- Protect driver's D3cold preference from being overwritten by user space
via sysfs (Lukas Wunner)
- Avoid PME from D3hot/D3cold for AMD Rembrandt and Phoenix USB4 to fix
wakeup by USB4-attached devices (Mario Limonciello)
* pci/pm:
x86/PCI: Avoid PME from D3hot/D3cold for AMD Rembrandt and Phoenix USB4
PCI/sysfs: Protect driver's D3cold preference from user space
The acpi_evaluate_dsm_typed() provides a way to check the type of the
object evaluated by _DSM call. Use it instead of open coded variant.
Link: https://lore.kernel.org/r/20231002135352.2603096-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: "Rafael J. Wysocki" <rafael@kernel.org>
struct pci_dev contains two flags which govern whether the device may
suspend to D3cold:
* no_d3cold provides an opt-out for drivers (e.g. if a device is known
to not wake from D3cold)
* d3cold_allowed provides an opt-out for user space (default is true,
user space may set to false)
Since commit 9d26d3a8f1 ("PCI: Put PCIe ports into D3 during suspend"),
the user space setting overwrites the driver setting. Essentially user
space is trusted to know better than the driver whether D3cold is
working.
That feels unsafe and wrong. Assume that the change was introduced
inadvertently and do not overwrite no_d3cold when d3cold_allowed is
modified. Instead, consider d3cold_allowed in addition to no_d3cold
when choosing a suspend state for the device.
That way, user space may opt out of D3cold if the driver hasn't, but it
may no longer force an opt in if the driver has opted out.
Fixes: 9d26d3a8f1 ("PCI: Put PCIe ports into D3 during suspend")
Link: https://lore.kernel.org/r/b8a7f4af2b73f6b506ad8ddee59d747cbf834606.1695025365.git.lukas@wunner.de
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Cc: stable@vger.kernel.org # v4.8+
ACPI r6.5, sec 6.5.4, describes how AML is unable to access an
OperationRegion unless _REG has been called to connect a handler:
The OS runs _REG control methods to inform AML code of a change in the
availability of an operation region. When an operation region handler is
unavailable, AML cannot access data fields in that region. (Operation
region writes will be ignored and reads will return indeterminate data.)
The PCI core does not call _REG at any time, leading to the undefined
behavior mentioned in the spec.
The spec explains that _REG should be executed to indicate whether a
given region can be accessed:
Once _REG has been executed for a particular operation region, indicating
that the operation region handler is ready, a control method can access
fields in the operation region. Conversely, control methods must not
access fields in operation regions when _REG method execution has not
indicated that the operation region handler is ready.
An example included in the spec demonstrates calling _REG when devices are
turned off: "when the host controller or bridge controller is turned off
or disabled, PCI Config Space Operation Regions for child devices are
no longer available. As such, ETH0’s _REG method will be run when it
is turned off and will again be run when PCI1 is turned off."
It is reported that ASMedia PCIe GPIO controllers fail functional tests
after the system has returning from suspend (S3 or s2idle). This is because
the BIOS checks whether the OSPM has called the _REG method to determine
whether it can interact with the OperationRegion assigned to the device as
part of the other AML called for the device.
To fix this issue, call acpi_evaluate_reg() when devices are transitioning
to D3cold or D0.
[bhelgaas: split pci_power_t checking to preliminary patch]
Link: https://uefi.org/specs/ACPI/6.5/06_Device_Configuration.html#reg-region
Link: https://lore.kernel.org/r/20230620140451.21007-1-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rafael J. Wysocki <rafael@kernel.org>
Previously acpi_pci_set_power_state() assumed the requested power state was
valid (PCI_D0 ... PCI_D3cold). If a caller supplied something else, we
could index outside the state_conv[] array and pass junk to
acpi_device_set_power().
Validate the pci_power_t parameter and return -EINVAL if it's invalid.
Link: https://lore.kernel.org/r/20230621222857.GA122930@bhelgaas
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
It is questionable to allow a PCI bridge to go into D3 if it has _S0W
returning D2 or a shallower power state, so modify acpi_pci_bridge_d3(() to
always take the return value of _S0W for the target bridge into account.
That is, make it return 'false' if _S0W returns D2 or a shallower power
state for the target bridge regardless of its ancestor Root Port
properties. Of course, this also causes 'false' to be returned if the Root
Port itself is the target and its _S0W returns D2 or a shallower power
state.
However, still allow bridges without _S0W that are power-manageable via
ACPI to enter D3 to retain the current code behavior in that case.
This fixes problems where a hotplug notification is missed because a bridge
is in D3. That means hot-added devices such as USB4 docks (and the devices
they contain) and Thunderbolt 3 devices may not work.
Link: https://lore.kernel.org/linux-pci/20221031223356.32570-1-mario.limonciello@amd.com/
Link: https://lore.kernel.org/r/12155458.O9o76ZdvQC@kreacher
Reported-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The previous link to the PCI firmware specification in the comments
for drivers/pci/pci-acpi.c no longer works. Update the comment
to a current link to this specification.
