Fedora 3.5.0 rc7 kernel with drm/nouveau driver....when running with no flags particular to nouveau, I see this error on bootup. Then when I try to do my startx X fails in libexa (I assume looking for acceleration that isn't there as a result of the GPU lockup). I can't boot the Fedora 3.5.0 kernels with nouveau.noaccel=1 at all; machine hangs hard and I have to power off/on to restart. Either the GPU lockup needs to be corrected (nv50 nVidia card) or I need to be able to disable the accleration in the nouveau driver.
Hi Kevin Would you mind providing more information as per our bug policy [1]. I can inform you that nouveau runs on quite a few gpus/cards without problems, thus the more information you provide the easier the resolution would be Regards Emil [1] http://nouveau.freedesktop.org/wiki/Bugs
Created attachment 64370 [details] dmesg showing drm/nouveau messages during the boot up and attempted running of X with kernel 3.5.0 dmesg output as requested.
Created attachment 64371 [details] log showing the stack trace for X under the 3.5.0 kernel.
Created attachment 64393 [details] vbios as requested.
Archlinux with gtx-560 and nouveau drivers currently running kernel linux-3.4.9-1 without problems. Yesterday I updated to kernel 3.5.3-1 and after booting got no login screen at all, when I changed tty to log in I discovered the messages: Failed to idle channel 1. PFIFO - playlist update failed PFIFO - playlist update failed I tried to fix it from there but it wouldn't allow me because it kept showing those messages all the time. I had to boot on run level 3 to be able to roll back to kernel 3.4.9-1.
Created attachment 66279 [details] kernel error I got
Created attachment 66280 [details] Parsed error file showing similar results.
Created attachment 66281 [details] Full system report from HardInfo 0.5.1 to aid invesitigation.
OS: Arch Linux Kernel Version: 3.5.3 Card: nvidia GTX570 Similar error of: GPU lockup - switching to software fbcon Downgrade to 3.4 resolved the issue.
Running arch linux with kernel 3.5.3-1-ARCH (x86_64) and nouveau xf86-video-nouveau 1.0.1-1 Although the machine boots and runs with the screen display apparently normal I do get in the kernel log: Sep 11 09:37:35 localhost kernel: [ 389.092546] [drm] nouveau 0000:01:00.0: PFIFO_CACHE_ERROR - Ch 3/0 Mthd 0x0060 Data 0x8000000f Sep 11 09:40:11 localhost kernel: [ 545.848393] [drm] nouveau 0000:01:00.0: PFIFO_CACHE_ERROR - Ch 3/5 Mthd 0x1b00 Data 0x00000000 However when I shutdown (after running KDE desktop with desktop effects switched on) I get a hang for some seconds after the X session ends, with a message on the console "GPU lockup - switching to software fbcon". However I don't see the GPU lockup message in the logs - and only on the console during shutdown. If there is any additional diagnostics that I can run to help pin down where the problem lies I will be happy to run tests.
I also have this in the kernel log but I don't know if it is indicative of the same issue: Sep 11 09:31:20 localhost kernel: [ 9.278108] [drm] nouveau 0000:01:00.0: allocated 1680x1050 fb: 0x2b0000, bo ffff88011284e400 Sep 11 09:31:20 localhost kernel: [ 9.278195] fbcon: nouveaufb (fb0) is primary device Sep 11 09:31:20 localhost kernel: [ 9.317007] Console: switching to colour frame buffer device 210x65 Sep 11 09:31:20 localhost kernel: [ 9.321037] fb0: nouveaufb frame buffer device Sep 11 09:31:20 localhost kernel: [ 9.321039] drm: registered panic notifier Sep 11 09:31:20 localhost kernel: [ 9.321044] [drm] Initialized nouveau 1.0.0 20120316 for 0000:01:00.0 on minor 0 Does the reference to panic notifier indicate a problem or is that expected?
I forgot to add that I am using card NVIDIA Corporation G86M [Quadro FX 360M] (rev a1)
https://bugs.freedesktop.org/show_bug.cgi?id=53101 and https://bugs.freedesktop.org/show_bug.cgi?id=53566 look to be a duplicates of this bug. Can confirm that reverting commit 1a46098e910b96337f0fe3838223db43b923bad4 'fixes'* the problem. I'm using 3.6.0-rc5 right now after doing just that. *Haven't extensively tested this yet, I've only just compiled and booted a 'fixed' kernel, but everything seems fine so far. In the second report, Vlad K suggests that reverting this commit may only delay the lockup.
