Tegra Endpoint exposes three 64-bit BARs at indices 0, 2, and 4:
- BAR0+BAR1: EPF test/data (programmable 64-bit BAR)
- BAR2+BAR3: MSI-X table (hardware-backed)
- BAR4+BAR5: DMA registers (hardware-backed)
Update tegra_pcie_epc_features so that BAR2 is BAR_RESERVED with
PCI_EPC_BAR_RSVD_MSIX_TBL_RAM (64 KB) & PCI_EPC_BAR_RSVD_MSIX_PBA_RAM
(64 KB) and BAR4 is BAR_RESERVED with PCI_EPC_BAR_RSVD_DMA_CTRL_MMIO (4KB).
This keeps CONSECUTIVE_BAR_TEST working while allowing the host to use
64-bit BAR2 (MSI-X) and BAR4 (DMA).
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260324080857.916263-4-mmaddireddy@nvidia.com
The Tegra194/234 Endpoint does not support the Resizable BAR capability,
but BAR0 can be programmed to different sizes via the DBI2 BAR registers
in dw_pcie_ep_set_bar_programmable(). The BAR0 size is set once during
initialization.
Remove the fixed 1MB limit from pci_epc_features so Endpoint function
drivers can configure the BAR0 size they need.
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260324080857.916263-3-mmaddireddy@nvidia.com
Add PCI_EPC_BAR_RSVD_MSIX_TBL_RAM and PCI_EPC_BAR_RSVD_MSIX_PBA_RAM to
enum pci_epc_bar_rsvd_region_type so that Endpoint controllers can
describe hardware-owned MSI-X Table and PBA (Pending Bit Array) regions
behind a BAR_RESERVED BAR.
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260324080857.916263-2-mmaddireddy@nvidia.com
The sub_size check ensures that each subrange is large enough for 32-bit
accesses. Subranges smaller than sizeof(u32) do not satisfy this
assumption, so this is a local sanity check rather than a resource
exhaustion case.
Return -EINVAL instead of -ENOSPC for this case.
Suggested-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260320140139.2415480-1-den@valinux.co.jp
When vNTB is used as a PCI endpoint function, the NTB device is backed
by a virtual PCI function. For DMA API allocations and mappings, NTB
clients must use the device that is associated with the IOMMU domain.
Implement ntb_dev_ops->get_dma_dev() for pci-epf-vntb and return the EPC
parent device.
Suggested-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://patch.msgid.link/20260306031443.1911860-4-den@valinux.co.jp
ntb_transport currently uses ndev->pdev->dev for coherent allocations
and frees.
Switch the coherent buffer allocation/free paths to use
ntb_get_dma_dev(), so ntb_transport can work with NTB implementations
where the NTB PCI function is not the right device to use for DMA
mappings.
Suggested-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://patch.msgid.link/20260306031443.1911860-3-den@valinux.co.jp
Some NTB implementations are backed by a PCI function that is not the right
struct device to use with DMA API helpers (e.g. due to IOMMU topology, or
because the NTB device is virtual).
Add an optional .get_dma_dev() callback to struct ntb_dev_ops and provide a
helper, ntb_get_dma_dev(), so NTB clients can use the appropriate struct
device for DMA allocations and mappings.
If the callback is not implemented, ntb_get_dma_dev() returns the current
default (ntb->dev.parent). Drivers that implement .get_dma_dev() must
return a non-NULL device.
Suggested-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
[bhelgaas: format doc]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://patch.msgid.link/20260306031443.1911860-2-den@valinux.co.jp
Include errno in the error prints and also use dev_err() where applicable.
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@oss.qualcomm.com>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20260305071236.23792-2-mani@kernel.org
Merely printing the error log without the actual EPF name will not give
much clue to the users about the failure. Hence, print the EPF name also.
Suggested-by: Bjorn Helgaas <helgaas@kernel.org>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@oss.qualcomm.com>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20260305071236.23792-1-mani@kernel.org
Running a test against a reserved BAR will result in the pci-epf-test
driver returning -ENOBUFS.
