linux-stable.git/drivers/pci/controller, branch v5.10.239

PCI: cadence: Fix runtime atomic count underflow

2025-06-27T10:04:06+00:00

[ Upstream commit 8805f32a96d3b97cef07999fa6f52112678f7e65 ]

If the call to pci_host_probe() in cdns_pcie_host_setup() fails, PM
runtime count is decremented in the error path using pm_runtime_put_sync().
But the runtime count is not incremented by this driver, but only by the
callers (cdns_plat_pcie_probe/j721e_pcie_probe). And the callers also
decrement the runtime PM count in their error path. So this leads to the
below warning from the PM core:

	"runtime PM usage count underflow!"

So fix it by getting rid of pm_runtime_put_sync() in the error path and
directly return the errno.

Fixes: 49e427e6bdd1 ("Merge branch 'pci/host-probe-refactor'")
Signed-off-by: Hans Zhang <18255117159@163.com>
Signed-off-by: Manivannan Sadhasivam 
Link: https://patch.msgid.link/20250419133058.162048-1-18255117159@163.com
Signed-off-by: Sasha Levin

PCI: brcmstb: Add a softdep to MIP MSI-X driver

2025-06-04T12:37:02+00:00

[ Upstream commit 2294059118c550464dd8906286324d90c33b152b ]

Then the brcmstb PCIe driver and MIP MSI-X interrupt controller
drivers are built as modules there could be a race in probing.

To avoid this, add a softdep to MIP driver to guarantee that
MIP driver will be load first.

Signed-off-by: Stanimir Varbanov 
Reviewed-by: Florian Fainelli 
Tested-by: Ivan T. Ivanov 
Link: https://lore.kernel.org/r/20250224083559.47645-5-svarbanov@suse.de
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński 
Signed-off-by: Sasha Levin

PCI: brcmstb: Expand inbound window size up to 64GB

2025-06-04T12:37:02+00:00

[ Upstream commit 25a98c727015638baffcfa236e3f37b70cedcf87 ]

The BCM2712 memory map can support up to 64GB of system memory, thus
expand the inbound window size in calculation helper function.

The change is safe for the currently supported SoCs that have smaller
inbound window sizes.

Signed-off-by: Stanimir Varbanov 
Reviewed-by: Florian Fainelli 
Reviewed-by: Jim Quinlan 
Tested-by: Ivan T. Ivanov 
Link: https://lore.kernel.org/r/20250224083559.47645-7-svarbanov@suse.de
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński 
Signed-off-by: Sasha Levin

PCI: imx6: Skip controller_id generation logic for i.MX7D

2025-06-04T12:36:52+00:00

commit f068ffdd034c93f0c768acdc87d4d2d7023c1379 upstream.

The i.MX7D only has one PCIe controller, so controller_id should always be
0. The previous code is incorrect although yielding the correct result.

Fix by removing "IMX7D" from the switch case branch.

Fixes: 2d8ed461dbc9 ("PCI: imx6: Add support for i.MX8MQ")
Link: https://lore.kernel.org/r/20241126075702.4099164-5-hongxing.zhu@nxp.com
Signed-off-by: Richard Zhu 
Signed-off-by: Krzysztof Wilczyński 
Signed-off-by: Bjorn Helgaas 
Reviewed-by: Manivannan Sadhasivam 
Reviewed-by: Frank Li 
[Because this switch case does more than just controller_id
 logic, move the "IMX7D" case label instead of removing it entirely.]
Signed-off-by: Ryan Matthews 
Signed-off-by: Greg Kroah-Hartman

PCI: brcmstb: Fix missing of_node_put() in brcm_pcie_probe()

2025-05-02T05:40:57+00:00

commit 2df181e1aea4628a8fd257f866026625d0519627 upstream.

A call to of_parse_phandle() is incrementing the refcount, and as such,
the of_node_put() must be called when the reference is no longer needed.

Thus, refactor the existing code and add a missing of_node_put() call
following the check to ensure that "msi_np" matches "pcie->np" and after
MSI initialization, but only if the MSI support is enabled system-wide.

Cc: stable@vger.kernel.org # v5.10+
Fixes: 40ca1bf580ef ("PCI: brcmstb: Add MSI support")
Signed-off-by: Stanimir Varbanov 
Reviewed-by: Florian Fainelli 
Reviewed-by: Manivannan Sadhasivam 
Link: https://lore.kernel.org/r/20250122222955.1752778-1-svarbanov@suse.de
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński 
Signed-off-by: Greg Kroah-Hartman

PCI: xilinx-cpm: Fix IRQ domain leak in error path of probe

2025-04-10T12:30:56+00:00

[ Upstream commit 57b0302240741e73fe51f88404b3866e0d2933ad ]

The IRQ domain allocated for the PCIe controller is not freed if
resource_list_first_type() returns NULL, leading to a resource leak.