It would be nice to see this fix appearing in the released version? Forgive my ignorance but does this go direct into the kernel or into libdrm-nouveau1 or nouveau-dri or xf86-video-nouveau? If it is in the kernel will it likely get o the next released 3.5.x? Thanks
I have been waiting for a patch too, my GTX 570 (NVIDIA NVc8) doesn't work in any 3.5 kernel.
This problem seems to be fixed in 3.6, I have been running 3.6.3 for a while now without any problems
Still occurs under kernel 3.7: [ 4.787773] nouveau [ DEVICE][0000:01:00.0] BOOT0 : 0x0a3880a2 [ 4.789963] nouveau [ DEVICE][0000:01:00.0] Chipset: GT215 (NVA3) [ 4.790141] nouveau [ DEVICE][0000:01:00.0] Family : NV50 [ 4.793661] nouveau [ VBIOS][0000:01:00.0] checking PRAMIN for image... [ 4.948004] nouveau [ VBIOS][0000:01:00.0] ... appears to be valid [ 4.948204] nouveau [ VBIOS][0000:01:00.0] using image from PRAMIN [ 4.948637] nouveau [ VBIOS][0000:01:00.0] BIT signature found [ 4.948824] nouveau [ VBIOS][0000:01:00.0] version 70.15.43.00 [ 4.998607] nouveau [ MXM][0000:01:00.0] no VBIOS data, nothing to do [ 4.998859] nouveau [ PFB][0000:01:00.0] RAM type: GDDR5 [ 4.999110] nouveau [ PFB][0000:01:00.0] RAM size: 1024 MiB [ 5.038919] nouveau [ DRM] VRAM: 1024 MiB [ 5.039095] nouveau [ DRM] GART: 512 MiB [ 5.039262] nouveau [ DRM] BIT BIOS found [ 5.039428] nouveau [ DRM] Bios version 70.15.43.00 [ 5.039594] nouveau [ DRM] TMDS table version 2.0 [ 5.039757] nouveau [ DRM] DCB version 4.0 [ 5.039933] nouveau [ DRM] DCB outp 00: 01000323 00010034 [ 5.040106] nouveau [ DRM] DCB outp 01: 02014300 00000000 [ 5.040274] nouveau [ DRM] DCB outp 02: 02021362 00020010 [ 5.040444] nouveau [ DRM] DCB outp 03: 080333c6 0f220010 [ 5.040611] nouveau [ DRM] DCB outp 04: 08033382 00020010 [ 5.040780] nouveau [ DRM] DCB conn 00: 00000340 [ 5.041031] nouveau [ DRM] DCB conn 01: 00001061 [ 5.041273] nouveau [ DRM] DCB conn 02: 00000147 [ 5.041510] nouveau [ DRM] DCB conn 03: 00202346 [ 5.041744] nouveau [ DRM] DCB conn 04: 00000400 [ 5.041989] nouveau [ DRM] DCB conn 05: 00000210 [ 5.042229] nouveau [ DRM] DCB conn 06: 00000211 [ 5.042468] nouveau [ DRM] DCB conn 07: 00000213 [ 5.042705] nouveau [ DRM] DCB conn 08: 00020846 [ 5.042948] nouveau [ DRM] DCB conn 09: 00000900 [ 5.083956] nouveau [ DRM] ACPI backlight interface available, not registering our own [ 6.448619] nouveau [ DRM] 3 available performance level(s) [ 6.448821] nouveau [ DRM] 0: core 135MHz shader 270MHz memory 135MHz voltage 800mV [ 6.449110] nouveau [ DRM] 1: core 405MHz shader 810MHz memory 324MHz voltage 850mV [ 6.449395] nouveau [ DRM] 3: core 576MHz shader 1436MHz memory 1600MHz voltage 1050mV [ 6.449720] nouveau [ DRM] c: core 405MHz shader 810MHz memory 324MHz voltage 850mV [ 6.506727] nouveau [ DRM] MM: using COPY for buffer copies [ 6.664752] nouveau [ DRM] allocated 1680x945 fb: 0xd0000, bo ffff88012d5d3f60 [ 6.665561] fbcon: nouveaufb (fb0) is primary device [ 7.130142] nouveau E[ DRM] GPU lockup - switching to software fbcon [ 7.138026] fb0: nouveaufb frame buffer device [ 7.138116] [drm] Initialized nouveau 1.1.0 20120801 for 0000:01:00.0 on minor 0
I can confirm this bug for kernel 3.7.3
This looks like a duplicate of 33165. Please correct me if I am wrong. *** This bug has been marked as a duplicate of bug 33165 ***
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.