channels. This patch removes the power status read and ensures the value is limited to the bit matching the channel index of the PI. Fixes: 20e6d190ffe1 ("net: pse-pd: Add TI TPS23881 PSE controller driver") Signed-off-by: Kory Maincent Acked-by: Oleksij Rempel Link: https://patch.msgid.link/20241220170400.291705-1-kory.maincent@bootlin.com Signed-off-by: Jakub Kicinski --- drivers/net/pse-pd/tps23881.c | 16 ++++------------ 1 file changed, 4 insertions(+), 12 deletions(-) diff --git a/drivers/net/pse-pd/tps23881.c b/drivers/net/pse-pd/tps23881.c index 5c4e88be46ee..8797ca1a8a21 100644 --- a/drivers/net/pse-pd/tps23881.c +++ b/drivers/net/pse-pd/tps23881.c @@ -64,15 +64,11 @@ static int tps23881_pi_enable(struct pse_controller_dev *pcdev, int id) if (id >= TPS23881_MAX_CHANS) return -ERANGE; - ret = i2c_smbus_read_word_data(client, TPS23881_REG_PW_STATUS); - if (ret < 0) - return ret; - chan = priv->port[id].chan[0]; if (chan < 4) - val = (u16)(ret | BIT(chan)); + val = BIT(chan); else - val = (u16)(ret | BIT(chan + 4)); + val = BIT(chan + 4); if (priv->port[id].is_4p) { chan = priv->port[id].chan[1]; @@ -100,15 +96,11 @@ static int tps23881_pi_disable(struct pse_controller_dev *pcdev, int id) if (id >= TPS23881_MAX_CHANS) return -ERANGE; - ret = i2c_smbus_read_word_data(client, TPS23881_REG_PW_STATUS); - if (ret < 0) - return ret; - chan = priv->port[id].chan[0]; if (chan < 4) - val = (u16)(ret | BIT(chan + 4)); + val = BIT(chan + 4); else - val = (u16)(ret | BIT(chan + 8)); + val = BIT(chan + 8); if (priv->port[id].is_4p) { chan = priv->port[id].chan[1]; From 050a4c011b0dfeb91664a5d7bd3647ff38db08ce Mon Sep 17 00:00:00 2001 From: Shahar Shitrit Date: Fri, 20 Dec 2024 10:15:02 +0200 Subject: [PATCH 076/196] net/mlx5: DR, select MSIX vector 0 for completion queue creation When creating a software steering completion queue (CQ), an arbitrary MSIX vector n is selected. This results in the CQ sharing the same Ethernet traffic channel n associated with the chosen vector. However, the value of n is often unpredictable, which can introduce complications for interrupt monitoring and verification tools. Moreover, SW steering uses polling rather than event-driven interrupts. Therefore, there is no need to select any MSIX vector other than the existing vector 0 for CQ creation. In light of these factors, and to enhance predictability, we modify the code to consistently select MSIX vector 0 for CQ creation. Fixes: 297cccebdc5a ("net/mlx5: DR, Expose an internal API to issue RDMA operations") Signed-off-by: Shahar Shitrit Reviewed-by: Yevgeny Kliteynik Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/20241220081505.1286093-2-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- .../net/ethernet/mellanox/mlx5/core/steering/sws/dr_send.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_send.c b/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_send.c index 6fa06ba2d346..f57c84e5128b 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_send.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/steering/sws/dr_send.c @@ -1067,7 +1067,6 @@ static struct mlx5dr_cq *dr_create_cq(struct mlx5_core_dev *mdev, int inlen, err, eqn; void *cqc, *in; __be64 *pas; - int vector; u32 i; cq = kzalloc(sizeof(*cq), GFP_KERNEL); @@ -1096,8 +1095,7 @@ static struct mlx5dr_cq *dr_create_cq(struct mlx5_core_dev *mdev, if (!in) goto err_cqwq; - vector = raw_smp_processor_id() % mlx5_comp_vectors_max(mdev); - err = mlx5_comp_eqn_get(mdev, vector, &eqn); + err = mlx5_comp_eqn_get(mdev, 0, &eqn); if (err) { kvfree(in); goto err_cqwq; From 8c6254479b3d5bd788d2b5fefaa48fb194331ed0 Mon Sep 17 00:00:00 2001 From: Dragos Tatulea Date: Fri, 20 Dec 2024 10:15:03 +0200 Subject: [PATCH 077/196] net/mlx5e: macsec: Maintain TX SA from encoding_sa In MACsec, it is possible to create multiple active TX SAs on a SC, but only one such SA can be used at a time for transmission. This SA is selected through the encoding_sa link parameter. When there are 2 or more active TX SAs configured (encoding_sa=0): ip macsec add macsec0 tx sa 0 pn 1 on key 00 ip macsec add macsec0 tx sa 1 pn 1 on key 00 ... the traffic should be still sent via TX SA 0 as the encoding_sa was not changed. However, the driver ignores the encoding_sa and overrides it to SA 1 by installing the flow steering id of the newly created TX SA into the SCI -> flow steering id hash map. The future packet tx descriptors will point to the incorrect flow steering rule (SA 1). This patch fixes the issue by avoiding the creation of the flow steering rule for an active TX SA that is not the encoding_sa. The driver side tx_sa object and the FW side macsec object are still created. When the encoding_sa link parameter is changed to another active TX SA, only the new flow steering rule will be created in the mlx5e_macsec_upd_txsa() handler. Fixes: 8ff0ac5be144 ("net/mlx5: Add MACsec offload Tx command support") Signed-off-by: Dragos Tatulea Reviewed-by: Cosmin Ratiu Reviewed-by: Lior Nahmanson Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/20241220081505.1286093-3-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c index cc9bcc420032..6ab02f3fc291 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_accel/macsec.c @@ -339,9 +339,13 @@ static int mlx5e_macsec_init_sa_fs(struct macsec_context *ctx, { struct mlx5e_priv *priv = macsec_netdev_priv(ctx->netdev); struct mlx5_macsec_fs *macsec_fs = priv->mdev->macsec_fs; + const struct macsec_tx_sc *tx_sc = &ctx->secy->tx_sc; struct mlx5_macsec_rule_attrs rule_attrs; union mlx5_macsec_rule *macsec_rule; + if (is_tx && tx_sc->encoding_sa != sa->assoc_num) + return 0; + rule_attrs.macsec_obj_id = sa->macsec_obj_id; rule_attrs.sci = sa->sci; rule_attrs.assoc_num = sa->assoc_num; From 5a03b368562a7ff5f5f1f63b5adf8309cbdbd5be Mon Sep 17 00:00:00 2001 From: Jianbo Liu Date: Fri, 20 Dec 2024 10:15:04 +0200 Subject: [PATCH 078/196] net/mlx5e: Skip restore TC rules for vport rep without loaded flag During driver unload, unregister_netdev is called after unloading vport rep. So, the mlx5e_rep_priv is already freed while trying to get rpriv->netdev, or walk rpriv->tc_ht, which results in use-after-free. So add the checking to make sure access the data of vport rep which is still loaded. Fixes: d1569537a837 ("net/mlx5e: Modify and restore TC rules for IPSec TX rules") Signed-off-by: Jianbo Liu Reviewed-by: Saeed Mahameed Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/20241220081505.1286093-4-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/mellanox/mlx5/core/esw/ipsec_fs.c | 6 +++--- drivers/net/ethernet/mellanox/mlx5/core/eswitch.h | 3 +++ drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c | 3 --- 3 files changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/esw/ipsec_fs.c b/drivers/net/ethernet/mellanox/mlx5/core/esw/ipsec_fs.c index 5a0047bdcb51..ed977ae75fab 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/esw/ipsec_fs.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/esw/ipsec_fs.c @@ -150,11 +150,11 @@ void mlx5_esw_ipsec_restore_dest_uplink(struct mlx5_core_dev *mdev) unsigned long i; int err; - xa_for_each(&esw->offloads.vport_reps, i, rep) { - rpriv = rep->rep_data[REP_ETH].priv; - if (!rpriv || !rpriv->netdev) + mlx5_esw_for_each_rep(esw, i, rep) { + if (atomic_read(&rep->rep_data[REP_ETH].state) != REP_LOADED) continue; + rpriv = rep->rep_data[REP_ETH].priv; rhashtable_walk_enter(&rpriv->tc_ht, &iter); rhashtable_walk_start(&iter); while ((flow = rhashtable_walk_next(&iter)) != NULL) { diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h index a83d41121db6..8573d36785f4 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.h @@ -714,6 +714,9 @@ void mlx5e_tc_clean_fdb_peer_flows(struct mlx5_eswitch *esw); MLX5_CAP_GEN_2((esw->dev), ec_vf_vport_base) +\ (last) - 1) +#define mlx5_esw_for_each_rep(esw, i, rep) \ + xa_for_each(&((esw)->offloads.vport_reps), i, rep) + struct mlx5_eswitch *__must_check mlx5_devlink_eswitch_get(struct devlink *devlink); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index d5b42b3a19fd..40359f320724 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -53,9 +53,6 @@ #include "lag/lag.h" #include "en/tc/post_meter.h" -#define mlx5_esw_for_each_rep(esw, i, rep) \ - xa_for_each(&((esw)->offloads.vport_reps), i, rep) - /* There are two match-all miss flows, one for unicast dst mac and * one for multicast. */ From 2a4f56fbcc473d8faeb29b73082df39efbe5893c Mon Sep 17 00:00:00 2001 From: Jianbo Liu Date: Fri, 20 Dec 2024 10:15:05 +0200 Subject: [PATCH 079/196] net/mlx5e: Keep netdev when leave switchdev for devlink set legacy only In the cited commit, when changing from switchdev to legacy mode, uplink representor's netdev is kept, and its profile is replaced with nic profile, so netdev is detached from old profile, then attach to new profile. During profile change, the hardware resources allocated by the old profile will be cleaned up. However, the cleanup is relying on the related kernel modules. And they may need to flush themselves first, which is triggered by netdev events, for example, NETDEV_UNREGISTER. However, netdev is kept, or netdev_register is called after the cleanup, which may cause troubles because the resources are still referred by kernel modules. The same process applies to all the caes when uplink is leaving switchdev mode, including devlink eswitch mode set legacy, driver unload and devlink reload. For the first one, it can be blocked and returns failure to users, whenever possible. But it's hard for the others. Besides, the attachment to nic profile is unnecessary as the netdev will be unregistered anyway for such cases. So in this patch, the original behavior is kept only for devlink eswitch set mode legacy. For the others, moves netdev unregistration before the profile change. Fixes: 7a9fb35e8c3a ("net/mlx5e: Do not reload ethernet ports when changing eswitch mode") Signed-off-by: Jianbo Liu Signed-off-by: Tariq Toukan Link: https://patch.msgid.link/20241220081505.1286093-5-tariqt@nvidia.com Signed-off-by: Jakub Kicinski --- .../net/ethernet/mellanox/mlx5/core/en_main.c | 19 +++++++++++++++++-- .../net/ethernet/mellanox/mlx5/core/en_rep.c | 15 +++++++++++++++ .../mellanox/mlx5/core/eswitch_offloads.c | 2 ++ include/linux/mlx5/driver.h | 1 + 4 files changed, 35 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c index dd16d73000c3..0ec17c276bdd 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_main.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_main.c @@ -6542,8 +6542,23 @@ static void _mlx5e_remove(struct auxiliary_device *adev) mlx5_core_uplink_netdev_set(mdev, NULL); mlx5e_dcbnl_delete_app(priv); - unregister_netdev(priv->netdev); - _mlx5e_suspend(adev, false); + /* When unload driver, the netdev is in registered state + * if it's from legacy mode. If from switchdev mode, it + * is already unregistered before changing to NIC profile. + */ + if (priv->netdev->reg_state == NETREG_REGISTERED) { + unregister_netdev(priv->netdev); + _mlx5e_suspend(adev, false); + } else { + struct mlx5_core_dev *pos; + int i; + + if (test_bit(MLX5E_STATE_DESTROYING, &priv->state)) + mlx5_sd_for_each_dev(i, mdev, pos) + mlx5e_destroy_mdev_resources(pos); + else + _mlx5e_suspend(adev, true); + } /* Avoid cleanup if profile rollback failed. */ if (priv->profile) priv->profile->cleanup(priv); diff --git a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c index 554f9cb5b53f..fdff9fd8a89e 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/en_rep.c @@ -1509,6 +1509,21 @@ mlx5e_vport_uplink_rep_unload(struct mlx5e_rep_priv *rpriv) priv = netdev_priv(netdev); + /* This bit is set when using devlink to change eswitch mode from + * switchdev to legacy. As need to keep uplink netdev ifindex, we + * detach uplink representor profile and attach NIC profile only. + * The netdev will be unregistered later when unload NIC auxiliary + * driver for this case. + * We explicitly block devlink eswitch mode change if any IPSec rules + * offloaded, but can't block other cases, such as driver unload + * and devlink reload. We have to unregister netdev before profile + * change for those cases. This is to avoid resource leak because + * the offloaded rules don't have the chance to be unoffloaded before + * cleanup which is triggered by detach uplink representor profile. + */ + if (!(priv->mdev->priv.flags & MLX5_PRIV_FLAGS_SWITCH_LEGACY)) + unregister_netdev(netdev); + mlx5e_netdev_attach_nic_profile(priv); } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c index 40359f320724..06076dd9ec64 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch_offloads.c @@ -3777,6 +3777,8 @@ int mlx5_devlink_eswitch_mode_set(struct devlink *devlink, u16 mode, esw->eswitch_operation_in_progress = true; up_write(&esw->mode_lock); + if (mode == DEVLINK_ESWITCH_MODE_LEGACY) + esw->dev->priv.flags |= MLX5_PRIV_FLAGS_SWITCH_LEGACY; mlx5_eswitch_disable_locked(esw); if (mode == DEVLINK_ESWITCH_MODE_SWITCHDEV) { if (mlx5_devlink_trap_get_num_active(esw->dev)) { diff --git a/include/linux/mlx5/driver.h b/include/linux/mlx5/driver.h index fc7e6153b73d..8f5991168ccd 100644 --- a/include/linux/mlx5/driver.h +++ b/include/linux/mlx5/driver.h @@ -524,6 +524,7 @@ enum { * creation/deletion on drivers rescan. Unset during device attach. */ MLX5_PRIV_FLAGS_DETACH = 1 << 2, + MLX5_PRIV_FLAGS_SWITCH_LEGACY = 1 << 3, }; struct mlx5_adev { From 2ac5415022d16d63d912a39a06f32f1f51140261 Mon Sep 17 00:00:00 2001 From: Zhu Yanjun Date: Fri, 20 Dec 2024 23:23:25 +0100 Subject: [PATCH 080/196] RDMA/rxe: Remove the direct link to net_device The similar patch in siw is in the link: https://git.kernel.org/rdma/rdma/c/16b87037b48889 This problem also occurred in RXE. The following analyze this problem. In the following Call Traces: " BUG: KASAN: slab-use-after-free in dev_get_flags+0x188/0x1d0 net/core/dev.c:8782 Read of size 4 at addr ffff8880554640b0 by task kworker/1:4/5295 CPU: 1 UID: 0 PID: 5295 Comm: kworker/1:4 Not tainted 6.12.0-rc3-syzkaller-00399-g9197b73fd7bb #0 Hardware name: Google Compute Engine/Google Compute Engine, BIOS Google 09/13/2024 Workqueue: infiniband ib_cache_event_task Call Trace: __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x241/0x360 lib/dump_stack.c:120 print_address_description mm/kasan/report.c:377 [inline] print_report+0x169/0x550 mm/kasan/report.c:488 kasan_report+0x143/0x180 mm/kasan/report.c:601 dev_get_flags+0x188/0x1d0 net/core/dev.c:8782 rxe_query_port+0x12d/0x260 drivers/infiniband/sw/rxe/rxe_verbs.c:60 __ib_query_port drivers/infiniband/core/device.c:2111 [inline] ib_query_port+0x168/0x7d0 drivers/infiniband/core/device.c:2143 ib_cache_update+0x1a9/0xb80 drivers/infiniband/core/cache.c:1494 ib_cache_event_task+0xf3/0x1e0 drivers/infiniband/core/cache.c:1568 process_one_work kernel/workqueue.c:3229 [inline] process_scheduled_works+0xa65/0x1850 kernel/workqueue.c:3310 worker_thread+0x870/0xd30 kernel/workqueue.c:3391 kthread+0x2f2/0x390 kernel/kthread.c:389 ret_from_fork+0x4d/0x80 arch/x86/kernel/process.c:147 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244 " 1). In the link [1], " infiniband syz2: set down " This means that on 839.350575, the event ib_cache_event_task was sent andi queued in ib_wq. 2). In the link [1], " team0 (unregistering): Port device team_slave_0 removed " It indicates that before 843.251853, the net device should be freed. 3). In the link [1], " BUG: KASAN: slab-use-after-free in dev_get_flags+0x188/0x1d0 " This means that on 850.559070, this slab-use-after-free problem occurred. In all, on 839.350575, the event ib_cache_event_task was sent and queued in ib_wq, before 843.251853, the net device veth was freed. on 850.559070, this event was executed, and the mentioned freed net device was called. Thus, the above call trace occurred. [1] https://syzkaller.appspot.com/x/log.txt?x=12e7025f980000 Reported-by: syzbot+4b87489410b4efd181bf@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=4b87489410b4efd181bf Fixes: 8700e3e7c485 ("Soft RoCE driver") Signed-off-by: Zhu Yanjun Link: https://patch.msgid.link/20241220222325.2487767-1-yanjun.zhu@linux.dev Signed-off-by: Leon Romanovsky --- drivers/infiniband/sw/rxe/rxe.c | 23 +++++++++++++++++++---- drivers/infiniband/sw/rxe/rxe.h | 3 ++- drivers/infiniband/sw/rxe/rxe_mcast.c | 22 ++++++++++++++++++++-- drivers/infiniband/sw/rxe/rxe_net.c | 24 ++++++++++++++++++++---- drivers/infiniband/sw/rxe/rxe_verbs.c | 26 +++++++++++++++++++++----- drivers/infiniband/sw/rxe/rxe_verbs.h | 11 ++++++++--- 6 files changed, 90 insertions(+), 19 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe.c b/drivers/infiniband/sw/rxe/rxe.c index 255677bc12b2..1ba4a0c8726a 100644 --- a/drivers/infiniband/sw/rxe/rxe.c +++ b/drivers/infiniband/sw/rxe/rxe.c @@ -40,6 +40,8 @@ void rxe_dealloc(struct ib_device *ib_dev) /* initialize rxe device parameters */ static void rxe_init_device_param(struct rxe_dev *rxe) { + struct net_device *ndev; + rxe->max_inline_data = RXE_MAX_INLINE_DATA; rxe->attr.vendor_id = RXE_VENDOR_ID; @@ -71,8 +73,15 @@ static void rxe_init_device_param(struct rxe_dev *rxe) rxe->attr.max_fast_reg_page_list_len = RXE_MAX_FMR_PAGE_LIST_LEN; rxe->attr.max_pkeys = RXE_MAX_PKEYS; rxe->attr.local_ca_ack_delay = RXE_LOCAL_CA_ACK_DELAY; + + ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); + if (!ndev) + return; + addrconf_addr_eui48((unsigned char *)&rxe->attr.sys_image_guid, - rxe->ndev->dev_addr); + ndev->dev_addr); + + dev_put(ndev); rxe->max_ucontext = RXE_MAX_UCONTEXT; } @@ -109,10 +118,15 @@ static void rxe_init_port_param(struct rxe_port *port) static void rxe_init_ports(struct rxe_dev *rxe) { struct rxe_port *port = &rxe->port; + struct net_device *ndev; rxe_init_port_param(port); + ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); + if (!ndev) + return; addrconf_addr_eui48((unsigned char *)&port->port_guid, - rxe->ndev->dev_addr); + ndev->dev_addr); + dev_put(ndev); spin_lock_init(&port->port_lock); } @@ -167,12 +181,13 @@ void rxe_set_mtu(struct rxe_dev *rxe, unsigned int ndev_mtu) /* called by ifc layer to create new rxe device. * The caller should allocate memory for rxe by calling ib_alloc_device. */ -int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name) +int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name, + struct net_device *ndev) { rxe_init(rxe); rxe_set_mtu(rxe, mtu); - return rxe_register_device(rxe, ibdev_name); + return rxe_register_device(rxe, ibdev_name, ndev); } static int rxe_newlink(const char *ibdev_name, struct net_device *ndev) diff --git a/drivers/infiniband/sw/rxe/rxe.h b/drivers/infiniband/sw/rxe/rxe.h index d8fb2c7af30a..fe7f97066732 100644 --- a/drivers/infiniband/sw/rxe/rxe.h +++ b/drivers/infiniband/sw/rxe/rxe.h @@ -139,7 +139,8 @@ enum resp_states { void rxe_set_mtu(struct rxe_dev *rxe, unsigned int dev_mtu); -int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name); +int rxe_add(struct rxe_dev *rxe, unsigned int mtu, const char *ibdev_name, + struct net_device *ndev); void rxe_rcv(struct sk_buff *skb); diff --git a/drivers/infiniband/sw/rxe/rxe_mcast.c b/drivers/infiniband/sw/rxe/rxe_mcast.c index 86cc2e18a7fd..07ff47bae31d 100644 --- a/drivers/infiniband/sw/rxe/rxe_mcast.c +++ b/drivers/infiniband/sw/rxe/rxe_mcast.c @@ -31,10 +31,19 @@ static int rxe_mcast_add(struct rxe_dev *rxe, union ib_gid *mgid) { unsigned char ll_addr[ETH_ALEN]; + struct net_device *ndev; + int ret; + + ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); + if (!ndev) + return -ENODEV; ipv6_eth_mc_map((struct in6_addr *)mgid->raw, ll_addr); - return dev_mc_add(rxe->ndev, ll_addr); + ret = dev_mc_add(ndev, ll_addr); + dev_put(ndev); + + return ret; } /** @@ -47,10 +56,19 @@ static int rxe_mcast_add(struct rxe_dev *rxe, union ib_gid *mgid) static int rxe_mcast_del(struct rxe_dev *rxe, union ib_gid *mgid) { unsigned char ll_addr[ETH_ALEN]; + struct net_device *ndev; + int ret; + + ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); + if (!ndev) + return -ENODEV; ipv6_eth_mc_map((struct in6_addr *)mgid->raw, ll_addr); - return dev_mc_del(rxe->ndev, ll_addr); + ret = dev_mc_del(ndev, ll_addr); + dev_put(ndev); + + return ret; } /** diff --git a/drivers/infiniband/sw/rxe/rxe_net.c b/drivers/infiniband/sw/rxe/rxe_net.c index 75d1407db52d..8cc64ceeb356 100644 --- a/drivers/infiniband/sw/rxe/rxe_net.c +++ b/drivers/infiniband/sw/rxe/rxe_net.c @@ -524,7 +524,16 @@ struct sk_buff *rxe_init_packet(struct rxe_dev *rxe, struct rxe_av *av, */ const char *rxe_parent_name(struct rxe_dev *rxe, unsigned int port_num) { - return rxe->ndev->name; + struct net_device *ndev; + char *ndev_name; + + ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); + if (!ndev) + return NULL; + ndev_name = ndev->name; + dev_put(ndev); + + return ndev_name; } int rxe_net_add(const char *ibdev_name, struct net_device *ndev) @@ -536,10 +545,9 @@ int rxe_net_add(const char *ibdev_name, struct net_device *ndev) if (!rxe) return -ENOMEM; - rxe->ndev = ndev; ib_mark_name_assigned_by_user(&rxe->ib_dev); - err = rxe_add(rxe, ndev->mtu, ibdev_name); + err = rxe_add(rxe, ndev->mtu, ibdev_name, ndev); if (err) { ib_dealloc_device(&rxe->ib_dev); return err; @@ -587,10 +595,18 @@ void rxe_port_down(struct rxe_dev *rxe) void rxe_set_port_state(struct rxe_dev *rxe) { - if (netif_running(rxe->ndev) && netif_carrier_ok(rxe->ndev)) + struct net_device *ndev; + + ndev = rxe_ib_device_get_netdev(&rxe->ib_dev); + if (!ndev) + return; + + if (netif_running(ndev) && netif_carrier_ok(ndev)) rxe_port_up(rxe); else rxe_port_down(rxe); + + dev_put(ndev); } static int rxe_notify(struct notifier_block *not_blk, diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c index 5c18f7e342f2..8a5fc20fd186 100644 --- a/drivers/infiniband/sw/rxe/rxe_verbs.c +++ b/drivers/infiniband/sw/rxe/rxe_verbs.c @@ -41,6 +41,7 @@ static int rxe_query_port(struct ib_device *ibdev, u32 port_num, struct ib_port_attr *attr) { struct rxe_dev *rxe = to_rdev(ibdev); + struct net_device *ndev; int err, ret; if (port_num != 1) { @@ -49,6 +50,12 @@ static int rxe_query_port(struct ib_device *ibdev, goto err_out; } + ndev = rxe_ib_device_get_netdev(ibdev); + if (!ndev) { + err = -ENODEV; + goto err_out; + } + memcpy(attr, &rxe->port.attr, sizeof(*attr)); mutex_lock(&rxe->usdev_lock); @@ -57,13 +64,14 @@ static int rxe_query_port(struct ib_device *ibdev, if (attr->state == IB_PORT_ACTIVE) attr->phys_state = IB_PORT_PHYS_STATE_LINK_UP; - else if (dev_get_flags(rxe->ndev) & IFF_UP) + else if (dev_get_flags(ndev) & IFF_UP) attr->phys_state = IB_PORT_PHYS_STATE_POLLING; else attr->phys_state = IB_PORT_PHYS_STATE_DISABLED; mutex_unlock(&rxe->usdev_lock); + dev_put(ndev); return ret; err_out: @@ -1425,9 +1433,16 @@ static const struct attribute_group rxe_attr_group = { static int rxe_enable_driver(struct ib_device *ib_dev) { struct rxe_dev *rxe = container_of(ib_dev, struct rxe_dev, ib_dev); + struct net_device *ndev; + + ndev = rxe_ib_device_get_netdev(ib_dev); + if (!ndev) + return -ENODEV; rxe_set_port_state(rxe); - dev_info(&rxe->ib_dev.dev, "added %s\n", netdev_name(rxe->ndev)); + dev_info(&rxe->ib_dev.dev, "added %s\n", netdev_name(ndev)); + + dev_put(ndev); return 0; } @@ -1495,7 +1510,8 @@ static const struct ib_device_ops rxe_dev_ops = { INIT_RDMA_OBJ_SIZE(ib_mw, rxe_mw, ibmw), }; -int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name) +int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name, + struct net_device *ndev) { int err; struct ib_device *dev = &rxe->ib_dev; @@ -1507,13 +1523,13 @@ int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name) dev->num_comp_vectors = num_possible_cpus(); dev->local_dma_lkey = 0; addrconf_addr_eui48((unsigned char *)&dev->node_guid, - rxe->ndev->dev_addr); + ndev->dev_addr); dev->uverbs_cmd_mask |= BIT_ULL(IB_USER_VERBS_CMD_POST_SEND) | BIT_ULL(IB_USER_VERBS_CMD_REQ_NOTIFY_CQ); ib_set_device_ops(dev, &rxe_dev_ops); - err = ib_device_set_netdev(&rxe->ib_dev, rxe->ndev, 1); + err = ib_device_set_netdev(&rxe->ib_dev, ndev, 1); if (err) return err; diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.h b/drivers/infiniband/sw/rxe/rxe_verbs.h index 3c1354f82283..6573ceec0ef5 100644 --- a/drivers/infiniband/sw/rxe/rxe_verbs.h +++ b/drivers/infiniband/sw/rxe/rxe_verbs.h @@ -370,6 +370,7 @@ struct rxe_port { u32 qp_gsi_index; }; +#define RXE_PORT 1 struct rxe_dev { struct ib_device ib_dev; struct ib_device_attr attr; @@ -377,8 +378,6 @@ struct rxe_dev { int max_inline_data; struct mutex usdev_lock; - struct net_device *ndev; - struct rxe_pool uc_pool; struct rxe_pool pd_pool; struct rxe_pool ah_pool; @@ -406,6 +405,11 @@ struct rxe_dev { struct crypto_shash *tfm; }; +static inline struct net_device *rxe_ib_device_get_netdev(struct ib_device *dev) +{ + return ib_device_get_netdev(dev, RXE_PORT); +} + static inline void rxe_counter_inc(struct rxe_dev *rxe, enum rxe_counters index) { atomic64_inc(&rxe->stats_counters[index]); @@ -471,6 +475,7 @@ static inline struct rxe_pd *rxe_mw_pd(struct rxe_mw *mw) return to_rpd(mw->ibmw.pd); } -int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name); +int rxe_register_device(struct rxe_dev *rxe, const char *ibdev_name, + struct net_device *ndev); #endif /* RXE_VERBS_H */ From 385a95cc72941c7f88630a7bc4176048cc03b395 Mon Sep 17 00:00:00 2001 From: Suraj Kandpal Date: Mon, 16 Dec 2024 23:45:54 +0530 Subject: [PATCH 081/196] drm/i915/cx0_phy: Fix C10 pll programming sequence According to spec VDR_CUSTOM_WIDTH register gets programmed after pll specific VDR registers and TX Lane programming registers are done. Moreover we only program into C10_VDR_CONTROL1 to update config and setup master lane once all VDR registers are written into. Bspec: 67636 Fixes: 51390cc0e00a ("drm/i915/mtl: Add Support for C10 PHY message bus and pll programming") Signed-off-by: Suraj Kandpal Reviewed-by: Ankit Nautiyal Link: https://patchwork.freedesktop.org/patch/msgid/20241216181554.2861381-1-suraj.kandpal@intel.com (cherry picked from commit f9d418552ba1e3a0e92487ff82eb515dab7516c0) Signed-off-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/display/intel_cx0_phy.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_cx0_phy.c b/drivers/gpu/drm/i915/display/intel_cx0_phy.c index 71dc659228ab..0c7aee13495a 100644 --- a/drivers/gpu/drm/i915/display/intel_cx0_phy.c +++ b/drivers/gpu/drm/i915/display/intel_cx0_phy.c @@ -2115,14 +2115,6 @@ static void intel_c10_pll_program(struct intel_display *display, 0, C10_VDR_CTRL_MSGBUS_ACCESS, MB_WRITE_COMMITTED); - /* Custom width needs to be programmed to 0 for both the phy lanes */ - intel_cx0_rmw(encoder, INTEL_CX0_BOTH_LANES, PHY_C10_VDR_CUSTOM_WIDTH, - C10_VDR_CUSTOM_WIDTH_MASK, C10_VDR_CUSTOM_WIDTH_8_10, - MB_WRITE_COMMITTED); - intel_cx0_rmw(encoder, INTEL_CX0_BOTH_LANES, PHY_C10_VDR_CONTROL(1), - 0, C10_VDR_CTRL_UPDATE_CFG, - MB_WRITE_COMMITTED); - /* Program the pll values only for the master lane */ for (i = 0; i < ARRAY_SIZE(pll_state->pll); i++) intel_cx0_write(encoder, INTEL_CX0_LANE0, PHY_C10_VDR_PLL(i), @@ -2132,6 +2124,10 @@ static void intel_c10_pll_program(struct intel_display *display, intel_cx0_write(encoder, INTEL_CX0_LANE0, PHY_C10_VDR_CMN(0), pll_state->cmn, MB_WRITE_COMMITTED); intel_cx0_write(encoder, INTEL_CX0_LANE0, PHY_C10_VDR_TX(0), pll_state->tx, MB_WRITE_COMMITTED); + /* Custom width needs to be programmed to 0 for both the phy lanes */ + intel_cx0_rmw(encoder, INTEL_CX0_BOTH_LANES, PHY_C10_VDR_CUSTOM_WIDTH, + C10_VDR_CUSTOM_WIDTH_MASK, C10_VDR_CUSTOM_WIDTH_8_10, + MB_WRITE_COMMITTED); intel_cx0_rmw(encoder, INTEL_CX0_LANE0, PHY_C10_VDR_CONTROL(1), 0, C10_VDR_CTRL_MASTER_LANE | C10_VDR_CTRL_UPDATE_CFG, MB_WRITE_COMMITTED); From 20e7c5313ffbf11c34a46395345677adbe890bee Mon Sep 17 00:00:00 2001 From: Rodrigo Vivi Date: Thu, 19 Dec 2024 16:00:19 -0500 Subject: [PATCH 082/196] drm/i915/dg1: Fix power gate sequence. sub-pipe PG is not present on DG1. Setting these bits can disable other power gates and cause GPU hangs on video playbacks. VLK: 16314, 4304 Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/13381 Fixes: 85a12d7eb8fe ("drm/i915/tgl: Fix Media power gate sequence.") Cc: Vinay Belgaumkar Cc: Himal Prasad Ghimiray Reviewed-by: Vinay Belgaumkar Reviewed-by: Himal Prasad Ghimiray Link: https://patchwork.freedesktop.org/patch/msgid/20241219210019.70532-1-rodrigo.vivi@intel.com Signed-off-by: Rodrigo Vivi (cherry picked from commit de7061947b4ed4be857d452c60d5fb795831d79e) Signed-off-by: Tvrtko Ursulin --- drivers/gpu/drm/i915/gt/intel_rc6.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/intel_rc6.c b/drivers/gpu/drm/i915/gt/intel_rc6.c index c864d101faf9..9378d5901c49 100644 --- a/drivers/gpu/drm/i915/gt/intel_rc6.c +++ b/drivers/gpu/drm/i915/gt/intel_rc6.c @@ -133,7 +133,7 @@ static void gen11_rc6_enable(struct intel_rc6 *rc6) GEN9_MEDIA_PG_ENABLE | GEN11_MEDIA_SAMPLER_PG_ENABLE; - if (GRAPHICS_VER(gt->i915) >= 12) { + if (GRAPHICS_VER(gt->i915) >= 12 && !IS_DG1(gt->i915)) { for (i = 0; i < I915_MAX_VCS; i++) if (HAS_ENGINE(gt, _VCS(i))) pg_enable |= (VDN_HCP_POWERGATE_ENABLE(i) | From d57212f281fda9056412cd6cca983d9d2eb89f53 Mon Sep 17 00:00:00 2001 From: Su Hui Date: Tue, 24 Dec 2024 12:43:58 +0800 Subject: [PATCH 083/196] workqueue: add printf attribute to __alloc_workqueue() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix a compiler warning with W=1: kernel/workqueue.c: error: function ‘__alloc_workqueue’ might be a candidate for ‘gnu_printf’ format attribute[-Werror=suggest-attribute=format] 5657 | name_len = vsnprintf(wq->name, sizeof(wq->name), fmt, args); | ^~~~~~~~ Fixes: 9b59a85a84dc ("workqueue: Don't call va_start / va_end twice") Signed-off-by: Su Hui Signed-off-by: Tejun Heo --- kernel/workqueue.c | 1 + 1 file changed, 1 insertion(+) diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 8336218ec4b8..f7d8fc204579 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -5645,6 +5645,7 @@ static void wq_adjust_max_active(struct workqueue_struct *wq) } while (activated); } +__printf(1, 0) static struct workqueue_struct *__alloc_workqueue(const char *fmt, unsigned int flags, int max_active, va_list args) From 35bf430e08a18fdab6eb94492a06d9ad14c6179b Mon Sep 17 00:00:00 2001 From: Henry Huang Date: Sun, 22 Dec 2024 23:43:16 +0800 Subject: [PATCH 084/196] sched_ext: initialize kit->cursor.flags struct bpf_iter_scx_dsq *it maybe not initialized. If we didn't call scx_bpf_dsq_move_set_vtime and scx_bpf_dsq_move_set_slice before scx_bpf_dsq_move, it would cause unexpected behaviors: 1. Assign a huge slice into p->scx.slice 2. Assign a invalid vtime into p->scx.dsq_vtime Signed-off-by: Henry Huang Fixes: 6462dd53a260 ("sched_ext: Compact struct bpf_iter_scx_dsq_kern") Cc: stable@vger.kernel.org # v6.12 Signed-off-by: Tejun Heo --- kernel/sched/ext.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/sched/ext.c b/kernel/sched/ext.c index 98519e6d0dcd..19d2699cf638 100644 --- a/kernel/sched/ext.c +++ b/kernel/sched/ext.c @@ -7013,7 +7013,7 @@ __bpf_kfunc int bpf_iter_scx_dsq_new(struct bpf_iter_scx_dsq *it, u64 dsq_id, return -ENOENT; INIT_LIST_HEAD(&kit->cursor.node); - kit->cursor.flags |= SCX_DSQ_LNODE_ITER_CURSOR | flags; + kit->cursor.flags = SCX_DSQ_LNODE_ITER_CURSOR | flags; kit->cursor.priv = READ_ONCE(kit->dsq->seq); return 0; From 542ed8145e6f9392e3d0a86a0e9027d2ffd183e4 Mon Sep 17 00:00:00 2001 From: Pablo Neira Ayuso Date: Sat, 21 Dec 2024 00:29:20 +0100 Subject: [PATCH 085/196] netfilter: nft_set_hash: unaligned atomic read on struct nft_set_ext Access to genmask field in struct nft_set_ext results in unaligned atomic read: [ 72.130109] Unable to handle kernel paging request at virtual address ffff0000c2bb708c [ 72.131036] Mem abort info: [ 72.131213] ESR = 0x0000000096000021 [ 72.131446] EC = 0x25: DABT (current EL), IL = 32 bits [ 72.132209] SET = 0, FnV = 0 [ 72.133216] EA = 0, S1PTW = 0 [ 72.134080] FSC = 0x21: alignment fault [ 72.135593] Data abort info: [ 72.137194] ISV = 0, ISS = 0x00000021, ISS2 = 0x00000000 [ 72.142351] CM = 0, WnR = 0, TnD = 0, TagAccess = 0 [ 72.145989] GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 [ 72.150115] swapper pgtable: 4k pages, 48-bit VAs, pgdp=0000000237d27000 [ 72.154893] [ffff0000c2bb708c] pgd=0000000000000000, p4d=180000023ffff403, pud=180000023f84b403, pmd=180000023f835403, +pte=0068000102bb7707 [ 72.163021] Internal error: Oops: 0000000096000021 [#1] SMP [...] [ 72.170041] CPU: 7 UID: 0 PID: 54 Comm: kworker/7:0 Tainted: G E 6.13.0-rc3+ #2 [ 72.170509] Tainted: [E]=UNSIGNED_MODULE [ 72.170720] Hardware name: QEMU QEMU Virtual Machine, BIOS edk2-stable202302-for-qemu 03/01/2023 [ 72.171192] Workqueue: events_power_efficient nft_rhash_gc [nf_tables] [ 72.171552] pstate: 21400005 (nzCv daif +PAN -UAO -TCO +DIT -SSBS BTYPE=--) [ 72.171915] pc : nft_rhash_gc+0x200/0x2d8 [nf_tables] [ 72.172166] lr : nft_rhash_gc+0x128/0x2d8 [nf_tables] [ 72.172546] sp : ffff800081f2bce0 [ 72.172724] x29: ffff800081f2bd40 x28: ffff0000c2bb708c x27: 0000000000000038 [ 72.173078] x26: ffff0000c6780ef0 x25: ffff0000c643df00 x24: ffff0000c6778f78 [ 72.173431] x23: 000000000000001a x22: ffff0000c4b1f000 x21: ffff0000c6780f78 [ 72.173782] x20: ffff0000c2bb70dc x19: ffff0000c2bb7080 x18: 0000000000000000 [ 72.174135] x17: ffff0000c0a4e1c0 x16: 0000000000003000 x15: 0000ac26d173b978 [ 72.174485] x14: ffffffffffffffff x13: 0000000000000030 x12: ffff0000c6780ef0 [ 72.174841] x11: 0000000000000000 x10: ffff800081f2bcf8 x9 : ffff0000c3000000 [ 72.175193] x8 : 00000000000004be x7 : 0000000000000000 x6 : 0000000000000000 [ 72.175544] x5 : 0000000000000040 x4 : ffff0000c3000010 x3 : 0000000000000000 [ 72.175871] x2 : 0000000000003a98 x1 : ffff0000c2bb708c x0 : 0000000000000004 [ 72.176207] Call trace: [ 72.176316] nft_rhash_gc+0x200/0x2d8 [nf_tables] (P) [ 72.176653] process_one_work+0x178/0x3d0 [ 72.176831] worker_thread+0x200/0x3f0 [ 72.176995] kthread+0xe8/0xf8 [ 72.177130] ret_from_fork+0x10/0x20 [ 72.177289] Code: 54fff984 d503201f d2800080 91003261 (f820303f) [ 72.177557] ---[ end trace 0000000000000000 ]--- Align struct nft_set_ext to word size to address this and documentation it. pahole reports that this increases the size of elements for rhash and pipapo in 8 bytes on x86_64. Fixes: 7ffc7481153b ("netfilter: nft_set_hash: skip duplicated elements pending gc run") Signed-off-by: Pablo Neira Ayuso --- include/net/netfilter/nf_tables.h | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/include/net/netfilter/nf_tables.h b/include/net/netfilter/nf_tables.h index 4afa64c81304..0027beca5cd5 100644 --- a/include/net/netfilter/nf_tables.h +++ b/include/net/netfilter/nf_tables.h @@ -733,15 +733,18 @@ struct nft_set_ext_tmpl { /** * struct nft_set_ext - set extensions * - * @genmask: generation mask + * @genmask: generation mask, but also flags (see NFT_SET_ELEM_DEAD_BIT) * @offset: offsets of individual extension types * @data: beginning of extension data + * + * This structure must be aligned to word size, otherwise atomic bitops + * on genmask field can cause alignment failure on some archs. */ struct nft_set_ext { u8 genmask; u8 offset[NFT_SET_EXT_NUM]; char data[]; -}; +} __aligned(BITS_PER_LONG / 8); static inline void nft_set_ext_prepare(struct nft_set_ext_tmpl *tmpl) { From ce2b93fc1dfa1c82f2576aa571731c4e5dcc8dd7 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Tue, 24 Dec 2024 14:09:15 -1000 Subject: [PATCH 086/196] sched_ext: Fix dsq_local_on selftest The dsp_local_on selftest expects the scheduler to fail by trying to schedule an e.g. CPU-affine task to the wrong CPU. However, this isn't guaranteed to happen in the 1 second window that the test is running. Besides, it's odd to have this particular exception path tested when there are no other tests that verify that the interface is working at all - e.g. the test would pass if dsp_local_on interface is completely broken and fails on any attempt. Flip the test so that it verifies that the feature works. While at it, fix a typo in the info message. Signed-off-by: Tejun Heo Reported-by: Ihor Solodrai Link: http://lkml.kernel.org/r/Z1n9v7Z6iNJ-wKmq@slm.duckdns.org Signed-off-by: Tejun Heo --- tools/testing/selftests/sched_ext/dsp_local_on.bpf.c | 5 ++++- tools/testing/selftests/sched_ext/dsp_local_on.c | 5 +++-- 2 files changed, 7 insertions(+), 3 deletions(-) diff --git a/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c b/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c index 6325bf76f47e..fbda6bf54671 100644 --- a/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c +++ b/tools/testing/selftests/sched_ext/dsp_local_on.bpf.c @@ -43,7 +43,10 @@ void BPF_STRUCT_OPS(dsp_local_on_dispatch, s32 cpu, struct task_struct *prev) if (!p) return; - target = bpf_get_prandom_u32() % nr_cpus; + if (p->nr_cpus_allowed == nr_cpus) + target = bpf_get_prandom_u32() % nr_cpus; + else + target = scx_bpf_task_cpu(p); scx_bpf_dsq_insert(p, SCX_DSQ_LOCAL_ON | target, SCX_SLICE_DFL, 0); bpf_task_release(p); diff --git a/tools/testing/selftests/sched_ext/dsp_local_on.c b/tools/testing/selftests/sched_ext/dsp_local_on.c index 472851b56854..0ff27e57fe43 100644 --- a/tools/testing/selftests/sched_ext/dsp_local_on.c +++ b/tools/testing/selftests/sched_ext/dsp_local_on.c @@ -34,9 +34,10 @@ static enum scx_test_status run(void *ctx) /* Just sleeping is fine, plenty of scheduling events happening */ sleep(1); - SCX_EQ(skel->data->uei.kind, EXIT_KIND(SCX_EXIT_ERROR)); bpf_link__destroy(link); + SCX_EQ(skel->data->uei.kind, EXIT_KIND(SCX_EXIT_UNREG)); + return SCX_TEST_PASS; } @@ -50,7 +51,7 @@ static void cleanup(void *ctx) struct scx_test dsp_local_on = { .name = "dsp_local_on", .description = "Verify we can directly dispatch tasks to a local DSQs " - "from osp.dispatch()", + "from ops.dispatch()", .setup = setup, .run = run, .cleanup = cleanup, From 6b830c6a023ff6e8fe05dbe47a9e5cd276df09ee Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Sat, 21 Dec 2024 12:09:14 +0100 Subject: [PATCH 087/196] netlink: specs: mptcp: add missing 'server-side' attr This attribute is added with the 'created' and 'established' events, but the documentation didn't mention it. The documentation in the UAPI header has been auto-generated by: ./tools/net/ynl/ynl-regen.sh Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241221-net-mptcp-netlink-specs-pm-doc-fixes-v2-1-e54f2db3f844@kernel.org Signed-off-by: Jakub Kicinski --- Documentation/netlink/specs/mptcp_pm.yaml | 6 ++++-- include/uapi/linux/mptcp_pm.h | 11 ++++++----- 2 files changed, 10 insertions(+), 7 deletions(-) diff --git a/Documentation/netlink/specs/mptcp_pm.yaml b/Documentation/netlink/specs/mptcp_pm.yaml index dc190bf838fe..fc0603f51665 100644 --- a/Documentation/netlink/specs/mptcp_pm.yaml +++ b/Documentation/netlink/specs/mptcp_pm.yaml @@ -23,7 +23,8 @@ definitions: - name: created doc: - token, family, saddr4 | saddr6, daddr4 | daddr6, sport, dport + token, family, saddr4 | saddr6, daddr4 | daddr6, sport, dport, + server-side A new MPTCP connection has been created. It is the good time to allocate memory and send ADD_ADDR if needed. Depending on the traffic-patterns it can take a long time until the @@ -31,7 +32,8 @@ definitions: - name: established doc: - token, family, saddr4 | saddr6, daddr4 | daddr6, sport, dport + token, family, saddr4 | saddr6, daddr4 | daddr6, sport, dport, + server-side A MPTCP connection is established (can start new subflows). - name: closed diff --git a/include/uapi/linux/mptcp_pm.h b/include/uapi/linux/mptcp_pm.h index 50589e5dd6a3..b34fd95b6f84 100644 --- a/include/uapi/linux/mptcp_pm.h +++ b/include/uapi/linux/mptcp_pm.h @@ -13,12 +13,13 @@ * enum mptcp_event_type * @MPTCP_EVENT_UNSPEC: unused event * @MPTCP_EVENT_CREATED: token, family, saddr4 | saddr6, daddr4 | daddr6, - * sport, dport A new MPTCP connection has been created. It is the good time - * to allocate memory and send ADD_ADDR if needed. Depending on the - * traffic-patterns it can take a long time until the MPTCP_EVENT_ESTABLISHED - * is sent. + * sport, dport, server-side A new MPTCP connection has been created. It is + * the good time to allocate memory and send ADD_ADDR if needed. Depending on + * the traffic-patterns it can take a long time until the + * MPTCP_EVENT_ESTABLISHED is sent. * @MPTCP_EVENT_ESTABLISHED: token, family, saddr4 | saddr6, daddr4 | daddr6, - * sport, dport A MPTCP connection is established (can start new subflows). + * sport, dport, server-side A MPTCP connection is established (can start new + * subflows). * @MPTCP_EVENT_CLOSED: token A MPTCP connection has stopped. * @MPTCP_EVENT_ANNOUNCED: token, rem_id, family, daddr4 | daddr6 [, dport] A * new address has been announced by the peer. From bea87657b5ee8e6f18af2833ee4b88212ef52d28 Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Sat, 21 Dec 2024 12:09:15 +0100 Subject: [PATCH 088/196] netlink: specs: mptcp: clearly mention attributes The rendered version of the MPTCP events [1] looked strange, because the whole content of the 'doc' was displayed in the same block. It was then not clear that the first words, not even ended by a period, were the attributes that are defined when such events are emitted. These attributes have now been moved to the end, prefixed by 'Attributes:' and ended with a period. Note that '>-' has been added after 'doc:' to allow ':' in the text below. The documentation in the UAPI header has been auto-generated by: ./tools/net/ynl/ynl-regen.sh Link: https://docs.kernel.org/networking/netlink_spec/mptcp_pm.html#event-type [1] Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241221-net-mptcp-netlink-specs-pm-doc-fixes-v2-2-e54f2db3f844@kernel.org Signed-off-by: Jakub Kicinski --- Documentation/netlink/specs/mptcp_pm.yaml | 50 ++++++++++----------- include/uapi/linux/mptcp_pm.h | 53 ++++++++++++----------- 2 files changed, 52 insertions(+), 51 deletions(-) diff --git a/Documentation/netlink/specs/mptcp_pm.yaml b/Documentation/netlink/specs/mptcp_pm.yaml index fc0603f51665..59087a230565 100644 --- a/Documentation/netlink/specs/mptcp_pm.yaml +++ b/Documentation/netlink/specs/mptcp_pm.yaml @@ -22,67 +22,67 @@ definitions: doc: unused event - name: created - doc: - token, family, saddr4 | saddr6, daddr4 | daddr6, sport, dport, - server-side + doc: >- A new MPTCP connection has been created. It is the good time to allocate memory and send ADD_ADDR if needed. Depending on the traffic-patterns it can take a long time until the MPTCP_EVENT_ESTABLISHED is sent. + Attributes: token, family, saddr4 | saddr6, daddr4 | daddr6, sport, + dport, server-side. - name: established - doc: - token, family, saddr4 | saddr6, daddr4 | daddr6, sport, dport, - server-side + doc: >- A MPTCP connection is established (can start new subflows). + Attributes: token, family, saddr4 | saddr6, daddr4 | daddr6, sport, + dport, server-side. - name: closed - doc: - token + doc: >- A MPTCP connection has stopped. + Attribute: token. - name: announced value: 6 - doc: - token, rem_id, family, daddr4 | daddr6 [, dport] + doc: >- A new address has been announced by the peer. + Attributes: token, rem_id, family, daddr4 | daddr6 [, dport]. - name: removed - doc: - token, rem_id + doc: >- An address has been lost by the peer. + Attributes: token, rem_id. - name: sub-established value: 10 - doc: - token, family, loc_id, rem_id, saddr4 | saddr6, daddr4 | daddr6, sport, - dport, backup, if_idx [, error] + doc: >- A new subflow has been established. 'error' should not be set. + Attributes: token, family, loc_id, rem_id, saddr4 | saddr6, daddr4 | + daddr6, sport, dport, backup, if_idx [, error]. - name: sub-closed - doc: - token, family, loc_id, rem_id, saddr4 | saddr6, daddr4 | daddr6, sport, - dport, backup, if_idx [, error] + doc: >- A subflow has been closed. An error (copy of sk_err) could be set if an error has been detected for this subflow. + Attributes: token, family, loc_id, rem_id, saddr4 | saddr6, daddr4 | + daddr6, sport, dport, backup, if_idx [, error]. - name: sub-priority value: 13 - doc: - token, family, loc_id, rem_id, saddr4 | saddr6, daddr4 | daddr6, sport, - dport, backup, if_idx [, error] + doc: >- The priority of a subflow has changed. 'error' should not be set. + Attributes: token, family, loc_id, rem_id, saddr4 | saddr6, daddr4 | + daddr6, sport, dport, backup, if_idx [, error]. - name: listener-created value: 15 - doc: - family, sport, saddr4 | saddr6 + doc: >- A new PM listener is created. + Attributes: family, sport, saddr4 | saddr6. - name: listener-closed - doc: - family, sport, saddr4 | saddr6 + doc: >- A PM listener is closed. + Attributes: family, sport, saddr4 | saddr6. attribute-sets: - diff --git a/include/uapi/linux/mptcp_pm.h b/include/uapi/linux/mptcp_pm.h index b34fd95b6f84..84fa8a21dfd0 100644 --- a/include/uapi/linux/mptcp_pm.h +++ b/include/uapi/linux/mptcp_pm.h @@ -12,32 +12,33 @@ /** * enum mptcp_event_type * @MPTCP_EVENT_UNSPEC: unused event - * @MPTCP_EVENT_CREATED: token, family, saddr4 | saddr6, daddr4 | daddr6, - * sport, dport, server-side A new MPTCP connection has been created. It is - * the good time to allocate memory and send ADD_ADDR if needed. Depending on - * the traffic-patterns it can take a long time until the - * MPTCP_EVENT_ESTABLISHED is sent. - * @MPTCP_EVENT_ESTABLISHED: token, family, saddr4 | saddr6, daddr4 | daddr6, - * sport, dport, server-side A MPTCP connection is established (can start new - * subflows). - * @MPTCP_EVENT_CLOSED: token A MPTCP connection has stopped. - * @MPTCP_EVENT_ANNOUNCED: token, rem_id, family, daddr4 | daddr6 [, dport] A - * new address has been announced by the peer. - * @MPTCP_EVENT_REMOVED: token, rem_id An address has been lost by the peer. - * @MPTCP_EVENT_SUB_ESTABLISHED: token, family, loc_id, rem_id, saddr4 | - * saddr6, daddr4 | daddr6, sport, dport, backup, if_idx [, error] A new - * subflow has been established. 'error' should not be set. - * @MPTCP_EVENT_SUB_CLOSED: token, family, loc_id, rem_id, saddr4 | saddr6, - * daddr4 | daddr6, sport, dport, backup, if_idx [, error] A subflow has been - * closed. An error (copy of sk_err) could be set if an error has been - * detected for this subflow. - * @MPTCP_EVENT_SUB_PRIORITY: token, family, loc_id, rem_id, saddr4 | saddr6, - * daddr4 | daddr6, sport, dport, backup, if_idx [, error] The priority of a - * subflow has changed. 'error' should not be set. - * @MPTCP_EVENT_LISTENER_CREATED: family, sport, saddr4 | saddr6 A new PM - * listener is created. - * @MPTCP_EVENT_LISTENER_CLOSED: family, sport, saddr4 | saddr6 A PM listener - * is closed. + * @MPTCP_EVENT_CREATED: A new MPTCP connection has been created. It is the + * good time to allocate memory and send ADD_ADDR if needed. Depending on the + * traffic-patterns it can take a long time until the MPTCP_EVENT_ESTABLISHED + * is sent. Attributes: token, family, saddr4 | saddr6, daddr4 | daddr6, + * sport, dport, server-side. + * @MPTCP_EVENT_ESTABLISHED: A MPTCP connection is established (can start new + * subflows). Attributes: token, family, saddr4 | saddr6, daddr4 | daddr6, + * sport, dport, server-side. + * @MPTCP_EVENT_CLOSED: A MPTCP connection has stopped. Attribute: token. + * @MPTCP_EVENT_ANNOUNCED: A new address has been announced by the peer. + * Attributes: token, rem_id, family, daddr4 | daddr6 [, dport]. + * @MPTCP_EVENT_REMOVED: An address has been lost by the peer. Attributes: + * token, rem_id. + * @MPTCP_EVENT_SUB_ESTABLISHED: A new subflow has been established. 'error' + * should not be set. Attributes: token, family, loc_id, rem_id, saddr4 | + * saddr6, daddr4 | daddr6, sport, dport, backup, if_idx [, error]. + * @MPTCP_EVENT_SUB_CLOSED: A subflow has been closed. An error (copy of + * sk_err) could be set if an error has been detected for this subflow. + * Attributes: token, family, loc_id, rem_id, saddr4 | saddr6, daddr4 | + * daddr6, sport, dport, backup, if_idx [, error]. + * @MPTCP_EVENT_SUB_PRIORITY: The priority of a subflow has changed. 