Created attachment 136980 [details]
lspci output

[  695.093933] radeon 0000:01:00.0: ring 0 stalled for more than 10150msec
[  695.100523] radeon 0000:01:00.0: GPU lockup (current fence id 0x000000000000befe last fence id 0x000000000000bf3e on ring 0)
[  695.100673] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100676] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100678] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100681] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100691] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100695] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100701] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100707] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100710] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100712] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100715] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100721] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100723] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100727] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100730] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100734] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100737] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100741] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100744] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100747] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100749] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100753] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100756] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100759] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.100762] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.100766] radeon 0000:01:00.0: failed to get a new IB (-35)
[  695.119228] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  695.968594] [drm:radeon_cs_ioctl] *ERROR* Failed to get ib !
[  696.004823] radeon 0000:01:00.0: Saved 2034 dwords of commands on ring 0.
[  696.004856] radeon 0000:01:00.0: GPU softreset: 0x00000008
[  696.004861] radeon 0000:01:00.0:   GRBM_STATUS               = 0xA0003828
[  696.004866] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[  696.004871] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[  696.004876] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200000C0
[  696.004880] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[  696.004885] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[  696.004889] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00004100
[  696.004894] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00020182
[  696.004899] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x80028243
[  696.004903] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44C83D57
[  696.054035] radeon 0000:01:00.0: GRBM_SOFT_RESET=0x00004001
[  696.054093] radeon 0000:01:00.0: SRBM_SOFT_RESET=0x00000100
[  696.055256] radeon 0000:01:00.0:   GRBM_STATUS               = 0x00003828
[  696.055260] radeon 0000:01:00.0:   GRBM_STATUS_SE0           = 0x00000007
[  696.055265] radeon 0000:01:00.0:   GRBM_STATUS_SE1           = 0x00000007
[  696.055269] radeon 0000:01:00.0:   SRBM_STATUS               = 0x200000C0
[  696.055274] radeon 0000:01:00.0:   SRBM_STATUS2              = 0x00000000
[  696.055278] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[  696.055282] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00000000
[  696.055286] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00000000
[  696.055290] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x00000000
[  696.055295] radeon 0000:01:00.0:   R_00D034_DMA_STATUS_REG   = 0x44C83D57
[  696.055332] radeon 0000:01:00.0: GPU reset succeeded, trying to resume
[  696.067624] [drm] PCIE gen 2 link speeds already enabled
[  696.129090] [drm] PCIE GART of 1024M enabled (table at 0x0000000000162000).
[  696.129270] radeon 0000:01:00.0: WB enabled
[  696.129277] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000020000c00 and cpu addr 0xffffffc4eaa7bc00
[  696.129283] radeon 0000:01:00.0: fence driver on ring 3 use gpu addr 0x0000000020000c0c and cpu addr 0xffffffc4eaa7bc0c
[  696.151289] radeon 0000:01:00.0: fence driver on ring 5 use gpu addr 0x0000000000072118 and cpu addr 0xffffff8018332118
[  696.168311] [drm] ring test on 0 succeeded in 2 usecs
[  696.168328] [drm] ring test on 3 succeeded in 7 usecs
[  696.345018] [drm] ring test on 5 succeeded in 2 usecs
[  696.345031] [drm] UVD initialized successfully.
[  696.379172] [drm] ib test on ring 0 succeeded in 0 usecs
[  696.379236] [drm] ib test on ring 3 succeeded in 0 usecs
[  696.530311] [drm] ib test on ring 5 succeeded


Sometimes when using qtcreator for a while a gpu lockup error occurred, on this basis, we guess that this gpulock error caused by qtcreator's welcome plugin, so we use a script to test the qtcreator's Welcome plugin as follows:

while true
do
for i in `seq 100`
do
qtcreator -test Welcome &
done
wait
done

The time of the error is random, and sometimes appear once a half hour, sometimes half a day, that is , not a fixed time. Currently we have tried a lot of ways to deal with cpu and gpu out of sync. E.g: dpm=0/pm_method=profile. Pipeline timeout. Use GTT for stream resources.
  


OS Details:
kernel 4.4 with 4.9drm.
xorg-server 1.18.3
xf86-video-ati 1:7.7
libdrm 2.4.65
mesa 11.2.0