Link: https://lore.kernel.org/r/20220722174754.27921-1-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
-----BEGIN PGP SIGNATURE-----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=hEti
-----END PGP SIGNATURE-----
Merge tag 'pci-v5.19-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci
Pull pci updates from Bjorn Helgaas:
"Resource management:
- Restrict E820 clipping to PCI host bridge windows (Bjorn Helgaas)
- Log E820 clipping better (Bjorn Helgaas)
- Add kernel cmdline options to enable/disable E820 clipping (Hans de
Goede)
- Disable E820 reserved region clipping for IdeaPads, Yoga, Yoga
Slip, Acer Spin 5, Clevo Barebone systems where clipping leaves no
usable address space for touchpads, Thunderbolt devices, etc (Hans
de Goede)
- Disable E820 clipping by default starting in 2023 (Hans de Goede)
PCI device hotplug:
- Include files to remove implicit dependencies (Christophe Leroy)
- Only put Root Ports in D3 if they can signal and wake from D3 so
AMD Yellow Carp doesn't miss hotplug events (Mario Limonciello)
Power management:
- Define pci_restore_standard_config() only for CONFIG_PM_SLEEP since
it's unused otherwise (Krzysztof Kozlowski)
- Power up devices completely, including anything platform firmware
needs to do, during runtime resume (Rafael J. Wysocki)
- Move pci_resume_bus() to PM callbacks so we observe the required
bridge power-up delays (Rafael J. Wysocki)
- Drop unneeded runtime_d3cold device flag (Rafael J. Wysocki)
- Split pci_raw_set_power_state() between pci_power_up() and a new
pci_set_low_power_state() (Rafael J. Wysocki)
- Set current_state to D3cold if config read returns ~0, indicating
the device is not accessible (Rafael J. Wysocki)
- Do not call pci_update_current_state() from pci_power_up() so BARs
and ASPM config are restored correctly (Rafael J. Wysocki)
- Write 0 to PMCSR in pci_power_up() in all cases (Rafael J. Wysocki)
- Split pci_power_up() to pci_set_full_power_state() to avoid some
redundant operations (Rafael J. Wysocki)
- Skip restoring BARs if device is not in D0 (Rafael J. Wysocki)
- Rearrange and clarify pci_set_power_state() (Rafael J. Wysocki)
- Remove redundant BAR restores from pci_pm_thaw_noirq() (Rafael J.
Wysocki)
Virtualization:
- Acquire device lock before config space access lock to avoid AB/BA
deadlock with sriov_numvfs_store() (Yicong Yang)
Error handling:
- Clear MULTI_ERR_COR/UNCOR_RCV bits, which a race could previously
leave permanently set (Kuppuswamy Sathyanarayanan)
Peer-to-peer DMA:
- Whitelist Intel Skylake-E Root Ports regardless of which devfn they
are (Shlomo Pongratz)
ASPM:
- Override L1 acceptable latency advertised by Intel DG2 so ASPM L1
can be enabled (Mika Westerberg)
Cadence PCIe controller driver:
- Set up device-specific register to allow PTM Responder to be
enabled by the normal architected bit (Christian Gmeiner)
- Override advertised FLR support since the controller doesn't
implement FLR correctly (Parshuram Thombare)
Cadence PCIe endpoint driver:
- Correct bitmap size for the ob_region_map of outbound window usage
(Dan Carpenter)
Freescale i.MX6 PCIe controller driver:
- Fix PERST# assertion/deassertion so we observe the required delays
before accessing device (Francesco Dolcini)
Freescale Layerscape PCIe controller driver:
- Add "big-endian" DT property (Hou Zhiqiang)
- Update SCFG DT property (Hou Zhiqiang)
- Add "aer", "pme", "intr" DT properties (Li Yang)
- Add DT compatible strings for ls1028a (Xiaowei Bao)
Intel VMD host bridge driver:
- Assign VMD IRQ domain before enumeration to avoid IOMMU interrupt
remapping errors when MSI-X remapping is disabled (Nirmal Patel)
- Revert VMD workaround that kept MSI-X remapping enabled when IOMMU
remapping was enabled (Nirmal Patel)
Marvell MVEBU PCIe controller driver:
- Add of_pci_get_slot_power_limit() to parse the
'slot-power-limit-milliwatt' DT property (Pali Rohár)
- Add mvebu support for sending Set_Slot_Power_Limit message (Pali
Rohár)
MediaTek PCIe controller driver:
- Fix refcount leak in mtk_pcie_subsys_powerup() (Miaoqian Lin)
MediaTek PCIe Gen3 controller driver:
- Reset PHY and MAC at probe time (AngeloGioacchino Del Regno)
Microchip PolarFlare PCIe controller driver:
- Add chained_irq_enter()/chained_irq_exit() calls to mc_handle_msi()
and mc_handle_intx() to avoid lost interrupts (Conor Dooley)
- Fix interrupt handling race (Daire McNamara)
NVIDIA Tegra194 PCIe controller driver:
- Drop tegra194 MSI register save/restore, which is unnecessary since
the DWC core does it (Jisheng Zhang)
Qualcomm PCIe controller driver:
- Add SM8150 SoC DT binding and support (Bhupesh Sharma)
- Fix pipe clock imbalance (Johan Hovold)
- Fix runtime PM imbalance on probe errors (Johan Hovold)
- Fix PHY init imbalance on probe errors (Johan Hovold)
- Convert DT binding to YAML (Dmitry Baryshkov)
- Update DT binding to show that resets aren't required for
MSM8996/APQ8096 platforms (Dmitry Baryshkov)
- Add explicit register names per chipset in DT binding (Dmitry
Baryshkov)
- Add sc7280-specific clock and reset definitions to DT binding
(Dmitry Baryshkov)
Rockchip PCIe controller driver:
- Fix bitmap size when searching for free outbound region (Dan
Carpenter)
Rockchip DesignWare PCIe controller driver:
- Remove "snps,dw-pcie" from rockchip-dwc DT "compatible" property
because it's not fully compatible with rockchip (Peter Geis)
- Reset rockchip-dwc controller at probe (Peter Geis)
- Add rockchip-dwc INTx support (Peter Geis)
Synopsys DesignWare PCIe controller driver:
- Return error instead of success if DMA mapping of MSI area fails
(Jiantao Zhang)
Miscellaneous:
- Change pci_set_dma_mask() documentation references to
dma_set_mask() (Alex Williamson)"
* tag 'pci-v5.19-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (64 commits)
dt-bindings: PCI: qcom: Add schema for sc7280 chipset
dt-bindings: PCI: qcom: Specify reg-names explicitly
dt-bindings: PCI: qcom: Do not require resets on msm8996 platforms
dt-bindings: PCI: qcom: Convert to YAML
PCI: qcom: Fix unbalanced PHY init on probe errors
PCI: qcom: Fix runtime PM imbalance on probe errors
PCI: qcom: Fix pipe clock imbalance
PCI: qcom: Add SM8150 SoC support
dt-bindings: pci: qcom: Document PCIe bindings for SM8150 SoC
x86/PCI: Disable E820 reserved region clipping starting in 2023
x86/PCI: Disable E820 reserved region clipping via quirks
x86/PCI: Add kernel cmdline options to use/ignore E820 reserved regions
PCI: microchip: Fix potential race in interrupt handling
PCI/AER: Clear MULTI_ERR_COR/UNCOR_RCV bits
PCI: cadence: Clear FLR in device capabilities register
PCI: cadence: Allow PTM Responder to be enabled
PCI: vmd: Revert 2565e5b69c ("PCI: vmd: Do not disable MSI-X remapping if interrupt remapping is enabled by IOMMU.")