'error' + * should not be set. Attributes: token, family, loc_id, rem_id, saddr4 | + * saddr6, daddr4 | daddr6, sport, dport, backup, if_idx [, error]. + * @MPTCP_EVENT_LISTENER_CREATED: A new PM listener is created. Attributes: + * family, sport, saddr4 | saddr6. + * @MPTCP_EVENT_LISTENER_CLOSED: A PM listener is closed. Attributes: family, + * sport, saddr4 | saddr6. */ enum mptcp_event_type { MPTCP_EVENT_UNSPEC, From 4f363fe9f6b28ed9b714cd7fe5ce880171927dab Mon Sep 17 00:00:00 2001 From: "Matthieu Baerts (NGI0)" Date: Sat, 21 Dec 2024 12:09:16 +0100 Subject: [PATCH 089/196] netlink: specs: mptcp: fix missing doc Two operations didn't have a small description. It looks like something that has been missed in the original commit introducing this file. Replace the two "todo" by a small and simple description: Create/Destroy subflow. While at it, also uniform the capital letters, avoid double spaces, and fix the "announce" event description: a new "address" has been announced, not a new "subflow". Reviewed-by: Geliang Tang Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241221-net-mptcp-netlink-specs-pm-doc-fixes-v2-3-e54f2db3f844@kernel.org Signed-off-by: Jakub Kicinski --- Documentation/netlink/specs/mptcp_pm.yaml | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/Documentation/netlink/specs/mptcp_pm.yaml b/Documentation/netlink/specs/mptcp_pm.yaml index 59087a230565..dfd017780d2f 100644 --- a/Documentation/netlink/specs/mptcp_pm.yaml +++ b/Documentation/netlink/specs/mptcp_pm.yaml @@ -308,8 +308,8 @@ operations: attributes: - addr - - name: flush-addrs - doc: flush addresses + name: flush-addrs + doc: Flush addresses attribute-set: endpoint dont-validate: [ strict ] flags: [ uns-admin-perm ] @@ -353,7 +353,7 @@ operations: - addr-remote - name: announce - doc: announce new sf + doc: Announce new address attribute-set: attr dont-validate: [ strict ] flags: [ uns-admin-perm ] @@ -364,7 +364,7 @@ operations: - token - name: remove - doc: announce removal + doc: Announce removal attribute-set: attr dont-validate: [ strict ] flags: [ uns-admin-perm ] @@ -375,7 +375,7 @@ operations: - loc-id - name: subflow-create - doc: todo + doc: Create subflow attribute-set: attr dont-validate: [ strict ] flags: [ uns-admin-perm ] @@ -387,7 +387,7 @@ operations: - addr-remote - name: subflow-destroy - doc: todo + doc: Destroy subflow attribute-set: attr dont-validate: [ strict ] flags: [ uns-admin-perm ] From a024e377efed31ecfb39210bed562932321345b3 Mon Sep 17 00:00:00 2001 From: Antonio Pastor Date: Tue, 24 Dec 2024 20:07:20 -0500 Subject: [PATCH 090/196] net: llc: reset skb->transport_header 802.2+LLC+SNAP frames received by napi_complete_done with GRO and DSA have skb->transport_header set two bytes short, or pointing 2 bytes before network_header & skb->data. As snap_rcv expects transport_header to point to SNAP header (OID:PID) after LLC processing advances offset over LLC header (llc_rcv & llc_fixup_skb), code doesn't find a match and packet is dropped. Between napi_complete_done and snap_rcv, transport_header is not used until __netif_receive_skb_core, where originally it was being reset. Commit fda55eca5a33 ("net: introduce skb_transport_header_was_set()") only does so if not set, on the assumption the value was set correctly by GRO (and also on assumption that "network stacks usually reset the transport header anyway"). Afterwards it is moved forward by llc_fixup_skb. Locally generated traffic shows up at __netif_receive_skb_core with no transport_header set and is processed without issue. On a setup with GRO but no DSA, transport_header and network_header are both set to point to skb->data which is also correct. As issue is LLC specific, to avoid impacting non-LLC traffic, and to follow up on original assumption made on previous code change, llc_fixup_skb to reset the offset after skb pull. llc_fixup_skb assumes the LLC header is at skb->data, and by definition SNAP header immediately follows. Fixes: fda55eca5a33 ("net: introduce skb_transport_header_was_set()") Signed-off-by: Antonio Pastor Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/20241225010723.2830290-1-antonio.pastor@gmail.com Signed-off-by: Jakub Kicinski --- net/llc/llc_input.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/net/llc/llc_input.c b/net/llc/llc_input.c index 51bccfb00a9c..61b0159b2fbe 100644 --- a/net/llc/llc_input.c +++ b/net/llc/llc_input.c @@ -124,8 +124,8 @@ static inline int llc_fixup_skb(struct sk_buff *skb) if (unlikely(!pskb_may_pull(skb, llc_len))) return 0; - skb->transport_header += llc_len; skb_pull(skb, llc_len); + skb_reset_transport_header(skb); if (skb->protocol == htons(ETH_P_802_2)) { __be16 pdulen; s32 data_size; From 4db3d750ac7e894278ef1cb1c53cc7d883060496 Mon Sep 17 00:00:00 2001 From: Leo Stone Date: Wed, 18 Dec 2024 10:49:57 -0800 Subject: [PATCH 091/196] nvmet: Don't overflow subsysnqn nvmet_root_discovery_nqn_store treats the subsysnqn string like a fixed size buffer, even though it is dynamically allocated to the size of the string. Create a new string with kstrndup instead of using the old buffer. Reported-by: syzbot+ff4aab278fa7e27e0f9e@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=ff4aab278fa7e27e0f9e Fixes: 95409e277d83 ("nvmet: implement unique discovery NQN") Signed-off-by: Leo Stone Reviewed-by: Sagi Grimberg Reviewed-by: Christoph Hellwig Signed-off-by: Keith Busch --- drivers/nvme/target/configfs.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/drivers/nvme/target/configfs.c b/drivers/nvme/target/configfs.c index eeee9e9b854c..9c109b93ffbf 100644 --- a/drivers/nvme/target/configfs.c +++ b/drivers/nvme/target/configfs.c @@ -2254,12 +2254,17 @@ static ssize_t nvmet_root_discovery_nqn_store(struct config_item *item, const char *page, size_t count) { struct list_head *entry; + char *old_nqn, *new_nqn; size_t len; len = strcspn(page, "\n"); if (!len || len > NVMF_NQN_FIELD_LEN - 1) return -EINVAL; + new_nqn = kstrndup(page, len, GFP_KERNEL); + if (!new_nqn) + return -ENOMEM; + down_write(&nvmet_config_sem); list_for_each(entry, &nvmet_subsystems_group.cg_children) { struct config_item *item = @@ -2268,13 +2273,15 @@ static ssize_t nvmet_root_discovery_nqn_store(struct config_item *item, if (!strncmp(config_item_name(item), page, len)) { pr_err("duplicate NQN %s\n", config_item_name(item)); up_write(&nvmet_config_sem); + kfree(new_nqn); return -EINVAL; } } - memset(nvmet_disc_subsys->subsysnqn, 0, NVMF_NQN_FIELD_LEN); - memcpy(nvmet_disc_subsys->subsysnqn, page, len); + old_nqn = nvmet_disc_subsys->subsysnqn; + nvmet_disc_subsys->subsysnqn = new_nqn; up_write(&nvmet_config_sem); + kfree(old_nqn); return len; } From b579d6fdc3a9149bb4d2b3133cc0767130ed13e6 Mon Sep 17 00:00:00 2001 From: Luis Chamberlain Date: Tue, 17 Dec 2024 18:33:25 -0800 Subject: [PATCH 092/196] nvmet: propagate npwg topology Ensure we propagate npwg to the target as well instead of assuming its the same logical blocks per physical block. This ensures devices with large IUs information properly propagated on the target. Signed-off-by: Luis Chamberlain Reviewed-by: Sagi Grimberg Signed-off-by: Keith Busch --- drivers/nvme/target/io-cmd-bdev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/nvme/target/io-cmd-bdev.c b/drivers/nvme/target/io-cmd-bdev.c index 0bda83d0fc3e..eaf31c823cbe 100644 --- a/drivers/nvme/target/io-cmd-bdev.c +++ b/drivers/nvme/target/io-cmd-bdev.c @@ -36,7 +36,7 @@ void nvmet_bdev_set_limits(struct block_device *bdev, struct nvme_id_ns *id) */ id->nsfeat |= 1 << 4; /* NPWG = Namespace Preferred Write Granularity. 0's based */ - id->npwg = lpp0b; + id->npwg = to0based(bdev_io_min(bdev) / bdev_logical_block_size(bdev)); /* NPWA = Namespace Preferred Write Alignment. 0's based */ id->npwa = id->npwg; /* NPDG = Namespace Preferred Deallocate Granularity. 0's based */ From 74d16965d7ac378d28ebd833ae6d6a097186a4ec Mon Sep 17 00:00:00 2001 From: Nilay Shroff Date: Wed, 11 Dec 2024 14:28:06 +0530 Subject: [PATCH 093/196] nvmet-loop: avoid using mutex in IO hotpath Using mutex lock in IO hot path causes the kernel BUG sleeping while atomic. Shinichiro[1], first encountered this issue while running blktest nvme/052 shown below: BUG: sleeping function called from invalid context at kernel/locking/mutex.c:585 in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 996, name: (udev-worker) preempt_count: 0, expected: 0 RCU nest depth: 1, expected: 0 2 locks held by (udev-worker)/996: #0: ffff8881004570c8 (mapping.invalidate_lock){.+.+}-{3:3}, at: page_cache_ra_unbounded+0x155/0x5c0 #1: ffffffff8607eaa0 (rcu_read_lock){....}-{1:2}, at: blk_mq_flush_plug_list+0xa75/0x1950 CPU: 2 UID: 0 PID: 996 Comm: (udev-worker) Not tainted 6.12.0-rc3+ #339 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.3-2.fc40 04/01/2014 Call Trace: dump_stack_lvl+0x6a/0x90 __might_resched.cold+0x1f7/0x23d ? __pfx___might_resched+0x10/0x10 ? vsnprintf+0xdeb/0x18f0 __mutex_lock+0xf4/0x1220 ? nvmet_subsys_nsid_exists+0xb9/0x150 [nvmet] ? __pfx_vsnprintf+0x10/0x10 ? __pfx___mutex_lock+0x10/0x10 ? snprintf+0xa5/0xe0 ? xas_load+0x1ce/0x3f0 ? nvmet_subsys_nsid_exists+0xb9/0x150 [nvmet] nvmet_subsys_nsid_exists+0xb9/0x150 [nvmet] ? __pfx_nvmet_subsys_nsid_exists+0x10/0x10 [nvmet] nvmet_req_find_ns+0x24e/0x300 [nvmet] nvmet_req_init+0x694/0xd40 [nvmet] ? blk_mq_start_request+0x11c/0x750 ? nvme_setup_cmd+0x369/0x990 [nvme_core] nvme_loop_queue_rq+0x2a7/0x7a0 [nvme_loop] ? __pfx___lock_acquire+0x10/0x10 ? __pfx_nvme_loop_queue_rq+0x10/0x10 [nvme_loop] __blk_mq_issue_directly+0xe2/0x1d0 ? __pfx___blk_mq_issue_directly+0x10/0x10 ? blk_mq_request_issue_directly+0xc2/0x140 blk_mq_plug_issue_direct+0x13f/0x630 ? lock_acquire+0x2d/0xc0 ? blk_mq_flush_plug_list+0xa75/0x1950 blk_mq_flush_plug_list+0xa9d/0x1950 ? __pfx_blk_mq_flush_plug_list+0x10/0x10 ? __pfx_mpage_readahead+0x10/0x10 __blk_flush_plug+0x278/0x4d0 ? __pfx___blk_flush_plug+0x10/0x10 ? lock_release+0x460/0x7a0 blk_finish_plug+0x4e/0x90 read_pages+0x51b/0xbc0 ? __pfx_read_pages+0x10/0x10 ? lock_release+0x460/0x7a0 page_cache_ra_unbounded+0x326/0x5c0 force_page_cache_ra+0x1ea/0x2f0 filemap_get_pages+0x59e/0x17b0 ? __pfx_filemap_get_pages+0x10/0x10 ? lock_is_held_type+0xd5/0x130 ? __pfx___might_resched+0x10/0x10 ? find_held_lock+0x2d/0x110 filemap_read+0x317/0xb70 ? up_write+0x1ba/0x510 ? __pfx_filemap_read+0x10/0x10 ? inode_security+0x54/0xf0 ? selinux_file_permission+0x36d/0x420 blkdev_read_iter+0x143/0x3b0 vfs_read+0x6ac/0xa20 ? __pfx_vfs_read+0x10/0x10 ? __pfx_vm_mmap_pgoff+0x10/0x10 ? __pfx___seccomp_filter+0x10/0x10 ksys_read+0xf7/0x1d0 ? __pfx_ksys_read+0x10/0x10 do_syscall_64+0x93/0x180 ? lockdep_hardirqs_on_prepare+0x16d/0x400 ? do_syscall_64+0x9f/0x180 ? lockdep_hardirqs_on+0x78/0x100 ? do_syscall_64+0x9f/0x180 ? lockdep_hardirqs_on_prepare+0x16d/0x400 entry_SYSCALL_64_after_hwframe+0x76/0x7e RIP: 0033:0x7f565bd1ce11 Code: 00 48 8b 15 09 90 0d 00 f7 d8 64 89 02 b8 ff ff ff ff eb bd e8 d0 ad 01 00 f3 0f 1e fa 80 3d 35 12 0e 00 00 74 13 31 c0 0f 05 <48> 3d 00 f0 ff ff 77 4f c3 66 0f 1f 44 00 00 55 48 89 e5 48 83 ec RSP: 002b:00007ffd6e7a20c8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000 RAX: ffffffffffffffda RBX: 0000000000001000 RCX: 00007f565bd1ce11 RDX: 0000000000001000 RSI: 00007f565babb000 RDI: 0000000000000014 RBP: 00007ffd6e7a2130 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000556000bfa610 R11: 0000000000000246 R12: 000000003ffff000 R13: 0000556000bfa5b0 R14: 0000000000000e00 R15: 0000556000c07328 Apparently, the above issue is caused due to using mutex lock while we're in IO hot path. It's a regression caused with commit 505363957fad ("nvmet: fix nvme status code when namespace is disabled"). The mutex ->su_mutex is used to find whether a disabled nsid exists in the config group or not. This is to differentiate between a nsid that is disabled vs non-existent. To mitigate the above issue, we've worked upon a fix[2] where we now insert nsid in subsys Xarray as soon as it's created under config group and later when that nsid is enabled, we add an Xarray mark on it and set ns->enabled to true. The Xarray mark is useful while we need to loop through all enabled namepsaces under a subsystem using xa_for_each_marked() API. If later a nsid is disabled then we clear Xarray mark from it and also set ns->enabled to false. It's only when nsid is deleted from the config group we delete it from the Xarray. So with this change, now we could easily differentiate a nsid is disabled (i.e. Xarray entry for ns exists but ns->enabled is set to false) vs non- existent (i.e.Xarray entry for ns doesn't exist). Link: https://lore.kernel.org/linux-nvme/20241022070252.GA11389@lst.de/ [2] Reported-by: Shinichiro Kawasaki Closes: https://lore.kernel.org/linux-nvme/tqcy3sveity7p56v7ywp7ssyviwcb3w4623cnxj3knoobfcanq@yxgt2mjkbkam/ [1] Fixes: 505363957fad ("nvmet: fix nvme status code when namespace is disabled") Fix-suggested-by: Christoph Hellwig Reviewed-by: Hannes Reinecke Reviewed-by: Chaitanya Kulkarni Reviewed-by: Sagi Grimberg Reviewed-by: Christoph Hellwig Signed-off-by: Nilay Shroff Signed-off-by: Keith Busch --- drivers/nvme/target/admin-cmd.c | 9 +-- drivers/nvme/target/configfs.c | 12 ---- drivers/nvme/target/core.c | 108 +++++++++++++++++++------------- drivers/nvme/target/nvmet.h | 7 +++ drivers/nvme/target/pr.c | 8 +-- 5 files changed, 79 insertions(+), 65 deletions(-) diff --git a/drivers/nvme/target/admin-cmd.c b/drivers/nvme/target/admin-cmd.c index 2962794ce881..fa89b0549c36 100644 --- a/drivers/nvme/target/admin-cmd.c +++ b/drivers/nvme/target/admin-cmd.c @@ -139,7 +139,7 @@ static u16 nvmet_get_smart_log_all(struct nvmet_req *req, unsigned long idx; ctrl = req->sq->ctrl; - xa_for_each(&ctrl->subsys->namespaces, idx, ns) { + nvmet_for_each_enabled_ns(&ctrl->subsys->namespaces, idx, ns) { /* we don't have the right data for file backed ns */ if (!ns->bdev) continue; @@ -331,9 +331,10 @@ static u32 nvmet_format_ana_group(struct nvmet_req *req, u32 grpid, u32 count = 0; if (!(req->cmd->get_log_page.lsp & NVME_ANA_LOG_RGO)) { - xa_for_each(&ctrl->subsys->namespaces, idx, ns) + nvmet_for_each_enabled_ns(&ctrl->subsys->namespaces, idx, ns) { if (ns->anagrpid == grpid) desc->nsids[count++] = cpu_to_le32(ns->nsid); + } } desc->grpid = cpu_to_le32(grpid); @@ -772,7 +773,7 @@ static void nvmet_execute_identify_endgrp_list(struct nvmet_req *req) goto out; } - xa_for_each(&ctrl->subsys->namespaces, idx, ns) { + nvmet_for_each_enabled_ns(&ctrl->subsys->namespaces, idx, ns) { if (ns->nsid <= min_endgid) continue; @@ -815,7 +816,7 @@ static void nvmet_execute_identify_nslist(struct nvmet_req *req, bool match_css) goto out; } - xa_for_each(&ctrl->subsys->namespaces, idx, ns) { + nvmet_for_each_enabled_ns(&ctrl->subsys->namespaces, idx, ns) { if (ns->nsid <= min_nsid) continue; if (match_css && req->ns->csi != req->cmd->identify.csi) diff --git a/drivers/nvme/target/configfs.c b/drivers/nvme/target/configfs.c index 9c109b93ffbf..2b030f0efc38 100644 --- a/drivers/nvme/target/configfs.c +++ b/drivers/nvme/target/configfs.c @@ -810,18 +810,6 @@ static struct configfs_attribute *nvmet_ns_attrs[] = { NULL, }; -bool nvmet_subsys_nsid_exists(struct nvmet_subsys *subsys, u32 nsid) -{ - struct config_item *ns_item; - char name[12]; - - snprintf(name, sizeof(name), "%u", nsid); - mutex_lock(&subsys->namespaces_group.cg_subsys->su_mutex); - ns_item = config_group_find_item(&subsys->namespaces_group, name); - mutex_unlock(&subsys->namespaces_group.cg_subsys->su_mutex); - return ns_item != NULL; -} - static void nvmet_ns_release(struct config_item *item) { struct nvmet_ns *ns = to_nvmet_ns(item); diff --git a/drivers/nvme/target/core.c b/drivers/nvme/target/core.c index 1f4e9989663b..fde6c555af61 100644 --- a/drivers/nvme/target/core.c +++ b/drivers/nvme/target/core.c @@ -127,7 +127,7 @@ static u32 nvmet_max_nsid(struct nvmet_subsys *subsys) unsigned long idx; u32 nsid = 0; - xa_for_each(&subsys->namespaces, idx, cur) + nvmet_for_each_enabled_ns(&subsys->namespaces, idx, cur) nsid = cur->nsid; return nsid; @@ -441,11 +441,14 @@ u16 nvmet_req_find_ns(struct nvmet_req *req) struct nvmet_subsys *subsys = nvmet_req_subsys(req); req->ns = xa_load(&subsys->namespaces, nsid); - if (unlikely(!req->ns)) { + if (unlikely(!req->ns || !req->ns->enabled)) { req->error_loc = offsetof(struct nvme_common_command, nsid); - if (nvmet_subsys_nsid_exists(subsys, nsid)) - return NVME_SC_INTERNAL_PATH_ERROR; - return NVME_SC_INVALID_NS | NVME_STATUS_DNR; + if (!req->ns) /* ns doesn't exist! */ + return NVME_SC_INVALID_NS | NVME_STATUS_DNR; + + /* ns exists but it's disabled */ + req->ns = NULL; + return NVME_SC_INTERNAL_PATH_ERROR; } percpu_ref_get(&req->ns->ref); @@ -583,8 +586,6 @@ int nvmet_ns_enable(struct nvmet_ns *ns) goto out_unlock; ret = -EMFILE; - if (subsys->nr_namespaces == NVMET_MAX_NAMESPACES) - goto out_unlock; ret = nvmet_bdev_ns_enable(ns); if (ret == -ENOTBLK) @@ -599,38 +600,19 @@ int nvmet_ns_enable(struct nvmet_ns *ns) list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) nvmet_p2pmem_ns_add_p2p(ctrl, ns); - ret = percpu_ref_init(&ns->ref, nvmet_destroy_namespace, - 0, GFP_KERNEL); - if (ret) - goto out_dev_put; - - if (ns->nsid > subsys->max_nsid) - subsys->max_nsid = ns->nsid; - - ret = xa_insert(&subsys->namespaces, ns->nsid, ns, GFP_KERNEL); - if (ret) - goto out_restore_subsys_maxnsid; - if (ns->pr.enable) { ret = nvmet_pr_init_ns(ns); if (ret) - goto out_remove_from_subsys; + goto out_dev_put; } - subsys->nr_namespaces++; - nvmet_ns_changed(subsys, ns->nsid); ns->enabled = true; + xa_set_mark(&subsys->namespaces, ns->nsid, NVMET_NS_ENABLED); ret = 0; out_unlock: mutex_unlock(&subsys->lock); return ret; - -out_remove_from_subsys: - xa_erase(&subsys->namespaces, ns->nsid); -out_restore_subsys_maxnsid: - subsys->max_nsid = nvmet_max_nsid(subsys); - percpu_ref_exit(&ns->ref); out_dev_put: list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) pci_dev_put(radix_tree_delete(&ctrl->p2p_ns_map, ns->nsid)); @@ -649,15 +631,37 @@ void nvmet_ns_disable(struct nvmet_ns *ns) goto out_unlock; ns->enabled = false; - xa_erase(&ns->subsys->namespaces, ns->nsid); - if (ns->nsid == subsys->max_nsid) - subsys->max_nsid = nvmet_max_nsid(subsys); + xa_clear_mark(&subsys->namespaces, ns->nsid, NVMET_NS_ENABLED); list_for_each_entry(ctrl, &subsys->ctrls, subsys_entry) pci_dev_put(radix_tree_delete(&ctrl->p2p_ns_map, ns->nsid)); mutex_unlock(&subsys->lock); + if (ns->pr.enable) + nvmet_pr_exit_ns(ns); + + mutex_lock(&subsys->lock); + nvmet_ns_changed(subsys, ns->nsid); + nvmet_ns_dev_disable(ns); +out_unlock: + mutex_unlock(&subsys->lock); +} + +void nvmet_ns_free(struct nvmet_ns *ns) +{ + struct nvmet_subsys *subsys = ns->subsys; + + nvmet_ns_disable(ns); + + mutex_lock(&subsys->lock); + + xa_erase(&subsys->namespaces, ns->nsid); + if (ns->nsid == subsys->max_nsid) + subsys->max_nsid = nvmet_max_nsid(subsys); + + mutex_unlock(&subsys->lock); + /* * Now that we removed the namespaces from the lookup list, we * can kill the per_cpu ref and wait for any remaining references @@ -671,21 +675,9 @@ void nvmet_ns_disable(struct nvmet_ns *ns) wait_for_completion(&ns->disable_done); percpu_ref_exit(&ns->ref); - if (ns->pr.enable) - nvmet_pr_exit_ns(ns); - mutex_lock(&subsys->lock); - subsys->nr_namespaces--; - nvmet_ns_changed(subsys, ns->nsid); - nvmet_ns_dev_disable(ns); -out_unlock: mutex_unlock(&subsys->lock); -} - -void nvmet_ns_free(struct nvmet_ns *ns) -{ - nvmet_ns_disable(ns); down_write(&nvmet_ana_sem); nvmet_ana_group_enabled[ns->anagrpid]--; @@ -699,15 +691,33 @@ struct nvmet_ns *nvmet_ns_alloc(struct nvmet_subsys *subsys, u32 nsid) { struct nvmet_ns *ns; + mutex_lock(&subsys->lock); + + if (subsys->nr_namespaces == NVMET_MAX_NAMESPACES) + goto out_unlock; + ns = kzalloc(sizeof(*ns), GFP_KERNEL); if (!ns) - return NULL; + goto out_unlock; init_completion(&ns->disable_done); ns->nsid = nsid; ns->subsys = subsys; + if (percpu_ref_init(&ns->ref, nvmet_destroy_namespace, 0, GFP_KERNEL)) + goto out_free; + + if (ns->nsid > subsys->max_nsid) + subsys->max_nsid = nsid; + + if (xa_insert(&subsys->namespaces, ns->nsid, ns, GFP_KERNEL)) + goto out_exit; + + subsys->nr_namespaces++; + + mutex_unlock(&subsys->lock); + down_write(&nvmet_ana_sem); ns->anagrpid = NVMET_DEFAULT_ANA_GRPID; nvmet_ana_group_enabled[ns->anagrpid]++; @@ -718,6 +728,14 @@ struct nvmet_ns *nvmet_ns_alloc(struct nvmet_subsys *subsys, u32 nsid) ns->csi = NVME_CSI_NVM; return ns; +out_exit: + subsys->max_nsid = nvmet_max_nsid(subsys); + percpu_ref_exit(&ns->ref); +out_free: + kfree(ns); +out_unlock: + mutex_unlock(&subsys->lock); + return NULL; } static void nvmet_update_sq_head(struct nvmet_req *req) @@ -1394,7 +1412,7 @@ static void nvmet_setup_p2p_ns_map(struct nvmet_ctrl *ctrl, ctrl->p2p_client = get_device(req->p2p_client); - xa_for_each(&ctrl->subsys->namespaces, idx, ns) + nvmet_for_each_enabled_ns(&ctrl->subsys->namespaces, idx, ns) nvmet_p2pmem_ns_add_p2p(ctrl, ns); } diff --git a/drivers/nvme/target/nvmet.h b/drivers/nvme/target/nvmet.h index 58328b35dc96..7233549f7c8a 100644 --- a/drivers/nvme/target/nvmet.h +++ b/drivers/nvme/target/nvmet.h @@ -24,6 +24,7 @@ #define NVMET_DEFAULT_VS NVME_VS(2, 1, 0) +#define NVMET_NS_ENABLED XA_MARK_1 #define NVMET_ASYNC_EVENTS 4 #define NVMET_ERROR_LOG_SLOTS 128 #define NVMET_NO_ERROR_LOC ((u16)-1) @@ -33,6 +34,12 @@ #define NVMET_FR_MAX_SIZE 8 #define NVMET_PR_LOG_QUEUE_SIZE 64 +#define nvmet_for_each_ns(xa, index, entry) \ + xa_for_each(xa, index, entry) + +#define nvmet_for_each_enabled_ns(xa, index, entry) \ + xa_for_each_marked(xa, index, entry, NVMET_NS_ENABLED) + /* * Supported optional AENs: */ diff --git a/drivers/nvme/target/pr.c b/drivers/nvme/target/pr.c index 90e9f5bbe581..cd22d8333314 100644 --- a/drivers/nvme/target/pr.c +++ b/drivers/nvme/target/pr.c @@ -60,7 +60,7 @@ u16 nvmet_set_feat_resv_notif_mask(struct nvmet_req *req, u32 mask) goto success; } - xa_for_each(&ctrl->subsys->namespaces, idx, ns) { + nvmet_for_each_enabled_ns(&ctrl->subsys->namespaces, idx, ns) { if (ns->pr.enable) WRITE_ONCE(ns->pr.notify_mask, mask); } @@ -1056,7 +1056,7 @@ int nvmet_ctrl_init_pr(struct nvmet_ctrl *ctrl) * nvmet_pr_init_ns(), see more details in nvmet_ns_enable(). * So just check ns->pr.enable. */ - xa_for_each(&subsys->namespaces, idx, ns) { + nvmet_for_each_enabled_ns(&subsys->namespaces, idx, ns) { if (ns->pr.enable) { ret = nvmet_pr_alloc_and_insert_pc_ref(ns, ctrl->cntlid, &ctrl->hostid); @@ -1067,7 +1067,7 @@ int nvmet_ctrl_init_pr(struct nvmet_ctrl *ctrl) return 0; free_per_ctrl_refs: - xa_for_each(&subsys->namespaces, idx, ns) { + nvmet_for_each_enabled_ns(&subsys->namespaces, idx, ns) { if (ns->pr.enable) { pc_ref = xa_erase(&ns->pr_per_ctrl_refs, ctrl->cntlid); if (pc_ref) @@ -1087,7 +1087,7 @@ void nvmet_ctrl_destroy_pr(struct nvmet_ctrl *ctrl) kfifo_free(&ctrl->pr_log_mgr.log_queue); mutex_destroy(&ctrl->pr_log_mgr.lock); - xa_for_each(&ctrl->subsys->namespaces, idx, ns) { + nvmet_for_each_enabled_ns(&ctrl->subsys->namespaces, idx, ns) { if (ns->pr.enable) { pc_ref = xa_erase(&ns->pr_per_ctrl_refs, ctrl->cntlid); if (pc_ref) From 36e3b1f9abe359b2bc25e81bc47b64354e42c9b1 Mon Sep 17 00:00:00 2001 From: "Chunguang.xu" Date: Tue, 3 Dec 2024 11:39:55 +0800 Subject: [PATCH 094/196] nvme-tcp: remove nvme_tcp_destroy_io_queues() Now when destroying the IO queue we call nvme_tcp_stop_io_queues() twice, nvme_tcp_destroy_io_queues() has an unnecessary call. Here we try to remove nvme_tcp_destroy_io_queues() and merge it into nvme_tcp_teardown_io_queues(), simplify the code and align with nvme-rdma, make it easy to maintaince. Signed-off-by: Chunguang.xu Reviewed-by: Sagi Grimberg Reviewed-by: Christoph Hellwig Signed-off-by: Keith Busch --- drivers/nvme/host/tcp.c | 18 +++++++----------- 1 file changed, 7 insertions(+), 11 deletions(-) diff --git a/drivers/nvme/host/tcp.c b/drivers/nvme/host/tcp.c index 28c76a3e1bd2..b127d41dbbfe 100644 --- a/drivers/nvme/host/tcp.c +++ b/drivers/nvme/host/tcp.c @@ -2024,14 +2024,6 @@ static int nvme_tcp_alloc_io_queues(struct nvme_ctrl *ctrl) return __nvme_tcp_alloc_io_queues(ctrl); } -static void nvme_tcp_destroy_io_queues(struct nvme_ctrl *ctrl, bool remove) -{ - nvme_tcp_stop_io_queues(ctrl); - if (remove) - nvme_remove_io_tag_set(ctrl); - nvme_tcp_free_io_queues(ctrl); -} - static int nvme_tcp_configure_io_queues(struct nvme_ctrl *ctrl, bool new) { int ret, nr_queues; @@ -2176,9 +2168,11 @@ static void nvme_tcp_teardown_io_queues(struct nvme_ctrl *ctrl, nvme_sync_io_queues(ctrl); nvme_tcp_stop_io_queues(ctrl); nvme_cancel_tagset(ctrl); - if (remove) + if (remove) { nvme_unquiesce_io_queues(ctrl); - nvme_tcp_destroy_io_queues(ctrl, remove); + nvme_remove_io_tag_set(ctrl); + } + nvme_tcp_free_io_queues(ctrl); } static void nvme_tcp_reconnect_or_remove(struct nvme_ctrl *ctrl, @@ -2267,7 +2261,9 @@ static int nvme_tcp_setup_ctrl(struct nvme_ctrl *ctrl, bool new) nvme_sync_io_queues(ctrl); nvme_tcp_stop_io_queues(ctrl); nvme_cancel_tagset(ctrl); - nvme_tcp_destroy_io_queues(ctrl, new); + if (new) + nvme_remove_io_tag_set(ctrl); + nvme_tcp_free_io_queues(ctrl); } destroy_admin: nvme_stop_keep_alive(ctrl); From 7a6c355b55c051eb37cb15d191241da3aa3d6cba Mon Sep 17 00:00:00 2001 From: Mostafa Saleh Date: Mon, 23 Dec 2024 12:44:53 +0000 Subject: [PATCH 095/196] scripts/mksysmap: Fix escape chars '$' Commit b18b047002b7 ("kbuild: change scripts/mksysmap into sed script") changed the invocation of the script, to call sed directly without shell. That means, the current extra escape that was added in: commit ec336aa83162 ("scripts/mksysmap: Fix badly escaped '$'") for the shell is not correct any more, at the moment the stack traces for nvhe are corrupted: [ 22.840904] kvm [190]: [] __kvm_nvhe_$x.220+0x58/0x9c [ 22.842913] kvm [190]: [] __kvm_nvhe_$x.9+0x44/0x50 [ 22.844112] kvm [190]: [] __kvm_nvhe___skip_pauth_save+0x4/0x4 With this patch: [ 25.793513] kvm [192]: nVHE call trace: [ 25.794141] kvm [192]: [] __kvm_nvhe_hyp_panic+0xb0/0xf4 [ 25.796590] kvm [192]: [] __kvm_nvhe_handle_trap+0xe4/0x188 [ 25.797553] kvm [192]: [] __kvm_nvhe___skip_pauth_save+0x4/0x4 Fixes: b18b047002b7 ("kbuild: change scripts/mksysmap into sed script") Signed-off-by: Mostafa Saleh Reviewed-by: Nathan Chancellor Signed-off-by: Masahiro Yamada --- scripts/mksysmap | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/scripts/mksysmap b/scripts/mksysmap index c12723a04655..3accbdb269ac 100755 --- a/scripts/mksysmap +++ b/scripts/mksysmap @@ -26,7 +26,7 @@ # (do not forget a space before each pattern) # local symbols for ARM, MIPS, etc. -/ \\$/d +/ \$/d # local labels, .LBB, .Ltmpxxx, .L__unnamed_xx, .LASANPC, etc. / \.L/d @@ -39,7 +39,7 @@ / __pi_\.L/d # arm64 local symbols in non-VHE KVM namespace -/ __kvm_nvhe_\\$/d +/ __kvm_nvhe_\$/d / __kvm_nvhe_\.L/d # lld arm/aarch64/mips thunks From bf36b4bf1b9a7a0015610e2f038ee84ddb085de2 Mon Sep 17 00:00:00 2001 From: Masahiro Yamada Date: Thu, 26 Dec 2024 00:33:35 +0900 Subject: [PATCH 096/196] modpost: fix the missed iteration for the max bit in do_input() This loop should iterate over the range from 'min' to 'max' inclusively. The last interation is missed. Fixes: 1d8f430c15b3 ("[PATCH] Input: add modalias support") Signed-off-by: Masahiro Yamada Tested-by: John Paul Adrian Glaubitz --- scripts/mod/file2alias.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/mod/file2alias.c b/scripts/mod/file2alias.c index 5b5745f00eb3..ff263c285977 100644 --- a/scripts/mod/file2alias.c +++ b/scripts/mod/file2alias.c @@ -656,7 +656,7 @@ static void do_input(char *alias, for (i = min / BITS_PER_LONG; i < max / BITS_PER_LONG + 1; i++) arr[i] = TO_NATIVE(arr[i]); - for (i = min; i < max; i++) + for (i = min; i <= max; i++) if (arr[i / BITS_PER_LONG] & (1ULL << (i%BITS_PER_LONG))) sprintf(alias + strlen(alias), "%X,*", i); } From e1352d7ead2b8803689823cd4059c1ec72609ed4 Mon Sep 17 00:00:00 2001 From: Masahiro Yamada Date: Thu, 26 Dec 2024 00:33:36 +0900 Subject: [PATCH 097/196] modpost: refactor do_vmbus_entry() Optimize the size of guid_name[], as it only requires 1 additional byte for '\0' instead of 2. Simplify the loop by incrementing the iterator by 1 instead of 2. Remove the unnecessary TO_NATIVE() call, as the guid is represented as a byte stream. Signed-off-by: Masahiro Yamada Tested-by: John Paul Adrian Glaubitz --- scripts/mod/file2alias.c | 8 +++----- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/scripts/mod/file2alias.c b/scripts/mod/file2alias.c index ff263c285977..2c7b76d4e8ec 100644 --- a/scripts/mod/file2alias.c +++ b/scripts/mod/file2alias.c @@ -812,15 +812,13 @@ static void do_virtio_entry(struct module *mod, void *symval) * Each byte of the guid will be represented by two hex characters * in the name. */ - static void do_vmbus_entry(struct module *mod, void *symval) { - int i; DEF_FIELD_ADDR(symval, hv_vmbus_device_id, guid); - char guid_name[(sizeof(*guid) + 1) * 2]; + char guid_name[sizeof(*guid) * 2 + 1]; - for (i = 0; i < (sizeof(*guid) * 2); i += 2) - sprintf(&guid_name[i], "%02x", TO_NATIVE((guid->b)[i/2])); + for (int i = 0; i < sizeof(*guid); i++) + sprintf(&guid_name[i * 2], "%02x", guid->b[i]); module_alias_printf(mod, false, "vmbus:%s", guid_name); } From 8fe1a63d3d99d86f1bdc034505aad6fc70424737 Mon Sep 17 00:00:00 2001 From: Masahiro Yamada Date: Thu, 26 Dec 2024 00:33:37 +0900 Subject: [PATCH 098/196] modpost: work around unaligned data access error With the latest binutils, modpost fails with a bus error on some architectures such as ARM and sparc64. Since binutils commit 1f1b5e506bf0 ("bfd/ELF: restrict file alignment for object files"), the byte offset to each section (sh_offset) in relocatable ELF is no longer guaranteed to be aligned. modpost parses MODULE_DEVICE_TABLE() data structures, which are usually located in the .rodata section. If it is not properly aligned, unaligned access errors may occur. To address the issue, this commit imports the get_unaligned() helper from include/linux/unaligned.h. The get_unaligned_native() helper caters to the endianness in addition to handling the unaligned access. I slightly refactored do_pcmcia_entry() and do_input() to avoid writing back to an unaligned address. (We would need the put_unaligned() helper to do that.) The addend_*_rel() functions need similar adjustments because the .text sections are not aligned either. It seems that the .symtab, .rel.* and .rela.* sections are still aligned. Keep normal pointer access for these sections to avoid unnecessary performance costs. Reported-by: Paulo Pisati Reported-by: Matthias Klose Closes: https://sourceware.org/bugzilla/show_bug.cgi?id=32435 Reported-by: John Paul Adrian Glaubitz Closes: https://sourceware.org/bugzilla/show_bug.cgi?id=32493 Signed-off-by: Masahiro Yamada Tested-by: John Paul Adrian Glaubitz --- scripts/mod/file2alias.c | 26 +++++++++++++------------- scripts/mod/modpost.c | 24 ++++++++++++------------ scripts/mod/modpost.h | 14 ++++++++++++++ 3 files changed, 39 insertions(+), 25 deletions(-) diff --git a/scripts/mod/file2alias.c b/scripts/mod/file2alias.c index 2c7b76d4e8ec..19ec72a69e90 100644 --- a/scripts/mod/file2alias.c +++ b/scripts/mod/file2alias.c @@ -132,7 +132,8 @@ struct devtable { * based at address m. */ #define DEF_FIELD(m, devid, f) \ - typeof(((struct devid *)0)->f) f = TO_NATIVE(*(typeof(f) *)((m) + OFF_##devid##_##f)) + typeof(((struct devid *)0)->f) f = \ + get_unaligned_native((typeof(f) *)((m) + OFF_##devid##_##f)) /* Define a variable f that holds the address of field f of struct devid * based at address m. Due to the way typeof works, for a field of type @@ -600,7 +601,7 @@ static void do_pnp_card_entry(struct module *mod, void *symval) static void do_pcmcia_entry(struct module *mod, void *symval) { char alias[256] = {}; - unsigned int i; + DEF_FIELD(symval, pcmcia_device_id, match_flags); DEF_FIELD(symval, pcmcia_device_id, manf_id); DEF_FIELD(symval, pcmcia_device_id, card_id); @@ -609,10 +610,6 @@ static void do_pcmcia_entry(struct module *mod, void *symval) DEF_FIELD(symval, pcmcia_device_id, device_no); DEF_FIELD_ADDR(symval, pcmcia_device_id, prod_id_hash); - for (i=0; i<4; i++) { - (*prod_id_hash)[i] = TO_NATIVE((*prod_id_hash)[i]); - } - ADD(alias, "m", match_flags & PCMCIA_DEV_ID_MATCH_MANF_ID, manf_id); ADD(alias, "c", match_flags & PCMCIA_DEV_ID_MATCH_CARD_ID, @@ -623,10 +620,14 @@ static void do_pcmcia_entry(struct module *mod, void *symval) function); ADD(alias, "pfn", match_flags & PCMCIA_DEV_ID_MATCH_DEVICE_NO, device_no); - ADD(alias, "pa", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID1, (*prod_id_hash)[0]); - ADD(alias, "pb", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID2, (*prod_id_hash)[1]); - ADD(alias, "pc", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID3, (*prod_id_hash)[2]); - ADD(alias, "pd", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID4, (*prod_id_hash)[3]); + ADD(alias, "pa", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID1, + get_unaligned_native(*prod_id_hash + 0)); + ADD(alias, "pb", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID2, + get_unaligned_native(*prod_id_hash + 1)); + ADD(alias, "pc", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID3, + get_unaligned_native(*prod_id_hash + 2)); + ADD(alias, "pd", match_flags & PCMCIA_DEV_ID_MATCH_PROD_ID4, + get_unaligned_native(*prod_id_hash + 3)); module_alias_printf(mod, true, "pcmcia:%s", alias); } @@ -654,10 +655,9 @@ static void do_input(char *alias, { unsigned int i; - for (i = min / BITS_PER_LONG; i < max / BITS_PER_LONG + 1; i++) - arr[i] = TO_NATIVE(arr[i]); for (i = min; i <= max; i++) - if (arr[i / BITS_PER_LONG] & (1ULL << (i%BITS_PER_LONG))) + if (get_unaligned_native(arr + i / BITS_PER_LONG) & + (1ULL << (i % BITS_PER_LONG))) sprintf(alias + strlen(alias), "%X,*", i); } diff --git a/scripts/mod/modpost.c b/scripts/mod/modpost.c index 94ee49207a45..7ea59dc4926b 100644 --- a/scripts/mod/modpost.c +++ b/scripts/mod/modpost.c @@ -1138,9 +1138,9 @@ static Elf_Addr addend_386_rel(uint32_t *location, unsigned int r_type) { switch (r_type) { case R_386_32: - return TO_NATIVE(*location); + return get_unaligned_native(location); case R_386_PC32: - return TO_NATIVE(*location) + 4; + return get_unaligned_native(location) + 4; } return (Elf_Addr)(-1); @@ -1161,24 +1161,24 @@ static Elf_Addr addend_arm_rel(void *loc, Elf_Sym *sym, unsigned int r_type) switch (r_type) { case R_ARM_ABS32: case R_ARM_REL32: - inst = TO_NATIVE(*(uint32_t *)loc); + inst = get_unaligned_native((uint32_t *)loc); return inst + sym->st_value; case R_ARM_MOVW_ABS_NC: case R_ARM_MOVT_ABS: - inst = TO_NATIVE(*(uint32_t *)loc); + inst = get_unaligned_native((uint32_t *)loc); offset = sign_extend32(((inst & 0xf0000) >> 4) | (inst & 0xfff), 15); return offset + sym->st_value; case R_ARM_PC24: case R_ARM_CALL: case R_ARM_JUMP24: - inst = TO_NATIVE(*(uint32_t *)loc); + inst = get_unaligned_native((uint32_t *)loc); offset = sign_extend32((inst & 0x00ffffff) << 2, 25); return offset + sym->st_value + 8; case R_ARM_THM_MOVW_ABS_NC: case R_ARM_THM_MOVT_ABS: - upper = TO_NATIVE(*(uint16_t *)loc); - lower = TO_NATIVE(*((uint16_t *)loc + 1)); + upper = get_unaligned_native((uint16_t *)loc); + lower = get_unaligned_native((uint16_t *)loc + 1); offset = sign_extend32(((upper & 0x000f) << 12) | ((upper & 0x0400) << 1) | ((lower & 0x7000) >> 4) | @@ -1195,8 +1195,8 @@ static Elf_Addr addend_arm_rel(void *loc, Elf_Sym *sym, unsigned int r_type) * imm11 = lower[10:0] * imm32 = SignExtend(S:J2:J1:imm6:imm11:'0') */ - upper = TO_NATIVE(*(uint16_t *)loc); - lower = TO_NATIVE(*((uint16_t *)loc + 1)); + upper = get_unaligned_native((uint16_t *)loc); + lower = get_unaligned_native((uint16_t *)loc + 1); sign = (upper >> 10) & 1; j1 = (lower >> 13) & 1; @@ -1219,8 +1219,8 @@ static Elf_Addr addend_arm_rel(void *loc, Elf_Sym *sym, unsigned int r_type) * I2 = NOT(J2 XOR S) * imm32 = SignExtend(S:I1:I2:imm10:imm11:'0') */ - upper = TO_NATIVE(*(uint16_t *)loc); - lower = TO_NATIVE(*((uint16_t *)loc + 1)); + upper = get_unaligned_native((uint16_t *)loc); + lower = get_unaligned_native((uint16_t *)loc + 1); sign = (upper >> 10) & 1; j1 = (lower >> 13) & 1; @@ -1241,7 +1241,7 @@ static Elf_Addr addend_mips_rel(uint32_t *location, unsigned int r_type) { uint32_t inst; - inst = TO_NATIVE(*location); + inst = get_unaligned_native(location); switch (r_type) { case R_MIPS_LO16: return inst & 0xffff; diff --git a/scripts/mod/modpost.h b/scripts/mod/modpost.h index 8b72c227ebf4..ffd0a52a606e 100644 --- a/scripts/mod/modpost.h +++ b/scripts/mod/modpost.h @@ -65,6 +65,20 @@ #define TO_NATIVE(x) \ (target_is_big_endian == host_is_big_endian ? x : bswap(x)) +#define __get_unaligned_t(type, ptr) ({ \ + const struct { type x; } __attribute__((__packed__)) *__pptr = \ + (typeof(__pptr))(ptr); \ + __pptr->x; \ +}) + +#define get_unaligned(ptr) __get_unaligned_t(typeof(*(ptr)), (ptr)) + +#define get_unaligned_native(ptr) \ +({ \ + typeof(*(ptr)) _val = get_unaligned(ptr); \ + TO_NATIVE(_val); \ +}) + #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0])) #define strstarts(str, prefix) (strncmp(str, prefix, strlen(prefix)) == 0) From 38fc96a58ce40257aec79b32e9b310c86907c63c Mon Sep 17 00:00:00 2001 From: Pavel Begunkov Date: Sat, 28 Dec 2024 17:44:52 +0000 Subject: [PATCH 099/196] io_uring/rw: fix downgraded mshot read The io-wq path can downgrade a multishot request to oneshot mode, however io_read_mshot() doesn't handle that and would still post multiple CQEs. That's not allowed, because io_req_post_cqe() requires stricter context requirements. The described can only happen with pollable files that don't support FMODE_NOWAIT, which is an odd combination, so if even allowed it should be fairly rare. Cc: stable@vger.kernel.org Reported-by: chase xd Fixes: bee1d5becdf5b ("io_uring: disable io-wq execution of multishot NOWAIT requests") Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/r/c5c8c4a50a882fd581257b81bf52eee260ac29fd.1735407848.git.asml.silence@gmail.com Signed-off-by: Jens Axboe --- io_uring/rw.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/io_uring/rw.c b/io_uring/rw.c index 0bcb83e4ce3c..29bb3010f9c0 100644 --- a/io_uring/rw.c +++ b/io_uring/rw.c @@ -983,6 +983,8 @@ int io_read_mshot(struct io_kiocb *req, unsigned int issue_flags) io_kbuf_recycle(req, issue_flags); if (ret < 0) req_set_fail(req); + } else if (!(req->flags & REQ_F_APOLL_MULTISHOT)) { + cflags = io_put_kbuf(req, ret, issue_flags); } else { /* * Any successful return value will keep the multishot read From b06a6187ef983f501e93faa56209169752d3bde3 Mon Sep 17 00:00:00 2001 From: Tanya Agarwal Date: Sun, 29 Dec 2024 11:32:42 +0530 Subject: [PATCH 100/196] ALSA: usb-audio: US16x08: Initialize array before use Initialize meter_urb array before use in mixer_us16x08.c. CID 1410197: (#1 of 1): Uninitialized scalar variable (UNINIT) uninit_use_in_call: Using uninitialized value *meter_urb when calling get_meter_levels_from_urb. Coverity Link: https://scan7.scan.coverity.com/#/project-view/52849/11354?selectedIssue=1410197 Fixes: d2bb390a2081 ("ALSA: usb-audio: Tascam US-16x08 DSP mixer quirk") Signed-off-by: Tanya Agarwal Link: https://patch.msgid.link/20241229060240.1642-1-tanyaagarwal25699@gmail.com Signed-off-by: Takashi Iwai --- sound/usb/mixer_us16x08.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/usb/mixer_us16x08.c b/sound/usb/mixer_us16x08.c index 6eb7d93b358d..20ac32635f1f 100644 --- a/sound/usb/mixer_us16x08.c +++ b/sound/usb/mixer_us16x08.c @@ -687,7 +687,7 @@ static int snd_us16x08_meter_get(struct snd_kcontrol *kcontrol, struct usb_mixer_elem_info *elem = kcontrol->private_data; struct snd_usb_audio *chip = elem->head.mixer->chip; struct snd_us16x08_meter_store *store = elem->private_data; - u8 meter_urb[64]; + u8 meter_urb[64] = {0}; switch (kcontrol->private_value) { case 0: { From 032fe9b0516702599c2dd990a4703f783d5716b8 Mon Sep 17 00:00:00 2001 From: Mingcong Bai Date: Thu, 26 Dec 2024 14:22:05 +0800 Subject: [PATCH 101/196] platform/x86: hp-wmi: mark 8A15 board for timed OMEN thermal profile MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The HP OMEN 8 (2022), corresponding to a board ID of 8A15, supports OMEN thermal profile and requires the timed profile quirk. Upon adding this ID to both the omen_thermal_profile_boards and omen_timed_thermal_profile_boards, significant bump in performance can be observed. For instance, SilverBench (https://silver.urih.com/) results improved from ~56,000 to ~69,000, as a result of higher power draws (and thus core frequencies) whilst under load: Package Power: - Before the patch: ~65W (dropping to about 55W under sustained load). - After the patch: ~115W (dropping to about 105W under sustained load). Core Power: - Before: ~60W (ditto above). - After: ~108W (ditto above). Add 8A15 to omen_thermal_profile_boards and omen_timed_thermal_profile_boards to improve performance. Signed-off-by: Xi Xiao <1577912515@qq.com> Signed-off-by: Mingcong Bai Link: https://lore.kernel.org/r/20241226062207.3352629-1-jeffbai@aosc.io Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen --- drivers/platform/x86/hp/hp-wmi.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/platform/x86/hp/hp-wmi.c b/drivers/platform/x86/hp/hp-wmi.c index 81ccc96ffe40..20c55bab3b8c 100644 --- a/drivers/platform/x86/hp/hp-wmi.c +++ b/drivers/platform/x86/hp/hp-wmi.c @@ -64,7 +64,7 @@ static const char * const omen_thermal_profile_boards[] = { "874A", "8603", "8604", "8748", "886B", "886C", "878A", "878B", "878C", "88C8", "88CB", "8786", "8787", "8788", "88D1", "88D2", "88F4", "88FD", "88F5", "88F6", "88F7", "88FE", "88FF", "8900", "8901", "8902", "8912", - "8917", "8918", "8949", "894A", "89EB", "8BAD", "8A42" + "8917", "8918", "8949", "894A", "89EB", "8BAD", "8A42", "8A15" }; /* DMI Board names of Omen laptops that are specifically set to be thermal @@ -80,7 +80,7 @@ static const char * const omen_thermal_profile_force_v0_boards[] = { * "balanced" when reaching zero. */ static const char * const omen_timed_thermal_profile_boards[] = { - "8BAD", "8A42" + "8BAD", "8A42", "8A15" }; /* DMI Board names of Victus laptops */ From 7e16ae558a87ac9099b6a93a43f19b42d809fd78 Mon Sep 17 00:00:00 2001 From: Vishnu Sankar Date: Sat, 28 Dec 2024 08:18:40 +0900 Subject: [PATCH 102/196] platform/x86: thinkpad-acpi: Add support for hotkey 0x1401 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit F8 mode key on Lenovo 2025 platforms use a different key code. Adding support for the new keycode 0x1401. Tested on X1 Carbon Gen 13 and X1 2-in-1 Gen 10. Signed-off-by: Vishnu Sankar Reviewed-by: Mark Pearson Link: https://lore.kernel.org/r/20241227231840.21334-1-vishnuocv@gmail.com Reviewed-by: Ilpo Järvinen Signed-off-by: Ilpo Järvinen --- Documentation/admin-guide/laptops/thinkpad-acpi.rst | 10 +++++++--- drivers/platform/x86/thinkpad_acpi.c | 4 +++- 2 files changed, 10 insertions(+), 4 deletions(-) diff --git a/Documentation/admin-guide/laptops/thinkpad-acpi.rst b/Documentation/admin-guide/laptops/thinkpad-acpi.rst index 7f674a6cfa8a..4ab0fef7d440 100644 --- a/Documentation/admin-guide/laptops/thinkpad-acpi.rst +++ b/Documentation/admin-guide/laptops/thinkpad-acpi.rst @@ -445,8 +445,10 @@ event code Key Notes 0x1008 0x07 FN+F8 IBM: toggle screen expand Lenovo: configure UltraNav, or toggle screen expand. - On newer platforms (2024+) - replaced by 0x131f (see below) + On 2024 platforms replaced by + 0x131f (see below) and on newer + platforms (2025 +) keycode is + replaced by 0x1401 (see below). 0x1009 0x08 FN+F9 - @@ -506,9 +508,11 @@ event code Key Notes 0x1019 0x18 unknown -0x131f ... FN+F8 Platform Mode change. +0x131f ... FN+F8 Platform Mode change (2024 systems). Implemented in driver. +0x1401 ... FN+F8 Platform Mode change (2025 + systems). + Implemented in driver. ... ... ... 0x1020 0x1F unknown diff --git a/drivers/platform/x86/thinkpad_acpi.c b/drivers/platform/x86/thinkpad_acpi.c index 6371a9f765c1..2cfb2ac3f465 100644 --- a/drivers/platform/x86/thinkpad_acpi.c +++ b/drivers/platform/x86/thinkpad_acpi.c @@ -184,7 +184,8 @@ enum tpacpi_hkey_event_t { */ TP_HKEY_EV_AMT_TOGGLE = 0x131a, /* Toggle AMT on/off */ TP_HKEY_EV_DOUBLETAP_TOGGLE = 0x131c, /* Toggle trackpoint doubletap on/off */ - TP_HKEY_EV_PROFILE_TOGGLE = 0x131f, /* Toggle platform profile */ + TP_HKEY_EV_PROFILE_TOGGLE = 0x131f, /* Toggle platform profile in 2024 systems */ + TP_HKEY_EV_PROFILE_TOGGLE2 = 0x1401, /* Toggle platform profile in 2025 + systems */ /* Reasons for waking up from S3/S4 */ TP_HKEY_EV_WKUP_S3_UNDOCK = 0x2304, /* undock requested, S3 */ @@ -11200,6 +11201,7 @@ static bool tpacpi_driver_event(const unsigned int hkey_event) tp_features.trackpoint_doubletap = !tp_features.trackpoint_doubletap; return true; case TP_HKEY_EV_PROFILE_TOGGLE: + case TP_HKEY_EV_PROFILE_TOGGLE2: platform_profile_cycle(); return true; } From 6a451e2c5c03e27aa3ec36be424fccaa286c3ccd Mon Sep 17 00:00:00 2001 From: Baojun Xu Date: Mon, 30 Dec 2024 14:49:10 +0800 Subject: [PATCH 103/196] ALSA: hda/tas2781: Ignore SUBSYS_ID not found for tas2563 projects Driver will return error if no SUBSYS_ID found in BIOS(acpi). It will cause error in tas2563 projects, which have no SUBSYS_ID. Fixes: 4e7035a75da9 ("ALSA: hda/tas2781: Add speaker id check for ASUS projects") Signed-off-by: Baojun Xu Link: https://lore.kernel.org/20241223225442.1358491-1-stuart.a.hayhurst@gmail.com Link: https://patch.msgid.link/20241230064910.1583-1-baojun.xu@ti.com Signed-off-by: Takashi Iwai --- sound/pci/hda/tas2781_hda_i2c.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/sound/pci/hda/tas2781_hda_i2c.c b/sound/pci/hda/tas2781_hda_i2c.c index 0af015806aba..0e42b87dadb8 100644 --- a/sound/pci/hda/tas2781_hda_i2c.c +++ b/sound/pci/hda/tas2781_hda_i2c.c @@ -142,6 +142,9 @@ static int tas2781_read_acpi(struct tasdevice_priv *p, const char *hid) } sub = acpi_get_subsystem_id(ACPI_HANDLE(physdev)); if (IS_ERR(sub)) { + /* No subsys id in older tas2563 projects. */ + if (!strncmp(hid, "INT8866", sizeof("INT8866"))) + goto end_2563; dev_err(p->dev, "Failed to get SUBSYS ID.\n"); ret = PTR_ERR(sub); goto err; @@ -164,6 +167,7 @@ static int tas2781_read_acpi(struct tasdevice_priv *p, const char *hid) p->speaker_id = NULL; } +end_2563: acpi_dev_free_resource_list(&resources); strscpy(p->dev_name, hid, sizeof(p->dev_name)); put_device(physdev); From ac9fae799eda81e24bbf2e0d5cb9e5c33fc9bdcb Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Sun, 29 Dec 2024 09:39:13 +0100 Subject: [PATCH 104/196] ALSA: compress_offload: Drop unneeded no_free_ptr() The error path for memdup_user() no longer needs the tricky wrap with no_free_ptr() and we can safely return the error pointer directly. Fixes: 04177158cf98 ("ALSA: compress_offload: introduce accel operation mode") Reported-by: kernel test robot Closes: https://lore.kernel.org/oe-kbuild-all/202412290846.cncnpGaw-lkp@intel.com/ Link: https://patch.msgid.link/20241229083917.14912-1-tiwai@suse.de Signed-off-by: Takashi Iwai --- sound/core/compress_offload.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/sound/core/compress_offload.c b/sound/core/compress_offload.c index edf5aadf38e5..4ed6cec5fd5c 100644 --- a/sound/core/compress_offload.c +++ b/sound/core/compress_offload.c @@ -1077,7 +1077,7 @@ static int snd_compr_task_create(struct snd_compr_stream *stream, unsigned long return -EPERM; task = memdup_user((void __user *)arg, sizeof(*task)); if (IS_ERR(task)) - return PTR_ERR(no_free_ptr(task)); + return PTR_ERR(task); retval = snd_compr_task_new(stream, task); if (retval >= 0) if (copy_to_user((void __user *)arg, task, sizeof(*task))) @@ -1138,7 +1138,7 @@ static int snd_compr_task_start_ioctl(struct snd_compr_stream *stream, unsigned return -EPERM; task = memdup_user((void __user *)arg, sizeof(*task)); if (IS_ERR(task)) - return PTR_ERR(no_free_ptr(task)); + return PTR_ERR(task); retval = snd_compr_task_start(stream, task); if (retval >= 0) if (copy_to_user((void __user *)arg, task, sizeof(*task))) @@ -1229,7 +1229,7 @@ static int snd_compr_task_status_ioctl(struct snd_compr_stream *stream, unsigned return -EPERM; status = memdup_user((void __user *)arg, sizeof(*status)); if (IS_ERR(status)) - return PTR_ERR(no_free_ptr(status)); + return PTR_ERR(status); retval = snd_compr_task_status(stream, status); if (retval >= 0) if (copy_to_user((void __user *)arg, status, sizeof(*status))) From 7439b395211874e20c24b2fe0e4903864357a3f5 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Sun, 29 Dec 2024 18:52:32 +0000 Subject: [PATCH 105/196] ALSA: compress_offload: fix remaining descriptor races in sound/core/compress_offload.c 3d3f43fab4cf ("ALSA: compress_offload: improve file descriptors installation for dma-buf") fixed some of descriptor races in snd_compr_task_new(), but there's a couple more left. We need to grab the references to dmabuf before moving them into descriptor table - trying to do that by descriptor afterwards might end up getting a different object, with a dangling reference left in task->{input,output} Fixes: 3d3f43fab4cf ("ALSA: compress_offload: improve file descriptors installation for dma-buf") Signed-off-by: Al Viro Link: https://patch.msgid.link/20241229185232.GA1977892@ZenIV Signed-off-by: Takashi Iwai --- sound/core/compress_offload.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/sound/core/compress_offload.c b/sound/core/compress_offload.c index 4ed6cec5fd5c..840bb9cfe789 100644 --- a/sound/core/compress_offload.c +++ b/sound/core/compress_offload.c @@ -1053,13 +1053,13 @@ static int snd_compr_task_new(struct snd_compr_stream *stream, struct snd_compr_ put_unused_fd(fd_i); goto cleanup; } + /* keep dmabuf reference until freed with task free ioctl */ + get_dma_buf(task->input); + get_dma_buf(task->output); fd_install(fd_i, task->input->file); fd_install(fd_o, task->output->file); utask->input_fd = fd_i; utask->output_fd = fd_o; - /* keep dmabuf reference until freed with task free ioctl */ - dma_buf_get(utask->input_fd); - dma_buf_get(utask->output_fd); list_add_tail(&task->list, &stream->runtime->tasks); stream->runtime->total_tasks++; return 0; From 0179488ca992d79908b8e26b9213f1554fc5bacc Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Mon, 30 Dec 2024 12:05:35 +0100 Subject: [PATCH 106/196] ALSA: seq: oss: Fix races at processing SysEx messages OSS sequencer handles the SysEx messages split in 6 bytes packets, and ALSA sequencer OSS layer tries to combine those. It stores the data in the internal buffer and this access is racy as of now, which may lead to the out-of-bounds access. As a temporary band-aid fix, introduce a mutex for serializing the process of the SysEx message packets. Reported-by: Kun Hu Closes: https://lore.kernel.org/2B7E93E4-B13A-4AE4-8E87-306A8EE9BBB7@m.fudan.edu.cn Cc: Link: https://patch.msgid.link/20241230110543.32454-1-tiwai@suse.de Signed-off-by: Takashi Iwai --- sound/core/seq/oss/seq_oss_synth.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/sound/core/seq/oss/seq_oss_synth.c b/sound/core/seq/oss/seq_oss_synth.c index e3394919daa0..51ee4c00a843 100644 --- a/sound/core/seq/oss/seq_oss_synth.c +++ b/sound/core/seq/oss/seq_oss_synth.c @@ -66,6 +66,7 @@ static struct seq_oss_synth midi_synth_dev = { }; static DEFINE_SPINLOCK(register_lock); +static DEFINE_MUTEX(sysex_mutex); /* * prototypes @@ -497,6 +498,7 @@ snd_seq_oss_synth_sysex(struct seq_oss_devinfo *dp, int dev, unsigned char *buf, if (!info) return -ENXIO; + guard(mutex)(&sysex_mutex); sysex = info->sysex; if (sysex == NULL) { sysex = kzalloc(sizeof(*sysex), GFP_KERNEL); From abbff41b6932cde359589fd51f4024b7c85f366b Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Mon, 30 Dec 2024 12:40:22 +0100 Subject: [PATCH 107/196] Revert "ALSA: ump: Don't enumeration invalid groups for legacy rawmidi" This reverts commit c2d188e137e77294323132a760a4608321a36a70. Although it's fine to filter the invalid UMP groups at the first probe time, this will become a problem when UMP groups are updated and (re-)activated. Then there is no way to re-add the substreams properly for the legacy rawmidi, and the new active groups will be still invisible. So let's revert the change. This will move back to showing the full 16 groups, but it's better than forever lost. Link: https://patch.msgid.link/20241230114023.3787-1-tiwai@suse.de Signed-off-by: Takashi Iwai --- sound/core/ump.