PASID usage requires PASID support in both device and IOMMU. Since the
iommu drivers always enable the PASID capability for the device if it
is supported, this extends the IOMMU_GET_HW_INFO to report the PASID
capability to userspace. Also, enhances the selftest accordingly.
Link: https://patch.msgid.link/r/20250321180143.8468-5-yi.l.liu@intel.com
Cc: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Tested-by: Zhangfei Gao <zhangfei.gao@linaro.org> #aarch64 platform
Tested-by: Nicolin Chen <nicolinc@nvidia.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
eetlp_prefix_path in the struct pci_dev tells if End-End TLP Prefixes
are supported by the path or not, and the value is only calculated if
CONFIG_PCI_PASID is set.
The Max End-End TLP Prefixes field in the Device Capabilities Register 2
also tells how many (1-4) End-End TLP Prefixes are supported (PCIe r6.2 sec
7.5.3.15). The number of supported End-End Prefixes is useful for reading
correct number of DWORDs from TLP Prefix Log register in AER capability
(PCIe r6.2 sec 7.8.4.12).
Replace eetlp_prefix_path with eetlp_prefix_max and determine the number of
supported End-End Prefixes regardless of CONFIG_PCI_PASID so that an
upcoming commit generalizing TLP Prefix Log register reading does not have
to read extra DWORDs for End-End Prefixes that never will be there.
The value stored into eetlp_prefix_max is directly derived from device's
Max End-End TLP Prefixes and does not consider limitations imposed by
bridges or the Root Port beyond supported/not supported flags. This is
intentional for two reasons:
1) PCIe r6.2 spec sections 2.2.10.4 & 6.2.4.4 indicate that a TLP is
malformed only if the number of prefixes exceed the number of Max
End-End TLP Prefixes, which seems to be the case even if the device
could never receive that many prefixes due to smaller maximum imposed
by a bridge or the Root Port. If TLP parsing is later added, this
distinction is significant in interpreting what is logged by the TLP
Prefix Log registers and the value matching to the Malformed TLP
threshold is going to be more useful.
2) TLP Prefix handling happens autonomously on a low layer and the value
in eetlp_prefix_max is not programmed anywhere by the kernel (i.e.,
there is no limiter OS can control to prevent sending more than N TLP
Prefixes).
Link: https://lore.kernel.org/r/20250114170840.1633-7-ilpo.jarvinen@linux.intel.com
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Yazen Ghannam <yazen.ghannam@amd.com>
-----BEGIN PGP SIGNATURE-----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=L3s9
-----END PGP SIGNATURE-----
Merge tag 'pci-v6.12-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci
Pull pci updates from Bjorn Helgaas:
"Enumeration:
- Wait for device readiness after reset by polling Vendor ID and
looking for Configuration RRS instead of polling the Command
register and looking for non-error completions, to avoid hardware
retries done for RRS on non-Vendor ID reads (Bjorn Helgaas)
- Rename CRS Completion Status to RRS ('Request Retry Status') to
match PCIe r6.0 spec usage (Bjorn Helgaas)
- Clear LBMS bit after a manual link retrain so we don't try to
retrain a link when there's no downstream device anymore (Maciej W.
Rozycki)
- Revert to the original link speed after retraining fails instead of
leaving it restricted to 2.5GT/s, so a future device has a chance
to use higher speeds (Maciej W. Rozycki)
- Wait for each level of downstream bus, not just the first, to
become accessible before restoring devices on that bus (Ilpo
Järvinen)
- Add ARCH_PCI_DEV_GROUPS so s390 can add its own attribute_groups
without having to stomp on the core's pdev->dev.groups (Lukas
Wunner)
Driver binding:
- Export pcim_request_region(), a managed counterpart of
pci_request_region(), for use by drivers (Philipp Stanner)
- Export pcim_iomap_region() and deprecate pcim_iomap_regions()
(Philipp Stanner)
- Request the PCI BAR used by xboxvideo (Philipp Stanner)
- Request and map drm/ast BARs with pcim_iomap_region() (Philipp
Stanner)
MSI:
- Add MSI_FLAG_NO_AFFINITY flag for devices that mux MSIs onto a
single IRQ line and cannot set the affinity of each MSI to a
specific CPU core (Marek Vasut)
- Use MSI_FLAG_NO_AFFINITY and remove unnecessary .irq_set_affinity()
