platform/kernel/linux-rpi.git
5 years agoMerge branch 'remotes/lorenzo/pci/layerscape'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:23 +0000 (16:10 -0500)]
Merge branch 'remotes/lorenzo/pci/layerscape'

  - Mark Layerscape endpoint BARs 2 and 4 as 64-bit (Xiaowei Bao)

  - Add CONFIG_PCI_LAYERSCAPE_EP so EP/RC can be built separately (Xiaowei
    Bao)

* remotes/lorenzo/pci/layerscape:
  PCI: layerscape: Add CONFIG_PCI_LAYERSCAPE_EP to build EP/RC separately
  PCI: layerscape: Add the bar_fixed_64bit property to the endpoint driver

5 years agoMerge branch 'remotes/lorenzo/pci/imx'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:21 +0000 (16:10 -0500)]
Merge branch 'remotes/lorenzo/pci/imx'

  - Reduce i.MX 6Quad DBI register length to avoid aborts from accessing
    invalid registers (Stefan Agner)

* remotes/lorenzo/pci/imx:
  PCI: imx6: Limit DBI register length

5 years agoMerge branch 'remotes/lorenzo/pci/hv'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:19 +0000 (16:10 -0500)]
Merge branch 'remotes/lorenzo/pci/hv'

  - Fix Hyper-V use-after-free in pci_dev removal (Dexuan Cui)

  - Fix Hyper-V build error in non-sysfs config (Randy Dunlap)

  - Reallocate to avoid Hyper-V domain number collisions (Haiyang Zhang)

  - Use Hyper-V instance ID bytes 4-5 to reduce domain collisions (Haiyang
    Zhang)

* remotes/lorenzo/pci/hv:
  PCI: hv: Use bytes 4 and 5 from instance ID as the PCI domain numbers
  PCI: hv: Detect and fix Hyper-V PCI domain number collision
  PCI: pci-hyperv: Fix build errors on non-SYSFS config
  PCI: hv: Avoid use of hv_pci_dev->pci_slot after freeing it

5 years agoMerge branch 'remotes/lorenzo/pci/dwc'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:18 +0000 (16:10 -0500)]
Merge branch 'remotes/lorenzo/pci/dwc'

  - Make kirin_dw_pcie_ops constant (Nishka Dasgupta)

  - Make DesignWare "num-lanes" property optional and remove from relevant
    DTs (Hou Zhiqiang)

* remotes/lorenzo/pci/dwc:
  arm64: dts: fsl: Remove num-lanes property from PCIe nodes
  ARM: dts: ls1021a: Remove num-lanes property from PCIe nodes
  PCI: dwc: Return directly when num-lanes is not found
  dt-bindings: PCI: designware: Remove the num-lanes from Required properties
  PCI: kirin: Make structure kirin_dw_pcie_ops constant

5 years agoMerge branch 'remotes/lorenzo/pci/al'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:17 +0000 (16:10 -0500)]
Merge branch 'remotes/lorenzo/pci/al'

  - Add driver for Amazon Annapurna Labs PCIe controller (Jonathan Chocron)

  - Disable MSI-X since Annapurna Labs advertises it, but it's broken
    (Jonathan Chocron)

  - Disable VPD since Annapurna Labs advertises it, but it's broken
    (Jonathan Chocron)

  - Add ACS quirk since Annapurna Labs doesn't support ACS but does provide
    some equivalent protections (Ali Saidi)

* remotes/lorenzo/pci/al:
  PCI: dwc: Add validation that PCIe core is set to correct mode
  PCI: dwc: al: Add Amazon Annapurna Labs PCIe controller driver
  dt-bindings: PCI: Add Amazon's Annapurna Labs PCIe host bridge binding
  PCI: Add quirk to disable MSI-X support for Amazon's Annapurna Labs Root Port
  PCI/VPD: Prevent VPD access for Amazon's Annapurna Labs Root Port
  PCI: Add ACS quirk for Amazon Annapurna Labs root ports
  PCI: Add Amazon's Annapurna Labs vendor ID

# Conflicts:
# drivers/pci/quirks.c

5 years agoMerge branch 'pci/resource'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:15 +0000 (16:10 -0500)]
Merge branch 'pci/resource'

  - Convert pci_resource_to_user() to a weak function to remove
    HAVE_ARCH_PCI_RESOURCE_TO_USER #defines (Denis Efremov)

  - Use PCI_SRIOV_NUM_BARS for idiomatic loop structure (Denis Efremov)

  - Fix Resizable BAR size suspend/restore for 1MB BARs (Sumit Saxena)

  - Correct "pci=resource_alignment" example in documentation (Alexey
    Kardashevskiy)

* pci/resource:
  PCI: Correct pci=resource_alignment parameter example
  PCI: Restore Resizable BAR size bits correctly for 1MB BARs
  PCI: Use PCI_SRIOV_NUM_BARS in loops instead of PCI_IOV_RESOURCE_END
  PCI: Convert pci_resource_to_user() to a weak function

# Conflicts:
# drivers/pci/pci.c

5 years agoMerge branch 'pci/pciehp'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:14 +0000 (16:10 -0500)]
Merge branch 'pci/pciehp'

  - Cleanup pciehp LED/indicator control with a new consolidated
    pciehp_set_indicators() interface that controls both Attention and
    Power Indicators (Denis Efremov)

* pci/pciehp:
  PCI: pciehp: Refer to "Indicators" instead of "LEDs" in comments
  PCI: pciehp: Remove pciehp_green_led_{on,off,blink}()
  PCI: pciehp: Remove pciehp_set_attention_status()
  PCI: pciehp: Combine adjacent indicator updates
  PCI: pciehp: Add pciehp_set_indicators() to set both indicators

5 years agoMerge branch 'pci/p2pdma'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:12 +0000 (16:10 -0500)]
Merge branch 'pci/p2pdma'

  - Move P2PCMA PCI bus offset from generic dev_pagemap to
    pci_p2pdma_pagemap (Logan Gunthorpe)

  - Add provider's pci_dev to pci_p2pdma_pagemap (Logan Gunthorpe)

  - Apply host bridge whitelist for ACS (Logan Gunthorpe)

  - Whitelist some Intel host bridges for P2PDMA (Logan Gunthorpe)

  - Add attrs to pci_p2pdma_map_sg() to match dma_map_sg() (Logan
    Gunthorpe)

  - Add pci_p2pdma_unmap_sg() (Logan Gunthorpe)

  - Store P2PDMA mapping method in xarray (Logan Gunthorpe)

  - Map requests that traverse a host bridge (Logan Gunthorpe)

  - Allow IOMMU for host bridge whitelist (Logan Gunthorpe)

* pci/p2pdma:
  PCI/P2PDMA: Update pci_p2pdma_distance_many() documentation
  PCI/P2PDMA: Allow IOMMU for host bridge whitelist
  PCI/P2PDMA: dma_map() requests that traverse the host bridge
  PCI/P2PDMA: Store mapping method in an xarray
  PCI/P2PDMA: Factor out __pci_p2pdma_map_sg()
  PCI/P2PDMA: Introduce pci_p2pdma_unmap_sg()
  PCI/P2PDMA: Add attrs argument to pci_p2pdma_map_sg()
  PCI/P2PDMA: Whitelist some Intel host bridges
  PCI/P2PDMA: Factor out host_bridge_whitelist()
  PCI/P2PDMA: Apply host bridge whitelist for ACS
  PCI/P2PDMA: Factor out __upstream_bridge_distance()
  PCI/P2PDMA: Add constants for map type results to upstream_bridge_distance()
  PCI/P2PDMA: Add provider's pci_dev to pci_p2pdma_pagemap struct
  PCI/P2PDMA: Introduce private pagemap structure

5 years agoMerge branch 'pci/msi'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:12 +0000 (16:10 -0500)]
Merge branch 'pci/msi'

  - Enable PCI_MSI_IRQ_DOMAIN support for RISC-V (Wesley Terpstra)

* pci/msi:
  PCI/MSI: Enable PCI_MSI_IRQ_DOMAIN support for RISC-V

5 years agoMerge branch 'pci/misc'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:10 +0000 (16:10 -0500)]
Merge branch 'pci/misc'

  - Use devm_add_action_or_reset() helper (Fuqian Huang)

  - Mark expected switch fall-through (Gustavo A. R. Silva)

  - Convert sysfs device attributes from __ATTR() to DEVICE_ATTR() (Kelsey
    Skunberg)

  - Convert sysfs file permissions from S_IRUSR etc to octal (Kelsey
    Skunberg)

  - Move SR-IOV sysfs functions to iov.c (Kelsey Skunberg)

  - Add pci_info_ratelimited() to ratelimit PCI messages separately
    (Krzysztof Wilczynski)

  - Fix "'static' not at beginning of declaration" warnings (Krzysztof
    Wilczynski)

  - Clean up resource_alignment parameter to not require static buffer
    (Logan Gunthorpe)

  - Add ACS quirk for iProc PAXB (Abhinav Ratna)

  - Add pci_irq_vector() and other stubs for !CONFIG_PCI (Herbert Xu)