PCI: vmd: Assign VMD IRQ domain before enumeration
PCI: Avoid pci_dev_lock() AB/BA deadlock with sriov_numvfs_store()
PCI: rockchip-dwc: Add legacy interrupt support
...
The initial configuration of ACPI power resources on some systems
implies that some PCI devices on them are initially in D3cold.
In some cases, especially for PCIe Root Ports, this is a "logical"
D3cold, meaning that the configuration space of the device is
accessible, but some of its functionality may be missing, but it
very well may be real D3cold, in which case the device will not
be accessible at all. However, the PCI bus type driver will need
to access its configuration space in order to enumerate it.
To prevent possible device enumeration failures that may ensue as
a result of ACPI power resources being initially in the "off"
state, power up all children of the host bridge ACPI device object
that hold valid _ADR objects (which indicates that they will be
enumerated by the PCI bus type driver) and do that to all children
of the ACPI device objects corresponding to PCI bridges (including
PCIe ports).
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
acpi_pci_bridge_d3(dev) returns "true" if "dev" is a hotplug bridge that
can handle hotplug events while in D3. Previously this meant either:
- "dev" has a _PS0 or _PR0 method (acpi_pci_power_manageable()), or
- The Root Port above "dev" has a _DSD with a "HotPlugSupportInD3"
property with value 1.
This did not consider _PRW, which tells us about wakeup GPEs (ACPI v6.4,
sec 7.3.13). Without a wakeup GPE, from an ACPI perspective the Root Port
has no way of generating wakeup signals, so hotplug events will be lost if
we use D3.
Similarly, it did not consider _S0W, which tells us the deepest D-state
from which a device can wake itself (sec 7.3.20). If _S0W tells us the
device cannot wake from D3, hotplug events will again be lost if we use D3.
Some platforms, e.g., AMD Yellow Carp, supply "HotPlugSupportInD3" without
_PRW or with an _S0W that says the Root Port cannot wake from D3. On those
platforms, we previously put bridges in D3hot, hotplug events were lost,
and hotplugged devices would not be recognized without manually rescanning.
Allow bridges to be put in D3 only if the Root Port can generate wakeup
GPEs (wakeup.flags.valid), it can wake from D3 (_S0W), AND it has the
"HotPlugSupportInD3" property.
Neither Windows 10 nor Windows 11 puts the bridge in D3 when the firmware
is configured this way, and this change aligns the handling of the
situation to be the same.
[bhelgaas: commit log, tidy "HotPlugSupportInD3" check and comment]
Link: https://uefi.org/htmlspecs/ACPI_Spec_6_4_html/07_Power_and_Performance_Mgmt/device-power-management-objects.html?highlight=s0w#s0w-s0-device-wake-state
Link: https://docs.microsoft.com/en-us/windows-hardware/drivers/pci/dsd-for-pcie-root-ports#identifying-pcie-root-ports-supporting-hot-plug-in-d3
Link: https://lore.kernel.org/r/20220401034003.3166-1-mario.limonciello@amd.com
Fixes: 26ad34d510 ("PCI / ACPI: Whitelist D3 for more PCIe hotplug ports")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Replace acpi_bus_get_device() that is going to be dropped with
acpi_fetch_acpi_dev().
No intentional functional impact.
Link: https://lore.kernel.org/r/11930209.O9o76ZdvQC@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Merge PCI device power management updates for 5.16-rc1:
- Make the association of ACPI device objects with PCI devices more
straightforward and simplify the code doing that for all devices
in general (Rafael Wysocki).
- Eliminate struct pci_platform_pm_ops and handle the both of its
users (PCI and Intel MID) directly in the PCI bus code (Rafael
Wysocki).
- Simplify and clarify ACPI PCI device PM helpers (Rafael Wysocki).
- Fix ordering of operations in pci_back_from_sleep() (Rafael
Wysocki).
* pm-pci:
PCI: PM: Fix ordering of operations in pci_back_from_sleep()
PCI: PM: Do not call platform_pci_power_manageable() unnecessarily
PCI: PM: Make pci_choose_state() call pci_target_state()
PCI: PM: Rearrange pci_target_state()
PCI: PM: Simplify acpi_pci_power_manageable()
PCI: PM: Drop struct pci_platform_pm_ops
PCI: ACPI: PM: Do not use pci_platform_pm_ops for ACPI
PCI: PM: Do not use pci_platform_pm_ops for Intel MID PM
ACPI: glue: Look for ACPI bus type only if ACPI companion is not known
ACPI: glue: Drop cleanup callback from struct acpi_bus_type
PCI: ACPI: Drop acpi_pci_bus
If acpi_pci_find_companion() is called for a device whose parent
pointer is NULL, it will crash when attempting to get the ACPI
companion of the parent due to a NULL pointer dereference in
the ACPI_COMPANION() macro.
This was not a problem before commit 375553a932 ("PCI: Setup ACPI
fwnode early and at the same time with OF") that made pci_setup_device()
call pci_set_acpi_fwnode() and so it allowed devices with NULL parent
pointers to be passed to acpi_pci_find_companion() which is the case
in pci_iov_add_virtfn(), for instance.
Fix this issue by making acpi_pci_find_companion() check the device's
parent pointer upfront and bail out if it is NULL.
While pci_iov_add_virtfn() can be changed to set the device's parent
pointer before calling pci_setup_device() for it, checking pointers
against NULL before dereferencing them is prudent anyway and looking
for ACPI companions of virtual functions isn't really useful.
Fixes: 375553a932 ("PCI: Setup ACPI fwnode early and at the same time with OF")
Link: https://lore.kernel.org/linux-acpi/8e4bbd5c59de31db71f718556654c0aa077df03d.camel@linux.ibm.com/
Reported-by: Niklas Schnelle <schnelle@linux.ibm.com>
Tested-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
The pci_choose_state() and pci_target_state() implementations are
somewhat divergent without a good reason, because they are used
for similar purposes.
