linux-dev/drivers/pci/access.c, branch linus/master

PCI: Reduce warnings on possible RW1C corruption

2022-03-04T21:59:52Z

For hardware that only supports 32-bit writes to PCI there is the possibility of clearing RW1C (write-one-to-clear) bits. A rate-limited messages was introduced by fb2659230120, but rate-limiting is not the best choice here. Some devices may not show the warnings they should if another device has just produced a bunch of warnings. Also, the number of messages can be a nuisance on devices which are otherwise working fine. Change the ratelimit to a single warning per bus. This ensures no bus is 'starved' of emitting a warning and also that there isn't a continuous stream of warnings. It would be preferable to have a warning per device, but the pci_dev structure is not available here, and a lookup from devfn would be far too slow. Suggested-by: Bjorn Helgaas Fixes: fb2659230120 ("PCI: Warn on possible RW1C corruption for sub-32 bit config writes") Link: https://lore.kernel.org/r/20200806041455.11070-1-mark.tomlinson@alliedtelesis.co.nz Signed-off-by: Mark Tomlinson Signed-off-by: Bjorn Helgaas Reviewed-by: Florian Fainelli Reviewed-by: Rob Herring Acked-by: Scott Branden

PCI: Use PCI_ERROR_RESPONSE to identify config read errors

2021-11-18T20:31:43Z

Include PCI_ERROR_RESPONSE along with 0xFFFF and 0xFFFFFFFF in the comment about identifying config read errors. This makes checks for config read errors easier to find. Comment change only. Link: https://lore.kernel.org/r/866e2db544df45af70df7e64659bf02e03998ae3.1637243717.git.naveennaidu479@gmail.com Signed-off-by: Naveen Naidu Signed-off-by: Bjorn Helgaas

PCI: Drop error data fabrication when config read fails

2021-11-18T19:38:20Z

If config pci_ops.read() methods return failure, the PCI_OP_READ() and PCI_USER_READ_CONFIG() wrappers use PCI_SET_ERROR_RESPONSE() to set the data value, so there's no need to set it in the pci_ops.read() methods themselves. Drop the unnecessary data value fabrication when pci_ops.read() fails. Link: https://lore.kernel.org/r/1b2edb060cf19b45f70645b331e6c08c9ba798c0.1637243717.git.naveennaidu479@gmail.com Signed-off-by: Naveen Naidu Signed-off-by: Bjorn Helgaas Reviewed-by: Rob Herring

PCI: Use PCI_SET_ERROR_RESPONSE() for disconnected devices

2021-11-18T19:37:34Z

A config read from a PCI device that doesn't exist or doesn't respond causes a PCI error. There's no real data to return to satisfy the CPU read, so most hardware fabricates ~0 data. Use PCI_SET_ERROR_RESPONSE() to set the error response when we think the device has already been disconnected. This helps unify PCI error response checking and make error checks consistent and easier to find. Link: https://lore.kernel.org/r/29db0a6874716db80757e4e3cdd03562f13eb0cb.1637243717.git.naveennaidu479@gmail.com Signed-off-by: Naveen Naidu Signed-off-by: Bjorn Helgaas

PCI: Set error response data when config read fails

2021-11-18T19:36:55Z

When a PCI config read fails, most PCI host bridges fabricate ~0 data to complete the CPU read. But some host bridges do not; their drivers may only return an error from the pci_ops.read() method. In PCI_OP_READ() and PCI_USER_READ_CONFIG(), use PCI_SET_ERROR_RESPONSE() to set the data value to indicate an error when pci_ops.read() fails. This means the host bridge driver no longer needs to fabricate error data when they detect errors. This makes error response fabrication consistent and helps in removal of a lot of repeated code. Suggested-by: Rob Herring Link: https://lore.kernel.org/r/4188fc5465631ce0d472d1423de3d9fb2f09b8ff.1637243717.git.naveennaidu479@gmail.com Signed-off-by: Naveen Naidu Signed-off-by: Bjorn Helgaas Reviewed-by: Rob Herring Reviewed-by: Pali Rohár

Merge branch 'pci/misc'