* pci/misc:
  PCI: Add pci_irq_vector() and other stubs when !CONFIG_PCI
  PCI: Add ACS quirk for iProc PAXB
  PCI: Force trailing new line to resource_alignment_param in sysfs
  PCI: Move pci_[get|set]_resource_alignment_param() into their callers
  PCI: Clean up resource_alignment parameter to not require static buffer
  PCI: Use static const struct, not const static struct
  PCI: Add pci_info_ratelimited() to ratelimit PCI separately
  PCI/IOV: Remove group write permission from sriov_numvfs, sriov_drivers_autoprobe
  PCI/IOV: Move sysfs SR-IOV functions to iov.c
  PCI: sysfs: Change permissions from symbolic to octal
  PCI: sysfs: Change DEVICE_ATTR() to DEVICE_ATTR_WO()
  PCI: sysfs: Define device attributes with DEVICE_ATTR*()
  PCI: Mark expected switch fall-through
  PCI: Use devm_add_action_or_reset()

5 years agoMerge branch 'pci/enumeration'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:08 +0000 (16:10 -0500)]
Merge branch 'pci/enumeration'

  - Consolidate _HPP & _HPX code in pci-acpi.h and remove unnecessary
    struct hotplug_program_ops (Krzysztof Wilczynski)

  - Fixup PCIe device types to remove the need for dev->has_secondary_link
    (Mika Westerberg)

* pci/enumeration:
  PCI: Get rid of dev->has_secondary_link flag
  PCI: Make pcie_downstream_port() available outside of access.c
  PCI/ACPI: Remove unnecessary struct hotplug_program_ops
  PCI/ACPI: Move _HPP & _HPX functions to pci-acpi.c
  PCI/ACPI: Rename _HPX structs from hpp_* to hpx_*

5 years agoMerge branch 'pci/encapsulate'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:07 +0000 (16:10 -0500)]
Merge branch 'pci/encapsulate'

  - Move many symbols from public linux/pci.h to subsystem-private
    drivers/pci/pci.h (Kelsey Skunberg)

  - Unexport pci_bus_get() and pci_bus_sem since they're not needed by
    modules (Kelsey Skunberg)

  - Remove unused pci_block_cfg_access() et al (Kelsey Skunberg)

* pci/encapsulate:
  PCI: Make pci_set_of_node(), etc private
  PCI: Make pci_enable_ptm() private
  PCI: Make pcie_set_ecrc_checking(), pcie_ecrc_get_policy() private
  PCI: Make pci_ats_init() private
  PCI: Make pcie_update_link_speed() private
  PCI: Make pci_bus_get(), pci_bus_put() private
  PCI: Make pci_hotplug_io_size, mem_size, and bus_size private
  PCI: Make pci_save_vc_state(), pci_restore_vc_state(), etc private
  PCI: Make pci_get_host_bridge_device(), pci_put_host_bridge_device() private
  PCI: Make pci_check_pme_status(), pci_pme_wakeup_bus() private
  PCI: Make PCI_PM_* delay times private
  PCI: Unexport pci_bus_sem
  PCI: Unexport pci_bus_get() and pci_bus_put()
  PCI: Remove pci_block_cfg_access() et al (unused)

5 years agoMerge branch 'pci/aspm'
Bjorn Helgaas [Mon, 23 Sep 2019 21:10:06 +0000 (16:10 -0500)]
Merge branch 'pci/aspm'

  - Consolidate ASPM interfaces in linux/pci.h (Krzysztof Wilczynski)

* pci/aspm:
  PCI: Move ASPM declarations to linux/pci.h

5 years agoPCI: Add pci_irq_vector() and other stubs when !CONFIG_PCI
Herbert Xu [Wed, 4 Sep 2019 12:26:00 +0000 (22:26 +1000)]
PCI: Add pci_irq_vector() and other stubs when !CONFIG_PCI

Add stub functions pci_alloc_irq_vectors_affinity() and pci_irq_vector()
when CONFIG_PCI is off so drivers can use them without resorting to ifdefs.

Also move the PCI_IRQ_* macros outside of the ifdefs so they are always
available.

Fixes: 625f269a5a7a ("crypto: inside-secure - add support for PCI based FPGA development board")
Link: https://lore.kernel.org/r/20190904122600.GA28660@gondor.apana.org.au
Reported-by: kbuild test robot <lkp@intel.com>
Reported-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: dwc: Add validation that PCIe core is set to correct mode
Jonathan Chocron [Thu, 12 Sep 2019 13:02:38 +0000 (16:02 +0300)]
PCI: dwc: Add validation that PCIe core is set to correct mode

Some PCIe controllers can be set to either Host or EP according to some
early boot FW. To make sure there is no discrepancy (e.g. FW configured
the port to EP mode while the DT specifies it as a host bridge or vice
versa), a check has been added for each mode.

Signed-off-by: Jonathan Chocron <jonnyc@amazon.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
Acked-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
5 years agoPCI: dwc: al: Add Amazon Annapurna Labs PCIe controller driver
Jonathan Chocron [Thu, 12 Sep 2019 13:02:37 +0000 (16:02 +0300)]
PCI: dwc: al: Add Amazon Annapurna Labs PCIe controller driver

This driver is DT based and utilizes the DesignWare APIs.

It allows using a smaller ECAM range for a larger bus range -
usually an entire bus uses 1MB of address space, but the driver
can use it for a larger number of buses. This is achieved by using a HW
mechanism which allows changing the BUS part of the "final" outgoing
config transaction. There are 2 HW regs, one which is basically a
bitmask determining which bits to take from the AXI transaction itself
and another which holds the complementary part programmed by the
driver.

All link initializations are handled by the boot FW.

Signed-off-by: Jonathan Chocron <jonnyc@amazon.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
5 years agodt-bindings: PCI: Add Amazon's Annapurna Labs PCIe host bridge binding
Jonathan Chocron [Thu, 12 Sep 2019 13:02:36 +0000 (16:02 +0300)]
dt-bindings: PCI: Add Amazon's Annapurna Labs PCIe host bridge binding

Document Amazon's Annapurna Labs PCIe host bridge.

Signed-off-by: Jonathan Chocron <jonnyc@amazon.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
5 years agoPCI: Add quirk to disable MSI-X support for Amazon's Annapurna Labs Root Port
Jonathan Chocron [Thu, 12 Sep 2019 13:00:42 +0000 (16:00 +0300)]
PCI: Add quirk to disable MSI-X support for Amazon's Annapurna Labs Root Port

The Root Port (identified by [1c36:0031]) doesn't support MSI-X. On some
platforms it is configured to not advertise the capability at all, while
on others it (mistakenly) does. This causes a panic during
initialization by the pcieport driver, since it tries to configure the
MSI-X capability. Specifically, when trying to access the MSI-X table
a "non-existing addr" exception occurs.

Example stacktrace snippet:

  SError Interrupt on CPU2, code 0xbf000000 -- SError
  CPU: 2 PID: 1 Comm: swapper/0 Not tainted 5.2.0-rc1-Jonny-14847-ge76f1d4a1828-dirty #33
  Hardware name: Annapurna Labs Alpine V3 EVP (DT)
  pstate: 80000005 (Nzcv daif -PAN -UAO)
  pc : __pci_enable_msix_range+0x4e4/0x608
  lr : __pci_enable_msix_range+0x498/0x608
  sp : ffffff80117db700
  x29: ffffff80117db700 x28: 0000000000000001
  x27: 0000000000000001 x26: 0000000000000000
  x25: ffffffd3e9d8c0b0 x24: 0000000000000000
  x23: 0000000000000000 x22: 0000000000000000
  x21: 0000000000000001 x20: 0000000000000000
  x19: ffffffd3e9d8c000 x18: ffffffffffffffff
  x17: 0000000000000000 x16: 0000000000000000
  x15: ffffff80116496c8 x14: ffffffd3e9844503
  x13: ffffffd3e9844502 x12: 0000000000000038
  x11: ffffffffffffff00 x10: 0000000000000040
  x9 : ffffff801165e270 x8 : ffffff801165e268
  x7 : 0000000000000002 x6 : 00000000000000b2
  x5 : ffffffd3e9d8c2c0 x4 : 0000000000000000
  x3 : 0000000000000000 x2 : 0000000000000000
  x1 : 0000000000000000 x0 : ffffffd3e9844680
  Kernel panic - not syncing: Asynchronous SError Interrupt
  CPU: 2 PID: 1 Comm: swapper/0 Not tainted 5.2.0-rc1-Jonny-14847-ge76f1d4a1828-dirty #33
  Hardware name: Annapurna Labs Alpine V3 EVP (DT)
  Call trace:
   dump_backtrace+0x0/0x140
   show_stack+0x14/0x20
   dump_stack+0xa8/0xcc
   panic+0x140/0x334
   nmi_panic+0x6c/0x70
   arm64_serror_panic+0x74/0x88
   __pte_error+0x0/0x28
   el1_error+0x84/0xf8
   __pci_enable_msix_range+0x4e4/0x608
   pci_alloc_irq_vectors_affinity+0xdc/0x150
   pcie_port_device_register+0x2b8/0x4e0
   pcie_portdrv_probe+0x34/0xf0

Notice that this quirk also disables MSI (which may work, but hasn't
been tested nor has a current use case), since currently there is no
standard way to disable only MSI-X.