Change the pci_choose_state() implementation to use pci_target_state()
internally except for transitions to the working state of the system
in which case it is expected to return D0.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Ferry Toth <fntoth@gmail.com>
Make acpi_pci_power_manageable() more straightforward.
No functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Ferry Toth <fntoth@gmail.com>
Using struct pci_platform_pm_ops for ACPI adds unnecessary
indirection to the interactions between the PCI core and ACPI PM,
which is also subject to retpolines.
Moreover, it is not particularly clear from the current code that,
as far as PCI PM is concerned, "platform" really means just ACPI
except for the special casess when Intel MID PCI PM is used or when
ACPI support is disabled (through the kernel config or command line,
or because there are no usable ACPI tables on the system).
To address the above, rework the PCI PM code to invoke ACPI PM
functions directly as needed and drop the acpi_pci_platform_pm
object that is not necessary any more.
Accordingly, update some of the ACPI PM functions in question to do
extra checks in case the ACPI support is disabled (which previously
was taken care of by avoiding to set the pci_platform_ops pointer
in those cases).
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Ferry Toth <fntoth@gmail.com>
The acpi_pci_bus structure was used primarily for running
acpi_pci_find_companion() during PCI device objects registration,
but after commit 375553a932 ("PCI: Setup ACPI fwnode early and at
the same time with OF") that function is called by pci_setup_device()
via pci_set_acpi_fwnode(), which happens before calling
pci_device_add() on the new PCI device object, so its ACPI companion
has been set already when acpi_device_notify() runs and it will never
call ->find_companion() from acpi_pci_bus.
For this reason, modify acpi_device_notify() and
acpi_device_notify_remove() to call pci_acpi_setup() and
pci_acpi_cleanup(), respectively, directly on PCI device objects
and drop acpi_pci_bus altogether.
While at it, notice that pci_acpi_setup() and pci_acpi_cleanup()
can obtain the ACPI companion pointer, which is guaranteed to not
be NULL, from their callers and modify them to work that way so
as to reduce the number of redundant checks somewhat.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Tested-by: Ferry Toth <fntoth@gmail.com>
Commit 375553a932 ("PCI: Setup ACPI fwnode early and at the same time
with OF") added a call to pci_set_acpi_fwnode() in pci_setup_device(),
which unconditionally clears any fwnode previously set by
pci_set_of_node().
pci_set_acpi_fwnode() looks for ACPI_COMPANION(), which only returns the
existing fwnode if it was set by ACPI_COMPANION_SET(). If it was set by
OF instead, ACPI_COMPANION() returns NULL and pci_set_acpi_fwnode()
accidentally clears the fwnode. To fix this, look for any fwnode instead
of just ACPI companions.
Fixes a virtio-iommu boot regression in v5.15-rc1.
Fixes: 375553a932 ("PCI: Setup ACPI fwnode early and at the same time with OF")
Link: https://lore.kernel.org/r/20210913172358.1775381-1-jean-philippe@linaro.org
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Rob Herring <robh@kernel.org>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
- Add ACPI support to the PCI VMD driver (Rafael Wysocki).
- Rearrange suspend-to-idle support code to reflect the platform
firmware expectations on some AMD platforms (Mario Limonciello).
- Make SSDT overlays documentation follow the code documented by it
more closely (Andy Shevchenko).
-----BEGIN PGP SIGNATURE-----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=8lPe
-----END PGP SIGNATURE-----
Merge tag 'acpi-5.15-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Pull more ACPI updates from Rafael Wysocki:
"These add ACPI support to the PCI VMD driver, improve suspend-to-idle
support for AMD platforms and update documentation.
Specifics:
- Add ACPI support to the PCI VMD driver (Rafael Wysocki)
- Rearrange suspend-to-idle support code to reflect the platform
firmware expectations on some AMD platforms (Mario Limonciello)
- Make SSDT overlays documentation follow the code documented by it
more closely (Andy Shevchenko)"
* tag 'acpi-5.15-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
ACPI: PM: s2idle: Run both AMD and Microsoft methods if both are supported
Documentation: ACPI: Align the SSDT overlays file with the code
PCI: VMD: ACPI: Make ACPI companion lookup work for VMD bus
On some systems, in order to get to the deepest low-power state of
the platform (which may be necessary to save significant enough
amounts of energy while suspended to idle. for example), devices on
the PCI bus exposed by the VMD driver need to be power-managed via
ACPI. However, the layout of the ACPI namespace below the VMD
controller device object does not reflect the layout of the PCI bus
under the VMD host bridge, so in order to identify the ACPI companion
objects for the devices on that bus, it is necessary to use a special
_ADR encoding on the ACPI side. In other words, acpi_pci_find_companion()
does not work for these devices, so it needs to be amended with a
special lookup logic specific to the VMD bus.
Address this issue by allowing the VMD driver to temporarily install
an ACPI companion lookup hook containing the code matching the devices
on the VMD PCI bus with the corresponding objects in the ACPI
namespace.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Jon Derrick <jonathan.derrick@intel.com>
Change the type of probe argument in functions which implement reset
methods from int to bool to make the context and intent clear.
Suggested-by: Alex Williamson <alex.williamson@redhat.com>
Link: https://lore.kernel.org/r/20210817180500.1253-10-ameynarkhede03@gmail.com
Signed-off-by: Amey Narkhede <ameynarkhede03@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
_RST is a standard ACPI method that performs a function level reset of a
device (ACPI v6.3, sec 7.3.25).
Add pci_dev_acpi_reset() to probe for _RST method and execute if present.
The default priority of this reset is set to below device-specific and
above hardware resets.