Make sure that the pci_endpoint_test selftest will return skip instead of
failure or success for reserved BARs.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Tested-by: Koichiro Den <den@valinux.co.jp>
Reviewed-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260312130229.2282001-22-cassel@kernel.org
Give reserved BARs a distinct error code, such that the pci_endpoint_test
selftest will be able to skip test cases that are run against reserved
BARs.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
[mani: Used __fls(CAP_BAR0_RESERVED) instead of PCI_ENDPOINT_CAP_BAR0_RESERVED_BIT]
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Tested-by: Koichiro Den <den@valinux.co.jp>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260312130229.2282001-21-cassel@kernel.org
Advertise reserved BARs as reserved in the Capabilities register,
such that the host side driver will be able to skip reserved BARs.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Koichiro Den <den@valinux.co.jp>
Tested-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260312130229.2282001-20-cassel@kernel.org
The current EPC core design relies on an EPC driver disabling all BARs by
default. An EPF driver will then enable the BARs that it wants to enabled.
This design is there because there is no epc->ops->disable_bar().
(There is a epc->ops->clear_bar(), but that is only to disable a BAR that
has been enabled using epc->ops->set_bar() first.)
By default, an EPF driver will not be able to get/enable BARs that are
marked as BAR_RESERVED or BAR_DISABLED (see pci_epc_get_next_free_bar()).
Since the current EPC code design requires an EPC driver to disable all
BARs by default, move this to DWC common code from each glue driver.
BAR_RESERVED BARs are not disabled by default because these BARs are
hardware backed, and should only be disabled explicitly by an EPF driver
if absolutely necessary for the EPF driver to function correctly.
(This is similar to how e.g. NVMe may have vendor specific BARs outside of
the mandatory BAR0 which contains the NVMe registers.)
Note that there is currently no EPC operation to disable a BAR that has not
first been programmed using pci_epc_set_bar(). If an EPF driver ever wants
to disable a BAR marked as BAR_RESERVED, a disable_bar() operation would
have to be added first.
No functional changes intended.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Tested-by: Koichiro Den <den@valinux.co.jp>
Reviewed-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260312130229.2282001-19-cassel@kernel.org
Most DWC based EPC glue drivers that have BARs marked as BAR_RESERVED in
epc_features also call dw_pcie_ep_reset_bar() for these reserved BARs in
ep->ops->init(). (The only exception is pci-keystone.c.)
An EPF driver will be able to get/enable BARs that have been disabled/reset
using dw_pcie_ep_reset_bar(), except if the BAR is marked as BAR_RESERVED
(see pci_epc_get_next_free_bar()).
Thus, all EPC drivers that have BARs marked as BAR_RESERVED in epc_features
and call dw_pcie_ep_reset_bar(), should really have these BARs marked as
BAR_DISABLED. If dw_pcie_ep_reset_bar() is not called by the glue driver,
the BARs are kept as BAR_RESERVED.
No EPC drivers outside drivers/pci/controllers/dwc mark their BARs as
BAR_RESERVED, so there is nothing to do in non-DWC based EPC drivers.
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Tested-by: Koichiro Den <den@valinux.co.jp>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260312130229.2282001-18-cassel@kernel.org
Add a pci_epc_bar_type BAR_DISABLED to more clearly differentiate from
BAR_RESERVED.
This BAR type will only be used to describe a BAR that the EPC driver
should disable, and will thus never be available to an EPF driver.
(Unlike BAR_RESERVED, which will never be disabled by default by an EPC
driver.)
Co-developed-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Koichiro Den <den@valinux.co.jp>
Tested-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Link: https://patch.msgid.link/20260312130229.2282001-17-cassel@kernel.org
On RK3588 PCIe3_4L in EP mode, the integrated DMA controller registers
are permanently mapped to BAR4 and must not be repurposed by EPF
drivers.