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/core/ump.c b/sound/core/ump.c index fe4d39ae1159..9198bff4768c 100644 --- a/sound/core/ump.c +++ b/sound/core/ump.c @@ -1244,7 +1244,7 @@ static int fill_legacy_mapping(struct snd_ump_endpoint *ump) num = 0; for (i = 0; i < SNDRV_UMP_MAX_GROUPS; i++) - if ((group_maps & (1U << i)) && ump->groups[i].valid) + if (group_maps & (1U << i)) ump->legacy_mapping[num++] = i; return num; From cc0dc9e871a91aadf5b26a2d7760fb762e0d9203 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Cl=C3=A9ment=20Le=20Goffic?= Date: Wed, 18 Dec 2024 10:22:27 +0100 Subject: [PATCH 108/196] watchdog: stm32_iwdg: fix error message during driver probe MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The commit 3ab1663af6c1 ("watchdog: stm32_iwdg: Add pretimeout support") introduces the support for the pre-timeout interrupt. The support for this interrupt is optional but the driver uses the platform_get_irq() which produces an error message during the driver probe if we don't have any `interrupts` property in the DT. Use the platform_get_irq_optional() API to get rid of the error message as this property is optional. Fixes: 3ab1663af6c1 ("watchdog: stm32_iwdg: Add pretimeout support") Signed-off-by: Clément Le Goffic Reviewed-by: Marek Vasut Reviewed-by: Guenter Roeck Link: https://lore.kernel.org/r/20241218092227.771133-1-clement.legoffic@foss.st.com Signed-off-by: Guenter Roeck Signed-off-by: Wim Van Sebroeck --- drivers/watchdog/stm32_iwdg.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/watchdog/stm32_iwdg.c b/drivers/watchdog/stm32_iwdg.c index d700e0d49bb9..8ad06b54c5ad 100644 --- a/drivers/watchdog/stm32_iwdg.c +++ b/drivers/watchdog/stm32_iwdg.c @@ -286,7 +286,7 @@ static int stm32_iwdg_irq_init(struct platform_device *pdev, if (!wdt->data->has_early_wakeup) return 0; - irq = platform_get_irq(pdev, 0); + irq = platform_get_irq_optional(pdev, 0); if (irq <= 0) return 0; From de30d74f58cbecb3894c7738985bd0086d04bec1 Mon Sep 17 00:00:00 2001 From: Steven Davis Date: Mon, 30 Dec 2024 19:34:31 +0000 Subject: [PATCH 109/196] cdrom: Fix typo, 'devicen' to 'device' Fix typo in cd_dbg line to add trailing newline character. Signed-off-by: Steven Davis Link: https://lore.kernel.org/lkml/20241229165744.21725-1-goldside000@outlook.com Reviewed-by: Phillip Potter Link: https://lore.kernel.org/lkml/Z3GV2W_MUOw5BrtR@equinox Signed-off-by: Phillip Potter Link: https://lore.kernel.org/r/20241230193431.441120-2-phil@philpotter.co.uk Signed-off-by: Jens Axboe --- drivers/cdrom/cdrom.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/cdrom/cdrom.c b/drivers/cdrom/cdrom.c index 6a99a459b80b..51745ed1bbab 100644 --- a/drivers/cdrom/cdrom.c +++ b/drivers/cdrom/cdrom.c @@ -1106,7 +1106,7 @@ int open_for_data(struct cdrom_device_info *cdi) } } - cd_dbg(CD_OPEN, "all seems well, opening the devicen"); + cd_dbg(CD_OPEN, "all seems well, opening the device\n"); /* all seems well, we can open the device */ ret = cdo->open(cdi, 0); /* open for data */ From a9c83a0ab66a5b02e914daed502fb8d3a8d3d619 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Mon, 30 Dec 2024 14:15:17 -0700 Subject: [PATCH 110/196] io_uring/timeout: flush timeouts outside of the timeout lock syzbot reports that a recent fix causes nesting issues between the (now) raw timeoutlock and the eventfd locking: ============================= [ BUG: Invalid wait context ] 6.13.0-rc4-00080-g9828a4c0901f #29 Not tainted ----------------------------- kworker/u32:0/68094 is trying to lock: ffff000014d7a520 (&ctx->wqh#2){..-.}-{3:3}, at: eventfd_signal_mask+0x64/0x180 other info that might help us debug this: context-{5:5} 6 locks held by kworker/u32:0/68094: #0: ffff0000c1d98148 ((wq_completion)iou_exit){+.+.}-{0:0}, at: process_one_work+0x4e8/0xfc0 #1: ffff80008d927c78 ((work_completion)(&ctx->exit_work)){+.+.}-{0:0}, at: process_one_work+0x53c/0xfc0 #2: ffff0000c59bc3d8 (&ctx->completion_lock){+.+.}-{3:3}, at: io_kill_timeouts+0x40/0x180 #3: ffff0000c59bc358 (&ctx->timeout_lock){-.-.}-{2:2}, at: io_kill_timeouts+0x48/0x180 #4: ffff800085127aa0 (rcu_read_lock){....}-{1:3}, at: rcu_lock_acquire+0x8/0x38 #5: ffff800085127aa0 (rcu_read_lock){....}-{1:3}, at: rcu_lock_acquire+0x8/0x38 stack backtrace: CPU: 7 UID: 0 PID: 68094 Comm: kworker/u32:0 Not tainted 6.13.0-rc4-00080-g9828a4c0901f #29 Hardware name: linux,dummy-virt (DT) Workqueue: iou_exit io_ring_exit_work Call trace: show_stack+0x1c/0x30 (C) __dump_stack+0x24/0x30 dump_stack_lvl+0x60/0x80 dump_stack+0x14/0x20 __lock_acquire+0x19f8/0x60c8 lock_acquire+0x1a4/0x540 _raw_spin_lock_irqsave+0x90/0xd0 eventfd_signal_mask+0x64/0x180 io_eventfd_signal+0x64/0x108 io_req_local_work_add+0x294/0x430 __io_req_task_work_add+0x1c0/0x270 io_kill_timeout+0x1f0/0x288 io_kill_timeouts+0xd4/0x180 io_uring_try_cancel_requests+0x2e8/0x388 io_ring_exit_work+0x150/0x550 process_one_work+0x5e8/0xfc0 worker_thread+0x7ec/0xc80 kthread+0x24c/0x300 ret_from_fork+0x10/0x20 because after the preempt-rt fix for the timeout lock nesting inside the io-wq lock, we now have the eventfd spinlock nesting inside the raw timeout spinlock. Rather than play whack-a-mole with other nesting on the timeout lock, split the deletion and killing of timeouts so queueing the task_work for the timeout cancelations can get done outside of the timeout lock. Reported-by: syzbot+b1fc199a40b65d601b65@syzkaller.appspotmail.com Fixes: 020b40f35624 ("io_uring: make ctx->timeout_lock a raw spinlock") Signed-off-by: Jens Axboe --- io_uring/timeout.c | 45 +++++++++++++++++++++++++++++++-------------- 1 file changed, 31 insertions(+), 14 deletions(-) diff --git a/io_uring/timeout.c b/io_uring/timeout.c index bbe58638eca7..362689b17ccc 100644 --- a/io_uring/timeout.c +++ b/io_uring/timeout.c @@ -85,7 +85,27 @@ static void io_timeout_complete(struct io_kiocb *req, struct io_tw_state *ts) io_req_task_complete(req, ts); } -static bool io_kill_timeout(struct io_kiocb *req, int status) +static __cold bool io_flush_killed_timeouts(struct list_head *list, int err) +{ + if (list_empty(list)) + return false; + + while (!list_empty(list)) { + struct io_timeout *timeout; + struct io_kiocb *req; + + timeout = list_first_entry(list, struct io_timeout, list); + list_del_init(&timeout->list); + req = cmd_to_io_kiocb(timeout); + if (err) + req_set_fail(req); + io_req_queue_tw_complete(req, err); + } + + return true; +} + +static void io_kill_timeout(struct io_kiocb *req, struct list_head *list) __must_hold(&req->ctx->timeout_lock) { struct io_timeout_data *io = req->async_data; @@ -93,21 +113,17 @@ static bool io_kill_timeout(struct io_kiocb *req, int status) if (hrtimer_try_to_cancel(&io->timer) != -1) { struct io_timeout *timeout = io_kiocb_to_cmd(req, struct io_timeout); - if (status) - req_set_fail(req); atomic_set(&req->ctx->cq_timeouts, atomic_read(&req->ctx->cq_timeouts) + 1); - list_del_init(&timeout->list); - io_req_queue_tw_complete(req, status); - return true; + list_move_tail(&timeout->list, list); } - return false; } __cold void io_flush_timeouts(struct io_ring_ctx *ctx) { - u32 seq; struct io_timeout *timeout, *tmp; + LIST_HEAD(list); + u32 seq; raw_spin_lock_irq(&ctx->timeout_lock); seq = ctx->cached_cq_tail - atomic_read(&ctx->cq_timeouts); @@ -131,10 +147,11 @@ __cold void io_flush_timeouts(struct io_ring_ctx *ctx) if (events_got < events_needed) break; - io_kill_timeout(req, 0); + io_kill_timeout(req, &list); } ctx->cq_last_tm_flush = seq; raw_spin_unlock_irq(&ctx->timeout_lock); + io_flush_killed_timeouts(&list, 0); } static void io_req_tw_fail_links(struct io_kiocb *link, struct io_tw_state *ts) @@ -661,7 +678,7 @@ __cold bool io_kill_timeouts(struct io_ring_ctx *ctx, struct io_uring_task *tctx bool cancel_all) { struct io_timeout *timeout, *tmp; - int canceled = 0; + LIST_HEAD(list); /* * completion_lock is needed for io_match_task(). Take it before @@ -672,11 +689,11 @@ __cold bool io_kill_timeouts(struct io_ring_ctx *ctx, struct io_uring_task *tctx list_for_each_entry_safe(timeout, tmp, &ctx->timeout_list, list) { struct io_kiocb *req = cmd_to_io_kiocb(timeout); - if (io_match_task(req, tctx, cancel_all) && - io_kill_timeout(req, -ECANCELED)) - canceled++; + if (io_match_task(req, tctx, cancel_all)) + io_kill_timeout(req, &list); } raw_spin_unlock_irq(&ctx->timeout_lock); spin_unlock(&ctx->completion_lock); - return canceled != 0; + + return io_flush_killed_timeouts(&list, -ECANCELED); } From b255ef45fcc2141c1bf98456796abb956d843a27 Mon Sep 17 00:00:00 2001 From: Vitalii Mordan Date: Fri, 27 Dec 2024 15:30:07 +0300 Subject: [PATCH 111/196] eth: bcmsysport: fix call balance of priv->clk handling routines Check the return value of clk_prepare_enable to ensure that priv->clk has been successfully enabled. If priv->clk was not enabled during bcm_sysport_probe, bcm_sysport_resume, or bcm_sysport_open, it must not be disabled in any subsequent execution paths. Fixes: 31bc72d97656 ("net: systemport: fetch and use clock resources") Signed-off-by: Vitalii Mordan Reviewed-by: Florian Fainelli Link: https://patch.msgid.link/20241227123007.2333397-1-mordan@ispras.ru Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/broadcom/bcmsysport.c | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/broadcom/bcmsysport.c b/drivers/net/ethernet/broadcom/bcmsysport.c index 42672c63f108..bc4e1f3b3752 100644 --- a/drivers/net/ethernet/broadcom/bcmsysport.c +++ b/drivers/net/ethernet/broadcom/bcmsysport.c @@ -1933,7 +1933,11 @@ static int bcm_sysport_open(struct net_device *dev) unsigned int i; int ret; - clk_prepare_enable(priv->clk); + ret = clk_prepare_enable(priv->clk); + if (ret) { + netdev_err(dev, "could not enable priv clock\n"); + return ret; + } /* Reset UniMAC */ umac_reset(priv); @@ -2591,7 +2595,11 @@ static int bcm_sysport_probe(struct platform_device *pdev) goto err_deregister_notifier; } - clk_prepare_enable(priv->clk); + ret = clk_prepare_enable(priv->clk); + if (ret) { + dev_err(&pdev->dev, "could not enable priv clock\n"); + goto err_deregister_netdev; + } priv->rev = topctrl_readl(priv, REV_CNTL) & REV_MASK; dev_info(&pdev->dev, @@ -2605,6 +2613,8 @@ static int bcm_sysport_probe(struct platform_device *pdev) return 0; +err_deregister_netdev: + unregister_netdev(dev); err_deregister_notifier: unregister_netdevice_notifier(&priv->netdev_notifier); err_deregister_fixed_link: @@ -2774,7 +2784,12 @@ static int __maybe_unused bcm_sysport_resume(struct device *d) if (!netif_running(dev)) return 0; - clk_prepare_enable(priv->clk); + ret = clk_prepare_enable(priv->clk); + if (ret) { + netdev_err(dev, "could not enable priv clock\n"); + return ret; + } + if (priv->wolopts) clk_disable_unprepare(priv->wol_clk); From fb3a9a1165cea104b5ab3753e88218e4497b01c1 Mon Sep 17 00:00:00 2001 From: Joshua Washington Date: Fri, 20 Dec 2024 19:28:06 -0800 Subject: [PATCH 112/196] gve: trigger RX NAPI instead of TX NAPI in gve_xsk_wakeup Commit ba0925c34e0f ("gve: process XSK TX descriptors as part of RX NAPI") moved XSK TX processing to be part of the RX NAPI. However, that commit did not include triggering the RX NAPI in gve_xsk_wakeup. This is necessary because the TX NAPI only processes TX completions, meaning that a TX wakeup would not actually trigger XSK descriptor processing. Also, the branch on XDP_WAKEUP_TX was supposed to have been removed, as the NAPI should be scheduled whether the wakeup is for RX or TX. Fixes: ba0925c34e0f ("gve: process XSK TX descriptors as part of RX NAPI") Cc: stable@vger.kernel.org Signed-off-by: Joshua Washington Signed-off-by: Praveen Kaligineedi Link: https://patch.msgid.link/20241221032807.302244-1-pkaligineedi@google.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/google/gve/gve_main.c | 21 +++++++-------------- 1 file changed, 7 insertions(+), 14 deletions(-) diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index 09fb7f16f73e..8a8f6ab12a98 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -1714,7 +1714,7 @@ static int gve_xsk_pool_disable(struct net_device *dev, static int gve_xsk_wakeup(struct net_device *dev, u32 queue_id, u32 flags) { struct gve_priv *priv = netdev_priv(dev); - int tx_queue_id = gve_xdp_tx_queue_id(priv, queue_id); + struct napi_struct *napi; if (!gve_get_napi_enabled(priv)) return -ENETDOWN; @@ -1722,19 +1722,12 @@ static int gve_xsk_wakeup(struct net_device *dev, u32 queue_id, u32 flags) if (queue_id >= priv->rx_cfg.num_queues || !priv->xdp_prog) return -EINVAL; - if (flags & XDP_WAKEUP_TX) { - struct gve_tx_ring *tx = &priv->tx[tx_queue_id]; - struct napi_struct *napi = - &priv->ntfy_blocks[tx->ntfy_id].napi; - - if (!napi_if_scheduled_mark_missed(napi)) { - /* Call local_bh_enable to trigger SoftIRQ processing */ - local_bh_disable(); - napi_schedule(napi); - local_bh_enable(); - } - - tx->xdp_xsk_wakeup++; + napi = &priv->ntfy_blocks[gve_rx_idx_to_ntfy(priv, queue_id)].napi; + if (!napi_if_scheduled_mark_missed(napi)) { + /* Call local_bh_enable to trigger SoftIRQ processing */ + local_bh_disable(); + napi_schedule(napi); + local_bh_enable(); } return 0; From ad5c318086e2e23b577eca33559c5ebf89bc7eb9 Mon Sep 17 00:00:00 2001 From: Joe Hattori Date: Sat, 21 Dec 2024 17:14:48 +0900 Subject: [PATCH 113/196] net: mv643xx_eth: fix an OF node reference leak Current implementation of mv643xx_eth_shared_of_add_port() calls of_parse_phandle(), but does not release the refcount on error. Call of_node_put() in the error path and in mv643xx_eth_shared_of_remove(). This bug was found by an experimental verification tool that I am developing. Fixes: 76723bca2802 ("net: mv643xx_eth: add DT parsing support") Signed-off-by: Joe Hattori Link: https://patch.msgid.link/20241221081448.3313163-1-joe@pf.is.s.u-tokyo.ac.jp Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/marvell/mv643xx_eth.c | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/drivers/net/ethernet/marvell/mv643xx_eth.c b/drivers/net/ethernet/marvell/mv643xx_eth.c index a06048719e84..67a6ff07c83d 100644 --- a/drivers/net/ethernet/marvell/mv643xx_eth.c +++ b/drivers/net/ethernet/marvell/mv643xx_eth.c @@ -2704,9 +2704,15 @@ static struct platform_device *port_platdev[3]; static void mv643xx_eth_shared_of_remove(void) { + struct mv643xx_eth_platform_data *pd; int n; for (n = 0; n < 3; n++) { + if (!port_platdev[n]) + continue; + pd = dev_get_platdata(&port_platdev[n]->dev); + if (pd) + of_node_put(pd->phy_node); platform_device_del(port_platdev[n]); port_platdev[n] = NULL; } @@ -2769,8 +2775,10 @@ static int mv643xx_eth_shared_of_add_port(struct platform_device *pdev, } ppdev = platform_device_alloc(MV643XX_ETH_NAME, dev_num); - if (!ppdev) - return -ENOMEM; + if (!ppdev) { + ret = -ENOMEM; + goto put_err; + } ppdev->dev.coherent_dma_mask = DMA_BIT_MASK(32); ppdev->dev.of_node = pnp; @@ -2792,6 +2800,8 @@ static int mv643xx_eth_shared_of_add_port(struct platform_device *pdev, port_err: platform_device_put(ppdev); +put_err: + of_node_put(ppd.phy_node); return ret; } From cbb26f7d8451fe56ccac802c6db48d16240feebd Mon Sep 17 00:00:00 2001 From: Paolo Abeni Date: Sat, 21 Dec 2024 09:51:46 +0100 Subject: [PATCH 114/196] mptcp: fix TCP options overflow. Syzbot reported the following splat: Oops: general protection fault, probably for non-canonical address 0xdffffc0000000001: 0000 [#1] PREEMPT SMP KASAN PTI KASAN: null-ptr-deref in range [0x0000000000000008-0x000000000000000f] CPU: 1 UID: 0 PID: 5836 Comm: sshd Not tainted 6.13.0-rc3-syzkaller #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 11/25/2024 RIP: 0010:_compound_head include/linux/page-flags.h:242 [inline] RIP: 0010:put_page+0x23/0x260 include/linux/mm.h:1552 Code: 90 90 90 90 90 90 90 55 41 57 41 56 53 49 89 fe 48 bd 00 00 00 00 00 fc ff df e8 f8 5e 12 f8 49 8d 5e 08 48 89 d8 48 c1 e8 03 <80> 3c 28 00 74 08 48 89 df e8 8f c7 78 f8 48 8b 1b 48 89 de 48 83 RSP: 0000:ffffc90003916c90 EFLAGS: 00010202 RAX: 0000000000000001 RBX: 0000000000000008 RCX: ffff888030458000 RDX: 0000000000000100 RSI: 0000000000000000 RDI: 0000000000000000 RBP: dffffc0000000000 R08: ffffffff898ca81d R09: 1ffff110054414ac R10: dffffc0000000000 R11: ffffed10054414ad R12: 0000000000000007 R13: ffff88802a20a542 R14: 0000000000000000 R15: 0000000000000000 FS: 00007f34f496e800(0000) GS:ffff8880b8700000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f9d6ec9ec28 CR3: 000000004d260000 CR4: 00000000003526f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: skb_page_unref include/linux/skbuff_ref.h:43 [inline] __skb_frag_unref include/linux/skbuff_ref.h:56 [inline] skb_release_data+0x483/0x8a0 net/core/skbuff.c:1119 skb_release_all net/core/skbuff.c:1190 [inline] __kfree_skb+0x55/0x70 net/core/skbuff.c:1204 tcp_clean_rtx_queue net/ipv4/tcp_input.c:3436 [inline] tcp_ack+0x2442/0x6bc0 net/ipv4/tcp_input.c:4032 tcp_rcv_state_process+0x8eb/0x44e0 net/ipv4/tcp_input.c:6805 tcp_v4_do_rcv+0x77d/0xc70 net/ipv4/tcp_ipv4.c:1939 tcp_v4_rcv+0x2dc0/0x37f0 net/ipv4/tcp_ipv4.c:2351 ip_protocol_deliver_rcu+0x22e/0x440 net/ipv4/ip_input.c:205 ip_local_deliver_finish+0x341/0x5f0 net/ipv4/ip_input.c:233 NF_HOOK+0x3a4/0x450 include/linux/netfilter.h:314 NF_HOOK+0x3a4/0x450 include/linux/netfilter.h:314 __netif_receive_skb_one_core net/core/dev.c:5672 [inline] __netif_receive_skb+0x2bf/0x650 net/core/dev.c:5785 process_backlog+0x662/0x15b0 net/core/dev.c:6117 __napi_poll+0xcb/0x490 net/core/dev.c:6883 napi_poll net/core/dev.c:6952 [inline] net_rx_action+0x89b/0x1240 net/core/dev.c:7074 handle_softirqs+0x2d4/0x9b0 kernel/softirq.c:561 __do_softirq kernel/softirq.c:595 [inline] invoke_softirq kernel/softirq.c:435 [inline] __irq_exit_rcu+0xf7/0x220 kernel/softirq.c:662 irq_exit_rcu+0x9/0x30 kernel/softirq.c:678 instr_sysvec_apic_timer_interrupt arch/x86/kernel/apic/apic.c:1049 [inline] sysvec_apic_timer_interrupt+0x57/0xc0 arch/x86/kernel/apic/apic.c:1049 asm_sysvec_apic_timer_interrupt+0x1a/0x20 arch/x86/include/asm/idtentry.h:702 RIP: 0033:0x7f34f4519ad5 Code: 85 d2 74 0d 0f 10 02 48 8d 54 24 20 0f 11 44 24 20 64 8b 04 25 18 00 00 00 85 c0 75 27 41 b8 08 00 00 00 b8 0f 01 00 00 0f 05 <48> 3d 00 f0 ff ff 76 75 48 8b 15 24 73 0d 00 f7 d8 64 89 02 48 83 RSP: 002b:00007ffec5b32ce0 EFLAGS: 00000246 RAX: 0000000000000001 RBX: 00000000000668a0 RCX: 00007f34f4519ad5 RDX: 00007ffec5b32d00 RSI: 0000000000000004 RDI: 0000564f4bc6cae0 RBP: 0000564f4bc6b5a0 R08: 0000000000000008 R09: 0000000000000000 R10: 00007ffec5b32de8 R11: 0000000000000246 R12: 0000564f48ea8aa4 R13: 0000000000000001 R14: 0000564f48ea93e8 R15: 00007ffec5b32d68 Eric noted a probable shinfo->nr_frags corruption, which indeed occurs. The root cause is a buggy MPTCP option len computation in some circumstances: the ADD_ADDR option should be mutually exclusive with DSS since the blamed commit. Still, mptcp_established_options_add_addr() tries to set the relevant info in mptcp_out_options, if the remaining space is large enough even when DSS is present. Since the ADD_ADDR infos and the DSS share the same union fields, adding first corrupts the latter. In the worst-case scenario, such corruption increases the DSS binary layout, exceeding the computed length and possibly overwriting the skb shared info. Address the issue by enforcing mutual exclusion in mptcp_established_options_add_addr(), too. Cc: stable@vger.kernel.org Reported-by: syzbot+38a095a81f30d82884c1@syzkaller.appspotmail.com Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/538 Fixes: 1bff1e43a30e ("mptcp: optimize out option generation") Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts (NGI0) Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/025d9df8cde3c9a557befc47e9bc08fbbe3476e5.1734771049.git.pabeni@redhat.com Signed-off-by: Jakub Kicinski --- net/mptcp/options.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/net/mptcp/options.c b/net/mptcp/options.c index 1603b3702e22..a62bc874bf1e 100644 --- a/net/mptcp/options.c +++ b/net/mptcp/options.c @@ -667,8 +667,15 @@ static bool mptcp_established_options_add_addr(struct sock *sk, struct sk_buff * &echo, &drop_other_suboptions)) return false; + /* + * Later on, mptcp_write_options() will enforce mutually exclusion with + * DSS, bail out if such option is set and we can't drop it. + */ if (drop_other_suboptions) remaining += opt_size; + else if (opts->suboptions & OPTION_MPTCP_DSS) + return false; + len = mptcp_add_addr_len(opts->addr.family, echo, !!opts->addr.port); if (remaining < len) return false; From 03c8d0af2e409e15c16130b185e12b5efba0a6b9 Mon Sep 17 00:00:00 2001 From: Pascal Hambourg Date: Mon, 23 Dec 2024 17:44:01 +0100 Subject: [PATCH 115/196] sky2: Add device ID 11ab:4373 for Marvell 88E8075 A Marvell 88E8075 ethernet controller has this device ID instead of 11ab:4370 and works fine with the sky2 driver. Signed-off-by: Pascal Hambourg Cc: stable@vger.kernel.org Link: https://patch.msgid.link/10165a62-99fb-4be6-8c64-84afd6234085@plouf.fr.eu.org Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/marvell/sky2.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/net/ethernet/marvell/sky2.c b/drivers/net/ethernet/marvell/sky2.c index 3914cd9210d4..988fa28cfb5f 100644 --- a/drivers/net/ethernet/marvell/sky2.c +++ b/drivers/net/ethernet/marvell/sky2.c @@ -130,6 +130,7 @@ static const struct pci_device_id sky2_id_table[] = { { PCI_DEVICE(PCI_VENDOR_ID_MARVELL, 0x436C) }, /* 88E8072 */ { PCI_DEVICE(PCI_VENDOR_ID_MARVELL, 0x436D) }, /* 88E8055 */ { PCI_DEVICE(PCI_VENDOR_ID_MARVELL, 0x4370) }, /* 88E8075 */ + { PCI_DEVICE(PCI_VENDOR_ID_MARVELL, 0x4373) }, /* 88E8075 */ { PCI_DEVICE(PCI_VENDOR_ID_MARVELL, 0x4380) }, /* 88E8057 */ { PCI_DEVICE(PCI_VENDOR_ID_MARVELL, 0x4381) }, /* 88E8059 */ { PCI_DEVICE(PCI_VENDOR_ID_MARVELL, 0x4382) }, /* 88E8079 */ From 8ec396d05d1b737c87311fb7311f753b02c2a6b1 Mon Sep 17 00:00:00 2001 From: Lorenzo Stoakes Date: Thu, 28 Nov 2024 15:06:17 +0000 Subject: [PATCH 116/196] mm: reinstate ability to map write-sealed memfd mappings read-only Patch series "mm: reinstate ability to map write-sealed memfd mappings read-only". In commit 158978945f31 ("mm: perform the mapping_map_writable() check after call_mmap()") (and preceding changes in the same series) it became possible to mmap() F_SEAL_WRITE sealed memfd mappings read-only. Commit 5de195060b2e ("mm: resolve faulty mmap_region() error path behaviour") unintentionally undid this logic by moving the mapping_map_writable() check before the shmem_mmap() hook is invoked, thereby regressing this change. This series reworks how we both permit write-sealed mappings being mapped read-only and disallow mprotect() from undoing the write-seal, fixing this regression. We also add a regression test to ensure that we do not accidentally regress this in future. Thanks to Julian Orth for reporting this regression. This patch (of 2): In commit 158978945f31 ("mm: perform the mapping_map_writable() check after call_mmap()") (and preceding changes in the same series) it became possible to mmap() F_SEAL_WRITE sealed memfd mappings read-only. This was previously unnecessarily disallowed, despite the man page documentation indicating that it would be, thereby limiting the usefulness of F_SEAL_WRITE logic. We fixed this by adapting logic that existed for the F_SEAL_FUTURE_WRITE seal (one which disallows future writes to the memfd) to also be used for F_SEAL_WRITE. For background - the F_SEAL_FUTURE_WRITE seal clears VM_MAYWRITE for a read-only mapping to disallow mprotect() from overriding the seal - an operation performed by seal_check_write(), invoked from shmem_mmap(), the f_op->mmap() hook used by shmem mappings. By extending this to F_SEAL_WRITE and critically - checking mapping_map_writable() to determine if we may map the memfd AFTER we invoke shmem_mmap() - the desired logic becomes possible. This is because mapping_map_writable() explicitly checks for VM_MAYWRITE, which we will have cleared. Commit 5de195060b2e ("mm: resolve faulty mmap_region() error path behaviour") unintentionally undid this logic by moving the mapping_map_writable() check before the shmem_mmap() hook is invoked, thereby regressing this change. We reinstate this functionality by moving the check out of shmem_mmap() and instead performing it in do_mmap() at the point at which VMA flags are being determined, which seems in any case to be a more appropriate place in which to make this determination. In order to achieve this we rework memfd seal logic to allow us access to this information using existing logic and eliminate the clearing of VM_MAYWRITE from seal_check_write() which we are performing in do_mmap() instead. Link: https://lkml.kernel.org/r/99fc35d2c62bd2e05571cf60d9f8b843c56069e0.1732804776.git.lorenzo.stoakes@oracle.com Fixes: 5de195060b2e ("mm: resolve faulty mmap_region() error path behaviour") Signed-off-by: Lorenzo Stoakes Reported-by: Julian Orth Closes: https://lore.kernel.org/all/CAHijbEUMhvJTN9Xw1GmbM266FXXv=U7s4L_Jem5x3AaPZxrYpQ@mail.gmail.com/ Cc: Jann Horn Cc: Liam R. Howlett Cc: Linus Torvalds Cc: Shuah Khan Cc: Vlastimil Babka Cc: Signed-off-by: Andrew Morton --- include/linux/memfd.h | 14 +++++++++++ include/linux/mm.h | 56 ++++++++++++++++++++++++++++++------------- mm/memfd.c | 2 +- mm/mmap.c | 4 ++++ 4 files changed, 58 insertions(+), 18 deletions(-) diff --git a/include/linux/memfd.h b/include/linux/memfd.h index 3f2cf339ceaf..d437e3070850 100644 --- a/include/linux/memfd.h +++ b/include/linux/memfd.h @@ -7,6 +7,7 @@ #ifdef CONFIG_MEMFD_CREATE extern long memfd_fcntl(struct file *file, unsigned int cmd, unsigned int arg); struct folio *memfd_alloc_folio(struct file *memfd, pgoff_t idx); +unsigned int *memfd_file_seals_ptr(struct file *file); #else static inline long memfd_fcntl(struct file *f, unsigned int c, unsigned int a) { @@ -16,6 +17,19 @@ static inline struct folio *memfd_alloc_folio(struct file *memfd, pgoff_t idx) { return ERR_PTR(-EINVAL); } + +static inline unsigned int *memfd_file_seals_ptr(struct file *file) +{ + return NULL; +} #endif +/* Retrieve memfd seals associated with the file, if any. */ +static inline unsigned int memfd_file_seals(struct file *file) +{ + unsigned int *sealsp = memfd_file_seals_ptr(file); + + return sealsp ? *sealsp : 0; +} + #endif /* __LINUX_MEMFD_H */ diff --git a/include/linux/mm.h b/include/linux/mm.h index 338a76ce9083..fb397918c43d 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -4101,6 +4101,37 @@ void mem_dump_obj(void *object); static inline void mem_dump_obj(void *object) {} #endif +static inline bool is_write_sealed(int seals) +{ + return seals & (F_SEAL_WRITE | F_SEAL_FUTURE_WRITE); +} + +/** + * is_readonly_sealed - Checks whether write-sealed but mapped read-only, + * in which case writes should be disallowing moving + * forwards. + * @seals: the seals to check + * @vm_flags: the VMA flags to check + * + * Returns whether readonly sealed, in which case writess should be disallowed + * going forward. + */ +static inline bool is_readonly_sealed(int seals, vm_flags_t vm_flags) +{ + /* + * Since an F_SEAL_[FUTURE_]WRITE sealed memfd can be mapped as + * MAP_SHARED and read-only, take care to not allow mprotect to + * revert protections on such mappings. Do this only for shared + * mappings. For private mappings, don't need to mask + * VM_MAYWRITE as we still want them to be COW-writable. + */ + if (is_write_sealed(seals) && + ((vm_flags & (VM_SHARED | VM_WRITE)) == VM_SHARED)) + return true; + + return false; +} + /** * seal_check_write - Check for F_SEAL_WRITE or F_SEAL_FUTURE_WRITE flags and * handle them. @@ -4112,24 +4143,15 @@ static inline void mem_dump_obj(void *object) {} */ static inline int seal_check_write(int seals, struct vm_area_struct *vma) { - if (seals & (F_SEAL_WRITE | F_SEAL_FUTURE_WRITE)) { - /* - * New PROT_WRITE and MAP_SHARED mmaps are not allowed when - * write seals are active. - */ - if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_WRITE)) - return -EPERM; + if (!is_write_sealed(seals)) + return 0; - /* - * Since an F_SEAL_[FUTURE_]WRITE sealed memfd can be mapped as - * MAP_SHARED and read-only, take care to not allow mprotect to - * revert protections on such mappings. Do this only for shared - * mappings. For private mappings, don't need to mask - * VM_MAYWRITE as we still want them to be COW-writable. - */ - if (vma->vm_flags & VM_SHARED) - vm_flags_clear(vma, VM_MAYWRITE); - } + /* + * New PROT_WRITE and MAP_SHARED mmaps are not allowed when + * write seals are active. + */ + if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_WRITE)) + return -EPERM; return 0; } diff --git a/mm/memfd.c b/mm/memfd.c index c17c3ea701a1..35a370d75c9a 100644 --- a/mm/memfd.c +++ b/mm/memfd.c @@ -170,7 +170,7 @@ static int memfd_wait_for_pins(struct address_space *mapping) return error; } -static unsigned int *memfd_file_seals_ptr(struct file *file) +unsigned int *memfd_file_seals_ptr(struct file *file) { if (shmem_file(file)) return &SHMEM_I(file_inode(file))->seals; diff --git a/mm/mmap.c b/mm/mmap.c index d32b7e701058..16f8e8be01f8 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -47,6 +47,7 @@ #include #include #include +#include #include #include @@ -368,6 +369,7 @@ unsigned long do_mmap(struct file *file, unsigned long addr, if (file) { struct inode *inode = file_inode(file); + unsigned int seals = memfd_file_seals(file); unsigned long flags_mask; if (!file_mmap_ok(file, inode, pgoff, len)) @@ -408,6 +410,8 @@ unsigned long do_mmap(struct file *file, unsigned long addr, vm_flags |= VM_SHARED | VM_MAYSHARE; if (!(file->f_mode & FMODE_WRITE)) vm_flags &= ~(VM_MAYWRITE | VM_SHARED); + else if (is_readonly_sealed(seals, vm_flags)) + vm_flags &= ~VM_MAYWRITE; fallthrough; case MAP_PRIVATE: if (!(file->f_mode & FMODE_READ)) From ea0916e01d0b0f2cce1369ac1494239a79827270 Mon Sep 17 00:00:00 2001 From: Lorenzo Stoakes Date: Thu, 28 Nov 2024 15:06:18 +0000 Subject: [PATCH 117/196] selftests/memfd: add test for mapping write-sealed memfd read-only Now we have reinstated the ability to map F_SEAL_WRITE mappings read-only, assert that we are able to do this in a test to ensure that we do not regress this again. Link: https://lkml.kernel.org/r/a6377ec470b14c0539b4600cf8fa24bf2e4858ae.1732804776.git.lorenzo.stoakes@oracle.com Signed-off-by: Lorenzo Stoakes Cc: Jann Horn Cc: Julian Orth Cc: Liam R. Howlett Cc: Linus Torvalds Cc: Shuah Khan Cc: Vlastimil Babka Signed-off-by: Andrew Morton --- tools/testing/selftests/memfd/memfd_test.c | 43 ++++++++++++++++++++++ 1 file changed, 43 insertions(+) diff --git a/tools/testing/selftests/memfd/memfd_test.c b/tools/testing/selftests/memfd/memfd_test.c index 0a0b55516028..c0c53451a16d 100644 --- a/tools/testing/selftests/memfd/memfd_test.c +++ b/tools/testing/selftests/memfd/memfd_test.c @@ -282,6 +282,24 @@ static void *mfd_assert_mmap_shared(int fd) return p; } +static void *mfd_assert_mmap_read_shared(int fd) +{ + void *p; + + p = mmap(NULL, + mfd_def_size, + PROT_READ, + MAP_SHARED, + fd, + 0); + if (p == MAP_FAILED) { + printf("mmap() failed: %m\n"); + abort(); + } + + return p; +} + static void *mfd_assert_mmap_private(int fd) { void *p; @@ -980,6 +998,30 @@ static void test_seal_future_write(void) close(fd); } +static void test_seal_write_map_read_shared(void) +{ + int fd; + void *p; + + printf("%s SEAL-WRITE-MAP-READ\n", memfd_str); + + fd = mfd_assert_new("kern_memfd_seal_write_map_read", + mfd_def_size, + MFD_CLOEXEC | MFD_ALLOW_SEALING); + + mfd_assert_add_seals(fd, F_SEAL_WRITE); + mfd_assert_has_seals(fd, F_SEAL_WRITE); + + p = mfd_assert_mmap_read_shared(fd); + + mfd_assert_read(fd); + mfd_assert_read_shared(fd); + mfd_fail_write(fd); + + munmap(p, mfd_def_size); + close(fd); +} + /* * Test SEAL_SHRINK * Test whether SEAL_SHRINK actually prevents shrinking @@ -1593,6 +1635,7 @@ int main(int argc, char **argv) test_seal_write(); test_seal_future_write(); + test_seal_write_map_read_shared(); test_seal_shrink(); test_seal_grow(); test_seal_resize(); From 6aaced5abd32e2a57cd94fd64f824514d0361da8 Mon Sep 17 00:00:00 2001 From: Seiji Nishikawa Date: Sun, 1 Dec 2024 01:12:34 +0900 Subject: [PATCH 118/196] mm: vmscan: account for free pages to prevent infinite Loop in throttle_direct_reclaim() The task sometimes continues looping in throttle_direct_reclaim() because allow_direct_reclaim(pgdat) keeps returning false. #0 [ffff80002cb6f8d0] __switch_to at ffff8000080095ac #1 [ffff80002cb6f900] __schedule at ffff800008abbd1c #2 [ffff80002cb6f990] schedule at ffff800008abc50c #3 [ffff80002cb6f9b0] throttle_direct_reclaim at ffff800008273550 #4 [ffff80002cb6fa20] try_to_free_pages at ffff800008277b68 #5 [ffff80002cb6fae0] __alloc_pages_nodemask at ffff8000082c4660 #6 [ffff80002cb6fc50] alloc_pages_vma at ffff8000082e4a98 #7 [ffff80002cb6fca0] do_anonymous_page at ffff80000829f5a8 #8 [ffff80002cb6fce0] __handle_mm_fault at ffff8000082a5974 #9 [ffff80002cb6fd90] handle_mm_fault at ffff8000082a5bd4 At this point, the pgdat contains the following two zones: NODE: 4 ZONE: 0 ADDR: ffff00817fffe540 NAME: "DMA32" SIZE: 20480 MIN/LOW/HIGH: 11/28/45 VM_STAT: NR_FREE_PAGES: 359 NR_ZONE_INACTIVE_ANON: 18813 NR_ZONE_ACTIVE_ANON: 0 NR_ZONE_INACTIVE_FILE: 50 NR_ZONE_ACTIVE_FILE: 0 NR_ZONE_UNEVICTABLE: 0 NR_ZONE_WRITE_PENDING: 0 NR_MLOCK: 0 NR_BOUNCE: 0 NR_ZSPAGES: 0 NR_FREE_CMA_PAGES: 0 NODE: 4 ZONE: 1 ADDR: ffff00817fffec00 NAME: "Normal" SIZE: 8454144 PRESENT: 98304 MIN/LOW/HIGH: 68/166/264 VM_STAT: NR_FREE_PAGES: 146 NR_ZONE_INACTIVE_ANON: 94668 NR_ZONE_ACTIVE_ANON: 3 NR_ZONE_INACTIVE_FILE: 735 NR_ZONE_ACTIVE_FILE: 78 NR_ZONE_UNEVICTABLE: 0 NR_ZONE_WRITE_PENDING: 0 NR_MLOCK: 0 NR_BOUNCE: 0 NR_ZSPAGES: 0 NR_FREE_CMA_PAGES: 0 In allow_direct_reclaim(), while processing ZONE_DMA32, the sum of inactive/active file-backed pages calculated in zone_reclaimable_pages() based on the result of zone_page_state_snapshot() is zero. Additionally, since this system lacks swap, the calculation of inactive/ active anonymous pages is skipped. crash> p nr_swap_pages nr_swap_pages = $1937 = { counter = 0 } As a result, ZONE_DMA32 is deemed unreclaimable and skipped, moving on to the processing of the next zone, ZONE_NORMAL, despite ZONE_DMA32 having free pages significantly exceeding the high watermark. The problem is that the pgdat->kswapd_failures hasn't been incremented. crash> px ((struct pglist_data *) 0xffff00817fffe540)->kswapd_failures $1935 = 0x0 This is because the node deemed balanced. The node balancing logic in balance_pgdat() evaluates all zones collectively. If one or more zones (e.g., ZONE_DMA32) have enough free pages to meet their watermarks, the entire node is deemed balanced. This causes balance_pgdat() to exit early before incrementing the kswapd_failures, as it considers the overall memory state acceptable, even though some zones (like ZONE_NORMAL) remain under significant pressure. The patch ensures that zone_reclaimable_pages() includes free pages (NR_FREE_PAGES) in its calculation when no other reclaimable pages are available (e.g., file-backed or anonymous pages). This change prevents zones like ZONE_DMA32, which have sufficient free pages, from being mistakenly deemed unreclaimable. By doing so, the patch ensures proper node balancing, avoids masking pressure on other zones like ZONE_NORMAL, and prevents infinite loops in throttle_direct_reclaim() caused by allow_direct_reclaim(pgdat) repeatedly returning false. The kernel hangs due to a task stuck in throttle_direct_reclaim(), caused by a node being incorrectly deemed balanced despite pressure in certain zones, such as ZONE_NORMAL. This issue arises from zone_reclaimable_pages() returning 0 for zones without reclaimable file- backed or anonymous pages, causing zones like ZONE_DMA32 with sufficient free pages to be skipped. The lack of swap or reclaimable pages results in ZONE_DMA32 being ignored during reclaim, masking pressure in other zones. Consequently, pgdat->kswapd_failures remains 0 in balance_pgdat(), preventing fallback mechanisms in allow_direct_reclaim() from being triggered, leading to an infinite loop in throttle_direct_reclaim(). This patch modifies zone_reclaimable_pages() to account for free pages (NR_FREE_PAGES) when no other reclaimable pages exist. This ensures zones with sufficient free pages are not skipped, enabling proper balancing and reclaim behavior. [akpm@linux-foundation.org: coding-style cleanups] Link: https://lkml.kernel.org/r/20241130164346.436469-1-snishika@redhat.com Link: https://lkml.kernel.org/r/20241130161236.433747-2-snishika@redhat.com Fixes: 5a1c84b404a7 ("mm: remove reclaim and compaction retry approximations") Signed-off-by: Seiji Nishikawa Cc: Mel Gorman Cc: Signed-off-by: Andrew Morton --- mm/vmscan.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) diff --git a/mm/vmscan.c b/mm/vmscan.c index 76378bc257e3..9a859b7d18d7 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -374,7 +374,14 @@ unsigned long zone_reclaimable_pages(struct zone *zone) if (can_reclaim_anon_pages(NULL, zone_to_nid(zone), NULL)) nr += zone_page_state_snapshot(zone, NR_ZONE_INACTIVE_ANON) + zone_page_state_snapshot(zone, NR_ZONE_ACTIVE_ANON); - + /* + * If there are no reclaimable file-backed or anonymous pages, + * ensure zones with sufficient free pages are not skipped. + * This prevents zones like DMA32 from being ignored in reclaim + * scenarios where they can still help alleviate memory pressure. + */ + if (nr == 0) + nr = zone_page_state_snapshot(zone, NR_FREE_PAGES); return nr; } From 34d7cf637c437d5c2a8a6ef23ea45193bad8a91c Mon Sep 17 00:00:00 2001 From: Kefeng Wang Date: Fri, 6 Dec 2024 15:03:45 +0800 Subject: [PATCH 119/196] mm: don't try THP alignment for FS without get_unmapped_area Commit ed48e87c7df3 ("thp: add thp_get_unmapped_area_vmflags()") changes thp_get_unmapped_area() to thp_get_unmapped_area_vmflags() in __get_unmapped_area(), which doesn't initialize local get_area for anonymous mappings. This leads to us always trying THP alignment even for file_operations which have a NULL ->get_unmapped_area() callback. Since commit efa7df3e3bb5 ("mm: align larger anonymous mappings on THP boundaries") we only want to enable THP alignment for anonymous mappings, so add a !file check to avoid attempting THP alignment for file mappings. Found issue by code inspection. THP alignment is used for easy or more pmd mappings, from vma side. This may cause unnecessary VMA fragmentation and potentially worse performance on filesystems that do not actually support THPs and thus cannot benefit from the alignment. Link: https://lkml.kernel.org/r/20241206070345.2526501-1-wangkefeng.wang@huawei.com Fixes: ed48e87c7df3 ("thp: add thp_get_unmapped_area_vmflags()") Signed-off-by: Kefeng Wang Reviewed-by: Vlastimil Babka Reviewed-by: Yang Shi Cc: Christophe Leroy Cc: David Hildenbrand Cc: Jann Horn Cc: Liam R. Howlett Cc: Lorenzo Stoakes Cc: Rick Edgecombe Cc: Ryan Roberts Signed-off-by: Andrew Morton --- mm/mmap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/mmap.c b/mm/mmap.c index 16f8e8be01f8..aec208f90337 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -892,7 +892,7 @@ __get_unmapped_area(struct file *file, unsigned long addr, unsigned long len, if (get_area) { addr = get_area(file, addr, len, pgoff, flags); - } else if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) + } else if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) && !file && !addr /* no hint */ && IS_ALIGNED(len, PMD_SIZE)) { /* Ensures that larger anonymous mappings are THP aligned. */ From 158cdce87c8c172787063998ad5dd3e2f658b963 Mon Sep 17 00:00:00 2001 From: Yafang Shao Date: Fri, 6 Dec 2024 16:30:25 +0800 Subject: [PATCH 120/196] mm/readahead: fix large folio support in async readahead When testing large folio support with XFS on our servers, we observed that only a few large folios are mapped when reading large files via mmap. After a thorough analysis, I identified it was caused by the `/sys/block/*/queue/read_ahead_kb` setting. On our test servers, this parameter is set to 128KB. After I tune it to 2MB, the large folio can work as expected. However, I believe the large folio behavior should not be dependent on the value of read_ahead_kb. It would be more robust if the kernel can automatically adopt to it. With /sys/block/*/queue/read_ahead_kb set to 128KB and performing a sequential read on a 1GB file using MADV_HUGEPAGE, the differences in /proc/meminfo are as follows: - before this patch FileHugePages: 18432 kB FilePmdMapped: 4096 kB - after this patch FileHugePages: 1067008 kB FilePmdMapped: 1048576 kB This shows that after applying the patch, the entire 1GB file is mapped to huge pages. The stable list is CCed, as without this patch, large folios don't function optimally in the readahead path. It's worth noting that if read_ahead_kb is set to a larger value that isn't aligned with huge page sizes (e.g., 4MB + 128KB), it may still fail to map to hugepages. Link: https://lkml.kernel.org/r/20241108141710.9721-1-laoar.shao@gmail.com Link: https://lkml.kernel.org/r/20241206083025.3478-1-laoar.shao@gmail.com Fixes: 4687fdbb805a ("mm/filemap: Support VM_HUGEPAGE for file mappings") Signed-off-by: Yafang Shao Tested-by: kernel test robot Cc: Matthew Wilcox Cc: David Hildenbrand Cc: Signed-off-by: Andrew Morton --- mm/readahead.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/mm/readahead.c b/mm/readahead.c index ea650b8b02fb..e151f4b13ca4 100644 --- a/mm/readahead.c +++ b/mm/readahead.c @@ -646,7 +646,11 @@ void page_cache_async_ra(struct readahead_control *ractl, 1UL << order); if (index == expected) { ra->start += ra->size; - ra->size = get_next_ra_size(ra, max_pages); + /* + * In the case of MADV_HUGEPAGE, the actual size might exceed + * the readahead window. + */ + ra->size = max(ra->size, get_next_ra_size(ra, max_pages)); ra->async_size = ra->size; goto readit; } From 1fd8bc7cd889bd73d07a83cb32d674ac68f99153 Mon Sep 17 00:00:00 2001 From: Yang Erkun Date: Sat, 14 Dec 2024 17:30:05 +0800 Subject: [PATCH 121/196] maple_tree: reload mas before the second call for mas_empty_area Change the LONG_MAX in simple_offset_add to 1024, and do latter: [root@fedora ~]# mkdir /tmp/dir [root@fedora ~]# for i in {1..1024}; do touch /tmp/dir/$i; done touch: cannot touch '/tmp/dir/1024': Device or resource busy [root@fedora ~]# rm /tmp/dir/123 [root@fedora ~]# touch /tmp/dir/1024 [root@fedora ~]# rm /tmp/dir/100 [root@fedora ~]# touch /tmp/dir/1025 touch: cannot touch '/tmp/dir/1025': Device or resource busy After we delete file 100, actually this is a empty entry, but the latter create failed unexpected. mas_alloc_cyclic has two chance to find empty entry. First find the entry with range range_lo and range_hi, if no empty entry exist, and range_lo > min, retry find with range min and range_hi. However, the first call mas_empty_area may mark mas as EBUSY, and the second call for mas_empty_area will return false directly. Fix this by reload mas before second call for mas_empty_area. [Liam.Howlett@Oracle.com: fix mas_alloc_cyclic() second search] Link: https://lore.kernel.org/all/20241216060600.287B4C4CED0@smtp.kernel.org/ Link: https://lkml.kernel.org/r/20241216190113.1226145-2-Liam.Howlett@oracle.com Link: https://lkml.kernel.org/r/20241214093005.72284-1-yangerkun@huaweicloud.com Fixes: 9b6713cc7522 ("maple_tree: Add mtree_alloc_cyclic()") Signed-off-by: Yang Erkun Signed-off-by: Liam R. Howlett Cc: Christian Brauner Cc: Chuck Lever says: Cc: Liam R. Howlett Cc: Signed-off-by: Andrew Morton --- lib/maple_tree.c | 1 + 1 file changed, 1 insertion(+) diff --git a/lib/maple_tree.c b/lib/maple_tree.c index d0ae808f3a14..047397136f15 100644 --- a/lib/maple_tree.c +++ b/lib/maple_tree.c @@ -4354,6 +4354,7 @@ int mas_alloc_cyclic(struct ma_state *mas, unsigned long *startp, ret = 1; } if (ret < 0 && range_lo > min) { + mas_reset(mas); ret = mas_empty_area(mas, min, range_hi, 1); if (ret == 0) ret = 1; From 59d9094df3d79443937add8700b2ef1a866b1081 Mon Sep 17 00:00:00 2001 From: Liu Shixin Date: Mon, 16 Dec 2024 15:11:47 +0800 Subject: [PATCH 122/196] mm: hugetlb: independent PMD page table shared count The folio refcount may be increased unexpectly through try_get_folio() by caller such as split_huge_pages. In huge_pmd_unshare(), we use refcount to check whether a pmd page table is shared. The check is incorrect if the refcount is increased by the above caller, and this can cause the page table leaked: BUG: Bad page state in process sh pfn:109324 page: refcount:0 mapcount:0 mapping:0000000000000000 index:0x66 pfn:0x109324 flags: 0x17ffff800000000(node=0|zone=2|lastcpupid=0xfffff) page_type: f2(table) raw: 017ffff800000000 0000000000000000 0000000000000000 0000000000000000 raw: 0000000000000066 0000000000000000 00000000f2000000 0000000000000000 page dumped because: nonzero mapcount ... CPU: 31 UID: 0 PID: 7515 Comm: sh Kdump: loaded Tainted: G B 6.13.0-rc2master+ #7 Tainted: [B]=BAD_PAGE Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015 Call trace: show_stack+0x20/0x38 (C) dump_stack_lvl+0x80/0xf8 dump_stack+0x18/0x28 bad_page+0x8c/0x130 free_page_is_bad_report+0xa4/0xb0 free_unref_page+0x3cc/0x620 __folio_put+0xf4/0x158 split_huge_pages_all+0x1e0/0x3e8 split_huge_pages_write+0x25c/0x2d8 full_proxy_write+0x64/0xd8 vfs_write+0xcc/0x280 ksys_write+0x70/0x110 __arm64_sys_write+0x24/0x38 invoke_syscall+0x50/0x120 el0_svc_common.constprop.0+0xc8/0xf0 do_el0_svc+0x24/0x38 el0_svc+0x34/0x128 el0t_64_sync_handler+0xc8/0xd0 el0t_64_sync+0x190/0x198 The issue may be triggered by damon, offline_page, page_idle, etc, which will increase the refcount of page table. 1. The page table itself will be discarded after reporting the "nonzero mapcount". 2. The HugeTLB page mapped by the page table miss freeing since we treat the page table as shared and a shared page table will not be unmapped. Fix it by introducing independent PMD page table shared count. As described by comment, pt_index/pt_mm/pt_frag_refcount are used for s390 gmap, x86 pgds and powerpc, pt_share_count is used for x86/arm64/riscv pmds, so we can reuse the field as pt_share_count. Link: https://lkml.kernel.org/r/20241216071147.3984217-1-liushixin2@huawei.com Fixes: 39dde65c9940 ("[PATCH] shared page table for hugetlb page") Signed-off-by: Liu Shixin Cc: Kefeng Wang Cc: Ken Chen Cc: Muchun Song Cc: Nanyong Sun Cc: Jane Chu Cc: Signed-off-by: Andrew Morton --- include/linux/mm.h | 1 + include/linux/mm_types.h | 30 ++++++++++++++++++++++++++++++ mm/hugetlb.c | 16 +++++++--------- 3 files changed, 38 insertions(+), 9 deletions(-) diff --git a/include/linux/mm.h b/include/linux/mm.h index fb397918c43d..b1c3db9cf355 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -3125,6 +3125,7 @@ static inline bool pagetable_pmd_ctor(struct ptdesc *ptdesc) if (!pmd_ptlock_init(ptdesc)) return false; __folio_set_pgtable(folio); + ptdesc_pmd_pts_init(ptdesc); lruvec_stat_add_folio(folio, NR_PAGETABLE); return true; } diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 7361a8f3ab68..332cee285662 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -445,6 +445,7 @@ FOLIO_MATCH(compound_head, _head_2a); * @pt_index: Used for s390 gmap. * @pt_mm: Used for x86 pgds. * @pt_frag_refcount: For fragmented page table tracking. Powerpc only. + * @pt_share_count: Used for HugeTLB PMD page table share count. * @_pt_pad_2: Padding to ensure proper alignment. * @ptl: Lock for the page table. * @__page_type: Same as page->page_type. Unused for page tables. @@ -471,6 +472,9 @@ struct ptdesc { pgoff_t pt_index; struct mm_struct *pt_mm; atomic_t pt_frag_refcount; +#ifdef CONFIG_HUGETLB_PMD_PAGE_TABLE_SHARING + atomic_t pt_share_count; +#endif }; union { @@ -516,6 +520,32 @@ static_assert(sizeof(struct ptdesc) <= sizeof(struct page)); const struct page *: (const struct ptdesc *)(p), \ struct page *: (struct ptdesc *)(p))) +#ifdef CONFIG_HUGETLB_PMD_PAGE_TABLE_SHARING +static inline void ptdesc_pmd_pts_init(struct ptdesc *ptdesc) +{ + atomic_set(&ptdesc->pt_share_count, 0); +} + +static inline void ptdesc_pmd_pts_inc(struct ptdesc *ptdesc) +{ + atomic_inc(&ptdesc->pt_share_count); +} + +static inline void ptdesc_pmd_pts_dec(struct ptdesc *ptdesc) +{ + atomic_dec(&ptdesc->pt_share_count); +} + +static inline int ptdesc_pmd_pts_count(struct ptdesc *ptdesc) +{ + return atomic_read(&ptdesc->pt_share_count); +} +#else +static inline void ptdesc_pmd_pts_init(struct ptdesc *ptdesc) +{ +} +#endif + /* * Used for sizing the vmemmap region on some architectures */ diff --git a/mm/hugetlb.c b/mm/hugetlb.c index cec4b121193f..c498874a7170 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -7211,7 +7211,7 @@ pte_t *huge_pmd_share(struct mm_struct *mm, struct vm_area_struct *vma, spte = hugetlb_walk(svma, saddr, vma_mmu_pagesize(svma)); if (spte) { - get_page(virt_to_page(spte)); + ptdesc_pmd_pts_inc(virt_to_ptdesc(spte)); break; } } @@ -7226,7 +7226,7 @@ pte_t *huge_pmd_share(struct mm_struct *mm, struct vm_area_struct *vma, (pmd_t *)((unsigned long)spte & PAGE_MASK)); mm_inc_nr_pmds(mm); } else { - put_page(virt_to_page(spte)); + ptdesc_pmd_pts_dec(virt_to_ptdesc(spte)); } spin_unlock(&mm->page_table_lock); out: @@ -7238,10 +7238,6 @@ pte_t *huge_pmd_share(struct mm_struct *mm, struct vm_area_struct *vma, /* * unmap huge page backed by shared pte. * - * Hugetlb pte page is ref counted at the time of mapping. If pte is shared - * indicated by page_count > 1, unmap is achieved by clearing pud and - * decrementing the ref count. If count == 1, the pte page is not shared. - * * Called with page table lock held. * * returns: 1 successfully unmapped a shared pte page @@ -7250,18 +7246,20 @@ pte_t *huge_pmd_share(struct mm_struct *mm, struct vm_area_struct *vma, int huge_pmd_unshare(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, pte_t *ptep) { + unsigned long sz = huge_page_size(hstate_vma(vma)); pgd_t *pgd = pgd_offset(mm, addr); p4d_t *p4d = p4d_offset(pgd, addr); pud_t *pud = pud_offset(p4d, addr); i_mmap_assert_write_locked(vma->vm_file->f_mapping); hugetlb_vma_assert_locked(vma); - BUG_ON(page_count(virt_to_page(ptep)) == 0); - if (page_count(virt_to_page(ptep)) == 1) + if (sz != PMD_SIZE) + return 0; + if (!ptdesc_pmd_pts_count(virt_to_ptdesc(ptep))) return 0; pud_clear(pud); - put_page(virt_to_page(ptep)); + ptdesc_pmd_pts_dec(virt_to_ptdesc(ptep)); mm_dec_nr_pmds(mm); return 1; } From cddc76b165161a02ff14c4d84d0f5266d9d32b9e Mon Sep 17 00:00:00 2001 From: Alessandro Carminati Date: Tue, 17 Dec 2024 14:20:33 +0000 Subject: [PATCH 123/196] mm/kmemleak: fix sleeping function called from invalid context at print message MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Address a bug in the kernel that triggers a "sleeping function called from invalid context" warning when /sys/kernel/debug/kmemleak is printed under specific conditions: - CONFIG_PREEMPT_RT=y - Set SELinux as the LSM for the system - Set kptr_restrict to 1 - kmemleak buffer contains at least one item BUG: sleeping function called from invalid context at kernel/locking/spinlock_rt.c:48 in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 136, name: cat preempt_count: 1, expected: 0 RCU nest depth: 2, expected: 2 6 locks held by cat/136: #0: ffff32e64bcbf950 (&p->lock){+.