Bug 100398

Summary: [agd5f] New errors when tonga card powers up
Product: DRI Reporter: Mike Lothian <mike>
Component: DRM/AMDgpuAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium CC: mike
Version: DRI git   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:

Description Mike Lothian 2017-03-26 01:09:01 UTC
I'm seeing the below messages on agd5f's drm-next-4.12-wip branch when the Tonga discret card powers up - this doesn't seem to prevent rendering but I'm not sure what each of the rings are

Possibly related to the recent KIQ changes 

[17689.424825] [drm] PCIE GART of 4096M enabled (table at 0x0000000000040000).
[17689.473369] [drm] ring test on 0 succeeded in 11 usecs
[17689.473531] [drm] ring test on 9 succeeded in 20 usecs
[17689.473646] [drm] ring test on 1 succeeded in 19 usecs
[17689.676500] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 2 test failed (scratch(0xC040)=0xCAFEDEAD)
[17689.825919] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 3 test failed (scratch(0xC040)=0xCAFEDEAD)
[17689.974629] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 4 test failed (scratch(0xC040)=0xCAFEDEAD)
[17690.127396] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 5 test failed (scratch(0xC040)=0xCAFEDEAD)
[17690.289456] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 6 test failed (scratch(0xC040)=0xCAFEDEAD)
[17690.439040] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 7 test failed (scratch(0xC040)=0xCAFEDEAD)
[17690.587548] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 8 test failed (scratch(0xC040)=0xCAFEDEAD)
[17690.587553] [drm:amdgpu_resume] *ERROR* resume of IP block <gfx_v8_0> failed -22
[17690.587556] [drm:amdgpu_device_resume] *ERROR* amdgpu_resume failed (-22).
[17690.595905] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
[17690.595935] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
[17696.119220] amdgpu 0000:01:00.0: GPU pci config reset
Comment 1 Alex Deucher 2017-03-27 22:05:21 UTC
Is this still an issue with the latest drm-next-4.12-wip branch?  I just a squashed in a fix for this.
Comment 2 Mike Lothian 2017-03-27 23:21:31 UTC
I'm still seeing issues with the latest code

[    9.633880] [drm] PCIE GART of 4096M enabled (table at 0x0000000000040000).
[    9.683789] [drm] ring test on 0 succeeded in 11 usecs
[    9.685580] [drm] ring test on 9 succeeded in 20 usecs
[    9.687286] [drm] ring test on 1 succeeded in 24 usecs
[    9.689011] [drm] ring test on 2 succeeded in 71 usecs
[    9.690736] [drm] ring test on 3 succeeded in 58 usecs
[    9.692426] [drm] ring test on 4 succeeded in 52 usecs
[    9.694053] [drm] ring test on 5 succeeded in 65 usecs
[    9.904171] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 6 test failed (scratch(0xC040)=0xCAFEDEAD)
[    9.905938] [drm] ring test on 7 succeeded in 72 usecs
[   10.116166] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 8 test failed (scratch(0xC040)=0xCAFEDEAD)
[   10.117710] [drm:amdgpu_resume] *ERROR* resume of IP block <gfx_v8_0> failed -22
[   10.119334] [drm:amdgpu_device_resume] *ERROR* amdgpu_resume failed (-22).
[   10.129232] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
[   10.415571] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
[   15.897045] IPv6: ADDRCONF(NETDEV_UP): wlan0: link is not ready
[   16.340635] amdgpu 0000:01:00.0: GPU pci config reset
[   21.099143] wlan0: authenticate with 78:3e:53:c0:2e:03
[   21.142404] wlan0: send auth to 78:3e:53:c0:2e:03 (try 1/3)
[   21.143010] wlan0: authenticated
[   21.145256] wlan0: associate with 78:3e:53:c0:2e:03 (try 1/3)
[   21.146259] wlan0: RX AssocResp from 78:3e:53:c0:2e:03 (capab=0x11 status=0 aid=3)
[   21.149536] wlan0: associated
[   21.149559] IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
[   21.152043] ath: EEPROM regdomain: 0x833a
[   21.152048] ath: EEPROM indicates we should expect a country code
[   21.152050] ath: doing EEPROM country->regdmn map search
[   21.152053] ath: country maps to regdmn code: 0x37
[   21.152055] ath: Country alpha2 being used: GB
[   21.152057] ath: Regpair used: 0x37
[   21.152059] ath: regdomain 0x833a dynamically updated by country IE
[   21.193897] wlan0: Limiting TX power to 23 (23 - 0) dBm as advertised by 78:3e:53:c0:2e:03
[   45.275130] [drm] Atomic update on pipe (A) took 210 us, max time under evasion is 100 us
[   72.704244] [drm] PCIE GART of 4096M enabled (table at 0x0000000000040000).
[   72.752795] [drm] ring test on 0 succeeded in 11 usecs
[   72.753031] [drm] ring test on 9 succeeded in 23 usecs
[   72.753141] [drm] ring test on 1 succeeded in 23 usecs
[   72.967686] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 2 test failed (scratch(0xC040)=0xCAFEDEAD)
[   73.183511] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 3 test failed (scratch(0xC040)=0xCAFEDEAD)
[   73.398194] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 4 test failed (scratch(0xC040)=0xCAFEDEAD)
[   73.611767] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 5 test failed (scratch(0xC040)=0xCAFEDEAD)
[   73.829620] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 6 test failed (scratch(0xC040)=0xCAFEDEAD)
[   73.988727] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 7 test failed (scratch(0xC040)=0xCAFEDEAD)
[   74.147664] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 8 test failed (scratch(0xC040)=0xCAFEDEAD)
[   74.147668] [drm:amdgpu_resume] *ERROR* resume of IP block <gfx_v8_0> failed -22
[   74.147670] [drm:amdgpu_device_resume] *ERROR* amdgpu_resume failed (-22).
[   74.155268] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
[   74.155289] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
Comment 3 Alex Deucher 2017-03-29 22:14:02 UTC
Can you try the latest drm-next-4.12-wip branch?
Comment 4 Mike Lothian 2017-03-29 23:31:59 UTC
I still get errors but they're slightly different this time:

[    9.641782] [drm] PCIE GART of 4096M enabled (table at 0x0000000000040000).
[    9.691719] [drm] ring test on 0 succeeded in 11 usecs
[    9.693744] [drm] ring test on 9 succeeded in 3 usecs
[    9.695366] [drm] ring test on 1 succeeded in 44 usecs
[    9.697032] [drm] ring test on 2 succeeded in 30 usecs
[    9.698691] [drm] ring test on 3 succeeded in 37 usecs
[    9.700281] [drm] ring test on 4 succeeded in 45 usecs
[    9.701869] [drm] ring test on 5 succeeded in 23 usecs
[    9.703452] [drm] ring test on 6 succeeded in 51 usecs
[    9.704944] [drm] ring test on 7 succeeded in 29 usecs
[    9.706528] [drm] ring test on 8 succeeded in 31 usecs
[    9.707917] [drm] ring test on 10 succeeded in 5 usecs
[    9.709369] [drm] ring test on 11 succeeded in 6 usecs
[    9.759220] [drm] ring test on 12 succeeded in 1 usecs
[    9.760550] [drm] UVD initialized successfully.
[    9.974226] [drm] ring test on 13 succeeded in 0 usecs
[    9.975629] [drm] ring test on 14 succeeded in 2 usecs
[    9.977047] [drm] ring test on 15 succeeded in 3 usecs
[    9.978340] [drm] VCE initialized successfully.
[   15.836977] IPv6: ADDRCONF(NETDEV_UP): wlan0: link is not ready
[   16.291409] amdgpu 0000:01:00.0: GPU pci config reset
[   21.034387] wlan0: authenticate with 78:3e:53:c0:2e:03
[   21.077809] wlan0: send auth to 78:3e:53:c0:2e:03 (try 1/3)
[   21.078758] wlan0: authenticated
[   21.080481] wlan0: associate with 78:3e:53:c0:2e:03 (try 1/3)
[   21.083555] wlan0: RX AssocResp from 78:3e:53:c0:2e:03 (capab=0x11 status=0 aid=3)
[   21.086443] wlan0: associated
[   21.086455] IPv6: ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
[   21.088115] ath: EEPROM regdomain: 0x833a
[   21.088116] ath: EEPROM indicates we should expect a country code
[   21.088118] ath: doing EEPROM country->regdmn map search
[   21.088119] ath: country maps to regdmn code: 0x37
[   21.088120] ath: Country alpha2 being used: GB
[   21.088121] ath: Regpair used: 0x37
[   21.088122] ath: regdomain 0x833a dynamically updated by country IE
[   21.241486] wlan0: Limiting TX power to 23 (23 - 0) dBm as advertised by 78:3e:53:c0:2e:03
[   40.917223] [drm] PCIE GART of 4096M enabled (table at 0x0000000000040000).
[   40.966588] [drm] ring test on 0 succeeded in 11 usecs
[   40.966966] [drm] ring test on 9 succeeded in 1 usecs
[   41.181390] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 1 test failed (scratch(0xC040)=0xCAFEDEAD)
[   41.340367] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 2 test failed (scratch(0xC040)=0xCAFEDEAD)
[   41.499547] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 3 test failed (scratch(0xC040)=0xCAFEDEAD)
[   41.657396] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 4 test failed (scratch(0xC040)=0xCAFEDEAD)
[   41.816496] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 5 test failed (scratch(0xC040)=0xCAFEDEAD)
[   41.976375] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 6 test failed (scratch(0xC040)=0xCAFEDEAD)
[   42.139328] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 7 test failed (scratch(0xC040)=0xCAFEDEAD)
[   42.296842] [drm:gfx_v8_0_ring_test_ring] *ERROR* amdgpu: ring 8 test failed (scratch(0xC040)=0xCAFEDEAD)
[   42.296845] [drm:amdgpu_resume] *ERROR* resume of IP block <gfx_v8_0> failed -22
[   42.296847] [drm:amdgpu_device_resume] *ERROR* amdgpu_resume failed (-22).
[   42.304956] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
[   42.304974] [drm:amdgpu_fill_buffer] *ERROR* Trying to clear memory with ring turned off.
Comment 5 Alex Deucher 2017-03-31 18:51:16 UTC
Please try again with drm-next-4.12-wip
Comment 6 Mike Lothian 2017-03-31 20:55:09 UTC
That fixes it thanks

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.