implementations in aardvark, altera, brcmstb, dwc, mediatek-gen3,
mediatek, mobiveil, plda, rcar, tegra, vmd, xilinx-nwl,
xilinx-xdma, and xilinx drivers to avoid 'IRQ: set affinity failed'
warnings (Marek Vasut)
Power management:
- Add pwrctl support for ATH11K inside the WCN6855 package (Konrad
Dybcio)
PCI device hotplug:
- Remove unnecessary hpc_ops struct from shpchp (ngn)
- Check for PCI_POSSIBLE_ERROR(), not 0xffffffff, in cpqphp
(weiyufeng)
Virtualization:
- Mark Creative Labs EMU20k2 INTx masking as broken (Alex Williamson)
- Add an ACS quirk for Qualcomm SA8775P, which doesn't advertise ACS
but does provide ACS-like features (Subramanian Ananthanarayanan)
IOMMU:
- Add function 0 DMA alias quirk for Glenfly Arise audio function,
which uses the function 0 Requester ID (WangYuli)
NPEM:
- Add Native PCIe Enclosure Management (NPEM) support for sysfs
control of NVMe RAID storage indicators (ok/fail/locate/
rebuild/etc) (Mariusz Tkaczyk)
- Add support for the ACPI _DSM PCIe SSD status LED management, which
is functionally similar to NPEM but mediated by platform firmware
(Mariusz Tkaczyk)
Device trees:
- Drop minItems and maxItems from ranges in PCI generic host binding
since host bridges may have several MMIO and I/O port apertures
(Frank Li)
- Add kirin, rcar-gen2, uniphier DT binding top-level constraints for
clocks (Krzysztof Kozlowski)
Altera PCIe controller driver:
- Convert altera DT bindings from text to YAML (Matthew Gerlach)
- Replace TLP_REQ_ID() with macro PCI_DEVID(), which does the same
thing and is what other drivers use (Jinjie Ruan)
Broadcom STB PCIe controller driver:
- Add DT binding maxItems for reset controllers (Jim Quinlan)
- Use the 'bridge' reset method if described in the DT (Jim Quinlan)
- Use the 'swinit' reset method if described in the DT (Jim Quinlan)
- Add 'has_phy' so the existence of a 'rescal' reset controller
doesn't imply software control of it (Jim Quinlan)
- Add support for many inbound DMA windows (Jim Quinlan)
- Rename SoC 'type' to 'soc_base' express the fact that SoCs come in
families of multiple similar devices (Jim Quinlan)
- Add Broadcom 7712 DT description and driver support (Jim Quinlan)
- Sort enums, pcie_offsets[], pcie_cfg_data, .compatible strings for
maintainability (Bjorn Helgaas)
Freescale i.MX6 PCIe controller driver:
- Add imx6q-pcie 'dbi2' and 'atu' reg-names for i.MX8M Endpoints
(Richard Zhu)
- Fix a code restructuring error that caused i.MX8MM and i.MX8MP
Endpoints to fail to establish link (Richard Zhu)
- Fix i.MX8MP Endpoint occasional failure to trigger MSI by enforcing
outbound alignment requirement (Richard Zhu)
- Call phy_power_off() in the .probe() error path (Frank Li)
- Rename internal names from imx6_* to imx_* since i.MX7/8/9 are also
supported (Frank Li)
- Manage Refclk by using SoC-specific callbacks instead of switch
statements (Frank Li)
- Manage core reset by using SoC-specific callbacks instead of switch
statements (Frank Li)
- Expand comments for erratum ERR010728 workaround (Frank Li)
- Use generic PHY APIs to configure mode, speed, and submode, which
is harmless for devices that implement their own internal PHY
management and don't set the generic imx_pcie->phy (Frank Li)
- Add i.MX8Q (i.MX8QM, i.MX8QXP, and i.MX8DXL) DT binding and driver
Root Complex support (Richard Zhu)
Freescale Layerscape PCIe controller driver:
- Replace layerscape-pcie DT binding compatible fsl,lx2160a-pcie with
fsl,lx2160ar2-pcie (Frank Li)
- Add layerscape-pcie DT binding deprecated 'num-viewport' property
to address a DT checker warning (Frank Li)
- Change layerscape-pcie DT binding 'fsl,pcie-scfg' to phandle-array
(Frank Li)
Loongson PCIe controller driver:
- Increase max PCI hosts to 8 for Loongson-3C6000 and newer chipsets
(Huacai Chen)
Marvell Aardvark PCIe controller driver:
- Fix issue with emulating Configuration RRS for two-byte reads of
Vendor ID; previously it only worked for four-byte reads (Bjorn
Helgaas)
MediaTek PCIe Gen3 controller driver:
- Add per-SoC struct mtk_gen3_pcie_pdata to support multiple SoC
types (Lorenzo Bianconi)
- Use reset_bulk APIs to manage PHY reset lines (Lorenzo Bianconi)
- Add DT and driver support for Airoha EN7581 PCIe controller
(Lorenzo Bianconi)
Qualcomm PCIe controller driver:
- Update qcom,pcie-sc7280 DT binding with eight interrupts (Rayyan
Ansari)
- Add back DT 'vddpe-3v3-supply', which was incorrectly removed
earlier (Johan Hovold)
- Drop endpoint redundant masking of global IRQ events (Manivannan
Sadhasivam)
- Clarify unknown global IRQ message and only log it once to avoid a
flood (Manivannan Sadhasivam)
- Add 'linux,pci-domain' property to endpoint DT binding (Manivannan
Sadhasivam)
- Assign PCI domain number for endpoint controllers (Manivannan
Sadhasivam)
- Add 'qcom_pcie_ep' and the PCI domain number to IRQ names for
endpoint controller (Manivannan Sadhasivam)
- Add global SPI interrupt for PCIe link events to DT binding
(Manivannan Sadhasivam)
- Add global RC interrupt handler to handle 'Link up' events and
automatically enumerate hot-added devices (Manivannan Sadhasivam)
- Avoid mirroring of DBI and iATU register space so it doesn't
overlap BAR MMIO space (Prudhvi Yarlagadda)
- Enable controller resources like PHY only after PERST# is
deasserted to partially avoid the problem that the endpoint SoC
crashes when accessing things when Refclk is absent (Manivannan
Sadhasivam)
- Add 16.0 GT/s equalization and RX lane margining settings (Shashank
Babu Chinta Venkata)
- Pass domain number to pci_bus_release_domain_nr() explicitly to
avoid a NULL pointer dereference (Manivannan Sadhasivam)
Renesas R-Car PCIe controller driver:
- Make the read-only const array 'check_addr' static (Colin Ian King)
- Add R-Car V4M (R8A779H0) PCIe host and endpoint to DT binding
(Yoshihiro Shimoda)
TI DRA7xx PCIe controller driver:
- Request IRQF_ONESHOT for 'dra7xx-pcie-main' IRQ since the primary
handler is NULL (Siddharth Vadapalli)
- Handle IRQ request errors during root port and endpoint probe
(Siddharth Vadapalli)
TI J721E PCIe driver:
- Add DT 'ti,syscon-acspcie-proxy-ctrl' and driver support to enable
the ACSPCIE module to drive Refclk for the Endpoint (Siddharth
Vadapalli)
- Extract the cadence link setup from cdns_pcie_host_setup() so link
setup can be done separately during resume (Thomas Richard)
- Add T_PERST_CLK_US definition for the mandatory delay between
Refclk becoming stable and PERST# being deasserted (Thomas Richard)
- Add j721e suspend and resume support (Théo Lebrun)
TI Keystone PCIe controller driver:
- Fix NULL pointer checking when applying MRRS limitation quirk for
AM65x SR 1.0 Errata #i2037 (Dan Carpenter)
Xilinx NWL PCIe controller driver:
- Fix off-by-one error in INTx IRQ handler that caused INTx
interrupts to be lost or delivered as the wrong interrupt (Sean
Anderson)
- Rate-limit misc interrupt messages (Sean Anderson)
- Turn off the clock on probe failure and device removal (Sean
Anderson)
- Add DT binding and driver support for enabling/disabling PHYs (Sean
Anderson)
- Add PCIe phy bindings for the ZCU102 (Sean Anderson)
Xilinx XDMA PCIe controller driver:
- Add support for Xilinx QDMA Soft IP PCIe Root Port Bridge to DT
binding and xilinx-dma-pl driver (Thippeswamy Havalige)
Miscellaneous:
- Fix buffer overflow in kirin_pcie_parse_port() (Alexandra Diupina)
- Fix minor kerneldoc issues and typos (Bjorn Helgaas)
- Use PCI_DEVID() macro in aer_inject() instead of open-coding it
(Jinjie Ruan)
- Check pcie_find_root_port() return in x86 fixups to avoid NULL
pointer dereferences (Samasth Norway Ananda)
- Make pci_bus_type constant (Kunwu Chan)
- Remove unused declarations of __pci_pme_wakeup() and
pci_vpd_release() (Yue Haibing)
- Remove any leftover .*.cmd files with make clean (zhang jiao)
- Remove unused BILLION macro (zhang jiao)"
* tag 'pci-v6.12-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/pci/pci: (132 commits)
PCI: Fix typos
dt-bindings: PCI: qcom: Allow 'vddpe-3v3-supply' again
tools: PCI: Remove unused BILLION macro
tools: PCI: Remove .*.cmd files with make clean
PCI: Pass domain number to pci_bus_release_domain_nr() explicitly
PCI: dra7xx: Fix error handling when IRQ request fails in probe
PCI: dra7xx: Fix threaded IRQ request for "dra7xx-pcie-main" IRQ
PCI: qcom: Add RX lane margining settings for 16.0 GT/s
PCI: qcom: Add equalization settings for 16.0 GT/s
PCI: dwc: Always cache the maximum link speed value in dw_pcie::max_link_speed
PCI: dwc: Rename 'dw_pcie::link_gen' to 'dw_pcie::max_link_speed'
PCI: qcom-ep: Enable controller resources like PHY only after refclk is available