Suggested-by: Alex Williamson <alex.williamson@redhat.com>
Link: https://lore.kernel.org/r/20210817180500.1253-9-ameynarkhede03@gmail.com
Signed-off-by: Shanker Donthineni <sdonthineni@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Sinan Kaya <okaya@kernel.org>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Previously, the ACPI_COMPANION() of a pci_dev was usually set by
acpi_bind_one() in this path:
pci_device_add
pci_configure_device
pci_init_capabilities
device_add
device_platform_notify
acpi_platform_notify
acpi_device_notify # KOBJ_ADD
acpi_bind_one
ACPI_COMPANION_SET
However, things like pci_configure_device() and pci_init_capabilities()
that run before device_add() need the ACPI_COMPANION, e.g.,
acpi_pci_bridge_d3() uses a _DSD method to learn about D3 support. These
places had special-case code to manually look up the ACPI_COMPANION.
Set the ACPI_COMPANION earlier, in pci_setup_device(), so it will be
available while configuring the device. This covers both paths to creating
pci_dev objects:
pci_scan_single_device # for normal non-SR-IOV devices
pci_scan_device
pci_setup_device
pci_set_acpi_fwnode
pci_device_add
pci_iov_add_virtfn # for SR-IOV virtual functions
pci_setup_device
pci_set_acpi_fwnode
Also move the OF fwnode setup to the same spot.
[bhelgaas: commit log]
Link: https://lore.kernel.org/r/20210817180500.1253-8-ameynarkhede03@gmail.com
Signed-off-by: Shanker Donthineni <sdonthineni@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Use acpi_pci_power_manageable() instead of duplicating the logic in
acpi_pci_bridge_d3(). No functional change intended.
[bhelgaas: split out from
https://lore.kernel.org/r/20210817180500.1253-8-ameynarkhede03@gmail.com]
Signed-off-by: Shanker Donthineni <sdonthineni@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Move the existing logic from acpi_pci_bridge_d3() to a separate function
pci_set_acpi_fwnode() to set the ACPI fwnode. No functional change
intended.
Link: https://lore.kernel.org/r/20210817180500.1253-7-ameynarkhede03@gmail.com
Signed-off-by: Shanker Donthineni <sdonthineni@nvidia.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
If PCI_D3cold is passed to acpi_pci_set_power_state() as the second
argument and there is no ACPI D3cold support for the given device, the
debug message will state that the device power state has been changed to
D3cold, while in fact it will be D3hot, because acpi_device_set_power()
falls back to D3hot automatically if D3cold is not supported without
returning an error.
To address this issue, modify the debug message to print the current power
state of the target PCI device's ACPI companion instead of printing the
target power state, which may not reflect the real final power state of the
device.
Link: https://lore.kernel.org/r/4319486.LvFx2qVVIh@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Krzysztof Wilczyński <kw@linux.com>
-----BEGIN PGP SIGNATURE-----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=xRTv
-----END PGP SIGNATURE-----
Merge tag 'pci-v5.11-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci
Pull PCI updates from Bjorn Helgaas:
"Enumeration:
- Decode PCIe 64 GT/s link speed (Gustavo Pimentel)
- Remove unused HAVE_PCI_SET_MWI (Heiner Kallweit)
- Reduce pci_set_cacheline_size() message to debug level (Heiner
Kallweit)
- Fix pci_slot_release() NULL pointer dereference (Jubin Zhong)
- Unify ECAM constants in native PCI Express drivers (Krzysztof
Wilczyński)
- Return u8 from pci_find_capability() and similar (Puranjay Mohan)
- Return u16 from pci_find_ext_capability() and similar (Bjorn
Helgaas)
- Fix ACPI companion lookup for device 0 on the root bus (Rafael J.
Wysocki)
Resource management:
- Keep both device and resource name for config space remaps
(Alexander Lobakin)
- Bounds-check command-line resource alignment requests (Bjorn
Helgaas)
- Fix overflow in command-line resource alignment requests (Colin Ian
King)
Driver binding:
- Avoid duplicate IDs in driver dynamic IDs list (Zhenzhong Duan)
Power management:
- Save/restore Precision Time Measurement Capability for
suspend/resume (David E. Box)
- Disable PTM during suspend to save power (David E. Box)
- Add sysfs attribute for device power state (Maximilian Luz)
- Rename pci_wakeup_bus() to pci_resume_bus() (Mika Westerberg)
- Do not generate wakeup event when runtime resuming device (Mika
Westerberg)
- Save/restore ASPM L1SS Capability for suspend/resume (Vidya Sagar)
Virtualization:
- Mark AMD Raven iGPU ATS as broken in some platforms (Alex Deucher)
- Add function 1 DMA alias quirk for Marvell 9215 SATA controller
(Bjorn Helgaas)
MSI:
- Disable MSI for Pericom PCIe-USB adapter (Andy Shevchenko)
- Improve warnings for 32-bit-limited MSI support (Vidya Sagar)
Error handling:
- Cache RCEC EA Capability offset in pci_init_capabilities() (Sean V
Kelley)
- Rename reset_link() to reset_subordinates() (Sean V Kelley)
- Write AER Capability only when we control it (Sean V Kelley)
- Clear AER status only when we control AER (Sean V Kelley)
- Bind RCEC devices to the Root Port driver (Qiuxu Zhuo)
- Recover from RCiEP AER errors (Qiuxu Zhuo)
- Recover from RCEC AER errors (Sean V Kelley)
- Add pcie_link_rcec() to associate RCiEPs (Sean V Kelley)
- Add pcie_walk_rcec() to RCEC AER handling (Sean V Kelley)
- Add pcie_walk_rcec() to RCEC PME handling (Sean V Kelley)
- Add RCEC AER error injection support (Qiuxu Zhuo)
Broadcom iProc PCIe controller driver:
- Fix out-of-bound array accesses (Bharat Gooty)
- Invalidate correct PAXB inbound windows (Roman Bacik)
- Enhance PCIe Link information display (Srinath Mannam)
Cadence PCIe controller driver:
- Make "cdns,max-outbound-regions" property optional (Kishon Vijay
Abraham I)
Intel VMD host bridge driver:
- Offset client MSI-X vectors (Jon Derrick)
- Update type of __iomem pointers (Krzysztof Wilczyński)
NVIDIA Tegra PCIe controller driver:
- Move "dbi" accesses to post common DWC initialization (Vidya Sagar)
- Read "dbi" base address to program in application logic (Vidya
Sagar)
- Fix ASPM-L1SS advertisement disable code (Vidya Sagar)
- Set DesignWare IP version (Vidya Sagar)
- Continue unconfig sequence even if parts fail (Vidya Sagar)
- Check return value of tegra_pcie_init_controller() (Vidya Sagar)
- Disable LTSSM during L2 entry (Vidya Sagar)
Qualcomm PCIe controller driver:
- Document PCIe bindings for SM8250 SoC (Manivannan Sadhasivam)
- Add SM8250 SoC support (Manivannan Sadhasivam)
- Add support for configuring BDF to SID mapping for SM8250
(Manivannan Sadhasivam)
Renesas R-Car PCIe controller driver:
- rcar: Drop unused members from struct rcar_pcie_host (Lad
Prabhakar)
- PCI: rcar-pci-host: Document r8a774e1 bindings (Lad Prabhakar)
- PCI: rcar-pci-host: Convert bindings to json-schema (Yoshihiro
Shimoda)
- PCI: rcar-pci-host: Document r8a77965 bindings (Yoshihiro Shimoda)
Samsung Exynos PCIe controller driver:
- Rework driver to support Exynos5433 PCIe PHY (Jaehoon Chung)
- Rework driver to support Exynos5433 variant (Jaehoon Chung)
- Drop samsung,exynos5440-pcie binding (Marek Szyprowski)
- Add the samsung,exynos-pcie binding (Marek Szyprowski)
- Add the samsung,exynos-pcie-phy binding (Marek Szyprowski)
Synopsys DesignWare PCIe controller driver:
- Support multiple ATU memory regions (Rob Herring)
- Move intel-gw ATU offset out of driver match data (Rob Herring)
- Move "dbi", "dbi2", and "addr_space" resource setup into common
code (Rob Herring)
- Remove intel-gw unneeded function wrappers (Rob Herring)
- Ensure all outbound ATU windows are reset (Rob Herring)
- Use the common MSI irq_chip in dra7xx (Rob Herring)
- Drop the .set_num_vectors() host op (Rob Herring)
- Move MSI interrupt setup into DWC common code (Rob Herring)
- Rework MSI initialization (Rob Herring)
- Move link handling into common code (Rob Herring)
- Move dw_pcie_msi_init() into core (Rob Herring)
- Move dw_pcie_setup_rc() to DWC common code (Rob Herring)
- Remove unnecessary wrappers around dw_pcie_host_init() (Rob
Herring)
- Drop keystone duplicated 'num-viewport'" (Rob Herring)
- Move inbound and outbound windows to common struct (Rob Herring)
- Detect number of iATU windows (Rob Herring)
- Warn if non-prefetchable memory aperture size is > 32-bit (Vidya
Sagar)
- Add support to program ATU for >4GB memory (Vidya Sagar)
- Set 32-bit DMA mask for MSI target address allocation (Vidya Sagar)
TI J721E PCIe driver:
- Fix "ti,syscon-pcie-ctrl" to take argument (Kishon Vijay Abraham I)
- Add host mode dt-bindings for TI's J7200 SoC (Kishon Vijay Abraham
I)
- Add EP mode dt-bindings for TI's J7200 SoC (Kishon Vijay Abraham I)
- Get offset within "syscon" from "ti,syscon-pcie-ctrl" phandle arg
(Kishon Vijay Abraham I)
TI Keystone PCIe controller driver:
- Enable compile-testing on !ARM (Alex Dewar)"
* tag 'pci-v5.11-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (100 commits)
PCI: Add function 1 DMA alias quirk for Marvell 9215 SATA controller
PCI/ACPI: Fix companion lookup for device 0 on the root bus
PCI: Keep both device and resource name for config space remaps
PCI: xgene: Removed unused ".bus_shift" initialisers from pci-xgene.c
PCI: vmd: Update type of the __iomem pointers
PCI: iproc: Convert to use the new ECAM constants
PCI: thunder-pem: Add constant for custom ".bus_shift" initialiser
PCI: Unify ECAM constants in native PCI Express drivers
PCI: Disable PTM during suspend to save power
PCI/PTM: Save/restore Precision Time Measurement Capability for suspend/resume
PCI: Mark AMD Raven iGPU ATS as broken in some platforms
PCI: j721e: Get offset within "syscon" from "ti,syscon-pcie-ctrl" phandle arg
dt-bindings: PCI: Add EP mode dt-bindings for TI's J7200 SoC
dt-bindings: PCI: Add host mode dt-bindings for TI's J7200 SoC
dt-bindings: pci: ti,j721e: Fix "ti,syscon-pcie-ctrl" to take argument
PCI: dwc: Set 32-bit DMA mask for MSI target address allocation
PCI: qcom: Add support for configuring BDF to SID mapping for SM8250
PCI: Reduce pci_set_cacheline_size() message to debug level
PCI: Remove unused HAVE_PCI_SET_MWI
PCI: qcom: Add SM8250 SoC support
...
In some cases acpi_pci_find_companion() returns an incorrect device object
as the ACPI companion for device 0 on the root bus (bus 0).
On the affected systems that device is the PCI interface to the host bridge
and the "ACPI companion" returned for it corresponds to a non-PCI device
located in the SoC (e.g. a sensor on an I2C bus). As a result of this, the
ACPI device object "attached" to PCI device 00:00.0 cannot be used for
enumerating the device that is really represented by it which (of course)
is problematic.
Address that issue by preventing acpi_pci_find_companion() from returning a
device object with a valid _HID (which by the spec should not be present
uder ACPI device objects corresponding to PCI devices) for PCI device
00:00.0.
[bhelgaas: use pci_is_root_bus()]
Link: https://lore.kernel.org/linux-acpi/1409ba0c-1580-dc09-e6fe-a0c9bcda6462@gmail.com/
Link: https://lore.kernel.org/r/4673285.9aE2nYKHPr@kreacher
Reported-by: Daniel Scally <djrscally@gmail.com>
Tested-by: Daniel Scally <djrscally@gmail.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Daniel Scally <djrscally@gmail.com>
The idea behind acpi_pm_set_bridge_wakeup() was to allow bridges to
be reference counted for wakeup enabling, because they may be enabled
to signal wakeup on behalf of their subordinate devices and that
may happen for multiple times in a row, whereas for the other devices
it only makes sense to enable wakeup signaling once.