Signed-off-by: Jonathan Chocron <jonnyc@amazon.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
5 years agoPCI/VPD: Prevent VPD access for Amazon's Annapurna Labs Root Port
Jonathan Chocron [Thu, 12 Sep 2019 13:00:41 +0000 (16:00 +0300)]
PCI/VPD: Prevent VPD access for Amazon's Annapurna Labs Root Port

The Amazon Annapurna Labs PCIe Root Port exposes the VPD capability,
but there is no actual support for it.

Trying to access the VPD (for example, as part of lspci -vv or when
reading the vpd sysfs file), results in the following warning print:

  pcieport 0001:00:00.0: VPD access failed.  This is likely a firmware bug on this device.  Contact the card vendor for a firmware update

Signed-off-by: Jonathan Chocron <jonnyc@amazon.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Add ACS quirk for Amazon Annapurna Labs root ports
Ali Saidi [Thu, 12 Sep 2019 13:00:40 +0000 (16:00 +0300)]
PCI: Add ACS quirk for Amazon Annapurna Labs root ports

The Amazon's Annapurna Labs root ports don't advertise an ACS
capability, but they don't allow peer-to-peer transactions and do
validate bus numbers through the SMMU. Additionally, it's not possible
for one RP to pass traffic to another RP.

Signed-off-by: Ali Saidi <alisaidi@amazon.com>
Signed-off-by: Jonathan Chocron <jonnyc@amazon.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Gustavo Pimentel <gustavo.pimentel@synopsys.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Add Amazon's Annapurna Labs vendor ID
Jonathan Chocron [Thu, 12 Sep 2019 13:00:39 +0000 (16:00 +0300)]
PCI: Add Amazon's Annapurna Labs vendor ID

Add Amazon's Annapurna Labs vendor ID to pci_ids.h.

Signed-off-by: Jonathan Chocron <jonnyc@amazon.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: hv: Use bytes 4 and 5 from instance ID as the PCI domain numbers
Haiyang Zhang [Thu, 15 Aug 2019 17:01:45 +0000 (17:01 +0000)]
PCI: hv: Use bytes 4 and 5 from instance ID as the PCI domain numbers

As recommended by Azure host team, the bytes 4, 5 have more uniqueness
(info entropy) than bytes 8, 9 so use them as the PCI domain numbers.

On older hosts, bytes 4, 5 can also be used -- no backward compatibility
issues are introduced and the chance of collision is greatly reduced.

In the rare cases of collision, the driver code detects and finds
another number that is not in use.

Suggested-by: Michael Kelley <mikelley@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Sasha Levin <sashal@kernel.org>
5 years agoPCI: Get rid of dev->has_secondary_link flag
Mika Westerberg [Thu, 22 Aug 2019 08:55:53 +0000 (11:55 +0300)]
PCI: Get rid of dev->has_secondary_link flag

In some systems, the Device/Port Type in the PCI Express Capabilities
register incorrectly identifies upstream ports as downstream ports.

d0751b98dfa3 ("PCI: Add dev->has_secondary_link to track downstream PCIe
links") addressed this by adding pci_dev.has_secondary_link, which is set
for downstream ports.  But this is confusing because pci_pcie_type()
sometimes gives the wrong answer, and it's not obvious that we should use
pci_dev.has_secondary_link instead.

Reduce the confusion by correcting the type of the port itself so that
pci_pcie_type() returns the actual type regardless of what the Device/Port
Type register claims it is.  Update the users to call pci_pcie_type() and
pcie_downstream_port() accordingly, and remove pci_dev.has_secondary_link
completely.

Link: https://lore.kernel.org/linux-pci/20190703133953.GK128603@google.com/
Suggested-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20190822085553.62697-2-mika.westerberg@linux.intel.com
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
5 years agoPCI: Make pcie_downstream_port() available outside of access.c
Mika Westerberg [Thu, 22 Aug 2019 08:55:52 +0000 (11:55 +0300)]
PCI: Make pcie_downstream_port() available outside of access.c

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 <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
5 years agoPCI: Add ACS quirk for iProc PAXB
Abhinav Ratna [Tue, 20 Aug 2019 04:39:45 +0000 (10:09 +0530)]
PCI: Add ACS quirk for iProc PAXB

iProc PAXB Root Ports don't advertise an ACS capability, but they do not
allow peer-to-peer transactions between Root Ports.  Add an ACS quirk so
each Root Port can be in a separate IOMMU group.

[bhelgaas: commit log, comment, use common implementation style]
Link: https://lore.kernel.org/r/1566275985-25670-1-git-send-email-srinath.mannam@broadcom.com
Signed-off-by: Abhinav Ratna <abhinav.ratna@broadcom.com>
Signed-off-by: Srinath Mannam <srinath.mannam@broadcom.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Scott Branden <scott.branden@broadcom.com>
5 years agoPCI: Force trailing new line to resource_alignment_param in sysfs
Logan Gunthorpe [Thu, 22 Aug 2019 16:10:13 +0000 (10:10 -0600)]
PCI: Force trailing new line to resource_alignment_param in sysfs

When 'pci=resource_alignment=' is specified on the command line, there is
no trailing new line.  Then, when it's read through the corresponding sysfs
attribute, there will be no newline and a cat command will not show
correctly in a shell. If the parameter is set through sysfs a new line will
be stored and it will 'cat' correctly.

To solve this, append a new line character in the show function if one does
not already exist.

Link: https://lore.kernel.org/r/20190822161013.5481-4-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Move pci_[get|set]_resource_alignment_param() into their callers
Logan Gunthorpe [Thu, 22 Aug 2019 16:10:12 +0000 (10:10 -0600)]
PCI: Move pci_[get|set]_resource_alignment_param() into their callers

Both the functions pci_get_resource_alignment_param() and
pci_set_resource_alignment_param() are now only called in one place:
resource_alignment_show() and resource_alignment_store() respectively.

There is no value in this extra set of functions so move both into their
callers respectively.

[bhelgaas: fold in "GFP_KERNEL while atomic" fix from Christoph Hellwig
<hch@infradead.org>
https://lore.kernel.org/r/20190902075006.GB754@infradead.org]
Link: https://lore.kernel.org/r/20190822161013.5481-3-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: pciehp: Refer to "Indicators" instead of "LEDs" in comments
Bjorn Helgaas [Thu, 5 Sep 2019 20:52:24 +0000 (15:52 -0500)]
PCI: pciehp: Refer to "Indicators" instead of "LEDs" in comments

The PCIe spec doesn't mention "green LEDs" or "amber LEDs".  Replace those
terms with "Power Indicator" and "Attention Indicator" so the comments
match the spec language.  Comment changes only.

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: pciehp: Remove pciehp_green_led_{on,off,blink}()
Denis Efremov [Tue, 3 Sep 2019 11:10:21 +0000 (14:10 +0300)]
PCI: pciehp: Remove pciehp_green_led_{on,off,blink}()

Remove pciehp_green_led_{on,off,blink}() and use pciehp_set_indicators()
instead, since the code is mostly the same.

[bhelgaas: drop set_power_indicator() wrapper to reduce the number of
interfaces]
Link: https://lore.kernel.org/r/20190903111021.1559-5-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
5 years agoPCI: pciehp: Remove pciehp_set_attention_status()
Denis Efremov [Tue, 3 Sep 2019 11:10:20 +0000 (14:10 +0300)]
PCI: pciehp: Remove pciehp_set_attention_status()

Remove pciehp_set_attention_status() and use pciehp_set_indicators()
instead, since the code is mostly the same.

Link: https://lore.kernel.org/r/20190903111021.1559-4-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
5 years agoPCI: pciehp: Combine adjacent indicator updates
Denis Efremov [Tue, 3 Sep 2019 11:10:19 +0000 (14:10 +0300)]
PCI: pciehp: Combine adjacent indicator updates

Combine adjacent updates of power and attention indicators into a single
pciehp_set_indicators() call.  This sends one command to the hotplug
controller instead of two.

Link: https://lore.kernel.org/r/20190903111021.1559-3-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Lukas Wunner <lukas@wunner.de>
5 years agoPCI: pciehp: Add pciehp_set_indicators() to set both indicators
Denis Efremov [Tue, 3 Sep 2019 11:10:18 +0000 (14:10 +0300)]
PCI: pciehp: Add pciehp_set_indicators() to set both indicators

Add pciehp_set_indicators() to set power and attention indicators with a
single register write.

This is a minor optimization because we frequently set both indicators and
this can do it with a single command.  It also reduces the number of
interfaces related to the indicators and makes them more discoverable
because callers use the PCI_EXP_SLTCTL_ATTN_IND_* and
PCI_EXP_SLTCTL_PWR_IND_* definitions directly.

[bhelgaas: extend commit log, s/PCI_EXP_SLTCTL_.*_IND_NONE/INDICATOR_NOOP/
so they don't look like things defined by the spec, add function doc, mask
commands to make it obvious we only send valid commands
(pcie_do_write_cmd() does mask it, but requires more effort to verify)]
Link: https://lore.kernel.org/r/20190903111021.1559-2-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Clean up resource_alignment parameter to not require static buffer
Logan Gunthorpe [Thu, 22 Aug 2019 16:10:11 +0000 (10:10 -0600)]
PCI: Clean up resource_alignment parameter to not require static buffer

Clean up the 'resource_alignment' parameter code to use kstrdup() in the
initcall routine instead of a static buffer that wastes memory regardless
of whether the feature is used.  This allows us to drop 'COMMAND_LINE_SIZE'
bytes (typically 256-4096 depending on architecture) of static data.

