PCI: Always enable ACS even if no ACS Capability

Some devices support ACS functionality even though they don't have a
spec-compliant ACS Capability; pci_enable_acs() has a quirk mechanism to
handle them.

We want to enable ACS whenever possible, but 52fbf5bdee ("PCI: Cache ACS
capability offset in device") inadvertently broke this by calling
pci_enable_acs() only if we find an ACS Capability.

This resulted in ACS not being enabled for these non-compliant devices,
which means devices can't be separated into different IOMMU groups, which
in turn means we may not be able to pass those devices through to VMs, as
reported by Boris V:

  https://lore.kernel.org/r/74aeea93-8a46-5f5a-343c-790d4c655da3@bstnet.org

Fixes: 52fbf5bdee ("PCI: Cache ACS capability offset in device")
Link: https://lore.kernel.org/r/20201028231545.4116866-1-rajatja@google.com
Reported-by: Boris V <borisvk@bstnet.org>
Signed-off-by: Rajat Jain <rajatja@google.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
This commit is contained in:
Rajat Jain 2020-10-28 16:15:45 -07:00 committed by Bjorn Helgaas
parent 3650b228f8
commit 462b58fb03

View File

@ -3516,8 +3516,13 @@ void pci_acs_init(struct pci_dev *dev)
{
dev->acs_cap = pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ACS);
if (dev->acs_cap)
pci_enable_acs(dev);
/*
* Attempt to enable ACS regardless of capability because some Root
* Ports (e.g. those quirked with *_intel_pch_acs_*) do not have
* the standard ACS capability but still support ACS via those
* quirks.
*/
pci_enable_acs(dev);
}
/**