Created attachment 69358 [details] Xorg log when the corruption occurs, not much here After logging into Gnome 3.6 display turns up corrupted (background garbled, letters missing in labels). Also several errors come up in the logs. Especially the ones that come up after logging out and logging back in look interesting. Restarting the gnome-shell with Alt-F2,r clears up the corruption. xorg-server 1.13.0 xf86-video-ati 6.14.6
Created attachment 69359 [details] Screenshot of the corruption after first login
Created attachment 69360 [details] Dmesg output after the first login, some errors
Created attachment 69361 [details] Screenshot of the corruption after the second login
Created attachment 69362 [details] Dmesg output after the second login, some more interesting errors
There's also one line that didn't end up in the attached logs here for some reason (Failed to allocate GEM object). --- [ 338.191642] [TTM] Failed to find memory space for buffer 0xffff8803f2c8fc48 eviction [ 338.191647] [TTM] No space for ffff8803f2c8fc48 (4160 pages, 16640K, 16M) [ 338.191649] [TTM] placement[0]=0x00070002 (1) [ 338.191650] [TTM] has_type: 1 [ 338.191651] [TTM] use_type: 1 [ 338.191652] [TTM] flags: 0x0000000A [ 338.191654] [TTM] gpu_offset: 0x40000000 [ 338.191655] [TTM] size: 131072 [ 338.191656] [TTM] available_caching: 0x00070000 [ 338.191657] [TTM] default_caching: 0x00010000 [ 338.191659] [TTM] 0x00000000-0x00000001: 1: used [ 338.191661] [TTM] 0x00000001-0x00000011: 16: used [ 338.191662] [TTM] 0x00000011-0x00000111: 256: used [ 338.191664] [TTM] 0x00000111-0x00000211: 256: used [ 338.191665] [TTM] 0x00000211-0x00000220: 15: free [ 338.191667] [TTM] 0x00000220-0x00001260: 4160: used [ 338.191668] [TTM] 0x00001260-0x000022a0: 4160: used [ 338.191670] [TTM] 0x000022a0-0x000032e0: 4160: used [ 338.191671] [TTM] 0x000032e0-0x00004320: 4160: used [ 338.191675] [TTM] 0x00004320-0x00005360: 4160: used [ 338.191678] [TTM] 0x00005360-0x00006300: 4000: used [ 338.191679] [TTM] 0x00006300-0x000072a0: 4000: used [ 338.191681] [TTM] 0x000072a0-0x00008240: 4000: used [ 338.191682] [TTM] 0x00008240-0x000082e0: 160: free [ 338.191683] [TTM] 0x000082e0-0x00009320: 4160: used [ 338.191685] [TTM] 0x00009320-0x0000a360: 4160: used [ 338.191686] [TTM] 0x0000a360-0x0000b2e0: 3968: free [ 338.191688] [TTM] 0x0000b2e0-0x0000c320: 4160: used [ 338.191689] [TTM] 0x0000c320-0x0000d060: 3392: free [ 338.191691] [TTM] 0x0000d060-0x0000e0a0: 4160: used [ 338.191692] [TTM] 0x0000e0a0-0x0000f040: 4000: used [ 338.191694] [TTM] 0x0000f040-0x0000ffe0: 4000: free [ 338.191695] [TTM] 0x0000ffe0-0x00010f80: 4000: used [ 338.191697] [TTM] 0x00010f80-0x00011de0: 3680: free [ 338.191701] [TTM] 0x00011de0-0x00012d80: 4000: used [ 338.191704] [TTM] 0x00012d80-0x00013d20: 4000: used [ 338.191705] [TTM] 0x00013d20-0x00014cc0: 4000: used [ 338.191707] [TTM] 0x00014cc0-0x00015c60: 4000: used [ 338.191708] [TTM] 0x00015c60-0x00016ca0: 4160: used [ 338.191710] [TTM] 0x00016ca0-0x00017ce0: 4160: used [ 338.191711] [TTM] 0x00017ce0-0x00018c80: 4000: used [ 338.191712] [TTM] 0x00018c80-0x00019c20: 4000: used [ 338.191714] [TTM] 0x00019c20-0x0001abc0: 4000: used [ 338.191715] [TTM] 0x0001abc0-0x0001bc00: 4160: used [ 338.191717] [TTM] 0x0001bc00-0x0001cba0: 4000: used [ 338.191718] [TTM] 0x0001cba0-0x0001dbe0: 4160: used [ 338.191720] [TTM] 0x0001dbe0-0x0001eb80: 4000: used [ 338.191721] [TTM] 0x0001eb80-0x0001fbc0: 4160: used [ 338.191723] [TTM] 0x0001fbc0-0x00020000: 1088: free [ 338.191724] [TTM] total: 131072, used 114769 free 16303 [ 338.191734] radeon 0000:01:00.0: object_init failed for (17039360, 0x00000006) [ 338.191737] [drm:radeon_gem_object_create] *ERROR* Failed to allocate GEM object (17039360, 4, 65536, -12) [ 338.257363] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Created attachment 69365 [details] One more image of the corruption I also had a dejavu. I think I had a similar issue also with an X1400, a different box and previous versions of Gnome. Same graphical symptoms, but more random (these are reproducable) and no errors in logs that time. The issue was never resolved.
