Bug 60890

Summary: Radeon HD 5700: GPU lockup in World in Conflict
Product: Mesa Reporter: Stefan Dösinger <stefandoesinger>
Component: Drivers/Gallium/r600Assignee: Default DRI bug account <dri-devel>
Status: RESOLVED DUPLICATE QA Contact:
Severity: normal    
Priority: medium    
Version: git   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:

Description Stefan Dösinger 2013-02-15 10:24:10 UTC
World in Conflict (running in Wine) causes GPU lockups on my Radeon HD 5770. This is a regression caused by 974b482acaf62ced1e8981761a8bda252bd51fe1:

974b482acaf62ced1e8981761a8bda252bd51fe1 is the first bad commit
commit 974b482acaf62ced1e8981761a8bda252bd51fe1
Author: Jerome Glisse <jglisse@redhat.com>
Date:   Fri Feb 8 16:02:32 2013 -0500

    r600g: fix lockup when hyperz & alpha test are enabled together. v3
    
:040000 040000 dda64e4a96123c0d2710354ce341b6e37a77e222 5e753aadebef5a00d66753b3b1c216d4ea483567 M      src

The corresponding dmesg lines are:
[30865.977733] radeon 0000:01:00.0: GPU lockup CP stall for more than 10000msec
[30865.977742] radeon 0000:01:00.0: GPU lockup (waiting for 0x00000000007f1db2 last fence id 0x00000000007f1dad)
[30865.978802] radeon 0000:01:00.0: Saved 759 dwords of commands on ring 0.
[30865.978809] radeon 0000:01:00.0: GPU softreset 
[30865.978815] radeon 0000:01:00.0:   GRBM_STATUS=0xF4002828
[30865.978820] radeon 0000:01:00.0:   GRBM_STATUS_SE0=0xC0000005
[30865.978825] radeon 0000:01:00.0:   GRBM_STATUS_SE1=0x00000007
[30865.978830] radeon 0000:01:00.0:   SRBM_STATUS=0x200000C0
[30865.978834] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[30865.978840] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x40040000
[30865.978845] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00048002
[30865.978851] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x80268643
[30865.983269] radeon 0000:01:00.0:   GRBM_SOFT_RESET=0x00007F6B
[30865.983375] radeon 0000:01:00.0:   GRBM_STATUS=0x00003828
[30865.983379] radeon 0000:01:00.0:   GRBM_STATUS_SE0=0x00000007
[30865.983382] radeon 0000:01:00.0:   GRBM_STATUS_SE1=0x00000007
[30865.983386] radeon 0000:01:00.0:   SRBM_STATUS=0x200000C0
[30865.983390] radeon 0000:01:00.0:   R_008674_CP_STALLED_STAT1 = 0x00000000
[30865.983394] radeon 0000:01:00.0:   R_008678_CP_STALLED_STAT2 = 0x00000000
[30865.983397] radeon 0000:01:00.0:   R_00867C_CP_BUSY_STAT     = 0x00000000
[30865.983401] radeon 0000:01:00.0:   R_008680_CP_STAT          = 0x00000000
[30866.000901] radeon 0000:01:00.0: GPU reset succeeded, trying to resume
[30866.087667] [drm] probing gen 2 caps for device 8086:d138 = 2/0
[30866.087672] [drm] PCIE gen 2 link speeds already enabled
[30866.089405] [drm] PCIE GART of 512M enabled (table at 0x0000000000040000).
[30866.089528] radeon 0000:01:00.0: WB enabled
[30866.089531] radeon 0000:01:00.0: fence driver on ring 0 use gpu addr 0x0000000040000c00 and cpu addr 0xffff8802352efc00
[30866.106118] [drm] ring test on 0 succeeded in 0 usecs
[30866.112724] [drm] ib test on ring 0 succeeded in 0 usecs

The game is executing the builtin benchmark in windowed mode, resolution 800x600, medium presets.
Comment 1 Jerome Glisse 2013-02-15 15:49:25 UTC
Please provide an apitrace capture trimed to the lockup issue
Comment 2 Jerome Glisse 2013-02-20 21:33:44 UTC

*** This bug has been marked as a duplicate of bug 60969 ***
Comment 3 Jerome Glisse 2013-02-20 21:34:35 UTC
Stefan patch for bug 60969 should fix it for you to. Please test.
Comment 4 Stefan Dösinger 2013-03-01 15:01:02 UTC
The bug is fixed.

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.