However, this becomes problematic if the bridge itself is suspended,
because it is treated as a "regular" device in that case and the
reference counting doesn't work.
For instance, suppose that there are two devices below a bridge and
they both can signal wakeup. Every time one of them is suspended,
wakeup signaling is enabled for the bridge, so when they both have
been suspended, the bridge's wakeup reference counter value is 2.
Say that the bridge is suspended subsequently and acpi_pci_wakeup()
is called for it. Because the bridge can signal wakeup, that
function will invoke acpi_pm_set_device_wakeup() to configure it
and __acpi_pm_set_device_wakeup() will be called with the last
argument equal to 1. This causes __acpi_device_wakeup_enable()
invoked by it to omit the reference counting, because the reference
counter of the target device (the bridge) is 2 at that time.
Now say that the bridge resumes and one of the device below it
resumes too, so the bridge's reference counter becomes 0 and
wakeup signaling is disabled for it, but there is still the other
suspended device which may need the bridge to signal wakeup on its
behalf and that is not going to work.
To address this scenario, use wakeup enable reference counting for
all devices, not just for bridges, so drop the last argument from
__acpi_device_wakeup_enable() and __acpi_pm_set_device_wakeup(),
which causes acpi_pm_set_device_wakeup() and
acpi_pm_set_bridge_wakeup() to become identical, so drop the latter
and use the former instead of it everywhere.
Fixes: 1ba51a7c14 ("ACPI / PCI / PM: Rework acpi_pci_propagate_wakeup()")
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: 4.14+ <stable@vger.kernel.org> # 4.14+
Recent laptops with dual AMD GPUs fail to suspend the discrete GPU, thus
causing lockups on system sleep and high power consumption at runtime.
The discrete GPU would normally be suspended to D3cold by turning off
ACPI _PR3 Power Resources of the Root Port above the GPU.
However on affected systems, the Root Port is hotplug-capable and
pci_bridge_d3_possible() only allows hotplug ports to go to D3 if they
belong to a Thunderbolt device or if the Root Port possesses a
"HotPlugSupportInD3" ACPI property. Neither is the case on affected
laptops. The reason for whitelisting only specific, known to work
hotplug ports for D3 is that there have been reports of SkyLake Xeon-SP
systems raising Hardware Error NMIs upon suspending their hotplug ports:
https://lore.kernel.org/linux-pci/20170503180426.GA4058@otc-nc-03/
But if a hotplug port is power manageable by ACPI (as can be detected
through presence of Power Resources and corresponding _PS0 and _PS3
methods) then it ought to be safe to suspend it to D3. To this end,
amend acpi_pci_bridge_d3() to whitelist such ports for D3.
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1222
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1252
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1304
Reported-and-tested-by: Arthur Borsboom <arthurborsboom@gmail.com>
Reported-and-tested-by: matoro <matoro@airmail.cc>
Reported-by: Aaron Zakhrov <aaron.zakhrov@gmail.com>
Reported-by: Michal Rostecki <mrostecki@suse.com>
Reported-by: Shai Coleman <git@shaicoleman.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Cc: 5.4+ <stable@vger.kernel.org> # 5.4+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
PCI devices support two variants of the D3 power state: D3hot (main power
present) D3cold (main power removed). Previously struct pci_dev contained:
unsigned int d3_delay; /* D3->D0 transition time in ms */
unsigned int d3cold_delay; /* D3cold->D0 transition time in ms */
"d3_delay" refers specifically to the D3hot state. Rename it to
"d3hot_delay" to avoid ambiguity and align with the ACPI "_DSM for
Specifying Device Readiness Durations" in the PCI Firmware spec r3.2,
sec 4.6.9.
There is no change to the functionality.
Link: https://lore.kernel.org/r/20200730210848.1578826-1-kw@linux.com
Signed-off-by: Krzysztof Wilczyński <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Rename PCI-related _DSM constants to align them with the PCI Firmware Spec,
r3.2, sec 4.6. No functional change intended.
Link: https://lore.kernel.org/r/20200526213905.2479381-1-kw@linux.com
Signed-off-by: Krzysztof Wilczyński <kw@linux.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Previously we used pcie_find_root_port() to find a Root Port from a PCIe
device and pci_find_pcie_root_port() to find a Root Port from a
Conventional PCI device.
Unify the two functions and use pcie_find_root_port() to find a Root Port
from either a Conventional PCI device or a PCIe device. Then there is no
need to distinguish the type of the device.
Link: https://lore.kernel.org/r/1589019568-5216-1-git-send-email-yangyicong@hisilicon.com
Signed-off-by: Yicong Yang <yangyicong@hisilicon.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless
Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com> # thunderbolt
- Move _HPX type array from stack to static data (Colin Ian King)
- Avoid an ASMedia XHCI USB PME# defect; apparently it doesn't assert
PME# when USB3.0 devices are hotplugged in D0 (Kai-Heng Feng)
- Revert sysfs "rescan" file renames that broke an application (Kelsey
Skunberg)
* pci/misc:
PCI: sysfs: Revert "rescan" file renames
PCI: Avoid ASMedia XHCI USB PME# from D0 defect
PCI/ACPI: Move pcie_to_hpx3_type[] from stack to static data
Error Disconnect Recover (EDR) is a feature that allows ACPI firmware to
notify OSPM that a device has been disconnected due to an error condition
(ACPI v6.3, sec 5.6.6). OSPM advertises its support for EDR on PCI devices
via _OSC (see [1], sec 4.5.1, table 4-4). The OSPM EDR notify handler
should invalidate software state associated with disconnected devices and
may attempt to recover them. OSPM communicates the status of recovery to
the firmware via _OST (sec 6.3.5.2).
For PCIe, firmware may use Downstream Port Containment (DPC) to support
EDR. Per [1], sec 4.5.1, table 4-6, even if firmware has retained control
of DPC, OSPM may read/write DPC control and status registers during the EDR
notification processing window, i.e., from the time it receives an EDR
notification until it clears the DPC Trigger Status.
