According to the PCIe standard the PERST# signal (reset-gpio in
fsl,imx* compatible dts) should be kept asserted for at least 100 usec
before the PCIe refclock is stable, should be kept asserted for at
least 100 msec after the power rails are stable and the host should wait
at least 100 msec after it is de-asserted before accessing the
configuration space of any attached device.
From PCIe CEM r2.0, sec 2.6.2
T-PVPERL: Power stable to PERST# inactive - 100 msec
T-PERST-CLK: REFCLK stable before PERST# inactive - 100 usec.
From PCIe r5.0, sec 6.6.1
With a Downstream Port that does not support Link speeds greater than
5.0 GT/s, software must wait a minimum of 100 ms before sending a
Configuration Request to the device immediately below that Port.
Failure to do so could prevent PCIe devices to be working correctly,
and this was experienced with real devices.
Move reset assert to imx6_pcie_assert_core_reset(), this way we ensure
that PERST# is asserted before enabling any clock, move de-assert to the
end of imx6_pcie_deassert_core_reset() after the clock is enabled and
deemed stable and add a new delay of 100 msec just afterward.
Link: https://lore.kernel.org/all/20220211152550.286821-1-francesco.dolcini@toradex.com
Link: https://lore.kernel.org/r/20220404081509.94356-1-francesco.dolcini@toradex.com
Fixes: bb38919ec5 ("PCI: imx6: Add support for i.MX6 PCIe controller")
Signed-off-by: Francesco Dolcini <francesco.dolcini@toradex.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Lucas Stach <l.stach@pengutronix.de>
Acked-by: Richard Zhu <hongxing.zhu@nxp.com>
Remove code duplication by moving the code related to enabling/disabling
the resources (PHY, CLK, Reset) to common functions so that they can be
called from multiple places.
[mani: renamed the functions and reworded the commit message]
Link: https://lore.kernel.org/r/20220502104938.97033-1-manivannan.sadhasivam@linaro.org
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Calling pci_set_power_state() to put the given device into D0 in
pci_pm_thaw_noirq() may cause it to restore the device's BARs, which is
redundant before calling pci_restore_state(), so replace it with a direct
pci_power_up() call followed by pci_update_current_state() if it returns a
nonzero value, in analogy with pci_pm_default_resume_early().
Avoid code duplication by introducing a wrapper function to contain the
repeating pattern and calling it in both places.
Link: https://lore.kernel.org/r/3639079.MHq7AAxBmi@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The part of pci_set_power_state() related to transitions into
low-power states is unnecessary convoluted, so clearly divide it
into the D3cold special case and the general case covering all of
the other states.
Also fix a potential issue with calling pci_bus_set_current_state()
to set the current state of all devices on the subordinate bus to
D3cold without checking if the power state of the parent bridge has
really changed to D3cold.
Link: https://lore.kernel.org/r/2139440.Mh6RI2rZIc@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>
Make the following assorted non-essential changes in
pci_set_low_power_state():
1. Drop two redundant checks from it (the caller takes care of these
conditions).
2. Change the log level of a messages printed by it to "debug",
because it only indicates a programming mistake.
Link: https://lore.kernel.org/r/2539071.Lt9SDvczpP@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>
Do not attempt to restore the device's BARs in
pci_set_full_power_state() if the actual current
power state of the device is not D0.
Link: https://lore.kernel.org/r/1849718.CQOukoFCf9@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
One of the two callers of pci_power_up() invokes
pci_update_current_state() and pci_restore_state() right after calling
it, in which case running the part of it happening after the mandatory
transition delays is redundant, so move that part out of it into a new
function called pci_set_full_power_state() that will be invoked from
pci_set_power_state() which is the other caller of pci_power_up().
Link: https://lore.kernel.org/r/1942150.usQuhbGJ8B@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Make pci_power_up() write 0 to the device's PCI_PM_CTRL register in
order to put it into D0 regardless of the power state returned by
the previous read from that register which should not matter.