When the remote peer needs to access these registers, it must use the
fixed BAR4 window instead of creating another inbound mapping in a
different BAR. Mixing the fixed window with an additional mapping can
lead to incorrect behavior.
Advertise the DMA controller MMIO window as a reserved BAR subregion so
EPF drivers can reuse it safely.
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Koichiro Den <den@valinux.co.jp>
Reviewed-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260312130229.2282001-16-cassel@kernel.org
Some endpoint controllers expose platform-owned, fixed register windows
within a BAR that EPF drivers must not reprogram (e.g. a BAR marked
BAR_RESERVED). Even in that case, EPF drivers may need to reference a
well-defined subset of that BAR, e.g. to reuse an integrated DMA
controller MMIO window as a doorbell target.
Introduce struct pci_epc_bar_rsvd_region and extend struct
pci_epc_bar_desc so EPC drivers can advertise such fixed subregions in a
controller-agnostic way.
No functional change for existing users.
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Tested-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Tested-by: Koichiro Den <den@valinux.co.jp>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260312130229.2282001-15-cassel@kernel.org
Remove the documentation that forbids setting only_64bit on a BAR of type
BAR_RESERVED.
When a reserved BAR is 64-bit by default, setting only_64bit is the most
accurate description. If we later add support to disable a reserved BAR
(e.g. disable_bar() for BARs that were never set via set_bar()), the
implementation will need to clear the adjacent BAR (upper 32 bits) as well;
having only_64bit set documents that requirement.
Signed-off-by: Manikanta Maddireddy <mmaddireddy@nvidia.com>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20260312130229.2282001-14-cassel@kernel.org
A BAR that can only be configured as a 64-bit BAR by an EPC driver is
marked as such using the "only_64bit" flag.
Currently, the documentation says that an EPC driver should explicitly
mark the BAR succeeding an "only_64bit" BAR as BAR_RESERVED.
However, a 64-bit BAR will always take up two BARs. It is thus redundant
to mark both BARs.
pci_epc_get_next_free_bar() already skips the BAR succeeding a "only_64bit"
BAR, regardless if the succeeding BAR is marked as BAR_RESERVED or not.
Thus, drop the BAR_RESERVED for a BAR succeeding a "only_64bit" BAR.
No functional changes.
Suggested-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20260312130229.2282001-13-cassel@kernel.org
pci_epf_make() overwrites the actual error returned by pci_epf_create()
with -EINVAL, which hides the real failure reason. Use PTR_ERR(epf) instead
and print the error code.
Signed-off-by: Alok Tiwari <alok.a.tiwari@oracle.com>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20260227030701.40533-1-alok.a.tiwari@oracle.com
The function description in kernel-doc refers to pci_epf_add_epf(), but the
correct function name is pci_epf_add_vepf(). Update it to match the
implementation.
Signed-off-by: Alok Tiwari <alok.a.tiwari@oracle.com>
[mani: commit log]
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20260224183927.1369124-1-alok.a.tiwari@oracle.com
Disable the delayed work before clearing BAR mappings and doorbells to
avoid running the handler after resources have been torn down.
Unable to handle kernel paging request at virtual address ffff800083f46004
[...]
Internal error: Oops: 0000000096000007 [#1] SMP
[...]
Call trace:
epf_ntb_cmd_handler+0x54/0x200 [pci_epf_vntb] (P)
process_one_work+0x154/0x3b0
worker_thread+0x2c8/0x400
kthread+0x148/0x210
ret_from_fork+0x10/0x20
Fixes: e35f56bb03 ("PCI: endpoint: Support NTB transfer between RC and EP")
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260226084142.2226875-4-den@valinux.co.jp
epf_ntb_epc_destroy() duplicates the teardown that the caller is
supposed to do later. This leads to an oops when .allow_link fails or
when .drop_link is performed. Remove the helper.
Also drop pci_epc_put(). EPC device refcounting is tied to configfs EPC
group lifetime, and pci_epc_put() in the .drop_link path is sufficient.
