On 10/24/2024 11:19 AM, Victor Zhao wrote: > during device init, under sriov, pci_restore_state happens after > fullaccess released, and it can have race condition with mmio protection > enable from host side. > > Since msix was toggled during pci_restore_state, if mmio protection > happens during this time, guest side msix will not be properly > programmed and leading to missing interrupts. > > So skip pci_restore_state during device init. > > Signed-off-by: Victor Zhao <Victor.Zhao@xxxxxxx> > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c > index 6c0ff1c2ae4c..52803cd91ef5 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c > @@ -4524,7 +4524,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, > dev_err(adev->dev, "amdgpu_pmu_init failed\n"); > > /* Have stored pci confspace at hand for restore in sudden PCI error */ > - if (amdgpu_device_cache_pci_state(adev->pdev)) > + if (!amdgpu_sriov_vf(adev) && amdgpu_device_cache_pci_state(adev->pdev)) This also prevents caching the state. If the intention is that way, put the check inside amdgpu_device_cache_pci_state to make it explicit that VFs avoid caching config space. Thanks, Lijo > pci_restore_state(pdev); > > /* if we have > 1 VGA cards, then disable the amdgpu VGA resources */