Link: https://lore.kernel.org/r/5748066.MhkbZ0Pkbq@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Notice that calling pci_update_current_state() from pci_power_up() is
redundant and may be harmful in some cases.
First, if the device is in a low-power state before pci_power_up()
gets called for it and platform_pci_set_power_state() successfully
changes its power state to D0, pci_update_current_state() will update
current_state to reflect that and pci_power_up() will return success
right away without restoring the device's BARs or reconfiguring ASPM
which may be necessary. This is arguably incorrect and definitely
inconsistent with the case when platform_pci_set_power_state() returns
an error (for example, because the device is not power-manageable by
the platform firmware).
Second, current_state should not be overwritten until the decision
whether or not to restore the device's BARs is made, because that
decision generally depends on its value. Again, calling
pci_update_current_state() in pci_power_up() is not consistent with
this observation.
Next, pci_power_up() attempts to read from the device's PCI_PM_CTRL
register regardless of the current_state value unless it is PCI_D0,
including the case when pci_update_current_state() sets current_state
to PCI_D3cold to indicate that the device is not accessible. If the
register read is not successful, current_state will be set to
PCI_D3cold anyway, so that pci_update_current_state() action is
redundant.
Further, if pci_update_current_state() reads the device's PCI_PM_CTRL
register, pci_power_up() will repeat that read going forward and
it is not necessary to update current_state in the meantime.
Finally, if pm_cap is not set (in which case the PCI_PM_CTRL register
is not present), the power state of the device should be determined
with the help of the platform firmware or set to D0 if that's not
possible and pci_update_current_state() does not do that.
Accordingly, rearrange pci_power_up() so as to address the above
shortcomings.
Link: https://lore.kernel.org/r/3695055.kQq0lBPeGt@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Some actions carried out by pci_platform_power_transition(() in
pci_power_up() are redundant, but before dealing with them, make
pci_power_up() call the pci_platform_power_transition() code directly
(and avoid a redundant check when pm_cap is unset while at it).
Link: https://lore.kernel.org/r/1922486.PYKUYFuaPT@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Make pci_power_up() and pci_set_low_power_state() change current_state
to PCI_D3cold when the device is not accessible along the lines of
pci_update_current_state().
Link: https://lore.kernel.org/r/10104376.nUPlyArG6x@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Because pci_set_power_state() is the only caller of
pci_set_low_power_state(), put the latter next to the former.
No functional impact.
Link: https://lore.kernel.org/r/3202976.44csPzL39Z@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The transitions from low-power states to D0 and the other way around
are unnecessarily tangled in pci_raw_set_power_state() which makes it
rather hard to follow.
Moreover, the only caller of pci_raw_set_power_state() passing PCI_D0
as its state argument is pci_power_up(), so the code carrying out
transitions into D0 can be put directly into that function.
Accordingly, move the code handling transitions from low-power states
into D0 directly into pci_power_up() and rename the remaining part
of pci_raw_set_power_state() to pci_set_low_power_state(), because
it only handles transitions into low-power state now.
While at it, fix up some white space, update some comments and modify
messages printed by pci_power_up() and pci_set_low_power_state() to
be less confusing (which is the only expected functional impact of
this change).
Link: https://lore.kernel.org/r/13038676.uLZWGnKmhe@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Save one config space access in pci_update_current_state() by testing the
retrieved PCI_PM_CTRL register value against PCI_POSSIBLE_ERROR() instead
of invoking pci_device_is_present() separately.
While at it, drop a pair of unnecessary parens.
No expected functional impact.
Link: https://lore.kernel.org/r/1917095.PYKUYFuaPT@kreacher
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The runtime_d3cold flag is not needed any more, so drop it.
Link: https://lore.kernel.org/r/8077784.T7Z3S40VBb@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>
Calling pci_resume_bus() on the secondary bus from pci_power_up() as it is
done now is questionable, because it depends on the mandatory bridge
power-up delays that are only covered by the PCI bus type PM callbacks.