Fixes: 8b821cf761 ("PCI: endpoint: Add EP function driver to provide NTB functionality")
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260226084142.2226875-3-den@valinux.co.jp
epf_ntb_epc_destroy() duplicates the teardown that the caller is
supposed to perform later. This leads to an oops when .allow_link fails
or when .drop_link is performed. The following is an example oops of the
former case:
Unable to handle kernel paging request at virtual address dead000000000108
[...]
[dead000000000108] address between user and kernel address ranges
Internal error: Oops: 0000000096000044 [#1] SMP
[...]
Call trace:
pci_epc_remove_epf+0x78/0xe0 (P)
pci_primary_epc_epf_link+0x88/0xa8
configfs_symlink+0x1f4/0x5a0
vfs_symlink+0x134/0x1d8
do_symlinkat+0x88/0x138
__arm64_sys_symlinkat+0x74/0xe0
[...]
Remove the helper, and drop pci_epc_put(). EPC device refcounting is
tied to the configfs EPC group lifetime, and pci_epc_put() in the
.drop_link path is sufficient.
Fixes: e35f56bb03 ("PCI: endpoint: Support NTB transfer between RC and EP")
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260226084142.2226875-2-den@valinux.co.jp
pci_epf_mhi_edma_read() and pci_epf_mhi_edma_write() start DMA
operations and wait for completion with a timeout.
On successful completion, they previously returned the remaining
timeout, which callers may treat as an error. In particular,
mhi_ep_ring_add_element(), which calls pci_epf_mhi_edma_write() via
mhi_cntrl->write_sync(), interprets any non-zero return value as
failure.
Return 0 on success instead of the remaining timeout to prevent
mhi_ep_ring_add_element() from treating successful completion as an
error.
Fixes: 7b99aaadda ("PCI: epf-mhi: Add eDMA support")
Signed-off-by: Daniel Hodges <git@danielhodges.dev>
[mani: changed commit log as per https://lore.kernel.org/linux-pci/20260227191510.GA3904799@bhelgaas]
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Krishna Chaitanya Chundru <krishna.chundru@oss.qualcomm.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/20260206200529.10784-1-git@danielhodges.dev
The mw1..mw4 configfs attributes are only valid when the MW index is
within the configured num_mws range. Return -ERANGE instead of -EINVAL
when a configfs MW size attribute is accessed with an out-of-range MW
index.
Suggested-by: Manivannan Sadhasivam <mani@kernel.org>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/20260219145633.4191325-1-den@valinux.co.jp
Follow common kernel idioms for indices derived from configfs attributes
and suppress Smatch warnings:
epf_ntb_mw1_show() warn: potential spectre issue 'ntb->mws_size' [r]
epf_ntb_mw1_store() warn: potential spectre issue 'ntb->mws_size' [w]
Also fix the error message for out-of-range MW indices and %lld format
for unsigned values.
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20260105075606.1253697-1-den@valinux.co.jp
'struct configfs_item_operations' and 'configfs_group_operations' are not
modified in this driver.
Constifying these structures moves some data to a read-only section, so
increases overall security, especially when the structure holds some
function pointers.
On a x86_64, with allmodconfig:
Before:
======
text data bss dec hex filename
27503 12184 256 39943 9c07 drivers/pci/endpoint/pci-ep-cfs.o
After:
=====
text data bss dec hex filename
27855 11832 256 39943 9c07 drivers/pci/endpoint/pci-ep-cfs.o
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Link: https://patch.msgid.link/f1f05f1c10c6caf37dd620fa12f508c53536996b.1765705512.git.christophe.jaillet@wanadoo.fr
PCITEST_DOORBELL may return -EOPNOTSUPP when the endpoint does not
advertise CAP_DYNAMIC_INBOUND_MAPPING.
Treat this like other optional capabilities and skip the doorbell test
instead of reporting a failure.
