[Mesa-dev] Radeon r600_ring_test/evergreen_resume errors

[Alex Deucher]

On Fri, Feb 21, 2014 at 12:22 PM, Dorrington, Albert
<albert.dorrington at lmco.com> wrote:
> I have started getting the errors shown below while experimenting with the
> openCL code:
>
>
>
> I am using a recent pull of Mesa (a 10.1-devel trunk pull, about a week old
> or so I think) and have updated to libdrm 2.4.52
>
> I'm stuck using kernel 3.10.0-54.0.1.el7.x86_64 for RHEL7.
>
>
>
> I have seen posts indicating similar problems with DPM enabled - but I am
> fairly certain that there is no DPM support in this RHEL7 kernel.
>

There is no dpm support in that kernel.

>
>
> The main issue is that when this happens, the kernel crashes and I end up
> having to reboot.
>
>
>
> I'm stumped at this moment, does anyone have any suggestions on what I
> should look into?
>
>

You are seeing a GPU hang and the driver attempts to reset it which
doesn't always work.  Probably a problem in the OpenGL or OpenCL
driver in mesa.

Alex

>
>
>
> Feb 21 11:25:03 bubba kernel: [ 5018.920026] radeon 0000:07:00.0: GPU lockup
> CP stall for more than 10000msec
>
> Feb 21 11:25:03 bubba kernel: [ 5018.920033] radeon 0000:07:00.0: GPU lockup
> (waiting for 0x000000000000159e last fence id 0x000000000000159d)
>
> Feb 21 11:25:03 bubba kernel: [ 5018.920045] [drm] Disabling audio 0 support
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927101] radeon 0000:07:00.0: Saved 55
> dwords of commands on ring 0.
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927113] radeon 0000:07:00.0: GPU
> softreset: 0x00000009
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927116] radeon 0000:07:00.0:
> GRBM_STATUS               = 0xB0433828
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927119] radeon 0000:07:00.0:
> GRBM_STATUS_SE0           = 0x08000007
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927122] radeon 0000:07:00.0:
> GRBM_STATUS_SE1           = 0x00000007
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927125] radeon 0000:07:00.0:
> SRBM_STATUS               = 0x200000C0
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927127] radeon 0000:07:00.0:
> SRBM_STATUS2              = 0x00000000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927130] radeon 0000:07:00.0:
> R_008674_CP_STALLED_STAT1 = 0x00000000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927133] radeon 0000:07:00.0:
> R_008678_CP_STALLED_STAT2 = 0x400C0000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927136] radeon 0000:07:00.0:
> R_00867C_CP_BUSY_STAT     = 0x00050000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927138] radeon 0000:07:00.0:
> R_008680_CP_STAT          = 0x80268643
>
> Feb 21 11:25:03 bubba kernel: [ 5018.927141] radeon 0000:07:00.0:
> R_00D034_DMA_STATUS_REG   = 0x44C83D57
>
> Feb 21 11:25:03 bubba kernel: [ 5018.940285] radeon 0000:07:00.0:
> GRBM_SOFT_RESET=0x00007F6B
>
> Feb 21 11:25:03 bubba kernel: [ 5018.940340] radeon 0000:07:00.0:
> SRBM_SOFT_RESET=0x00000100
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941497] radeon 0000:07:00.0:
> GRBM_STATUS               = 0x00003828
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941500] radeon 0000:07:00.0:
> GRBM_STATUS_SE0           = 0x00000007
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941502] radeon 0000:07:00.0:
> GRBM_STATUS_SE1           = 0x00000007
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941505] radeon 0000:07:00.0:
> SRBM_STATUS               = 0x200000C0
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941508] radeon 0000:07:00.0:
> SRBM_STATUS2              = 0x00000000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941511] radeon 0000:07:00.0:
> R_008674_CP_STALLED_STAT1 = 0x00000000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941513] radeon 0000:07:00.0:
> R_008678_CP_STALLED_STAT2 = 0x00000000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941516] radeon 0000:07:00.0:
> R_00867C_CP_BUSY_STAT     = 0x00000000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941519] radeon 0000:07:00.0:
> R_008680_CP_STAT          = 0x00000000
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941521] radeon 0000:07:00.0:
> R_00D034_DMA_STATUS_REG   = 0x44C83D57
>
> Feb 21 11:25:03 bubba kernel: [ 5018.941530] radeon 0000:07:00.0: GPU reset
> succeeded, trying to resume
>
> Feb 21 11:25:03 bubba kernel: [ 5018.963883] [drm] PCIE GART of 1024M
> enabled (table at 0x0000000000273000).
>
> Feb 21 11:25:03 bubba kernel: [ 5018.963990] radeon 0000:07:00.0: WB enabled
>
> Feb 21 11:25:03 bubba kernel: [ 5018.963995] radeon 0000:07:00.0: fence
> driver on ring 0 use gpu addr 0x0000000040000c00 and cpu addr
> 0xffff880126601c00
>
> Feb 21 11:25:03 bubba kernel: [ 5018.963998] radeon 0000:07:00.0: fence
> driver on ring 3 use gpu addr 0x0000000040000c0c and cpu addr
> 0xffff880126601c0c
>
> Feb 21 11:25:03 bubba kernel: [ 5018.965558] radeon 0000:07:00.0: fence
> driver on ring 5 use gpu addr 0x0000000000072118 and cpu addr
> 0xffffc90010a32118
>
> Feb 21 11:25:03 bubba kernel: [ 5019.179402] [drm:r600_ring_test] *ERROR*
> radeon: ring 0 test failed (scratch(0x8504)=0xCAFEDEAD)
>
> Feb 21 11:25:03 bubba kernel: [ 5019.179405] [drm:evergreen_resume] *ERROR*
> evergreen startup failed on resume
>
>
>
>
>
> Al Dorrington
>
> Software Engineer Sr
>
> Lockheed Martin, Mission Systems and Training
>
>
>
>
> _______________________________________________
> mesa-dev mailing list
> mesa-dev at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/mesa-dev
>