This fix ensures properly cleaning up the allocated IRQ domain in
the error path.

Fixes: 49e427e6bdd1 ("Merge branch 'pci/host-probe-refactor'")
Signed-off-by: Thippeswamy Havalige 
[kwilczynski: added missing Fixes: tag, refactored to use one of the goto labels]
Signed-off-by: Krzysztof Wilczyński 
Link: https://lore.kernel.org/r/20250224155025.782179-2-thippeswamy.havalige@amd.com
Signed-off-by: Sasha Levin

PCI: brcmstb: Use internal register to change link capability

2025-04-10T12:30:56+00:00

[ Upstream commit 0c97321e11e0e9e18546f828492758f6aaecec59 ]

The driver has been mistakenly writing to a read-only (RO)
configuration space register (PCI_EXP_LNKCAP) to change the
PCIe link capability.

Although harmless in this case, the proper write destination
is an internal register that is reflected by PCI_EXP_LNKCAP.

Thus, fix the brcm_pcie_set_gen() function to correctly update
the link capability.

Fixes: c0452137034b ("PCI: brcmstb: Add Broadcom STB PCIe host controller driver")
Signed-off-by: Jim Quinlan 
Reviewed-by: Florian Fainelli 
Reviewed-by: Manivannan Sadhasivam 
Link: https://lore.kernel.org/r/20250214173944.47506-3-james.quinlan@broadcom.com
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński 
Signed-off-by: Sasha Levin

PCI: cadence-ep: Fix the driver to send MSG TLP for INTx without data payload

2025-04-10T12:30:56+00:00

[ Upstream commit 3ac47fbf4f6e8c3a7c3855fac68cc3246f90f850 ]

Per the Cadence's "PCIe Controller IP for AX14" user guide, Version
1.04, Section 9.1.7.1, "AXI Subordinate to PCIe Address Translation
Registers", Table 9.4, the bit 16 of the AXI Subordinate Address
(axi_s_awaddr) when set corresponds to MSG with data, and when not set,
to MSG without data.

However, the driver is currently doing the opposite and due to this,
the INTx is never received on the host.

So, fix the driver to reflect the documentation and also make INTx work.

Fixes: 37dddf14f1ae ("PCI: cadence: Add EndPoint Controller driver for Cadence PCIe controller")
Signed-off-by: Hans Zhang <18255117159@163.com>
Signed-off-by: Hans Zhang 
Reviewed-by: Manivannan Sadhasivam 
Link: https://lore.kernel.org/r/20250214165724.184599-1-18255117159@163.com
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński 
Signed-off-by: Sasha Levin

PCI: rcar-ep: Fix incorrect variable used when calling devm_request_mem_region()

2025-03-13T11:46:55+00:00

[ Upstream commit 2d2da5a4c1b4509f6f7e5a8db015cd420144beb4 ]

The rcar_pcie_parse_outbound_ranges() uses the devm_request_mem_region()
macro to request a needed resource. A string variable that lives on the
stack is then used to store a dynamically computed resource name, which
is then passed on as one of the macro arguments. This can lead to
undefined behavior.

Depending on the current contents of the memory, the manifestations of
errors may vary. One possible output may be as follows:

  $ cat /proc/iomem
  30000000-37ffffff :
  38000000-3fffffff :

Sometimes, garbage may appear after the colon.

In very rare cases, if no NULL-terminator is found in memory, the system
might crash because the string iterator will overrun which can lead to
access of unmapped memory above the stack.

Thus, fix this by replacing outbound_name with the name of the previously
requested resource. With the changes applied, the output will be as
follows:

  $ cat /proc/iomem
  30000000-37ffffff : memory2
  38000000-3fffffff : memory3

Fixes: 2a6d0d63d999 ("PCI: rcar: Add endpoint mode support")
Link: https://lore.kernel.org/r/tencent_DBDCC19D60F361119E76919ADAB25EC13C06@qq.com
Tested-by: Lad Prabhakar 
Signed-off-by: King Dix 
[kwilczynski: commit log]
Signed-off-by: Krzysztof Wilczyński 
Reviewed-by: Lad Prabhakar 
Reviewed-by: Manivannan Sadhasivam 
Signed-off-by: Sasha Levin

Revert "PCI: Use preserve_config in place of pci_flags"

2025-02-01T17:22:30+00:00

This reverts commit 0dde3ae52a0dcc5cdfe2185ec58ec52b43fda22e which is
commit 7246a4520b4bf1494d7d030166a11b5226f6d508 upstream.

This patch causes a regression in cuttlefish/crossvm boot on arm64.

The patch was part of a series that when applied will not cause a regression
but this patch was backported to the 5.10 branch by itself.

The other patches do not apply cleanly to the 5.10 branch.

Signed-off-by: Terry Tritton 
Signed-off-by: Greg Kroah-Hartman