Suggested-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260215150334.3391943-4-den@valinux.co.jp
The doorbell test relies on the endpoint being able to update inbound
translations at runtime so the host can reach the doorbell target
through a BAR mapping.
If the endpoint does not advertise CAP_DYNAMIC_INBOUND_MAPPING, return
-EOPNOTSUPP from PCITEST_DOORBELL.
This avoids confusing failures in user space and kselftests when the
required capability is not available.
Suggested-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260215150334.3391943-3-den@valinux.co.jp
The doorbell test requires the EPC driver to support dynamic inbound
mapping so the host can map the doorbell target address into a BAR
aperture.
Expose epc_features->dynamic_inbound_mapping via a new
CAP_DYNAMIC_INBOUND_MAPPING bit in the pci-epf-test capability register,
so the host-side pci_endpoint_test driver can detect missing support and
return -EOPNOTSUPP instead of running the test fruitlessly.
Suggested-by: Niklas Cassel <cassel@kernel.org>
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260215150334.3391943-2-den@valinux.co.jp
pci_epf_alloc_doorbell() stores the allocated doorbell message array in
epf->db_msg/epf->num_db before requesting MSI vectors. If MSI allocation
fails, the array is freed but the EPF state may still point to freed
memory.
Clear epf->db_msg and epf->num_db on the MSI allocation failure path so
that later cleanup cannot double-free the array and callers can retry
allocation.
Also return -EBUSY when doorbells have already been allocated to prevent
leaking or overwriting an existing allocation.
Fixes: 1c3b002c6b ("PCI: endpoint: Add RC-to-EP doorbell support using platform MSI controller")
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260217063856.3759713-4-den@valinux.co.jp
pci_epf_test_doorbell_cleanup() unconditionally calls free_irq() for the
doorbell virq, which can trigger "Trying to free already-free IRQ"
warnings when the IRQ was never requested or when request_threaded_irq()
failed.
Move free_irq() out of pci_epf_test_doorbell_cleanup() and invoke it
only after a successful request, so that free_irq() is not called for
an unrequested IRQ.
Fixes: eff0c286aa ("PCI: endpoint: pci-epf-test: Add doorbell test support")
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260217063856.3759713-3-den@valinux.co.jp
epf_ntb_db_bar_init_msi_doorbell() requests ntb->db_count doorbell IRQs
and then performs additional MSI doorbell setup that may still fail.
The error path unwinds the requested IRQs, but it uses a loop variable
that is reused later in the function. When a later step fails, the
unwind can run with an unexpected index value and leave some IRQs
requested.
Track the number of successfully requested IRQs separately and use that
counter for the unwind so all previously requested IRQs are freed on
failure.
Fixes: dc693d6066 ("PCI: endpoint: pci-epf-vntb: Add MSI doorbell support")
Signed-off-by: Koichiro Den <den@valinux.co.jp>
Signed-off-by: Manivannan Sadhasivam <mani@kernel.org>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Reviewed-by: Niklas Cassel <cassel@kernel.org>
Link: https://patch.msgid.link/20260217063856.3759713-2-den@valinux.co.jp
- Fix a build error on parisc
- Remove the non-large-folio-aware function fsverity_verify_page()
-----BEGIN PGP SIGNATURE-----
iIoEABYIADIWIQSacvsUNc7UX4ntmEPzXCl4vpKOKwUCaZtg+xQcZWJpZ2dlcnNA
a2VybmVsLm9yZwAKCRDzXCl4vpKOKwQ+AQCiXEYAibl3vHRgQo7qEPCC5or4FtkF
HZ0ERRArhsU17AD/TKHE/AJkyFrwK4rGTb6I9Wi1OXnpG7jihZlYjj03Ag4=
=CUql
-----END PGP SIGNATURE-----
Merge tag 'fsverity-for-linus' of git://git.kernel.org/pub/scm/fs/fsverity/linux
Pull fsverity fixes from Eric Biggers:
- Fix a build error on parisc
- Remove the non-large-folio-aware function fsverity_verify_page()
* tag 'fsverity-for-linus' of git://git.kernel.org/pub/scm/fs/fsverity/linux:
fsverity: fix build error by adding fsverity_readahead() stub
fsverity: remove fsverity_verify_page()
f2fs: make f2fs_verify_cluster() partially large-folio-aware
f2fs: remove unnecessary ClearPageUptodate in f2fs_verify_cluster()
Fix a big endian specific issue in the PPC64-optimized AES code.
