linux/kernel/rcu
Yao Kai d41e37f26b rcu: Fix rcu_read_unlock() deadloop due to softirq
Commit 5f5fa7ea89 ("rcu: Don't use negative nesting depth in
__rcu_read_unlock()") removes the recursion-protection code from
__rcu_read_unlock(). Therefore, we could invoke the deadloop in
raise_softirq_irqoff() with ftrace enabled as follows:

WARNING: CPU: 0 PID: 0 at kernel/trace/trace.c:3021 __ftrace_trace_stack.constprop.0+0x172/0x180
Modules linked in: my_irq_work(O)
CPU: 0 UID: 0 PID: 0 Comm: swapper/0 Tainted: G O 6.18.0-rc7-dirty #23 PREEMPT(full)
Tainted: [O]=OOT_MODULE
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04/01/2014
RIP: 0010:__ftrace_trace_stack.constprop.0+0x172/0x180
RSP: 0018:ffffc900000034a8 EFLAGS: 00010002
RAX: 0000000000000000 RBX: 0000000000000004 RCX: 0000000000000000
RDX: 0000000000000003 RSI: ffffffff826d7b87 RDI: ffffffff826e9329
RBP: 0000000000090009 R08: 0000000000000005 R09: ffffffff82afbc4c
R10: 0000000000000008 R11: 0000000000011d7a R12: 0000000000000000
R13: ffff888003874100 R14: 0000000000000003 R15: ffff8880038c1054
FS:  0000000000000000(0000) GS:ffff8880fa8ea000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000055b31fa7f540 CR3: 00000000078f4005 CR4: 0000000000770ef0
PKRU: 55555554
Call Trace:
 <IRQ>
 trace_buffer_unlock_commit_regs+0x6d/0x220
 trace_event_buffer_commit+0x5c/0x260
 trace_event_raw_event_softirq+0x47/0x80
 raise_softirq_irqoff+0x6e/0xa0
 rcu_read_unlock_special+0xb1/0x160
 unwind_next_frame+0x203/0x9b0
 __unwind_start+0x15d/0x1c0
 arch_stack_walk+0x62/0xf0
 stack_trace_save+0x48/0x70
 __ftrace_trace_stack.constprop.0+0x144/0x180
 trace_buffer_unlock_commit_regs+0x6d/0x220
 trace_event_buffer_commit+0x5c/0x260
 trace_event_raw_event_softirq+0x47/0x80
 raise_softirq_irqoff+0x6e/0xa0
 rcu_read_unlock_special+0xb1/0x160
 unwind_next_frame+0x203/0x9b0
 __unwind_start+0x15d/0x1c0
 arch_stack_walk+0x62/0xf0
 stack_trace_save+0x48/0x70
 __ftrace_trace_stack.constprop.0+0x144/0x180
 trace_buffer_unlock_commit_regs+0x6d/0x220
 trace_event_buffer_commit+0x5c/0x260
 trace_event_raw_event_softirq+0x47/0x80
 raise_softirq_irqoff+0x6e/0xa0
 rcu_read_unlock_special+0xb1/0x160
 unwind_next_frame+0x203/0x9b0
 __unwind_start+0x15d/0x1c0
 arch_stack_walk+0x62/0xf0
 stack_trace_save+0x48/0x70
 __ftrace_trace_stack.constprop.0+0x144/0x180
 trace_buffer_unlock_commit_regs+0x6d/0x220
 trace_event_buffer_commit+0x5c/0x260
 trace_event_raw_event_softirq+0x47/0x80
 raise_softirq_irqoff+0x6e/0xa0
 rcu_read_unlock_special+0xb1/0x160
 __is_insn_slot_addr+0x54/0x70
 kernel_text_address+0x48/0xc0
 __kernel_text_address+0xd/0x40
 unwind_get_return_address+0x1e/0x40
 arch_stack_walk+0x9c/0xf0
 stack_trace_save+0x48/0x70
 __ftrace_trace_stack.constprop.0+0x144/0x180
 trace_buffer_unlock_commit_regs+0x6d/0x220
 trace_event_buffer_commit+0x5c/0x260
 trace_event_raw_event_softirq+0x47/0x80
 __raise_softirq_irqoff+0x61/0x80
 __flush_smp_call_function_queue+0x115/0x420
 __sysvec_call_function_single+0x17/0xb0
 sysvec_call_function_single+0x8c/0xc0
 </IRQ>