PCI: Mark Creative Labs EMU20k2 INTx masking as broken
dt-bindings: PCI: imx6q-pcie: Add reg-name "dbi2" and "atu" for i.MX8M PCIe Endpoint
dt-bindings: PCI: altera: msi: Convert to YAML
PCI: imx6: Add i.MX8Q PCIe Root Complex (RC) support
PCI: Rename CRS Completion Status to RRS
PCI: aardvark: Correct Configuration RRS checking
PCI: Wait for device readiness with Configuration RRS
PCI: brcmstb: Sort enums, pcie_offsets[], pcie_cfg_data, .compatible strings
...
PCI ATS has a global Smallest Translation Unit field that is located in
the PF but shared by all of the VFs.
The expectation is that the STU will be set to the root port's global STU
capability which is driven by the IO page table configuration of the iommu
HW. Today it becomes set when the iommu driver first enables ATS.
Thus, to enable ATS on the VF, the PF must have already had the correct
STU programmed, even if ATS is off on the PF.
Unfortunately the PF only programs the STU when the PF enables ATS. The
iommu drivers tend to leave ATS disabled when IDENTITY translation is
being used.
Thus we can get into a state where the PF is setup to use IDENTITY with
the DMA API while the VF would like to use VFIO with a PAGING domain and
have ATS turned on. This fails because the PF never loaded a PAGING domain
and so it never setup the STU, and the VF can't do it.
The simplest solution is to have the iommu driver set the ATS STU when it
probes the device. This way the ATS STU is loaded immediately at boot time
to all PFs and there is no issue when a VF comes to use it.
Add a new call pci_prepare_ats() which should be called by iommu drivers
in their probe_device() op for every PCI device if the iommu driver
supports ATS. This will setup the STU based on whatever page size
capability the iommu HW has.
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
Link: https://lore.kernel.org/r/0-v1-0fb4d2ab6770+7e706-ats_vf_jgg@nvidia.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Use FIELD_GET() to remove dependences on the field position, i.e., the
shift value. No functional change intended.
Link: https://lore.kernel.org/r/20231010204436.1000644-6-helgaas@kernel.org
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
The Requester ID/Process Address Space ID (PASID) combination
identifies an address space distinct from the PCI bus address space,
e.g., an address space defined by an IOMMU.
But the PCIe fabric routes Memory Requests based on the TLP address,
ignoring any PASID (PCIe r6.0, sec 2.2.10.4), so a TLP with PASID that
SHOULD go upstream to the IOMMU may instead be routed as a P2P
Request if its address falls in a bridge window.
To ensure that all Memory Requests with PASID are routed upstream,
only enable PASID if ACS P2P Request Redirect and Upstream Forwarding
are enabled for the path leading to the device.
Suggested-by: Jason Gunthorpe <jgg@nvidia.com>
Suggested-by: Kevin Tian <kevin.tian@intel.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Tested-by: Tony Zhu <tony.zhu@intel.com>
Link: https://lore.kernel.org/r/20221031005917.45690-5-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Some systems, e.g., HiSilicon KunPeng920 and KunPeng930, have devices that
appear as PCI but are actually on the AMBA bus. Some of these fake PCI
devices support a PASID-like feature and they do have a working PASID
capability even though they do not use the PCIe Transport Layer Protocol
and do not support TLP prefixes.
Add a pasid_no_tlp bit for this "PASID works without TLP prefixes" case and
update pci_enable_pasid() so it can enable PASID on these devices.
Set this bit for HiSilicon KunPeng920 and KunPeng930.
[bhelgaas: squashed, commit log]
Suggested-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/1626144876-11352-2-git-send-email-zhangfei.gao@linaro.org
Link: https://lore.kernel.org/r/1626144876-11352-3-git-send-email-zhangfei.gao@linaro.org
Signed-off-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Signed-off-by: Zhou Wang <wangzhou1@hisilicon.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
For SR-IOV, the PF PRI is shared between the PF and any associated VFs, and
the PRI Capability is allowed for PFs but not for VFs. Searching for the
PRI Capability on a VF always fails, even if its associated PF supports
PRI.