-----BEGIN PGP SIGNATURE-----
iIoEABYIADIWIQSacvsUNc7UX4ntmEPzXCl4vpKOKwUCaZtiphQcZWJpZ2dlcnNA
a2VybmVsLm9yZwAKCRDzXCl4vpKOK/z2AQD1j15Ao3iDW3yBSyTS+tFJaRUVDtjg
bNostoNJAjMM9wD/X0oAnt95WTNkdHexs+2aMzQ4ULKFRcwQrUTarKx7IgY=
=Dmfx
-----END PGP SIGNATURE-----
Merge tag 'libcrypto-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux
Pull crypto library fix from Eric Biggers:
"Fix a big endian specific issue in the PPC64-optimized AES code"
* tag 'libcrypto-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiggers/linux:
lib/crypto: powerpc/aes: Fix rndkey_from_vsx() on big endian CPUs
Stephen retired and stepped back from -next maintainership, update his
entry in CREDITS to recognise his 18 years of hard work making it what
it is today and all the impact it's had on our development process.
Also update to his current GnuPG key while we're here.
Acked-by: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: SeongJae Park <sj@kernel.org>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
Acked-by: Krzysztof Kozlowski <krzk@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The x509 public key code gained a dependency on the sha256 hash
implementation, causing a rare link time failure in randconfig
builds:
arm-linux-gnueabi-ld: crypto/asymmetric_keys/x509_public_key.o: in function `x509_get_sig_params':
x509_public_key.c:(.text.x509_get_sig_params+0x12): undefined reference to `sha256'
arm-linux-gnueabi-ld: (sha256): Unknown destination type (ARM/Thumb) in crypto/asymmetric_keys/x509_public_key.o
x509_public_key.c:(.text.x509_get_sig_params+0x12): dangerous relocation: unsupported relocation
Select the necessary library code from Kconfig.
Fixes: 2c62068ac8 ("x509: Separately calculate sha256 for blacklist")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Eric Biggers <ebiggers@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Align to the commit bf4afc53b7 ("Convert 'alloc_obj' family to use the
new default GFP_KERNEL argument") update the 'kmalloc_obj' declaration
for userspace to fix below compile error:
In file included from arch/arm/boot/compressed/../../../../lib/decompress_unxz.c:241,
from arch/arm/boot/compressed/decompress.c:56:
arch/arm/boot/compressed/../../../../lib/xz/xz_dec_stream.c: In function 'xz_dec_init':
arch/arm/boot/compressed/../../../../lib/xz/xz_dec_stream.c:787:28: error: implicit declaration of function 'kmalloc_obj'; did you mean 'kmalloc'? [-Wimplicit-function-declaration]
787 | struct xz_dec *s = kmalloc_obj(*s);
| ^~~~~~~~~~~
| kmalloc
Signed-off-by: Haiyue Wang <haiyuewa@163.com>
Fixes: 69050f8d6d ("treewide: Replace kmalloc with kmalloc_obj for non-scalar types")
Fixes: bf4afc53b7 ("Convert 'alloc_obj' family to use the new default GFP_KERNEL argument")
Reviewed-by: Kees Cook <kees@kernel.org>
Acked-by: Lasse Collin <lasse.collin@tukaani.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Toolchain and infrastructure:
- Pass '-Zunstable-options' flag required by the future Rust 1.95.0.
- Fix 'objtool' warning for Rust 1.84.0.