Commit b41642c877 ("rcu: Fix rcu_read_unlock() deadloop due to IRQ work")
fixed the infinite loop in rcu_read_unlock_special() for IRQ work by
setting a flag before calling irq_work_queue_on(). We fix this issue by
setting the same flag before calling raise_softirq_irqoff() and rename the
flag to defer_qs_pending for more common.

Fixes: 5f5fa7ea89 ("rcu: Don't use negative nesting depth in __rcu_read_unlock()")
Reported-by: Tengda Wu <wutengda2@huawei.com>
Signed-off-by: Yao Kai <yaokai34@huawei.com>
Reviewed-by: Joel Fernandes <joelagnelf@nvidia.com>
Tested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Joel Fernandes <joelagnelf@nvidia.com>
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
2026-01-07 21:58:37 +08:00
..
Kconfig srcu: Make FORCE_NEED_SRCU_NMI_SAFE depend on RCU_EXPERT 2025-03-28 21:19:17 -07:00
Kconfig.debug rcu: Add a small-width RCU watching counter debug option 2025-10-28 17:32:56 +01:00
Makefile rcuperf: Change rcuperf to rcuscale 2020-08-24 18:39:24 -07:00
rcu_segcblist.c rcu/nocb: Simplify (de-)offloading state machine 2024-09-09 00:03:55 +05:30
rcu_segcblist.h rcu: Remove unused declaration rcu_segcblist_offload() 2024-10-22 15:36:56 +02:00
rcu.h Merge branches 'rcu/misc-for-6.16', 'rcu/seq-counters-for-6.16' and 'rcu/torture-for-6.16' into rcu/for-next 2025-05-16 11:18:16 -04:00
rcuscale.c mm: replace (20 - PAGE_SHIFT) with common macros for pages<->MB conversion 2025-09-13 16:54:42 -07:00
rcutorture.c rcutorture: Correctly compute probability to invoke ->exp_current() 2026-01-07 21:58:34 +08:00
refscale.c Merge branch 'rcu/refscale' into next 2025-11-28 23:30:38 +01:00
srcutiny.c srcu: Permit Tiny SRCU srcu_read_unlock() with interrupts disabled 2025-11-05 23:58:11 +01:00
srcutree.c srcu: Create an SRCU-fast-updown API 2025-11-27 14:22:31 +01:00
sync.c rcu: Eliminate lockless accesses to rcu_sync->gp_count 2024-07-04 13:48:57 -07:00
tasks.h rcu: replace use of system_wq with system_percpu_wq 2025-09-23 02:01:18 -07:00
tiny.c sched: Provide and use set_need_resched_current() 2025-11-20 22:26:09 +01:00
tree_exp.h rcu: Make expedited RCU CPU stall warnings detect stall-end races 2026-01-07 21:58:26 +08:00
tree_nocb.h rcu: Fix delayed execution of hurry callbacks 2025-07-18 09:25:34 +05:30
tree_plugin.h rcu: Fix rcu_read_unlock() deadloop due to softirq 2026-01-07 21:58:37 +08:00
tree_stall.h sched: Provide and use set_need_resched_current() 2025-11-20 22:26:09 +01:00
tree.c RCU pull request for v6.19 2025-12-03 12:18:07 -08:00
tree.h rcu: Fix rcu_read_unlock() deadloop due to softirq 2026-01-07 21:58:37 +08:00
update.c rcu: Mark diagnostic functions as notrace 2025-11-07 13:57:37 +01:00