Random display freeze AMD TURKS (DRM 2.50.0 / 5.4.0-77-generic, LLVM 11.0.0)

Sat Sep 4 06:17:11 UTC 2021

I have some more info on it:

Sometimes I can go to another terminal screen, sometimes I can still 
login through ssh, sometimes the system fully locks up, but I can reboot 
with sysreq-alt-b.

The whole thing seems to happen random, but mostly after the system has 
been running one or two days.

OpenGL Renderer: AMD TURKS (DRM 2.50.0 / 5.4.0-77-generic, LLVM 12.0.0)

Last part of dmeg --syslog latest time it happened:

...

[21385.583783] audit: type=1400 audit(1630676128.271:208): 
apparmor="ALLOWED" operation="rename_src" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F6C7531393238303462336C616B2E746D70 
pid=19280 comm="soffice.bin" requested_mask="wrd" denied_mask="wrd" 
fsuid=1000 ouid=1000
[21536.024463] kauditd_printk_skb: 7 callbacks suppressed
[21536.024465] audit: type=1400 audit(1630676278.701:216): 
apparmor="ALLOWED" operation="open" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F2E7E6C6F636B2E54696D6573686565742045484D2076312E302E786C737823 
pid=19280 comm="soffice.bin" requested_mask="r" denied_mask="r" 
fsuid=1000 ouid=1000
[21536.029431] CIFS VFS: SMB signature verification returned error = -13
[21536.031847] audit: type=1400 audit(1630676278.709:217): 
apparmor="ALLOWED" operation="unlink" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F2E7E6C6F636B2E54696D6573686565742045484D2076312E302E786C737823 
pid=19280 comm="soffice.bin" requested_mask="d" denied_mask="d" 
fsuid=1000 ouid=1000
[25333.210567] hrtimer: interrupt took 25127 ns
[30072.138275] audit: type=1400 audit(1630684814.956:218): 
apparmor="DENIED" operation="capable" profile="/usr/sbin/cups-browsed" 
pid=33554 comm="cups-browsed" capability=23  capname="sys_nice"
[30859.256212] radeon 0000:01:00.0: ring 3 stalled for more than 10224msec
[30859.256220] radeon 0000:01:00.0: GPU lockup (current fence id 
0x0000000000002058 last fence id 0x000000000000205a on ring 3)
[30859.366220] radeon 0000:01:00.0: Saved 1746 dwords of commands on ring 0.
[30859.366234] radeon 0000:01:00.0: GPU softreset: 0x0000008C
[30859.366235] radeon 0000:01:00.0:   GRBM_STATUS               = 0xA0003828
[30859.366236] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[30859.366237] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[30859.366238] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200440C0
[30859.366239] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[30859.366240] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[30859.366241] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00010000
[30859.366242] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00000002
[30859.366243] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x80010243
[30859.366244] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44483146
[30859.379995] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00004001
[30859.380047] radeon 0000:01:00.0: SRBM_SOFT_RESET=0x00108100
[30859.381206] radeon 0000:01:00.0:   GRBM_STATUS               = 0x00003828
[30859.381207] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[30859.381208] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[30859.381209] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200400C0
[30859.381210] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[30859.381211] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[30859.381212] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00000000
[30859.381213] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00000000
[30859.381214] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x00000000
[30859.381215] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44C83D57
[30859.381227] radeon 0000:01:00.0: GPU reset succeeded, trying to resume
[30859.403380] [drm] enabling PCIE gen 2 link speeds, disable with 
radeon.pcie_gen2=0
[30859.407610] [drm] PCIE GART of 1024M enabled (table at 
0x0000000000162000).
[30859.407704] radeon 0000:01:00.0: WB enabled
[30859.407706] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 
0x00000000e0000c00 and cpu addr 0x0000000068b59d95
[30859.407706] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 
0x00000000e0000c0c and cpu addr 0x000000005a048e55
[30859.408486] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 
0x0000000000072118 and cpu addr 0x000000003f7c7eb9
[30859.424829] [drm] ring test on 0 succeeded in 3 usecs
[30859.424840] [drm] ring test on 3 succeeded in 7 usecs
[30859.600585] [drm] ring test on 5 succeeded in 2 usecs
[30859.600594] [drm] UVD initialized successfully.
[30860.760281] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait 
timed out.
[30860.760342] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: 
failed testing IB on GFX ring (-110).

Here the graphical (desktop) display was fully locked, but I was able to 
access a terminal screen if I remember correctly and reboot the system 
'normally'. After reboot the system always runs okay.

Earlier in the log:

...