For this reason, move the subordinate bus resume to those callbacks too and
use the observation that if a bridge is being powered-up during resume from
system-wide suspend, it may be still desirable to runtime-resume its
subordinate bus after completing the system-wide transition (in case the
resume of the devices on that bus is skipped during it).
Link: https://lore.kernel.org/r/3190097.aeNJFYEL58@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>
Currently, endpoint devices may not be powered up entirely during runtime
resume that follows a D3hot -> D0 transition of the parent bridge.
Namely, even if the power state of an endpoint device, as indicated by its
PCI_PM_CTRL register, is D0 after powering up its parent bridge, it may be
still necessary to bring its ACPI companion into D0 and that should be done
before accessing it. However, the current code assumes that reading the
PCI_PM_CTRL register is sufficient to establish the endpoint device's power
state, which may lead to problems.
Address that by forcing a power-up of all PCI devices, including the
platform firmware part of it, during runtime resume.
Link: https://lore.kernel.org/linux-pm/11967527.O9o76ZdvQC@kreacher
Fixes: 5775b843a6 ("PCI: Restore config space on runtime resume despite being unbound")
Link: https://lore.kernel.org/r/2652115.mvXUDI8C0e@kreacher
Reported-by: Abhishek Sahu <abhsahu@nvidia.com>
Tested-by: Abhishek Sahu <abhsahu@nvidia.com>
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>
pci_restore_standard_config() was defined under CONFIG_PM but called only
by pci_pm_resume() (defined under CONFIG_SUSPEND) and pci_pm_restore()
(defined under CONFIG_HIBERNATE_CALLBACKS). A configuration with only
CONFIG_PM leads to a warning:
drivers/pci/pci-driver.c:533:12: error: ‘pci_restore_standard_config’ defined but not used [-Werror=unused-function]
CONFIG_PM_SLEEP depends on CONFIG_SUSPEND and CONFIG_HIBERNATE_CALLBACKS,
so define pci_restore_standard_config() under that instead.
Link: https://lore.kernel.org/r/20220420141135.444820-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
The Qualcomm SC8180X platform was piggy-backing on the SM8250
qcom_pcie_cfg, but SC8180X doesn't have the ddrss_sf_tbu clock, so
it now fails to probe due to the missing clock.
Give SC8180X its own qcom_pcie_cfg, without the ddrss_sf_tbu flag set.
Fixes: 0614f98bbb ("PCI: qcom: Add ddrss_sf_tbu flag")
Link: https://lore.kernel.org/r/20220331013415.592748-1-bjorn.andersson@linaro.org
Tested-by: Steev Klimaszewski <steev@kali.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Acked-by: Stanimir Varbanov <svarbanov@mm-sol.com>
Currently when the pci-hyperv driver finishes probing and initializing the
PCI device, it sets the PCI_COMMAND_MEMORY bit; later when the PCI device
is registered to the core PCI subsystem, the core PCI driver's BAR detection
and initialization code toggles the bit multiple times, and each toggling of
the bit causes the hypervisor to unmap/map the virtual BARs from/to the
physical BARs, which can be slow if the BAR sizes are huge, e.g., a Linux VM
with 14 GPU devices has to spend more than 3 minutes on BAR detection and
initialization, causing a long boot time.
Reduce the boot time by not setting the PCI_COMMAND_MEMORY bit when we
register the PCI device (there is no need to have it set in the first place).
The bit stays off till the PCI device driver calls pci_enable_device().
With this change, the boot time of such a 14-GPU VM is reduced by almost
3 minutes.
Link: https://lore.kernel.org/lkml/20220419220007.26550-1-decui@microsoft.com/
Tested-by: Boqun Feng (Microsoft) <boqun.feng@gmail.com>
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Acked-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Jake Oshins <jakeo@microsoft.com>
Link: https://lore.kernel.org/r/20220502074255.16901-1-decui@microsoft.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
In the multi-MSI case, hv_arch_irq_unmask() will only operate on the first
MSI of the N allocated. This is because only the first msi_desc is cached
and it is shared by all the MSIs of the multi-MSI block. This means that
hv_arch_irq_unmask() gets the correct address, but the wrong data (always
0).