Add pci_pri_supported() to check whether device or its associated PF
supports PRI.
[bhelgaas: commit log, avoid "!!"]
Fixes: b16d0cb9e2 ("iommu/vt-d: Always enable PASID/PRI PCI capabilities before ATS")
Link: https://lore.kernel.org/r/1595543849-19692-1-git-send-email-ashok.raj@intel.com
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Lu Baolu <baolu.lu@linux.intel.com>
Acked-by: Joerg Roedel <jroedel@suse.de>
Cc: stable@vger.kernel.org # v4.4+
Add pci_ats_supported(), which checks whether a device has an ATS
capability, and whether it is trusted. A device is untrusted if it is
plugged into an external-facing port such as Thunderbolt and could be
spoofing an existing device to exploit weaknesses in the IOMMU
configuration. PCIe ATS is one such weaknesses since it allows
endpoints to cache IOMMU translations and emit transactions with
'Translated' Address Type (10b) that partially bypass the IOMMU
translation.
The SMMUv3 and VT-d IOMMU drivers already disallow ATS and transactions
with 'Translated' Address Type for untrusted devices. Add the check to
pci_enable_ats() to let other drivers (AMD IOMMU for now) benefit from
it.
By checking ats_cap, the pci_ats_supported() helper also returns whether
ATS was globally disabled with pci=noats, and could later include more
things, for example whether the whole PCIe hierarchy down to the
endpoint supports ATS.
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20200520152201.3309416-2-jean-philippe@linaro.org
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The Arm SMMUv3 driver uses pci_{enable,disable}_pasid() and related
functions. Export them to allow the driver to be built as a module.
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Signed-off-by: Will Deacon <will@kernel.org>
-----BEGIN PGP SIGNATURE-----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=tcy+
-----END PGP SIGNATURE-----
Merge tag 'pci-v5.6-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci
Pull PCI fixes from Bjorn Helgaas:
- Define to_pci_sysdata() always to fix build breakage when !CONFIG_PCI
(Jason A. Donenfeld)
- Use PF PASID for VFs to fix VF IOMMU bind failures (Kuppuswamy
Sathyanarayanan)
* tag 'pci-v5.6-fixes-1' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci:
PCI/ATS: Use PF PASID for VFs
x86/PCI: Define to_pci_sysdata() even when !CONFIG_PCI
Per PCIe r5.0, sec 9.3.7.14, if a PF implements the PASID Capability, the
PF PASID configuration is shared by its VFs, and VFs must not implement
their own PASID Capability. But commit 751035b8dc ("PCI/ATS: Cache PASID
Capability offset") changed pci_max_pasids() and pci_pasid_features() to
use the PASID Capability of the VF device instead of the associated PF
device. This leads to IOMMU bind failures when pci_max_pasids() and
pci_pasid_features() are called for VFs.
In pci_max_pasids() and pci_pasid_features(), always use the PF PASID
Capability.
Fixes: 751035b8dc ("PCI/ATS: Cache PASID Capability offset")
Link: https://lore.kernel.org/r/fe891f9755cb18349389609e7fed9940fc5b081a.1580325170.git.sathyanarayanan.kuppuswamy@linux.intel.com
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: stable@vger.kernel.org # v5.5+
Commit d355bb2097 ("PCI/ATS: Remove unnecessary EXPORT_SYMBOL_GPL()")
unexported a bunch of symbols from the PCI core since the only external
users were non-modular IOMMU drivers. Although most of those symbols
can remain private for now, 'pci_{enable,disable_ats()' is required for
the ARM SMMUv3 driver to build as a module, otherwise we get a build
failure as follows:
| ERROR: "pci_enable_ats" [drivers/iommu/arm-smmu-v3.ko] undefined!
| ERROR: "pci_disable_ats" [drivers/iommu/arm-smmu-v3.ko] undefined!
Re-export these two functions so that the ARM SMMUv3 driver can be build
as a module.