[ 3053.103898] audit: type=1400 audit(1630657795.570:146): 
apparmor="ALLOWED" operation="file_lock" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F54696D6573686565742045484D2076312E302E786C7378 
pid=6030 comm="soffice.bin" requested_mask="wk" denied_mask="wk" 
fsuid=1000 ouid=1000
[ 3053.108517] audit: type=1400 audit(1630657795.578:147): 
apparmor="ALLOWED" operation="mknod" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F2E7E6C6F636B2E54696D6573686565742045484D2076312E302E786C737823 
pid=6030 comm="soffice.bin" requested_mask="c" denied_mask="c" 
fsuid=1000 ouid=1000
[ 3053.112153] audit: type=1400 audit(1630657795.582:148): 
apparmor="ALLOWED" operation="open" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F2E7E6C6F636B2E54696D6573686565742045484D2076312E302E786C737823 
pid=6030 comm="soffice.bin" requested_mask="wrc" denied_mask="wrc" 
fsuid=1000 ouid=1000
[ 3053.151664] CIFS VFS: SMB signature verification returned error = -13
[ 3053.303590] CIFS VFS: SMB signature verification returned error = -13
[ 3053.319941] CIFS VFS: SMB signature verification returned error = -13
[ 3053.334053] audit: type=1400 audit(1630657795.802:149): 
apparmor="ALLOWED" operation="rename_src" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F6C75363033303432646D7A362E746D70 
pid=6030 comm="soffice.bin" requested_mask="wrd" denied_mask="wrd" 
fsuid=1000 ouid=1000
[ 7476.284522] perf: interrupt took too long (2508 > 2500), lowering 
kernel.perf_event_max_sample_rate to 79500
[ 7930.584695] radeon 0000:01:00.0: ring 0 stalled for more than 10120msec
[ 7930.584703] radeon 0000:01:00.0: GPU lockup (current fence id 
0x000000000004351e last fence id 0x000000000004353d on ring 0)
[ 7931.096709] radeon 0000:01:00.0: ring 0 stalled for more than 10632msec
[ 7931.096717] radeon 0000:01:00.0: GPU lockup (current fence id 
0x000000000004351e last fence id 0x0000000000043542 on ring 0)
[ 7931.128660] radeon 0000:01:00.0: ring 3 stalled for more than 10240msec
[ 7931.128664] radeon 0000:01:00.0: GPU lockup (current fence id 
0x0000000000000b16 last fence id 0x0000000000000b18 on ring 3)
[ 7931.608731] radeon 0000:01:00.0: ring 0 stalled for more than 11144msec
[ 7931.608740] radeon 0000:01:00.0: GPU lockup (current fence id 
0x000000000004351e last fence id 0x0000000000043546 on ring 0)
[ 7931.644708] radeon 0000:01:00.0: ring 3 stalled for more than 10756msec
[ 7931.644712] radeon 0000:01:00.0: GPU lockup (current fence id 
0x0000000000000b16 last fence id 0x0000000000000b18 on ring 3)
[ 7932.120676] radeon 0000:01:00.0: ring 0 stalled for more than 11656msec
[ 7932.120680] radeon 0000:01:00.0: GPU lockup (current fence id 
0x000000000004351e last fence id 0x000000000004354c on ring 0)
[ 7932.128800] radeon 0000:01:00.0: failed to get a new IB (-35)
[ 7932.128825] [drm:radeon_cs_ioctl [radeon]] *ERROR* Failed to get ib !
[ 7932.249318] radeon 0000:01:00.0: Saved 1522 dwords of commands on ring 0.
[ 7932.249329] radeon 0000:01:00.0: GPU softreset: 0x0000000C
[ 7932.249330] radeon 0000:01:00.0:   GRBM_STATUS               = 0xA0003828
[ 7932.249331] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[ 7932.249332] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[ 7932.249333] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200000C0
[ 7932.249334] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[ 7932.249335] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[ 7932.249336] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00010002
[ 7932.249337] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00020186
[ 7932.249338] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x80038647
[ 7932.249339] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44483146
[ 7932.256904] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00004001
[ 7932.256955] radeon 0000:01:00.0: SRBM_SOFT_RESET=0x00100100
[ 7932.258100] radeon 0000:01:00.0:   GRBM_STATUS               = 0x00003828
[ 7932.258101] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[ 7932.258102] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[ 7932.258103] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200000C0
[ 7932.258104] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[ 7932.258105] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[ 7932.258106] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00000000
[ 7932.258107] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00000000
[ 7932.258108] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x00000000
[ 7932.258109] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44C83D57
[ 7932.258122] radeon 0000:01:00.0: GPU reset succeeded, trying to resume
[ 7932.280351] [drm] enabling PCIE gen 2 link speeds, disable with 
radeon.pcie_gen2=0
[ 7932.284667] [drm] PCIE GART of 1024M enabled (table at 
0x0000000000162000).
[ 7932.284775] radeon 0000:01:00.0: WB enabled
[ 7932.284777] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 
0x00000000e0000c00 and cpu addr 0x0000000068b59d95
[ 7932.284778] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 