This is similar to what has been done for the 'disable_acs_redir'
parameter.

Link: https://lore.kernel.org/r/20190822161013.5481-2-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Use static const struct, not const static struct
Krzysztof Wilczynski [Mon, 26 Aug 2019 15:14:36 +0000 (17:14 +0200)]
PCI: Use static const struct, not const static struct

Move the static keyword to the front of declarations of pci_regs_behavior[]
and pcie_cap_regs_behavior[], which resolves compiler warnings when
building with "W=1":

  drivers/pci/pci-bridge-emul.c:41:1: warning: ‘static’ is not at beginning of
  declaration [-Wold-style-declaration]
   const static struct pci_bridge_reg_behavior pci_regs_behavior[] = {
   ^
  drivers/pci/pci-bridge-emul.c:176:1: warning: ‘static’ is not at beginning of
  declaration [-Wold-style-declaration]
   const static struct pci_bridge_reg_behavior pcie_cap_regs_behavior[] = {
   ^

Link: https://lore.kernel.org/r/20190826151436.4672-1-kw@linux.com
Link: https://lore.kernel.org/r/20190828131733.5817-1-kw@linux.com
Signed-off-by: Krzysztof Wilczynski <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
5 years agoPCI: Add pci_info_ratelimited() to ratelimit PCI separately
Krzysztof Wilczynski [Sun, 25 Aug 2019 22:46:16 +0000 (00:46 +0200)]
PCI: Add pci_info_ratelimited() to ratelimit PCI separately

Do not use printk_ratelimit() in drivers/pci/pci.c as it shares the rate
limiting state with all other callers to the printk_ratelimit().

Add pci_info_ratelimited() (similar to pci_notice_ratelimited() added in
the commit a88a7b3eb076 ("vfio: Use dev_printk() when possible")) and use
it instead of printk_ratelimit() + pci_info().

Link: https://lore.kernel.org/r/20190825224616.8021-1-kw@linux.com
Signed-off-by: Krzysztof Wilczynski <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI/IOV: Remove group write permission from sriov_numvfs, sriov_drivers_autoprobe
Kelsey Skunberg [Thu, 5 Sep 2019 06:32:26 +0000 (00:32 -0600)]
PCI/IOV: Remove group write permission from sriov_numvfs, sriov_drivers_autoprobe

Previously the sriov_numvfs and sriov_drivers_autoprobe sysfs files had
0664 permissions, which allowed group write.  libvirt runs as root when
dealing with PCI, and it chowns files needed by qemu, so group write
permission should not be needed.

Change these permissions from 0664 to 0644, which is what DEVICE_ATTR_RW()
does by default.

[bhelgaas: commit log]
Link: https://lore.kernel.org/r/20190905063226.43269-1-skunberg.kelsey@gmail.com
Link: https://lore.kernel.org/r/850cf536-0b72-d78c-efaf-855dcb391087@redhat.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Donald Dutile <ddutile@redhat.com>
5 years agoPCI/ACPI: Remove unnecessary struct hotplug_program_ops
Krzysztof Wilczynski [Tue, 27 Aug 2019 09:49:51 +0000 (11:49 +0200)]
PCI/ACPI: Remove unnecessary struct hotplug_program_ops

Move the ACPI-specific structs hpx_type0, hpx_type1, hpx_type2 and
hpx_type3 to drivers/pci/pci-acpi.c as they are not used anywhere else.
Then remove the struct hotplug_program_ops that has been shared between
drivers/pci/probe.c and drivers/pci/pci-acpi.c from drivers/pci/pci.h as it
is no longer needed.

The struct hotplug_program_ops was added by 87fcf12e846a ("PCI/ACPI: Remove
the need for 'struct hotplug_params'") and replaced previously used struct
hotplug_params enabling the support for the _HPX Type 3 Setting Record that
was added by f873c51a155a ("PCI/ACPI: Implement _HPX Type 3 Setting
Record").

The new struct allowed for the static functions such program_hpx_type0(),
program_hpx_type1(), etc., from the drivers/pci/probe.c to be called from
the function pci_acpi_program_hp_params() in the drivers/pci/pci-acpi.c.

Previously a programming of _HPX Type 0 was as follows:

  drivers/pci/probe.c:

    program_hpx_type0()
    ...
    pci_configure_device()
      hp_ops = {
        .program_type0 = program_hpx_type0,
        ...
      }
      pci_acpi_program_hp_params(&hp_ops)

  drivers/pci/pci-acpi.c:

    pci_acpi_program_hp_params(&hp_ops)
      acpi_run_hpx(hp_ops)
        decode_type0_hpx_record()
          hp_ops->program_type0     # program_hpx_type0() called via hp_ops

After the ACPI-specific functions, structs, enums, etc., have been moved to
drivers/pci/pci-acpi.c there is no need for the hotplug_program_ops as all
of the _HPX Type 0, 1, 2 and 3 are directly accessible.

Link: https://lore.kernel.org/r/20190827094951.10613-4-kw@linux.com
Signed-off-by: Krzysztof Wilczynski <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI/ACPI: Move _HPP & _HPX functions to pci-acpi.c
Krzysztof Wilczynski [Tue, 27 Aug 2019 09:49:50 +0000 (11:49 +0200)]
PCI/ACPI: Move _HPP & _HPX functions to pci-acpi.c

Move program_hpx_type0(), program_hpx_type1(), etc., and enums
hpx_type3_dev_type, hpx_type3_fn_type and hpx_type3_cfg_loc to
drivers/pci/pci-acpi.c as these functions and enums are ACPI-specific.

Move structs hpx_type0, hpx_type1, hpx_type2 and hpx_type3 to
drivers/pci/pci.h as these are shared between drivers/pci/pci-acpi.c and
drivers/pci/probe.c.

Link: https://lore.kernel.org/r/20190827094951.10613-3-kw@linux.com
Signed-off-by: Krzysztof Wilczynski <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI/ACPI: Rename _HPX structs from hpp_* to hpx_*
Krzysztof Wilczynski [Tue, 27 Aug 2019 09:49:49 +0000 (11:49 +0200)]
PCI/ACPI: Rename _HPX structs from hpp_* to hpx_*

The names of the hpp_type0, hpp_type1 and hpp_type2 structs suggest that
they're related to _HPP, when in fact they're related to _HPX.

The struct hpp_type0 denotes an _HPX Type 0 setting record that supersedes
the _HPP setting record, and it has been used interchangeably for _HPP as
per the ACPI specification (see version 6.3, section 6.2.9.1) which states
that it should be applied to PCI, PCI-X and PCI Express devices, with
settings being ignored if they are not applicable.

Rename them to hpx_type0, hpx_type1 and hpx_type2 to reflect their relation
to _HPX rather than _HPP.

Link: https://lore.kernel.org/r/20190827094951.10613-2-kw@linux.com
Signed-off-by: Krzysztof Wilczynski <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Move ASPM declarations to linux/pci.h
Krzysztof Wilczynski [Tue, 27 Aug 2019 09:56:20 +0000 (11:56 +0200)]
PCI: Move ASPM declarations to linux/pci.h

Move ASPM definitions and function prototypes from include/linux/pci-aspm.h
to include/linux/pci.h so users only need to include <linux/pci.h>:

  PCIE_LINK_STATE_L0S
  PCIE_LINK_STATE_L1
  PCIE_LINK_STATE_CLKPM
  pci_disable_link_state()
  pci_disable_link_state_locked()
  pcie_no_aspm()

No functional changes intended.

Link: https://lore.kernel.org/r/20190827095620.11213-1-kw@linux.com
Signed-off-by: Krzysztof Wilczynski <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoarm64: dts: fsl: Remove num-lanes property from PCIe nodes
Hou Zhiqiang [Tue, 20 Aug 2019 07:29:01 +0000 (07:29 +0000)]
arm64: dts: fsl: Remove num-lanes property from PCIe nodes

Remove the num-lanes property to avoid the driver setting the
link width.

On FSL Layerscape SoCs, the number of lanes assigned to PCIe
controller is not fixed, it is determined by the selected SerDes
protocol in the RCW (Reset Configuration Word).

The PCIe link training is completed automatically through the selected
SerDes protocol - the link width set-up is updated by hardware after
power on reset, so the num-lanes property is not needed for Layerscape
PCIe.

The current num-lanes property was added erroneously, which actually
indicates the maximum lanes the PCIe controller can support up to,
instead of the lanes assigned to the PCIe controller. The link width set
by SerDes protocol will be overridden by the num-lanes property, hence
the subsequent re-training will fail when the assigned lanes do not
match the value in the num-lanes property.

Remove the property to fix the issue

Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
5 years agoARM: dts: ls1021a: Remove num-lanes property from PCIe nodes
Hou Zhiqiang [Tue, 20 Aug 2019 07:28:55 +0000 (07:28 +0000)]
ARM: dts: ls1021a: Remove num-lanes property from PCIe nodes

Remove the num-lanes property to avoid the driver setting the
link width.

On FSL Layerscape SoCs, the number of lanes assigned to PCIe
controller is not fixed, it is determined by the selected SerDes
protocol in the RCW (Reset Configuration Word).