This can break MSIs.
Lets assume MSI0 is vector 34 on CPU0, and MSI1 is vector 33 on CPU0.
hv_arch_irq_unmask() is called on MSI0. It uses a hypercall to configure
the MSI address and data (0) to vector 34 of CPU0. This is correct. Then
hv_arch_irq_unmask is called on MSI1. It uses another hypercall to
configure the MSI address and data (0) to vector 33 of CPU0. This is
wrong, and results in both MSI0 and MSI1 being routed to vector 33. Linux
will observe extra instances of MSI1 and no instances of MSI0 despite the
endpoint device behaving correctly.
For the multi-MSI case, we need unique address and data info for each MSI,
but the cached msi_desc does not provide that. However, that information
can be gotten from the int_desc cached in the chip_data by
compose_msi_msg(). Fix the multi-MSI case to use that cached information
instead. Since hv_set_msi_entry_from_desc() is no longer applicable,
remove it.
Signed-off-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Link: https://lore.kernel.org/r/1651068453-29588-1-git-send-email-quic_jhugo@quicinc.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
If a switch lacks ACS P2P Request Redirect, a device below the switch can
bypass the IOMMU and DMA directly to other devices below the switch, so
all the downstream devices must be in the same IOMMU group as the switch
itself.
The existing VFIO framework allows the portdrv driver to be bound to the
bridge while its downstream devices are assigned to user space. The
pci_dma_configure() marks the IOMMU group as containing only devices
with kernel drivers that manage DMA. Avoid this default behavior for the
portdrv driver in order for compatibility with the current VFIO usage.
We achieve this by setting ".driver_managed_dma = true" in pci_driver
structure. It is safe because the portdrv driver meets below criteria:
- This driver doesn't use DMA, as you can't find any related calls like
pci_set_master() or any kernel DMA API (dma_map_*() and etc.).
- It doesn't use MMIO as you can't find ioremap() or similar calls. It's
tolerant to userspace possibly also touching the same MMIO registers
via P2P DMA access.
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>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20220418005000.897664-7-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The current VFIO implementation allows pci-stub driver to be bound to
a PCI device with other devices in the same IOMMU group being assigned
to userspace. The pci-stub driver has no dependencies on DMA or the
IOVA mapping of the device, but it does prevent the user from having
direct access to the device, which is useful in some circumstances.
The pci_dma_configure() marks the iommu_group as containing only devices
with kernel drivers that manage DMA. For compatibility with the VFIO
usage, avoid this default behavior for the pci_stub. This allows the
pci_stub still able to be used by the admin to block driver binding after
applying the DMA ownership to VFIO.
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lore.kernel.org/r/20220418005000.897664-6-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
The devices on platform/amba/fsl-mc/PCI buses could be bound to drivers
with the device DMA managed by kernel drivers or user-space applications.
Unfortunately, multiple devices may be placed in the same IOMMU group
because they cannot be isolated from each other. The DMA on these devices
must either be entirely under kernel control or userspace control, never
a mixture. Otherwise the driver integrity is not guaranteed because they
could access each other through the peer-to-peer accesses which by-pass
the IOMMU protection.
This checks and sets the default DMA mode during driver binding, and
cleanups during driver unbinding. In the default mode, the device DMA is
managed by the device driver which handles DMA operations through the
kernel DMA APIs (see Documentation/core-api/dma-api.rst).
For cases where the devices are assigned for userspace control through the
userspace driver framework(i.e. VFIO), the drivers(for example, vfio_pci/
vfio_platfrom etc.) may set a new flag (driver_managed_dma) to skip this
default setting in the assumption that the drivers know what they are
doing with the device DMA.