0x00000000e0000c0c and cpu addr 0x000000005a048e55
[ 7932.285539] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 
0x0000000000072118 and cpu addr 0x000000003f7c7eb9
[ 7932.301773] [drm] ring test on 0 succeeded in 2 usecs
[ 7932.301783] [drm] ring test on 3 succeeded in 7 usecs
[ 7932.477565] [drm] ring test on 5 succeeded in 2 usecs
[ 7932.477573] [drm] UVD initialized successfully.
[ 7933.624806] [drm:r600_ib_test [radeon]] *ERROR* radeon: fence wait 
timed out.
[ 7933.624864] [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: 
failed testing IB on GFX ring (-110).
[ 7933.737625] radeon 0000:01:00.0: GPU softreset: 0x00000008
[ 7933.737627] radeon 0000:01:00.0:   GRBM_STATUS               = 0xA0003828
[ 7933.737628] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[ 7933.737629] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[ 7933.737630] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200000C0
[ 7933.737631] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[ 7933.737631] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[ 7933.737633] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00010100
[ 7933.737633] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00020182
[ 7933.737634] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x80038243
[ 7933.737636] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44C83D57
[ 7933.748431] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00004001
[ 7933.748483] radeon 0000:01:00.0: SRBM_SOFT_RESET=0x00000100
[ 7933.749628] radeon 0000:01:00.0:   GRBM_STATUS               = 0x00003828
[ 7933.749629] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[ 7933.749630] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[ 7933.749631] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200000C0
[ 7933.749632] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[ 7933.749633] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[ 7933.749644] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00000000
[ 7933.749645] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00000000
[ 7933.749646] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x00000000
[ 7933.749647] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44C83D57
[ 7933.749659] radeon 0000:01:00.0: GPU reset succeeded, trying to resume
[ 7933.771913] [drm] enabling PCIE gen 2 link speeds, disable with 
radeon.pcie_gen2=0
[ 7933.776106] [drm] PCIE GART of 1024M enabled (table at 
0x0000000000162000).
[ 7933.776199] radeon 0000:01:00.0: WB enabled
[ 7933.776201] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 
0x00000000e0000c00 and cpu addr 0x0000000068b59d95
[ 7933.776201] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 
0x00000000e0000c0c and cpu addr 0x000000005a048e55
[ 7933.776982] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 
0x0000000000072118 and cpu addr 0x000000003f7c7eb9
[ 7933.793238] [drm] ring test on 0 succeeded in 3 usecs
[ 7933.793249] [drm] ring test on 3 succeeded in 7 usecs
[ 7933.969032] [drm] ring test on 5 succeeded in 2 usecs
[ 7933.969041] [drm] UVD initialized successfully.
[ 7934.106544] [drm] ib test on ring 0 succeeded in 0 usecs
[ 7934.106582] [drm] ib test on ring 3 succeeded in 0 usecs
[ 7935.292759] [drm] ib test on ring 5 succeeded
[ 7970.036738] kauditd_printk_skb: 7 callbacks suppressed
[ 7970.036740] audit: type=1400 audit(1630662712.521:157): 
apparmor="ALLOWED" operation="open" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F2E7E6C6F636B2E54696D6573686565742045484D2076312E302E786C737823 
pid=6030 comm="soffice.bin" requested_mask="r" denied_mask="r" 
fsuid=1000 ouid=1000
[ 7970.040368] CIFS VFS: SMB signature verification returned error = -13
[ 7970.042470] audit: type=1400 audit(1630662712.525:158): 
apparmor="ALLOWED" operation="unlink" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F2E7E6C6F636B2E54696D6573686565742045484D2076312E302E786C737823 
pid=6030 comm="soffice.bin" requested_mask="d" denied_mask="d" 
fsuid=1000 ouid=1000
[ 7970.047342] audit: type=1400 audit(1630662712.529:159): 
apparmor="ALLOWED" operation="mknod" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F6C75363033303432646D7A672E746D70 
pid=6030 comm="soffice.bin" requested_mask="c" denied_mask="c" 
fsuid=1000 ouid=1000
[ 7970.051025] audit: type=1400 audit(1630662712.533:160): 
apparmor="ALLOWED" operation="open" profile="libreoffice-soffice" 
name=2F6E6574776F726B2F636F6D70616E792D6F6E2D6D6F6F6E2F437573746F6D657220646F63756D656E74732F456967656E20486F72656361204D616B656C6161722F6C75363033303432646D7A672E746D70 
pid=6030 comm="soffice.bin" requested_mask="wrc" denied_mask="wrc" 
fsuid=1000 ouid=1000

...

On 8/3/21 12:54 AM, Guus Ellenkamp wrote:
> My display freezes randomly on an Ubuntu 20.04 system with a Radeon 
> AMD Turks graphics card.
>
> Before the final freeze I often get warnings by the display suddenly 
> turning black and then turning on again.
>
> Not sure if it's the driver or the (cheap) graphics card. How can I 
> find out and is there any solution?
>
> Restarting the display manager does not have any effect.
>