The PCIe link training is completed automatically through the selected
SerDes protocol - the link width set-up is updated by hardware after
power on reset, so the num-lanes property is not needed for Layerscape
PCIe.

The current num-lanes property was added erroneously, which actually
indicates the maximum lanes the PCIe controller can support up to,
instead of the lanes assigned to the PCIe controller. The link width set
by SerDes protocol will be overridden by the num-lanes property, hence
the subsequent re-training will fail when the assigned lanes do not
match the value in the num-lanes property.

Remove the property to fix the issue.

Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
5 years agoPCI: dwc: Return directly when num-lanes is not found
Hou Zhiqiang [Tue, 20 Aug 2019 07:28:49 +0000 (07:28 +0000)]
PCI: dwc: Return directly when num-lanes is not found

The num-lanes is optional since it is not needed on some platforms
that bring up the link in firmware.

The link programming is based on the num-lanes properties (which is
optional); if it is not present code must return instead of fiddling
with the lanes value to print an error message.

Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
5 years agodt-bindings: PCI: designware: Remove the num-lanes from Required properties
Hou Zhiqiang [Tue, 20 Aug 2019 07:28:43 +0000 (07:28 +0000)]
dt-bindings: PCI: designware: Remove the num-lanes from Required properties

The num-lanes is not a mandatory property, e.g. on FSL
Layerscape SoCs, the PCIe link training is completed
automatically based on the selected SerDes protocol, it
does not need the num-lanes to set-up the link width.

Currently it is both a Required and Optional property,
let's remove it from the Required properties.

Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
5 years agoPCI: kirin: Make structure kirin_dw_pcie_ops constant
Nishka Dasgupta [Mon, 19 Aug 2019 07:39:46 +0000 (13:09 +0530)]
PCI: kirin: Make structure kirin_dw_pcie_ops constant

Static variable kirin_dw_pcie_ops, of type dw_pcie_ops, is used only
once, when it is assigned to the constant field ops of variable pci
(having type dw_pcie) so kirin_dw_pcie_ops is never modified.

Make it constant to protect it from unintended modification.

Issue found with Coccinelle.

Signed-off-by: Nishka Dasgupta <nishkadg.linux@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
5 years agoPCI: hv: Detect and fix Hyper-V PCI domain number collision
Haiyang Zhang [Thu, 15 Aug 2019 17:01:37 +0000 (17:01 +0000)]
PCI: hv: Detect and fix Hyper-V PCI domain number collision

Currently in Azure cloud, for passthrough devices, the host sets the
device instance ID's bytes 8 - 15 to a value derived from the host HWID,
which is the same on all devices in a VM. So, the device instance ID's
bytes 8 and 9 provided by the host are no longer unique. This affects
all Azure hosts since July 2018, and can cause device passthrough to VMs
to fail because the bytes 8 and 9 are used as PCI domain number.
Collision of domain numbers will cause the second device with the same
domain number fail to load.

In the cases of collision, we will detect and find another number that is
not in use.

Suggested-by: Michael Kelley <mikelley@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Sasha Levin <sashal@kernel.org>
5 years agoPCI/IOV: Move sysfs SR-IOV functions to iov.c
Kelsey Skunberg [Tue, 13 Aug 2019 20:45:13 +0000 (14:45 -0600)]
PCI/IOV: Move sysfs SR-IOV functions to iov.c

The sysfs SR-IOV functions are only needed when the kernel is built with
SR-IOV support.  Rather than put them in pci-sysfs.c under #ifdef
CONFIG_PCI_IOV, move them to iov.c, which is only compiled when
CONFIG_PCI_IOV=y.

Link: https://lore.kernel.org/r/20190813204513.4790-4-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Donald Dutile <ddutile@redhat.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
5 years agoPCI: sysfs: Change permissions from symbolic to octal
Kelsey Skunberg [Tue, 13 Aug 2019 20:45:12 +0000 (14:45 -0600)]
PCI: sysfs: Change permissions from symbolic to octal

We prefer octal permissions over symbolic permissions such as "(S_IWUSR |
S_IWGRP)".  Change all symbolic permissions to octal permissions, e.g.,

  - (S_IWUSR | S_IWGRP)
  + 0220

Link: https://lore.kernel.org/r/20190813204513.4790-3-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Donald Dutile <ddutile@redhat.com>
5 years agoPCI: sysfs: Change DEVICE_ATTR() to DEVICE_ATTR_WO()
Kelsey Skunberg [Thu, 15 Aug 2019 15:33:52 +0000 (09:33 -0600)]
PCI: sysfs: Change DEVICE_ATTR() to DEVICE_ATTR_WO()

DEVICE_ATTR() should only be used when files have unusual permissions.
Change DEVICE_ATTR() with '0220' write-only permissions to
DEVICE_ATTR_WO(), e.g.,

  - static DEVICE_ATTR(_name, (S_IWUSR | S_IWGRP), NULL, _store);
  + static DEVICE_ATTR_WO(_name);

Since _store is no longer passed, make the _name passed by DEVICE_ATTR_WO()
and the related _name##_store() name match with each other, e.g.,

  DEVICE_ATTR_WO(bus_rescan) must be able to call bus_rescan_store()

Link: https://lore.kernel.org/r/20190815153352.86143-4-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Donald Dutile <ddutile@redhat.com>
5 years agoPCI: sysfs: Define device attributes with DEVICE_ATTR*()
Kelsey Skunberg [Tue, 13 Aug 2019 20:45:11 +0000 (14:45 -0600)]
PCI: sysfs: Define device attributes with DEVICE_ATTR*()

Device attributes should be defined using DEVICE_ATTR*(_name, _mode, _show,
_store).  Convert them all from __ATTR*() to DEVICE_ATTR*(), e.g.,

  - struct device_attribute dev_attr_##_name = __ATTR(_name, _mode, _show,
                                                      _store)
  + static DEVICE_ATTR(foo, S_IWUSR | S_IRUGO, show_foo, store_foo)

Link: https://lore.kernel.org/r/20190813204513.4790-2-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Donald Dutile <ddutile@redhat.com>
5 years agoPCI/P2PDMA: Update pci_p2pdma_distance_many() documentation
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:48 +0000 (11:30 -0600)]
PCI/P2PDMA: Update pci_p2pdma_distance_many() documentation

The comment describing pci_p2pdma_distance_many() still referred to
the devices being behind the same root port. This no longer applies
so reword the documentation.

Link: https://lore.kernel.org/r/20190730163545.4915-15-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-15-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Allow IOMMU for host bridge whitelist
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:47 +0000 (11:30 -0600)]
PCI/P2PDMA: Allow IOMMU for host bridge whitelist

Now that we map the requests correctly we can remove the iommu_present()
restriction.

Link: https://lore.kernel.org/r/20190730163545.4915-14-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-14-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: dma_map() requests that traverse the host bridge
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:46 +0000 (11:30 -0600)]
PCI/P2PDMA: dma_map() requests that traverse the host bridge

Any requests that traverse the host bridge will need to be mapped into the
IOMMU, so call dma_map_sg() inside pci_p2pdma_map_sg() when appropriate.

Similarly, call dma_unmap_sg() inside pci_p2pdma_unmap_sg().

Link: https://lore.kernel.org/r/20190730163545.4915-13-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-13-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Store mapping method in an xarray
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:45 +0000 (11:30 -0600)]
PCI/P2PDMA: Store mapping method in an xarray

When upstream_bridge_distance() is called, store the method required to map
the DMA transfers in an xarray so it can be looked up efficiently on the
hot path in pci_p2pdma_map_sg().

Link: https://lore.kernel.org/r/20190730163545.4915-12-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-12-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Factor out __pci_p2pdma_map_sg()
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:44 +0000 (11:30 -0600)]
PCI/P2PDMA: Factor out __pci_p2pdma_map_sg()

Factor out the bus-only mapping into its own static function.  No
functional changes. The original pci_p2pdma_map_sg_attrs() will be used to
decide whether this is an appropriate way to map.

Link: https://lore.kernel.org/r/20190730163545.4915-11-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-11-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Introduce pci_p2pdma_unmap_sg()
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:43 +0000 (11:30 -0600)]
PCI/P2PDMA: Introduce pci_p2pdma_unmap_sg()

Add pci_p2pdma_unmap_sg() to the two places that call pci_p2pdma_map_sg().

This is a prep patch to introduce correct mappings for p2pdma transactions
that go through the root complex.

Link: https://lore.kernel.org/r/20190730163545.4915-10-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-10-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Add attrs argument to pci_p2pdma_map_sg()
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:42 +0000 (11:30 -0600)]
PCI/P2PDMA: Add attrs argument to pci_p2pdma_map_sg()

This is to match the dma_map_sg() API which this function will have to call
in an future patch.

Add a pci_p2pdma_map_sg_attrs() function and helper to call it with no
attributes just like the dma_map_sg() function.

Link: https://lore.kernel.org/r/20190730163545.4915-9-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-9-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Whitelist some Intel host bridges
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:41 +0000 (11:30 -0600)]
PCI/P2PDMA: Whitelist some Intel host bridges

Intel devices do not have good support for P2P requests that span different
host bridges as the transactions will cross the QPI/UPI bus and this does
not perform well.