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
[will: rewrote commit message]
Signed-off-by: Will Deacon <will@kernel.org>
Tested-by: John Garry <john.garry@huawei.com> # smmu v3
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The following functions are only used by the PCI core or by IOMMU drivers
that cannot be modular, so there's no need to export them at all:
pci_enable_ats()
pci_disable_ats()
pci_restore_ats_state()
pci_ats_queue_depth()
pci_ats_page_aligned()
pci_enable_pri()
pci_restore_pri_state()
pci_reset_pri()
pci_prg_resp_pasid_required()
pci_enable_pasid()
pci_disable_pasid()
pci_restore_pasid_state()
pci_pasid_features()
pci_max_pasids()
Remove the unnecessary EXPORT_SYMBOL_GPL()s.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
The PRG Response PASID Required bit in the PRI Capability is read-only.
Read it once when we enumerate the device and cache the value so we don't
need to read it again.
Based-on-patch-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Previously we didn't disable the PF ATS until all associated VFs had
disabled it. But per PCIe spec r5.0, sec 9.3.7.8, the ATS Capability in
VFs and associated PFs may be enabled independently. Leaving ATS enabled
in the PF unnecessarily may have power and performance impacts.
Remove this dependency logic in the ATS enable/disable code.
[bhelgaas: commit log]
Suggested-by: Ashok Raj <ashok.raj@intel.com>
Link: https://lore.kernel.org/r/8163ab8fa66afd2cba514ae95d29ab12104781aa.1567029860.git.sathyanarayanan.kuppuswamy@linux.intel.com
Link: https://lore.kernel.org/r/20190905193146.90250-4-helgaas@kernel.org
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Keith Busch <keith.busch@intel.com>
Per PCIe r5.0, sec 9.3.7.14, if a PF implements the PASID Capability, the
PF PASID configuration is shared by its VFs. VFs must not implement their
own PASID Capability. Since VFs don't have a PASID Capability,
pci_enable_pasid() always failed, which caused IOMMU setup to fail.
Update the PASID interfaces so for VFs they reflect the state of the PF
PASID.
[bhelgaas: rebase without pasid_cap caching, commit log]
Suggested-by: Ashok Raj <ashok.raj@intel.com>
Link: https://lore.kernel.org/r/8ba1ac192e4ac737508b6ac15002158e176bab91.1567029860.git.sathyanarayanan.kuppuswamy@linux.intel.com
Link: https://lore.kernel.org/r/20190905193146.90250-3-helgaas@kernel.org
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Keith Busch <keith.busch@intel.com>
Per PCIe r5.0, sec 9.3.7.11, VFs must not implement the PRI Capability. If
the PF implements PRI, it is shared by the VFs. Since VFs don't have a PRI
Capability, pci_enable_pri() always failed, which caused IOMMU setup to
fail.
Update the PRI interfaces so for VFs they reflect the state of the PF PRI.
[bhelgaas: rebase without pri_cap caching, commit log]
Suggested-by: Ashok Raj <ashok.raj@intel.com>
Link: https://lore.kernel.org/r/b971e31f8695980da8e4a7f93e3b6a3edba3edaa.1567029860.git.sathyanarayanan.kuppuswamy@linux.intel.com
Link: https://lore.kernel.org/r/20190905193146.90250-2-helgaas@kernel.org
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Keith Busch <keith.busch@intel.com>
pci_prg_resp_pasid_required() returns the value of the "PRG Response PASID
Required" bit from the PRI capability, but the interface was previously
defined under #ifdef CONFIG_PCI_PASID.
Move it from CONFIG_PCI_PASID to CONFIG_PCI_PRI so it's with the other
PRI-related things.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Fix typos in drivers/pci. Comment and whitespace changes only.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Return the Page Aligned Request bit in the ATS Capability Register.
As per PCIe spec r4.0, sec 10.5.1.2, if the Page Aligned Request bit is
set, it indicates the Untranslated Addresses generated by the device are
always aligned to a 4096 byte boundary.
An IOMMU that can only translate page-aligned addresses can only be used
with devices that always produce aligned Untranslated Addresses. This
interface will be used by drivers for such IOMMUs to determine whether
devices can use the ATS service.
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Jacob Pan <jacob.jun.pan@linux.intel.com>
Cc: Keith Busch <keith.busch@intel.com>
Suggested-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Return the PRG Response PASID Required bit in the Page Request
Status Register.
As per PCIe spec r4.0, sec 10.5.2.3, if this bit is Set, the device
expects a PASID TLP Prefix on PRG Response Messages when the
corresponding Page Requests had a PASID TLP Prefix. If Clear, the device
does not expect PASID TLP Prefixes on any PRG Response Message, and the
device behavior is undefined if the device receives a PRG Response Message
with a PASID TLP Prefix. Also the device behavior is undefined if this
bit is Set and the device receives a PRG Response Message with no PASID TLP
Prefix when the corresponding Page Requests had a PASID TLP Prefix.