+.}-{3:3}, at: seq_read_iter+0xb8/0xe30 #1: ffffafe6aaa9dea0 (scan_mutex){+.+.}-{3:3}, at: kmemleak_seq_start+0x34/0x128 #3: ffff32e6546b1cd0 (&object->lock){....}-{2:2}, at: kmemleak_seq_show+0x3c/0x1e0 #4: ffffafe6aa8d8560 (rcu_read_lock){....}-{1:2}, at: has_ns_capability_noaudit+0x8/0x1b0 #5: ffffafe6aabbc0f8 (notif_lock){+.+.}-{2:2}, at: avc_compute_av+0xc4/0x3d0 irq event stamp: 136660 hardirqs last enabled at (136659): [] _raw_spin_unlock_irqrestore+0xa8/0xd8 hardirqs last disabled at (136660): [] _raw_spin_lock_irqsave+0x8c/0xb0 softirqs last enabled at (0): [] copy_process+0x11d8/0x3df8 softirqs last disabled at (0): [<0000000000000000>] 0x0 Preemption disabled at: [] kmemleak_seq_show+0x3c/0x1e0 CPU: 1 UID: 0 PID: 136 Comm: cat Tainted: G E 6.11.0-rt7+ #34 Tainted: [E]=UNSIGNED_MODULE Hardware name: linux,dummy-virt (DT) Call trace: dump_backtrace+0xa0/0x128 show_stack+0x1c/0x30 dump_stack_lvl+0xe8/0x198 dump_stack+0x18/0x20 rt_spin_lock+0x8c/0x1a8 avc_perm_nonode+0xa0/0x150 cred_has_capability.isra.0+0x118/0x218 selinux_capable+0x50/0x80 security_capable+0x7c/0xd0 has_ns_capability_noaudit+0x94/0x1b0 has_capability_noaudit+0x20/0x30 restricted_pointer+0x21c/0x4b0 pointer+0x298/0x760 vsnprintf+0x330/0xf70 seq_printf+0x178/0x218 print_unreferenced+0x1a4/0x2d0 kmemleak_seq_show+0xd0/0x1e0 seq_read_iter+0x354/0xe30 seq_read+0x250/0x378 full_proxy_read+0xd8/0x148 vfs_read+0x190/0x918 ksys_read+0xf0/0x1e0 __arm64_sys_read+0x70/0xa8 invoke_syscall.constprop.0+0xd4/0x1d8 el0_svc+0x50/0x158 el0t_64_sync+0x17c/0x180 %pS and %pK, in the same back trace line, are redundant, and %pS can void %pK service in certain contexts. %pS alone already provides the necessary information, and if it cannot resolve the symbol, it falls back to printing the raw address voiding the original intent behind the %pK. Additionally, %pK requires a privilege check CAP_SYSLOG enforced through the LSM, which can trigger a "sleeping function called from invalid context" warning under RT_PREEMPT kernels when the check occurs in an atomic context. This issue may also affect other LSMs. This change avoids the unnecessary privilege check and resolves the sleeping function warning without any loss of information. Link: https://lkml.kernel.org/r/20241217142032.55793-1-acarmina@redhat.com Fixes: 3a6f33d86baa ("mm/kmemleak: use %pK to display kernel pointers in backtrace") Signed-off-by: Alessandro Carminati Acked-by: Sebastian Andrzej Siewior Acked-by: Catalin Marinas Cc: Clément Léger Cc: Alessandro Carminati Cc: Eric Chanudet Cc: Gabriele Paoloni Cc: Juri Lelli Cc: Steven Rostedt Cc: Thomas Weißschuh Cc: Signed-off-by: Andrew Morton --- mm/kmemleak.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/kmemleak.c b/mm/kmemleak.c index 2a945c07ae99..737af23f4f4e 100644 --- a/mm/kmemleak.c +++ b/mm/kmemleak.c @@ -373,7 +373,7 @@ static void print_unreferenced(struct seq_file *seq, for (i = 0; i < nr_entries; i++) { void *ptr = (void *)entries[i]; - warn_or_seq_printf(seq, " [<%pK>] %pS\n", ptr, ptr); + warn_or_seq_printf(seq, " %pS\n", ptr); } } From 4d9b90df2eb49ab9becdbfd1fd60071bb107406e Mon Sep 17 00:00:00 2001 From: Mathieu Othacehe Date: Tue, 17 Dec 2024 11:09:21 +0100 Subject: [PATCH 124/196] mailmap: modify the entry for Mathieu Othacehe Set my gnu address as the main one. Link: https://lkml.kernel.org/r/20241217100924.7821-1-othacehe@gnu.org Signed-off-by: Mathieu Othacehe Cc: Alex Elder Cc: David S. Miller Cc: Geliang Tang Cc: Kees Cook Cc: Matthieu Baerts (NGI0) Cc: Neeraj Upadhyay Cc: Quentin Monnet Signed-off-by: Andrew Morton --- .mailmap | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.mailmap b/.mailmap index 7efe43237ca8..f5f97f947020 100644 --- a/.mailmap +++ b/.mailmap @@ -435,7 +435,7 @@ Martin Kepplinger Martin Kepplinger Martin Kepplinger Martyna Szapar-Mudlaw -Mathieu Othacehe +Mathieu Othacehe Mat Martineau Mat Martineau Matthew Wilcox From 472098f23323c39cc6269d7b7bf76cba62830a4c Mon Sep 17 00:00:00 2001 From: Baolin Wang Date: Tue, 17 Dec 2024 16:55:39 +0800 Subject: [PATCH 125/196] docs: mm: fix the incorrect 'FileHugeMapped' field The '/proc/PID/smaps' does not have the 'FileHugeMapped' field to count the file transparent huge pages, instead, the 'FilePmdMapped' field should be used. Fix it. Link: https://lkml.kernel.org/r/d520ce3aba2b03b088be30bece732426a939049a.1734425264.git.baolin.wang@linux.alibaba.com Signed-off-by: Baolin Wang Acked-by: David Hildenbrand Signed-off-by: Andrew Morton --- Documentation/admin-guide/mm/transhuge.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/admin-guide/mm/transhuge.rst b/Documentation/admin-guide/mm/transhuge.rst index 5034915f4e8e..8872203df088 100644 --- a/Documentation/admin-guide/mm/transhuge.rst +++ b/Documentation/admin-guide/mm/transhuge.rst @@ -436,7 +436,7 @@ AnonHugePmdMapped). The number of file transparent huge pages mapped to userspace is available by reading ShmemPmdMapped and ShmemHugePages fields in ``/proc/meminfo``. To identify what applications are mapping file transparent huge pages, it -is necessary to read ``/proc/PID/smaps`` and count the FileHugeMapped fields +is necessary to read ``/proc/PID/smaps`` and count the FilePmdMapped fields for each mapping. Note that reading the smaps file is expensive and reading it From cb0ca08b326aa03f87fe94bb91872ce8d2ef1ed8 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 17 Dec 2024 08:18:10 +0100 Subject: [PATCH 126/196] kcov: mark in_softirq_really() as __always_inline If gcc decides not to inline in_softirq_really(), objtool warns about a function call with UACCESS enabled: kernel/kcov.o: warning: objtool: __sanitizer_cov_trace_pc+0x1e: call to in_softirq_really() with UACCESS enabled kernel/kcov.o: warning: objtool: check_kcov_mode+0x11: call to in_softirq_really() with UACCESS enabled Mark this as __always_inline to avoid the problem. Link: https://lkml.kernel.org/r/20241217071814.2261620-1-arnd@kernel.org Fixes: 7d4df2dad312 ("kcov: properly check for softirq context") Signed-off-by: Arnd Bergmann Reviewed-by: Marco Elver Cc: Aleksandr Nogikh Cc: Andrey Konovalov Cc: Dmitry Vyukov Cc: Josh Poimboeuf Cc: Peter Zijlstra Cc: Signed-off-by: Andrew Morton --- kernel/kcov.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/kcov.c b/kernel/kcov.c index 28a6be6e64fd..187ba1b80bda 100644 --- a/kernel/kcov.c +++ b/kernel/kcov.c @@ -166,7 +166,7 @@ static void kcov_remote_area_put(struct kcov_remote_area *area, * Unlike in_serving_softirq(), this function returns false when called during * a hardirq or an NMI that happened in the softirq context. */ -static inline bool in_softirq_really(void) +static __always_inline bool in_softirq_really(void) { return in_serving_softirq() && !in_hardirq() && !in_nmi(); } From 3754137d263f52f4b507cf9ae913f8f0497d1b0e Mon Sep 17 00:00:00 2001 From: David Hildenbrand Date: Tue, 17 Dec 2024 20:50:00 +0100 Subject: [PATCH 127/196] fs/proc/task_mmu: fix pagemap flags with PMD THP entries on 32bit Entries (including flags) are u64, even on 32bit. So right now we are cutting of the flags on 32bit. This way, for example the cow selftest complains about: # ./cow ... Bail Out! read and ioctl return unmatched results for populated: 0 1 Link: https://lkml.kernel.org/r/20241217195000.1734039-1-david@redhat.com Fixes: 2c1f057e5be6 ("fs/proc/task_mmu: properly detect PM_MMAP_EXCLUSIVE per page of PMD-mapped THPs") Signed-off-by: David Hildenbrand Cc: Oscar Salvador Cc: Signed-off-by: Andrew Morton --- fs/proc/task_mmu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 38a5a3e9cba2..f02cd362309a 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -1810,7 +1810,7 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end, } for (; addr != end; addr += PAGE_SIZE, idx++) { - unsigned long cur_flags = flags; + u64 cur_flags = flags; pagemap_entry_t pme; if (folio && (flags & PM_PRESENT) && From 5f3fd772d152229d94602bca243fbb658068a597 Mon Sep 17 00:00:00 2001 From: Dennis Lam Date: Tue, 17 Dec 2024 21:39:25 -0500 Subject: [PATCH 128/196] ocfs2: fix slab-use-after-free due to dangling pointer dqi_priv When mounting ocfs2 and then remounting it as read-only, a slab-use-after-free occurs after the user uses a syscall to quota_getnextquota. Specifically, sb_dqinfo(sb, type)->dqi_priv is the dangling pointer. During the remounting process, the pointer dqi_priv is freed but is never set as null leaving it to be accessed. Additionally, the read-only option for remounting sets the DQUOT_SUSPENDED flag instead of setting the DQUOT_USAGE_ENABLED flags. Moreover, later in the process of getting the next quota, the function ocfs2_get_next_id is called and only checks the quota usage flags and not the quota suspended flags. To fix this, I set dqi_priv to null when it is freed after remounting with read-only and put a check for DQUOT_SUSPENDED in ocfs2_get_next_id. [akpm@linux-foundation.org: coding-style cleanups] Link: https://lkml.kernel.org/r/20241218023924.22821-2-dennis.lamerice@gmail.com Fixes: 8f9e8f5fcc05 ("ocfs2: Fix Q_GETNEXTQUOTA for filesystem without quotas") Signed-off-by: Dennis Lam Reported-by: syzbot+d173bf8a5a7faeede34c@syzkaller.appspotmail.com Tested-by: syzbot+d173bf8a5a7faeede34c@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/6731d26f.050a0220.1fb99c.014b.GAE@google.com/T/ Reviewed-by: Joseph Qi Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Jun Piao Cc: Signed-off-by: Andrew Morton --- fs/ocfs2/quota_global.c | 2 +- fs/ocfs2/quota_local.c | 1 + 2 files changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index 2b0daced98eb..3404e7a30c33 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -893,7 +893,7 @@ static int ocfs2_get_next_id(struct super_block *sb, struct kqid *qid) int status = 0; trace_ocfs2_get_next_id(from_kqid(&init_user_ns, *qid), type); - if (!sb_has_quota_loaded(sb, type)) { + if (!sb_has_quota_active(sb, type)) { status = -ESRCH; goto out; } diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index 73d3367c533b..2956d888c131 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -867,6 +867,7 @@ static int ocfs2_local_free_info(struct super_block *sb, int type) brelse(oinfo->dqi_libh); brelse(oinfo->dqi_lqi_bh); kfree(oinfo); + info->dqi_priv = NULL; return status; } From eaebeb93922ca6ab0dd92027b73d0112701706ef Mon Sep 17 00:00:00 2001 From: Yosry Ahmed Date: Thu, 19 Dec 2024 21:24:37 +0000 Subject: [PATCH 129/196] mm: zswap: fix race between [de]compression and CPU hotunplug In zswap_compress() and zswap_decompress(), the per-CPU acomp_ctx of the current CPU at the beginning of the operation is retrieved and used throughout. However, since neither preemption nor migration are disabled, it is possible that the operation continues on a different CPU. If the original CPU is hotunplugged while the acomp_ctx is still in use, we run into a UAF bug as the resources attached to the acomp_ctx are freed during hotunplug in zswap_cpu_comp_dead(). The problem was introduced in commit 1ec3b5fe6eec ("mm/zswap: move to use crypto_acomp API for hardware acceleration") when the switch to the crypto_acomp API was made. Prior to that, the per-CPU crypto_comp was retrieved using get_cpu_ptr() which disables preemption and makes sure the CPU cannot go away from under us. Preemption cannot be disabled with the crypto_acomp API as a sleepable context is needed. Commit 8ba2f844f050 ("mm/zswap: change per-cpu mutex and buffer to per-acomp_ctx") increased the UAF surface area by making the per-CPU buffers dynamic, adding yet another resource that can be freed from under zswap compression/decompression by CPU hotunplug. There are a few ways to fix this: (a) Add a refcount for acomp_ctx. (b) Disable migration while using the per-CPU acomp_ctx. (c) Disable CPU hotunplug while using the per-CPU acomp_ctx by holding the CPUs read lock. Implement (c) since it's simpler than (a), and (b) involves using migrate_disable() which is apparently undesired (see huge comment in include/linux/preempt.h). Link: https://lkml.kernel.org/r/20241219212437.2714151-1-yosryahmed@google.com Fixes: 1ec3b5fe6eec ("mm/zswap: move to use crypto_acomp API for hardware acceleration") Signed-off-by: Yosry Ahmed Reported-by: Johannes Weiner Closes: https://lore.kernel.org/lkml/20241113213007.GB1564047@cmpxchg.org/ Reported-by: Sam Sun Closes: https://lore.kernel.org/lkml/CAEkJfYMtSdM5HceNsXUDf5haghD5+o2e7Qv4OcuruL4tPg6OaQ@mail.gmail.com/ Reviewed-by: Chengming Zhou Acked-by: Barry Song Reviewed-by: Nhat Pham Cc: Vitaly Wool Cc: Signed-off-by: Andrew Morton --- mm/zswap.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/mm/zswap.c b/mm/zswap.c index f6316b66fb23..5a27af8d86ea 100644 --- a/mm/zswap.c +++ b/mm/zswap.c @@ -880,6 +880,18 @@ static int zswap_cpu_comp_dead(unsigned int cpu, struct hlist_node *node) return 0; } +/* Prevent CPU hotplug from freeing up the per-CPU acomp_ctx resources */ +static struct crypto_acomp_ctx *acomp_ctx_get_cpu(struct crypto_acomp_ctx __percpu *acomp_ctx) +{ + cpus_read_lock(); + return raw_cpu_ptr(acomp_ctx); +} + +static void acomp_ctx_put_cpu(void) +{ + cpus_read_unlock(); +} + static bool zswap_compress(struct page *page, struct zswap_entry *entry, struct zswap_pool *pool) { @@ -893,8 +905,7 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry, gfp_t gfp; u8 *dst; - acomp_ctx = raw_cpu_ptr(pool->acomp_ctx); - + acomp_ctx = acomp_ctx_get_cpu(pool->acomp_ctx); mutex_lock(&acomp_ctx->mutex); dst = acomp_ctx->buffer; @@ -950,6 +961,7 @@ static bool zswap_compress(struct page *page, struct zswap_entry *entry, zswap_reject_alloc_fail++; mutex_unlock(&acomp_ctx->mutex); + acomp_ctx_put_cpu(); return comp_ret == 0 && alloc_ret == 0; } @@ -960,7 +972,7 @@ static void zswap_decompress(struct zswap_entry *entry, struct folio *folio) struct crypto_acomp_ctx *acomp_ctx; u8 *src; - acomp_ctx = raw_cpu_ptr(entry->pool->acomp_ctx); + acomp_ctx = acomp_ctx_get_cpu(entry->pool->acomp_ctx); mutex_lock(&acomp_ctx->mutex); src = zpool_map_handle(zpool, entry->handle, ZPOOL_MM_RO); @@ -990,6 +1002,7 @@ static void zswap_decompress(struct zswap_entry *entry, struct folio *folio) if (src != acomp_ctx->buffer) zpool_unmap_handle(zpool, entry->handle); + acomp_ctx_put_cpu(); } /********************************* From 11673247700e2af3a6a95f7b3f1bb80b691c950e Mon Sep 17 00:00:00 2001 From: Gal Pressman Date: Thu, 19 Dec 2024 14:18:28 +0200 Subject: [PATCH 130/196] percpu: remove intermediate variable in PERCPU_PTR() The intermediate variable in the PERCPU_PTR() macro results in a kernel panic on boot [1] due to a compiler bug seen when compiling the kernel (+ KASAN) with gcc 11.3.1, but not when compiling with latest gcc (v14.2)/clang(v18.1). To solve it, remove the intermediate variable (which is not needed) and keep the casting that resolves the address space checks. [1] Oops: general protection fault, probably for non-canonical address 0xdffffc0000000003: 0000 [#1] SMP KASAN KASAN: null-ptr-deref in range [0x0000000000000018-0x000000000000001f] CPU: 0 UID: 0 PID: 547 Comm: iptables Not tainted 6.13.0-rc1_external_tested-master #1 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.org 04/01/2014 RIP: 0010:nf_ct_netns_do_get+0x139/0x540 Code: 03 00 00 48 81 c4 88 00 00 00 5b 5d 41 5c 41 5d 41 5e 41 5f c3 4d 8d 75 08 48 b8 00 00 00 00 00 fc ff df 4c 89 f2 48 c1 ea 03 <0f> b6 04 02 84 c0 74 08 3c 03 0f 8e 27 03 00 00 41 8b 45 08 83 c0 RSP: 0018:ffff888116df75e8 EFLAGS: 00010207 RAX: dffffc0000000000 RBX: 1ffff11022dbeebe RCX: ffffffff839a2382 RDX: 0000000000000003 RSI: 0000000000000008 RDI: ffff88842ec46d10 RBP: 0000000000000002 R08: 0000000000000000 R09: fffffbfff0b0860c R10: ffff888116df75e8 R11: 0000000000000001 R12: ffffffff879d6a80 R13: 0000000000000016 R14: 000000000000001e R15: ffff888116df7908 FS: 00007fba01646740(0000) GS:ffff88842ec00000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 000055bd901800d8 CR3: 00000001205f0003 CR4: 0000000000172eb0 Call Trace: ? die_addr+0x3d/0xa0 ? exc_general_protection+0x144/0x220 ? asm_exc_general_protection+0x22/0x30 ? __mutex_lock+0x2c2/0x1d70 ? nf_ct_netns_do_get+0x139/0x540 ? nf_ct_netns_do_get+0xb5/0x540 ? net_generic+0x1f0/0x1f0 ? __create_object+0x5e/0x80 xt_check_target+0x1f0/0x930 ? textify_hooks.constprop.0+0x110/0x110 ? pcpu_alloc_noprof+0x7cd/0xcf0 ? xt_find_target+0x148/0x1e0 find_check_entry.constprop.0+0x6c0/0x920 ? get_info+0x380/0x380 ? __virt_addr_valid+0x1df/0x3b0 ? kasan_quarantine_put+0xe3/0x200 ? kfree+0x13e/0x3d0 ? translate_table+0xaf5/0x1750 translate_table+0xbd8/0x1750 ? ipt_unregister_table_exit+0x30/0x30 ? __might_fault+0xbb/0x170 do_ipt_set_ctl+0x408/0x1340 ? nf_sockopt_find.constprop.0+0x17b/0x1f0 ? lock_downgrade+0x680/0x680 ? lockdep_hardirqs_on_prepare+0x284/0x400 ? ipt_register_table+0x440/0x440 ? bit_wait_timeout+0x160/0x160 nf_setsockopt+0x6f/0xd0 raw_setsockopt+0x7e/0x200 ? raw_bind+0x590/0x590 ? do_user_addr_fault+0x812/0xd20 do_sock_setsockopt+0x1e2/0x3f0 ? move_addr_to_user+0x90/0x90 ? lock_downgrade+0x680/0x680 __sys_setsockopt+0x9e/0x100 __x64_sys_setsockopt+0xb9/0x150 ? do_syscall_64+0x33/0x140 do_syscall_64+0x6d/0x140 entry_SYSCALL_64_after_hwframe+0x4b/0x53 RIP: 0033:0x7fba015134ce Code: 0f 1f 40 00 48 8b 15 59 69 0e 00 f7 d8 64 89 02 48 c7 c0 ff ff ff ff eb b1 0f 1f 00 f3 0f 1e fa 49 89 ca b8 36 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 0a c3 66 0f 1f 84 00 00 00 00 00 48 8b 15 21 RSP: 002b:00007ffd9de6f388 EFLAGS: 00000246 ORIG_RAX: 0000000000000036 RAX: ffffffffffffffda RBX: 000055bd9017f490 RCX: 00007fba015134ce RDX: 0000000000000040 RSI: 0000000000000000 RDI: 0000000000000004 RBP: 0000000000000500 R08: 0000000000000560 R09: 0000000000000052 R10: 000055bd901800e0 R11: 0000000000000246 R12: 000055bd90180140 R13: 000055bd901800e0 R14: 000055bd9017f498 R15: 000055bd9017ff10 Modules linked in: xt_MASQUERADE nf_conntrack_netlink nfnetlink xt_addrtype iptable_nat nf_nat br_netfilter rpcsec_gss_krb5 auth_rpcgss oid_registry overlay zram zsmalloc mlx4_ib mlx4_en mlx4_core rpcrdma rdma_ucm ib_uverbs ib_iser libiscsi scsi_transport_iscsi fuse ib_umad rdma_cm ib_ipoib iw_cm ib_cm ib_core ---[ end trace 0000000000000000 ]--- [akpm@linux-foundation.org: simplification, per Uros] Link: https://lkml.kernel.org/r/20241219121828.2120780-1-gal@nvidia.com Fixes: dabddd687c9e ("percpu: cast percpu pointer in PERCPU_PTR() via unsigned long") Signed-off-by: Gal Pressman Closes: https://lore.kernel.org/all/7590f546-4021-4602-9252-0d525de35b52@nvidia.com Cc: Uros Bizjak Cc: Bill Wendling Cc: Christoph Lameter Cc: Dennis Zhou Cc: Justin Stitt Cc: Nathan Chancellor Cc: Nick Desaulniers Cc: Tejun Heo Signed-off-by: Andrew Morton --- include/linux/percpu-defs.h | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/include/linux/percpu-defs.h b/include/linux/percpu-defs.h index 35842d1e3879..5b520fe86b60 100644 --- a/include/linux/percpu-defs.h +++ b/include/linux/percpu-defs.h @@ -221,10 +221,7 @@ do { \ } while (0) #define PERCPU_PTR(__p) \ -({ \ - unsigned long __pcpu_ptr = (__force unsigned long)(__p); \ - (typeof(*(__p)) __force __kernel *)(__pcpu_ptr); \ -}) + (typeof(*(__p)) __force __kernel *)((__force unsigned long)(__p)) #ifdef CONFIG_SMP From d0e6983a6d1719738cf8d13982a68094f0a1872a Mon Sep 17 00:00:00 2001 From: Baolin Wang Date: Thu, 19 Dec 2024 15:30:08 +0800 Subject: [PATCH 131/196] mm: shmem: fix incorrect index alignment for within_size policy With enabling the shmem per-size within_size policy, using an incorrect 'order' size to round_up() the index can lead to incorrect i_size checks, resulting in an inappropriate large orders being returned. Changing to use '1 << order' to round_up() the index to fix this issue. Additionally, adding an 'aligned_index' variable to avoid affecting the index checks. Link: https://lkml.kernel.org/r/77d8ef76a7d3d646e9225e9af88a76549a68aab1.1734593154.git.baolin.wang@linux.alibaba.com Fixes: e7a2ab7b3bb5 ("mm: shmem: add mTHP support for anonymous shmem") Signed-off-by: Baolin Wang Acked-by: David Hildenbrand Cc: Hugh Dickins Cc: Signed-off-by: Andrew Morton --- mm/shmem.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/mm/shmem.c b/mm/shmem.c index f6fb053ac50d..dec659e84562 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -1689,6 +1689,7 @@ unsigned long shmem_allowable_huge_orders(struct inode *inode, unsigned long mask = READ_ONCE(huge_shmem_orders_always); unsigned long within_size_orders = READ_ONCE(huge_shmem_orders_within_size); unsigned long vm_flags = vma ? vma->vm_flags : 0; + pgoff_t aligned_index; bool global_huge; loff_t i_size; int order; @@ -1723,9 +1724,9 @@ unsigned long shmem_allowable_huge_orders(struct inode *inode, /* Allow mTHP that will be fully within i_size. */ order = highest_order(within_size_orders); while (within_size_orders) { - index = round_up(index + 1, order); + aligned_index = round_up(index + 1, 1 << order); i_size = round_up(i_size_read(inode), PAGE_SIZE); - if (i_size >> PAGE_SHIFT >= index) { + if (i_size >> PAGE_SHIFT >= aligned_index) { mask |= within_size_orders; break; } From d77b90d2b2642655b5f60953c36ad887257e1802 Mon Sep 17 00:00:00 2001 From: Baolin Wang Date: Thu, 19 Dec 2024 15:30:09 +0800 Subject: [PATCH 132/196] mm: shmem: fix the update of 'shmem_falloc->nr_unswapped' The 'shmem_falloc->nr_unswapped' is used to record how many writepage refused to swap out because fallocate() is allocating, but after shmem supports large folio swap out, the update of 'shmem_falloc->nr_unswapped' does not use the correct number of pages in the large folio, which may lead to fallocate() not exiting as soon as possible. Anyway, this is found through code inspection, and I am not sure whether it would actually cause serious issues. Link: https://lkml.kernel.org/r/f66a0119d0564c2c37c84f045835b870d1b2196f.1734593154.git.baolin.wang@linux.alibaba.com Fixes: 809bc86517cc ("mm: shmem: support large folio swap out") Signed-off-by: Baolin Wang Acked-by: David Hildenbrand Cc: Hugh Dickins Cc: Signed-off-by: Andrew Morton --- mm/shmem.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/shmem.c b/mm/shmem.c index dec659e84562..ac58d4fb2e6f 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -1535,7 +1535,7 @@ static int shmem_writepage(struct page *page, struct writeback_control *wbc) !shmem_falloc->waitq && index >= shmem_falloc->start && index < shmem_falloc->next) - shmem_falloc->nr_unswapped++; + shmem_falloc->nr_unswapped += nr_pages; else shmem_falloc = NULL; spin_unlock(&inode->i_lock); From adcfb264c3ed51fbbf5068ddf10d309a63683868 Mon Sep 17 00:00:00 2001 From: Koichiro Den Date: Sat, 21 Dec 2024 12:33:20 +0900 Subject: [PATCH 133/196] vmstat: disable vmstat_work on vmstat_cpu_down_prep() Even after mm/vmstat:online teardown, shepherd may still queue work for the dying cpu until the cpu is removed from online mask. While it's quite rare, this means that after unbind_workers() unbinds a per-cpu kworker, it potentially runs vmstat_update for the dying CPU on an irrelevant cpu before entering atomic AP states. When CONFIG_DEBUG_PREEMPT=y, it results in the following error with the backtrace. BUG: using smp_processor_id() in preemptible [00000000] code: \ kworker/7:3/1702 caller is refresh_cpu_vm_stats+0x235/0x5f0 CPU: 0 UID: 0 PID: 1702 Comm: kworker/7:3 Tainted: G Tainted: [N]=TEST Workqueue: mm_percpu_wq vmstat_update Call Trace: dump_stack_lvl+0x8d/0xb0 check_preemption_disabled+0xce/0xe0 refresh_cpu_vm_stats+0x235/0x5f0 vmstat_update+0x17/0xa0 process_one_work+0x869/0x1aa0 worker_thread+0x5e5/0x1100 kthread+0x29e/0x380 ret_from_fork+0x2d/0x70 ret_from_fork_asm+0x1a/0x30 So, for mm/vmstat:online, disable vmstat_work reliably on teardown and symmetrically enable it on startup. Link: https://lkml.kernel.org/r/20241221033321.4154409-1-koichiro.den@canonical.com Signed-off-by: Koichiro Den Cc: Sebastian Andrzej Siewior Cc: Signed-off-by: Andrew Morton --- mm/vmstat.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/mm/vmstat.c b/mm/vmstat.c index 4d016314a56c..0889b75cef14 100644 --- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -2148,13 +2148,14 @@ static int vmstat_cpu_online(unsigned int cpu) if (!node_state(cpu_to_node(cpu), N_CPU)) { node_set_state(cpu_to_node(cpu), N_CPU); } + enable_delayed_work(&per_cpu(vmstat_work, cpu)); return 0; } static int vmstat_cpu_down_prep(unsigned int cpu) { - cancel_delayed_work_sync(&per_cpu(vmstat_work, cpu)); + disable_delayed_work_sync(&per_cpu(vmstat_work, cpu)); return 0; } From 98a6abc6cec186bdc3d94c162227cc8e003de76c Mon Sep 17 00:00:00 2001 From: Kairui Song Date: Mon, 23 Dec 2024 23:09:07 +0800 Subject: [PATCH 134/196] mm/list_lru: fix false warning of negative counter commit 2788cf0c401c ("memcg: reparent list_lrus and free kmemcg_id on css offline") removed sanity checks for the nr_items counter's value because it implemented list_lru re-parenting in a way that will redirect children's list_lru to the parent before re-parenting the items in list_lru. This will make item counter uncharging happen in the parent while the item is still being held by the child. As a result, the parent's counter value may become negative. This is acceptable because re-parenting will sum up the children's counter values, and the parent's counter will be fixed. Later commit fb56fdf8b9a2 ("mm/list_lru: split the lock to per-cgroup scope") reworked the re-parenting process, and removed the redirect. So it added the sanity check back, assuming that as long as items are still in the children's list_lru, parent's counter will not be uncharged. But that assumption is incorrect. The xas_store in memcg_reparent_list_lrus will set children's list_lru to NULL before re-parenting the items, it redirects list_lru helpers to use parent's list_lru just like before. But still, it's not a problem as re-parenting will fix the counter. Therefore, remove this sanity check, but add a new check to ensure that the counter won't go negative in a different way: the child's list_lru being re-parented should never have a negative counter, since re-parenting should occur in order and fixes counters. Link: https://lkml.kernel.org/r/20241223150907.1591-1-ryncsn@gmail.com Fixes: fb56fdf8b9a2 ("mm/list_lru: split the lock to per-cgroup scope") Signed-off-by: Kairui Song Closes: https://lore.kernel.org/lkml/Z2Bz9t92Be9l1xqj@lappy/ Cc: Chengming Zhou Cc: Johannes Weiner Cc: Matthew Wilcox Cc: Michal Hocko Cc: Muchun Song Cc: Qi Zheng Cc: Roman Gushchin Cc: Sasha Levin Cc: Shakeel Butt Cc: Yu Zhao Signed-off-by: Andrew Morton --- mm/list_lru.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/list_lru.c b/mm/list_lru.c index f93ada6a207b..7d69434c70e0 100644 --- a/mm/list_lru.c +++ b/mm/list_lru.c @@ -77,7 +77,6 @@ lock_list_lru_of_memcg(struct list_lru *lru, int nid, struct mem_cgroup *memcg, spin_lock(&l->lock); nr_items = READ_ONCE(l->nr_items); if (likely(nr_items != LONG_MIN)) { - WARN_ON(nr_items < 0); rcu_read_unlock(); return l; } @@ -450,6 +449,7 @@ static void memcg_reparent_list_lru_one(struct list_lru *lru, int nid, list_splice_init(&src->list, &dst->list); if (src->nr_items) { + WARN_ON(src->nr_items < 0); dst->nr_items += src->nr_items; set_shrinker_bit(dst_memcg, nid, lru_shrinker_id(lru)); } From 8debfc5b1aa569d3d2ac836af2553da037611c61 Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Sun, 22 Dec 2024 15:12:21 -0800 Subject: [PATCH 135/196] mm/damon/core: fix new damon_target objects leaks on damon_commit_targets() Patch series "mm/damon/core: fix memory leaks and ignored inputs from damon_commit_ctx()". Due to two bugs in damon_commit_targets() and damon_commit_schemes(), which are called from damon_commit_ctx(), some user inputs can be ignored, and some mmeory objects can be leaked. Fix those. Note that only DAMON sysfs interface users are affected. Other DAMON core API user modules that more focused more on simple and dedicated production usages, including DAMON_RECLAIM and DAMON_LRU_SORT are not using the buggy function in the way, so not affected. This patch (of 2): When new DAMON targets are added via damon_commit_targets(), the newly created targets are not deallocated when updating the internal data (damon_commit_target()) is failed. Worse yet, even if the setup is successfully done, the new target is not linked to the context. Hence, the new targets are always leaked regardless of the internal data setup failure. Fix the leaks. Link: https://lkml.kernel.org/r/20241222231222.85060-2-sj@kernel.org Fixes: 9cb3d0b9dfce ("mm/damon/core: implement DAMON context commit function") Signed-off-by: SeongJae Park Cc: Signed-off-by: Andrew Morton --- mm/damon/core.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/mm/damon/core.c b/mm/damon/core.c index 8b8e2933dcd4..dc52361f1863 100644 --- a/mm/damon/core.c +++ b/mm/damon/core.c @@ -961,8 +961,11 @@ static int damon_commit_targets( return -ENOMEM; err = damon_commit_target(new_target, false, src_target, damon_target_has_pid(src)); - if (err) + if (err) { + damon_destroy_target(new_target); return err; + } + damon_add_target(dst, new_target); } return 0; } From 7d390b53067ef745e2d9bee5a9683df4c96b80a0 Mon Sep 17 00:00:00 2001 From: SeongJae Park Date: Sun, 22 Dec 2024 15:12:22 -0800 Subject: [PATCH 136/196] mm/damon/core: fix ignored quota goals and filters of newly committed schemes damon_commit_schemes() ignores quota goals and filters of the newly committed schemes. This makes users confused about the behaviors. Correctly handle those inputs. Link: https://lkml.kernel.org/r/20241222231222.85060-3-sj@kernel.org Fixes: 9cb3d0b9dfce ("mm/damon/core: implement DAMON context commit function") Signed-off-by: SeongJae Park Cc: Signed-off-by: Andrew Morton --- mm/damon/core.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/mm/damon/core.c b/mm/damon/core.c index dc52361f1863..0776452a1abb 100644 --- a/mm/damon/core.c +++ b/mm/damon/core.c @@ -868,6 +868,11 @@ static int damon_commit_schemes(struct damon_ctx *dst, struct damon_ctx *src) NUMA_NO_NODE); if (!new_scheme) return -ENOMEM; + err = damos_commit(new_scheme, src_scheme); + if (err) { + damon_destroy_scheme(new_scheme); + return err; + } damon_add_scheme(dst, new_scheme); } return 0; From 62e72d2cf702a5e2fb53d9c46ed900d9384e4a06 Mon Sep 17 00:00:00 2001 From: Kairui Song Date: Sun, 22 Dec 2024 20:29:36 +0800 Subject: [PATCH 137/196] mm, madvise: fix potential workingset node list_lru leaks Since commit 5abc1e37afa0 ("mm: list_lru: allocate list_lru_one only when needed"), all list_lru users need to allocate the items using the new infrastructure that provides list_lru info for slab allocation, ensuring that the corresponding memcg list_lru is allocated before use. For workingset shadow nodes (which are xa_node), users are converted to use the new infrastructure by commit 9bbdc0f32409 ("xarray: use kmem_cache_alloc_lru to allocate xa_node"). The xas->xa_lru will be set correctly for filemap users. However, there is a missing case: xa_node allocations caused by madvise(..., MADV_COLLAPSE). madvise(..., MADV_COLLAPSE) will also read in the absent parts of file map, and there will be xa_nodes allocated for the caller's memcg (assuming it's not rootcg). However, these allocations won't trigger memcg list_lru allocation because the proper xas info was not set. If nothing else has allocated other xa_nodes for that memcg to trigger list_lru creation, and memory pressure starts to evict file pages, workingset_update_node will try to add these xa_nodes to their corresponding memcg list_lru, and it does not exist (NULL). So they will be added to rootcg's list_lru instead. This shouldn't be a significant issue in practice, but it is indeed unexpected behavior, and these xa_nodes will not be reclaimed effectively. And may lead to incorrect counting of the list_lru->nr_items counter. This problem wasn't exposed until recent commit 28e98022b31ef ("mm/list_lru: simplify reparenting and initial allocation") added a sanity check: only dying memcg could have a NULL list_lru when list_lru_{add,del} is called. This problem triggered this WARNING. So make madvise(..., MADV_COLLAPSE) also call xas_set_lru() to pass the list_lru which we may want to insert xa_node into later. And move mapping_set_update to mm/internal.h, and turn into a macro to avoid including extra headers in mm/internal.h. Link: https://lkml.kernel.org/r/20241222122936.67501-1-ryncsn@gmail.com Fixes: 9bbdc0f32409 ("xarray: use kmem_cache_alloc_lru to allocate xa_node") Reported-by: syzbot+38a0cbd267eff2d286ff@syzkaller.appspotmail.com Closes: https://lore.kernel.org/lkml/675d01e9.050a0220.37aaf.00be.GAE@google.com/ Signed-off-by: Kairui Song Cc: Chengming Zhou Cc: Johannes Weiner Cc: Matthew Wilcox (Oracle) Cc: Michal Hocko Cc: Muchun Song Cc: Qi Zheng Cc: Roman Gushchin Cc: Sasha Levin Cc: Shakeel Butt Cc: Yu Zhao Signed-off-by: Andrew Morton --- mm/filemap.c | 9 --------- mm/internal.h | 6 ++++++ mm/khugepaged.c | 3 +++ 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/mm/filemap.c b/mm/filemap.c index f61cf51c2238..33b60d448fca 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -124,15 +124,6 @@ * ->private_lock (zap_pte_range->block_dirty_folio) */ -static void mapping_set_update(struct xa_state *xas, - struct address_space *mapping) -{ - if (dax_mapping(mapping) || shmem_mapping(mapping)) - return; - xas_set_update(xas, workingset_update_node); - xas_set_lru(xas, &shadow_nodes); -} - static void page_cache_delete(struct address_space *mapping, struct folio *folio, void *shadow) { diff --git a/mm/internal.h b/mm/internal.h index 3bd08bafad04..9826f7dce607 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -1504,6 +1504,12 @@ static inline void shrinker_debugfs_remove(struct dentry *debugfs_entry, /* Only track the nodes of mappings with shadow entries */ void workingset_update_node(struct xa_node *node); extern struct list_lru shadow_nodes; +#define mapping_set_update(xas, mapping) do { \ + if (!dax_mapping(mapping) && !shmem_mapping(mapping)) { \ + xas_set_update(xas, workingset_update_node); \ + xas_set_lru(xas, &shadow_nodes); \ + } \ +} while (0) /* mremap.c */ unsigned long move_page_tables(struct vm_area_struct *vma, diff --git a/mm/khugepaged.c b/mm/khugepaged.c index 6f8d46d107b4..653dbb1ff05c 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -19,6 +19,7 @@ #include #include #include +#include #include #include @@ -1837,6 +1838,8 @@ static int collapse_file(struct mm_struct *mm, unsigned long addr, if (result != SCAN_SUCCEED) goto out; + mapping_set_update(&xas, mapping); + __folio_set_locked(new_folio); if (is_shmem) __folio_set_swapbacked(new_folio); From dd2a5b5514ab0e690f018595e34dd1fcb981d345 Mon Sep 17 00:00:00 2001 From: Tetsuo Handa Date: Sat, 21 Dec 2024 16:47:29 +0900 Subject: [PATCH 138/196] mm/util: make memdup_user_nul() similar to memdup_user() Since the string data to copy from userspace is likely less than PAGE_SIZE bytes, replace GFP_KERNEL with GFP_USER like commit 6c2c97a24f09 ("memdup_user(): switch to GFP_USER") does and add __GFP_NOWARN like commit 6c8fcc096be9 ("mm: don't let userspace spam allocations warnings") does. Also, use dedicated slab buckets like commit d73778e4b867 ("mm/util: Use dedicated slab buckets for memdup_user()") does. Link: https://lkml.kernel.org/r/014cd694-cc27-4a07-a34a-2ae95d744515@I-love.SAKURA.ne.jp Reported-by: syzbot+7e12e97b36154c54414b@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=7e12e97b36154c54414b Signed-off-by: Tetsuo Handa Signed-off-by: Andrew Morton --- mm/util.c | 7 +------ 1 file changed, 1 insertion(+), 6 deletions(-) diff --git a/mm/util.c b/mm/util.c index c1c3b06ab4f9..60aa40f612b8 100644 --- a/mm/util.c +++ b/mm/util.c @@ -297,12 +297,7 @@ void *memdup_user_nul(const void __user *src, size_t len) { char *p; - /* - * Always use GFP_KERNEL, since copy_from_user() can sleep and - * cause pagefault, which makes it pointless to use GFP_NOFS - * or GFP_ATOMIC. - */ - p = kmalloc_track_caller(len + 1, GFP_KERNEL); + p = kmem_buckets_alloc_track_caller(user_buckets, len + 1, GFP_USER | __GFP_NOWARN); if (!p) return ERR_PTR(-ENOMEM); From 0210d251162f4033350a94a43f95b1c39ec84a90 Mon Sep 17 00:00:00 2001 From: Kuan-Wei Chiu Date: Thu, 26 Dec 2024 22:03:32 +0800 Subject: [PATCH 139/196] scripts/sorttable: fix orc_sort_cmp() to maintain symmetry and transitivity The orc_sort_cmp() function, used with qsort(), previously violated the symmetry and transitivity rules required by the C standard. Specifically, when both entries are ORC_TYPE_UNDEFINED, it could result in both a < b and b < a, which breaks the required symmetry and transitivity. This can lead to undefined behavior and incorrect sorting results, potentially causing memory corruption in glibc implementations [1]. Symmetry: If x < y, then y > x. Transitivity: If x < y and y < z, then x < z. Fix the comparison logic to return 0 when both entries are ORC_TYPE_UNDEFINED, ensuring compliance with qsort() requirements. Link: https://www.qualys.com/2024/01/30/qsort.txt [1] Link: https://lkml.kernel.org/r/20241226140332.2670689-1-visitorckw@gmail.com Fixes: 57fa18994285 ("scripts/sorttable: Implement build-time ORC unwind table sorting") Fixes: fb799447ae29 ("x86,objtool: Split UNWIND_HINT_EMPTY in two") Signed-off-by: Kuan-Wei Chiu Cc: Ching-Chun (Jim) Huang Cc: Cc: Ingo Molnar Cc: Josh Poimboeuf Cc: Peter Zijlstra Cc: Shile Zhang Cc: Steven Rostedt Cc: Signed-off-by: Andrew Morton --- scripts/sorttable.h | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/scripts/sorttable.h b/scripts/sorttable.h index 7bd0184380d3..a7c5445baf00 100644 --- a/scripts/sorttable.h +++ b/scripts/sorttable.h @@ -110,7 +110,7 @@ static inline unsigned long orc_ip(const int *ip) static int orc_sort_cmp(const void *_a, const void *_b) { - struct orc_entry *orc_a; + struct orc_entry *orc_a, *orc_b; const int *a = g_orc_ip_table + *(int *)_a; const int *b = g_orc_ip_table + *(int *)_b; unsigned long a_val = orc_ip(a); @@ -128,6 +128,9 @@ static int orc_sort_cmp(const void *_a, const void *_b) * whitelisted .o files which didn't get objtool generation. */ orc_a = g_orc_table + (a - g_orc_ip_table); + orc_b = g_orc_table + (b - g_orc_ip_table); + if (orc_a->type == ORC_TYPE_UNDEFINED && orc_b->type == ORC_TYPE_UNDEFINED) + return 0; return orc_a->type == ORC_TYPE_UNDEFINED ? -1 : 1; } From e7404921818d676da4d7143ce78659456b05e2af Mon Sep 17 00:00:00 2001 From: "Chester A. Unal" Date: Wed, 25 Dec 2024 15:50:41 +0300 Subject: [PATCH 140/196] =?UTF-8?q?MAINTAINERS:=20change=20Ar=C4=B1n=C3=A7?= =?UTF-8?q?=20=5FNAL's=20name=20and=20email=20address?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit My legal name now includes Chester. Change the name and the email address sections to reflect that. Link: https://lkml.kernel.org/r/20241225-for-unknown-upstream-v1-1-3e35e4d5e161@arinc9.com Signed-off-by: Chester A. Unal Signed-off-by: Andrew Morton --- MAINTAINERS | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/MAINTAINERS b/MAINTAINERS index 910305c11e8a..22fa261cb60e 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -14756,7 +14756,7 @@ F: drivers/memory/mtk-smi.c F: include/soc/mediatek/smi.h MEDIATEK SWITCH DRIVER -M: Arınç ÜNAL +M: Chester A. Unal M: Daniel Golle M: DENG Qingfang M: Sean Wang @@ -18460,7 +18460,7 @@ F: Documentation/devicetree/bindings/pinctrl/mediatek,mt8183-pinctrl.yaml F: drivers/pinctrl/mediatek/ PIN CONTROLLER - MEDIATEK MIPS -M: Arınç ÜNAL +M: Chester A. Unal M: Sergio Paracuellos L: linux-mediatek@lists.infradead.org (moderated for non-subscribers) L: linux-mips@vger.kernel.org @@ -19504,7 +19504,7 @@ S: Maintained F: arch/mips/ralink RALINK MT7621 MIPS ARCHITECTURE -M: Arınç ÜNAL +M: Chester A. Unal M: Sergio Paracuellos L: linux-mips@vger.kernel.org S: Maintained From 4f619d518db9cd1a933c3a095a5f95d0c1584ae8 Mon Sep 17 00:00:00 2001 From: Jinjian Song Date: Tue, 24 Dec 2024 12:15:52 +0800 Subject: [PATCH 141/196] net: wwan: t7xx: Fix FSM command timeout issue When driver processes the internal state change command, it use an asynchronous thread to process the command operation. If the main thread detects that the task has timed out, the asynchronous thread will panic when executing the completion notification because the main thread completion object has been released. BUG: unable to handle page fault for address: fffffffffffffff8 PGD 1f283a067 P4D 1f283a067 PUD 1f283c067 PMD 0 Oops: 0000 [#1] PREEMPT SMP NOPTI RIP: 0010:complete_all+0x3e/0xa0 [...] Call Trace: ? __die_body+0x68/0xb0 ? page_fault_oops+0x379/0x3e0 ? exc_page_fault+0x69/0xa0 ? asm_exc_page_fault+0x22/0x30 ? complete_all+0x3e/0xa0 fsm_main_thread+0xa3/0x9c0 [mtk_t7xx (HASH:1400 5)] ? __pfx_autoremove_wake_function+0x10/0x10 kthread+0xd8/0x110 ? __pfx_fsm_main_thread+0x10/0x10 [mtk_t7xx (HASH:1400 5)] ? __pfx_kthread+0x10/0x10 ret_from_fork+0x38/0x50 ? __pfx_kthread+0x10/0x10 ret_from_fork_asm+0x1b/0x30 [...] CR2: fffffffffffffff8 ---[ end trace 0000000000000000 ]--- Use the reference counter to ensure safe release as Sergey suggests: https://lore.kernel.org/all/da90f64c-260a-4329-87bf-1f9ff20a5951@gmail.com/ Fixes: 13e920d93e37 ("net: wwan: t7xx: Add core components") Signed-off-by: Jinjian Song Acked-by: Sergey Ryazanov Link: https://patch.msgid.link/20241224041552.8711-1-jinjian.song@fibocom.com Signed-off-by: Jakub Kicinski --- drivers/net/wwan/t7xx/t7xx_state_monitor.c | 26 ++++++++++++++-------- drivers/net/wwan/t7xx/t7xx_state_monitor.h | 5 +++-- 2 files changed, 20 insertions(+), 11 deletions(-) diff --git a/drivers/net/wwan/t7xx/t7xx_state_monitor.c b/drivers/net/wwan/t7xx/t7xx_state_monitor.c index 3931c7a13f5a..cbdbb91e8381 100644 --- a/drivers/net/wwan/t7xx/t7xx_state_monitor.c +++ b/drivers/net/wwan/t7xx/t7xx_state_monitor.c @@ -104,14 +104,21 @@ void t7xx_fsm_broadcast_state(struct t7xx_fsm_ctl *ctl, enum md_state state) fsm_state_notify(ctl->md, state); } +static void fsm_release_command(struct kref *ref) +{ + struct t7xx_fsm_command *cmd = container_of(ref, typeof(*cmd), refcnt); + + kfree(cmd); +} + static void fsm_finish_command(struct t7xx_fsm_ctl *ctl, struct t7xx_fsm_command *cmd, int result) { if (cmd->flag & FSM_CMD_FLAG_WAIT_FOR_COMPLETION) { - *cmd->ret = result; - complete_all(cmd->done); + cmd->result = result; + complete_all(&cmd->done); } - kfree(cmd); + kref_put(&cmd->refcnt, fsm_release_command); } static void fsm_del_kf_event(struct t7xx_fsm_event *event) @@ -475,7 +482,6 @@ static int fsm_main_thread(void *data) int t7xx_fsm_append_cmd(struct t7xx_fsm_ctl *ctl, enum t7xx_fsm_cmd_state cmd_id, unsigned int flag) { - DECLARE_COMPLETION_ONSTACK(done); struct t7xx_fsm_command *cmd; unsigned long flags; int ret; @@ -487,11 +493,13 @@ int t7xx_fsm_append_cmd(struct t7xx_fsm_ctl *ctl, enum t7xx_fsm_cmd_state cmd_id INIT_LIST_HEAD(&cmd->entry); cmd->cmd_id = cmd_id; cmd->flag = flag; + kref_init(&cmd->refcnt); if (flag & FSM_CMD_FLAG_WAIT_FOR_COMPLETION) { - cmd->done = &done; - cmd->ret = &ret; + init_completion(&cmd->done); + kref_get(&cmd->refcnt); } + kref_get(&cmd->refcnt); spin_lock_irqsave(&ctl->command_lock, flags); list_add_tail(&cmd->entry, &ctl->command_queue); spin_unlock_irqrestore(&ctl->command_lock, flags); @@ -501,11 +509,11 @@ int t7xx_fsm_append_cmd(struct t7xx_fsm_ctl *ctl, enum t7xx_fsm_cmd_state cmd_id if (flag & FSM_CMD_FLAG_WAIT_FOR_COMPLETION) { unsigned long wait_ret; - wait_ret = wait_for_completion_timeout(&done, + wait_ret = wait_for_completion_timeout(&cmd->done, msecs_to_jiffies(FSM_CMD_TIMEOUT_MS)); - if (!wait_ret) - return -ETIMEDOUT; + ret = wait_ret ? cmd->result : -ETIMEDOUT; + kref_put(&cmd->refcnt, fsm_release_command); return ret; } diff --git a/drivers/net/wwan/t7xx/t7xx_state_monitor.h b/drivers/net/wwan/t7xx/t7xx_state_monitor.h index 7b0a9baf488c..6e0601bb752e 100644 --- a/drivers/net/wwan/t7xx/t7xx_state_monitor.h +++ b/drivers/net/wwan/t7xx/t7xx_state_monitor.h @@ -110,8 +110,9 @@ struct t7xx_fsm_command { struct list_head entry; enum t7xx_fsm_cmd_state cmd_id; unsigned int flag; - struct completion *done; - int *ret; + struct completion done; + int result; + struct kref refcnt; }; struct t7xx_fsm_notifier { From afc6717628f959941d7b33728570568b4af1c4b8 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 31 Dec 2024 00:06:46 -0500 Subject: [PATCH 142/196] tracing: Have process_string() also allow arrays In order to catch a common bug where a TRACE_EVENT() TP_fast_assign() assigns an address of an allocated string to the ring buffer and then references it in TP_printk(), which can be executed hours later when the string is free, the function test_event_printk() runs on all events as they are registered to make sure there's no unwanted dereferencing. It calls process_string() to handle cases in TP_printk() format that has "%s". It returns whether or not the string is safe. But it can have some false positives. For instance, xe_bo_move() has: TP_printk("move_lacks_source:%s, migrate object %p [size %zu] from %s to %s device_id:%s", __entry->move_lacks_source ? "yes" : "no", __entry->bo, __entry->size, xe_mem_type_to_name[__entry->old_placement], xe_mem_type_to_name[__entry->new_placement], __get_str(device_id)) Where the "%s" references into xe_mem_type_to_name[]. This is an array of pointers that should be safe for the event to access. Instead of flagging this as a bad reference, if a reference points to an array, where the record field is the index, consider it safe. Link: https://lore.kernel.org/all/9dee19b6185d325d0e6fa5f7cbba81d007d99166.camel@sapience.com/ Cc: stable@vger.kernel.org Cc: Masami Hiramatsu Cc: Mathieu Desnoyers Link: https://lore.kernel.org/20241231000646.324fb5f7@gandalf.local.home Fixes: 65a25d9f7ac02 ("tracing: Add "%s" check in test_event_printk()") Reported-by: Genes Lists Tested-by: Gene C Signed-off-by: Steven Rostedt (Google) --- kernel/trace/trace_events.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 1545cc8b49d0..770e7ed91716 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -364,6 +364,18 @@ static bool process_string(const char *fmt, int len, struct trace_event_call *ca s = r + 1; } while (s < e); + /* + * Check for arrays. If the argument has: foo[REC->val] + * then it is very likely that foo is an array of strings + * that are safe to use. + */ + r = strstr(s, "["); + if (r && r < e) { + r = strstr(r, "REC->"); + if (r && r < e) + return true; + } + /* * If there's any strings in the argument consider this arg OK as it * could be: REC->field ? "foo" : "bar" and we don't want to get into From 7b509910b3ad6d7aacead24c8744de10daf8715d Mon Sep 17 00:00:00 2001 From: Daniel Schaefer Date: Tue, 31 Dec 2024 12:59:58 +0800 Subject: [PATCH 143/196] ALSA hda/realtek: Add quirk for Framework F111:000C Similar to commit eb91c456f371 ("ALSA: hda/realtek: Add Framework Laptop 13 (Intel Core Ultra) to quirks") and previous quirks for Framework systems with Realtek codecs. 000C is a new platform that will also have an ALC285 codec and needs the same quirk. Cc: Jaroslav Kysela Cc: Takashi Iwai Cc: linux@frame.work Cc: Dustin L. Howett Signed-off-by: Daniel Schaefer Cc: Link: https://patch.msgid.link/20241231045958.14545-1-dhs@frame.work Signed-off-by: Takashi Iwai --- sound/pci/hda/patch_realtek.