Calling iommu_device_use_default_domain() before {of,acpi}_dma_configure
is currently a problem. As things stand, the IOMMU driver ignored the
initial iommu_probe_device() call when the device was added, since at
that point it had no fwspec yet. In this situation,
{of,acpi}_iommu_configure() are retriggering iommu_probe_device() after
the IOMMU driver has seen the firmware data via .of_xlate to learn that
it actually responsible for the given device. As the result, before
that gets fixed, iommu_use_default_domain() goes at the end, and calls
arch_teardown_dma_ops() if it fails.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Bjorn Helgaas <bhelgaas@google.com>
Cc: Stuart Yoder <stuyoder@gmail.com>
Cc: Laurentiu Tudor <laurentiu.tudor@nxp.com>
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20220418005000.897664-5-baolu.lu@linux.intel.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Variable retval is being assigned a value that is never read, the
variable is redundant and can be removed.
Cleans up clang scan build warning:
drivers/pci/controller/pci-versatile.c:37:10: warning: Although the value
stored to 'retval' is used in the enclosing expression, the value is never
actually read from 'retval' [deadcode.DeadStores]
Link: https://lore.kernel.org/r/20220418144416.86121-1-colin.i.king@gmail.com
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Dexuan wrote:
"[...] when we disable AccelNet, the host PCI VSP driver sends a
PCI_EJECT message first, and the channel callback may set
hpdev->state to hv_pcichild_ejecting on a different CPU. This can
cause hv_compose_msi_msg() to exit from the loop and 'return', and
the on-stack variable 'ctxt' is invalid. Now, if the response
message from the host arrives, the channel callback will try to
access the invalid 'ctxt' variable, and this may cause a crash."
Schematically:
Hyper-V sends PCI_EJECT msg
hv_pci_onchannelcallback()
state = hv_pcichild_ejecting
hv_compose_msi_msg()
alloc and init comp_pkt
state == hv_pcichild_ejecting
Hyper-V sends VM_PKT_COMP msg
hv_pci_onchannelcallback()
retrieve address of comp_pkt
'free' comp_pkt and return
comp_pkt->completion_func()
Dexuan also showed how the crash can be triggered after introducing
suitable delays in the driver code, thus validating the 'assumption'
that the host can still normally respond to the guest's compose_msi
request after the host has started to eject the PCI device.
Fix the synchronization by leveraging the requestor lock as follows:
- Before 'return'-ing in hv_compose_msi_msg(), remove the ID (while
holding the requestor lock) associated to the completion packet.
- Retrieve the address *and call ->completion_func() within a same
(requestor) critical section in hv_pci_onchannelcallback().
Reported-by: Wei Hu <weh@microsoft.com>
Reported-by: Dexuan Cui <decui@microsoft.com>
Suggested-by: Michael Kelley <mikelley@microsoft.com>
Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Link: https://lore.kernel.org/r/20220419122325.10078-7-parri.andrea@gmail.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
Currently, pointers to guest memory are passed to Hyper-V as transaction
IDs in hv_pci. In the face of errors or malicious behavior in Hyper-V,
hv_pci should not expose or trust the transaction IDs returned by
Hyper-V to be valid guest memory addresses. Instead, use small integers
generated by vmbus_requestor as request (transaction) IDs.
Suggested-by: Michael Kelley <mikelley@microsoft.com>
Signed-off-by: Andrea Parri (Microsoft) <parri.andrea@gmail.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Link: https://lore.kernel.org/r/20220419122325.10078-3-parri.andrea@gmail.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
If the allocation of multiple MSI vectors for multi-MSI fails in the core
PCI framework, the framework will retry the allocation as a single MSI
vector, assuming that meets the min_vecs specified by the requesting
driver.
Hyper-V advertises that multi-MSI is supported, but reuses the VECTOR
domain to implement that for x86. The VECTOR domain does not support
multi-MSI, so the alloc will always fail and fallback to a single MSI
allocation.
In short, Hyper-V advertises a capability it does not implement.
Hyper-V can support multi-MSI because it coordinates with the hypervisor
to map the MSIs in the IOMMU's interrupt remapper, which is something the
VECTOR domain does not have. Therefore the fix is simple - copy what the
x86 IOMMU drivers (AMD/Intel-IR) do by removing
X86_IRQ_ALLOC_CONTIGUOUS_VECTORS after calling the VECTOR domain's
pci_msi_prepare().