This function will be used by drivers like IOMMU, if it is required to
check the status of the PRG Response PASID Required bit before enabling
the PASID support of the device.
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Jacob Pan <jacob.jun.pan@linux.intel.com>
Cc: Keith Busch <keith.busch@intel.com>
Suggested-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
A PCIe endpoint carries the process address space identifier (PASID) in
the TLP prefix as part of the memory read/write transaction. The address
information in the TLP is relevant only for a given PASID context.
An IOMMU takes PASID value and the address information from the
TLP to look up the physical address in the system.
PASID is an End-End TLP Prefix (PCIe r4.0, sec 6.20). Sec 2.2.10.2 says
It is an error to receive a TLP with an End-End TLP Prefix by a
Receiver that does not support End-End TLP Prefixes. A TLP in
violation of this rule is handled as a Malformed TLP. This is a
reported error associated with the Receiving Port (see Section 6.2).
Prevent error condition by proactively requiring End-End TLP prefix to be
supported on the entire data path between the endpoint and the root port
before enabling PASID.
Signed-off-by: Sinan Kaya <okaya@codeaurora.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Adds a "pci=noats" boot parameter. When supplied, all ATS related
functions fail immediately and the IOMMU is configured to not use
device-IOTLB.
Any function that checks for ATS capabilities directly against the devices
should also check this flag. Currently, such functions exist only in IOMMU
drivers, and they are covered by this patch.
The motivation behind this patch is the existence of malicious devices.
Lots of research has been done about how to use the IOMMU as protection
from such devices. When ATS is supported, any I/O device can access any
physical address by faking device-IOTLB entries. Adding the ability to
ignore these entries lets sysadmins enhance system security.
Signed-off-by: Gil Kupfer <gilkup@cs.technion.ac.il>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Joerg Roedel <jroedel@suse.de>
Remove pointless comments that tell us the file name, remove blank line
comments, follow multi-line comment conventions. No functional change
intended.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
b24413180f ("License cleanup: add SPDX GPL-2.0 license identifier to
files with no license") added SPDX GPL-2.0 to several PCI files that
previously contained no license information.
Add SPDX GPL-2.0 to all other PCI files that did not contain any license
information and hence were under the default GPL version 2 license of the
kernel.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
After a Function-Level Reset, PCI states need to be restored. Save PASID
features and PRI reqs cached.
[bhelgaas: search for capability only if PRI/PASID were enabled]
Signed-off-by: CQ Tang <cq.tang@intel.com>
Signed-off-by: Ashok Raj <ashok.raj@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Jean-Phillipe Brucker <jean-philippe.brucker@arm.com>
Cc: David Woodhouse <dwmw2@infradead.org>
Device drivers need to check if an IOMMU enabled ATS, PRI and PASID in
order to know when they can use the SVM API. Cache PRI and PASID bits in
the pci_dev structure, similarly to what is currently done for ATS.
Signed-off-by: Jean-Philippe Brucker <jean-philippe.brucker@arm.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Remove pci_ats_enabled(). There are no callers outside the ATS code
itself. We don't need to check ats_cap, because if we don't find an ATS
capability, we'll never set ats_enabled.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Stop caching the Invalidate Queue Depth in struct pci_dev.
pci_ats_queue_depth() is typically called only once per device, and it
returns a fixed value per-device, so callers who need the value frequently
can cache it themselves.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
There's no need to BUG() if we enable ATS when it's already enabled. We
don't need to BUG() when disabling ATS on a device that doesn't support ATS
or if it's already disabled. If ATS is enabled, certainly we found an ATS
capability in the past, so it should still be there now.
Clean up these error paths.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Use the pci_physfn() helper rather than looking up physfn by hand.
No functional change.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
The ATS setup code in ats_alloc_one() is only used by pci_ats_init(), so
inline it there. No functional change.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
We previously returned -ENODEV for devices that don't support ATS (except
that we always returned 0 for VFs, whether or not they support ATS).
For consistency, always return -EINVAL (not -ENODEV) if the device doesn't
support ATS. Return zero for VFs that support ATS.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
The pci_ats struct is small and will get smaller, so I don't think it's
worth allocating it separately from the pci_dev struct.
Embed the ATS fields directly into struct pci_dev.
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
Previously, we allocated pci_ats structures when an IOMMU driver called
pci_enable_ats(). An SR-IOV VF shares the STU setting with its PF, so when
enabling ATS on the VF, we allocated a pci_ats struct for the PF if it
didn't already have one. We held the sriov->lock to serialize threads
concurrently enabling ATS on several VFS so only one would allocate the PF
pci_ats.