Which version of Mesa are you using? Is this a regression, i.e. did it only start after you upgraded some software packages?
mesa 9.0 libgl 9.0 I haven't been using Gnome for a while, but decided to try it out now that 3.6 came out. If the issue had occurred just after a straightforward upgrade, I could've started bisecting immediately. Now the case is that I have know idea where to look, unfortunately.
Oh yeah. And I tried downgrading to mesa 8.0.4. No change.
I get the same corruption after updating the following packages, as well as the same kernel messages from the Radeon driver. [2012-10-31 17:32] upgraded glib2 (2.32.4-1 -> 2.34.1-1) [2012-10-31 17:32] upgraded clutter (1.10.6-1 -> 1.12.2-1) [2012-10-31 17:32] upgraded clutter-gst (1.6.0-1 -> 1.9.92-1) [2012-10-31 17:32] upgraded clutter-gtk (1.2.0-1 -> 1.4.0-1) [2012-10-31 17:32] upgraded libgee (0.6.5-1 -> 0.6.6-1) [2012-10-31 17:32] upgraded gnome-video-effects (0.4.0-1 -> 0.4.0-2) [2012-10-31 17:32] upgraded libcroco (0.6.5-1 -> 0.6.8-1) [2012-10-31 17:32] upgraded librsvg (2.36.3-1 -> 2.36.4-1) [2012-10-31 17:32] upgraded gnome-desktop (1:3.4.2-1 -> 1:3.6.1-1) [2012-10-31 17:33] upgraded gnome-backgrounds (3.4.2-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-color-manager (3.4.2-2 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-contacts (3.4.1-1 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-menus (3.4.2-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-control-center (3.4.2-1 -> 3.6.2-1) [2012-10-31 17:33] upgraded gnome-disk-utility (3.4.1-1 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-documents (0.4.2-1 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-font-viewer (3.4.0-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-games (3.4.2-2 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-keyring (3.4.1-3 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-nettool (3.1.1-1 -> 3.2.0-1) [2012-10-31 17:33] upgraded gnome-panel (3.4.2.1-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-power-manager (3.4.0-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded libxklavier (5.2.1-2 -> 5.3-1) [2012-10-31 17:33] upgraded libgnomekbd (3.4.0.2-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-screensaver (3.4.4-1 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-screenshot (3.4.1-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded zenity (3.4.0-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded mutter (3.4.1-2 -> 3.6.1-2) [2012-10-31 17:33] upgraded nautilus (3.4.2-1 -> 3.6.2-1) [2012-10-31 17:33] upgraded gnome-shell (3.4.2-1 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-system-log (3.4.1-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gtkmm3 (3.4.1-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-system-monitor (3.4.1-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-terminal (3.4.1.1-1 -> 3.6.0-1) [2012-10-31 17:33] upgraded gnome-themes-standard (3.4.2-1 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-tweak-tool (3.4.0.1-1 -> 3.6.1-1) [2012-10-31 17:33] upgraded gnome-user-docs (3.4.2-1 -> 3.6.0-1) [2012-10-31 17:34] upgraded xorg-xinit (1.3.2-1 -> 1.3.2-3)