Therefore, enable support for these devices only if the host bridges match.

Add Intel devices that have been tested and are known to work. There are
likely many others out there that will need to be tested and added.

Link: https://lore.kernel.org/r/20190730163545.4915-8-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-8-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Factor out host_bridge_whitelist()
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:40 +0000 (11:30 -0600)]
PCI/P2PDMA: Factor out host_bridge_whitelist()

Push both PCI devices into the whitelist checking function seeing some
hardware will require us ensuring they are on the same host bridge.

At the same time we rename root_complex_whitelist() to
host_bridge_whitelist() to match the terminology used in the code.

Link: https://lore.kernel.org/r/20190730163545.4915-7-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-7-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Apply host bridge whitelist for ACS
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:39 +0000 (11:30 -0600)]
PCI/P2PDMA: Apply host bridge whitelist for ACS

When a P2PDMA transfer is rejected due to ACS being set, we can also check
the whitelist and allow the transactions.

Do this by pushing the whitelist check into the upstream_bridge_distance()
function.

Link: https://lore.kernel.org/r/20190730163545.4915-6-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-6-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Factor out __upstream_bridge_distance()
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:38 +0000 (11:30 -0600)]
PCI/P2PDMA: Factor out __upstream_bridge_distance()

This is a prep patch to create a second level helper. There are no
functional changes.

The root complex whitelist code will be moved into this function in a
subsequent patch.

Link: https://lore.kernel.org/r/20190730163545.4915-5-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-5-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Add constants for map type results to upstream_bridge_distance()
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:37 +0000 (11:30 -0600)]
PCI/P2PDMA: Add constants for map type results to upstream_bridge_distance()

Add constant flags to indicate how two devices will be mapped or if they
are unsupported. upstream_bridge_distance() will now return the
mapping type and the distance in a passed-by-reference argument.

This helps annotate the code better, but the main reason is so we can use
the information to store the required mapping method in an xarray.

Link: https://lore.kernel.org/r/20190730163545.4915-4-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-4-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Add provider's pci_dev to pci_p2pdma_pagemap struct
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:36 +0000 (11:30 -0600)]
PCI/P2PDMA: Add provider's pci_dev to pci_p2pdma_pagemap struct

The provider will be needed to figure out how to map a device.

Link: https://lore.kernel.org/r/20190730163545.4915-3-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-3-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI/P2PDMA: Introduce private pagemap structure
Logan Gunthorpe [Mon, 12 Aug 2019 17:30:35 +0000 (11:30 -0600)]
PCI/P2PDMA: Introduce private pagemap structure

Move the PCI bus offset from the generic dev_pagemap structure to a
specific pci_p2pdma_pagemap structure.

This structure will grow in subsequent patches.

Link: https://lore.kernel.org/r/20190730163545.4915-2-logang@deltatee.com
Link: https://lore.kernel.org/r/20190812173048.9186-2-logang@deltatee.com
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
5 years agoPCI: imx6: Limit DBI register length
Stefan Agner [Fri, 26 Jul 2019 14:40:07 +0000 (16:40 +0200)]
PCI: imx6: Limit DBI register length

Define the length of the DBI registers and limit config space to its
length. This makes sure that the kernel does not access registers
beyond that point, avoiding the following abort on a i.MX 6Quad:

  # cat /sys/devices/soc0/soc/1ffc000.pcie/pci0000\:00/0000\:00\:00.0/config
  [  100.021433] Unhandled fault: imprecise external abort (0x1406) at 0xb6ea7000
  ...
  [  100.056423] PC is at dw_pcie_read+0x50/0x84
  [  100.060790] LR is at dw_pcie_rd_own_conf+0x44/0x48
  ...

Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
5 years agoPCI: pci-hyperv: Fix build errors on non-SYSFS config
Randy Dunlap [Fri, 12 Jul 2019 15:53:19 +0000 (08:53 -0700)]
PCI: pci-hyperv: Fix build errors on non-SYSFS config

Fix build errors when building almost-allmodconfig but with SYSFS
not set (not enabled). Fixes these build errors:

ERROR: "pci_destroy_slot" [drivers/pci/controller/pci-hyperv.ko] undefined!
ERROR: "pci_create_slot" [drivers/pci/controller/pci-hyperv.ko] undefined!

drivers/pci/slot.o is only built when SYSFS is enabled, so
pci-hyperv.o has an implicit dependency on SYSFS.
Make that explicit.

Also, depending on X86 && X86_64 is not needed, so just change that
to depend on X86_64.

Fixes: a15f2c08c708 ("PCI: hv: support reporting serial number as slot information")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Haiyang Zhang <haiyangz@microsoft.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Jake Oshins <jakeo@microsoft.com>
Cc: "K. Y. Srinivasan" <kys@microsoft.com>
Cc: Haiyang Zhang <haiyangz@microsoft.com>
Cc: Stephen Hemminger <sthemmin@microsoft.com>
Cc: Stephen Hemminger <stephen@networkplumber.org>
Cc: Sasha Levin <sashal@kernel.org>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: linux-pci@vger.kernel.org
Cc: linux-hyperv@vger.kernel.org
Cc: Dexuan Cui <decui@microsoft.com>
5 years agoPCI: layerscape: Add CONFIG_PCI_LAYERSCAPE_EP to build EP/RC separately
Xiaowei Bao [Wed, 14 Aug 2019 02:03:30 +0000 (10:03 +0800)]
PCI: layerscape: Add CONFIG_PCI_LAYERSCAPE_EP to build EP/RC separately

Add CONFIG_PCI_LAYERSCAPE_EP so that endpoint and host controller
drivers can be built separately.

Signed-off-by: Xiaowei Bao <xiaowei.bao@nxp.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
5 years agoPCI: layerscape: Add the bar_fixed_64bit property to the endpoint driver
Xiaowei Bao [Wed, 14 Aug 2019 02:03:29 +0000 (10:03 +0800)]
PCI: layerscape: Add the bar_fixed_64bit property to the endpoint driver

The layerscape PCIe controller have 4 BARs.

 BAR0 and BAR1 are 32bit, BAR2 and BAR4 are 64bit and that's a
fixed hardware configuration.

Set the bar_fixed_64bit variable accordingly.

Signed-off-by: Xiaowei Bao <xiaowei.bao@nxp.com>
[lorenzo.pieralisi@arm.com: commit log]
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Kishon Vijay Abraham I <kishon@ti.com>
5 years agoPCI: hv: Avoid use of hv_pci_dev->pci_slot after freeing it
Dexuan Cui [Fri, 2 Aug 2019 22:50:20 +0000 (22:50 +0000)]
PCI: hv: Avoid use of hv_pci_dev->pci_slot after freeing it

The slot must be removed before the pci_dev is removed, otherwise a panic
can happen due to use-after-free.

Fixes: 15becc2b56c6 ("PCI: hv: Add hv_pci_remove_slots() when we unload the driver")
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: stable@vger.kernel.org
5 years agoPCI/MSI: Enable PCI_MSI_IRQ_DOMAIN support for RISC-V
Wesley Terpstra [Thu, 25 Jul 2019 21:28:07 +0000 (14:28 -0700)]
PCI/MSI: Enable PCI_MSI_IRQ_DOMAIN support for RISC-V

Add RISC-V as an arch that supports PCI_MSI_IRQ_DOMAIN.  The related change
to generate asm/msi.h is 251a44888183 ("riscv: include generic support for
MSI irqdomains").

Link: https://lore.kernel.org/r/alpine.DEB.2.21.9999.1907251426450.32766@viisi.sifive.com
Signed-off-by: Wesley Terpstra <wesley@sifive.com>
[paul.walmsley@sifive.com: wrote patch description; split this
 patch from the arch/riscv patch]
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
5 years agoPCI: Mark expected switch fall-through
Gustavo A. R. Silva [Fri, 2 Aug 2019 01:22:48 +0000 (20:22 -0500)]
PCI: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