Fixes: 4daace0d8c ("PCI: hv: Add paravirtual PCI front-end for Microsoft Hyper-V VMs")
Signed-off-by: Jeffrey Hugo <quic_jhugo@quicinc.com>
Reviewed-by: Dexuan Cui <decui@microsoft.com>
Link: https://lore.kernel.org/r/1649856981-14649-1-git-send-email-quic_jhugo@quicinc.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
If the driver is configured as a module (after allowing this by changing
PCIE_MICROCHIP_HOST from bool to tristate) the missing semicolon makes the
compiler very unhappy. While there isn't a real problem as
MODULE_DEVICE_TABLE always evaluates to nothing for a built-in driver,
do it right for consistency with other drivers.
Link: https://lore.kernel.org/r/20220420065832.14173-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Daire McNamara <daire.mcnamara@microchip.com>
If DT supplies the 'slot-power-limit-milliwatt' property, program
the value in the Slot Power Limit in the Slot Capabilities register
and program the Root Port to send a Set_Slot_Power_Limit Message
when the Link transitions to DL_Up.
Link: https://lore.kernel.org/r/20220412094946.27069-5-pali@kernel.org
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Add function of_pci_get_slot_power_limit(), which parses the
'slot-power-limit-milliwatt' DT property, returning the value in
milliwatts and in format ready for the PCIe Slot Capabilities Register.
Link: https://lore.kernel.org/r/20220412094946.27069-4-pali@kernel.org
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Bjorn Helgaas <bhelgaas@google.com>
Use a helper to set driver_override to the reduce amount of duplicated
code. Make the driver_override field const char, because it is not
modified by the core and it matches other subsystems.
Reviewed-by: Andy Shevchenko <andy.shevchenko@gmail.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20220419113435.246203-6-krzysztof.kozlowski@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
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>
In 7b94b53db3 ("PCI/P2PDMA: Add Intel Sky Lake-E Root Ports B, C, D to
the whitelist"), Andrew Maier added Skylake-E 2031, 2032, and 2033 Root
Ports to the pci_p2pdma_whitelist[], so we assume P2PDMA between devices
below these ports works.
Previously we only checked the whitelist for a device at devfn 00.0 on the
root bus, which is often a "host bridge". But these Skylake Root Ports may
be at any devfn and there may be no "host bridge" device.
Generalize pci_host_bridge_dev() so we check the first device on the root
bus, whether it is devfn 00.0 or a PCIe Root Port, against the whitelist.
[bhelgaas: commit log, comment]
Link: https://lore.kernel.org/r/20220410105213.690-2-shlomop@pliops.com
Tested-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Shlomo Pongratz <shlomop@pliops.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: Andrew Maier <andrew.maier@eideticom.com>
If a Root Port on Elo i2 is put into D3cold and then back into D0, the
downstream device becomes permanently inaccessible, so add a bridge D3 DMI
quirk for that system.
This was exposed by 14858dcc3b ("PCI: Use pci_update_current_state() in
pci_enable_device_flags()"), but before that commit the Root Port in
question had never been put into D3cold for real due to a mismatch between
its power state retrieved from the PCI_PM_CTRL register (which was
accessible even though the platform firmware indicated that the port was in
D3cold) and the state of an ACPI power resource involved in its power
management.
BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=215715
Link: https://lore.kernel.org/r/11980172.O9o76ZdvQC@kreacher
Reported-by: Stefan Gottwald <gottwald@igel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Cc: stable@vger.kernel.org # v5.15+
The integrated MSI Receiver enable register is always initialized in
dw_pcie_setup_rc() which is also called in resume code path, so we
don't need to save/restore the enable register during suspend/resume.
Link: https://lore.kernel.org/r/20211226074910.2722-1-jszhang@kernel.org
Signed-off-by: Jisheng Zhang <jszhang@kernel.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Acked-by: Vidya Sagar <vidyas@nvidia.com>
When dma_mapping_error() returns error because of no enough memory,
but dw_pcie_host_init() returns success, which will mislead the callers.