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 61ba5dc35b8b..b74b566f675e 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -11009,6 +11009,7 @@ static const struct hda_quirk alc269_fixup_tbl[] = { SND_PCI_QUIRK(0xf111, 0x0001, "Framework Laptop", ALC295_FIXUP_FRAMEWORK_LAPTOP_MIC_NO_PRESENCE), SND_PCI_QUIRK(0xf111, 0x0006, "Framework Laptop", ALC295_FIXUP_FRAMEWORK_LAPTOP_MIC_NO_PRESENCE), SND_PCI_QUIRK(0xf111, 0x0009, "Framework Laptop", ALC295_FIXUP_FRAMEWORK_LAPTOP_MIC_NO_PRESENCE), + SND_PCI_QUIRK(0xf111, 0x000c, "Framework Laptop", ALC295_FIXUP_FRAMEWORK_LAPTOP_MIC_NO_PRESENCE), #if 0 /* Below is a quirk table taken from the old code. From fb514b31395946022f13a08e06a435f53cf9e8b3 Mon Sep 17 00:00:00 2001 From: Li Zhijian Date: Tue, 31 Dec 2024 09:34:16 +0800 Subject: [PATCH 144/196] RDMA/rtrs: Ensure 'ib_sge list' is accessible Move the declaration of the 'ib_sge list' variable outside the 'always_invalidate' block to ensure it remains accessible for use throughout the function. Previously, 'ib_sge list' was declared within the 'always_invalidate' block, limiting its accessibility, then caused a 'BUG: kernel NULL pointer dereference'[1]. ? __die_body.cold+0x19/0x27 ? page_fault_oops+0x15a/0x2d0 ? search_module_extables+0x19/0x60 ? search_bpf_extables+0x5f/0x80 ? exc_page_fault+0x7e/0x180 ? asm_exc_page_fault+0x26/0x30 ? memcpy_orig+0xd5/0x140 rxe_mr_copy+0x1c3/0x200 [rdma_rxe] ? rxe_pool_get_index+0x4b/0x80 [rdma_rxe] copy_data+0xa5/0x230 [rdma_rxe] rxe_requester+0xd9b/0xf70 [rdma_rxe] ? finish_task_switch.isra.0+0x99/0x2e0 rxe_sender+0x13/0x40 [rdma_rxe] do_task+0x68/0x1e0 [rdma_rxe] process_one_work+0x177/0x330 worker_thread+0x252/0x390 ? __pfx_worker_thread+0x10/0x10 This change ensures the variable is available for subsequent operations that require it. [1] https://lore.kernel.org/linux-rdma/6a1f3e8f-deb0-49f9-bc69-a9b03ecfcda7@fujitsu.com/ Fixes: 9cb837480424 ("RDMA/rtrs: server: main functionality") Signed-off-by: Li Zhijian Link: https://patch.msgid.link/20241231013416.1290920-1-lizhijian@fujitsu.com Signed-off-by: Leon Romanovsky --- drivers/infiniband/ulp/rtrs/rtrs-srv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/infiniband/ulp/rtrs/rtrs-srv.c b/drivers/infiniband/ulp/rtrs/rtrs-srv.c index e83d95647852..ef4abdea3c2d 100644 --- a/drivers/infiniband/ulp/rtrs/rtrs-srv.c +++ b/drivers/infiniband/ulp/rtrs/rtrs-srv.c @@ -349,6 +349,7 @@ static int send_io_resp_imm(struct rtrs_srv_con *con, struct rtrs_srv_op *id, struct rtrs_srv_mr *srv_mr; bool need_inval = false; enum ib_send_flags flags; + struct ib_sge list; u32 imm; int err; @@ -401,7 +402,6 @@ static int send_io_resp_imm(struct rtrs_srv_con *con, struct rtrs_srv_op *id, imm = rtrs_to_io_rsp_imm(id->msg_id, errno, need_inval); imm_wr.wr.next = NULL; if (always_invalidate) { - struct ib_sge list; struct rtrs_msg_rkey_rsp *msg; srv_mr = &srv_path->mrs[id->msg_id]; From e6178bf78d0378c2d397a6aafaf4882d0af643fa Mon Sep 17 00:00:00 2001 From: Kalesh AP Date: Tue, 31 Dec 2024 08:20:08 +0530 Subject: [PATCH 145/196] RDMA/bnxt_re: Fix error recovery sequence Fixed to return ENXIO from __send_message_basic_sanity() to indicate that device is in error state. In the case of ERR_DEVICE_DETACHED state, the driver should not post the commands to the firmware as it will time out eventually. Removed bnxt_re_modify_qp() call from bnxt_re_dev_stop() as it is a no-op. Fixes: cc5b9b48d447 ("RDMA/bnxt_re: Recover the device when FW error is detected") Signed-off-by: Kalesh AP Signed-off-by: Kashyap Desai Link: https://patch.msgid.link/20241231025008.2267162-1-kalesh-anakkur.purayil@broadcom.com Reviewed-by: Selvin Xavier Signed-off-by: Leon Romanovsky --- drivers/infiniband/hw/bnxt_re/main.c | 8 +------- drivers/infiniband/hw/bnxt_re/qplib_rcfw.c | 5 +++-- 2 files changed, 4 insertions(+), 9 deletions(-) diff --git a/drivers/infiniband/hw/bnxt_re/main.c b/drivers/infiniband/hw/bnxt_re/main.c index b7af0d5ff3b6..c143f273b759 100644 --- a/drivers/infiniband/hw/bnxt_re/main.c +++ b/drivers/infiniband/hw/bnxt_re/main.c @@ -1715,11 +1715,8 @@ static bool bnxt_re_is_qp1_or_shadow_qp(struct bnxt_re_dev *rdev, static void bnxt_re_dev_stop(struct bnxt_re_dev *rdev) { - int mask = IB_QP_STATE; - struct ib_qp_attr qp_attr; struct bnxt_re_qp *qp; - qp_attr.qp_state = IB_QPS_ERR; mutex_lock(&rdev->qp_lock); list_for_each_entry(qp, &rdev->qp_list, list) { /* Modify the state of all QPs except QP1/Shadow QP */ @@ -1727,12 +1724,9 @@ static void bnxt_re_dev_stop(struct bnxt_re_dev *rdev) if (qp->qplib_qp.state != CMDQ_MODIFY_QP_NEW_STATE_RESET && qp->qplib_qp.state != - CMDQ_MODIFY_QP_NEW_STATE_ERR) { + CMDQ_MODIFY_QP_NEW_STATE_ERR) bnxt_re_dispatch_event(&rdev->ibdev, &qp->ib_qp, 1, IB_EVENT_QP_FATAL); - bnxt_re_modify_qp(&qp->ib_qp, &qp_attr, mask, - NULL); - } } } mutex_unlock(&rdev->qp_lock); diff --git a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c index 5e90ea232de8..17e62f22683b 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_rcfw.c @@ -424,7 +424,8 @@ static int __send_message_basic_sanity(struct bnxt_qplib_rcfw *rcfw, /* Prevent posting if f/w is not in a state to process */ if (test_bit(ERR_DEVICE_DETACHED, &rcfw->cmdq.flags)) - return bnxt_qplib_map_rc(opcode); + return -ENXIO; + if (test_bit(FIRMWARE_STALL_DETECTED, &cmdq->flags)) return -ETIMEDOUT; @@ -493,7 +494,7 @@ static int __bnxt_qplib_rcfw_send_message(struct bnxt_qplib_rcfw *rcfw, rc = __send_message_basic_sanity(rcfw, msg, opcode); if (rc) - return rc; + return rc == -ENXIO ? bnxt_qplib_map_rc(opcode) : rc; rc = __send_message(rcfw, msg, opcode); if (rc) From 8765429279e7d3d68d39ace5f84af2815174bb1e Mon Sep 17 00:00:00 2001 From: Takashi Iwai Date: Tue, 31 Dec 2024 15:53:58 +0100 Subject: [PATCH 146/196] ALSA: seq: Check UMP support for midi_version change When the kernel is built without UMP support but a user-space app requires the midi_version > 0, the kernel should return an error. Otherwise user-space assumes as if it were possible to deal, eventually hitting serious errors later. Fixes: 46397622a3fa ("ALSA: seq: Add UMP support") Cc: Link: https://patch.msgid.link/20241231145358.21946-1-tiwai@suse.de Signed-off-by: Takashi Iwai --- sound/core/seq/seq_clientmgr.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/sound/core/seq/seq_clientmgr.c b/sound/core/seq/seq_clientmgr.c index 3930e2f9082f..77b6ac9b5c11 100644 --- a/sound/core/seq/seq_clientmgr.c +++ b/sound/core/seq/seq_clientmgr.c @@ -1275,10 +1275,16 @@ static int snd_seq_ioctl_set_client_info(struct snd_seq_client *client, if (client->type != client_info->type) return -EINVAL; - /* check validity of midi_version field */ - if (client->user_pversion >= SNDRV_PROTOCOL_VERSION(1, 0, 3) && - client_info->midi_version > SNDRV_SEQ_CLIENT_UMP_MIDI_2_0) - return -EINVAL; + if (client->user_pversion >= SNDRV_PROTOCOL_VERSION(1, 0, 3)) { + /* check validity of midi_version field */ + if (client_info->midi_version > SNDRV_SEQ_CLIENT_UMP_MIDI_2_0) + return -EINVAL; + + /* check if UMP is supported in kernel */ + if (!IS_ENABLED(CONFIG_SND_SEQ_UMP) && + client_info->midi_version > 0) + return -EINVAL; + } /* fill the info fields */ if (client_info->name[0]) From f0ed39830e6064d62f9c5393505677a26569bb56 Mon Sep 17 00:00:00 2001 From: Umesh Nerlige Ramappa Date: Fri, 20 Dec 2024 09:19:18 -0800 Subject: [PATCH 147/196] xe/oa: Fix query mode of operation for OAR/OAC MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is a set of squashed commits to facilitate smooth applying to stable. Each commit message is retained for reference. 1) Allow a GGTT mapped batch to be submitted to user exec queue For a OA use case, one of the HW registers needs to be modified by submitting an MI_LOAD_REGISTER_IMM command to the users exec queue, so that the register is modified in the user's hardware context. In order to do this a batch that is mapped in GGTT, needs to be submitted to the user exec queue. Since all user submissions use q->vm and hence PPGTT, add some plumbing to enable submission of batches mapped in GGTT. v2: ggtt is zero-initialized, so no need to set it false (Matt Brost) 2) xe/oa: Use MI_LOAD_REGISTER_IMMEDIATE to enable OAR/OAC To enable OAR/OAC, a bit in RING_CONTEXT_CONTROL needs to be set. Setting this bit cause the context image size to change and if not done correct, can cause undesired hangs. Current code uses a separate exec_queue to modify this bit and is error-prone. As per HW recommendation, submit MI_LOAD_REGISTER_IMM to the target hardware context to modify the relevant bit. In v2 version, an attempt to submit everything to the user-queue was made, but it failed the unprivileged-single-ctx-counters test. It appears that the OACTXCONTROL must be modified from a remote context. In v3 version, all context specific register configurations were moved to use LOAD_REGISTER_IMMEDIATE and that seems to work well. This is a cleaner way, since we can now submit all configuration to user exec_queue and the fence handling is simplified. v2: (Matt) - set job->ggtt to true if create job is successful - unlock vm on job error (Ashutosh) - don't wait on job submission - use kernel exec queue where possible v3: (Ashutosh) - Fix checkpatch issues - Remove extra spaces/new-lines - Add Fixes: and Cc: tags - Reset context control bit when OA stream is closed - Submit all config via MI_LOAD_REGISTER_IMMEDIATE (Umesh) - Update commit message for v3 experiment - Squash patches for easier port to stable v4: (Ashutosh) - No need to pass q to xe_oa_submit_bb - Do not support exec queues with width > 1 - Fix disabling of CTX_CTRL_OAC_CONTEXT_ENABLE v5: (Ashutosh) - Drop reg_lri related comments - Use XE_OA_SUBMIT_NO_DEPS in xe_oa_load_with_lri Fixes: 8135f1c09dd2 ("drm/xe/oa: Don't reset OAC_CONTEXT_ENABLE on OA stream close") Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Matthew Brost # commit 1 Reviewed-by: Ashutosh Dixit Cc: stable@vger.kernel.org Reviewed-by: Jonathan Cavitt Signed-off-by: Ashutosh Dixit Link: https://patchwork.freedesktop.org/patch/msgid/20241220171919.571528-2-umesh.nerlige.ramappa@intel.com (cherry picked from commit 55039832f98c7e05f1cf9e0d8c12b2490abd0f16) Signed-off-by: Thomas Hellström --- drivers/gpu/drm/xe/xe_oa.c | 134 ++++++++---------------- drivers/gpu/drm/xe/xe_ring_ops.c | 5 +- drivers/gpu/drm/xe/xe_sched_job_types.h | 2 + 3 files changed, 51 insertions(+), 90 deletions(-) diff --git a/drivers/gpu/drm/xe/xe_oa.c b/drivers/gpu/drm/xe/xe_oa.c index 8dd55798ab31..5cc0f6f9bc11 100644 --- a/drivers/gpu/drm/xe/xe_oa.c +++ b/drivers/gpu/drm/xe/xe_oa.c @@ -74,12 +74,6 @@ struct xe_oa_config { struct rcu_head rcu; }; -struct flex { - struct xe_reg reg; - u32 offset; - u32 value; -}; - struct xe_oa_open_param { struct xe_file *xef; u32 oa_unit_id; @@ -596,19 +590,38 @@ static __poll_t xe_oa_poll(struct file *file, poll_table *wait) return ret; } +static void xe_oa_lock_vma(struct xe_exec_queue *q) +{ + if (q->vm) { + down_read(&q->vm->lock); + xe_vm_lock(q->vm, false); + } +} + +static void xe_oa_unlock_vma(struct xe_exec_queue *q) +{ + if (q->vm) { + xe_vm_unlock(q->vm); + up_read(&q->vm->lock); + } +} + static struct dma_fence *xe_oa_submit_bb(struct xe_oa_stream *stream, enum xe_oa_submit_deps deps, struct xe_bb *bb) { + struct xe_exec_queue *q = stream->exec_q ?: stream->k_exec_q; struct xe_sched_job *job; struct dma_fence *fence; int err = 0; - /* Kernel configuration is issued on stream->k_exec_q, not stream->exec_q */ - job = xe_bb_create_job(stream->k_exec_q, bb); + xe_oa_lock_vma(q); + + job = xe_bb_create_job(q, bb); if (IS_ERR(job)) { err = PTR_ERR(job); goto exit; } + job->ggtt = true; if (deps == XE_OA_SUBMIT_ADD_DEPS) { for (int i = 0; i < stream->num_syncs && !err; i++) @@ -623,10 +636,13 @@ static struct dma_fence *xe_oa_submit_bb(struct xe_oa_stream *stream, enum xe_oa fence = dma_fence_get(&job->drm.s_fence->finished); xe_sched_job_push(job); + xe_oa_unlock_vma(q); + return fence; err_put_job: xe_sched_job_put(job); exit: + xe_oa_unlock_vma(q); return ERR_PTR(err); } @@ -675,63 +691,19 @@ static void xe_oa_free_configs(struct xe_oa_stream *stream) dma_fence_put(stream->last_fence); } -static void xe_oa_store_flex(struct xe_oa_stream *stream, struct xe_lrc *lrc, - struct xe_bb *bb, const struct flex *flex, u32 count) -{ - u32 offset = xe_bo_ggtt_addr(lrc->bo); - - do { - bb->cs[bb->len++] = MI_STORE_DATA_IMM | MI_SDI_GGTT | MI_SDI_NUM_DW(1); - bb->cs[bb->len++] = offset + flex->offset * sizeof(u32); - bb->cs[bb->len++] = 0; - bb->cs[bb->len++] = flex->value; - - } while (flex++, --count); -} - -static int xe_oa_modify_ctx_image(struct xe_oa_stream *stream, struct xe_lrc *lrc, - const struct flex *flex, u32 count) +static int xe_oa_load_with_lri(struct xe_oa_stream *stream, struct xe_oa_reg *reg_lri, u32 count) { struct dma_fence *fence; struct xe_bb *bb; int err; - bb = xe_bb_new(stream->gt, 4 * count, false); + bb = xe_bb_new(stream->gt, 2 * count + 1, false); if (IS_ERR(bb)) { err = PTR_ERR(bb); goto exit; } - xe_oa_store_flex(stream, lrc, bb, flex, count); - - fence = xe_oa_submit_bb(stream, XE_OA_SUBMIT_NO_DEPS, bb); - if (IS_ERR(fence)) { - err = PTR_ERR(fence); - goto free_bb; - } - xe_bb_free(bb, fence); - dma_fence_put(fence); - - return 0; -free_bb: - xe_bb_free(bb, NULL); -exit: - return err; -} - -static int xe_oa_load_with_lri(struct xe_oa_stream *stream, struct xe_oa_reg *reg_lri) -{ - struct dma_fence *fence; - struct xe_bb *bb; - int err; - - bb = xe_bb_new(stream->gt, 3, false); - if (IS_ERR(bb)) { - err = PTR_ERR(bb); - goto exit; - } - - write_cs_mi_lri(bb, reg_lri, 1); + write_cs_mi_lri(bb, reg_lri, count); fence = xe_oa_submit_bb(stream, XE_OA_SUBMIT_NO_DEPS, bb); if (IS_ERR(fence)) { @@ -751,71 +723,55 @@ static int xe_oa_load_with_lri(struct xe_oa_stream *stream, struct xe_oa_reg *re static int xe_oa_configure_oar_context(struct xe_oa_stream *stream, bool enable) { const struct xe_oa_format *format = stream->oa_buffer.format; - struct xe_lrc *lrc = stream->exec_q->lrc[0]; - u32 regs_offset = xe_lrc_regs_offset(lrc) / sizeof(u32); u32 oacontrol = __format_to_oactrl(format, OAR_OACONTROL_COUNTER_SEL_MASK) | (enable ? OAR_OACONTROL_COUNTER_ENABLE : 0); - struct flex regs_context[] = { + struct xe_oa_reg reg_lri[] = { { OACTXCONTROL(stream->hwe->mmio_base), - stream->oa->ctx_oactxctrl_offset[stream->hwe->class] + 1, enable ? OA_COUNTER_RESUME : 0, }, + { + OAR_OACONTROL, + oacontrol, + }, { RING_CONTEXT_CONTROL(stream->hwe->mmio_base), - regs_offset + CTX_CONTEXT_CONTROL, - _MASKED_BIT_ENABLE(CTX_CTRL_OAC_CONTEXT_ENABLE), + _MASKED_FIELD(CTX_CTRL_OAC_CONTEXT_ENABLE, + enable ? CTX_CTRL_OAC_CONTEXT_ENABLE : 0) }, }; - struct xe_oa_reg reg_lri = { OAR_OACONTROL, oacontrol }; - int err; - /* Modify stream hwe context image with regs_context */ - err = xe_oa_modify_ctx_image(stream, stream->exec_q->lrc[0], - regs_context, ARRAY_SIZE(regs_context)); - if (err) - return err; - - /* Apply reg_lri using LRI */ - return xe_oa_load_with_lri(stream, ®_lri); + return xe_oa_load_with_lri(stream, reg_lri, ARRAY_SIZE(reg_lri)); } static int xe_oa_configure_oac_context(struct xe_oa_stream *stream, bool enable) { const struct xe_oa_format *format = stream->oa_buffer.format; - struct xe_lrc *lrc = stream->exec_q->lrc[0]; - u32 regs_offset = xe_lrc_regs_offset(lrc) / sizeof(u32); u32 oacontrol = __format_to_oactrl(format, OAR_OACONTROL_COUNTER_SEL_MASK) | (enable ? OAR_OACONTROL_COUNTER_ENABLE : 0); - struct flex regs_context[] = { + struct xe_oa_reg reg_lri[] = { { OACTXCONTROL(stream->hwe->mmio_base), - stream->oa->ctx_oactxctrl_offset[stream->hwe->class] + 1, enable ? OA_COUNTER_RESUME : 0, }, + { + OAC_OACONTROL, + oacontrol + }, { RING_CONTEXT_CONTROL(stream->hwe->mmio_base), - regs_offset + CTX_CONTEXT_CONTROL, - _MASKED_BIT_ENABLE(CTX_CTRL_OAC_CONTEXT_ENABLE) | + _MASKED_FIELD(CTX_CTRL_OAC_CONTEXT_ENABLE, + enable ? CTX_CTRL_OAC_CONTEXT_ENABLE : 0) | _MASKED_FIELD(CTX_CTRL_RUN_ALONE, enable ? CTX_CTRL_RUN_ALONE : 0), }, }; - struct xe_oa_reg reg_lri = { OAC_OACONTROL, oacontrol }; - int err; /* Set ccs select to enable programming of OAC_OACONTROL */ xe_mmio_write32(&stream->gt->mmio, __oa_regs(stream)->oa_ctrl, __oa_ccs_select(stream)); - /* Modify stream hwe context image with regs_context */ - err = xe_oa_modify_ctx_image(stream, stream->exec_q->lrc[0], - regs_context, ARRAY_SIZE(regs_context)); - if (err) - return err; - - /* Apply reg_lri using LRI */ - return xe_oa_load_with_lri(stream, ®_lri); + return xe_oa_load_with_lri(stream, reg_lri, ARRAY_SIZE(reg_lri)); } static int xe_oa_configure_oa_context(struct xe_oa_stream *stream, bool enable) @@ -2066,8 +2022,8 @@ int xe_oa_stream_open_ioctl(struct drm_device *dev, u64 data, struct drm_file *f if (XE_IOCTL_DBG(oa->xe, !param.exec_q)) return -ENOENT; - if (param.exec_q->width > 1) - drm_dbg(&oa->xe->drm, "exec_q->width > 1, programming only exec_q->lrc[0]\n"); + if (XE_IOCTL_DBG(oa->xe, param.exec_q->width > 1)) + return -EOPNOTSUPP; } /* diff --git a/drivers/gpu/drm/xe/xe_ring_ops.c b/drivers/gpu/drm/xe/xe_ring_ops.c index 0be4f489d3e1..9f327f27c072 100644 --- a/drivers/gpu/drm/xe/xe_ring_ops.c +++ b/drivers/gpu/drm/xe/xe_ring_ops.c @@ -221,7 +221,10 @@ static int emit_pipe_imm_ggtt(u32 addr, u32 value, bool stall_only, u32 *dw, static u32 get_ppgtt_flag(struct xe_sched_job *job) { - return job->q->vm ? BIT(8) : 0; + if (job->q->vm && !job->ggtt) + return BIT(8); + + return 0; } static int emit_copy_timestamp(struct xe_lrc *lrc, u32 *dw, int i) diff --git a/drivers/gpu/drm/xe/xe_sched_job_types.h b/drivers/gpu/drm/xe/xe_sched_job_types.h index f13f333f00be..d942b20a9f29 100644 --- a/drivers/gpu/drm/xe/xe_sched_job_types.h +++ b/drivers/gpu/drm/xe/xe_sched_job_types.h @@ -56,6 +56,8 @@ struct xe_sched_job { u32 migrate_flush_flags; /** @ring_ops_flush_tlb: The ring ops need to flush TLB before payload. */ bool ring_ops_flush_tlb; + /** @ggtt: mapped in ggtt. */ + bool ggtt; /** @ptrs: per instance pointers. */ struct xe_job_ptrs ptrs[]; }; From 0bc21e701a6ffacfdde7f04f87d664d82e8a13bf Mon Sep 17 00:00:00 2001 From: Olof Johansson Date: Thu, 2 Jan 2025 06:30:03 -0800 Subject: [PATCH 148/196] MAINTAINERS: Remove Olof from SoC maintainers I haven't been an active participant for a couple of years now, and after discussions at Linux Plumbers in 2024, Arnd is getting fresh help from a few more participants. It's time to remove myself, and spare myself from patches and pull requests in my inbox. Signed-off-by: Olof Johansson Cc: Arnd Bergmann Signed-off-by: Linus Torvalds --- MAINTAINERS | 1 - 1 file changed, 1 deletion(-) diff --git a/MAINTAINERS b/MAINTAINERS index 910305c11e8a..c575de4903db 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -1797,7 +1797,6 @@ F: include/uapi/linux/if_arcnet.h ARM AND ARM64 SoC SUB-ARCHITECTURES (COMMON PARTS) M: Arnd Bergmann -M: Olof Johansson L: linux-arm-kernel@lists.infradead.org (moderated for non-subscribers) L: soc@lists.linux.dev S: Maintained From d65474033740ded0a4fe9a097fce72328655b41d Mon Sep 17 00:00:00 2001 From: Zilin Guan Date: Tue, 31 Dec 2024 11:37:31 +0000 Subject: [PATCH 149/196] fgraph: Add READ_ONCE() when accessing fgraph_array[] In __ftrace_return_to_handler(), a loop iterates over the fgraph_array[] elements, which are fgraph_ops. The loop checks if an element is a fgraph_stub to prevent using a fgraph_stub afterward. However, if the compiler reloads fgraph_array[] after this check, it might race with an update to fgraph_array[] that introduces a fgraph_stub. This could result in the stub being processed, but the stub contains a null "func_hash" field, leading to a NULL pointer dereference. To ensure that the gops compared against the fgraph_stub matches the gops processed later, add a READ_ONCE(). A similar patch appears in commit 63a8dfb ("function_graph: Add READ_ONCE() when accessing fgraph_array[]"). Cc: stable@vger.kernel.org Fixes: 37238abe3cb47 ("ftrace/function_graph: Pass fgraph_ops to function graph callbacks") Link: https://lore.kernel.org/20241231113731.277668-1-zilin@seu.edu.cn Signed-off-by: Zilin Guan Signed-off-by: Steven Rostedt (Google) --- kernel/trace/fgraph.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/trace/fgraph.c b/kernel/trace/fgraph.c index ddedcb50917f..30e3ddc8a8a8 100644 --- a/kernel/trace/fgraph.c +++ b/kernel/trace/fgraph.c @@ -833,7 +833,7 @@ static unsigned long __ftrace_return_to_handler(struct fgraph_ret_regs *ret_regs #endif { for_each_set_bit(i, &bitmap, sizeof(bitmap) * BITS_PER_BYTE) { - struct fgraph_ops *gops = fgraph_array[i]; + struct fgraph_ops *gops = READ_ONCE(fgraph_array[i]); if (gops == &fgraph_stub) continue; From 789a8cff8d2dbe4b5c617c3004b5eb63fa7a3b35 Mon Sep 17 00:00:00 2001 From: Kohei Enju Date: Thu, 2 Jan 2025 04:08:20 +0900 Subject: [PATCH 150/196] ftrace: Fix function profiler's filtering functionality Commit c132be2c4fcc ("function_graph: Have the instances use their own ftrace_ops for filtering"), function profiler (enabled via function_profile_enabled) has been showing statistics for all functions, ignoring set_ftrace_filter settings. While tracers are instantiated, the function profiler is not. Therefore, it should use the global set_ftrace_filter for consistency. This patch modifies the function profiler to use the global filter, fixing the filtering functionality. Before (filtering not working): ``` root@localhost:~# echo 'vfs*' > /sys/kernel/tracing/set_ftrace_filter root@localhost:~# echo 1 > /sys/kernel/tracing/function_profile_enabled root@localhost:~# sleep 1 root@localhost:~# echo 0 > /sys/kernel/tracing/function_profile_enabled root@localhost:~# head /sys/kernel/tracing/trace_stat/* Function Hit Time Avg s^2 -------- --- ---- --- --- schedule 314 22290594 us 70989.15 us 40372231 us x64_sys_call 1527 8762510 us 5738.382 us 3414354 us schedule_hrtimeout_range 176 8665356 us 49234.98 us 405618876 us __x64_sys_ppoll 324 5656635 us 17458.75 us 19203976 us do_sys_poll 324 5653747 us 17449.83 us 19214945 us schedule_timeout 67 5531396 us 82558.15 us 2136740827 us __x64_sys_pselect6 12 3029540 us 252461.7 us 63296940171 us do_pselect.constprop.0 12 3029532 us 252461.0 us 63296952931 us ``` After (filtering working): ``` root@localhost:~# echo 'vfs*' > /sys/kernel/tracing/set_ftrace_filter root@localhost:~# echo 1 > /sys/kernel/tracing/function_profile_enabled root@localhost:~# sleep 1 root@localhost:~# echo 0 > /sys/kernel/tracing/function_profile_enabled root@localhost:~# head /sys/kernel/tracing/trace_stat/* Function Hit Time Avg s^2 -------- --- ---- --- --- vfs_write 462 68476.43 us 148.217 us 25874.48 us vfs_read 641 9611.356 us 14.994 us 28868.07 us vfs_fstat 890 878.094 us 0.986 us 1.667 us vfs_fstatat 227 757.176 us 3.335 us 18.928 us vfs_statx 226 610.610 us 2.701 us 17.749 us vfs_getattr_nosec 1187 460.919 us 0.388 us 0.326 us vfs_statx_path 297 343.287 us 1.155 us 11.116 us vfs_rename 6 291.575 us 48.595 us 9889.236 us ``` Cc: stable@vger.kernel.org Link: https://lore.kernel.org/20250101190820.72534-1-enjuk@amazon.com Fixes: c132be2c4fcc ("function_graph: Have the instances use their own ftrace_ops for filtering") Signed-off-by: Kohei Enju Signed-off-by: Steven Rostedt (Google) --- kernel/trace/ftrace.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index 9b17efb1a87d..2e113f8b13a2 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -902,16 +902,13 @@ static void profile_graph_return(struct ftrace_graph_ret *trace, } static struct fgraph_ops fprofiler_ops = { - .ops = { - .flags = FTRACE_OPS_FL_INITIALIZED, - INIT_OPS_HASH(fprofiler_ops.ops) - }, .entryfunc = &profile_graph_entry, .retfunc = &profile_graph_return, }; static int register_ftrace_profiler(void) { + ftrace_ops_set_global_filter(&fprofiler_ops.ops); return register_ftrace_graph(&fprofiler_ops); } @@ -922,12 +919,11 @@ static void unregister_ftrace_profiler(void) #else static struct ftrace_ops ftrace_profile_ops __read_mostly = { .func = function_profile_call, - .flags = FTRACE_OPS_FL_INITIALIZED, - INIT_OPS_HASH(ftrace_profile_ops) }; static int register_ftrace_profiler(void) { + ftrace_ops_set_global_filter(&ftrace_profile_ops); return register_ftrace_function(&ftrace_profile_ops); } From c6e60a0a68b7e6b3c7e33863a16e8e88ba9eee6f Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Thu, 2 Jan 2025 16:32:51 -0700 Subject: [PATCH 151/196] io_uring/net: always initialize kmsg->msg.msg_inq upfront syzbot reports that ->msg_inq may get used uinitialized from the following path: BUG: KMSAN: uninit-value in io_recv_buf_select io_uring/net.c:1094 [inline] BUG: KMSAN: uninit-value in io_recv+0x930/0x1f90 io_uring/net.c:1158 io_recv_buf_select io_uring/net.c:1094 [inline] io_recv+0x930/0x1f90 io_uring/net.c:1158 io_issue_sqe+0x420/0x2130 io_uring/io_uring.c:1740 io_queue_sqe io_uring/io_uring.c:1950 [inline] io_req_task_submit+0xfa/0x1d0 io_uring/io_uring.c:1374 io_handle_tw_list+0x55f/0x5c0 io_uring/io_uring.c:1057 tctx_task_work_run+0x109/0x3e0 io_uring/io_uring.c:1121 tctx_task_work+0x6d/0xc0 io_uring/io_uring.c:1139 task_work_run+0x268/0x310 kernel/task_work.c:239 io_run_task_work+0x43a/0x4a0 io_uring/io_uring.h:343 io_cqring_wait io_uring/io_uring.c:2527 [inline] __do_sys_io_uring_enter io_uring/io_uring.c:3439 [inline] __se_sys_io_uring_enter+0x204f/0x4ce0 io_uring/io_uring.c:3330 __x64_sys_io_uring_enter+0x11f/0x1a0 io_uring/io_uring.c:3330 x64_sys_call+0xce5/0x3c30 arch/x86/include/generated/asm/syscalls_64.h:427 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xcd/0x1e0 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f and it is correct, as it's never initialized upfront. Hence the first submission can end up using it uninitialized, if the recv wasn't successful and the networking stack didn't honor ->msg_get_inq being set and filling in the output value of ->msg_inq as requested. Set it to 0 upfront when it's allocated, just to silence this KMSAN warning. There's no side effect of using it uninitialized, it'll just potentially cause the next receive to use a recv value hint that's not accurate. Fixes: c6f32c7d9e09 ("io_uring/net: get rid of ->prep_async() for receive side") Reported-by: syzbot+068ff190354d2f74892f@syzkaller.appspotmail.com Signed-off-by: Jens Axboe --- io_uring/net.c | 1 + 1 file changed, 1 insertion(+) diff --git a/io_uring/net.c b/io_uring/net.c index df1f7dc6f1c8..c6cd38cc5dc4 100644 --- a/io_uring/net.c +++ b/io_uring/net.c @@ -754,6 +754,7 @@ static int io_recvmsg_prep_setup(struct io_kiocb *req) if (req->opcode == IORING_OP_RECV) { kmsg->msg.msg_name = NULL; kmsg->msg.msg_namelen = 0; + kmsg->msg.msg_inq = 0; kmsg->msg.msg_control = NULL; kmsg->msg.msg_get_inq = 1; kmsg->msg.msg_controllen = 0; From a8620de72e5676993ec3a3b975f7c10908f5f60f Mon Sep 17 00:00:00 2001 From: Liang Jie Date: Mon, 30 Dec 2024 17:37:09 +0800 Subject: [PATCH 152/196] net: sfc: Correct key_len for efx_tc_ct_zone_ht_params In efx_tc_ct_zone_ht_params, the key_len was previously set to offsetof(struct efx_tc_ct_zone, linkage). This calculation is incorrect because it includes any padding between the zone field and the linkage field due to structure alignment, which can vary between systems. This patch updates key_len to use sizeof_field(struct efx_tc_ct_zone, zone) , ensuring that the hash table correctly uses the zone as the key. This fix prevents potential hash lookup errors and improves connection tracking reliability. Fixes: c3bb5c6acd4e ("sfc: functions to register for conntrack zone offload") Signed-off-by: Liang Jie Acked-by: Edward Cree Link: https://patch.msgid.link/20241230093709.3226854-1-buaajxlj@163.com Signed-off-by: Jakub Kicinski --- drivers/net/ethernet/sfc/tc_conntrack.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/ethernet/sfc/tc_conntrack.c b/drivers/net/ethernet/sfc/tc_conntrack.c index d90206f27161..c0603f54cec3 100644 --- a/drivers/net/ethernet/sfc/tc_conntrack.c +++ b/drivers/net/ethernet/sfc/tc_conntrack.c @@ -16,7 +16,7 @@ static int efx_tc_flow_block(enum tc_setup_type type, void *type_data, void *cb_priv); static const struct rhashtable_params efx_tc_ct_zone_ht_params = { - .key_len = offsetof(struct efx_tc_ct_zone, linkage), + .key_len = sizeof_field(struct efx_tc_ct_zone, zone), .key_offset = 0, .head_offset = offsetof(struct efx_tc_ct_zone, linkage), }; From 68e068cabd2c6c533ef934c2e5151609cf6ecc6d Mon Sep 17 00:00:00 2001 From: Willem de Bruijn Date: Wed, 1 Jan 2025 11:47:40 -0500 Subject: [PATCH 153/196] net: reenable NETIF_F_IPV6_CSUM offload for BIG TCP packets The blamed commit disabled hardware offoad of IPv6 packets with extension headers on devices that advertise NETIF_F_IPV6_CSUM, based on the definition of that feature in skbuff.h: * * - %NETIF_F_IPV6_CSUM * - Driver (device) is only able to checksum plain * TCP or UDP packets over IPv6. These are specifically * unencapsulated packets of the form IPv6|TCP or * IPv6|UDP where the Next Header field in the IPv6 * header is either TCP or UDP. IPv6 extension headers * are not supported with this feature. This feature * cannot be set in features for a device with * NETIF_F_HW_CSUM also set. This feature is being * DEPRECATED (see below). The change causes skb_warn_bad_offload to fire for BIG TCP packets. [ 496.310233] WARNING: CPU: 13 PID: 23472 at net/core/dev.c:3129 skb_warn_bad_offload+0xc4/0xe0 [ 496.310297] ? skb_warn_bad_offload+0xc4/0xe0 [ 496.310300] skb_checksum_help+0x129/0x1f0 [ 496.310303] skb_csum_hwoffload_help+0x150/0x1b0 [ 496.310306] validate_xmit_skb+0x159/0x270 [ 496.310309] validate_xmit_skb_list+0x41/0x70 [ 496.310312] sch_direct_xmit+0x5c/0x250 [ 496.310317] __qdisc_run+0x388/0x620 BIG TCP introduced an IPV6_TLV_JUMBO IPv6 extension header to communicate packet length, as this is an IPv6 jumbogram. But, the feature is only enabled on devices that support BIG TCP TSO. The header is only present for PF_PACKET taps like tcpdump, and not transmitted by physical devices. For this specific case of extension headers that are not transmitted, return to the situation before the blamed commit and support hardware offload. ipv6_has_hopopt_jumbo() tests not only whether this header is present, but also that it is the only extension header before a terminal (L4) header. Fixes: 04c20a9356f2 ("net: skip offload for NETIF_F_IPV6_CSUM if ipv6 header contains extension") Reported-by: syzbot Reported-by: Eric Dumazet Closes: https://lore.kernel.org/netdev/CANn89iK1hdC3Nt8KPhOtTF8vCPc1AHDCtse_BTNki1pWxAByTQ@mail.gmail.com/ Signed-off-by: Willem de Bruijn Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/20250101164909.1331680-1-willemdebruijn.kernel@gmail.com Signed-off-by: Jakub Kicinski --- net/core/dev.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/net/core/dev.c b/net/core/dev.c index 45a8c3dd4a64..faa23042df38 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3642,8 +3642,10 @@ int skb_csum_hwoffload_help(struct sk_buff *skb, if (features & (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM)) { if (vlan_get_protocol(skb) == htons(ETH_P_IPV6) && - skb_network_header_len(skb) != sizeof(struct ipv6hdr)) + skb_network_header_len(skb) != sizeof(struct ipv6hdr) && + !ipv6_has_hopopt_jumbo(skb)) goto sw_checksum; + switch (skb->csum_offset) { case offsetof(struct tcphdr, check): case offsetof(struct udphdr, check): From 5b0af621c3f6ef9261cf6067812f2fd9943acb4b Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Tue, 31 Dec 2024 16:05:27 +0000 Subject: [PATCH 154/196] net: restrict SO_REUSEPORT to inet sockets After blamed commit, crypto sockets could accidentally be destroyed from RCU call back, as spotted by zyzbot [1]. Trying to acquire a mutex in RCU callback is not allowed. Restrict SO_REUSEPORT socket option to inet sockets. v1 of this patch supported TCP, UDP and SCTP sockets, but fcnal-test.sh test needed RAW and ICMP support. [1] BUG: sleeping function called from invalid context at kernel/locking/mutex.c:562 in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 24, name: ksoftirqd/1 preempt_count: 100, expected: 0 RCU nest depth: 0, expected: 0 1 lock held by ksoftirqd/1/24: #0: ffffffff8e937ba0 (rcu_callback){....}-{0:0}, at: rcu_lock_acquire include/linux/rcupdate.h:337 [inline] #0: ffffffff8e937ba0 (rcu_callback){....}-{0:0}, at: rcu_do_batch kernel/rcu/tree.c:2561 [inline] #0: ffffffff8e937ba0 (rcu_callback){....}-{0:0}, at: rcu_core+0xa37/0x17a0 kernel/rcu/tree.c:2823 Preemption disabled at: [] softirq_handle_begin kernel/softirq.c:402 [inline] [] handle_softirqs+0x128/0x9b0 kernel/softirq.c:537 CPU: 1 UID: 0 PID: 24 Comm: ksoftirqd/1 Not tainted 6.13.0-rc3-syzkaller-00174-ga024e377efed #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/13/2024 Call Trace: __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x241/0x360 lib/dump_stack.c:120 __might_resched+0x5d4/0x780 kernel/sched/core.c:8758 __mutex_lock_common kernel/locking/mutex.c:562 [inline] __mutex_lock+0x131/0xee0 kernel/locking/mutex.c:735 crypto_put_default_null_skcipher+0x18/0x70 crypto/crypto_null.c:179 aead_release+0x3d/0x50 crypto/algif_aead.c:489 alg_do_release crypto/af_alg.c:118 [inline] alg_sock_destruct+0x86/0xc0 crypto/af_alg.c:502 __sk_destruct+0x58/0x5f0 net/core/sock.c:2260 rcu_do_batch kernel/rcu/tree.c:2567 [inline] rcu_core+0xaaa/0x17a0 kernel/rcu/tree.c:2823 handle_softirqs+0x2d4/0x9b0 kernel/softirq.c:561 run_ksoftirqd+0xca/0x130 kernel/softirq.c:950 smpboot_thread_fn+0x544/0xa30 kernel/smpboot.c:164 kthread+0x2f0/0x390 kernel/kthread.c:389 ret_from_fork+0x4b/0x80 arch/x86/kernel/process.c:147 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244 Fixes: 8c7138b33e5c ("net: Unpublish sk from sk_reuseport_cb before call_rcu") Reported-by: syzbot+b3e02953598f447d4d2a@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6772f2f4.050a0220.2f3838.04cb.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Cc: Martin KaFai Lau Reviewed-by: Kuniyuki Iwashima Link: https://patch.msgid.link/20241231160527.3994168-1-edumazet@google.com Signed-off-by: Jakub Kicinski --- net/core/sock.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/net/core/sock.c b/net/core/sock.c index 74729d20cd00..be84885f9290 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -1295,7 +1295,10 @@ int sk_setsockopt(struct sock *sk, int level, int optname, sk->sk_reuse = (valbool ? SK_CAN_REUSE : SK_NO_REUSE); break; case SO_REUSEPORT: - sk->sk_reuseport = valbool; + if (valbool && !sk_is_inet(sk)) + ret = -EOPNOTSUPP; + else + sk->sk_reuseport = valbool; break; case SO_DONTROUTE: sock_valbool_flag(sk, SOCK_LOCALROUTE, valbool); From a7af435df0e04cfb4a4004136d597c42639a2ae7 Mon Sep 17 00:00:00 2001 From: "Maciej S. Szmigiero" Date: Sun, 29 Dec 2024 17:46:58 +0100 Subject: [PATCH 155/196] net: wwan: iosm: Properly check for valid exec stage in ipc_mmio_init() ipc_mmio_init() used the post-decrement operator in its loop continuing condition of "retries" counter being "> 0", which meant that when this condition caused loop exit "retries" counter reached -1. But the later valid exec stage failure check only tests for "retries" counter being exactly zero, so it didn't trigger in this case (but would wrongly trigger if the code reaches a valid exec stage in the very last loop iteration). Fix this by using the pre-decrement operator instead, so the loop counter is exactly zero on valid exec stage failure. Fixes: dc0514f5d828 ("net: iosm: mmio scratchpad") Signed-off-by: Maciej S. Szmigiero Link: https://patch.msgid.link/8b19125a825f9dcdd81c667c1e5c48ba28d505a6.1735490770.git.mail@maciej.szmigiero.name Signed-off-by: Jakub Kicinski --- drivers/net/wwan/iosm/iosm_ipc_mmio.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/wwan/iosm/iosm_ipc_mmio.c b/drivers/net/wwan/iosm/iosm_ipc_mmio.c index 63eb08c43c05..6764c13530b9 100644 --- a/drivers/net/wwan/iosm/iosm_ipc_mmio.c +++ b/drivers/net/wwan/iosm/iosm_ipc_mmio.c @@ -104,7 +104,7 @@ struct iosm_mmio *ipc_mmio_init(void __iomem *mmio, struct device *dev) break; msleep(20); - } while (retries-- > 0); + } while (--retries > 0); if (!retries) { dev_err(ipc_mmio->dev, "invalid exec stage %X", stage); From 77ee7a6d16b6ec07b5c3ae2b6b60a24c1afbed09 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 30 Dec 2024 16:10:03 +0000 Subject: [PATCH 156/196] af_packet: fix vlan_get_tci() vs MSG_PEEK Blamed commit forgot MSG_PEEK case, allowing a crash [1] as found by syzbot. Rework vlan_get_tci() to not touch skb at all, so that it can be used from many cpus on the same skb. Add a const qualifier to skb argument. [1] skbuff: skb_under_panic: text:ffffffff8a8da482 len:32 put:14 head:ffff88807a1d5800 data:ffff88807a1d5810 tail:0x14 end:0x140 dev: ------------[ cut here ]------------ kernel BUG at net/core/skbuff.c:206 ! Oops: invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI CPU: 0 UID: 0 PID: 5880 Comm: syz-executor172 Not tainted 6.13.0-rc3-syzkaller-00762-g9268abe611b0 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/13/2024 RIP: 0010:skb_panic net/core/skbuff.c:206 [inline] RIP: 0010:skb_under_panic+0x14b/0x150 net/core/skbuff.c:216 Code: 0b 8d 48 c7 c6 9e 6c 26 8e 48 8b 54 24 08 8b 0c 24 44 8b 44 24 04 4d 89 e9 50 41 54 41 57 41 56 e8 3a 5a 79 f7 48 83 c4 20 90 <0f> 0b 0f 1f 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 RSP: 0018:ffffc90003baf5b8 EFLAGS: 00010286 RAX: 0000000000000087 RBX: dffffc0000000000 RCX: 8565c1eec37aa000 RDX: 0000000000000000 RSI: 0000000080000000 RDI: 0000000000000000 RBP: ffff88802616fb50 R08: ffffffff817f0a4c R09: 1ffff92000775e50 R10: dffffc0000000000 R11: fffff52000775e51 R12: 0000000000000140 R13: ffff88807a1d5800 R14: ffff88807a1d5810 R15: 0000000000000014 FS: 00007fa03261f6c0(0000) GS:ffff8880b8600000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007ffd65753000 CR3: 0000000031720000 CR4: 00000000003526f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: skb_push+0xe5/0x100 net/core/skbuff.c:2636 vlan_get_tci+0x272/0x550 net/packet/af_packet.c:565 packet_recvmsg+0x13c9/0x1ef0 net/packet/af_packet.c:3616 sock_recvmsg_nosec net/socket.c:1044 [inline] sock_recvmsg+0x22f/0x280 net/socket.c:1066 ____sys_recvmsg+0x1c6/0x480 net/socket.c:2814 ___sys_recvmsg net/socket.c:2856 [inline] do_recvmmsg+0x426/0xab0 net/socket.c:2951 __sys_recvmmsg net/socket.c:3025 [inline] __do_sys_recvmmsg net/socket.c:3048 [inline] __se_sys_recvmmsg net/socket.c:3041 [inline] __x64_sys_recvmmsg+0x199/0x250 net/socket.c:3041 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83 Fixes: 79eecf631c14 ("af_packet: Handle outgoing VLAN packets without hardware offloading") Reported-by: syzbot+8400677f3fd43f37d3bc@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6772c485.050a0220.2f3838.04c6.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Cc: Chengen Du Reviewed-by: Willem de Bruijn Link: https://patch.msgid.link/20241230161004.2681892-1-edumazet@google.com Signed-off-by: Jakub Kicinski --- net/packet/af_packet.c | 12 +++--------- 1 file changed, 3 insertions(+), 9 deletions(-) diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index 886c0dd47b66..e2e34a49e98d 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -538,10 +538,8 @@ static void *packet_current_frame(struct packet_sock *po, return packet_lookup_frame(po, rb, rb->head, status); } -static u16 vlan_get_tci(struct sk_buff *skb, struct net_device *dev) +static u16 vlan_get_tci(const struct sk_buff *skb, struct net_device *dev) { - u8 *skb_orig_data = skb->data; - int skb_orig_len = skb->len; struct vlan_hdr vhdr, *vh; unsigned int header_len; @@ -562,12 +560,8 @@ static u16 vlan_get_tci(struct sk_buff *skb, struct net_device *dev) else return 0; - skb_push(skb, skb->data - skb_mac_header(skb)); - vh = skb_header_pointer(skb, header_len, sizeof(vhdr), &vhdr); - if (skb_orig_data != skb->data) { - skb->data = skb_orig_data; - skb->len = skb_orig_len; - } + vh = skb_header_pointer(skb, skb_mac_offset(skb) + header_len, + sizeof(vhdr), &vhdr); if (unlikely(!vh)) return 0; From f91a5b8089389eb408501af2762f168c3aaa7b79 Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 30 Dec 2024 16:10:04 +0000 Subject: [PATCH 157/196] af_packet: fix vlan_get_protocol_dgram() vs MSG_PEEK Blamed commit forgot MSG_PEEK case, allowing a crash [1] as found by syzbot. Rework vlan_get_protocol_dgram() to not touch skb at all, so that it can be used from many cpus on the same skb. Add a const qualifier to skb argument. [1] skbuff: skb_under_panic: text:ffffffff8a8ccd05 len:29 put:14 head:ffff88807fc8e400 data:ffff88807fc8e3f4 tail:0x11 end:0x140 dev: ------------[ cut here ]------------ kernel BUG at net/core/skbuff.c:206 ! Oops: invalid opcode: 0000 [#1] PREEMPT SMP KASAN PTI CPU: 1 UID: 0 PID: 5892 Comm: syz-executor883 Not tainted 6.13.0-rc4-syzkaller-00054-gd6ef8b40d075 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/13/2024 RIP: 0010:skb_panic net/core/skbuff.c:206 [inline] RIP: 0010:skb_under_panic+0x14b/0x150 net/core/skbuff.c:216 Code: 0b 8d 48 c7 c6 86 d5 25 8e 48 8b 54 24 08 8b 0c 24 44 8b 44 24 04 4d 89 e9 50 41 54 41 57 41 56 e8 5a 69 79 f7 48 83 c4 20 90 <0f> 0b 0f 1f 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 RSP: 0018:ffffc900038d7638 EFLAGS: 00010282 RAX: 0000000000000087 RBX: dffffc0000000000 RCX: 609ffd18ea660600 RDX: 0000000000000000 RSI: 0000000080000000 RDI: 0000000000000000 RBP: ffff88802483c8d0 R08: ffffffff817f0a8c R09: 1ffff9200071ae60 R10: dffffc0000000000 R11: fffff5200071ae61 R12: 0000000000000140 R13: ffff88807fc8e400 R14: ffff88807fc8e3f4 R15: 0000000000000011 FS: 00007fbac5e006c0(0000) GS:ffff8880b8700000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007fbac5e00d58 CR3: 000000001238e000 CR4: 00000000003526f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: skb_push+0xe5/0x100 net/core/skbuff.c:2636 vlan_get_protocol_dgram+0x165/0x290 net/packet/af_packet.c:585 packet_recvmsg+0x948/0x1ef0 net/packet/af_packet.c:3552 sock_recvmsg_nosec net/socket.c:1033 [inline] sock_recvmsg+0x22f/0x280 net/socket.c:1055 ____sys_recvmsg+0x1c6/0x480 net/socket.c:2803 ___sys_recvmsg net/socket.c:2845 [inline] do_recvmmsg+0x426/0xab0 net/socket.c:2940 __sys_recvmmsg net/socket.c:3014 [inline] __do_sys_recvmmsg net/socket.c:3037 [inline] __se_sys_recvmmsg net/socket.c:3030 [inline] __x64_sys_recvmmsg+0x199/0x250 net/socket.c:3030 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f Fixes: 79eecf631c14 ("af_packet: Handle outgoing VLAN packets without hardware offloading") Reported-by: syzbot+74f70bb1cb968bf09e4f@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6772c485.050a0220.2f3838.04c5.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Cc: Chengen Du Reviewed-by: Willem de Bruijn Link: https://patch.msgid.link/20241230161004.2681892-2-edumazet@google.com Signed-off-by: Jakub Kicinski --- include/linux/if_vlan.h | 16 +++++++++++++--- net/packet/af_packet.c | 16 ++++------------ 2 files changed, 17 insertions(+), 15 deletions(-) diff --git a/include/linux/if_vlan.h b/include/linux/if_vlan.h index c1645c86eed9..d65b5d71b93b 100644 --- a/include/linux/if_vlan.h +++ b/include/linux/if_vlan.h @@ -585,13 +585,16 @@ static inline int vlan_get_tag(const struct sk_buff *skb, u16 *vlan_tci) * vlan_get_protocol - get protocol EtherType. * @skb: skbuff to query * @type: first vlan protocol + * @mac_offset: MAC offset * @depth: buffer to store length of eth and vlan tags in bytes * * Returns the EtherType of the packet, regardless of whether it is * vlan encapsulated (normal or hardware accelerated) or not. */ -static inline __be16 __vlan_get_protocol(const struct sk_buff *skb, __be16 type, - int *depth) +static inline __be16 __vlan_get_protocol_offset(const struct sk_buff *skb, + __be16 type, + int mac_offset, + int *depth) { unsigned int vlan_depth = skb->mac_len, parse_depth = VLAN_MAX_DEPTH; @@ -610,7 +613,8 @@ static inline __be16 __vlan_get_protocol(const struct sk_buff *skb, __be16 type, do { struct vlan_hdr vhdr, *vh; - vh = skb_header_pointer(skb, vlan_depth, sizeof(vhdr), &vhdr); + vh = skb_header_pointer(skb, mac_offset + vlan_depth, + sizeof(vhdr), &vhdr); if (unlikely(!vh || !--parse_depth)) return 0; @@ -625,6 +629,12 @@ static inline __be16 __vlan_get_protocol(const struct sk_buff *skb, __be16 type, return type; } +static inline __be16 __vlan_get_protocol(const struct sk_buff *skb, __be16 type, + int *depth) +{ + return __vlan_get_protocol_offset(skb, type, 0, depth); +} + /** * vlan_get_protocol - get protocol EtherType. * @skb: skbuff to query diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index e2e34a49e98d..2d73769d67f4 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -568,21 +568,13 @@ static u16 vlan_get_tci(const struct sk_buff *skb, struct net_device *dev) return ntohs(vh->h_vlan_TCI); } -static __be16 vlan_get_protocol_dgram(struct sk_buff *skb) +static __be16 vlan_get_protocol_dgram(const struct sk_buff *skb) { __be16 proto = skb->protocol; - if (unlikely(eth_type_vlan(proto))) { - u8 *skb_orig_data = skb->data; - int skb_orig_len = skb->len; - - skb_push(skb, skb->data - skb_mac_header(skb)); - proto = __vlan_get_protocol(skb, proto, NULL); - if (skb_orig_data != skb->data) { - skb->data = skb_orig_data; - skb->len = skb_orig_len; - } - } + if (unlikely(eth_type_vlan(proto))) + proto = __vlan_get_protocol_offset(skb, proto, + skb_mac_offset(skb), NULL); return proto; } From 260466b576bca0081a7d4acecc8e93687aa22d0e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Mon, 30 Dec 2024 16:28:49 +0000 Subject: [PATCH 158/196] ila: serialize calls to nf_register_net_hooks() syzbot found a race in ila_add_mapping() [1] commit 031ae72825ce ("ila: call nf_unregister_net_hooks() sooner") attempted to fix a similar issue. Looking at the syzbot repro, we have concurrent ILA_CMD_ADD commands. Add a mutex to make sure at most one thread is calling nf_register_net_hooks(). [1] BUG: KASAN: slab-use-after-free in rht_key_hashfn include/linux/rhashtable.h:159 [inline] BUG: KASAN: slab-use-after-free in __rhashtable_lookup.constprop.0+0x426/0x550 include/linux/rhashtable.h:604 Read of size 4 at addr ffff888028f40008 by task dhcpcd/5501 CPU: 1 UID: 0 PID: 5501 Comm: dhcpcd Not tainted 6.13.0-rc4-syzkaller-00054-gd6ef8b40d075 #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 09/13/2024 Call Trace: __dump_stack lib/dump_stack.c:94 [inline] dump_stack_lvl+0x116/0x1f0 lib/dump_stack.c:120 print_address_description mm/kasan/report.c:378 [inline] print_report+0xc3/0x620 mm/kasan/report.c:489 kasan_report+0xd9/0x110 mm/kasan/report.c:602 rht_key_hashfn include/linux/rhashtable.h:159 [inline] __rhashtable_lookup.constprop.0+0x426/0x550 include/linux/rhashtable.h:604 rhashtable_lookup include/linux/rhashtable.h:646 [inline] rhashtable_lookup_fast include/linux/rhashtable.h:672 [inline] ila_lookup_wildcards net/ipv6/ila/ila_xlat.c:127 [inline] ila_xlat_addr net/ipv6/ila/ila_xlat.c:652 [inline] ila_nf_input+0x1ee/0x620 net/ipv6/ila/ila_xlat.c:185 nf_hook_entry_hookfn include/linux/netfilter.h:154 [inline] nf_hook_slow+0xbb/0x200 net/netfilter/core.c:626 nf_hook.constprop.0+0x42e/0x750 include/linux/netfilter.h:269 NF_HOOK include/linux/netfilter.h:312 [inline] ipv6_rcv+0xa4/0x680 net/ipv6/ip6_input.c:309 __netif_receive_skb_one_core+0x12e/0x1e0 net/core/dev.c:5672 __netif_receive_skb+0x1d/0x160 net/core/dev.c:5785 process_backlog+0x443/0x15f0 net/core/dev.c:6117 __napi_poll.constprop.0+0xb7/0x550 net/core/dev.c:6883 napi_poll net/core/dev.c:6952 [inline] net_rx_action+0xa94/0x1010 net/core/dev.c:7074 handle_softirqs+0x213/0x8f0 kernel/softirq.c:561 __do_softirq kernel/softirq.c:595 [inline] invoke_softirq kernel/softirq.c:435 [inline] __irq_exit_rcu+0x109/0x170 kernel/softirq.c:662 irq_exit_rcu+0x9/0x30 kernel/softirq.c:678 instr_sysvec_apic_timer_interrupt arch/x86/kernel/apic/apic.c:1049 [inline] sysvec_apic_timer_interrupt+0xa4/0xc0 arch/x86/kernel/apic/apic.c:1049 Fixes: 7f00feaf1076 ("ila: Add generic ILA translation facility") Reported-by: syzbot+47e761d22ecf745f72b9@syzkaller.appspotmail.com Closes: https://lore.kernel.org/netdev/6772c9ae.050a0220.2f3838.04c7.GAE@google.com/T/#u Signed-off-by: Eric Dumazet Cc: Florian Westphal Cc: Tom Herbert Link: https://patch.msgid.link/20241230162849.2795486-1-edumazet@google.com Signed-off-by: Jakub Kicinski --- net/ipv6/ila/ila_xlat.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) diff --git a/net/ipv6/ila/ila_xlat.c b/net/ipv6/ila/ila_xlat.c index 7646e401c630..1d41b2ab4884 100644 --- a/net/ipv6/ila/ila_xlat.c +++ b/net/ipv6/ila/ila_xlat.c @@ -195,6 +195,8 @@ static const struct nf_hook_ops ila_nf_hook_ops[] = { }, }; +static DEFINE_MUTEX(ila_mutex); + static int ila_add_mapping(struct net *net, struct ila_xlat_params *xp) { struct ila_net *ilan = net_generic(net, ila_net_id); @@ -202,16 +204,20 @@ static int ila_add_mapping(struct net *net, struct ila_xlat_params *xp) spinlock_t *lock = ila_get_lock(ilan, xp->ip.locator_match); int err = 0, order; - if (!ilan->xlat.hooks_registered) { + if (!READ_ONCE(ilan->xlat.hooks_registered)) { /* We defer registering net hooks in the namespace until the * first mapping is added. */ - err = nf_register_net_hooks(net, ila_nf_hook_ops, - ARRAY_SIZE(ila_nf_hook_ops)); + mutex_lock(&ila_mutex); + if (!ilan->xlat.hooks_registered) { + err = nf_register_net_hooks(net, ila_nf_hook_ops, + ARRAY_SIZE(ila_nf_hook_ops)); + if (!err) + WRITE_ONCE(ilan->xlat.hooks_registered, true); + } + mutex_unlock(&ila_mutex); if (err) return err; - - ilan->xlat.hooks_registered = true; } ila = kzalloc(sizeof(*ila), GFP_KERNEL); From 449e6912a2522af672e99992e1201a454910864e Mon Sep 17 00:00:00 2001 From: Paolo Abeni Date: Mon, 30 Dec 2024 19:12:30 +0100 Subject: [PATCH 159/196] mptcp: fix recvbuffer adjust on sleeping rcvmsg If the recvmsg() blocks after receiving some data - i.e. due to SO_RCVLOWAT - the MPTCP code will attempt multiple times to adjust the receive buffer size, wrongly accounting every time the cumulative of received data - instead of accounting only for the delta. Address the issue moving mptcp_rcv_space_adjust just after the data reception and passing it only the just received bytes. This also removes an unneeded difference between the TCP and MPTCP RX code path implementation. Fixes: 581302298524 ("mptcp: error out earlier on disconnect") Cc: stable@vger.kernel.org Signed-off-by: Paolo Abeni Reviewed-by: Mat Martineau Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241230-net-mptcp-rbuf-fixes-v1-1-8608af434ceb@kernel.org Signed-off-by: Jakub Kicinski --- net/mptcp/protocol.