Same corruption with same paul's gnome 3.4 > 3.6 upgrade on archlinux x86_64, with a Radeon HD3650 [ 119.497525] [TTM] Failed to find memory space for buffer 0xffff88010949d048 eviction [ 119.497531] [TTM] No space for ffff88010949d048 (4500 pages, 18000K, 17M) [ 119.497533] [TTM] placement[0]=0x00070002 (1) [ 119.497535] [TTM] has_type: 1 [ 119.497536] [TTM] use_type: 1 [ 119.497537] [TTM] flags: 0x0000000A [ 119.497538] [TTM] gpu_offset: 0x20000000 [ 119.497540] [TTM] size: 131072 [ 119.497541] [TTM] available_caching: 0x00070000 [ 119.497543] [TTM] default_caching: 0x00010000 [ 119.497545] [TTM] 0x00000000-0x00000001: 1: used [ 119.497547] [TTM] 0x00000001-0x00000011: 16: used [ 119.497549] [TTM] 0x00000011-0x00000111: 256: used [ 119.497550] [TTM] 0x00000111-0x00000211: 256: used [ 119.497552] [TTM] 0x00000211-0x00000219: 8: free [ 119.497554] [TTM] 0x00000219-0x000013ad: 4500: used [ 119.497555] [TTM] 0x000013ad-0x00001514: 359: free [ 119.497557] [TTM] 0x00001514-0x000026a8: 4500: used [ 119.497558] [TTM] 0x000026a8-0x0000383c: 4500: used [ 119.497560] [TTM] 0x0000383c-0x000038c6: 138: free [ 119.497562] [TTM] 0x000038c6-0x00004a5a: 4500: used [ 119.497564] [TTM] 0x00004a5a-0x00005bee: 4500: used [ 119.497565] [TTM] 0x00005bee-0x00006d82: 4500: used [ 119.497567] [TTM] 0x00006d82-0x00007f16: 4500: used [ 119.497569] [TTM] 0x00007f16-0x000090aa: 4500: used [ 119.497571] [TTM] 0x000090aa-0x0000a23e: 4500: used [ 119.497572] [TTM] 0x0000a23e-0x0000b3d2: 4500: used [ 119.497574] [TTM] 0x0000b3d2-0x0000c566: 4500: used [ 119.497576] [TTM] 0x0000c566-0x0000d6fa: 4500: used [ 119.497577] [TTM] 0x0000d6fa-0x0000e88e: 4500: used [ 119.497579] [TTM] 0x0000e88e-0x0000fa22: 4500: used [ 119.497580] [TTM] 0x0000fa22-0x00010bb6: 4500: used [ 119.497582] [TTM] 0x00010bb6-0x00010bb7: 1: used [ 119.497584] [TTM] 0x00010bb7-0x00010bb8: 1: used [ 119.497585] [TTM] 0x00010bb8-0x00010bb9: 1: used [ 119.497587] [TTM] 0x00010bb9-0x00010bba: 1: used [ 119.497589] [TTM] 0x00010bba-0x00010bbb: 1: used [ 119.497590] [TTM] 0x00010bbb-0x00010bbc: 1: used [ 119.497592] [TTM] 0x00010bbc-0x00011b90: 4052: free [ 119.497593] [TTM] 0x00011b90-0x00012d24: 4500: used [ 119.497595] [TTM] 0x00012d24-0x00013eb8: 4500: used [ 119.497596] [TTM] 0x00013eb8-0x0001504c: 4500: used [ 119.497598] [TTM] 0x0001504c-0x000161e0: 4500: used [ 119.497600] [TTM] 0x000161e0-0x00017374: 4500: used [ 119.497601] [TTM] 0x00017374-0x00018508: 4500: used [ 119.497603] [TTM] 0x00018508-0x0001969c: 4500: used [ 119.497605] [TTM] 0x0001969c-0x0001a830: 4500: used [ 119.497606] [TTM] 0x0001a830-0x0001b9c4: 4500: used [ 119.497608] [TTM] 0x0001b9c4-0x0001cb58: 4500: used [ 119.497610] [TTM] 0x0001cb58-0x0001dcec: 4500: used [ 119.497611] [TTM] 0x0001dcec-0x0001ee80: 4500: used [ 119.497613] [TTM] 0x0001ee80-0x00020000: 4480: free [ 119.497615] [TTM] total: 131072, used 122035 free 9037 [ 119.497618] [drm:radeon_cs_ioctl] *ERROR* Failed to parse relocation -12!
Can you try booting with radeon.gartsize=1024? Ideally with a 3.7 kernel.
I tried the mainline kernel 3.7rc4 with radeon.gartsize=1024 No corruption, but the kernel error messages are still there.
Created attachment 69767 [details] Kernel error messages with 3.7rc4 and radeon.gartsize=1024 Here are the messages I get (without corruption).
I'm not experiencing corruptions nor errors in dmesg after adding radeon.gartsize=1024
Interesting. I still do (get errors I mean - no corruption). Upgraded to 3.7rc5 and xf86-video-ati-7.0.0, but no change.
I have to correct my previous comment: the error messages reappeared in dmesg (however without all the [TTM] lines, only the "Failed to parse relocation -12" one) and the corruption too. The beahavior is still quite random, the error appears about once every 5 boots, the corruption once every 10 boots (or logout/login cycles)
This might be related to <https://bugzilla.kernel.org/show_bug.cgi?id=49981>
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/driver/xf86-video-ati/issues/48.
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.