We finally root caused this to a GPU configuration issue. On some harvest configurations, a driver needs to properly initialize some of the caches on the GPU for instances that are harvested (parts of the chip that are disabled due to silicon flaws). The necessary code to fix this up is too complex to add as a quirk. Bug: https://gitlab.freedesktop.org/drm/amd/issues/1015 Signed-off-by: Alex Deucher <alexander.deucher@xxxxxxx> --- drivers/pci/quirks.c | 24 +++++++++++++++++------- 1 file changed, 17 insertions(+), 7 deletions(-) diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 4937a088d7d8..6569dacbb48b 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -5074,18 +5074,28 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_SERVERWORKS, 0x0422, quirk_no_ext_tags); #ifdef CONFIG_PCI_ATS /* - * Some devices have a broken ATS implementation causing IOMMU stalls. - * Don't use ATS for those devices. + * Some devices require additional driver setup to enable ATS. + * Don't use ATS for those devices as ATS will be enabled before + * the driver has had a chance to load and properly configure + * the device. */ -static void quirk_no_ats(struct pci_dev *pdev) +static void quirk_amd_harvest_no_ats(struct pci_dev *pdev) { - pci_info(pdev, "disabling ATS (broken on this device)\n"); - pdev->ats_cap = 0; + switch (pdev->device) { + case 0x98e4: + case 0x6900: + pci_info(pdev, "disabling ATS\n"); + pdev->ats_cap = 0; + break; + default: + break; + } } /* AMD Stoney platform GPU */ -DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x98e4, quirk_no_ats); -DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x6900, quirk_no_ats); +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x98e4, quirk_amd_harvest_no_ats); +/* AMD Iceland dGPU */ +DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_ATI, 0x6900, quirk_amd_harvest_no_ats); #endif /* CONFIG_PCI_ATS */ /* Freescale PCIe doesn't support MSI in RC mode */ -- 2.24.1