Link: https://lore.kernel.org/r/30170911-0e2f-98ce-9266-70465b9073e5@huawei.com
Fixes: 07940c369a ("PCI: dwc: Fix MSI page leakage in suspend/resume")
Signed-off-by: Jianrong Zhang <zhangjianrong5@huawei.com>
Signed-off-by: Jiantao Zhang <water.zhangjiantao@huawei.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
The of_find_compatible_node() function returns a node pointer with
refcount incremented, We should use of_node_put() on it when done
Add the missing of_node_put() to release the refcount.
Link: https://lore.kernel.org/r/20220309091953.5630-1-linmq006@gmail.com
Fixes: 87e8657ba9 ("PCI: mediatek: Add new method to get shared pcie-cfg base address")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Miles Chen <miles.chen@mediatek.com>
Acked-by: Rob Herring <robh@kernel.org>
The ep->ob_region_map bitmap is a long and it has BITS_PER_LONG bits.
Link: https://lore.kernel.org/r/20220315065944.GB13572@kili
Fixes: cf590b0783 ("PCI: rockchip: Add EP driver for Rockchip PCIe controller")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
The ep->ob_region_map bitmap is a long and it has BITS_PER_LONG bits.
Link: https://lore.kernel.org/r/20220315065829.GA13572@kili
Fixes: 37dddf14f1 ("PCI: cadence: Add EndPoint Controller driver for Cadence PCIe controller")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Intel DG2 discrete graphics PCIe endpoints advertise L1 acceptable exit
latency to be < 1us even though they can actually tolerate unlimited exit
latencies just fine. Quirk the L1 acceptable exit latency for these
endpoints to be unlimited so ASPM L1 can be enabled.
[bhelgaas: use FIELD_GET/FIELD_PREP, wordsmith comment & commit log]
Link: https://lore.kernel.org/r/20220405093810.76613-1-mika.westerberg@linux.intel.com
Signed-off-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@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>
Fix the following build error:
drivers/pci/controller/pci-hyperv.c:769:13: error: ‘hv_set_msi_entry_from_desc’ defined but not used [-Werror=unused-function]
769 | static void hv_set_msi_entry_from_desc(union hv_msi_entry *msi_entry,
The arm64 implementation of hv_set_msi_entry_from_desc() is not used after
d06957d7a6 ("PCI: hv: Avoid the retarget interrupt hypercall in
irq_unmask() on ARM64"), so remove it.
Fixes: d06957d7a6 ("PCI: hv: Avoid the retarget interrupt hypercall in irq_unmask() on ARM64")
Link: https://lore.kernel.org/r/20220317085130.36388-1-yuehaibing@huawei.com
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Acked-by: Boqun Feng <boqun.feng@gmail.com>
PCI pass-thru devices in a Hyper-V VM are represented as a VMBus
device and as a PCI device. The coherence of the VMbus device is
set based on the VMbus node in ACPI, but the PCI device has no
ACPI node and defaults to not hardware coherent. This results
in extra software coherence management overhead on ARM64 when
devices are hardware coherent.
Fix this by setting up the PCI host bus so that normal
PCI mechanisms will propagate the coherence of the VMbus
device to the PCI device. There's no effect on x86/x64 where
devices are always hardware coherent.