This fixes the following warning (Building: allmodconfig i386):

  drivers/pci/hotplug/ibmphp_res.c: In function ‘update_bridge_ranges’:
  drivers/pci/hotplug/ibmphp_res.c:1943:16: warning: this statement may fall through [-Wimplicit-fallthrough=]
 function = 0x8;
 ~~~~~~~~~^~~~~
  drivers/pci/hotplug/ibmphp_res.c:1944:6: note: here
case PCI_HEADER_TYPE_MULTIBRIDGE:
^~~~

Link: https://lore.kernel.org/r/20190802012248.GA22622@embeddedor
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
5 years agoPCI: Correct pci=resource_alignment parameter example
Alexey Kardashevskiy [Thu, 6 Jun 2019 03:25:57 +0000 (13:25 +1000)]
PCI: Correct pci=resource_alignment parameter example

The "pci=resource_alignment" parameter is described as requiring an order
(not a size) and the code in pci_specified_resource_alignment() expects an
order.

But the example wrongly shows a size.  Convert the example to an order.

Fixes: 8b078c603249 ("PCI: Update "pci=resource_alignment" documentation")
Link: https://lore.kernel.org/r/20190606032557.107542-1-aik@ozlabs.ru
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Restore Resizable BAR size bits correctly for 1MB BARs
Sumit Saxena [Thu, 25 Jul 2019 19:25:52 +0000 (00:55 +0530)]
PCI: Restore Resizable BAR size bits correctly for 1MB BARs

In a Resizable BAR Control Register, bits 13:8 control the size of the BAR.
The encoded values of these bits are as follows (see PCIe r5.0, sec
7.8.6.3):

  Value    BAR size
     0     1 MB (2^20 bytes)
     1     2 MB (2^21 bytes)
     2     4 MB (2^22 bytes)
   ...
    43     8 EB (2^63 bytes)

Previously we incorrectly set the BAR size bits for a 1 MB BAR to 0x1f
instead of 0, so devices that support that size, e.g., new megaraid_sas and
mpt3sas adapters, fail to initialize during resume from S3 sleep.

Correctly calculate the BAR size bits for Resizable BAR control registers.

Link: https://lore.kernel.org/r/20190725192552.24295-1-sumit.saxena@broadcom.com
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=203939
Fixes: d3252ace0bc6 ("PCI: Restore resized BAR state on resume")
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Cc: stable@vger.kernel.org # v4.19+
5 years agoPCI: Use PCI_SRIOV_NUM_BARS in loops instead of PCI_IOV_RESOURCE_END
Denis Efremov [Tue, 6 Aug 2019 14:07:15 +0000 (17:07 +0300)]
PCI: Use PCI_SRIOV_NUM_BARS in loops instead of PCI_IOV_RESOURCE_END

Writing loop conditions as "i < NUM" is a common C idiom; using "i <= END"
is unusual and thus prone to errors.  Change loops to use the former.

Link: https://lore.kernel.org/r/20190806140715.19847-1-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
5 years agoPCI: Convert pci_resource_to_user() to a weak function
Denis Efremov [Mon, 29 Jul 2019 10:13:57 +0000 (13:13 +0300)]
PCI: Convert pci_resource_to_user() to a weak function

Convert pci_resource_to_user() to a weak function so the existing
architecture-specific implementations will automatically override the
generic one.  This allows us to remove HAVE_ARCH_PCI_RESOURCE_TO_USER
definitions and avoid the conditional compilation for this single function.

Link: https://lore.kernel.org/r/20190729101401.28068-1-efremov@linux.com
Link: https://lore.kernel.org/r/20190729101401.28068-2-efremov@linux.com
Link: https://lore.kernel.org/r/20190729101401.28068-3-efremov@linux.com
Link: https://lore.kernel.org/r/20190729101401.28068-4-efremov@linux.com
Link: https://lore.kernel.org/r/20190729101401.28068-5-efremov@linux.com
Link: https://lore.kernel.org/r/20190729101401.28068-6-efremov@linux.com
Signed-off-by: Denis Efremov <efremov@linux.com>
[bhelgaas: squash into one commit]
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Paul Burton <paul.burton@mips.com> # MIPS
5 years agoPCI: Make pci_set_of_node(), etc private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:48 +0000 (17:38 -0600)]
PCI: Make pci_set_of_node(), etc private

These interfaces:

  void pci_set_of_node(struct pci_dev *dev);
  void pci_release_of_node(struct pci_dev *dev);
  void pci_set_bus_of_node(struct pci_bus *bus);
  void pci_release_bus_of_node(struct pci_bus *bus);

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-12-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pci_enable_ptm() private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:47 +0000 (17:38 -0600)]
PCI: Make pci_enable_ptm() private

This interface:

  int pci_enable_ptm(struct pci_dev *dev, u8 *granularity);

is only used in drivers/pci/ and does not need to be seen by the rest of
the kernel.  Move it to drivers/pci/pci.h so it's private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-11-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pcie_set_ecrc_checking(), pcie_ecrc_get_policy() private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:46 +0000 (17:38 -0600)]
PCI: Make pcie_set_ecrc_checking(), pcie_ecrc_get_policy() private

These interfaces:

  void pcie_set_ecrc_checking(struct pci_dev *dev);
  void pcie_ecrc_get_policy(char *str);

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-10-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pci_ats_init() private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:45 +0000 (17:38 -0600)]
PCI: Make pci_ats_init() private

This interface:

  void pci_ats_init(struct pci_dev *dev);

is only used in drivers/pci/ and does not need to be seen by the rest of
the kernel.  Move it to drivers/pci/pci.h so it's private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-9-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pcie_update_link_speed() private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:44 +0000 (17:38 -0600)]
PCI: Make pcie_update_link_speed() private

This interface:

  void pcie_update_link_speed(struct pci_bus *bus, u16 link_status);

is only used in drivers/pci/ and does not need to be seen by the rest of
the kernel.  Move it to drivers/pci/pci.h so it's private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-8-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pci_bus_get(), pci_bus_put() private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:43 +0000 (17:38 -0600)]
PCI: Make pci_bus_get(), pci_bus_put() private

These interfaces:

  struct pci_bus *pci_bus_get(struct pci_bus *bus);
  void pci_bus_put(struct pci_bus *bus);

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-7-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pci_hotplug_io_size, mem_size, and bus_size private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:42 +0000 (17:38 -0600)]
PCI: Make pci_hotplug_io_size, mem_size, and bus_size private

These symbols:

  extern unsigned long pci_hotplug_io_size;
  extern unsigned long pci_hotplug_mem_size;
  extern unsigned long pci_hotplug_bus_size;

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-6-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pci_save_vc_state(), pci_restore_vc_state(), etc private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:41 +0000 (17:38 -0600)]
PCI: Make pci_save_vc_state(), pci_restore_vc_state(), etc private

These Virtual Channel interfaces:

  int pci_save_vc_state(struct pci_dev *dev);
  void pci_restore_vc_state(struct pci_dev *dev);
  void pci_allocate_vc_save_buffers(struct pci_dev *dev);

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-5-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pci_get_host_bridge_device(), pci_put_host_bridge_device() private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:40 +0000 (17:38 -0600)]
PCI: Make pci_get_host_bridge_device(), pci_put_host_bridge_device() private

These interfaces:

  struct device *pci_get_host_bridge_device(struct pci_dev *dev);
  void pci_put_host_bridge_device(struct device *dev);

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-4-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make pci_check_pme_status(), pci_pme_wakeup_bus() private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:39 +0000 (17:38 -0600)]
PCI: Make pci_check_pme_status(), pci_pme_wakeup_bus() private

These interfaces:

  bool pci_check_pme_status(struct pci_dev *dev);
  void pci_pme_wakeup_bus(struct pci_bus *bus);

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-3-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Make PCI_PM_* delay times private
Kelsey Skunberg [Wed, 24 Jul 2019 23:38:38 +0000 (17:38 -0600)]
PCI: Make PCI_PM_* delay times private

These delay time definitions:

  #define PCI_PM_D2_DELAY         200
  #define PCI_PM_D3_WAIT          10
  #define PCI_PM_D3COLD_WAIT      100
  #define PCI_PM_BUS_WAIT         50

are only used in drivers/pci/ and do not need to be seen by the rest of the
kernel.  Move them to drivers/pci/pci.h so they're private to the PCI
subsystem.

Link: https://lore.kernel.org/r/20190724233848.73327-2-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Use devm_add_action_or_reset()
Fuqian Huang [Mon, 8 Jul 2019 12:33:54 +0000 (20:33 +0800)]
PCI: Use devm_add_action_or_reset()

devm_add_action_or_reset() is a helper function which internally calls
devm_add_action().  If the devm_add_action() fails, it will execute the
action mentioned and return the error code.

Use devm_add_action_or_reset() to reduce source code size (avoid writing
the action twice) and reduce the likelihood of bugs.

Link: https://lore.kernel.org/r/20190708123354.12127-1-huangfq.daxian@gmail.com
Signed-off-by: Fuqian Huang <huangfq.daxian@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Unexport pci_bus_sem
Kelsey Skunberg [Thu, 18 Jul 2019 03:29:52 +0000 (21:29 -0600)]
PCI: Unexport pci_bus_sem

