Hi, I have a Cape Verde GPU card in my pc and after git bisecting a situation where, at the time of the amdgpu module, the monitor would lose signal and the pc locked up so that it only responded to a magic sysreq boot (with no logging due to it happening before the root filesystem was writeable), the above commit was identified as the culprit. The GPU card is a Gigabyte R7 250 with pci-id 1002:682b (rev 87). With the 5.17.0 kernel and a kernel command line of: amdgpu.audio=1 amdgpu.si_support=1 the following dmesg output was received: [ 76.118991] [drm] amdgpu kernel modesetting enabled. [ 76.119100] amdgpu 0000:01:00.0: vgaarb: deactivate vga console [ 76.120004] Console: switching to colour dummy device 80x25 [ 76.120203] [drm] initializing kernel modesetting (VERDE 0x1002:0x682B 0x1458:0x22CA 0x87). [ 76.120211] amdgpu 0000:01:00.0: amdgpu: Trusted Memory Zone (TMZ) feature not supported [ 76.120235] [drm] register mmio base: 0xFE8C0000 [ 76.120238] [drm] register mmio size: 262144 [ 76.120245] [drm] add ip block number 0 <si_common> [ 76.120248] [drm] add ip block number 1 <gmc_v6_0> [ 76.120251] [drm] add ip block number 2 <si_ih> [ 76.120253] [drm] add ip block number 3 <gfx_v6_0> [ 76.120256] [drm] add ip block number 4 <si_dma> [ 76.120258] [drm] add ip block number 5 <si_dpm> [ 76.120261] [drm] add ip block number 6 <dce_v6_0> [ 76.120264] [drm] add ip block number 7 <uvd_v3_1> [ 76.163659] [drm] BIOS signature incorrect 5b 7 [ 76.163669] resource sanity check: requesting [mem 0x000c0000-0x000dffff], which spans more than PCI Bus 0000:00 [mem 0x000d0000-0x000dffff window] [ 76.163677] caller pci_map_rom+0x68/0x1b0 mapping multiple BARs [ 76.163691] amdgpu 0000:01:00.0: No more image in the PCI ROM [ 76.164996] amdgpu 0000:01:00.0: amdgpu: Fetched VBIOS from ROM BAR [ 76.165001] amdgpu: ATOM BIOS: xxx-xxx-xxx [ 76.165018] amdgpu 0000:01:00.0: amdgpu: PCIE atomic ops is not supported [ 76.165270] [drm] vm size is 64 GB, 2 levels, block size is 10-bit, fragment size is 9-bit [ 76.349679] amdgpu 0000:01:00.0: amdgpu: VRAM: 2048M 0x000000F400000000 - 0x000000F47FFFFFFF (2048M used) [ 76.349716] amdgpu 0000:01:00.0: amdgpu: GART: 1024M 0x000000FF00000000 - 0x000000FF3FFFFFFF [ 76.349753] [drm] Detected VRAM RAM=2048M, BAR=256M [ 76.349764] [drm] RAM width 128bits DDR3 [ 76.349940] [drm] amdgpu: 2048M of VRAM memory ready [ 76.349953] [drm] amdgpu: 3072M of GTT memory ready. [ 76.349992] [drm] GART: num cpu pages 262144, num gpu pages 262144 [ 76.350506] amdgpu 0000:01:00.0: amdgpu: PCIE GART of 1024M enabled (table at 0x000000F400900000). [ 76.495343] [drm] Internal thermal controller with fan control [ 76.495391] [drm] amdgpu: dpm initialized [ 76.495637] [drm] AMDGPU Display Connectors [ 76.495647] [drm] Connector 0: [ 76.495655] [drm] HDMI-A-1 [ 76.495662] [drm] HPD1 [ 76.495668] [drm] DDC: 0x194c 0x194c 0x194d 0x194d 0x194e 0x194e 0x194f 0x194f [ 76.495685] [drm] Encoders: [ 76.495691] [drm] DFP1: INTERNAL_UNIPHY [ 76.495699] [drm] Connector 1: [ 76.495706] [drm] DVI-D-1 [ 76.495712] [drm] HPD2 [ 76.495718] [drm] DDC: 0x1950 0x1950 0x1951 0x1951 0x1952 0x1952 0x1953 0x1953 [ 76.495733] [drm] Encoders: [ 76.495739] [drm] DFP2: INTERNAL_UNIPHY [ 76.495746] [drm] Connector 2: [ 76.495753] [drm] VGA-1 [ 76.495758] [drm] DDC: 0x1970 0x1970 0x1971 0x1971 0x1972 0x1972 0x1973 0x1973 [ 76.495773] [drm] Encoders: [ 76.495779] [drm] CRT1: INTERNAL_KLDSCP_DAC1 [ 76.599604] [drm] Found UVD firmware Version: 64.0 Family ID: 13 [ 76.603443] [drm] PCIE gen 2 link speeds already enabled [ 77.149564] [drm] UVD initialized successfully. [ 77.149578] amdgpu 0000:01:00.0: amdgpu: SE 1, SH per SE 2, CU per SH 5, active_cu_number 8 [ 77.456492] RTL8211B Gigabit Ethernet r8169-0-300:00: attached PHY driver (mii_bus:phy_addr=r8169-0-300:00, irq=MAC) [ 77.486245] [drm] Initialized amdgpu 3.44.0 20150101 for 0000:01:00.0 on minor 0 [ 77.521555] r8169 0000:03:00.0 eth0: Link is Down [ 77.547158] fbcon: amdgpudrmfb (fb0) is primary device [ 77.591226] Console: switching to colour frame buffer device 240x67 [ 77.600296] amdgpu 0000:01:00.0: [drm] fb0: amdgpudrmfb frame buffer device I can supply extra details but found no logging from the sessions that experienced the lock-up. Regards, Arthur Marsh.