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 08a72242428c..27afdb7e2071 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -1939,6 +1939,8 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len) goto out; } +static void mptcp_rcv_space_adjust(struct mptcp_sock *msk, int copied); + static int __mptcp_recvmsg_mskq(struct mptcp_sock *msk, struct msghdr *msg, size_t len, int flags, @@ -1992,6 +1994,7 @@ static int __mptcp_recvmsg_mskq(struct mptcp_sock *msk, break; } + mptcp_rcv_space_adjust(msk, copied); return copied; } @@ -2268,7 +2271,6 @@ static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, } pr_debug("block timeout %ld\n", timeo); - mptcp_rcv_space_adjust(msk, copied); err = sk_wait_data(sk, &timeo, NULL); if (err < 0) { err = copied ? : err; @@ -2276,8 +2278,6 @@ static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, } } - mptcp_rcv_space_adjust(msk, copied); - out_err: if (cmsg_flags && copied >= 0) { if (cmsg_flags & MPTCP_CMSG_TS) From 551844f26da2a9f76c0a698baaffa631d1178645 Mon Sep 17 00:00:00 2001 From: Paolo Abeni Date: Mon, 30 Dec 2024 19:12:31 +0100 Subject: [PATCH 160/196] mptcp: don't always assume copied data in mptcp_cleanup_rbuf() Under some corner cases the MPTCP protocol can end-up invoking mptcp_cleanup_rbuf() when no data has been copied, but such helper assumes the opposite condition. Explicitly drop such assumption and performs the costly call only when strictly needed - before releasing the msk socket lock. Fixes: fd8976790a6c ("mptcp: be careful on MPTCP-level ack.") Cc: stable@vger.kernel.org Signed-off-by: Paolo Abeni Reviewed-by: Mat Martineau Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241230-net-mptcp-rbuf-fixes-v1-2-8608af434ceb@kernel.org Signed-off-by: Jakub Kicinski --- net/mptcp/protocol.c | 18 +++++++++--------- 1 file changed, 9 insertions(+), 9 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 27afdb7e2071..5307fff9d995 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -528,13 +528,13 @@ static void mptcp_send_ack(struct mptcp_sock *msk) mptcp_subflow_send_ack(mptcp_subflow_tcp_sock(subflow)); } -static void mptcp_subflow_cleanup_rbuf(struct sock *ssk) +static void mptcp_subflow_cleanup_rbuf(struct sock *ssk, int copied) { bool slow; slow = lock_sock_fast(ssk); if (tcp_can_send_ack(ssk)) - tcp_cleanup_rbuf(ssk, 1); + tcp_cleanup_rbuf(ssk, copied); unlock_sock_fast(ssk, slow); } @@ -551,7 +551,7 @@ static bool mptcp_subflow_could_cleanup(const struct sock *ssk, bool rx_empty) (ICSK_ACK_PUSHED2 | ICSK_ACK_PUSHED))); } -static void mptcp_cleanup_rbuf(struct mptcp_sock *msk) +static void mptcp_cleanup_rbuf(struct mptcp_sock *msk, int copied) { int old_space = READ_ONCE(msk->old_wspace); struct mptcp_subflow_context *subflow; @@ -559,14 +559,14 @@ static void mptcp_cleanup_rbuf(struct mptcp_sock *msk) int space = __mptcp_space(sk); bool cleanup, rx_empty; - cleanup = (space > 0) && (space >= (old_space << 1)); - rx_empty = !__mptcp_rmem(sk); + cleanup = (space > 0) && (space >= (old_space << 1)) && copied; + rx_empty = !__mptcp_rmem(sk) && copied; mptcp_for_each_subflow(msk, subflow) { struct sock *ssk = mptcp_subflow_tcp_sock(subflow); if (cleanup || mptcp_subflow_could_cleanup(ssk, rx_empty)) - mptcp_subflow_cleanup_rbuf(ssk); + mptcp_subflow_cleanup_rbuf(ssk, copied); } } @@ -2220,9 +2220,6 @@ static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, copied += bytes_read; - /* be sure to advertise window change */ - mptcp_cleanup_rbuf(msk); - if (skb_queue_empty(&msk->receive_queue) && __mptcp_move_skbs(msk)) continue; @@ -2271,6 +2268,7 @@ static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, } pr_debug("block timeout %ld\n", timeo); + mptcp_cleanup_rbuf(msk, copied); err = sk_wait_data(sk, &timeo, NULL); if (err < 0) { err = copied ? : err; @@ -2278,6 +2276,8 @@ static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, } } + mptcp_cleanup_rbuf(msk, copied); + out_err: if (cmsg_flags && copied >= 0) { if (cmsg_flags & MPTCP_CMSG_TS) From 56b824eb49d6258aa0bad09a406ceac3f643cdae Mon Sep 17 00:00:00 2001 From: Paolo Abeni Date: Mon, 30 Dec 2024 19:12:32 +0100 Subject: [PATCH 161/196] mptcp: prevent excessive coalescing on receive Currently the skb size after coalescing is only limited by the skb layout (the skb must not carry frag_list). A single coalesced skb covering several MSS can potentially fill completely the receive buffer. In such a case, the snd win will zero until the receive buffer will be empty again, affecting tput badly. Fixes: 8268ed4c9d19 ("mptcp: introduce and use mptcp_try_coalesce()") Cc: stable@vger.kernel.org # please delay 2 weeks after 6.13-final release Signed-off-by: Paolo Abeni Reviewed-by: Mat Martineau Signed-off-by: Matthieu Baerts (NGI0) Link: https://patch.msgid.link/20241230-net-mptcp-rbuf-fixes-v1-3-8608af434ceb@kernel.org Signed-off-by: Jakub Kicinski --- net/mptcp/protocol.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 5307fff9d995..1b2e7cbb577f 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -136,6 +136,7 @@ static bool mptcp_try_coalesce(struct sock *sk, struct sk_buff *to, int delta; if (MPTCP_SKB_CB(from)->offset || + ((to->len + from->len) > (sk->sk_rcvbuf >> 3)) || !skb_try_coalesce(to, from, &fragstolen, &delta)) return false; From 9facce84f4062f782ebde18daa7006a23d40b607 Mon Sep 17 00:00:00 2001 From: MD Danish Anwar Date: Mon, 23 Dec 2024 20:45:49 +0530 Subject: [PATCH 162/196] net: ti: icssg-prueth: Fix firmware load sequence. Timesync related operations are ran in PRU0 cores for both ICSSG SLICE0 and SLICE1. Currently whenever any ICSSG interface comes up we load the respective firmwares to PRU cores and whenever interface goes down, we stop the resective cores. Due to this, when SLICE0 goes down while SLICE1 is still active, PRU0 firmwares are unloaded and PRU0 core is stopped. This results in clock jump for SLICE1 interface as the timesync related operations are no longer running. As there are interdependencies between SLICE0 and SLICE1 firmwares, fix this by running both PRU0 and PRU1 firmwares as long as at least 1 ICSSG interface is up. Add new flag in prueth struct to check if all firmwares are running and remove the old flag (fw_running). Use emacs_initialized as reference count to load the firmwares for the first and last interface up/down. Moving init_emac_mode and fw_offload_mode API outside of icssg_config to icssg_common_start API as they need to be called only once per firmware boot. Change prueth_emac_restart() to return error code and add error prints inside the caller of this functions in case of any failures. Move prueth_emac_stop() from common to sr1 driver. sr1 and sr2 drivers have different logic handling for stopping the firmwares. While sr1 driver is dependent on emac structure to stop the corresponding pru cores for that slice, for sr2 all the pru cores of both the slices are stopped and is not dependent on emac. So the prueth_emac_stop() function is no longer common and can be moved to sr1 driver. Fixes: c1e0230eeaab ("net: ti: icss-iep: Add IEP driver") Signed-off-by: MD Danish Anwar Signed-off-by: Meghana Malladi Signed-off-by: David S. Miller --- drivers/net/ethernet/ti/icssg/icssg_common.c | 25 -- drivers/net/ethernet/ti/icssg/icssg_config.c | 41 ++- drivers/net/ethernet/ti/icssg/icssg_config.h | 1 + drivers/net/ethernet/ti/icssg/icssg_prueth.c | 281 ++++++++++++------ drivers/net/ethernet/ti/icssg/icssg_prueth.h | 5 +- .../net/ethernet/ti/icssg/icssg_prueth_sr1.c | 24 +- 6 files changed, 246 insertions(+), 131 deletions(-) diff --git a/drivers/net/ethernet/ti/icssg/icssg_common.c b/drivers/net/ethernet/ti/icssg/icssg_common.c index fdebeb2f84e0..74f0f200a89d 100644 --- a/drivers/net/ethernet/ti/icssg/icssg_common.c +++ b/drivers/net/ethernet/ti/icssg/icssg_common.c @@ -855,31 +855,6 @@ irqreturn_t prueth_rx_irq(int irq, void *dev_id) } EXPORT_SYMBOL_GPL(prueth_rx_irq); -void prueth_emac_stop(struct prueth_emac *emac) -{ - struct prueth *prueth = emac->prueth; - int slice; - - switch (emac->port_id) { - case PRUETH_PORT_MII0: - slice = ICSS_SLICE0; - break; - case PRUETH_PORT_MII1: - slice = ICSS_SLICE1; - break; - default: - netdev_err(emac->ndev, "invalid port\n"); - return; - } - - emac->fw_running = 0; - if (!emac->is_sr1) - rproc_shutdown(prueth->txpru[slice]); - rproc_shutdown(prueth->rtu[slice]); - rproc_shutdown(prueth->pru[slice]); -} -EXPORT_SYMBOL_GPL(prueth_emac_stop); - void prueth_cleanup_tx_ts(struct prueth_emac *emac) { int i; diff --git a/drivers/net/ethernet/ti/icssg/icssg_config.c b/drivers/net/ethernet/ti/icssg/icssg_config.c index 5d2491c2943a..ddfd1c02a885 100644 --- a/drivers/net/ethernet/ti/icssg/icssg_config.c +++ b/drivers/net/ethernet/ti/icssg/icssg_config.c @@ -397,7 +397,7 @@ static int prueth_emac_buffer_setup(struct prueth_emac *emac) return 0; } -static void icssg_init_emac_mode(struct prueth *prueth) +void icssg_init_emac_mode(struct prueth *prueth) { /* When the device is configured as a bridge and it is being brought * back to the emac mode, the host mac address has to be set as 0. @@ -406,9 +406,6 @@ static void icssg_init_emac_mode(struct prueth *prueth) int i; u8 mac[ETH_ALEN] = { 0 }; - if (prueth->emacs_initialized) - return; - /* Set VLAN TABLE address base */ regmap_update_bits(prueth->miig_rt, FDB_GEN_CFG1, SMEM_VLAN_OFFSET_MASK, addr << SMEM_VLAN_OFFSET); @@ -423,15 +420,13 @@ static void icssg_init_emac_mode(struct prueth *prueth) /* Clear host MAC address */ icssg_class_set_host_mac_addr(prueth->miig_rt, mac); } +EXPORT_SYMBOL_GPL(icssg_init_emac_mode); -static void icssg_init_fw_offload_mode(struct prueth *prueth) +void icssg_init_fw_offload_mode(struct prueth *prueth) { u32 addr = prueth->shram.pa + EMAC_ICSSG_SWITCH_DEFAULT_VLAN_TABLE_OFFSET; int i; - if (prueth->emacs_initialized) - return; - /* Set VLAN TABLE address base */ regmap_update_bits(prueth->miig_rt, FDB_GEN_CFG1, SMEM_VLAN_OFFSET_MASK, addr << SMEM_VLAN_OFFSET); @@ -448,6 +443,7 @@ static void icssg_init_fw_offload_mode(struct prueth *prueth) icssg_class_set_host_mac_addr(prueth->miig_rt, prueth->hw_bridge_dev->dev_addr); icssg_set_pvid(prueth, prueth->default_vlan, PRUETH_PORT_HOST); } +EXPORT_SYMBOL_GPL(icssg_init_fw_offload_mode); int icssg_config(struct prueth *prueth, struct prueth_emac *emac, int slice) { @@ -455,11 +451,6 @@ int icssg_config(struct prueth *prueth, struct prueth_emac *emac, int slice) struct icssg_flow_cfg __iomem *flow_cfg; int ret; - if (prueth->is_switch_mode || prueth->is_hsr_offload_mode) - icssg_init_fw_offload_mode(prueth); - else - icssg_init_emac_mode(prueth); - memset_io(config, 0, TAS_GATE_MASK_LIST0); icssg_miig_queues_init(prueth, slice); @@ -786,3 +777,27 @@ void icssg_set_pvid(struct prueth *prueth, u8 vid, u8 port) writel(pvid, prueth->shram.va + EMAC_ICSSG_SWITCH_PORT0_DEFAULT_VLAN_OFFSET); } EXPORT_SYMBOL_GPL(icssg_set_pvid); + +int emac_fdb_flow_id_updated(struct prueth_emac *emac) +{ + struct mgmt_cmd_rsp fdb_cmd_rsp = { 0 }; + int slice = prueth_emac_slice(emac); + struct mgmt_cmd fdb_cmd = { 0 }; + int ret; + + fdb_cmd.header = ICSSG_FW_MGMT_CMD_HEADER; + fdb_cmd.type = ICSSG_FW_MGMT_FDB_CMD_TYPE_RX_FLOW; + fdb_cmd.seqnum = ++(emac->prueth->icssg_hwcmdseq); + fdb_cmd.param = 0; + + fdb_cmd.param |= (slice << 4); + fdb_cmd.cmd_args[0] = 0; + + ret = icssg_send_fdb_msg(emac, &fdb_cmd, &fdb_cmd_rsp); + if (ret) + return ret; + + WARN_ON(fdb_cmd.seqnum != fdb_cmd_rsp.seqnum); + return fdb_cmd_rsp.status == 1 ? 0 : -EINVAL; +} +EXPORT_SYMBOL_GPL(emac_fdb_flow_id_updated); diff --git a/drivers/net/ethernet/ti/icssg/icssg_config.h b/drivers/net/ethernet/ti/icssg/icssg_config.h index 92c2deaa3068..c884e9fa099e 100644 --- a/drivers/net/ethernet/ti/icssg/icssg_config.h +++ b/drivers/net/ethernet/ti/icssg/icssg_config.h @@ -55,6 +55,7 @@ struct icssg_rxq_ctx { #define ICSSG_FW_MGMT_FDB_CMD_TYPE 0x03 #define ICSSG_FW_MGMT_CMD_TYPE 0x04 #define ICSSG_FW_MGMT_PKT 0x80000000 +#define ICSSG_FW_MGMT_FDB_CMD_TYPE_RX_FLOW 0x05 struct icssg_r30_cmd { u32 cmd[4]; diff --git a/drivers/net/ethernet/ti/icssg/icssg_prueth.c b/drivers/net/ethernet/ti/icssg/icssg_prueth.c index c568c84a032b..d76fe6d05e10 100644 --- a/drivers/net/ethernet/ti/icssg/icssg_prueth.c +++ b/drivers/net/ethernet/ti/icssg/icssg_prueth.c @@ -164,11 +164,26 @@ static struct icssg_firmwares icssg_emac_firmwares[] = { } }; -static int prueth_emac_start(struct prueth *prueth, struct prueth_emac *emac) +static int prueth_start(struct rproc *rproc, const char *fw_name) +{ + int ret; + + ret = rproc_set_firmware(rproc, fw_name); + if (ret) + return ret; + return rproc_boot(rproc); +} + +static void prueth_shutdown(struct rproc *rproc) +{ + rproc_shutdown(rproc); +} + +static int prueth_emac_start(struct prueth *prueth) { struct icssg_firmwares *firmwares; struct device *dev = prueth->dev; - int slice, ret; + int ret, slice; if (prueth->is_switch_mode) firmwares = icssg_switch_firmwares; @@ -177,49 +192,126 @@ static int prueth_emac_start(struct prueth *prueth, struct prueth_emac *emac) else firmwares = icssg_emac_firmwares; - slice = prueth_emac_slice(emac); - if (slice < 0) { - netdev_err(emac->ndev, "invalid port\n"); - return -EINVAL; + for (slice = 0; slice < PRUETH_NUM_MACS; slice++) { + ret = prueth_start(prueth->pru[slice], firmwares[slice].pru); + if (ret) { + dev_err(dev, "failed to boot PRU%d: %d\n", slice, ret); + goto unwind_slices; + } + + ret = prueth_start(prueth->rtu[slice], firmwares[slice].rtu); + if (ret) { + dev_err(dev, "failed to boot RTU%d: %d\n", slice, ret); + rproc_shutdown(prueth->pru[slice]); + goto unwind_slices; + } + + ret = prueth_start(prueth->txpru[slice], firmwares[slice].txpru); + if (ret) { + dev_err(dev, "failed to boot TX_PRU%d: %d\n", slice, ret); + rproc_shutdown(prueth->rtu[slice]); + rproc_shutdown(prueth->pru[slice]); + goto unwind_slices; + } } - ret = icssg_config(prueth, emac, slice); - if (ret) - return ret; - - ret = rproc_set_firmware(prueth->pru[slice], firmwares[slice].pru); - ret = rproc_boot(prueth->pru[slice]); - if (ret) { - dev_err(dev, "failed to boot PRU%d: %d\n", slice, ret); - return -EINVAL; - } - - ret = rproc_set_firmware(prueth->rtu[slice], firmwares[slice].rtu); - ret = rproc_boot(prueth->rtu[slice]); - if (ret) { - dev_err(dev, "failed to boot RTU%d: %d\n", slice, ret); - goto halt_pru; - } - - ret = rproc_set_firmware(prueth->txpru[slice], firmwares[slice].txpru); - ret = rproc_boot(prueth->txpru[slice]); - if (ret) { - dev_err(dev, "failed to boot TX_PRU%d: %d\n", slice, ret); - goto halt_rtu; - } - - emac->fw_running = 1; return 0; -halt_rtu: - rproc_shutdown(prueth->rtu[slice]); - -halt_pru: - rproc_shutdown(prueth->pru[slice]); +unwind_slices: + while (--slice >= 0) { + prueth_shutdown(prueth->txpru[slice]); + prueth_shutdown(prueth->rtu[slice]); + prueth_shutdown(prueth->pru[slice]); + } return ret; } +static void prueth_emac_stop(struct prueth *prueth) +{ + int slice; + + for (slice = 0; slice < PRUETH_NUM_MACS; slice++) { + prueth_shutdown(prueth->txpru[slice]); + prueth_shutdown(prueth->rtu[slice]); + prueth_shutdown(prueth->pru[slice]); + } +} + +static int prueth_emac_common_start(struct prueth *prueth) +{ + struct prueth_emac *emac; + int ret = 0; + int slice; + + if (!prueth->emac[ICSS_SLICE0] && !prueth->emac[ICSS_SLICE1]) + return -EINVAL; + + /* clear SMEM and MSMC settings for all slices */ + memset_io(prueth->msmcram.va, 0, prueth->msmcram.size); + memset_io(prueth->shram.va, 0, ICSSG_CONFIG_OFFSET_SLICE1 * PRUETH_NUM_MACS); + + icssg_class_default(prueth->miig_rt, ICSS_SLICE0, 0, false); + icssg_class_default(prueth->miig_rt, ICSS_SLICE1, 0, false); + + if (prueth->is_switch_mode || prueth->is_hsr_offload_mode) + icssg_init_fw_offload_mode(prueth); + else + icssg_init_emac_mode(prueth); + + for (slice = 0; slice < PRUETH_NUM_MACS; slice++) { + emac = prueth->emac[slice]; + if (!emac) + continue; + ret = icssg_config(prueth, emac, slice); + if (ret) + goto disable_class; + } + + ret = prueth_emac_start(prueth); + if (ret) + goto disable_class; + + emac = prueth->emac[ICSS_SLICE0] ? prueth->emac[ICSS_SLICE0] : + prueth->emac[ICSS_SLICE1]; + ret = icss_iep_init(emac->iep, &prueth_iep_clockops, + emac, IEP_DEFAULT_CYCLE_TIME_NS); + if (ret) { + dev_err(prueth->dev, "Failed to initialize IEP module\n"); + goto stop_pruss; + } + + return 0; + +stop_pruss: + prueth_emac_stop(prueth); + +disable_class: + icssg_class_disable(prueth->miig_rt, ICSS_SLICE0); + icssg_class_disable(prueth->miig_rt, ICSS_SLICE1); + + return ret; +} + +static int prueth_emac_common_stop(struct prueth *prueth) +{ + struct prueth_emac *emac; + + if (!prueth->emac[ICSS_SLICE0] && !prueth->emac[ICSS_SLICE1]) + return -EINVAL; + + icssg_class_disable(prueth->miig_rt, ICSS_SLICE0); + icssg_class_disable(prueth->miig_rt, ICSS_SLICE1); + + prueth_emac_stop(prueth); + + emac = prueth->emac[ICSS_SLICE0] ? prueth->emac[ICSS_SLICE0] : + prueth->emac[ICSS_SLICE1]; + icss_iep_exit(emac->iep); + + return 0; +} + /* called back by PHY layer if there is change in link state of hw port*/ static void emac_adjust_link(struct net_device *ndev) { @@ -374,9 +466,6 @@ static void prueth_iep_settime(void *clockops_data, u64 ns) u32 cycletime; int timeout; - if (!emac->fw_running) - return; - sc_descp = emac->prueth->shram.va + TIMESYNC_FW_WC_SETCLOCK_DESC_OFFSET; cycletime = IEP_DEFAULT_CYCLE_TIME_NS; @@ -543,23 +632,17 @@ static int emac_ndo_open(struct net_device *ndev) { struct prueth_emac *emac = netdev_priv(ndev); int ret, i, num_data_chn = emac->tx_ch_num; + struct icssg_flow_cfg __iomem *flow_cfg; struct prueth *prueth = emac->prueth; int slice = prueth_emac_slice(emac); struct device *dev = prueth->dev; int max_rx_flows; int rx_flow; - /* clear SMEM and MSMC settings for all slices */ - if (!prueth->emacs_initialized) { - memset_io(prueth->msmcram.va, 0, prueth->msmcram.size); - memset_io(prueth->shram.va, 0, ICSSG_CONFIG_OFFSET_SLICE1 * PRUETH_NUM_MACS); - } - /* set h/w MAC as user might have re-configured */ ether_addr_copy(emac->mac_addr, ndev->dev_addr); icssg_class_set_mac_addr(prueth->miig_rt, slice, emac->mac_addr); - icssg_class_default(prueth->miig_rt, slice, 0, false); icssg_ft1_set_mac_addr(prueth->miig_rt, slice, emac->mac_addr); /* Notify the stack of the actual queue counts. */ @@ -597,18 +680,23 @@ static int emac_ndo_open(struct net_device *ndev) goto cleanup_napi; } - /* reset and start PRU firmware */ - ret = prueth_emac_start(prueth, emac); - if (ret) - goto free_rx_irq; + if (!prueth->emacs_initialized) { + ret = prueth_emac_common_start(prueth); + if (ret) + goto free_rx_irq; + } + + flow_cfg = emac->dram.va + ICSSG_CONFIG_OFFSET + PSI_L_REGULAR_FLOW_ID_BASE_OFFSET; + writew(emac->rx_flow_id_base, &flow_cfg->rx_base_flow); + ret = emac_fdb_flow_id_updated(emac); + + if (ret) { + netdev_err(ndev, "Failed to update Rx Flow ID %d", ret); + goto stop; + } icssg_mii_update_mtu(prueth->mii_rt, slice, ndev->max_mtu); - if (!prueth->emacs_initialized) { - ret = icss_iep_init(emac->iep, &prueth_iep_clockops, - emac, IEP_DEFAULT_CYCLE_TIME_NS); - } - ret = request_threaded_irq(emac->tx_ts_irq, NULL, prueth_tx_ts_irq, IRQF_ONESHOT, dev_name(dev), emac); if (ret) @@ -653,7 +741,8 @@ static int emac_ndo_open(struct net_device *ndev) free_tx_ts_irq: free_irq(emac->tx_ts_irq, emac); stop: - prueth_emac_stop(emac); + if (!prueth->emacs_initialized) + prueth_emac_common_stop(prueth); free_rx_irq: free_irq(emac->rx_chns.irq[rx_flow], emac); cleanup_napi: @@ -689,8 +778,6 @@ static int emac_ndo_stop(struct net_device *ndev) if (ndev->phydev) phy_stop(ndev->phydev); - icssg_class_disable(prueth->miig_rt, prueth_emac_slice(emac)); - if (emac->prueth->is_hsr_offload_mode) __dev_mc_unsync(ndev, icssg_prueth_hsr_del_mcast); else @@ -728,11 +815,9 @@ static int emac_ndo_stop(struct net_device *ndev) /* Destroying the queued work in ndo_stop() */ cancel_delayed_work_sync(&emac->stats_work); - if (prueth->emacs_initialized == 1) - icss_iep_exit(emac->iep); - /* stop PRUs */ - prueth_emac_stop(emac); + if (prueth->emacs_initialized == 1) + prueth_emac_common_stop(prueth); free_irq(emac->tx_ts_irq, emac); @@ -1053,10 +1138,11 @@ static void prueth_offload_fwd_mark_update(struct prueth *prueth) } } -static void prueth_emac_restart(struct prueth *prueth) +static int prueth_emac_restart(struct prueth *prueth) { struct prueth_emac *emac0 = prueth->emac[PRUETH_MAC0]; struct prueth_emac *emac1 = prueth->emac[PRUETH_MAC1]; + int ret; /* Detach the net_device for both PRUeth ports*/ if (netif_running(emac0->ndev)) @@ -1065,36 +1151,46 @@ static void prueth_emac_restart(struct prueth *prueth) netif_device_detach(emac1->ndev); /* Disable both PRUeth ports */ - icssg_set_port_state(emac0, ICSSG_EMAC_PORT_DISABLE); - icssg_set_port_state(emac1, ICSSG_EMAC_PORT_DISABLE); + ret = icssg_set_port_state(emac0, ICSSG_EMAC_PORT_DISABLE); + ret |= icssg_set_port_state(emac1, ICSSG_EMAC_PORT_DISABLE); + if (ret) + return ret; /* Stop both pru cores for both PRUeth ports*/ - prueth_emac_stop(emac0); - prueth->emacs_initialized--; - prueth_emac_stop(emac1); - prueth->emacs_initialized--; + ret = prueth_emac_common_stop(prueth); + if (ret) { + dev_err(prueth->dev, "Failed to stop the firmwares"); + return ret; + } /* Start both pru cores for both PRUeth ports */ - prueth_emac_start(prueth, emac0); - prueth->emacs_initialized++; - prueth_emac_start(prueth, emac1); - prueth->emacs_initialized++; + ret = prueth_emac_common_start(prueth); + if (ret) { + dev_err(prueth->dev, "Failed to start the firmwares"); + return ret; + } /* Enable forwarding for both PRUeth ports */ - icssg_set_port_state(emac0, ICSSG_EMAC_PORT_FORWARD); - icssg_set_port_state(emac1, ICSSG_EMAC_PORT_FORWARD); + ret = icssg_set_port_state(emac0, ICSSG_EMAC_PORT_FORWARD); + ret |= icssg_set_port_state(emac1, ICSSG_EMAC_PORT_FORWARD); /* Attache net_device for both PRUeth ports */ netif_device_attach(emac0->ndev); netif_device_attach(emac1->ndev); + + return ret; } static void icssg_change_mode(struct prueth *prueth) { struct prueth_emac *emac; - int mac; + int mac, ret; - prueth_emac_restart(prueth); + ret = prueth_emac_restart(prueth); + if (ret) { + dev_err(prueth->dev, "Failed to restart the firmwares, aborting the process"); + return; + } for (mac = PRUETH_MAC0; mac < PRUETH_NUM_MACS; mac++) { emac = prueth->emac[mac]; @@ -1173,13 +1269,18 @@ static void prueth_netdevice_port_unlink(struct net_device *ndev) { struct prueth_emac *emac = netdev_priv(ndev); struct prueth *prueth = emac->prueth; + int ret; prueth->br_members &= ~BIT(emac->port_id); if (prueth->is_switch_mode) { prueth->is_switch_mode = false; emac->port_vlan = 0; - prueth_emac_restart(prueth); + ret = prueth_emac_restart(prueth); + if (ret) { + dev_err(prueth->dev, "Failed to restart the firmwares, aborting the process"); + return; + } } prueth_offload_fwd_mark_update(prueth); @@ -1228,6 +1329,7 @@ static void prueth_hsr_port_unlink(struct net_device *ndev) struct prueth *prueth = emac->prueth; struct prueth_emac *emac0; struct prueth_emac *emac1; + int ret; emac0 = prueth->emac[PRUETH_MAC0]; emac1 = prueth->emac[PRUETH_MAC1]; @@ -1238,7 +1340,11 @@ static void prueth_hsr_port_unlink(struct net_device *ndev) emac0->port_vlan = 0; emac1->port_vlan = 0; prueth->hsr_dev = NULL; - prueth_emac_restart(prueth); + ret = prueth_emac_restart(prueth); + if (ret) { + dev_err(prueth->dev, "Failed to restart the firmwares, aborting the process"); + return; + } netdev_dbg(ndev, "Disabling HSR Offload mode\n"); } } @@ -1413,13 +1519,10 @@ static int prueth_probe(struct platform_device *pdev) prueth->pa_stats = NULL; } - if (eth0_node) { + if (eth0_node || eth1_node) { ret = prueth_get_cores(prueth, ICSS_SLICE0, false); if (ret) goto put_cores; - } - - if (eth1_node) { ret = prueth_get_cores(prueth, ICSS_SLICE1, false); if (ret) goto put_cores; @@ -1618,14 +1721,12 @@ static int prueth_probe(struct platform_device *pdev) pruss_put(prueth->pruss); put_cores: - if (eth1_node) { - prueth_put_cores(prueth, ICSS_SLICE1); - of_node_put(eth1_node); - } - - if (eth0_node) { + if (eth0_node || eth1_node) { prueth_put_cores(prueth, ICSS_SLICE0); of_node_put(eth0_node); + + prueth_put_cores(prueth, ICSS_SLICE1); + of_node_put(eth1_node); } return ret; diff --git a/drivers/net/ethernet/ti/icssg/icssg_prueth.h b/drivers/net/ethernet/ti/icssg/icssg_prueth.h index f5c1d473e9f9..5473315ea204 100644 --- a/drivers/net/ethernet/ti/icssg/icssg_prueth.h +++ b/drivers/net/ethernet/ti/icssg/icssg_prueth.h @@ -140,7 +140,6 @@ struct prueth_rx_chn { /* data for each emac port */ struct prueth_emac { bool is_sr1; - bool fw_running; struct prueth *prueth; struct net_device *ndev; u8 mac_addr[6]; @@ -361,6 +360,8 @@ int icssg_set_port_state(struct prueth_emac *emac, enum icssg_port_state_cmd state); void icssg_config_set_speed(struct prueth_emac *emac); void icssg_config_half_duplex(struct prueth_emac *emac); +void icssg_init_emac_mode(struct prueth *prueth); +void icssg_init_fw_offload_mode(struct prueth *prueth); /* Buffer queue helpers */ int icssg_queue_pop(struct prueth *prueth, u8 queue); @@ -377,6 +378,7 @@ void icssg_vtbl_modify(struct prueth_emac *emac, u8 vid, u8 port_mask, u8 untag_mask, bool add); u16 icssg_get_pvid(struct prueth_emac *emac); void icssg_set_pvid(struct prueth *prueth, u8 vid, u8 port); +int emac_fdb_flow_id_updated(struct prueth_emac *emac); #define prueth_napi_to_tx_chn(pnapi) \ container_of(pnapi, struct prueth_tx_chn, napi_tx) @@ -407,7 +409,6 @@ void emac_rx_timestamp(struct prueth_emac *emac, struct sk_buff *skb, u32 *psdata); enum netdev_tx icssg_ndo_start_xmit(struct sk_buff *skb, struct net_device *ndev); irqreturn_t prueth_rx_irq(int irq, void *dev_id); -void prueth_emac_stop(struct prueth_emac *emac); void prueth_cleanup_tx_ts(struct prueth_emac *emac); int icssg_napi_rx_poll(struct napi_struct *napi_rx, int budget); int prueth_prepare_rx_chan(struct prueth_emac *emac, diff --git a/drivers/net/ethernet/ti/icssg/icssg_prueth_sr1.c b/drivers/net/ethernet/ti/icssg/icssg_prueth_sr1.c index 5024f0647a0d..3dc86397c367 100644 --- a/drivers/net/ethernet/ti/icssg/icssg_prueth_sr1.c +++ b/drivers/net/ethernet/ti/icssg/icssg_prueth_sr1.c @@ -440,7 +440,6 @@ static int prueth_emac_start(struct prueth *prueth, struct prueth_emac *emac) goto halt_pru; } - emac->fw_running = 1; return 0; halt_pru: @@ -449,6 +448,29 @@ static int prueth_emac_start(struct prueth *prueth, struct prueth_emac *emac) return ret; } +static void prueth_emac_stop(struct prueth_emac *emac) +{ + struct prueth *prueth = emac->prueth; + int slice; + + switch (emac->port_id) { + case PRUETH_PORT_MII0: + slice = ICSS_SLICE0; + break; + case PRUETH_PORT_MII1: + slice = ICSS_SLICE1; + break; + default: + netdev_err(emac->ndev, "invalid port\n"); + return; + } + + if (!emac->is_sr1) + rproc_shutdown(prueth->txpru[slice]); + rproc_shutdown(prueth->rtu[slice]); + rproc_shutdown(prueth->pru[slice]); +} + /** * emac_ndo_open - EMAC device open * @ndev: network adapter device From 9b115361248dc6cce182a2dc030c1c70b0a9639e Mon Sep 17 00:00:00 2001 From: Meghana Malladi Date: Mon, 23 Dec 2024 20:45:50 +0530 Subject: [PATCH 163/196] net: ti: icssg-prueth: Fix clearing of IEP_CMP_CFG registers during iep_init When ICSSG interfaces are brought down and brought up again, the pru cores are shut down and booted again, flushing out all the memories and start again in a clean state. Hence it is expected that the IEP_CMP_CFG register needs to be flushed during iep_init() to ensure that the existing residual configuration doesn't cause any unusual behavior. If the register is not cleared, existing IEP_CMP_CFG set for CMP1 will result in SYNC0_OUT signal based on the SYNC_OUT register values. After bringing the interface up, calling PPS enable doesn't work as the driver believes PPS is already enabled, (iep->pps_enabled is not cleared during interface bring down) and driver will just return true even though there is no signal. Fix this by disabling pps and perout. Fixes: c1e0230eeaab ("net: ti: icss-iep: Add IEP driver") Signed-off-by: Meghana Malladi Reviewed-by: Roger Quadros Signed-off-by: David S. Miller --- drivers/net/ethernet/ti/icssg/icss_iep.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/net/ethernet/ti/icssg/icss_iep.c b/drivers/net/ethernet/ti/icssg/icss_iep.c index 5d6d1cf78e93..768578c0d958 100644 --- a/drivers/net/ethernet/ti/icssg/icss_iep.c +++ b/drivers/net/ethernet/ti/icssg/icss_iep.c @@ -215,6 +215,9 @@ static void icss_iep_enable_shadow_mode(struct icss_iep *iep) for (cmp = IEP_MIN_CMP; cmp < IEP_MAX_CMP; cmp++) { regmap_update_bits(iep->map, ICSS_IEP_CMP_STAT_REG, IEP_CMP_STATUS(cmp), IEP_CMP_STATUS(cmp)); + + regmap_update_bits(iep->map, ICSS_IEP_CMP_CFG_REG, + IEP_CMP_CFG_CMP_EN(cmp), 0); } /* enable reset counter on CMP0 event */ @@ -780,6 +783,11 @@ int icss_iep_exit(struct icss_iep *iep) } icss_iep_disable(iep); + if (iep->pps_enabled) + icss_iep_pps_enable(iep, false); + else if (iep->perout_enabled) + icss_iep_perout_enable(iep, NULL, false); + return 0; } EXPORT_SYMBOL_GPL(icss_iep_exit); From 45d339fefaa3dcd237038769e0d34584fb867390 Mon Sep 17 00:00:00 2001 From: Mark Zhang Date: Thu, 19 Dec 2024 14:23:36 +0200 Subject: [PATCH 164/196] RDMA/mlx5: Enable multiplane mode only when it is supported Driver queries vport_cxt.num_plane and enables multiplane when it is greater then 0, but some old FWs (versions from x.40.1000 till x.42.1000), report vport_cxt.num_plane = 1 unexpectedly. Fix it by querying num_plane only when HCA_CAP2.multiplane bit is set. Fixes: 2a5db20fa532 ("RDMA/mlx5: Add support to multi-plane device and port") Link: https://patch.msgid.link/r/1ef901acdf564716fcf550453cf5e94f343777ec.1734610916.git.leon@kernel.org Cc: stable@vger.kernel.org Reported-by: Francesco Poli Closes: https://lore.kernel.org/all/nvs4i2v7o6vn6zhmtq4sgazy2hu5kiulukxcntdelggmznnl7h@so3oul6uwgbl/ Signed-off-by: Mark Zhang Signed-off-by: Leon Romanovsky Reviewed-by: Michal Swiatkowski Signed-off-by: Jason Gunthorpe --- drivers/infiniband/hw/mlx5/main.c | 2 +- include/linux/mlx5/mlx5_ifc.h | 4 +++- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/infiniband/hw/mlx5/main.c b/drivers/infiniband/hw/mlx5/main.c index c2314797afc9..f5b59d02f4d3 100644 --- a/drivers/infiniband/hw/mlx5/main.c +++ b/drivers/infiniband/hw/mlx5/main.c @@ -2839,7 +2839,7 @@ static int mlx5_ib_get_plane_num(struct mlx5_core_dev *mdev, u8 *num_plane) int err; *num_plane = 0; - if (!MLX5_CAP_GEN(mdev, ib_virt)) + if (!MLX5_CAP_GEN(mdev, ib_virt) || !MLX5_CAP_GEN_2(mdev, multiplane)) return 0; err = mlx5_query_hca_vport_context(mdev, 0, 1, 0, &vport_ctx); diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 4fbbcf35498b..48d47181c7cd 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -2119,7 +2119,9 @@ struct mlx5_ifc_cmd_hca_cap_2_bits { u8 migration_in_chunks[0x1]; u8 reserved_at_d1[0x1]; u8 sf_eq_usage[0x1]; - u8 reserved_at_d3[0xd]; + u8 reserved_at_d3[0x5]; + u8 multiplane[0x1]; + u8 reserved_at_d9[0x7]; u8 cross_vhca_object_to_object_supported[0x20]; From ed123c948d06688d10f3b10a7bce1d6fbfd1ed07 Mon Sep 17 00:00:00 2001 From: Jens Axboe Date: Fri, 3 Jan 2025 09:29:09 -0700 Subject: [PATCH 165/196] io_uring/kbuf: use pre-committed buffer address for non-pollable file For non-pollable files, buffer ring consumption will commit upfront. This is fine, but io_ring_buffer_select() will return the address of the buffer after having committed it. For incrementally consumed buffers, this is incorrect as it will modify the buffer address. Store the pre-committed value and return that. If that isn't done, then the initial part of the buffer is not used and the application will correctly assume the content arrived at the start of the userspace buffer, but the kernel will have put it later in the buffer. Or it can cause a spurious -EFAULT returned in the CQE, depending on the buffer size. As bounds are suitably checked for doing the actual IO, no adverse side effects are possible - it's just a data misplacement within the existing buffer. Reported-by: Gwendal Fernet Cc: stable@vger.kernel.org Fixes: ae98dbf43d75 ("io_uring/kbuf: add support for incremental buffer consumption") Signed-off-by: Jens Axboe --- io_uring/kbuf.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/io_uring/kbuf.c b/io_uring/kbuf.c index d407576ddfb7..eec5eb7de843 100644 --- a/io_uring/kbuf.c +++ b/io_uring/kbuf.c @@ -139,6 +139,7 @@ static void __user *io_ring_buffer_select(struct io_kiocb *req, size_t *len, struct io_uring_buf_ring *br = bl->buf_ring; __u16 tail, head = bl->head; struct io_uring_buf *buf; + void __user *ret; tail = smp_load_acquire(&br->tail); if (unlikely(tail == head)) @@ -153,6 +154,7 @@ static void __user *io_ring_buffer_select(struct io_kiocb *req, size_t *len, req->flags |= REQ_F_BUFFER_RING | REQ_F_BUFFERS_COMMIT; req->buf_list = bl; req->buf_index = buf->bid; + ret = u64_to_user_ptr(buf->addr); if (issue_flags & IO_URING_F_UNLOCKED || !io_file_can_poll(req)) { /* @@ -168,7 +170,7 @@ static void __user *io_ring_buffer_select(struct io_kiocb *req, size_t *len, io_kbuf_commit(req, bl, *len, 1); req->buf_list = NULL; } - return u64_to_user_ptr(buf->addr); + return ret; } void __user *io_buffer_select(struct io_kiocb *req, size_t *len, From 385443057f475e775fe1c66e77d4be9727f40973 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Thomas=20Wei=C3=9Fschuh?= Date: Fri, 3 Jan 2025 19:20:23 +0100 Subject: [PATCH 166/196] kbuild: pacman-pkg: provide versioned linux-api-headers package MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The Arch Linux glibc package contains a versioned dependency on "linux-api-headers". If the linux-api-headers package provided by pacman-pkg does not specify an explicit version this dependency is not satisfied. Fix the dependency by providing an explicit version. Fixes: c8578539deba ("kbuild: add script and target to generate pacman package") Signed-off-by: Thomas Weißschuh Reviewed-by: Nathan Chancellor Signed-off-by: Masahiro Yamada --- scripts/package/PKGBUILD | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/package/PKGBUILD b/scripts/package/PKGBUILD index f83493838cf9..dca706617adc 100644 --- a/scripts/package/PKGBUILD +++ b/scripts/package/PKGBUILD @@ -103,7 +103,7 @@ _package-headers() { _package-api-headers() { pkgdesc="Kernel headers sanitized for use in userspace" - provides=(linux-api-headers) + provides=(linux-api-headers="${pkgver}") conflicts=(linux-api-headers) _prologue From 9d89551994a430b50c4fffcb1e617a057fa76e20 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Sun, 5 Jan 2025 14:13:40 -0800 Subject: [PATCH 167/196] Linux 6.13-rc6 --- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Makefile b/Makefile index 48e89108aa58..7904d5d88088 100644 --- a/Makefile +++ b/Makefile @@ -2,7 +2,7 @@ VERSION = 6 PATCHLEVEL = 13 SUBLEVEL = 0 -EXTRAVERSION = -rc5 +EXTRAVERSION = -rc6 NAME = Baby Opossum Posse # *DOCUMENTATION* From 440ae6f03098133014daa9dd1390afaa8ad530fa Mon Sep 17 00:00:00 2001 From: Wolfram Sang Date: Mon, 2 Dec 2024 09:26:58 +0100 Subject: [PATCH 168/196] i2c: keba: drop check because i2c_unregister_device() is NULL safe No need to check the argument of i2c_unregister_device() because the function itself does it. Signed-off-by: Wolfram Sang Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-keba.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/drivers/i2c/busses/i2c-keba.c b/drivers/i2c/busses/i2c-keba.c index 759732a07ef0..7b9ed2592f5b 100644 --- a/drivers/i2c/busses/i2c-keba.c +++ b/drivers/i2c/busses/i2c-keba.c @@ -464,12 +464,8 @@ static void ki2c_unregister_devices(struct ki2c *ki2c) { int i; - for (i = 0; i < ki2c->client_size; i++) { - struct i2c_client *client = ki2c->client[i]; - - if (client) - i2c_unregister_device(client); - } + for (i = 0; i < ki2c->client_size; i++) + i2c_unregister_device(ki2c->client[i]); } static int ki2c_register_devices(struct ki2c *ki2c) From cf30675e4d0631eac6634b2ae3f77b7b0eafe447 Mon Sep 17 00:00:00 2001 From: Carlos Song Date: Mon, 25 Nov 2024 22:15:21 +0800 Subject: [PATCH 169/196] i2c: imx: fix divide by zero warning Add "i2c_clk_rate / 2" check to avoid "divide by zero warning". i2c_clk_rate may be zero if i2c clock is disabled. Signed-off-by: Carlos Song Signed-off-by: Clark Wang Signed-off-by: Haibo Chen Reviewed-by: Frank Li Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-imx.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c index 5c9a8dfbc4a0..da77ba5ee579 100644 --- a/drivers/i2c/busses/i2c-imx.c +++ b/drivers/i2c/busses/i2c-imx.c @@ -621,8 +621,8 @@ static int i2c_imx_acked(struct imx_i2c_struct *i2c_imx) return 0; } -static void i2c_imx_set_clk(struct imx_i2c_struct *i2c_imx, - unsigned int i2c_clk_rate) +static int i2c_imx_set_clk(struct imx_i2c_struct *i2c_imx, + unsigned int i2c_clk_rate) { struct imx_i2c_clk_pair *i2c_clk_div = i2c_imx->hwdata->clk_div; unsigned int div; @@ -637,7 +637,11 @@ static void i2c_imx_set_clk(struct imx_i2c_struct *i2c_imx, /* Divider value calculation */ if (i2c_imx->cur_clk == i2c_clk_rate) - return; + return 0; + + /* Keep the denominator of the following program always NOT equal to 0. */ + if (!(i2c_clk_rate / 2)) + return -EINVAL; i2c_imx->cur_clk = i2c_clk_rate; @@ -668,6 +672,8 @@ static void i2c_imx_set_clk(struct imx_i2c_struct *i2c_imx, dev_dbg(&i2c_imx->adapter.dev, "IFDR[IC]=0x%x, REAL DIV=%d\n", i2c_clk_div[i].val, i2c_clk_div[i].div); #endif + + return 0; } static int i2c_imx_clk_notifier_call(struct notifier_block *nb, @@ -677,11 +683,12 @@ static int i2c_imx_clk_notifier_call(struct notifier_block *nb, struct imx_i2c_struct *i2c_imx = container_of(nb, struct imx_i2c_struct, clk_change_nb); + int ret = 0; if (action & POST_RATE_CHANGE) - i2c_imx_set_clk(i2c_imx, ndata->new_rate); + ret = i2c_imx_set_clk(i2c_imx, ndata->new_rate); - return NOTIFY_OK; + return notifier_from_errno(ret); } static int i2c_imx_start(struct imx_i2c_struct *i2c_imx, bool atomic) @@ -1780,7 +1787,11 @@ static int i2c_imx_probe(struct platform_device *pdev) i2c_imx->bitrate = pdata->bitrate; i2c_imx->clk_change_nb.notifier_call = i2c_imx_clk_notifier_call; clk_notifier_register(i2c_imx->clk, &i2c_imx->clk_change_nb); - i2c_imx_set_clk(i2c_imx, clk_get_rate(i2c_imx->clk)); + ret = i2c_imx_set_clk(i2c_imx, clk_get_rate(i2c_imx->clk)); + if (ret < 0) { + dev_err(&pdev->dev, "can't get I2C clock\n"); + goto clk_notifier_unregister; + } i2c_imx_reset_regs(i2c_imx); From 358025ac091e5a54f9819b33ee9c7cb07c55ee5d Mon Sep 17 00:00:00 2001 From: Carlos Song Date: Mon, 25 Nov 2024 22:21:08 +0800 Subject: [PATCH 170/196] i2c: imx: make controller available until system suspend_noirq() and from resume_noirq() Put runtime PM to resume state between suspend() and suspend_noirq(), resume_noirq() and resume(), because some I2C devices need the controller on to perform communication during this period. The controller can't be woken up once runtime pm is disabled and in runtime autosuspended state. The problem can be easily reproduced on the I.MX8MQ platform: the PMIC needs to be used to enable regulator when the system resumes. When the PMIC uses the I2C controller, I2C runtime pm has not been enabled, so in i2c xfer(), pm_runtime_resume_and_get() will return an error, which causes data transfer to fail. Therefore, regulators cannot be enabled and system resume hangs. Here is resume error log: [ 53.888902] galcore 38000000.gpu3d: PM: calling genpd_resume_noirq @ 529, parent: platform [ 53.897203] i2c_imx_xfer, pm_runtime_resume_and_get is -13 [ 53.902713] imx-pgc imx-pgc-domain.5: failed to enable regulator: -EACCES [ 53.909518] galcore 38000000.gpu3d: PM: genpd_resume_noirq returned 0 after 12331 usecs [ 53.917545] mxc_hantro 38300000.vpu: PM: calling genpd_resume_noirq @ 529, parent: soc@0 [ 53.925659] i2c_imx_xfer, pm_runtime_resume_and_get is -13 [ 53.931157] imx-pgc imx-pgc-domain.6: failed to enable regulator: -EACCES I.MX8MQ system resume normally after applying the fix. Here is resume log: [ 71.068807] galcore 38000000.gpu3d: PM: calling genpd_resume_noirq @ 530, parent: platform [ 71.077103] i2c_imx_xfer, pm_runtime_resume_and_get is 0 [ 71.083578] galcore 38000000.gpu3d: PM: genpd_resume_noirq returned 0 after 6490 usecs [ 71.091526] mxc_hantro 38300000.vpu: PM: calling genpd_resume_noirq @ 530, parent: soc@0 [ 71.099638] i2c_imx_xfer, pm_runtime_resume_and_get is 0 [ 71.106091] mxc_hantro 38300000.vpu: PM: genpd_resume_noirq returned 0 after 6458 usecs Signed-off-by: Carlos Song Signed-off-by: Haibo Chen Reviewed-by: Frank Li Link: https://lore.kernel.org/r/20241125142108.1613016-1-carlos.song@nxp.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-imx.c | 39 +++++++++++++++++++++++++++++++++++- 1 file changed, 38 insertions(+), 1 deletion(-) diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c index da77ba5ee579..167985f60d27 100644 --- a/drivers/i2c/busses/i2c-imx.c +++ b/drivers/i2c/busses/i2c-imx.c @@ -1767,7 +1767,8 @@ static int i2c_imx_probe(struct platform_device *pdev) goto rpm_disable; /* Request IRQ */ - ret = request_irq(irq, i2c_imx_isr, IRQF_SHARED, pdev->name, i2c_imx); + ret = request_irq(irq, i2c_imx_isr, IRQF_SHARED | IRQF_NO_SUSPEND, + pdev->name, i2c_imx); if (ret) { dev_err(&pdev->dev, "can't claim irq %d\n", irq); goto rpm_disable; @@ -1885,7 +1886,43 @@ static int i2c_imx_runtime_resume(struct device *dev) return ret; } +static int i2c_imx_suspend(struct device *dev) +{ + /* + * Some I2C devices may need the I2C controller to remain active + * during resume_noirq() or suspend_noirq(). If the controller is + * autosuspended, there is no way to wake it up once runtime PM is + * disabled (in suspend_late()). + * + * During system resume, the I2C controller will be available only + * after runtime PM is re-enabled (in resume_early()). However, this + * may be too late for some devices. + * + * Wake up the controller in the suspend() callback while runtime PM + * is still enabled. The I2C controller will remain available until + * the suspend_noirq() callback (pm_runtime_force_suspend()) is + * called. During resume, the I2C controller can be restored by the + * resume_noirq() callback (pm_runtime_force_resume()). + * + * Finally, the resume() callback re-enables autosuspend, ensuring + * the I2C controller remains available until the system enters + * suspend_noirq() and from resume_noirq(). + */ + return pm_runtime_resume_and_get(dev); +} + +static int i2c_imx_resume(struct device *dev) +{ + pm_runtime_mark_last_busy(dev); + pm_runtime_put_autosuspend(dev); + + return 0; +} + static const struct dev_pm_ops i2c_imx_pm_ops = { + NOIRQ_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, + pm_runtime_force_resume) + SYSTEM_SLEEP_PM_OPS(i2c_imx_suspend, i2c_imx_resume) RUNTIME_PM_OPS(i2c_imx_runtime_suspend, i2c_imx_runtime_resume, NULL) }; From 576eba03c99435380d155e5f71d5d7603b9178f6 Mon Sep 17 00:00:00 2001 From: Carlos Song Date: Mon, 23 Dec 2024 11:43:43 +0800 Subject: [PATCH 171/196] i2c: imx: switch different pinctrl state in different system power status Switch different pinctrl state in different system power status. Signed-off-by: Carlos Song Reviewed-by: Frank Li Acked-by: Oleksij Rempel Link: https://lore.kernel.org/r/20241223034343.544002-1-carlos.song@nxp.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-imx.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c index 167985f60d27..1f441227bfdc 100644 --- a/drivers/i2c/busses/i2c-imx.c +++ b/drivers/i2c/busses/i2c-imx.c @@ -1870,8 +1870,7 @@ static int i2c_imx_runtime_suspend(struct device *dev) struct imx_i2c_struct *i2c_imx = dev_get_drvdata(dev); clk_disable(i2c_imx->clk); - - return 0; + return pinctrl_pm_select_sleep_state(dev); } static int i2c_imx_runtime_resume(struct device *dev) @@ -1879,6 +1878,10 @@ static int i2c_imx_runtime_resume(struct device *dev) struct imx_i2c_struct *i2c_imx = dev_get_drvdata(dev); int ret; + ret = pinctrl_pm_select_default_state(dev); + if (ret) + return ret; + ret = clk_enable(i2c_imx->clk); if (ret) dev_err(dev, "can't enable I2C clock, ret=%d\n", ret); From 4a34b9bf5f6f3f3ed910863fd1b4741a62864d4b Mon Sep 17 00:00:00 2001 From: Tyrone Ting Date: Thu, 19 Dec 2024 17:08:56 +0800 Subject: [PATCH 172/196] i2c: npcm: Modify timeout evaluation mechanism The users want to connect a lot of masters on the same bus. This timeout is used to determine the time it takes to take bus ownership. The transactions are very long, so waiting 35ms is not enough. Increase the timeout and treat it as the total timeout, including retries. The total timeout is 2 seconds now. The i2c core layer will have chances to retry to call the i2c driver transfer function if the i2c driver reports that the bus is busy and returns -EAGAIN. Signed-off-by: Tyrone Ting Reviewed-by: Tali Perry Link: https://lore.kernel.org/r/20241219090859.18722-2-kfting@nuvoton.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-npcm7xx.c | 24 +++++++++++++++--------- 1 file changed, 15 insertions(+), 9 deletions(-) diff --git a/drivers/i2c/busses/i2c-npcm7xx.c b/drivers/i2c/busses/i2c-npcm7xx.c index 482a0074d448..c96a25d37c14 100644 --- a/drivers/i2c/busses/i2c-npcm7xx.c +++ b/drivers/i2c/busses/i2c-npcm7xx.c @@ -2132,19 +2132,12 @@ static int npcm_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, } } - /* - * Adaptive TimeOut: estimated time in usec + 100% margin: - * 2: double the timeout for clock stretching case - * 9: bits per transaction (including the ack/nack) - */ - timeout_usec = (2 * 9 * USEC_PER_SEC / bus->bus_freq) * (2 + nread + nwrite); - timeout = max_t(unsigned long, bus->adap.timeout, usecs_to_jiffies(timeout_usec)); if (nwrite >= 32 * 1024 || nread >= 32 * 1024) { dev_err(bus->dev, "i2c%d buffer too big\n", bus->num); return -EINVAL; } - time_left = jiffies + timeout + 1; + time_left = jiffies + bus->adap.timeout / bus->adap.retries + 1; do { /* * we must clear slave address immediately when the bus is not @@ -2192,6 +2185,14 @@ static int npcm_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, if (npcm_i2c_master_start_xmit(bus, slave_addr, nwrite, nread, write_data, read_data, read_PEC, read_block)) { + /* + * Adaptive TimeOut: estimated time in usec + 100% margin: + * 2: double the timeout for clock stretching case + * 9: bits per transaction (including the ack/nack) + */ + timeout_usec = (2 * 9 * USEC_PER_SEC / bus->bus_freq) * (2 + nread + nwrite); + timeout = max_t(unsigned long, bus->adap.timeout / bus->adap.retries, + usecs_to_jiffies(timeout_usec)); time_left = wait_for_completion_timeout(&bus->cmd_complete, timeout); @@ -2317,7 +2318,12 @@ static int npcm_i2c_probe_bus(struct platform_device *pdev) adap = &bus->adap; adap->owner = THIS_MODULE; adap->retries = 3; - adap->timeout = msecs_to_jiffies(35); + /* + * The users want to connect a lot of masters on the same bus. + * This timeout is used to determine the time it takes to take bus ownership. + * The transactions are very long, so waiting 35ms is not enough. + */ + adap->timeout = 2 * HZ; adap->algo = &npcm_i2c_algo; adap->quirks = &npcm_i2c_quirks; adap->algo_data = bus; From 862dd4f7733a694b12ad9e023ce639e9c21d6abd Mon Sep 17 00:00:00 2001 From: Tyrone Ting Date: Thu, 19 Dec 2024 17:08:57 +0800 Subject: [PATCH 173/196] i2c: npcm: Assign client address earlier for `i2c_recover_bus()` Store the client address earlier since it might get called in the i2c_recover_bus() logic flow at the early stage of npcm_i2c_master_xfer(). Signed-off-by: Tyrone Ting Reviewed-by: Tali Perry Link: https://lore.kernel.org/r/20241219090859.18722-3-kfting@nuvoton.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-npcm7xx.c | 23 +++++++++++++++++------ 1 file changed, 17 insertions(+), 6 deletions(-) diff --git a/drivers/i2c/busses/i2c-npcm7xx.c b/drivers/i2c/busses/i2c-npcm7xx.c index c96a25d37c14..760608fdd075 100644 --- a/drivers/i2c/busses/i2c-npcm7xx.c +++ b/drivers/i2c/busses/i2c-npcm7xx.c @@ -2035,7 +2035,7 @@ static irqreturn_t npcm_i2c_bus_irq(int irq, void *dev_id) } static bool npcm_i2c_master_start_xmit(struct npcm_i2c *bus, - u8 slave_addr, u16 nwrite, u16 nread, + u16 nwrite, u16 nread, u8 *write_data, u8 *read_data, bool use_PEC, bool use_read_block) { @@ -2043,7 +2043,6 @@ static bool npcm_i2c_master_start_xmit(struct npcm_i2c *bus, bus->cmd_err = -EBUSY; return false; } - bus->dest_addr = slave_addr << 1; bus->wr_buf = write_data; bus->wr_size = nwrite; bus->wr_ind = 0; @@ -2086,7 +2085,6 @@ static int npcm_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, unsigned long time_left, flags; u16 nwrite, nread; u8 *write_data, *read_data; - u8 slave_addr; unsigned long timeout; bool read_block = false; bool read_PEC = false; @@ -2099,7 +2097,6 @@ static int npcm_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, } msg0 = &msgs[0]; - slave_addr = msg0->addr; if (msg0->flags & I2C_M_RD) { /* read */ nwrite = 0; write_data = NULL; @@ -2155,6 +2152,21 @@ static int npcm_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, } while (time_is_after_jiffies(time_left) && bus_busy); + /* + * Store the address early in a global position to ensure it is + * accessible for a potential call to i2c_recover_bus(). + * + * Since the transfer might be a read operation, remove the I2C_M_RD flag + * from the bus->dest_addr for the i2c_recover_bus() call later. + * + * The i2c_recover_bus() uses the address in a write direction to recover + * the i2c bus if some error condition occurs. + * + * Remove the I2C_M_RD flag from the address since npcm_i2c_master_start_xmit() + * handles the read/write operation internally. + */ + bus->dest_addr = i2c_8bit_addr_from_msg(msg0) & ~I2C_M_RD; + /* * Check the BER (bus error) state, when ber_state is true, it means that the module * detects the bus error which is caused by some factor like that the electricity @@ -2172,7 +2184,6 @@ static int npcm_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, } npcm_i2c_init_params(bus); - bus->dest_addr = slave_addr; bus->msgs = msgs; bus->msgs_num = num; bus->cmd_err = 0; @@ -2182,7 +2193,7 @@ static int npcm_i2c_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, npcm_i2c_int_enable(bus, true); - if (npcm_i2c_master_start_xmit(bus, slave_addr, nwrite, nread, + if (npcm_i2c_master_start_xmit(bus, nwrite, nread, write_data, read_data, read_PEC, read_block)) { /* From 6ddc9132bc42a7dfcaca0cce4ff0f94e69bc5114 Mon Sep 17 00:00:00 2001 From: Tyrone Ting Date: Thu, 19 Dec 2024 17:08:58 +0800 Subject: [PATCH 174/196] i2c: npcm: use i2c frequency table Modify i2c frequency from table parameters for NPCM i2c modules. Supported frequencies are: 1. 100KHz 2. 400KHz 3. 