Note that per [1], sec 4.5.1 and 4.5.2.4,
1. If the OS supports EDR, it should advertise that to firmware by
setting OSC_PCI_EDR_SUPPORT in _OSC Support.
2. If the OS sets OSC_PCI_EXPRESS_DPC_CONTROL in _OSC Control to request
control of the DPC capability, it must also set OSC_PCI_EDR_SUPPORT in
_OSC Support.
Add an EDR notify handler to attempt recovery.
[1] Downstream Port Containment Related Enhancements ECN, Jan 28, 2019,
affecting PCI Firmware Specification, Rev. 3.2
https://members.pcisig.com/wg/PCI-SIG/document/12888
[bhelgaas: squash add/enable patches into one]
Link: https://lore.kernel.org/r/90f91fe6d25c13f9d2255d2ce97ca15be307e1bb.1585000084.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: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Len Brown <lenb@kernel.org>
Move pcie_to_hpx3_type[] from the stack to static data. This reduces stack
usage and also makes the object code slightly smaller.
Link: https://lore.kernel.org/r/20200210085256.319424-1-colin.king@canonical.com
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
- 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_*
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 87fcf12e84 ("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 f873c51a15 ("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>
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>
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>
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>
In pci_pm_complete() there are checks to decide whether or not to
resume devices that were left in runtime-suspend during the preceding
system-wide transition into a sleep state. They involve checking the
current power state of the device and comparing it with the power
state of it set before the preceding system-wide transition, but the
platform component of the device's power state is not handled
correctly in there.
Namely, on platforms with ACPI, the device power state information
needs to be updated with care, so that the reference counters of
power resources used by the device (if any) are set to ensure that
the refreshed power state of it will be maintained going forward.
To that end, introduce a new ->refresh_state() platform PM callback
for PCI devices, for asking the platform to refresh the device power
state data and ensure that the corresponding power state will be
maintained going forward, make it invoke acpi_device_update_power()
(for devices with ACPI PM) on platforms with ACPI and make
pci_pm_complete() use it, through a new pci_refresh_power_state()
wrapper function.
Fixes: a0d2a959d3 (PCI: Avoid unnecessary resume after direct-complete)
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
If otherwise unrelated PCI devices share ACPI power resources turning
them on causes the devices to enter D0uninitialized power state which may
cause problems.
For example in Intel Ice Lake two root ports (RP0 and RP1), Thunderbolt
controller (NHI) and xHCI controller all share power resources as can be
ween in the topology below where power resources are marked with []:
Host bridge
|
+- RP0 ---\
+- RP1 ---|--+--> [TBT]
+- NHI --/ |
| |
| v
+- xHCI --> [D3C]
In a situation where all devices sharing the power resources are in
D3cold (the power resources are turned off) and for example the
Thunderbolt controller is runtime resumed resulting that the power
resources are turned on. This means that the other devices sharing them
(RP0, RP1 and xHCI) are transitioned into D0uninitialized state. If they
were configured to trigger wake (PME) on a certain event that
configuration gets lost after reset so we would need to re-initialize
them to get the wakeup working as expected again. To do so we would need
to runtime resume all of them to make sure their registers get restored
properly before we can runtime suspend them again.
Since we just added concept of "_PR0 dependent device" we can solve this
by calling the relevant add/remove functions when the PCI device is bind
to its ACPI representation. If it has power resources the PCI device
will be added as dependent device to them and runtime resumed whenever
they are physically turned on. This should make sure PCI core can
reconfigure wakes after the device is transitioned into D0uninitialized.
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The ACPI power state returned by acpi_device_get_power() may depend on
the configuration of ACPI power resources in the system which may change
any time after acpi_device_get_power() has returned, unless the
reference counters of the ACPI power resources in question are set to
prevent that from happening. Thus it is invalid to use acpi_device_get_power()
in acpi_pci_get_power_state() the way it is done now and the value of
the ->power.state field in the corresponding struct acpi_device objects
(which reflects the ACPI power resources reference counting, among other
things) should be used instead.
As an example where this becomes an issue is Intel Ice Lake where the
Thunderbolt controller (NHI), two PCIe root ports (RP0 and RP1) and xHCI
all share the same power resources. The following picture with power
resources marked with [] shows the topology:
Host bridge
|
+- RP0 ---\
+- RP1 ---|--+--> [TBT]
+- NHI --/ |
| |
| v
+- xHCI --> [D3C]
Here TBT and D3C are the shared ACPI power resources. ACPI _PR3() method
of the devices in question returns either TBT or D3C or both.
Say we runtime suspend first the root ports RP0 and RP1, then NHI. Now
since the TBT power resource is still on when the root ports are runtime
suspended their dev->current_state is set to D3hot. When NHI is runtime
suspended TBT is finally turned off but state of the root ports remain
to be D3hot. Now when the xHCI is runtime suspended D3C gets also turned
off. PCI core thus has power states of these devices cached in their
dev->current_state as follows:
RP0 -> D3hot
RP1 -> D3hot
NHI -> D3cold
xHCI -> D3cold
If the user now runs lspci for instance, the result is all 1's like in
the below output (00:07.0 is the first root port, RP0):
00:07.0 PCI bridge: Intel Corporation Device 8a1d (rev ff) (prog-if ff)
!!! Unknown header type 7f
Kernel driver in use: pcieport
In short the hardware state is not in sync with the software state
anymore. The exact same thing happens with the PME polling thread which
ends up bringing the root ports back into D0 after they are runtime
suspended.
For this reason, modify acpi_pci_get_power_state() so that it uses the
ACPI device power state that was cached by the ACPI core. This makes the
PCI device power state match the ACPI device power state regardless of
state of the shared power resources which may still be on at this point.
Link: https://lore.kernel.org/r/20190618161858.77834-2-mika.westerberg@linux.intel.com
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Both acpi_pci_need_resume() and acpi_dev_needs_resume() check if the
current ACPI wakeup configuration of the device matches what is
expected as far as system wakeup from sleep states is concerned, as
reflected by the device_may_wakeup() return value for the device.
However, they only should do that if wakeup.flags.valid is set for
the device's ACPI companion, because otherwise the wakeup.prepare_count
value for it is meaningless.
Add the missing wakeup.flags.valid checks to these functions.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