pci_bus_sem is not used by a loadable kernel module and does not need to be
exported.

It was exported by ce29ca3ea407 ("PCI: acpiphp: remove all functions in
slot, even without ACPI _EJx"), which added a use of pci_bus_sem in
acpiphp, which could be built as a module at that time.  But since
6037a803b05e ("PCI: acpiphp: Convert acpiphp to be builtin only, not
modular"), it can no longer be built as a module.

Link: https://lore.kernel.org/r/20190718032951.40188-1-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
5 years agoPCI: Unexport pci_bus_get() and pci_bus_put()
Kelsey Skunberg [Wed, 17 Jul 2019 18:23:53 +0000 (12:23 -0600)]
PCI: Unexport pci_bus_get() and pci_bus_put()

pci_bus_get() and pci_bus_put() are not used by a loadable kernel module
and do not need to be exported.

These were exported by fe830ef62ac6 ("PCI: Introduce pci_bus_{get|put}() to
manage PCI bus reference count"), but there are no loadable modules in the
tree that use them.

Link: https://lore.kernel.org/r/20190717182353.45557-1-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
5 years agoPCI: Remove pci_block_cfg_access() et al (unused)
Kelsey Skunberg [Mon, 15 Jul 2019 20:34:16 +0000 (14:34 -0600)]
PCI: Remove pci_block_cfg_access() et al (unused)

Remove the following unused functions from include/linux/pci.h:

  pci_block_cfg_access()
  pci_block_cfg_access_in_atomic()
  pci_unblock_cfg_access()

These were added by fb51ccbf217c ("PCI: Rework config space blocking
services"), though no callers were added.

Link: https://lore.kernel.org/r/20190715203416.37547-1-skunberg.kelsey@gmail.com
Signed-off-by: Kelsey Skunberg <skunberg.kelsey@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
5 years agoLinus 5.3-rc1
Linus Torvalds [Sun, 21 Jul 2019 21:05:38 +0000 (14:05 -0700)]
Linus 5.3-rc1

5 years agoMerge tag 'devicetree-fixes-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 21 Jul 2019 17:28:39 +0000 (10:28 -0700)]
Merge tag 'devicetree-fixes-for-5.3' of git://git./linux/kernel/git/robh/linux

Pull Devicetree fixes from Rob Herring:
 "Fix several warnings/errors in validation of binding schemas"

* tag 'devicetree-fixes-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux:
  dt-bindings: pinctrl: stm32: Fix missing 'clocks' property in examples
  dt-bindings: iio: ad7124: Fix dtc warnings in example
  dt-bindings: iio: avia-hx711: Fix avdd-supply typo in example
  dt-bindings: pinctrl: aspeed: Fix AST2500 example errors
  dt-bindings: pinctrl: aspeed: Fix 'compatible' schema errors
  dt-bindings: riscv: Limit cpus schema to only check RiscV 'cpu' nodes
  dt-bindings: Ensure child nodes are of type 'object'

5 years agoMerge branch 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Sun, 21 Jul 2019 17:09:43 +0000 (10:09 -0700)]
Merge branch 'work.misc' of git://git./linux/kernel/git/viro/vfs

Pull vfs documentation typo fix from Al Viro.

* 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  typo fix: it's d_make_root, not d_make_inode...

5 years agoMerge tag '5.3-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 21 Jul 2019 17:01:17 +0000 (10:01 -0700)]
Merge tag '5.3-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6

Pull cifs fixes from Steve French:
 "Two fixes for stable, one that had dependency on earlier patch in this
  merge window and can now go in, and a perf improvement in SMB3 open"

* tag '5.3-smb3-fixes' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: update internal module number
  cifs: flush before set-info if we have writeable handles
  smb3: optimize open to not send query file internal info
  cifs: copy_file_range needs to strip setuid bits and update timestamps
  CIFS: fix deadlock in cached root handling

5 years agoiommu/amd: fix a crash in iova_magazine_free_pfns
Qian Cai [Thu, 11 Jul 2019 16:17:45 +0000 (12:17 -0400)]
iommu/amd: fix a crash in iova_magazine_free_pfns

The commit b3aa14f02254 ("iommu: remove the mapping_error dma_map_ops
method") incorrectly changed the checking from dma_ops_alloc_iova() in
map_sg() causes a crash under memory pressure as dma_ops_alloc_iova()
never return DMA_MAPPING_ERROR on failure but 0, so the error handling
is all wrong.

   kernel BUG at drivers/iommu/iova.c:801!
    Workqueue: kblockd blk_mq_run_work_fn
    RIP: 0010:iova_magazine_free_pfns+0x7d/0xc0
    Call Trace:
     free_cpu_cached_iovas+0xbd/0x150
     alloc_iova_fast+0x8c/0xba
     dma_ops_alloc_iova.isra.6+0x65/0xa0
     map_sg+0x8c/0x2a0
     scsi_dma_map+0xc6/0x160
     pqi_aio_submit_io+0x1f6/0x440 [smartpqi]
     pqi_scsi_queue_command+0x90c/0xdd0 [smartpqi]
     scsi_queue_rq+0x79c/0x1200
     blk_mq_dispatch_rq_list+0x4dc/0xb70
     blk_mq_sched_dispatch_requests+0x249/0x310
     __blk_mq_run_hw_queue+0x128/0x200
     blk_mq_run_work_fn+0x27/0x30
     process_one_work+0x522/0xa10
     worker_thread+0x63/0x5b0
     kthread+0x1d2/0x1f0
     ret_from_fork+0x22/0x40

Fixes: b3aa14f02254 ("iommu: remove the mapping_error dma_map_ops method")
Signed-off-by: Qian Cai <cai@lca.pw>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agohexagon: switch to generic version of pte allocation
Mike Rapoport [Tue, 30 Apr 2019 14:27:50 +0000 (17:27 +0300)]
hexagon: switch to generic version of pte allocation

The hexagon implementation pte_alloc_one(), pte_alloc_one_kernel(),
pte_free_kernel() and pte_free() is identical to the generic except of
lack of __GFP_ACCOUNT for the user PTEs allocation.

Switch hexagon to use generic version of these functions.

Signed-off-by: Mike Rapoport <rppt@linux.ibm.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
5 years agoMerge tag 'ntb-5.3' of git://github.com/jonmason/ntb
Linus Torvalds [Sun, 21 Jul 2019 16:46:59 +0000 (09:46 -0700)]
Merge tag 'ntb-5.3' of git://github.com/jonmason/ntb

Pull NTB updates from Jon Mason:
 "New feature to add support for NTB virtual MSI interrupts, the ability
  to test and use this feature in the NTB transport layer.

  Also, bug fixes for the AMD and Switchtec drivers, as well as some
  general patches"

* tag 'ntb-5.3' of git://github.com/jonmason/ntb: (22 commits)
  NTB: Describe the ntb_msi_test client in the documentation.
  NTB: Add MSI interrupt support to ntb_transport
  NTB: Add ntb_msi_test support to ntb_test
  NTB: Introduce NTB MSI Test Client
  NTB: Introduce MSI library
  NTB: Rename ntb.c to support multiple source files in the module
  NTB: Introduce functions to calculate multi-port resource index
  NTB: Introduce helper functions to calculate logical port number
  PCI/switchtec: Add module parameter to request more interrupts
  PCI/MSI: Support allocating virtual MSI interrupts
  ntb_hw_switchtec: Fix setup MW with failure bug
  ntb_hw_switchtec: Skip unnecessary re-setup of shared memory window for crosslink case
  ntb_hw_switchtec: Remove redundant steps of switchtec_ntb_reinit_peer() function
  NTB: correct ntb_dev_ops and ntb_dev comment typos
  NTB: amd: Silence shift wrapping warning in amd_ntb_db_vector_mask()
  ntb_hw_switchtec: potential shift wrapping bug in switchtec_ntb_init_sndev()
  NTB: ntb_transport: Ensure qp->tx_mw_dma_addr is initaliazed
  NTB: ntb_hw_amd: set peer limit register
  NTB: ntb_perf: Clear stale values in doorbell and command SPAD register
  NTB: ntb_perf: Disable NTB link after clearing peer XLAT registers
  ...

5 years agotypo fix: it's d_make_root, not d_make_inode...
Al Viro [Sun, 21 Jul 2019 03:17:30 +0000 (23:17 -0400)]
typo fix: it's d_make_root, not d_make_inode...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
5 years agodt-bindings: pinctrl: stm32: Fix missing 'clocks' property in examples
Rob Herring [Tue, 16 Jul 2019 21:34:40 +0000 (15:34 -0600)]
dt-bindings: pinctrl: stm32: Fix missing 'clocks' property in examples

Now that examples are validated against the DT schema, an error with
required 'clocks' property missing is exposed:

Documentation/devicetree/bindings/pinctrl/st,stm32-pinctrl.example.dt.yaml: \
pinctrl@40020000: gpio@0: 'clocks' is a required property
Documentation/devicetree/bindings/pinctrl/st,stm32-pinctrl.example.dt.yaml: \
pinctrl@50020000: gpio@1000: 'clocks' is a required property
Documentation/devicetree/bindings/pinctrl/st,stm32-pinctrl.example.dt.yaml: \
pinctrl@50020000: gpio@2000: 'clocks' is a required property

Add the missing 'clocks' properties to the examples to fix the errors.

Fixes: 2c9239c125f0 ("dt-bindings: pinctrl: Convert stm32 pinctrl bindings to json-schema")
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: linux-gpio@vger.kernel.org
Cc: linux-stm32@st-md-mailman.stormreply.com
Acked-by: Alexandre TORGUE <alexandre.torgue@st.com>
Signed-off-by: Rob Herring <robh@kernel.org>
5 years agodt-bindings: iio: ad7124: Fix dtc warnings in example
Rob Herring [Tue, 16 Jul 2019 20:21:56 +0000 (14:21 -0600)]
dt-bindings: iio: ad7124: Fix dtc warnings in example

With the conversion to DT schema, the examples are now compiled with
dtc. The ad7124 binding example has the following warning:

Documentation/devicetree/bindings/iio/adc/adi,ad7124.example.dts:19.11-21: \
Warning (reg_format): /example-0/adc@0:reg: property has invalid length (4 bytes) (#address-cells == 1, #size-cells == 1)

There's a default #size-cells and #address-cells values of 1 for
examples. For examples needing different values such as this one on a
SPI bus, they need to provide a SPI bus parent node.

Fixes: 26ae15e62d3c ("Convert AD7124 bindings documentation to YAML format.")

Cc: Jonathan Cameron <jic23@kernel.org>
Cc: linux-iio@vger.kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>