Signed-off-by: Michael Kelley <mikelley@microsoft.com>
Acked-by: Boqun Feng <boqun.feng@gmail.com>
Acked-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/1648138492-2191-3-git-send-email-mikelley@microsoft.com
Signed-off-by: Wei Liu <wei.liu@kernel.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYIAB0WIQRTLbB6QfY48x44uB6AXGG7T9hjvgUCYkF9UwAKCRCAXGG7T9hj
vsXpAPwKXI4WIQcvnVCdULQfuXpA1TbD5XZuS9OuiN/OxWHbzAEA1VHWTmS+tpZ1
ptOyoGhAWhTGeplToobDSGz5qTXEPAI=
=FaKX
-----END PGP SIGNATURE-----
Merge tag 'for-linus-5.18-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip
Pull xen updates from Juergen Gross:
- A bunch of minor cleanups
- A fix for kexec in Xen dom0 when executed on a high cpu number
- A fix for resuming after suspend of a Xen guest with assigned PCI
devices
- A fix for a crash due to not disabled preemption when resuming as Xen
dom0
* tag 'for-linus-5.18-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: fix is_xen_pmu()
xen: don't hang when resuming PCI device
arch:x86:xen: Remove unnecessary assignment in xen_apic_read()
xen/grant-table: remove readonly parameter from functions
xen/grant-table: remove gnttab_*transfer*() functions
drivers/xen: use helper macro __ATTR_RW
x86/xen: Fix kerneldoc warning
xen: delay xen_hvm_init_time_ops() if kdump is boot on vcpu>=32
xen: use time_is_before_eq_jiffies() instead of open coding it
Here is the big set of char/misc and other small driver subsystem
updates for 5.18-rc1.
Included in here are merges from driver subsystems which contain:
- iio driver updates and new drivers
- fsi driver updates
- fpga driver updates
- habanalabs driver updates and support for new hardware
- soundwire driver updates and new drivers
- phy driver updates and new drivers
- coresight driver updates
- icc driver updates
Individual changes include:
- mei driver updates
- interconnect driver updates
- new PECI driver subsystem added
- vmci driver updates
- lots of tiny misc/char driver updates
There will be two merge conflicts with your tree, one in MAINTAINERS
which is obvious to fix up, and one in drivers/phy/freescale/Kconfig
which also should be easy to resolve.
All of these have been in linux-next for a while with no reported
problems.
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-----BEGIN PGP SIGNATURE-----
iG0EABECAC0WIQT0tgzFv3jCIUoxPcsxR9QN2y37KQUCYkG3fQ8cZ3JlZ0Brcm9h
aC5jb20ACgkQMUfUDdst+ykNEgCfaRG8CRxewDXOO4+GSeA3NGK+AIoAnR89donC
R4bgCjfg8BWIBcVVXg3/
=WWXC
-----END PGP SIGNATURE-----
Merge tag 'char-misc-5.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc
Pull char/misc and other driver updates from Greg KH:
"Here is the big set of char/misc and other small driver subsystem
updates for 5.18-rc1.
Included in here are merges from driver subsystems which contain:
- iio driver updates and new drivers
- fsi driver updates
- fpga driver updates
- habanalabs driver updates and support for new hardware
- soundwire driver updates and new drivers
- phy driver updates and new drivers
- coresight driver updates
- icc driver updates
Individual changes include:
- mei driver updates
- interconnect driver updates
- new PECI driver subsystem added
- vmci driver updates
- lots of tiny misc/char driver updates
All of these have been in linux-next for a while with no reported
problems"
* tag 'char-misc-5.18-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (556 commits)
firmware: google: Properly state IOMEM dependency
kgdbts: fix return value of __setup handler
firmware: sysfb: fix platform-device leak in error path
firmware: stratix10-svc: add missing callback parameter on RSU
arm64: dts: qcom: add non-secure domain property to fastrpc nodes
misc: fastrpc: Add dma handle implementation
misc: fastrpc: Add fdlist implementation
misc: fastrpc: Add helper function to get list and page
misc: fastrpc: Add support to secure memory map
dt-bindings: misc: add fastrpc domain vmid property
misc: fastrpc: check before loading process to the DSP
misc: fastrpc: add secure domain support
dt-bindings: misc: add property to support non-secure DSP
misc: fastrpc: Add support to get DSP capabilities
misc: fastrpc: add support for FASTRPC_IOCTL_MEM_MAP/UNMAP
misc: fastrpc: separate fastrpc device from channel context
dt-bindings: nvmem: brcm,nvram: add basic NVMEM cells
dt-bindings: nvmem: make "reg" property optional
nvmem: brcm_nvram: parse NVRAM content into NVMEM cells
nvmem: dt-bindings: Fix the error of dt-bindings check
...