'kernel' crate:
- 'irq' module: add missing bound detected by the future Rust 1.95.0.
- 'list' module: add missing 'unsafe' blocks and placeholder safety
comments to macros (an issue for future callers within the crate).
'pin-init' crate:
- Clean Clippy warning that changed behavior in the future Rust 1.95.0.
-----BEGIN PGP SIGNATURE-----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=DOG1
-----END PGP SIGNATURE-----
Merge tag 'rust-fixes-7.0' of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux
Pull rust fixes from Miguel Ojeda:
"Toolchain and infrastructure:
- Pass '-Zunstable-options' flag required by the future Rust 1.95.0
- Fix 'objtool' warning for Rust 1.84.0
'kernel' crate:
- 'irq' module: add missing bound detected by the future Rust 1.95.0
- 'list' module: add missing 'unsafe' blocks and placeholder safety
comments to macros (an issue for future callers within the crate)
'pin-init' crate:
- Clean Clippy warning that changed behavior in the future Rust
1.95.0"
* tag 'rust-fixes-7.0' of git://git.kernel.org/pub/scm/linux/kernel/git/ojeda/linux:
rust: list: Add unsafe blocks for container_of and safety comments
rust: pin-init: replace clippy `expect` with `allow`
rust: irq: add `'static` bounds to irq callbacks
objtool/rust: add one more `noreturn` Rust function
rust: kbuild: pass `-Zunstable-options` for Rust 1.95.0
Summary of changes:
- Fix multiple definition of __pcpu_unique_da_mon_this
After refactoring monitors, we used static per-cpu variables with the
same names across different per-cpu monitors. This is explicitly
disallowed for modules on some architectures (alpha) or if
CONFIG_DEBUG_FORCE_WEAK_PER_CPU is enabled (e.g. Fedora's debug
kernel). Make sure all those variables have different names to avoid
compilation issues.
-----BEGIN PGP SIGNATURE-----
iIoEABYKADIWIQRRSw7ePDh/lE+zeZMp5XQQmuv6qgUCaZoJ1BQccm9zdGVkdEBn
b29kbWlzLm9yZwAKCRAp5XQQmuv6qkykAP41lj9VAe0dM81f20IfOBhqIBZPS6MP
ajRGZ9bz1ZiJwwD/dHm+APheD7umImzoDXpW1IJxgNPlZTnpte3WyhIYaQs=
=HxRz
-----END PGP SIGNATURE-----
Merge tag 'trace-rv-7.0-2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull runtime verifier fix from Steven Rostedt:
- Fix multiple definition of __pcpu_unique_da_mon_this
After refactoring monitors, we used static per-cpu variables with the
same names across different per-cpu monitors. This is explicitly
disallowed for modules on some architectures (alpha) or if
CONFIG_DEBUG_FORCE_WEAK_PER_CPU is enabled (e.g. Fedora's debug
kernel). Make sure all those variables have different names to avoid
compilation issues.
* tag 'trace-rv-7.0-2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
rv: Fix multiple definition of __pcpu_unique_da_mon_this
This converts some of the visually simpler cases that have been split
over multiple lines. I only did the ones that are easy to verify the
resulting diff by having just that final GFP_KERNEL argument on the next
line.
Somebody should probably do a proper coccinelle script for this, but for
me the trivial script actually resulted in an assertion failure in the
middle of the script. I probably had made it a bit _too_ trivial.
So after fighting that far a while I decided to just do some of the
syntactically simpler cases with variations of the previous 'sed'
scripts.
The more syntactically complex multi-line cases would mostly really want
whitespace cleanup anyway.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This is the exact same thing as the 'alloc_obj()' version, only much
smaller because there are a lot fewer users of the *alloc_flex()
interface.