1MHz The original equations were tested on a variety of chips and base clocks. Since we added devices that use higher frequencies of the module we saw that there is a mismatch between the equation and the actual results on the bus itself, measured on scope. Meanwhile, the equations were not accurate to begin with. They are an approximation of the ideal value. The ideal value is calculated per frequency of the core module. So instead of using the equations we did an optimization per module frequency, verified on a device. Most of the work was focused on the rise time of the SCL and SDA, which depends on external load of the bus and PU. Different PCB designs, or specifically to this case: the number and type of targets on the bus, impact the required values for the timing registers. Users can recalculate the numbers for each bus and get an even better optimization, but our users chose not to. We manually picked values per frequency that match the entire valid range of targets (from 1 to max number). Then we check against the AMR described in SMB spec and make sure that none of the values is exceeding. This process was led by the chip architect and included a lot of testing. Signed-off-by: Tyrone Ting Reviewed-by: Tali Perry Link: https://lore.kernel.org/r/20241219090859.18722-4-kfting@nuvoton.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-npcm7xx.c | 374 ++++++++++++++++++++++++------- 1 file changed, 288 insertions(+), 86 deletions(-) diff --git a/drivers/i2c/busses/i2c-npcm7xx.c b/drivers/i2c/busses/i2c-npcm7xx.c index 760608fdd075..1aae1a8a8055 100644 --- a/drivers/i2c/busses/i2c-npcm7xx.c +++ b/drivers/i2c/busses/i2c-npcm7xx.c @@ -263,6 +263,265 @@ static const int npcm_i2caddr[I2C_NUM_OWN_ADDR] = { #define I2C_FREQ_MIN_HZ 10000 #define I2C_FREQ_MAX_HZ I2C_MAX_FAST_MODE_PLUS_FREQ +struct smb_timing_t { + u32 core_clk; + u8 hldt; + u8 dbcnt; + u16 sclfrq; + u8 scllt; + u8 sclht; + bool fast_mode; +}; + +static struct smb_timing_t smb_timing_100khz[] = { + { + .core_clk = 100000000, .hldt = 0x2A, .dbcnt = 0x4, + .sclfrq = 0xFB, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 62500000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x9D, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 50000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x7E, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 48000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x79, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 40000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x65, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 30000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x4C, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 29000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x49, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 26000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x42, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 25000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x3F, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 24000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x3D, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 20000000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x33, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 16180000, .hldt = 0x2A, .dbcnt = 0x1, + .sclfrq = 0x29, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 15000000, .hldt = 0x23, .dbcnt = 0x1, + .sclfrq = 0x26, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 13000000, .hldt = 0x1D, .dbcnt = 0x1, + .sclfrq = 0x21, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 12000000, .hldt = 0x1B, .dbcnt = 0x1, + .sclfrq = 0x1F, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 10000000, .hldt = 0x18, .dbcnt = 0x1, + .sclfrq = 0x1A, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 9000000, .hldt = 0x16, .dbcnt = 0x1, + .sclfrq = 0x17, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 8090000, .hldt = 0x14, .dbcnt = 0x1, + .sclfrq = 0x15, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 7500000, .hldt = 0x7, .dbcnt = 0x1, + .sclfrq = 0x13, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 6500000, .hldt = 0xE, .dbcnt = 0x1, + .sclfrq = 0x11, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, + { + .core_clk = 4000000, .hldt = 0x9, .dbcnt = 0x1, + .sclfrq = 0xB, .scllt = 0x0, .sclht = 0x0, + .fast_mode = false, + }, +}; + +static struct smb_timing_t smb_timing_400khz[] = { + { + .core_clk = 100000000, .hldt = 0x2A, .dbcnt = 0x3, + .sclfrq = 0x0, .scllt = 0x47, .sclht = 0x35, + .fast_mode = true, + }, + { + .core_clk = 62500000, .hldt = 0x2A, .dbcnt = 0x2, + .sclfrq = 0x0, .scllt = 0x2C, .sclht = 0x22, + .fast_mode = true, + }, + { + .core_clk = 50000000, .hldt = 0x21, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x24, .sclht = 0x1B, + .fast_mode = true, + }, + { + .core_clk = 48000000, .hldt = 0x1E, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x24, .sclht = 0x19, + .fast_mode = true, + }, + { + .core_clk = 40000000, .hldt = 0x1B, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x1E, .sclht = 0x14, + .fast_mode = true, + }, + { + .core_clk = 33000000, .hldt = 0x15, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x19, .sclht = 0x11, + .fast_mode = true, + }, + { + .core_clk = 30000000, .hldt = 0x15, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x19, .sclht = 0xD, + .fast_mode = true, + }, + { + .core_clk = 29000000, .hldt = 0x11, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x15, .sclht = 0x10, + .fast_mode = true, + }, + { + .core_clk = 26000000, .hldt = 0x10, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x13, .sclht = 0xE, + .fast_mode = true, + }, + { + .core_clk = 25000000, .hldt = 0xF, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x13, .sclht = 0xD, + .fast_mode = true, + }, + { + .core_clk = 24000000, .hldt = 0xD, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x12, .sclht = 0xD, + .fast_mode = true, + }, + { + .core_clk = 20000000, .hldt = 0xB, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0xF, .sclht = 0xA, + .fast_mode = true, + }, + { + .core_clk = 16180000, .hldt = 0xA, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0xC, .sclht = 0x9, + .fast_mode = true, + }, + { + .core_clk = 15000000, .hldt = 0x9, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0xB, .sclht = 0x8, + .fast_mode = true, + }, + { + .core_clk = 13000000, .hldt = 0x7, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0xA, .sclht = 0x7, + .fast_mode = true, + }, + { + .core_clk = 12000000, .hldt = 0x7, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0xA, .sclht = 0x6, + .fast_mode = true, + }, + { + .core_clk = 10000000, .hldt = 0x6, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x8, .sclht = 0x5, + .fast_mode = true, + }, +}; + +static struct smb_timing_t smb_timing_1000khz[] = { + { + .core_clk = 100000000, .hldt = 0x15, .dbcnt = 0x4, + .sclfrq = 0x0, .scllt = 0x1C, .sclht = 0x15, + .fast_mode = true, + }, + { + .core_clk = 62500000, .hldt = 0xF, .dbcnt = 0x3, + .sclfrq = 0x0, .scllt = 0x11, .sclht = 0xE, + .fast_mode = true, + }, + { + .core_clk = 50000000, .hldt = 0xA, .dbcnt = 0x2, + .sclfrq = 0x0, .scllt = 0xE, .sclht = 0xB, + .fast_mode = true, + }, + { + .core_clk = 48000000, .hldt = 0x9, .dbcnt = 0x2, + .sclfrq = 0x0, .scllt = 0xD, .sclht = 0xB, + .fast_mode = true, + }, + { + .core_clk = 41000000, .hldt = 0x9, .dbcnt = 0x2, + .sclfrq = 0x0, .scllt = 0xC, .sclht = 0x9, + .fast_mode = true, + }, + { + .core_clk = 40000000, .hldt = 0x8, .dbcnt = 0x2, + .sclfrq = 0x0, .scllt = 0xB, .sclht = 0x9, + .fast_mode = true, + }, + { + .core_clk = 33000000, .hldt = 0x7, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0xA, .sclht = 0x7, + .fast_mode = true, + }, + { + .core_clk = 25000000, .hldt = 0x4, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x7, .sclht = 0x6, + .fast_mode = true, + }, + { + .core_clk = 24000000, .hldt = 0x7, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x8, .sclht = 0x5, + .fast_mode = true, + }, + { + .core_clk = 20000000, .hldt = 0x4, .dbcnt = 0x1, + .sclfrq = 0x0, .scllt = 0x6, .sclht = 0x4, + .fast_mode = true, + }, +}; + struct npcm_i2c_data { u8 fifo_size; u32 segctl_init_val; @@ -1805,102 +2064,45 @@ static void npcm_i2c_recovery_init(struct i2c_adapter *_adap) */ static int npcm_i2c_init_clk(struct npcm_i2c *bus, u32 bus_freq_hz) { - u32 k1 = 0; - u32 k2 = 0; - u8 dbnct = 0; - u32 sclfrq = 0; - u8 hldt = 7; + struct smb_timing_t *smb_timing; + u8 scl_table_cnt = 0, table_size = 0; u8 fast_mode = 0; - u32 src_clk_khz; - u32 bus_freq_khz; - src_clk_khz = bus->apb_clk / 1000; - bus_freq_khz = bus_freq_hz / 1000; bus->bus_freq = bus_freq_hz; - /* 100KHz and below: */ - if (bus_freq_hz <= I2C_MAX_STANDARD_MODE_FREQ) { - sclfrq = src_clk_khz / (bus_freq_khz * 4); - - if (sclfrq < SCLFRQ_MIN || sclfrq > SCLFRQ_MAX) - return -EDOM; - - if (src_clk_khz >= 40000) - hldt = 17; - else if (src_clk_khz >= 12500) - hldt = 15; - else - hldt = 7; - } - - /* 400KHz: */ - else if (bus_freq_hz <= I2C_MAX_FAST_MODE_FREQ) { - sclfrq = 0; + switch (bus_freq_hz) { + case I2C_MAX_STANDARD_MODE_FREQ: + smb_timing = smb_timing_100khz; + table_size = ARRAY_SIZE(smb_timing_100khz); + break; + case I2C_MAX_FAST_MODE_FREQ: + smb_timing = smb_timing_400khz; + table_size = ARRAY_SIZE(smb_timing_400khz); fast_mode = I2CCTL3_400K_MODE; - - if (src_clk_khz < 7500) - /* 400KHZ cannot be supported for core clock < 7.5MHz */ - return -EDOM; - - else if (src_clk_khz >= 50000) { - k1 = 80; - k2 = 48; - hldt = 12; - dbnct = 7; - } - - /* Master or Slave with frequency > 25MHz */ - else if (src_clk_khz > 25000) { - hldt = clk_coef(src_clk_khz, 300) + 7; - k1 = clk_coef(src_clk_khz, 1600); - k2 = clk_coef(src_clk_khz, 900); - } - } - - /* 1MHz: */ - else if (bus_freq_hz <= I2C_MAX_FAST_MODE_PLUS_FREQ) { - sclfrq = 0; + break; + case I2C_MAX_FAST_MODE_PLUS_FREQ: + smb_timing = smb_timing_1000khz; + table_size = ARRAY_SIZE(smb_timing_1000khz); fast_mode = I2CCTL3_400K_MODE; - - /* 1MHZ cannot be supported for core clock < 24 MHz */ - if (src_clk_khz < 24000) - return -EDOM; - - k1 = clk_coef(src_clk_khz, 620); - k2 = clk_coef(src_clk_khz, 380); - - /* Core clk > 40 MHz */ - if (src_clk_khz > 40000) { - /* - * Set HLDT: - * SDA hold time: (HLDT-7) * T(CLK) >= 120 - * HLDT = 120/T(CLK) + 7 = 120 * FREQ(CLK) + 7 - */ - hldt = clk_coef(src_clk_khz, 120) + 7; - } else { - hldt = 7; - dbnct = 2; - } + break; + default: + return -EINVAL; } - /* Frequency larger than 1 MHz is not supported */ - else + for (scl_table_cnt = 0; scl_table_cnt < table_size; scl_table_cnt++) + if (bus->apb_clk >= smb_timing[scl_table_cnt].core_clk) + break; + + if (scl_table_cnt == table_size) return -EINVAL; - if (bus_freq_hz >= I2C_MAX_FAST_MODE_FREQ) { - k1 = round_up(k1, 2); - k2 = round_up(k2 + 1, 2); - if (k1 < SCLFRQ_MIN || k1 > SCLFRQ_MAX || - k2 < SCLFRQ_MIN || k2 > SCLFRQ_MAX) - return -EDOM; - } - /* write sclfrq value. bits [6:0] are in I2CCTL2 reg */ - iowrite8(FIELD_PREP(I2CCTL2_SCLFRQ6_0, sclfrq & 0x7F), + iowrite8(FIELD_PREP(I2CCTL2_SCLFRQ6_0, smb_timing[scl_table_cnt].sclfrq & 0x7F), bus->reg + NPCM_I2CCTL2); /* bits [8:7] are in I2CCTL3 reg */ - iowrite8(fast_mode | FIELD_PREP(I2CCTL3_SCLFRQ8_7, (sclfrq >> 7) & 0x3), + iowrite8(FIELD_PREP(I2CCTL3_SCLFRQ8_7, (smb_timing[scl_table_cnt].sclfrq >> 7) & 0x3) | + fast_mode, bus->reg + NPCM_I2CCTL3); /* Select Bank 0 to access NPCM_I2CCTL4/NPCM_I2CCTL5 */ @@ -1912,13 +2114,13 @@ static int npcm_i2c_init_clk(struct npcm_i2c *bus, u32 bus_freq_hz) * k1 = 2 * SCLLT7-0 -> Low Time = k1 / 2 * k2 = 2 * SCLLT7-0 -> High Time = k2 / 2 */ - iowrite8(k1 / 2, bus->reg + NPCM_I2CSCLLT); - iowrite8(k2 / 2, bus->reg + NPCM_I2CSCLHT); + iowrite8(smb_timing[scl_table_cnt].scllt, bus->reg + NPCM_I2CSCLLT); + iowrite8(smb_timing[scl_table_cnt].sclht, bus->reg + NPCM_I2CSCLHT); - iowrite8(dbnct, bus->reg + NPCM_I2CCTL5); + iowrite8(smb_timing[scl_table_cnt].dbcnt, bus->reg + NPCM_I2CCTL5); } - iowrite8(hldt, bus->reg + NPCM_I2CCTL4); + iowrite8(smb_timing[scl_table_cnt].hldt, bus->reg + NPCM_I2CCTL4); /* Return to Bank 1, and stay there by default: */ npcm_i2c_select_bank(bus, I2C_BANK_1); From 8b56f98b28e6238bcfe00de743b4361d26a7f980 Mon Sep 17 00:00:00 2001 From: Charles Boyer Date: Thu, 19 Dec 2024 17:08:59 +0800 Subject: [PATCH 175/196] i2c: npcm: Enable slave in eob interrupt Nuvoton slave enable was in user space API call master_xfer, so it is subject to delays from the OS scheduler. If the BMC is not enabled for slave mode in time for master to send response, then it will NAK the address match. Then the PLDM request timeout occurs. If the slave enable is moved to the EOB interrupt service routine, then the BMC can be ready in slave mode by the time it needs to receive a response. Signed-off-by: Charles Boyer Signed-off-by: Vivekanand Veeracholan Signed-off-by: Tyrone Ting Reviewed-by: Tali Perry Link: https://lore.kernel.org/r/20241219090859.18722-5-kfting@nuvoton.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-npcm7xx.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/i2c/busses/i2c-npcm7xx.c b/drivers/i2c/busses/i2c-npcm7xx.c index 1aae1a8a8055..3ca08b8ef8af 100644 --- a/drivers/i2c/busses/i2c-npcm7xx.c +++ b/drivers/i2c/busses/i2c-npcm7xx.c @@ -1925,6 +1925,12 @@ static int npcm_i2c_int_master_handler(struct npcm_i2c *bus) (FIELD_GET(NPCM_I2CCST3_EO_BUSY, ioread8(bus->reg + NPCM_I2CCST3)))) { npcm_i2c_irq_handle_eob(bus); +#if IS_ENABLED(CONFIG_I2C_SLAVE) + /* reenable slave if it was enabled */ + if (bus->slave) + iowrite8(bus->slave->addr | NPCM_I2CADDR_SAEN, + bus->reg + NPCM_I2CADDR1); +#endif return 0; } From d9a13c08f88d696218a44658f290b1b433e787b6 Mon Sep 17 00:00:00 2001 From: Carlos Song Date: Thu, 26 Dec 2024 14:25:21 +0800 Subject: [PATCH 176/196] i2c: imx: support DMA defer probing Return -EPROBE_DEFER when dma_request_slave_channel() because DMA driver have not ready yet. Move i2c_imx_dma_request() before registering I2C adapter to avoid infinite loop of .probe() calls to the same driver, see "e8c220fac415 Revert "i2c: imx: improve the error handling in i2c_imx_dma_request()"" and "Documentation/driver-api/driver-model/driver.rst". Use CPU mode to avoid stuck registering i2c adapter when DMA resources are unavailable. Signed-off-by: Carlos Song Signed-off-by: Clark Wang Acked-by: Ahmad Fatoum Acked-by: Oleksij Rempel Link: https://lore.kernel.org/r/20241226062521.1004809-1-carlos.song@nxp.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-imx.c | 30 ++++++++++++++++++++++-------- 1 file changed, 22 insertions(+), 8 deletions(-) diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c index 1f441227bfdc..ee0d25b498cb 100644 --- a/drivers/i2c/busses/i2c-imx.c +++ b/drivers/i2c/busses/i2c-imx.c @@ -397,17 +397,16 @@ static void i2c_imx_reset_regs(struct imx_i2c_struct *i2c_imx) } /* Functions for DMA support */ -static void i2c_imx_dma_request(struct imx_i2c_struct *i2c_imx, - dma_addr_t phy_addr) +static int i2c_imx_dma_request(struct imx_i2c_struct *i2c_imx, dma_addr_t phy_addr) { struct imx_i2c_dma *dma; struct dma_slave_config dma_sconfig; - struct device *dev = &i2c_imx->adapter.dev; + struct device *dev = i2c_imx->adapter.dev.parent; int ret; dma = devm_kzalloc(dev, sizeof(*dma), GFP_KERNEL); if (!dma) - return; + return -ENOMEM; dma->chan_tx = dma_request_chan(dev, "tx"); if (IS_ERR(dma->chan_tx)) { @@ -452,7 +451,7 @@ static void i2c_imx_dma_request(struct imx_i2c_struct *i2c_imx, dev_info(dev, "using %s (tx) and %s (rx) for DMA transfers\n", dma_chan_name(dma->chan_tx), dma_chan_name(dma->chan_rx)); - return; + return 0; fail_rx: dma_release_channel(dma->chan_rx); @@ -460,6 +459,8 @@ static void i2c_imx_dma_request(struct imx_i2c_struct *i2c_imx, dma_release_channel(dma->chan_tx); fail_al: devm_kfree(dev, dma); + + return ret; } static void i2c_imx_dma_callback(void *arg) @@ -1802,6 +1803,22 @@ static int i2c_imx_probe(struct platform_device *pdev) if (ret == -EPROBE_DEFER) goto clk_notifier_unregister; + /* + * DMA mode should be optional for I2C, when encountering DMA errors, + * no need to exit I2C probe. Only print warning to show DMA error and + * use PIO mode directly to ensure I2C bus available as much as possible. + */ + ret = i2c_imx_dma_request(i2c_imx, phy_addr); + if (ret) { + if (ret == -EPROBE_DEFER) + goto clk_notifier_unregister; + else if (ret == -ENODEV) + dev_dbg(&pdev->dev, "Only use PIO mode\n"); + else + dev_warn(&pdev->dev, "Failed to setup DMA (%pe), only use PIO mode\n", + ERR_PTR(ret)); + } + /* Add I2C adapter */ ret = i2c_add_numbered_adapter(&i2c_imx->adapter); if (ret < 0) @@ -1816,9 +1833,6 @@ static int i2c_imx_probe(struct platform_device *pdev) i2c_imx->adapter.name); dev_info(&i2c_imx->adapter.dev, "IMX I2C adapter registered\n"); - /* Init DMA config if supported */ - i2c_imx_dma_request(i2c_imx, phy_addr); - return 0; /* Return OK */ clk_notifier_unregister: From 567ef9035f820ecd3e6fcc142003f396848b31f7 Mon Sep 17 00:00:00 2001 From: Biju Das Date: Mon, 16 Dec 2024 20:24:34 +0000 Subject: [PATCH 177/196] dt-bindings: i2c: renesas,riic: Document the R9A09G047 support Document support for the I2C Bus Interface (RIIC) found in the Renesas RZ/G3E (R9A09G047) SoC. This IP is compatible with Renesas RZ/V2H (R9A09G057) RIIC IP. Acked-by: Conor Dooley Reviewed-by: Geert Uytterhoeven Signed-off-by: Biju Das Link: https://lore.kernel.org/r/20241216202436.185773-1-biju.das.jz@bp.renesas.com Signed-off-by: Andi Shyti --- Documentation/devicetree/bindings/i2c/renesas,riic.yaml | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/Documentation/devicetree/bindings/i2c/renesas,riic.yaml b/Documentation/devicetree/bindings/i2c/renesas,riic.yaml index 505a8ec92266..1b7fed232642 100644 --- a/Documentation/devicetree/bindings/i2c/renesas,riic.yaml +++ b/Documentation/devicetree/bindings/i2c/renesas,riic.yaml @@ -26,7 +26,9 @@ properties: - const: renesas,riic-rz # RZ/A or RZ/G2L - items: - - const: renesas,riic-r9a08g045 # RZ/G3S + - enum: + - renesas,riic-r9a08g045 # RZ/G3S + - renesas,riic-r9a09g047 # RZ/G3E - const: renesas,riic-r9a09g057 # RZ/V2H(P) - const: renesas,riic-r9a09g057 # RZ/V2H(P) @@ -89,6 +91,7 @@ if: - renesas,riic-r9a07g043 - renesas,riic-r9a07g044 - renesas,riic-r9a07g054 + - renesas,riic-r9a09g057 then: required: - resets From 4d51ac5b9b4e175abda13617a48639406d3ed7b1 Mon Sep 17 00:00:00 2001 From: Andi Shyti Date: Tue, 10 Dec 2024 23:16:41 +0100 Subject: [PATCH 178/196] i2c: nomadik: Enable compile testing for the Nomadik driver The AMBA bus header files do not generate compilation errors even when the AMBA bus is not enabled in Kconfig via ARM_AMBA. This allows the Nomadik driver to be compiled for testing purposes with the COMPILE_TEST option enabled. Cc: Linus Walleij Link: https://lore.kernel.org/r/20241210221641.2661577-1-andi.shyti@kernel.org Signed-off-by: Andi Shyti --- drivers/i2c/busses/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig index ceb3ecdf884b..ae4772faa876 100644 --- a/drivers/i2c/busses/Kconfig +++ b/drivers/i2c/busses/Kconfig @@ -910,7 +910,7 @@ config I2C_MXS config I2C_NOMADIK tristate "ST-Ericsson Nomadik/Ux500 I2C Controller" - depends on ARM_AMBA + depends on ARM_AMBA || COMPILE_TEST help If you say yes to this option, support will be included for the I2C interface from ST-Ericsson's Nomadik and Ux500 architectures, From f524707b44c96254c045dfbfe998f1a33e8595f5 Mon Sep 17 00:00:00 2001 From: Geert Uytterhoeven Date: Tue, 3 Dec 2024 17:19:50 +0100 Subject: [PATCH 179/196] i2c: I2C_BRCMSTB should not default to y when compile-testing Merely enabling compile-testing should not enable additional functionality. Signed-off-by: Geert Uytterhoeven Acked-by: Florian Fainelli Link: https://lore.kernel.org/r/36b2923241f285595f43bb9565da352fd366c63e.1733242595.git.geert+renesas@glider.be Signed-off-by: Andi Shyti --- drivers/i2c/busses/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig index ae4772faa876..eec95c724b25 100644 --- a/drivers/i2c/busses/Kconfig +++ b/drivers/i2c/busses/Kconfig @@ -503,7 +503,7 @@ config I2C_BRCMSTB tristate "BRCM Settop/DSL I2C controller" depends on ARCH_BCM2835 || ARCH_BCMBCA || ARCH_BRCMSTB || \ BMIPS_GENERIC || COMPILE_TEST - default y + default ARCH_BCM2835 || ARCH_BCMBCA || ARCH_BRCMSTB || BMIPS_GENERIC help If you say yes to this option, support will be included for the I2C interface on the Broadcom Settop/DSL SoCs. From 2ee409c6a12d74cd1b16ca3690baa7d5ffffeb7d Mon Sep 17 00:00:00 2001 From: Manikanta Guntupalli Date: Tue, 10 Dec 2024 15:22:41 +0530 Subject: [PATCH 180/196] i2c: xiic: Relocate xiic_i2c_runtime_suspend and xiic_i2c_runtime_resume to facilitate atomic mode Relocate xiic_i2c_runtime_suspend and xiic_i2c_runtime_resume functions to avoid prototype statements in atomic mode changes. Signed-off-by: Manikanta Guntupalli Acked-by: Michal Simek Link: https://lore.kernel.org/r/20241210095242.1982770-2-manikanta.guntupalli@amd.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-xiic.c | 46 +++++++++++++++++------------------ 1 file changed, 23 insertions(+), 23 deletions(-) diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c index c4d3eb02da09..f48e73f905e5 100644 --- a/drivers/i2c/busses/i2c-xiic.c +++ b/drivers/i2c/busses/i2c-xiic.c @@ -238,6 +238,29 @@ static const struct timing_regs timing_reg_values[] = { static int xiic_start_xfer(struct xiic_i2c *i2c, struct i2c_msg *msgs, int num); static void __xiic_start_xfer(struct xiic_i2c *i2c); +static int __maybe_unused xiic_i2c_runtime_suspend(struct device *dev) +{ + struct xiic_i2c *i2c = dev_get_drvdata(dev); + + clk_disable(i2c->clk); + + return 0; +} + +static int __maybe_unused xiic_i2c_runtime_resume(struct device *dev) +{ + struct xiic_i2c *i2c = dev_get_drvdata(dev); + int ret; + + ret = clk_enable(i2c->clk); + if (ret) { + dev_err(dev, "Cannot enable clock.\n"); + return ret; + } + + return 0; +} + /* * For the register read and write functions, a little-endian and big-endian * version are necessary. Endianness is detected during the probe function. @@ -1365,29 +1388,6 @@ static void xiic_i2c_remove(struct platform_device *pdev) pm_runtime_dont_use_autosuspend(&pdev->dev); } -static int __maybe_unused xiic_i2c_runtime_suspend(struct device *dev) -{ - struct xiic_i2c *i2c = dev_get_drvdata(dev); - - clk_disable(i2c->clk); - - return 0; -} - -static int __maybe_unused xiic_i2c_runtime_resume(struct device *dev) -{ - struct xiic_i2c *i2c = dev_get_drvdata(dev); - int ret; - - ret = clk_enable(i2c->clk); - if (ret) { - dev_err(dev, "Cannot enable clock.\n"); - return ret; - } - - return 0; -} - static const struct dev_pm_ops xiic_dev_pm_ops = { SET_RUNTIME_PM_OPS(xiic_i2c_runtime_suspend, xiic_i2c_runtime_resume, NULL) From e22daca457d5bec579f62ba84e826a51c26cce31 Mon Sep 17 00:00:00 2001 From: Manikanta Guntupalli Date: Tue, 10 Dec 2024 15:22:42 +0530 Subject: [PATCH 181/196] i2c: xiic: Add atomic transfer support Rework the read and write code paths in the driver to support operation in atomic contexts. Similar changes have been implemented in other drivers, including: commit 3a5ee18d2a32 ("i2c: imx: implement master_xfer_atomic callback") commit 445094c8a9fb ("i2c: exynos5: add support for atomic transfers") commit ede2299f7101 ("i2c: tegra: Support atomic transfers") commit fe402bd09049 ("i2c: meson: implement the master_xfer_atomic callback") Signed-off-by: Manikanta Guntupalli Acked-by: Michal Simek Link: https://lore.kernel.org/r/20241210095242.1982770-3-manikanta.guntupalli@amd.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-xiic.c | 241 +++++++++++++++++++++++++++++----- 1 file changed, 207 insertions(+), 34 deletions(-) diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c index f48e73f905e5..dc1e46d834dc 100644 --- a/drivers/i2c/busses/i2c-xiic.c +++ b/drivers/i2c/busses/i2c-xiic.c @@ -30,6 +30,8 @@ #include #include #include +#include +#include #define DRIVER_NAME "xiic-i2c" #define DYNAMIC_MODE_READ_BROKEN_BIT BIT(0) @@ -74,6 +76,9 @@ enum i2c_scl_freq { * @smbus_block_read: Flag to handle block read * @input_clk: Input clock to I2C controller * @i2c_clk: I2C SCL frequency + * @atomic: Mode of transfer + * @atomic_lock: Lock for atomic transfer mode + * @atomic_xfer_state: See STATE_ */ struct xiic_i2c { struct device *dev; @@ -96,6 +101,9 @@ struct xiic_i2c { bool smbus_block_read; unsigned long input_clk; unsigned int i2c_clk; + bool atomic; + spinlock_t atomic_lock; /* Lock for atomic transfer mode */ + enum xilinx_i2c_state atomic_xfer_state; }; struct xiic_version_data { @@ -224,6 +232,8 @@ static const struct timing_regs timing_reg_values[] = { #define XIIC_I2C_TIMEOUT (msecs_to_jiffies(1000)) /* timeout waiting for the controller finish transfers */ #define XIIC_XFER_TIMEOUT (msecs_to_jiffies(10000)) +/* timeout waiting for the controller finish transfers in micro seconds */ +#define XIIC_XFER_TIMEOUT_US 10000000 /* * The following constant is used for the device global interrupt enable @@ -238,7 +248,7 @@ static const struct timing_regs timing_reg_values[] = { static int xiic_start_xfer(struct xiic_i2c *i2c, struct i2c_msg *msgs, int num); static void __xiic_start_xfer(struct xiic_i2c *i2c); -static int __maybe_unused xiic_i2c_runtime_suspend(struct device *dev) +static int xiic_i2c_runtime_suspend(struct device *dev) { struct xiic_i2c *i2c = dev_get_drvdata(dev); @@ -247,7 +257,7 @@ static int __maybe_unused xiic_i2c_runtime_suspend(struct device *dev) return 0; } -static int __maybe_unused xiic_i2c_runtime_resume(struct device *dev) +static int xiic_i2c_runtime_resume(struct device *dev) { struct xiic_i2c *i2c = dev_get_drvdata(dev); int ret; @@ -397,9 +407,10 @@ static int xiic_setclk(struct xiic_i2c *i2c) unsigned int index = 0; u32 reg_val; - dev_dbg(i2c->adap.dev.parent, - "%s entry, i2c->input_clk: %ld, i2c->i2c_clk: %d\n", - __func__, i2c->input_clk, i2c->i2c_clk); + if (!i2c->atomic) + dev_dbg(i2c->adap.dev.parent, + "%s entry, i2c->input_clk: %ld, i2c->i2c_clk: %d\n", + __func__, i2c->input_clk, i2c->i2c_clk); /* If not specified in DT, do not configure in SW. Rely only on Vivado design */ if (!i2c->i2c_clk || !i2c->input_clk) @@ -490,7 +501,8 @@ static int xiic_reinit(struct xiic_i2c *i2c) return ret; /* Enable interrupts */ - xiic_setreg32(i2c, XIIC_DGIER_OFFSET, XIIC_GINTR_ENABLE_MASK); + if (!i2c->atomic) + xiic_setreg32(i2c, XIIC_DGIER_OFFSET, XIIC_GINTR_ENABLE_MASK); xiic_irq_clr_en(i2c, XIIC_INTR_ARB_LOST_MASK); @@ -572,11 +584,12 @@ static void xiic_read_rx(struct xiic_i2c *i2c) bytes_in_fifo = xiic_getreg8(i2c, XIIC_RFO_REG_OFFSET) + 1; - dev_dbg(i2c->adap.dev.parent, - "%s entry, bytes in fifo: %d, rem: %d, SR: 0x%x, CR: 0x%x\n", - __func__, bytes_in_fifo, xiic_rx_space(i2c), - xiic_getreg8(i2c, XIIC_SR_REG_OFFSET), - xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); + if (!i2c->atomic) + dev_dbg(i2c->adap.dev.parent, + "%s entry, bytes in fifo: %d, rem: %d, SR: 0x%x, CR: 0x%x\n", + __func__, bytes_in_fifo, xiic_rx_space(i2c), + xiic_getreg8(i2c, XIIC_SR_REG_OFFSET), + xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); if (bytes_in_fifo > xiic_rx_space(i2c)) bytes_in_fifo = xiic_rx_space(i2c); @@ -635,6 +648,26 @@ static void xiic_read_rx(struct xiic_i2c *i2c) } } +static bool xiic_error_check(struct xiic_i2c *i2c) +{ + bool status = false; + u32 pend, isr, ier; + + isr = xiic_getreg32(i2c, XIIC_IISR_OFFSET); + ier = xiic_getreg32(i2c, XIIC_IIER_OFFSET); + pend = isr & ier; + + if ((pend & XIIC_INTR_ARB_LOST_MASK) || + ((pend & XIIC_INTR_TX_ERROR_MASK) && + !(pend & XIIC_INTR_RX_FULL_MASK))) { + xiic_reinit(i2c); + status = true; + if (i2c->tx_msg || i2c->rx_msg) + i2c->atomic_xfer_state = STATE_ERROR; + } + return status; +} + static int xiic_tx_fifo_space(struct xiic_i2c *i2c) { /* return the actual space left in the FIFO */ @@ -648,8 +681,9 @@ static void xiic_fill_tx_fifo(struct xiic_i2c *i2c) len = (len > fifo_space) ? fifo_space : len; - dev_dbg(i2c->adap.dev.parent, "%s entry, len: %d, fifo space: %d\n", - __func__, len, fifo_space); + if (!i2c->atomic) + dev_dbg(i2c->adap.dev.parent, "%s entry, len: %d, fifo space: %d\n", + __func__, len, fifo_space); while (len--) { u16 data = i2c->tx_msg->buf[i2c->tx_pos++]; @@ -672,9 +706,13 @@ static void xiic_fill_tx_fifo(struct xiic_i2c *i2c) xiic_setreg8(i2c, XIIC_CR_REG_OFFSET, cr & ~XIIC_CR_MSMS_MASK); } - dev_dbg(i2c->adap.dev.parent, "%s TX STOP\n", __func__); + if (!i2c->atomic) + dev_dbg(i2c->adap.dev.parent, "%s TX STOP\n", __func__); } xiic_setreg16(i2c, XIIC_DTR_REG_OFFSET, data); + + if (i2c->atomic && xiic_error_check(i2c)) + return; } } @@ -877,22 +915,51 @@ static int xiic_wait_not_busy(struct xiic_i2c *i2c) */ err = xiic_bus_busy(i2c); while (err && tries--) { - msleep(1); + if (i2c->atomic) + udelay(1000); + else + usleep_range(1000, 1100); err = xiic_bus_busy(i2c); } return err; } +static void xiic_recv_atomic(struct xiic_i2c *i2c) +{ + while (xiic_rx_space(i2c)) { + if (xiic_getreg32(i2c, XIIC_IISR_OFFSET) & XIIC_INTR_RX_FULL_MASK) { + xiic_read_rx(i2c); + + /* Clear Rx full and Tx error interrupts. */ + xiic_irq_clr_en(i2c, XIIC_INTR_RX_FULL_MASK | + XIIC_INTR_TX_ERROR_MASK); + } + if (xiic_error_check(i2c)) + return; + } + + i2c->rx_msg = NULL; + xiic_irq_clr_en(i2c, XIIC_INTR_TX_ERROR_MASK); + + /* send next message if this wasn't the last. */ + if (i2c->nmsgs > 1) { + i2c->nmsgs--; + i2c->tx_msg++; + __xiic_start_xfer(i2c); + } +} + static void xiic_start_recv(struct xiic_i2c *i2c) { u16 rx_watermark; u8 cr = 0, rfd_set = 0; struct i2c_msg *msg = i2c->rx_msg = i2c->tx_msg; - dev_dbg(i2c->adap.dev.parent, "%s entry, ISR: 0x%x, CR: 0x%x\n", - __func__, xiic_getreg32(i2c, XIIC_IISR_OFFSET), - xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); + if (!i2c->atomic) + dev_dbg(i2c->adap.dev.parent, "%s entry, ISR: 0x%x, CR: 0x%x\n", + __func__, xiic_getreg32(i2c, XIIC_IISR_OFFSET), + xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); /* Disable Tx interrupts */ xiic_irq_dis(i2c, XIIC_INTR_TX_HALF_MASK | XIIC_INTR_TX_EMPTY_MASK); @@ -990,9 +1057,10 @@ static void xiic_start_recv(struct xiic_i2c *i2c) XIIC_CR_MSMS_MASK) & ~(XIIC_CR_DIR_IS_TX_MASK)); } - dev_dbg(i2c->adap.dev.parent, "%s end, ISR: 0x%x, CR: 0x%x\n", - __func__, xiic_getreg32(i2c, XIIC_IISR_OFFSET), - xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); + if (!i2c->atomic) + dev_dbg(i2c->adap.dev.parent, "%s end, ISR: 0x%x, CR: 0x%x\n", + __func__, xiic_getreg32(i2c, XIIC_IISR_OFFSET), + xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); } if (i2c->nmsgs == 1) @@ -1002,10 +1070,55 @@ static void xiic_start_recv(struct xiic_i2c *i2c) /* the message is tx:ed */ i2c->tx_pos = msg->len; - /* Enable interrupts */ - xiic_setreg32(i2c, XIIC_DGIER_OFFSET, XIIC_GINTR_ENABLE_MASK); - i2c->prev_msg_tx = false; + + /* Enable interrupts */ + if (!i2c->atomic) + xiic_setreg32(i2c, XIIC_DGIER_OFFSET, XIIC_GINTR_ENABLE_MASK); + else + xiic_recv_atomic(i2c); +} + +static void xiic_send_rem_atomic(struct xiic_i2c *i2c) +{ + while (xiic_tx_space(i2c)) { + if (xiic_tx_fifo_space(i2c)) { + u16 data; + + data = i2c->tx_msg->buf[i2c->tx_pos]; + i2c->tx_pos++; + if (!xiic_tx_space(i2c) && i2c->nmsgs == 1) { + /* last message in transfer -> STOP */ + if (i2c->dynamic) { + data |= XIIC_TX_DYN_STOP_MASK; + } else { + u8 cr; + int status; + + /* Wait till FIFO is empty so STOP is sent last */ + status = xiic_wait_tx_empty(i2c); + if (status) + return; + + /* Write to CR to stop */ + cr = xiic_getreg8(i2c, XIIC_CR_REG_OFFSET); + xiic_setreg8(i2c, XIIC_CR_REG_OFFSET, cr & + ~XIIC_CR_MSMS_MASK); + } + } + xiic_setreg16(i2c, XIIC_DTR_REG_OFFSET, data); + } + if (xiic_error_check(i2c)) + return; + } + + if (i2c->nmsgs > 1) { + i2c->nmsgs--; + i2c->tx_msg++; + __xiic_start_xfer(i2c); + } else { + xiic_irq_dis(i2c, XIIC_INTR_TX_HALF_MASK); + } } static void xiic_start_send(struct xiic_i2c *i2c) @@ -1014,11 +1127,13 @@ static void xiic_start_send(struct xiic_i2c *i2c) u16 data; struct i2c_msg *msg = i2c->tx_msg; - dev_dbg(i2c->adap.dev.parent, "%s entry, msg: %p, len: %d", - __func__, msg, msg->len); - dev_dbg(i2c->adap.dev.parent, "%s entry, ISR: 0x%x, CR: 0x%x\n", - __func__, xiic_getreg32(i2c, XIIC_IISR_OFFSET), - xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); + if (!i2c->atomic) { + dev_dbg(i2c->adap.dev.parent, "%s entry, msg: %p, len: %d", + __func__, msg, msg->len); + dev_dbg(i2c->adap.dev.parent, "%s entry, ISR: 0x%x, CR: 0x%x\n", + __func__, xiic_getreg32(i2c, XIIC_IISR_OFFSET), + xiic_getreg8(i2c, XIIC_CR_REG_OFFSET)); + } if (i2c->dynamic) { /* write the address */ @@ -1083,19 +1198,27 @@ static void xiic_start_send(struct xiic_i2c *i2c) XIIC_INTR_TX_ERROR_MASK | XIIC_INTR_BNB_MASK); } + i2c->prev_msg_tx = true; + + if (i2c->atomic && !i2c->atomic_xfer_state) + xiic_send_rem_atomic(i2c); } static void __xiic_start_xfer(struct xiic_i2c *i2c) { int fifo_space = xiic_tx_fifo_space(i2c); - dev_dbg(i2c->adap.dev.parent, "%s entry, msg: %p, fifos space: %d\n", - __func__, i2c->tx_msg, fifo_space); + if (!i2c->atomic) + dev_dbg(i2c->adap.dev.parent, "%s entry, msg: %p, fifos space: %d\n", + __func__, i2c->tx_msg, fifo_space); if (!i2c->tx_msg) return; + if (i2c->atomic && xiic_error_check(i2c)) + return; + i2c->rx_pos = 0; i2c->tx_pos = 0; i2c->state = STATE_START; @@ -1112,7 +1235,10 @@ static int xiic_start_xfer(struct xiic_i2c *i2c, struct i2c_msg *msgs, int num) bool broken_read, max_read_len, smbus_blk_read; int ret, count; - mutex_lock(&i2c->lock); + if (i2c->atomic) + spin_lock(&i2c->atomic_lock); + else + mutex_lock(&i2c->lock); if (i2c->tx_msg || i2c->rx_msg) { dev_err(i2c->adap.dev.parent, @@ -1121,6 +1247,8 @@ static int xiic_start_xfer(struct xiic_i2c *i2c, struct i2c_msg *msgs, int num) goto out; } + i2c->atomic_xfer_state = STATE_DONE; + /* In single master mode bus can only be busy, when in use by this * driver. If the register indicates bus being busy for some reason we * should ignore it, since bus will never be released and i2c will be @@ -1147,7 +1275,9 @@ static int xiic_start_xfer(struct xiic_i2c *i2c, struct i2c_msg *msgs, int num) i2c->tx_msg = msgs; i2c->rx_msg = NULL; i2c->nmsgs = num; - init_completion(&i2c->completion); + + if (!i2c->atomic) + init_completion(&i2c->completion); /* Decide standard mode or Dynamic mode */ i2c->dynamic = true; @@ -1182,7 +1312,10 @@ static int xiic_start_xfer(struct xiic_i2c *i2c, struct i2c_msg *msgs, int num) __xiic_start_xfer(i2c); out: - mutex_unlock(&i2c->lock); + if (i2c->atomic) + spin_unlock(&i2c->atomic_lock); + else + mutex_unlock(&i2c->lock); return ret; } @@ -1221,6 +1354,44 @@ static int xiic_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num) return err; } +static int xiic_xfer_atomic(struct i2c_adapter *adap, struct i2c_msg *msgs, int num) +{ + struct xiic_i2c *i2c = i2c_get_adapdata(adap); + u32 status_reg; + int err; + + err = xiic_i2c_runtime_resume(i2c->dev); + if (err) + return err; + + i2c->atomic = true; + err = xiic_start_xfer(i2c, msgs, num); + if (err < 0) + return err; + + err = readl_poll_timeout_atomic(i2c->base + XIIC_SR_REG_OFFSET, + status_reg, !(status_reg & XIIC_SR_BUS_BUSY_MASK), + 1, XIIC_XFER_TIMEOUT_US); + + if (err) /* Timeout */ + err = -ETIMEDOUT; + + spin_lock(&i2c->atomic_lock); + if (err || i2c->state) { + i2c->tx_msg = NULL; + i2c->rx_msg = NULL; + i2c->nmsgs = 0; + } + + err = (i2c->atomic_xfer_state == STATE_DONE) ? num : -EIO; + spin_unlock(&i2c->atomic_lock); + + i2c->atomic = false; + xiic_i2c_runtime_suspend(i2c->dev); + + return err; +} + static u32 xiic_func(struct i2c_adapter *adap) { return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL | I2C_FUNC_SMBUS_BLOCK_DATA; @@ -1228,6 +1399,7 @@ static u32 xiic_func(struct i2c_adapter *adap) static const struct i2c_algorithm xiic_algorithm = { .master_xfer = xiic_xfer, + .master_xfer_atomic = xiic_xfer_atomic, .functionality = xiic_func, }; @@ -1291,6 +1463,7 @@ static int xiic_i2c_probe(struct platform_device *pdev) DRIVER_NAME " %s", pdev->name); mutex_init(&i2c->lock); + spin_lock_init(&i2c->atomic_lock); i2c->clk = devm_clk_get_enabled(&pdev->dev, NULL); if (IS_ERR(i2c->clk)) From 525defdb608a153108c434e127b0bd1d83a5d8c1 Mon Sep 17 00:00:00 2001 From: Chen Ni Date: Tue, 26 Nov 2024 10:38:39 +0800 Subject: [PATCH 182/196] i2c: isch: Convert comma to semicolon Replace comma between expressions with semicolons. Using a ',' in place of a ';' can have unintended side effects. Although that is not the case here, it is seems best to use ';' unless ',' is intended. Found by inspection. No functional change intended. Compile tested only. Signed-off-by: Chen Ni Link: https://lore.kernel.org/r/20241126023839.251922-1-nichen@iscas.ac.cn Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-isch.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/i2c/busses/i2c-isch.c b/drivers/i2c/busses/i2c-isch.c index 2b3b65ef2900..a2ac992f9cb0 100644 --- a/drivers/i2c/busses/i2c-isch.c +++ b/drivers/i2c/busses/i2c-isch.c @@ -291,9 +291,9 @@ static int smbus_sch_probe(struct platform_device *pdev) /* Set up the sysfs linkage to our parent device */ priv->adapter.dev.parent = dev; - priv->adapter.owner = THIS_MODULE, - priv->adapter.class = I2C_CLASS_HWMON, - priv->adapter.algo = &smbus_algorithm, + priv->adapter.owner = THIS_MODULE; + priv->adapter.class = I2C_CLASS_HWMON; + priv->adapter.algo = &smbus_algorithm; snprintf(priv->adapter.name, sizeof(priv->adapter.name), "SMBus SCH adapter at %04x", (unsigned short)res->start); From a09c8b3f9047959e60dde90c51dcca17f66778c1 Mon Sep 17 00:00:00 2001 From: Carlos Song Date: Mon, 25 Nov 2024 22:29:09 +0800 Subject: [PATCH 183/196] i2c: imx-lpi2c: add eDMA mode support for LPI2C Add eDMA mode support for LPI2C. There are some differences between TX DMA mode and RX DMA mode. LPI2C MTDR register is Controller Transmit Data Register. When lpi2c send data, it is tx cmd register and tx data fifo. When lpi2c receive data, it is just a rx cmd register. LPI2C MRDR register is Controller Receive Data Register, received data are stored in this. MTDR[8:10] is CMD field and MTDR[0:7] is DATA filed. +-----------+-------------------------------+ | C M D | D A T A | +---+---+---+---+---+---+---+---+---+---+---+ | 10| 9 | 8 | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 | +---+---+---+---+---+---+---+---+---+---+---+ MRDR is Controller Receive Data Register. MRDR[0:7] is DATA filed. +-------------------------------+ | D A T A | +---+---+---+---+---+---+---+---+ | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 | +---+---+---+---+---+---+---+---+ When the LPI2C controller needs to send data, tx cmd and 8-bit data should be written into MTDR: CMD: 000b: Transmit the value in DATA[7:0]. DATA: 8-bit data. If lpi2c controller needs to send N 8-bit data, just write N times (CMD(W) + DATA(u8)) to MTDR. When the LPI2C controller needs to receive data, rx cmd should be written into MTDR, the received data will be stored in the MRDR. MTDR(CMD): 001b: Receive (DATA[7:0] + 1) 8-bit data. MTDR(DATA): byte counter. MRDR(DATA): 8-bit data. So when lpi2c controller needs to receive N 8-bit data, 1. N <= 256: Write 1 time (CMD(R) + BYTE COUNT(N-1)) into MTDR and receive data from MRDR. 2. N > 256: Write N/256 times (CMD(R) + BYTE COUNT(255)) + 1 time (CMD(R) + BYTE COUNT(N%256)) into MTDR and receive data from MRDR. Due to these differences, when lpi2c is in DMA TX mode, only enable TX channel to send data. But when lpi2c is in DMA RX mode, TX and RX channel are both enabled, TX channel is used to send RX cmd and RX channel is used to receive data. Signed-off-by: Carlos Song Signed-off-by: Frank Li Link: https://lore.kernel.org/r/20241125142909.1613245-1-carlos.song@nxp.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-imx-lpi2c.c | 524 ++++++++++++++++++++++++++++- 1 file changed, 517 insertions(+), 7 deletions(-) diff --git a/drivers/i2c/busses/i2c-imx-lpi2c.c b/drivers/i2c/busses/i2c-imx-lpi2c.c index 8adf2963d764..9778dc8f6d36 100644 --- a/drivers/i2c/busses/i2c-imx-lpi2c.c +++ b/drivers/i2c/busses/i2c-imx-lpi2c.c @@ -8,6 +8,8 @@ #include #include #include +#include +#include #include #include #include @@ -29,6 +31,7 @@ #define LPI2C_MCR 0x10 /* i2c contrl register */ #define LPI2C_MSR 0x14 /* i2c status register */ #define LPI2C_MIER 0x18 /* i2c interrupt enable */ +#define LPI2C_MDER 0x1C /* i2c DMA enable */ #define LPI2C_MCFGR0 0x20 /* i2c master configuration */ #define LPI2C_MCFGR1 0x24 /* i2c master configuration */ #define LPI2C_MCFGR2 0x28 /* i2c master configuration */ @@ -70,11 +73,14 @@ #define MCFGR1_AUTOSTOP BIT(8) #define MCFGR1_IGNACK BIT(9) #define MRDR_RXEMPTY BIT(14) +#define MDER_TDDE BIT(0) +#define MDER_RDDE BIT(1) #define I2C_CLK_RATIO 2 #define CHUNK_DATA 256 #define I2C_PM_TIMEOUT 10 /* ms */ +#define I2C_DMA_THRESHOLD 8 /* bytes */ enum lpi2c_imx_mode { STANDARD, /* 100+Kbps */ @@ -91,6 +97,24 @@ enum lpi2c_imx_pincfg { FOUR_PIN_PP, }; +struct lpi2c_imx_dma { + bool using_pio_mode; + u8 rx_cmd_buf_len; + u8 *dma_buf; + u16 *rx_cmd_buf; + unsigned int dma_len; + unsigned int tx_burst_num; + unsigned int rx_burst_num; + unsigned long dma_msg_flag; + resource_size_t phy_addr; + dma_addr_t dma_tx_addr; + dma_addr_t dma_addr; + enum dma_data_direction dma_data_dir; + enum dma_transfer_direction dma_transfer_dir; + struct dma_chan *chan_tx; + struct dma_chan *chan_rx; +}; + struct lpi2c_imx_struct { struct i2c_adapter adapter; int num_clks; @@ -108,6 +132,8 @@ struct lpi2c_imx_struct { unsigned int rxfifosize; enum lpi2c_imx_mode mode; struct i2c_bus_recovery_info rinfo; + bool can_use_dma; + struct lpi2c_imx_dma *dma; }; static void lpi2c_imx_intctrl(struct lpi2c_imx_struct *lpi2c_imx, @@ -305,7 +331,7 @@ static int lpi2c_imx_master_disable(struct lpi2c_imx_struct *lpi2c_imx) return 0; } -static int lpi2c_imx_msg_complete(struct lpi2c_imx_struct *lpi2c_imx) +static int lpi2c_imx_pio_msg_complete(struct lpi2c_imx_struct *lpi2c_imx) { unsigned long time_left; @@ -451,6 +477,425 @@ static void lpi2c_imx_read(struct lpi2c_imx_struct *lpi2c_imx, lpi2c_imx_intctrl(lpi2c_imx, MIER_RDIE | MIER_NDIE); } +static bool is_use_dma(struct lpi2c_imx_struct *lpi2c_imx, struct i2c_msg *msg) +{ + if (!lpi2c_imx->can_use_dma) + return false; + + /* + * When the length of data is less than I2C_DMA_THRESHOLD, + * cpu mode is used directly to avoid low performance. + */ + return !(msg->len < I2C_DMA_THRESHOLD); +} + +static int lpi2c_imx_pio_xfer(struct lpi2c_imx_struct *lpi2c_imx, + struct i2c_msg *msg) +{ + reinit_completion(&lpi2c_imx->complete); + + if (msg->flags & I2C_M_RD) + lpi2c_imx_read(lpi2c_imx, msg); + else + lpi2c_imx_write(lpi2c_imx, msg); + + return lpi2c_imx_pio_msg_complete(lpi2c_imx); +} + +static int lpi2c_imx_dma_timeout_calculate(struct lpi2c_imx_struct *lpi2c_imx) +{ + unsigned long time = 0; + + time = 8 * lpi2c_imx->dma->dma_len * 1000 / lpi2c_imx->bitrate; + + /* Add extra second for scheduler related activities */ + time += 1; + + /* Double calculated time */ + return msecs_to_jiffies(time * MSEC_PER_SEC); +} + +static int lpi2c_imx_alloc_rx_cmd_buf(struct lpi2c_imx_struct *lpi2c_imx) +{ + struct lpi2c_imx_dma *dma = lpi2c_imx->dma; + u16 rx_remain = dma->dma_len; + int cmd_num; + u16 temp; + + /* + * Calculate the number of rx command words via the DMA TX channel + * writing into command register based on the i2c msg len, and build + * the rx command words buffer. + */ + cmd_num = DIV_ROUND_UP(rx_remain, CHUNK_DATA); + dma->rx_cmd_buf = kcalloc(cmd_num, sizeof(u16), GFP_KERNEL); + dma->rx_cmd_buf_len = cmd_num * sizeof(u16); + + if (!dma->rx_cmd_buf) { + dev_err(&lpi2c_imx->adapter.dev, "Alloc RX cmd buffer failed\n"); + return -ENOMEM; + } + + for (int i = 0; i < cmd_num ; i++) { + temp = rx_remain > CHUNK_DATA ? CHUNK_DATA - 1 : rx_remain - 1; + temp |= (RECV_DATA << 8); + rx_remain -= CHUNK_DATA; + dma->rx_cmd_buf[i] = temp; + } + + return 0; +} + +static int lpi2c_imx_dma_msg_complete(struct lpi2c_imx_struct *lpi2c_imx) +{ + unsigned long time_left, time; + + time = lpi2c_imx_dma_timeout_calculate(lpi2c_imx); + time_left = wait_for_completion_timeout(&lpi2c_imx->complete, time); + if (time_left == 0) { + dev_err(&lpi2c_imx->adapter.dev, "I/O Error in DMA Data Transfer\n"); + return -ETIMEDOUT; + } + + return 0; +} + +static void lpi2c_dma_unmap(struct lpi2c_imx_dma *dma) +{ + struct dma_chan *chan = dma->dma_data_dir == DMA_FROM_DEVICE + ? dma->chan_rx : dma->chan_tx; + + dma_unmap_single(chan->device->dev, dma->dma_addr, + dma->dma_len, dma->dma_data_dir); + + dma->dma_data_dir = DMA_NONE; +} + +static void lpi2c_cleanup_rx_cmd_dma(struct lpi2c_imx_dma *dma) +{ + dmaengine_terminate_sync(dma->chan_tx); + dma_unmap_single(dma->chan_tx->device->dev, dma->dma_tx_addr, + dma->rx_cmd_buf_len, DMA_TO_DEVICE); +} + +static void lpi2c_cleanup_dma(struct lpi2c_imx_dma *dma) +{ + if (dma->dma_data_dir == DMA_FROM_DEVICE) + dmaengine_terminate_sync(dma->chan_rx); + else if (dma->dma_data_dir == DMA_TO_DEVICE) + dmaengine_terminate_sync(dma->chan_tx); + + lpi2c_dma_unmap(dma); +} + +static void lpi2c_dma_callback(void *data) +{ + struct lpi2c_imx_struct *lpi2c_imx = (struct lpi2c_imx_struct *)data; + + complete(&lpi2c_imx->complete); +} + +static int lpi2c_dma_rx_cmd_submit(struct lpi2c_imx_struct *lpi2c_imx) +{ + struct dma_async_tx_descriptor *rx_cmd_desc; + struct lpi2c_imx_dma *dma = lpi2c_imx->dma; + struct dma_chan *txchan = dma->chan_tx; + dma_cookie_t cookie; + + dma->dma_tx_addr = dma_map_single(txchan->device->dev, + dma->rx_cmd_buf, dma->rx_cmd_buf_len, + DMA_TO_DEVICE); + if (dma_mapping_error(txchan->device->dev, dma->dma_tx_addr)) { + dev_err(&lpi2c_imx->adapter.dev, "DMA map failed, use pio\n"); + return -EINVAL; + } + + rx_cmd_desc = dmaengine_prep_slave_single(txchan, dma->dma_tx_addr, + dma->rx_cmd_buf_len, DMA_MEM_TO_DEV, + DMA_PREP_INTERRUPT | DMA_CTRL_ACK); + if (!rx_cmd_desc) { + dev_err(&lpi2c_imx->adapter.dev, "DMA prep slave sg failed, use pio\n"); + goto desc_prepare_err_exit; + } + + cookie = dmaengine_submit(rx_cmd_desc); + if (dma_submit_error(cookie)) { + dev_err(&lpi2c_imx->adapter.dev, "submitting DMA failed, use pio\n"); + goto submit_err_exit; + } + + dma_async_issue_pending(txchan); + + return 0; + +desc_prepare_err_exit: + dma_unmap_single(txchan->device->dev, dma->dma_tx_addr, + dma->rx_cmd_buf_len, DMA_TO_DEVICE); + return -EINVAL; + +submit_err_exit: + dma_unmap_single(txchan->device->dev, dma->dma_tx_addr, + dma->rx_cmd_buf_len, DMA_TO_DEVICE); + dmaengine_desc_free(rx_cmd_desc); + return -EINVAL; +} + +static int lpi2c_dma_submit(struct lpi2c_imx_struct *lpi2c_imx) +{ + struct lpi2c_imx_dma *dma = lpi2c_imx->dma; + struct dma_async_tx_descriptor *desc; + struct dma_chan *chan; + dma_cookie_t cookie; + + if (dma->dma_msg_flag & I2C_M_RD) { + chan = dma->chan_rx; + dma->dma_data_dir = DMA_FROM_DEVICE; + dma->dma_transfer_dir = DMA_DEV_TO_MEM; + } else { + chan = dma->chan_tx; + dma->dma_data_dir = DMA_TO_DEVICE; + dma->dma_transfer_dir = DMA_MEM_TO_DEV; + } + + dma->dma_addr = dma_map_single(chan->device->dev, + dma->dma_buf, dma->dma_len, dma->dma_data_dir); + if (dma_mapping_error(chan->device->dev, dma->dma_addr)) { + dev_err(&lpi2c_imx->adapter.dev, "DMA map failed, use pio\n"); + return -EINVAL; + } + + desc = dmaengine_prep_slave_single(chan, dma->dma_addr, + dma->dma_len, dma->dma_transfer_dir, + DMA_PREP_INTERRUPT | DMA_CTRL_ACK); + if (!desc) { + dev_err(&lpi2c_imx->adapter.dev, "DMA prep slave sg failed, use pio\n"); + goto desc_prepare_err_exit; + } + + reinit_completion(&lpi2c_imx->complete); + desc->callback = lpi2c_dma_callback; + desc->callback_param = lpi2c_imx; + + cookie = dmaengine_submit(desc); + if (dma_submit_error(cookie)) { + dev_err(&lpi2c_imx->adapter.dev, "submitting DMA failed, use pio\n"); + goto submit_err_exit; + } + + /* Can't switch to PIO mode when DMA have started transfer */ + dma->using_pio_mode = false; + + dma_async_issue_pending(chan); + + return 0; + +desc_prepare_err_exit: + lpi2c_dma_unmap(dma); + return -EINVAL; + +submit_err_exit: + lpi2c_dma_unmap(dma); + dmaengine_desc_free(desc); + return -EINVAL; +} + +static int lpi2c_imx_find_max_burst_num(unsigned int fifosize, unsigned int len) +{ + unsigned int i; + + for (i = fifosize / 2; i > 0; i--) + if (!