| Summary: | [drm:r100_ring_test] *ERROR* radeon: ring test failed - since linux-3.17_rc1 on RS690/RS740 [Radeon 2100] | ||
|---|---|---|---|
| Product: | DRI | Reporter: | jospezial <jospezial> |
| Component: | DRM/Radeon | Assignee: | Default DRI bug account <dri-devel> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | medium | ||
| Version: | DRI git | ||
| Hardware: | Other | ||
| OS: | Linux (All) | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
| Attachments: | |||
|
Description
jospezial
2014-09-17 13:42:06 UTC
Created attachment 106432 [details] [review] Test for reverting the patch in question. It's unlikely that the patch you mentioned causes such problems, but the patch directly before this one could. The attached patch (on top of Alex drm-fixes-3.17 branch) reverts this patch and adds a few extra debug log lines. So please test and report back with a new dmesg log. Created attachment 106469 [details] dmesg for linux-3.17_rc5 with 0001-Revert-drm-radeon-Remove-radeon_gart_restore.patch I used a fresh version of 3.17-rc5 and applied your patch. As you can see, your patch did not fix the bug. I can't see new debug messages you mentioned. Do I have to tell dmesg to be more verbose? How? I'd like to test a revert of the commit that my bisect said it is bad. But as I wrote, my first try ended in compilation fail. (undefined things). What part of http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/gpu/drm/radeon?id=77497f2735ad6e29c55475e15e9790dbfa2c2ef8 should I try to revert? Tia Created attachment 106470 [details] [review] Fix typo 'addr' -> 'entry' in rs400_gart_set_page Does this patch fix the problem? Please test it without Christian's patch, that one can't work correctly because it doesn't set the correct GART page flags in all cases. (In reply to comment #3) > Created attachment 106470 [details] [review] [review] > Fix typo 'addr' -> 'entry' in rs400_gart_set_page > > Does this patch fix the problem? > > Please test it without Christian's patch, that one can't work correctly > because it doesn't set the correct GART page flags in all cases. This fixes the bug. Thank you all so much! Will I see the commit in -rc6 ? dmesg for linux-3.17_rc5 patched with rs400-gart_set_page-typo.diff: [ 6.929979] [drm] Initialized drm 1.1.0 20060810 [ 7.565290] [drm] radeon kernel modesetting enabled. [ 7.566263] [drm] initializing kernel modesetting (RS740 0x1002:0x796E 0x105B:0x0E13). [ 7.566289] [drm] register mmio base: 0xFEAF0000 [ 7.566293] [drm] register mmio size: 65536 [ 7.566880] ATOM BIOS: ATI [ 7.566893] radeon 0000:01:05.0: VRAM: 128M 0x00000000D8000000 - 0x00000000DFFFFFFF (128M used) [ 7.566895] radeon 0000:01:05.0: GTT: 512M 0x00000000A0000000 - 0x00000000BFFFFFFF [ 7.566907] [drm] Detected VRAM RAM=128M, BAR=128M [ 7.566908] [drm] RAM width 128bits DDR [ 7.566971] [TTM] Zone kernel: Available graphics memory: 1959278 kiB [ 7.566972] [TTM] Initializing pool allocator [ 7.566976] [TTM] Initializing DMA pool allocator [ 7.566997] [drm] radeon: 128M of VRAM memory ready [ 7.566998] [drm] radeon: 512M of GTT memory ready. [ 7.567010] [drm] GART: num cpu pages 131072, num gpu pages 131072 [ 7.582020] [drm] radeon: 1 quad pipes, 1 z pipes initialized. [ 7.582025] [drm] PCIE GART of 512M enabled (table at 0x00000000D4F00000). [ 7.582119] radeon 0000:01:05.0: WB enabled [ 7.582138] radeon 0000:01:05.0: fence driver on ring 0 use gpu addr 0x00000000a0000000 and cpu addr 0xffff8800d7077000 [ 7.582146] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013). [ 7.582149] [drm] Driver supports precise vblank timestamp query. [ 7.582173] [drm] radeon: irq initialized. [ 7.582201] [drm] Loading RS690/RS740 Microcode [ 7.734500] [drm] radeon: ring at 0x00000000A0001000 [ 7.734530] [drm] ring test succeeded in 0 usecs [ 7.734793] [drm] ib test succeeded in 0 usecs [ 7.736412] [drm] Radeon Display Connectors [ 7.736416] [drm] Connector 0: [ 7.736418] [drm] VGA-1 [ 7.736421] [drm] DDC: 0x7e50 0x7e40 0x7e54 0x7e44 0x7e58 0x7e48 0x7e5c 0x7e4c [ 7.736423] [drm] Encoders: [ 7.736425] [drm] CRT1: INTERNAL_KLDSCP_DAC1 [ 7.736426] [drm] Connector 1: [ 7.736428] [drm] DVI-D-1 [ 7.736429] [drm] HPD2 [ 7.736432] [drm] DDC: 0x7e40 0x7e60 0x7e44 0x7e64 0x7e48 0x7e68 0x7e4c 0x7e6c [ 7.736433] [drm] Encoders: [ 7.736435] [drm] DFP2: INTERNAL_DDI [ 7.736436] [drm] Connector 2: [ 7.736437] [drm] DVI-D-2 [ 7.736440] [drm] DDC: 0x7e40 0x7e50 0x7e44 0x7e54 0x7e48 0x7e58 0x7e4c 0x7e5c [ 7.736441] [drm] Encoders: [ 7.736443] [drm] DFP3: INTERNAL_LVTM1 [ 7.788582] [drm] fb mappable at 0xF0040000 [ 7.788591] [drm] vram apper at 0xF0000000 [ 7.788595] [drm] size 9216000 [ 7.788598] [drm] fb depth is 24 [ 7.788601] [drm] pitch is 7680 [ 7.788796] fbcon: radeondrmfb (fb0) is primary device [ 7.818230] Console: switching to colour frame buffer device 240x75 [ 7.843203] radeon 0000:01:05.0: fb0: radeondrmfb frame buffer device [ 7.843210] radeon 0000:01:05.0: registered panic notifier [ 7.851229] [drm] Initialized radeon 2.40.0 20080528 for 0000:01:05.0 on minor 0 (In reply to comment #4) > This fixes the bug. Thank you all so much! Great, thank you for testing it. > Will I see the commit in -rc6 ? Not sure it'll make rc6, but hopefully before 3.17 final. https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=1734a6e47fbad93f896644ad8ad12574fadba49d Merge branch 'drm-fixes' of git://people.freedesktop.org/~airlied/linux drm/radeon: Fix typo 'addr' -> 'entry' in rs400_gart_set_page |
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.