As with alloc_obj() version, this was done entirely with mindless brute
force, using the same script, except using 'flex' in the pattern rather
than 'objs*'.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This was done entirely with mindless brute force, using
git grep -l '\<k[vmz]*alloc_objs*(.*, GFP_KERNEL)' |
xargs sed -i 's/\(alloc_objs*(.*\), GFP_KERNEL)/\1)/'
to convert the new alloc_obj() users that had a simple GFP_KERNEL
argument to just drop that argument.
Note that due to the extreme simplicity of the scripting, any slightly
more complex cases spread over multiple lines would not be triggered:
they definitely exist, but this covers the vast bulk of the cases, and
the resulting diff is also then easier to check automatically.
For the same reason the 'flex' versions will be done as a separate
conversion.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Most simple allocations use GFP_KERNEL, and with the new allocation
helpers being introduced, let's just take advantage of that to simplify
that default case.
It's a numbers game:
git grep 'alloc_obj(' |
sed 's/.*\(GFP_[_A-Z]*\).*/\1/' |
sort | uniq -c | sort -n | tail
shows that about 90% of all those new allocator instances just use that
standard GFP_KERNEL.
Those helpers are already macros, and we can easily just make it be the
default case when the gfp argument is missing.
And yes, we could do that for all the legacy interfaces too, but let's
keep it to just the new ones at least for now, since those all got
converted recently anyway, so this is not any "extra" noise outside of
that limited conversion.
And, in fact, I want to do this before doing the -rc1 release, exactly
so that we don't get extra merge conflicts.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Commit 69050f8d6d ("treewide: Replace kmalloc with kmalloc_obj for
non-scalar types") started using the new allocation helpers, and in the
process showed that they were completely non-working.
The overflow logic in overflows_flex_counter_type() is completely the
wrong way around, and that broke __alloc_flex() completely. By chance,
the resulting code was then such a mess that clang generated
sufficiently garbage code that objtool warned about it all. Which made
it somewhat quicker to narrow things down.
While fixing overflows_flex_counter_type() would presumably fix this
all, I'm excising the whole broken overflow logic from __alloc_flex(),
because we don't want that kind of code in basic allocation functions
anyway.
That (no longer) broken overflows_flex_counter_type() thing needs to be
inserted into the actual __set_flex_counter() logic in the unlikely case
that we ever want this at all. And made conditional.
Fixes: 81cee9166a ("compiler_types: Introduce __flex_counter() and family")
Fixes: 69050f8d6d ("treewide: Replace kmalloc with kmalloc_obj for non-scalar types")
Cc: Kees Cook <kees@kernel.org>
Link: https://lore.kernel.org/all/CAHk-=whEd020BYzGTzYrENjD9Z5_82xx6h8HsQvH5xDSnv0=Hw@mail.gmail.com/
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQRSPkdeREjth1dHnSE2KwveOeQkuwUCaZl14wAKCRA2KwveOeQk
uz8aAQCBFLYlij3Y3ivVADkBxuVF3xECaznFya41ENYsBwlHdwEArXqMyNrw+DiG
TvWCK/tiddNmGIRpI2sxBFzyRpsHfAY=
=rVD3
-----END PGP SIGNATURE-----
Merge tag 'kmalloc_obj-treewide-v7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux
Pull kmalloc_obj conversion from Kees Cook:
"This does the tree-wide conversion to kmalloc_obj() and friends using
coccinelle, with a subsequent small manual cleanup of whitespace
alignment that coccinelle does not handle.
This uncovered a clang bug in __builtin_counted_by_ref(), so the
conversion is preceded by disabling that for current versions of
clang. The imminent clang 22.1 release has the fix.
I've done allmodconfig build tests for x86_64, arm64, i386, and arm. I
did defconfig builds for alpha, m68k, mips, parisc, powerpc, riscv,
s390, sparc, sh, arc, csky, xtensa, hexagon, and openrisc"
* tag 'kmalloc_obj-treewide-v7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
kmalloc_obj: Clean up after treewide replacements
treewide: Replace kmalloc with kmalloc_obj for non-scalar types
compiler_types: Disable __builtin_counted_by_ref for Clang