2020-08-05T23:24:16Z

- Convert PCIe capability PCIBIOS errors to errno (Bolarinwa Olayemi Saheed) - Align PCIe capability and PCI accessor return values (Bolarinwa Olayemi Saheed) - Replace http:// links with https:// (Alexander A. Klimov) - Replace lkml.org, spinics, gmane with lore.kernel.org (Bjorn Helgaas) - Update panic message to mention kzalloc(), not kmalloc() (Liao Pingfang) - Move PCI_VENDOR_ID_REDHAT definition to pci_ids.h (Huacai Chen) - Remove unused pci_lost_interrupt() (Heiner Kallweit) * pci/misc: PCI: Remove unused pci_lost_interrupt() PCI: Move PCI_VENDOR_ID_REDHAT definition to pci_ids.h PCI: Fix error in panic message PCI: Replace lkml.org, spinics, gmane with lore.kernel.org PCI: Replace http:// links with https:// PCI: Align PCIe capability and PCI accessor return values PCI: Convert PCIe capability PCIBIOS errors to errno

PCI: Fix pci_cfg_wait queue locking problem

2020-07-07T20:09:01Z

The pci_cfg_wait queue is used to prevent user-space config accesses to devices while they are recovering from reset. Previously we used these operations on pci_cfg_wait: __add_wait_queue(&pci_cfg_wait, ...) __remove_wait_queue(&pci_cfg_wait, ...) wake_up_all(&pci_cfg_wait) The wake_up acquires the wait queue lock, but the add and remove do not. Originally these were all protected by the pci_lock, but cdcb33f98244 ("PCI: Avoid possible deadlock on pci_lock and p->pi_lock"), moved wake_up_all() outside pci_lock, so it could race with add/remove operations, which caused occasional kernel panics, e.g., during vfio-pci hotplug/unplug testing: Unable to handle kernel read from unreadable memory at virtual address ffff802dac469000 Resolve this by using wait_event() instead of __add_wait_queue() and __remove_wait_queue(). The wait queue lock is held by both wait_event() and wake_up_all(), so it provides mutual exclusion. Fixes: cdcb33f98244 ("PCI: Avoid possible deadlock on pci_lock and p->pi_lock") Link: https://lore.kernel.org/linux-pci/79827f2f-9b43-4411-1376-b9063b67aee3@huawei.com/T/#u Based-on: https://lore.kernel.org/linux-pci/20191210031527.40136-1-zhengxiang9@huawei.com/ Based-on-patch-by: Xiang Zheng Signed-off-by: Bjorn Helgaas Tested-by: Xiang Zheng Cc: Heyi Guo Cc: Biaoxiang Ye

PCI: Align PCIe capability and PCI accessor return values

2020-06-26T18:50:41Z

The PCI config accessors (pci_read_config_word(), et al) return PCIBIOS_SUCCESSFUL (zero) or positive error values like PCIBIOS_FUNC_NOT_SUPPORTED. The PCIe capability accessors similarly return PCIBIOS errors, but in addition, they can return -EINVAL. This makes it harder than it should be to check for errors. Return PCIBIOS_BAD_REGISTER_NUMBER instead of -EINVAL in all PCIe capability accessors. Suggested-by: Bjorn Helgaas Link: https://lore.kernel.org/r/20200615073225.24061-9-refactormyself@gmail.com Signed-off-by: Bolarinwa Olayemi Saheed Signed-off-by: Bjorn Helgaas

PCI/AER: Save AER Capability for suspend/resume

2019-10-18T22:05:42Z

Previously we did not save and restore the AER configuration on suspend/resume, so the configuration may be lost after resume. Save the AER configuration during suspend and restore it during resume. [bhelgaas: commit log] Link: https://lore.kernel.org/r/92EBB4272BF81E4089A7126EC1E7B28492C3B007@IRSMSX101.ger.corp.intel.com Signed-off-by: Mayurkumar Patel Signed-off-by: Kuppuswamy Sathyanarayanan Signed-off-by: Bjorn Helgaas Reviewed-by: Andy Shevchenko

PCI: Make pcie_downstream_port() available outside of access.c

2019-09-07T12:45:25Z

pcie_downstream_port() is useful in other places where code needs to determine whether the PCIe port is downstream so make it available outside of access.c. Link: https://lore.kernel.org/r/20190822085553.62697-1-mika.westerberg@linux.intel.com Signed-off-by: Mika Westerberg Signed-off-by: Bjorn Helgaas Reviewed-by: Andy Shevchenko