(len % i)) + break; + + return i; +} + +/* + * For a highest DMA efficiency, tx/rx burst number should be calculated according + * to the FIFO depth. + */ +static void lpi2c_imx_dma_burst_num_calculate(struct lpi2c_imx_struct *lpi2c_imx) +{ + struct lpi2c_imx_dma *dma = lpi2c_imx->dma; + unsigned int cmd_num; + + if (dma->dma_msg_flag & I2C_M_RD) { + /* + * One RX cmd word can trigger DMA receive no more than 256 bytes. + * The number of RX cmd words should be calculated based on the data + * length. + */ + cmd_num = DIV_ROUND_UP(dma->dma_len, CHUNK_DATA); + dma->tx_burst_num = lpi2c_imx_find_max_burst_num(lpi2c_imx->txfifosize, + cmd_num); + dma->rx_burst_num = lpi2c_imx_find_max_burst_num(lpi2c_imx->rxfifosize, + dma->dma_len); + } else { + dma->tx_burst_num = lpi2c_imx_find_max_burst_num(lpi2c_imx->txfifosize, + dma->dma_len); + } +} + +static int lpi2c_dma_config(struct lpi2c_imx_struct *lpi2c_imx) +{ + struct lpi2c_imx_dma *dma = lpi2c_imx->dma; + struct dma_slave_config rx = {}, tx = {}; + int ret; + + lpi2c_imx_dma_burst_num_calculate(lpi2c_imx); + + if (dma->dma_msg_flag & I2C_M_RD) { + tx.dst_addr = dma->phy_addr + LPI2C_MTDR; + tx.dst_addr_width = DMA_SLAVE_BUSWIDTH_2_BYTES; + tx.dst_maxburst = dma->tx_burst_num; + tx.direction = DMA_MEM_TO_DEV; + ret = dmaengine_slave_config(dma->chan_tx, &tx); + if (ret < 0) + return ret; + + rx.src_addr = dma->phy_addr + LPI2C_MRDR; + rx.src_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE; + rx.src_maxburst = dma->rx_burst_num; + rx.direction = DMA_DEV_TO_MEM; + ret = dmaengine_slave_config(dma->chan_rx, &rx); + if (ret < 0) + return ret; + } else { + tx.dst_addr = dma->phy_addr + LPI2C_MTDR; + tx.dst_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE; + tx.dst_maxburst = dma->tx_burst_num; + tx.direction = DMA_MEM_TO_DEV; + ret = dmaengine_slave_config(dma->chan_tx, &tx); + if (ret < 0) + return ret; + } + + return 0; +} + +static void lpi2c_dma_enable(struct lpi2c_imx_struct *lpi2c_imx) +{ + struct lpi2c_imx_dma *dma = lpi2c_imx->dma; + /* + * TX interrupt will be triggered when the number of words in + * the transmit FIFO is equal or less than TX watermark. + * RX interrupt will be triggered when the number of words in + * the receive FIFO is greater than RX watermark. + * In order to trigger the DMA interrupt, TX watermark should be + * set equal to the DMA TX burst number but RX watermark should + * be set less than the DMA RX burst number. + */ + if (dma->dma_msg_flag & I2C_M_RD) { + /* Set I2C TX/RX watermark */ + writel(dma->tx_burst_num | (dma->rx_burst_num - 1) << 16, + lpi2c_imx->base + LPI2C_MFCR); + /* Enable I2C DMA TX/RX function */ + writel(MDER_TDDE | MDER_RDDE, lpi2c_imx->base + LPI2C_MDER); + } else { + /* Set I2C TX watermark */ + writel(dma->tx_burst_num, lpi2c_imx->base + LPI2C_MFCR); + /* Enable I2C DMA TX function */ + writel(MDER_TDDE, lpi2c_imx->base + LPI2C_MDER); + } + + /* Enable NACK detected */ + lpi2c_imx_intctrl(lpi2c_imx, MIER_NDIE); +}; + +/* + * When lpi2c is in TX DMA mode we can use one DMA TX channel to write + * data word into TXFIFO, but in RX DMA mode it is different. + * + * The LPI2C MTDR register is a command data and transmit data register. + * Bits 8-10 are the command data field and Bits 0-7 are the transmit + * data field. When the LPI2C master needs to read data, the number of + * bytes to read should be set in the command field and RECV_DATA should + * be set into the command data field to receive (DATA[7:0] + 1) bytes. + * The recv data command word is made of RECV_DATA in the command data + * field and the number of bytes to read in transmit data field. When the + * length of data to be read exceeds 256 bytes, recv data command word + * needs to be written to TXFIFO multiple times. + * + * So when in RX DMA mode, the TX channel also must to be configured to + * send RX command words and the RX command word must be set in advance + * before transmitting. + */ +static int lpi2c_imx_dma_xfer(struct lpi2c_imx_struct *lpi2c_imx, + struct i2c_msg *msg) +{ + struct lpi2c_imx_dma *dma = lpi2c_imx->dma; + int ret; + + /* When DMA mode fails before transferring, CPU mode can be used. */ + dma->using_pio_mode = true; + + dma->dma_len = msg->len; + dma->dma_msg_flag = msg->flags; + dma->dma_buf = i2c_get_dma_safe_msg_buf(msg, I2C_DMA_THRESHOLD); + if (!dma->dma_buf) + return -ENOMEM; + + ret = lpi2c_dma_config(lpi2c_imx); + if (ret) { + dev_err(&lpi2c_imx->adapter.dev, "Failed to configure DMA (%d)\n", ret); + goto disable_dma; + } + + lpi2c_dma_enable(lpi2c_imx); + + ret = lpi2c_dma_submit(lpi2c_imx); + if (ret) { + dev_err(&lpi2c_imx->adapter.dev, "DMA submission failed (%d)\n", ret); + goto disable_dma; + } + + if (dma->dma_msg_flag & I2C_M_RD) { + ret = lpi2c_imx_alloc_rx_cmd_buf(lpi2c_imx); + if (ret) + goto disable_cleanup_data_dma; + + ret = lpi2c_dma_rx_cmd_submit(lpi2c_imx); + if (ret) + goto disable_cleanup_data_dma; + } + + ret = lpi2c_imx_dma_msg_complete(lpi2c_imx); + if (ret) + goto disable_cleanup_all_dma; + + /* When encountering NACK in transfer, clean up all DMA transfers */ + if ((readl(lpi2c_imx->base + LPI2C_MSR) & MSR_NDF) && !ret) { + ret = -EIO; + goto disable_cleanup_all_dma; + } + + if (dma->dma_msg_flag & I2C_M_RD) + dma_unmap_single(dma->chan_tx->device->dev, dma->dma_tx_addr, + dma->rx_cmd_buf_len, DMA_TO_DEVICE); + lpi2c_dma_unmap(dma); + + goto disable_dma; + +disable_cleanup_all_dma: + if (dma->dma_msg_flag & I2C_M_RD) + lpi2c_cleanup_rx_cmd_dma(dma); +disable_cleanup_data_dma: + lpi2c_cleanup_dma(dma); +disable_dma: + /* Disable I2C DMA function */ + writel(0, lpi2c_imx->base + LPI2C_MDER); + + if (dma->dma_msg_flag & I2C_M_RD) + kfree(dma->rx_cmd_buf); + + if (ret) + i2c_put_dma_safe_msg_buf(dma->dma_buf, msg, false); + else + i2c_put_dma_safe_msg_buf(dma->dma_buf, msg, true); + + return ret; +} + static int lpi2c_imx_xfer(struct i2c_adapter *adapter, struct i2c_msg *msgs, int num) { @@ -477,12 +922,14 @@ static int lpi2c_imx_xfer(struct i2c_adapter *adapter, lpi2c_imx->msglen = msgs[i].len; init_completion(&lpi2c_imx->complete); - if (msgs[i].flags & I2C_M_RD) - lpi2c_imx_read(lpi2c_imx, &msgs[i]); - else - lpi2c_imx_write(lpi2c_imx, &msgs[i]); + if (is_use_dma(lpi2c_imx, &msgs[i])) { + result = lpi2c_imx_dma_xfer(lpi2c_imx, &msgs[i]); + if (result && lpi2c_imx->dma->using_pio_mode) + result = lpi2c_imx_pio_xfer(lpi2c_imx, &msgs[i]); + } else { + result = lpi2c_imx_pio_xfer(lpi2c_imx, &msgs[i]); + } - result = lpi2c_imx_msg_complete(lpi2c_imx); if (result) goto stop; @@ -546,6 +993,58 @@ static int lpi2c_imx_init_recovery_info(struct lpi2c_imx_struct *lpi2c_imx, return 0; } +static void dma_exit(struct device *dev, struct lpi2c_imx_dma *dma) +{ + if (dma->chan_rx) + dma_release_channel(dma->chan_rx); + + if (dma->chan_tx) + dma_release_channel(dma->chan_tx); + + devm_kfree(dev, dma); +} + +static int lpi2c_dma_init(struct device *dev, dma_addr_t phy_addr) +{ + struct lpi2c_imx_struct *lpi2c_imx = dev_get_drvdata(dev); + struct lpi2c_imx_dma *dma; + int ret; + + dma = devm_kzalloc(dev, sizeof(*dma), GFP_KERNEL); + if (!dma) + return -ENOMEM; + + dma->phy_addr = phy_addr; + + /* Prepare for TX DMA: */ + dma->chan_tx = dma_request_chan(dev, "tx"); + if (IS_ERR(dma->chan_tx)) { + ret = PTR_ERR(dma->chan_tx); + if (ret != -ENODEV && ret != -EPROBE_DEFER) + dev_err(dev, "can't request DMA tx channel (%d)\n", ret); + dma->chan_tx = NULL; + goto dma_exit; + } + + /* Prepare for RX DMA: */ + dma->chan_rx = dma_request_chan(dev, "rx"); + if (IS_ERR(dma->chan_rx)) { + ret = PTR_ERR(dma->chan_rx); + if (ret != -ENODEV && ret != -EPROBE_DEFER) + dev_err(dev, "can't request DMA rx channel (%d)\n", ret); + dma->chan_rx = NULL; + goto dma_exit; + } + + lpi2c_imx->can_use_dma = true; + lpi2c_imx->dma = dma; + return 0; + +dma_exit: + dma_exit(dev, dma); + return ret; +} + static u32 lpi2c_imx_func(struct i2c_adapter *adapter) { return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL | @@ -566,6 +1065,8 @@ MODULE_DEVICE_TABLE(of, lpi2c_imx_of_match); static int lpi2c_imx_probe(struct platform_device *pdev) { struct lpi2c_imx_struct *lpi2c_imx; + struct resource *res; + dma_addr_t phy_addr; unsigned int temp; int irq, ret; @@ -573,7 +1074,7 @@ static int lpi2c_imx_probe(struct platform_device *pdev) if (!lpi2c_imx) return -ENOMEM; - lpi2c_imx->base = devm_platform_ioremap_resource(pdev, 0); + lpi2c_imx->base = devm_platform_get_and_ioremap_resource(pdev, 0, &res); if (IS_ERR(lpi2c_imx->base)) return PTR_ERR(lpi2c_imx->base); @@ -587,6 +1088,7 @@ static int lpi2c_imx_probe(struct platform_device *pdev) lpi2c_imx->adapter.dev.of_node = pdev->dev.of_node; strscpy(lpi2c_imx->adapter.name, pdev->name, sizeof(lpi2c_imx->adapter.name)); + phy_addr = (dma_addr_t)res->start; ret = devm_clk_bulk_get_all(&pdev->dev, &lpi2c_imx->clks); if (ret < 0) @@ -640,6 +1142,14 @@ static int lpi2c_imx_probe(struct platform_device *pdev) if (ret == -EPROBE_DEFER) goto rpm_disable; + /* Init DMA */ + ret = lpi2c_dma_init(&pdev->dev, phy_addr); + if (ret) { + if (ret == -EPROBE_DEFER) + goto rpm_disable; + dev_info(&pdev->dev, "use pio mode\n"); + } + ret = i2c_add_adapter(&lpi2c_imx->adapter); if (ret) goto rpm_disable; From 01aa028151380556a46225d840898fd4c590119f Mon Sep 17 00:00:00 2001 From: Ivaylo Ivanov Date: Sat, 28 Dec 2024 13:15:08 +0200 Subject: [PATCH 184/196] dt-bindings: i2c: exynos5: Add samsung,exynos8895-hsi2c compatible Add samsung,exynos8895-hsi2c dedicated compatible for representing I2C of Exynos8895 SoC. Since there are I2C buses that aren't implemented as a part of USIv1 blocks, they only require a single clock. Signed-off-by: Ivaylo Ivanov Reviewed-by: Krzysztof Kozlowski Link: https://lore.kernel.org/r/20241228111509.896502-2-ivo.ivanov.ivanov1@gmail.com Signed-off-by: Andi Shyti --- Documentation/devicetree/bindings/i2c/i2c-exynos5.yaml | 1 + 1 file changed, 1 insertion(+) diff --git a/Documentation/devicetree/bindings/i2c/i2c-exynos5.yaml b/Documentation/devicetree/bindings/i2c/i2c-exynos5.yaml index cc8bba5537b9..70cc2ee9ee27 100644 --- a/Documentation/devicetree/bindings/i2c/i2c-exynos5.yaml +++ b/Documentation/devicetree/bindings/i2c/i2c-exynos5.yaml @@ -25,6 +25,7 @@ properties: - samsung,exynos5250-hsi2c # Exynos5250 and Exynos5420 - samsung,exynos5260-hsi2c # Exynos5260 - samsung,exynos7-hsi2c # Exynos7 + - samsung,exynos8895-hsi2c - samsung,exynosautov9-hsi2c - items: - enum: From f8cef982c2ee2ba1042e72d5d33dd1be2de313dd Mon Sep 17 00:00:00 2001 From: Ivaylo Ivanov Date: Sat, 28 Dec 2024 13:15:09 +0200 Subject: [PATCH 185/196] i2c: exynos5: Add support for Exynos8895 SoC Exynos8895 functioning logic mostly follows I2C_TYPE_EXYNOS7, but timing and temp calculations are slightly different according to the following logic: FPCLK / FI2C = (CLK_DIV + 1) * (TSCLK_L + TSCLK_H + 2) + 2 * ((FLT_CYCLE + 3) - (FLT_CYCLE + 3) % (CLK_DIV + 1)) temp := (FPCLK / FI2C) - (FLT_CYCLE + 3) * 2 Signed-off-by: Ivaylo Ivanov Link: https://lore.kernel.org/r/20241228111509.896502-3-ivo.ivanov.ivanov1@gmail.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-exynos5.c | 35 ++++++++++++++++++++++++++++---- 1 file changed, 31 insertions(+), 4 deletions(-) diff --git a/drivers/i2c/busses/i2c-exynos5.c b/drivers/i2c/busses/i2c-exynos5.c index e330015087ab..6cdd957ea7e4 100644 --- a/drivers/i2c/busses/i2c-exynos5.c +++ b/drivers/i2c/busses/i2c-exynos5.c @@ -168,6 +168,7 @@ enum i2c_type_exynos { I2C_TYPE_EXYNOS5, I2C_TYPE_EXYNOS7, I2C_TYPE_EXYNOSAUTOV9, + I2C_TYPE_EXYNOS8895, }; struct exynos5_i2c { @@ -240,6 +241,11 @@ static const struct exynos_hsi2c_variant exynosautov9_hsi2c_data = { .hw = I2C_TYPE_EXYNOSAUTOV9, }; +static const struct exynos_hsi2c_variant exynos8895_hsi2c_data = { + .fifo_depth = 64, + .hw = I2C_TYPE_EXYNOS8895, +}; + static const struct of_device_id exynos5_i2c_match[] = { { .compatible = "samsung,exynos5-hsi2c", @@ -256,6 +262,9 @@ static const struct of_device_id exynos5_i2c_match[] = { }, { .compatible = "samsung,exynosautov9-hsi2c", .data = &exynosautov9_hsi2c_data + }, { + .compatible = "samsung,exynos8895-hsi2c", + .data = &exynos8895_hsi2c_data }, {}, }; MODULE_DEVICE_TABLE(of, exynos5_i2c_match); @@ -331,6 +340,14 @@ static int exynos5_i2c_set_timing(struct exynos5_i2c *i2c, bool hs_timings) * clk_cycle := TSCLK_L + TSCLK_H * temp := (CLK_DIV + 1) * (clk_cycle + 2) * + * In case of HSI2C controllers in Exynos8895 + * FPCLK / FI2C = + * (CLK_DIV + 1) * (TSCLK_L + TSCLK_H + 2) + + * 2 * ((FLT_CYCLE + 3) - (FLT_CYCLE + 3) % (CLK_DIV + 1)) + * + * clk_cycle := TSCLK_L + TSCLK_H + * temp := (FPCLK / FI2C) - (FLT_CYCLE + 3) * 2 + * * Constraints: 4 <= temp, 0 <= CLK_DIV < 256, 2 <= clk_cycle <= 510 * * To split SCL clock into low, high periods appropriately, one @@ -352,11 +369,19 @@ static int exynos5_i2c_set_timing(struct exynos5_i2c *i2c, bool hs_timings) * */ t_ftl_cycle = (readl(i2c->regs + HSI2C_CONF) >> 16) & 0x7; - temp = clkin / op_clk - 8 - t_ftl_cycle; - if (i2c->variant->hw != I2C_TYPE_EXYNOS7) - temp -= t_ftl_cycle; + if (i2c->variant->hw == I2C_TYPE_EXYNOS8895) + temp = clkin / op_clk - (t_ftl_cycle + 3) * 2; + else if (i2c->variant->hw == I2C_TYPE_EXYNOS7) + temp = clkin / op_clk - 8 - t_ftl_cycle; + else + temp = clkin / op_clk - 8 - (t_ftl_cycle * 2); div = temp / 512; - clk_cycle = temp / (div + 1) - 2; + + if (i2c->variant->hw == I2C_TYPE_EXYNOS8895) + clk_cycle = (temp + ((t_ftl_cycle + 3) % (div + 1)) * 2) / + (div + 1) - 2; + else + clk_cycle = temp / (div + 1) - 2; if (temp < 4 || div >= 256 || clk_cycle < 2) { dev_err(i2c->dev, "%s clock set-up failed\n", hs_timings ? "HS" : "FS"); @@ -491,6 +516,8 @@ static irqreturn_t exynos5_i2c_irq(int irqno, void *dev_id) switch (i2c->variant->hw) { case I2C_TYPE_EXYNOSAUTOV9: fallthrough; + case I2C_TYPE_EXYNOS8895: + fallthrough; case I2C_TYPE_EXYNOS7: if (int_status & HSI2C_INT_TRANS_DONE) { i2c->trans_done = 1; From bfb3ddfdc5fb7904ba5c7870cfec9efabc9a4f83 Mon Sep 17 00:00:00 2001 From: Andi Shyti Date: Fri, 27 Dec 2024 23:32:29 +0100 Subject: [PATCH 186/196] i2c: qcom-geni: Use dev_err_probe in the probe function Replace classical dev_err with dev_err_probe in the probe function for better error reporting. Also, use dev_err_probe in cases where the error number is clear (e.g., -EIO or -EINVAL) to maintain consistency. Additionally, remove redundant logging to simplify the code. Reviewed-by: Vladimir Zapolskiy Acked-by: Mukesh Kumar Savaliya Link: https://lore.kernel.org/r/20241227223230.462395-2-andi.shyti@kernel.org Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-qcom-geni.c | 33 +++++++++++++----------------- 1 file changed, 14 insertions(+), 19 deletions(-) diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c index 7a22e1f46e60..01db24188e29 100644 --- a/drivers/i2c/busses/i2c-qcom-geni.c +++ b/drivers/i2c/busses/i2c-qcom-geni.c @@ -823,11 +823,9 @@ static int geni_i2c_probe(struct platform_device *pdev) return gi2c->irq; ret = geni_i2c_clk_map_idx(gi2c); - if (ret) { - dev_err(dev, "Invalid clk frequency %d Hz: %d\n", - gi2c->clk_freq_out, ret); - return ret; - } + if (ret) + return dev_err_probe(dev, ret, "Invalid clk frequency %d Hz\n", + gi2c->clk_freq_out); gi2c->adap.algo = &geni_i2c_algo; init_completion(&gi2c->done); @@ -837,11 +835,10 @@ static int geni_i2c_probe(struct platform_device *pdev) /* Keep interrupts disabled initially to allow for low-power modes */ ret = devm_request_irq(dev, gi2c->irq, geni_i2c_irq, IRQF_NO_AUTOEN, dev_name(dev), gi2c); - if (ret) { - dev_err(dev, "Request_irq failed:%d: err:%d\n", - gi2c->irq, ret); - return ret; - } + if (ret) + return dev_err_probe(dev, ret, + "Request_irq failed: %d\n", gi2c->irq); + i2c_set_adapdata(&gi2c->adap, gi2c); gi2c->adap.dev.parent = dev; gi2c->adap.dev.of_node = dev->of_node; @@ -870,16 +867,14 @@ static int geni_i2c_probe(struct platform_device *pdev) ret = geni_se_resources_on(&gi2c->se); if (ret) { - dev_err(dev, "Error turning on resources %d\n", ret); clk_disable_unprepare(gi2c->core_clk); - return ret; + return dev_err_probe(dev, ret, "Error turning on resources\n"); } proto = geni_se_read_proto(&gi2c->se); if (proto != GENI_SE_I2C) { - dev_err(dev, "Invalid proto %d\n", proto); geni_se_resources_off(&gi2c->se); clk_disable_unprepare(gi2c->core_clk); - return -ENXIO; + return dev_err_probe(dev, -ENXIO, "Invalid proto %d\n", proto); } if (desc && desc->no_dma_support) @@ -894,7 +889,7 @@ static int geni_i2c_probe(struct platform_device *pdev) if (ret) { geni_se_resources_off(&gi2c->se); clk_disable_unprepare(gi2c->core_clk); - return dev_err_probe(dev, ret, "Failed to setup GPI DMA mode\n"); + return ret; } dev_dbg(dev, "Using GPI DMA mode for I2C\n"); @@ -907,10 +902,10 @@ static int geni_i2c_probe(struct platform_device *pdev) tx_depth = desc->tx_fifo_depth; if (!tx_depth) { - dev_err(dev, "Invalid TX FIFO depth\n"); geni_se_resources_off(&gi2c->se); clk_disable_unprepare(gi2c->core_clk); - return -EINVAL; + return dev_err_probe(dev, -EINVAL, + "Invalid TX FIFO depth\n"); } gi2c->tx_wm = tx_depth - 1; @@ -924,7 +919,7 @@ static int geni_i2c_probe(struct platform_device *pdev) clk_disable_unprepare(gi2c->core_clk); ret = geni_se_resources_off(&gi2c->se); if (ret) { - dev_err(dev, "Error turning off resources %d\n", ret); + dev_err_probe(dev, ret, "Error turning off resources\n"); goto err_dma; } @@ -940,7 +935,7 @@ static int geni_i2c_probe(struct platform_device *pdev) ret = i2c_add_adapter(&gi2c->adap); if (ret) { - dev_err(dev, "Error adding i2c adapter %d\n", ret); + dev_err_probe(dev, ret, "Error adding i2c adapter\n"); pm_runtime_disable(gi2c->se.dev); goto err_dma; } From 494fed6809585b8386072cf96ef2426e7a7c6326 Mon Sep 17 00:00:00 2001 From: Andi Shyti Date: Fri, 27 Dec 2024 23:32:30 +0100 Subject: [PATCH 187/196] i2c: qcom-geni: Simplify error handling in probe function Avoid repeating the error handling pattern: geni_se_resources_off(&gi2c->se); clk_disable_unprepare(gi2c->core_clk); return; Introduce a single 'goto' exit label for cleanup in case of errors. While there are currently two distinct exit points, there is no overlap in their handling, allowing both branches to coexist cleanly. Reviewed-by: Vladimir Zapolskiy Link: https://lore.kernel.org/r/20241227223230.462395-3-andi.shyti@kernel.org Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-qcom-geni.c | 33 ++++++++++++++++-------------- 1 file changed, 18 insertions(+), 15 deletions(-) diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c index 01db24188e29..7bbd478171e0 100644 --- a/drivers/i2c/busses/i2c-qcom-geni.c +++ b/drivers/i2c/busses/i2c-qcom-geni.c @@ -867,14 +867,13 @@ static int geni_i2c_probe(struct platform_device *pdev) ret = geni_se_resources_on(&gi2c->se); if (ret) { - clk_disable_unprepare(gi2c->core_clk); - return dev_err_probe(dev, ret, "Error turning on resources\n"); + dev_err_probe(dev, ret, "Error turning on resources\n"); + goto err_clk; } proto = geni_se_read_proto(&gi2c->se); if (proto != GENI_SE_I2C) { - geni_se_resources_off(&gi2c->se); - clk_disable_unprepare(gi2c->core_clk); - return dev_err_probe(dev, -ENXIO, "Invalid proto %d\n", proto); + ret = dev_err_probe(dev, -ENXIO, "Invalid proto %d\n", proto); + goto err_resources; } if (desc && desc->no_dma_support) @@ -886,11 +885,8 @@ static int geni_i2c_probe(struct platform_device *pdev) /* FIFO is disabled, so we can only use GPI DMA */ gi2c->gpi_mode = true; ret = setup_gpi_dma(gi2c); - if (ret) { - geni_se_resources_off(&gi2c->se); - clk_disable_unprepare(gi2c->core_clk); - return ret; - } + if (ret) + goto err_resources; dev_dbg(dev, "Using GPI DMA mode for I2C\n"); } else { @@ -902,10 +898,9 @@ static int geni_i2c_probe(struct platform_device *pdev) tx_depth = desc->tx_fifo_depth; if (!tx_depth) { - geni_se_resources_off(&gi2c->se); - clk_disable_unprepare(gi2c->core_clk); - return dev_err_probe(dev, -EINVAL, - "Invalid TX FIFO depth\n"); + ret = dev_err_probe(dev, -EINVAL, + "Invalid TX FIFO depth\n"); + goto err_resources; } gi2c->tx_wm = tx_depth - 1; @@ -942,10 +937,18 @@ static int geni_i2c_probe(struct platform_device *pdev) dev_dbg(dev, "Geni-I2C adaptor successfully added\n"); - return 0; + return ret; + +err_resources: + geni_se_resources_off(&gi2c->se); +err_clk: + clk_disable_unprepare(gi2c->core_clk); + + return ret; err_dma: release_gpi_dma(gi2c); + return ret; } From febe15218072eee5361a36f4510711a7cf7403e1 Mon Sep 17 00:00:00 2001 From: Bryan O'Donoghue Date: Fri, 27 Dec 2024 13:11:34 +0000 Subject: [PATCH 188/196] dt-bindings: i2c: qcom-cci: Document x1e80100 compatible Add the x1e80100 CCI device string compatible. Signed-off-by: Bryan O'Donoghue Reviewed-by: Krzysztof Kozlowski Link: https://lore.kernel.org/r/20241227-b4-linux-next-24-11-18-dtsi-x1e80100-camss-v2-1-06fdd5a7d5bb@linaro.org Signed-off-by: Andi Shyti --- Documentation/devicetree/bindings/i2c/qcom,i2c-cci.yaml | 2 ++ 1 file changed, 2 insertions(+) diff --git a/Documentation/devicetree/bindings/i2c/qcom,i2c-cci.yaml b/Documentation/devicetree/bindings/i2c/qcom,i2c-cci.yaml index ef26ba6eda28..73144473b9b2 100644 --- a/Documentation/devicetree/bindings/i2c/qcom,i2c-cci.yaml +++ b/Documentation/devicetree/bindings/i2c/qcom,i2c-cci.yaml @@ -34,6 +34,7 @@ properties: - qcom,sm8450-cci - qcom,sm8550-cci - qcom,sm8650-cci + - qcom,x1e80100-cci - const: qcom,msm8996-cci # CCI v2 "#address-cells": @@ -224,6 +225,7 @@ allOf: enum: - qcom,sm8550-cci - qcom,sm8650-cci + - qcom,x1e80100-cci then: properties: clocks: From 1ee867e465c1b0216ec02d4c0d334c4e89919cec Mon Sep 17 00:00:00 2001 From: Carlos Song Date: Mon, 30 Dec 2024 11:38:46 +0800 Subject: [PATCH 189/196] i2c: imx-lpi2c: add target mode support LPI2C support master controller and target controller enabled simultaneously. Both controllers share the same SDA/SCL lines and interrupt source but has a separate control and status registers. When target is enabled and an interrupt has been triggered, target register status will be checked to determine IRQ source. Then enter the corresponding interrupt handler function of master or target to handle the interrupt event. This patch supports basic target data read/write operations in 7-bit target address. LPI2C target mode can be enabled by using I2C slave backend. I2C slave backend behaves like a standard I2C client. For simple use and test, Linux I2C slave EEPROM backend can be used. Signed-off-by: Carlos Song Link: https://lore.kernel.org/r/20241230033846.2302500-1-carlos.song@nxp.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-imx-lpi2c.c | 251 ++++++++++++++++++++++++++++- 1 file changed, 247 insertions(+), 4 deletions(-) diff --git a/drivers/i2c/busses/i2c-imx-lpi2c.c b/drivers/i2c/busses/i2c-imx-lpi2c.c index 9778dc8f6d36..dc25d7dfe1ab 100644 --- a/drivers/i2c/busses/i2c-imx-lpi2c.c +++ b/drivers/i2c/busses/i2c-imx-lpi2c.c @@ -43,6 +43,20 @@ #define LPI2C_MTDR 0x60 /* i2c master TX data register */ #define LPI2C_MRDR 0x70 /* i2c master RX data register */ +#define LPI2C_SCR 0x110 /* i2c target control register */ +#define LPI2C_SSR 0x114 /* i2c target status register */ +#define LPI2C_SIER 0x118 /* i2c target interrupt enable */ +#define LPI2C_SDER 0x11C /* i2c target DMA enable */ +#define LPI2C_SCFGR0 0x120 /* i2c target configuration */ +#define LPI2C_SCFGR1 0x124 /* i2c target configuration */ +#define LPI2C_SCFGR2 0x128 /* i2c target configuration */ +#define LPI2C_SAMR 0x140 /* i2c target address match */ +#define LPI2C_SASR 0x150 /* i2c target address status */ +#define LPI2C_STAR 0x154 /* i2c target transmit ACK */ +#define LPI2C_STDR 0x160 /* i2c target transmit data */ +#define LPI2C_SRDR 0x170 /* i2c target receive data */ +#define LPI2C_SRDROR 0x178 /* i2c target receive data read only */ + /* i2c command */ #define TRAN_DATA 0X00 #define RECV_DATA 0X01 @@ -76,6 +90,42 @@ #define MDER_TDDE BIT(0) #define MDER_RDDE BIT(1) +#define SCR_SEN BIT(0) +#define SCR_RST BIT(1) +#define SCR_FILTEN BIT(4) +#define SCR_RTF BIT(8) +#define SCR_RRF BIT(9) +#define SSR_TDF BIT(0) +#define SSR_RDF BIT(1) +#define SSR_AVF BIT(2) +#define SSR_TAF BIT(3) +#define SSR_RSF BIT(8) +#define SSR_SDF BIT(9) +#define SSR_BEF BIT(10) +#define SSR_FEF BIT(11) +#define SSR_SBF BIT(24) +#define SSR_BBF BIT(25) +#define SSR_CLEAR_BITS (SSR_RSF | SSR_SDF | SSR_BEF | SSR_FEF) +#define SIER_TDIE BIT(0) +#define SIER_RDIE BIT(1) +#define SIER_AVIE BIT(2) +#define SIER_TAIE BIT(3) +#define SIER_RSIE BIT(8) +#define SIER_SDIE BIT(9) +#define SIER_BEIE BIT(10) +#define SIER_FEIE BIT(11) +#define SIER_AM0F BIT(12) +#define SCFGR1_RXSTALL BIT(1) +#define SCFGR1_TXDSTALL BIT(2) +#define SCFGR2_FILTSDA_SHIFT 24 +#define SCFGR2_FILTSCL_SHIFT 16 +#define SCFGR2_CLKHOLD(x) (x) +#define SCFGR2_FILTSDA(x) ((x) << SCFGR2_FILTSDA_SHIFT) +#define SCFGR2_FILTSCL(x) ((x) << SCFGR2_FILTSCL_SHIFT) +#define SASR_READ_REQ 0x1 +#define SLAVE_INT_FLAG (SIER_TDIE | SIER_RDIE | SIER_AVIE | \ + SIER_SDIE | SIER_BEIE) + #define I2C_CLK_RATIO 2 #define CHUNK_DATA 256 @@ -134,6 +184,7 @@ struct lpi2c_imx_struct { struct i2c_bus_recovery_info rinfo; bool can_use_dma; struct lpi2c_imx_dma *dma; + struct i2c_client *target; }; static void lpi2c_imx_intctrl(struct lpi2c_imx_struct *lpi2c_imx, @@ -957,9 +1008,56 @@ static int lpi2c_imx_xfer(struct i2c_adapter *adapter, return (result < 0) ? result : num; } -static irqreturn_t lpi2c_imx_isr(int irq, void *dev_id) +static irqreturn_t lpi2c_imx_target_isr(struct lpi2c_imx_struct *lpi2c_imx, + u32 ssr, u32 sier_filter) +{ + u8 value; + u32 sasr; + + /* Arbitration lost */ + if (sier_filter & SSR_BEF) { + writel(0, lpi2c_imx->base + LPI2C_SIER); + return IRQ_HANDLED; + } + + /* Address detected */ + if (sier_filter & SSR_AVF) { + sasr = readl(lpi2c_imx->base + LPI2C_SASR); + if (SASR_READ_REQ & sasr) { + /* Read request */ + i2c_slave_event(lpi2c_imx->target, I2C_SLAVE_READ_REQUESTED, &value); + writel(value, lpi2c_imx->base + LPI2C_STDR); + goto ret; + } else { + /* Write request */ + i2c_slave_event(lpi2c_imx->target, I2C_SLAVE_WRITE_REQUESTED, &value); + } + } + + if (sier_filter & SSR_SDF) + /* STOP */ + i2c_slave_event(lpi2c_imx->target, I2C_SLAVE_STOP, &value); + + if (sier_filter & SSR_TDF) { + /* Target send data */ + i2c_slave_event(lpi2c_imx->target, I2C_SLAVE_READ_PROCESSED, &value); + writel(value, lpi2c_imx->base + LPI2C_STDR); + } + + if (sier_filter & SSR_RDF) { + /* Target receive data */ + value = readl(lpi2c_imx->base + LPI2C_SRDR); + i2c_slave_event(lpi2c_imx->target, I2C_SLAVE_WRITE_RECEIVED, &value); + } + +ret: + /* Clear SSR */ + writel(ssr & SSR_CLEAR_BITS, lpi2c_imx->base + LPI2C_SSR); + return IRQ_HANDLED; +} + +static irqreturn_t lpi2c_imx_master_isr(struct lpi2c_imx_struct *lpi2c_imx) { - struct lpi2c_imx_struct *lpi2c_imx = dev_id; unsigned int enabled; unsigned int temp; @@ -979,6 +1077,124 @@ static irqreturn_t lpi2c_imx_isr(int irq, void *dev_id) return IRQ_HANDLED; } +static irqreturn_t lpi2c_imx_isr(int irq, void *dev_id) +{ + struct lpi2c_imx_struct *lpi2c_imx = dev_id; + + if (lpi2c_imx->target) { + u32 scr = readl(lpi2c_imx->base + LPI2C_SCR); + u32 ssr = readl(lpi2c_imx->base + LPI2C_SSR); + u32 sier_filter = ssr & readl(lpi2c_imx->base + LPI2C_SIER); + + /* + * The target is enabled and an interrupt has been triggered. + * Enter the target's irq handler. + */ + if ((scr & SCR_SEN) && sier_filter) + return lpi2c_imx_target_isr(lpi2c_imx, ssr, sier_filter); + } + + /* + * Otherwise the interrupt has been triggered by the master. + * Enter the master's irq handler. + */ + return lpi2c_imx_master_isr(lpi2c_imx); +} + +static void lpi2c_imx_target_init(struct lpi2c_imx_struct *lpi2c_imx) +{ + u32 temp; + + /* reset target module */ + writel(SCR_RST, lpi2c_imx->base + LPI2C_SCR); + writel(0, lpi2c_imx->base + LPI2C_SCR); + + /* Set target address */ + writel((lpi2c_imx->target->addr << 1), lpi2c_imx->base + LPI2C_SAMR); + + writel(SCFGR1_RXSTALL | SCFGR1_TXDSTALL, lpi2c_imx->base + LPI2C_SCFGR1); + + /* + * set SCFGR2: FILTSDA, FILTSCL and CLKHOLD + * + * FILTSCL/FILTSDA can eliminate signal skew. It should generally be + * set to the same value and should be set >= 50ns. + * + * CLKHOLD is only used when clock stretching is enabled, but it will + * extend the clock stretching to ensure there is an additional delay + * between the target driving SDA and the target releasing the SCL pin. + * + * CLKHOLD setting is crucial for lpi2c target. When master read data + * from target, if there is a delay caused by cpu idle, excessive load, + * or other delays between two bytes in one message transmission, it + * will cause a short interval time between the driving SDA signal and + * releasing SCL signal. The lpi2c master will mistakenly think it is a stop + * signal resulting in an arbitration failure. This issue can be avoided + * by setting CLKHOLD. + * + * In order to ensure lpi2c function normally when the lpi2c speed is as + * low as 100kHz, CLKHOLD should be set to 3 and it is also compatible with + * higher clock frequency like 400kHz and 1MHz. + */ + temp = SCFGR2_FILTSDA(2) | SCFGR2_FILTSCL(2) | SCFGR2_CLKHOLD(3); + writel(temp, lpi2c_imx->base + LPI2C_SCFGR2); + + /* + * Enable module: + * SCR_FILTEN can enable digital filter and output delay counter for LPI2C + * target mode. So SCR_FILTEN need be asserted when enable SDA/SCL FILTER + * and CLKHOLD. + */ + writel(SCR_SEN | SCR_FILTEN, lpi2c_imx->base + LPI2C_SCR); + + /* Enable interrupt from i2c module */ + writel(SLAVE_INT_FLAG, lpi2c_imx->base + LPI2C_SIER); +} + +static int lpi2c_imx_register_target(struct i2c_client *client) +{ + struct lpi2c_imx_struct *lpi2c_imx = i2c_get_adapdata(client->adapter); + int ret; + + if (lpi2c_imx->target) + return -EBUSY; + + lpi2c_imx->target = client; + + ret = pm_runtime_resume_and_get(lpi2c_imx->adapter.dev.parent); + if (ret < 0) { + dev_err(&lpi2c_imx->adapter.dev, "failed to resume i2c controller"); + return ret; + } + + lpi2c_imx_target_init(lpi2c_imx); + + return 0; +} + +static int lpi2c_imx_unregister_target(struct i2c_client *client) +{ + struct lpi2c_imx_struct *lpi2c_imx = i2c_get_adapdata(client->adapter); + int ret; + + if (!lpi2c_imx->target) + return -EINVAL; + + /* Reset target address. */ + writel(0, lpi2c_imx->base + LPI2C_SAMR); + + writel(SCR_RST, lpi2c_imx->base + LPI2C_SCR); + writel(0, lpi2c_imx->base + LPI2C_SCR); + + lpi2c_imx->target = NULL; + + ret = pm_runtime_put_sync(lpi2c_imx->adapter.dev.parent); + if (ret < 0) + dev_err(&lpi2c_imx->adapter.dev, "failed to suspend i2c controller"); + + return ret; +} + static int lpi2c_imx_init_recovery_info(struct lpi2c_imx_struct *lpi2c_imx, struct platform_device *pdev) { @@ -1054,6 +1270,8 @@ static u32 lpi2c_imx_func(struct i2c_adapter *adapter) static const struct i2c_algorithm lpi2c_imx_algo = { .master_xfer = lpi2c_imx_xfer, .functionality = lpi2c_imx_func, + .reg_target = lpi2c_imx_register_target, + .unreg_target = lpi2c_imx_unregister_target, }; static const struct of_device_id lpi2c_imx_of_match[] = { @@ -1204,9 +1422,34 @@ static int __maybe_unused lpi2c_runtime_resume(struct device *dev) return 0; } +static int __maybe_unused lpi2c_suspend_noirq(struct device *dev) +{ + return pm_runtime_force_suspend(dev); +} + +static int __maybe_unused lpi2c_resume_noirq(struct device *dev) +{ + struct lpi2c_imx_struct *lpi2c_imx = dev_get_drvdata(dev); + int ret; + + ret = pm_runtime_force_resume(dev); + if (ret) + return ret; + + /* + * If the I2C module powers down during system suspend, + * the register values will be lost. Therefore, reinitialize + * the target when the system resumes. + */ + if (lpi2c_imx->target) + lpi2c_imx_target_init(lpi2c_imx); + + return 0; +} + static const struct dev_pm_ops lpi2c_pm_ops = { - SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, - pm_runtime_force_resume) + SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(lpi2c_suspend_noirq, + lpi2c_resume_noirq) SET_RUNTIME_PM_OPS(lpi2c_runtime_suspend, lpi2c_runtime_resume, NULL) }; From 70f3d3669c074efbcee32867a1ab71f5f7ead385 Mon Sep 17 00:00:00 2001 From: Atharva Tiwari Date: Wed, 1 Jan 2025 16:04:22 +0530 Subject: [PATCH 190/196] i2c: amd756: Fix endianness handling for word data Ensure correct handling of "endianness" for word-sized data in amd756_access - Convert word data into little-endian using cpu_to_le16 - Convert word data from little-endian to cpu native format using le16_to_cpu This fixes poteential issues on big-endian systems and ensure proper byte ordering for SMBus word transacitions Signed-off-by: Atharva Tiwari Link: https://lore.kernel.org/r/20250101103422.30523-1-evepolonium@gmail.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-amd756.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/i2c/busses/i2c-amd756.c b/drivers/i2c/busses/i2c-amd756.c index fa0d5a2c3732..e551d63e96b1 100644 --- a/drivers/i2c/busses/i2c-amd756.c +++ b/drivers/i2c/busses/i2c-amd756.c @@ -31,6 +31,7 @@ #include #include #include +#include /* AMD756 SMBus address offsets */ #define SMB_ADDR_OFFSET 0xE0 @@ -211,7 +212,7 @@ static s32 amd756_access(struct i2c_adapter * adap, u16 addr, SMB_HOST_ADDRESS); outb_p(command, SMB_HOST_COMMAND); if (read_write == I2C_SMBUS_WRITE) - outw_p(data->word, SMB_HOST_DATA); /* TODO: endian???? */ + outw_p(cpu_to_le16((u16)data->word), SMB_HOST_DATA); size = AMD756_WORD_DATA; break; case I2C_SMBUS_BLOCK_DATA: @@ -256,7 +257,7 @@ static s32 amd756_access(struct i2c_adapter * adap, u16 addr, data->byte = inw_p(SMB_HOST_DATA); break; case AMD756_WORD_DATA: - data->word = inw_p(SMB_HOST_DATA); /* TODO: endian???? */ + data->word = (u16)le16_to_cpu(inw_p(SMB_HOST_DATA)); break; case AMD756_BLOCK_DATA: data->block[0] = inw_p(SMB_HOST_DATA) & 0x3f; From b04ce63859793e3439b394976b8d29e785d4d69a Mon Sep 17 00:00:00 2001 From: Bartosz Golaszewski Date: Wed, 11 Dec 2024 11:23:35 +0100 Subject: [PATCH 191/196] i2c: davinci: kill platform data There are no more board file users of this driver. The platform data structure is only used internally. Two of the four fields it stores are not used at all anymore. Pull the remainder into the driver data struct and shrink code by removing parts that are now dead code. Signed-off-by: Bartosz Golaszewski Link: https://lore.kernel.org/r/20241211102337.37956-1-brgl@bgdev.pl Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-davinci.c | 80 +++++------------------ include/linux/platform_data/i2c-davinci.h | 26 -------- 2 files changed, 17 insertions(+), 89 deletions(-) delete mode 100644 include/linux/platform_data/i2c-davinci.h diff --git a/drivers/i2c/busses/i2c-davinci.c b/drivers/i2c/busses/i2c-davinci.c index 71dc0a6688b7..4b499931fdfd 100644 --- a/drivers/i2c/busses/i2c-davinci.c +++ b/drivers/i2c/busses/i2c-davinci.c @@ -26,7 +26,6 @@ #include #include #include -#include #include /* ----- global defines ----------------------------------------------- */ @@ -117,6 +116,8 @@ /* timeout for pm runtime autosuspend */ #define DAVINCI_I2C_PM_TIMEOUT 1000 /* ms */ +#define DAVINCI_I2C_DEFAULT_BUS_FREQ 100 + struct davinci_i2c_dev { struct device *dev; void __iomem *base; @@ -132,13 +133,10 @@ struct davinci_i2c_dev { #ifdef CONFIG_CPU_FREQ struct notifier_block freq_transition; #endif - struct davinci_i2c_platform_data *pdata; -}; - -/* default platform data to use if not supplied in the platform_device */ -static struct davinci_i2c_platform_data davinci_i2c_platform_data_default = { - .bus_freq = 100, - .bus_delay = 0, + /* standard bus frequency (kHz) */ + unsigned int bus_freq; + /* Chip has a ICPFUNC register */ + bool has_pfunc; }; static inline void davinci_i2c_write_reg(struct davinci_i2c_dev *i2c_dev, @@ -168,7 +166,6 @@ static inline void davinci_i2c_reset_ctrl(struct davinci_i2c_dev *i2c_dev, static void i2c_davinci_calc_clk_dividers(struct davinci_i2c_dev *dev) { - struct davinci_i2c_platform_data *pdata = dev->pdata; u16 psc; u32 clk; u32 d; @@ -212,16 +209,16 @@ static void i2c_davinci_calc_clk_dividers(struct davinci_i2c_dev *dev) if (of_node && of_device_is_compatible(of_node, "ti,keystone-i2c")) d = 6; - clk = ((input_clock / (psc + 1)) / (pdata->bus_freq * 1000)); + clk = ((input_clock / (psc + 1)) / (dev->bus_freq * 1000)); /* Avoid driving the bus too fast because of rounding errors above */ - if (input_clock / (psc + 1) / clk > pdata->bus_freq * 1000) + if (input_clock / (psc + 1) / clk > dev->bus_freq * 1000) clk++; /* * According to I2C-BUS Spec 2.1, in FAST-MODE LOW period should be at * least 1.3uS, which is not the case with 50% duty cycle. Driving HIGH * to LOW ratio as 1 to 2 is more safe. */ - if (pdata->bus_freq > 100) + if (dev->bus_freq > 100) clkl = (clk << 1) / 3; else clkl = (clk >> 1); @@ -255,8 +252,6 @@ static void i2c_davinci_calc_clk_dividers(struct davinci_i2c_dev *dev) */ static int i2c_davinci_init(struct davinci_i2c_dev *dev) { - struct davinci_i2c_platform_data *pdata = dev->pdata; - /* put I2C into reset */ davinci_i2c_reset_ctrl(dev, 0); @@ -274,8 +269,7 @@ static int i2c_davinci_init(struct davinci_i2c_dev *dev) davinci_i2c_read_reg(dev, DAVINCI_I2C_CLKL_REG)); dev_dbg(dev->dev, "CLKH = %d\n", davinci_i2c_read_reg(dev, DAVINCI_I2C_CLKH_REG)); - dev_dbg(dev->dev, "bus_freq = %dkHz, bus_delay = %d\n", - pdata->bus_freq, pdata->bus_delay); + dev_dbg(dev->dev, "bus_freq = %dkHz\n", dev->bus_freq); /* Take the I2C module out of reset: */ @@ -309,12 +303,6 @@ static void davinci_i2c_unprepare_recovery(struct i2c_adapter *adap) i2c_davinci_init(dev); } -static struct i2c_bus_recovery_info davinci_i2c_gpio_recovery_info = { - .recover_bus = i2c_generic_scl_recovery, - .prepare_recovery = davinci_i2c_prepare_recovery, - .unprepare_recovery = davinci_i2c_unprepare_recovery, -}; - static void davinci_i2c_set_scl(struct i2c_adapter *adap, int val) { struct davinci_i2c_dev *dev = i2c_get_adapdata(adap); @@ -414,7 +402,6 @@ static int i2c_davinci_xfer_msg(struct i2c_adapter *adap, struct i2c_msg *msg, int stop) { struct davinci_i2c_dev *dev = i2c_get_adapdata(adap); - struct davinci_i2c_platform_data *pdata = dev->pdata; u32 flag; u16 w; unsigned long time_left; @@ -424,10 +411,6 @@ i2c_davinci_xfer_msg(struct i2c_adapter *adap, struct i2c_msg *msg, int stop) return -EADDRNOTAVAIL; } - /* Introduce a delay, required for some boards (e.g Davinci EVM) */ - if (pdata->bus_delay) - udelay(pdata->bus_delay); - /* set the target address */ davinci_i2c_write_reg(dev, DAVINCI_I2C_SAR_REG, msg->addr); @@ -758,8 +741,8 @@ static int davinci_i2c_probe(struct platform_device *pdev) { struct davinci_i2c_dev *dev; struct i2c_adapter *adap; - struct i2c_bus_recovery_info *rinfo; int r, irq; + u32 prop; irq = platform_get_irq(pdev, 0); if (irq < 0) @@ -773,29 +756,15 @@ static int davinci_i2c_probe(struct platform_device *pdev) dev->dev = &pdev->dev; dev->irq = irq; - dev->pdata = dev_get_platdata(&pdev->dev); platform_set_drvdata(pdev, dev); - if (!dev->pdata && pdev->dev.of_node) { - u32 prop; + r = device_property_read_u32(&pdev->dev, "clock-frequency", &prop); + if (r) + prop = DAVINCI_I2C_DEFAULT_BUS_FREQ; - dev->pdata = devm_kzalloc(&pdev->dev, - sizeof(struct davinci_i2c_platform_data), GFP_KERNEL); - if (!dev->pdata) - return -ENOMEM; + dev->bus_freq = prop / 1000; - memcpy(dev->pdata, &davinci_i2c_platform_data_default, - sizeof(struct davinci_i2c_platform_data)); - if (!of_property_read_u32(pdev->dev.of_node, "clock-frequency", - &prop)) - dev->pdata->bus_freq = prop / 1000; - - dev->pdata->has_pfunc = - of_property_read_bool(pdev->dev.of_node, - "ti,has-pfunc"); - } else if (!dev->pdata) { - dev->pdata = &davinci_i2c_platform_data_default; - } + dev->has_pfunc = device_property_present(&pdev->dev, "ti,has-pfunc"); dev->clk = devm_clk_get(&pdev->dev, NULL); if (IS_ERR(dev->clk)) @@ -843,23 +812,8 @@ static int davinci_i2c_probe(struct platform_device *pdev) adap->timeout = DAVINCI_I2C_TIMEOUT; adap->dev.of_node = pdev->dev.of_node; - if (dev->pdata->has_pfunc) + if (dev->has_pfunc) adap->bus_recovery_info = &davinci_i2c_scl_recovery_info; - else if (dev->pdata->gpio_recovery) { - rinfo = &davinci_i2c_gpio_recovery_info; - adap->bus_recovery_info = rinfo; - rinfo->scl_gpiod = devm_gpiod_get(&pdev->dev, "scl", - GPIOD_OUT_HIGH_OPEN_DRAIN); - if (IS_ERR(rinfo->scl_gpiod)) { - r = PTR_ERR(rinfo->scl_gpiod); - goto err_unuse_clocks; - } - rinfo->sda_gpiod = devm_gpiod_get(&pdev->dev, "sda", GPIOD_IN); - if (IS_ERR(rinfo->sda_gpiod)) { - r = PTR_ERR(rinfo->sda_gpiod); - goto err_unuse_clocks; - } - } adap->nr = pdev->id; r = i2c_add_numbered_adapter(adap); diff --git a/include/linux/platform_data/i2c-davinci.h b/include/linux/platform_data/i2c-davinci.h deleted file mode 100644 index 98967df07468..000000000000 --- a/include/linux/platform_data/i2c-davinci.h +++ /dev/null @@ -1,26 +0,0 @@ -/* - * DaVinci I2C controller platform_device info - * - * Author: Vladimir Barinov, MontaVista Software, Inc. - * - * 2007 (c) MontaVista Software, Inc. This file is licensed under - * the terms of the GNU General Public License version 2. This program - * is licensed "as is" without any warranty of any kind, whether express - * or implied. -*/ - -#ifndef __ASM_ARCH_I2C_H -#define __ASM_ARCH_I2C_H - -/* All frequencies are expressed in kHz */ -struct davinci_i2c_platform_data { - unsigned int bus_freq; /* standard bus frequency (kHz) */ - unsigned int bus_delay; /* post-transaction delay (usec) */ - bool gpio_recovery; /* Use GPIO recovery method */ - bool has_pfunc; /* Chip has a ICPFUNC register */ -}; - -/* for board setup code */ -void davinci_init_i2c(struct davinci_i2c_platform_data *); - -#endif /* __ASM_ARCH_I2C_H */ From 7d2b31f833b7faf2f17f9d54cf3820367ba85a23 Mon Sep 17 00:00:00 2001 From: Bartosz Golaszewski Date: Wed, 11 Dec 2024 11:23:36 +0100 Subject: [PATCH 192/196] i2c: davinci: order includes alphabetically For better readability order included headers alphabetically. Signed-off-by: Bartosz Golaszewski Reviewed-by: Andi Shyti Link: https://lore.kernel.org/r/20241211102337.37956-2-brgl@bgdev.pl Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-davinci.c | 25 +++++++++++++------------ 1 file changed, 13 insertions(+), 12 deletions(-) diff --git a/drivers/i2c/busses/i2c-davinci.c b/drivers/i2c/busses/i2c-davinci.c index 4b499931fdfd..bab9f785eeec 100644 --- a/drivers/i2c/busses/i2c-davinci.c +++ b/drivers/i2c/busses/i2c-davinci.c @@ -11,22 +11,23 @@ * * ---------------------------------------------------------------------------- */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include #include #include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include #include +#include #include +#include +#include /* ----- global defines ----------------------------------------------- */ From 8abbc711da605e530a0f06bed7feb76df35e99f7 Mon Sep 17 00:00:00 2001 From: Bartosz Golaszewski Date: Wed, 11 Dec 2024 11:23:37 +0100 Subject: [PATCH 193/196] i2c: davinci: use generic device property accessors Don't use generic OF APIs if the generic device-level ones will do. Signed-off-by: Bartosz Golaszewski Reviewed-by: Andi Shyti Link: https://lore.kernel.org/r/20241211102337.37956-3-brgl@bgdev.pl Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-davinci.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/drivers/i2c/busses/i2c-davinci.c b/drivers/i2c/busses/i2c-davinci.c index bab9f785eeec..6a909d339681 100644 --- a/drivers/i2c/busses/i2c-davinci.c +++ b/drivers/i2c/busses/i2c-davinci.c @@ -23,9 +23,9 @@ #include #include #include -#include #include #include +#include #include #include @@ -173,7 +173,6 @@ static void i2c_davinci_calc_clk_dividers(struct davinci_i2c_dev *dev) u32 clkh; u32 clkl; u32 input_clock = clk_get_rate(dev->clk); - struct device_node *of_node = dev->dev->of_node; /* NOTE: I2C Clock divider programming info * As per I2C specs the following formulas provide prescaler @@ -207,7 +206,7 @@ static void i2c_davinci_calc_clk_dividers(struct davinci_i2c_dev *dev) psc++; /* better to run under spec than over */ d = (psc >= 2) ? 5 : 7 - psc; - if (of_node && of_device_is_compatible(of_node, "ti,keystone-i2c")) + if (device_is_compatible(dev->dev, "ti,keystone-i2c")) d = 6; clk = ((input_clock / (psc + 1)) / (dev->bus_freq * 1000)); @@ -811,7 +810,7 @@ static int davinci_i2c_probe(struct platform_device *pdev) adap->algo = &i2c_davinci_algo; adap->dev.parent = &pdev->dev; adap->timeout = DAVINCI_I2C_TIMEOUT; - adap->dev.of_node = pdev->dev.of_node; + adap->dev.of_node = dev_of_node(&pdev->dev); if (dev->has_pfunc) adap->bus_recovery_info = &davinci_i2c_scl_recovery_info; From 4262df2a69c32a0bce87d75342f6f8e3a535d7cb Mon Sep 17 00:00:00 2001 From: Carlos Song Date: Fri, 27 Dec 2024 16:47:36 +0800 Subject: [PATCH 194/196] i2c: imx-lpi2c: make controller available until the system enters suspend_noirq() and from resume_noirq(). This is a general i2c controller feature. Some I2C devices may need the I2C controller to remain active during resume_noirq() or suspend_noirq(). If the controller is autosuspended, there is no way to wake it up once runtime PM disabled (in suspend_late()). During system resume, the I2C controller will be available only after runtime PM is re-enabled (in resume_early()). However, this may be too late for some devices. Wake up the controller in the suspend() callback while runtime PM is still enabled. The I2C controller will remain available until the suspend_noirq() callback (pm_runtime_force_suspend()) is called. During resume, the I2C controller can be restored by the resume_noirq() callback (pm_runtime_force_resume()). Finally, the resume() callback re-enables autosuspend. As a result, the I2C controller can remain available until the system enters suspend_noirq() and from resume_noirq(). Signed-off-by: Carlos Song Link: https://lore.kernel.org/r/20241227084736.1323943-1-carlos.song@nxp.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-imx-lpi2c.c | 36 +++++++++++++++++++++++++++++- 1 file changed, 35 insertions(+), 1 deletion(-) diff --git a/drivers/i2c/busses/i2c-imx-lpi2c.c b/drivers/i2c/busses/i2c-imx-lpi2c.c index dc25d7dfe1ab..0d4b3935e687 100644 --- a/drivers/i2c/busses/i2c-imx-lpi2c.c +++ b/drivers/i2c/busses/i2c-imx-lpi2c.c @@ -1318,7 +1318,7 @@ static int lpi2c_imx_probe(struct platform_device *pdev) if (ret) lpi2c_imx->bitrate = I2C_MAX_STANDARD_MODE_FREQ; - ret = devm_request_irq(&pdev->dev, irq, lpi2c_imx_isr, 0, + ret = devm_request_irq(&pdev->dev, irq, lpi2c_imx_isr, IRQF_NO_SUSPEND, pdev->name, lpi2c_imx); if (ret) return dev_err_probe(&pdev->dev, ret, "can't claim irq %d\n", irq); @@ -1447,9 +1447,43 @@ static int __maybe_unused lpi2c_resume_noirq(struct device *dev) return 0; } +static int lpi2c_suspend(struct device *dev) +{ + /* + * Some I2C devices may need the I2C controller to remain active + * during resume_noirq() or suspend_noirq(). If the controller is + * autosuspended, there is no way to wake it up once runtime PM is + * disabled (in suspend_late()). + * + * During system resume, the I2C controller will be available only + * after runtime PM is re-enabled (in resume_early()). However, this + * may be too late for some devices. + * + * Wake up the controller in the suspend() callback while runtime PM + * is still enabled. The I2C controller will remain available until + * the suspend_noirq() callback (pm_runtime_force_suspend()) is + * called. During resume, the I2C controller can be restored by the + * resume_noirq() callback (pm_runtime_force_resume()). + * + * Finally, the resume() callback re-enables autosuspend, ensuring + * the I2C controller remains available until the system enters + * suspend_noirq() and from resume_noirq(). + */ + return pm_runtime_resume_and_get(dev); +} + +static int lpi2c_resume(struct device *dev) +{ + pm_runtime_mark_last_busy(dev); + pm_runtime_put_autosuspend(dev); + + return 0; +} + static const struct dev_pm_ops lpi2c_pm_ops = { SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(lpi2c_suspend_noirq, lpi2c_resume_noirq) + SYSTEM_SLEEP_PM_OPS(lpi2c_suspend, lpi2c_resume) SET_RUNTIME_PM_OPS(lpi2c_runtime_suspend, lpi2c_runtime_resume, NULL) }; From 65cba48d295216d0636ee835631142099374d905 Mon Sep 17 00:00:00 2001 From: Philipp Stanner Date: Thu, 21 Nov 2024 20:56:25 +0100 Subject: [PATCH 195/196] i2c: i801: Remove unnecessary PCI function call Since the changes in commit f748a07a0b64 ("PCI: Remove legacy pcim_release()") all pcim_enable_device() does is set up a callback that disables the device from being disabled from driver detach. The function pcim_pin_device() prevents said disabling. pcim_enable_device(), therefore, sets up an action that is removed immediately afterwards by pcim_pin_device(). Replace pcim_enable_device() with pci_enable_device() and remove the unnecessary call to pcim_pin_device(). Signed-off-by: Philipp Stanner Link: https://lore.kernel.org/r/20241121195624.144839-2-pstanner@redhat.com Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-i801.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/drivers/i2c/busses/i2c-i801.c b/drivers/i2c/busses/i2c-i801.c index 75dab01d43a7..d9d354ce231d 100644 --- a/drivers/i2c/busses/i2c-i801.c +++ b/drivers/i2c/busses/i2c-i801.c @@ -1682,13 +1682,16 @@ static int i801_probe(struct pci_dev *dev, const struct pci_device_id *id) if (!(priv->features & FEATURE_BLOCK_BUFFER)) priv->features &= ~FEATURE_BLOCK_PROC; - err = pcim_enable_device(dev); + /* + * Do not call pcim_enable_device(), because the device has to remain + * enabled on driver detach. See i801_remove() for the reasoning. + */ + err = pci_enable_device(dev); if (err) { dev_err(&dev->dev, "Failed to enable SMBus PCI device (%d)\n", err); return err; } - pcim_pin_device(dev); /* Determine the address of the SMBus area */ priv->smba = pci_resource_start(dev, SMBBAR); From fd4e9308312972d5bd11fe1d8c032b95b44748cb Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Patrick=20H=C3=B6hn?= Date: Tue, 12 Mar 2024 20:31:31 +0100 Subject: [PATCH 196/196] i2c: i801: Add lis3lv02d for Dell Precision M6800 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On the Dell Precision M6800/OXD1M5, BIOS A26 06/13/2029, Linux prints the warning below. i801_smbus 0000:00:1f.4: Accelerometer lis3lv02d is present on SMBus but its address is unknown, skipping registration Following the same suggestions by Wolfram Sang as for the Dell Precision 3540 [1], the accelerometer can be successfully found on I2C bus 0 at address 0x29. $ echo lis3lv02d 0x29 | sudo tee /sys/bus/i2c/devices/i2c-0/new_device lis3lv02d 0x29 $ dmesg | tail -5 [1185.385204] lis3lv02d_i2c 0-0029: supply Vdd not found, using dummy regulator [1185.385235] lis3lv02d_i2c 0-0029: supply Vdd_IO not found, using dummy regulator [1185.399689] lis3lv02d: 8 bits 3DC sensor found [1185.449391] input: ST LIS3LV02DL Accelerometer as /devices/platform/lis3lv02d/input/input371 [1185.449577] i2c i2c-0: new_device: Instantiated device lis3lv02d at 0x29 So, the device has that accelerometer. Add the I2C address to the mapping list, and test it successfully on the device. [1]: https://lore.kernel.org/linux-i2c/97708c11-ac85-fb62-2c8e-d37739ca826f@molgen.mpg.de/ Signed-off-by: Patrick Höhn Acked-by: Pali Rohár Link: https://lore.kernel.org/r/20240312193132.26518-1-hoehnp@gmx.de Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-i801.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/i2c/busses/i2c-i801.c b/drivers/i2c/busses/i2c-i801.c index d9d354ce231d..b6d1140eb678 100644 --- a/drivers/i2c/busses/i2c-i801.c +++ b/drivers/i2c/busses/i2c-i801.c @@ -1252,6 +1252,7 @@ static const struct { */ { "Latitude 5480", 0x29 }, { "Precision 3540", 0x29 }, + { "Precision M6800", 0x29 }, { "Vostro V131", 0x1d }, { "Vostro 5568", 0x29 }, { "XPS 15 7590", 0x29 },