Gregor reported a deadlock here:
pci_enable_sriov
sriov_enable
virtfn_add
mutex_lock(dev->sriov->lock) # acquire sriov->lock
pci_device_add
device_add
BUS_NOTIFY_ADD_DEVICE notifier chain
iommu_bus_notifier
amd_iommu_add_device # iommu_ops.add_device
init_iommu_group
iommu_group_get_for_dev
iommu_group_add_device
__iommu_attach_device
amd_iommu_attach_device # iommu_ops.attach_device
attach_device
pci_enable_ats
mutex_lock(dev->sriov->lock) # deadlock
There's no reason to delay allocating the pci_ats struct, and if we
allocate it for each device at enumeration-time, there's no need for
locking in pci_enable_ats().
Allocate pci_ats struct during enumeration, when we initialize other
capabilities.
Note that this implementation requires ATS to be enabled on the PF first,
before on any of the VFs because the PF controls the STU for all the VFs.
Link: http://permalink.gmane.org/gmane.linux.kernel.iommu/9433
Reported-by: Gregor Dick <gdick@solarflare.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Joerg Roedel <jroedel@suse.de>
My philosophy is unused code is dead code. And dead code is subject to bit
rot and is a likely source of bugs. Use it or lose it.
This reverts parts of c320b976d7 ("PCI: Add implementation for PRI
capability"), removing these interfaces:
pci_pri_enabled()
pci_pri_stopped()
pci_pri_status()
[bhelgaas: split to separate patch]
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: Joerg Roedel <joro@8bytes.org>
Fix whitespace, capitalization, and spelling errors. No functional change.
I know "busses" is not an error, but "buses" was more common, so I used it
consistently.
Signed-off-by: Marta Rybczynska <rybczynska@gmail.com> (pci_reset_bridge_secondary_bus())
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
* 'linux-next' of git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/pci: (80 commits)
x86/PCI: Expand the x86_msi_ops to have a restore MSIs.
PCI: Increase resource array mask bit size in pcim_iomap_regions()
PCI: DEVICE_COUNT_RESOURCE should be equal to PCI_NUM_RESOURCES
PCI: pci_ids: add device ids for STA2X11 device (aka ConneXT)
PNP: work around Dell 1536/1546 BIOS MMCONFIG bug that breaks USB
x86/PCI: amd: factor out MMCONFIG discovery
PCI: Enable ATS at the device state restore
PCI: msi: fix imbalanced refcount of msi irq sysfs objects
PCI: kconfig: English typo in pci/pcie/Kconfig
PCI/PM/Runtime: make PCI traces quieter
PCI: remove pci_create_bus()
xtensa/PCI: convert to pci_scan_root_bus() for correct root bus resources
x86/PCI: convert to pci_create_root_bus() and pci_scan_root_bus()
x86/PCI: use pci_scan_bus() instead of pci_scan_bus_parented()
x86/PCI: read Broadcom CNB20LE host bridge info before PCI scan
sparc32, leon/PCI: convert to pci_scan_root_bus() for correct root bus resources
sparc/PCI: convert to pci_create_root_bus()
sh/PCI: convert to pci_scan_root_bus() for correct root bus resources
powerpc/PCI: convert to pci_create_root_bus()
powerpc/PCI: split PHB part out of pcibios_map_io_space()
...
Fix up conflicts in drivers/pci/msi.c and include/linux/pci_regs.h due
to the same patches being applied in other branches.
During S3 or S4 resume or PCI reset, ATS regs aren't restored correctly.
This patch enables ATS at the device state restore if PCI device has ATS
capability.
Signed-off-by: Xudong Hao <xudong.hao@intel.com>
Signed-off-by: Xiantao Zhang <xiantao.zhang@intel.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
These are extended capabilities, rename and move to proper
group for consistency.
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
I get this compile failure on parisc:
drivers/pci/ats.c: In function 'ats_alloc_one':
drivers/pci/ats.c:29: error: implicit declaration of function 'kzalloc'
drivers/pci/ats.c:29: warning: assignment makes pointer from integer without a cast
drivers/pci/ats.c: In function 'ats_free_one':
drivers/pci/ats.c:45: error: implicit declaration of function 'kfree'
Because ats.c is missing linux/slab.h as an include. This patch fixes it
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
More consistency cleanups. Drop the _OFF, separate and indent
CTRL/CAP/STATUS bit definitions. This helped find the previous
mis-use of bit 0 in the PASID capability register.
Reviewed-by: Joerg Roedel <joerg.roedel@amd.com>
Tested-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Jesse Barnes <jbarnes@virtuousgeek.org>
