Bug 69952

mmio traces:

filed in another bug (https://bugs.freedesktop.org/show_bug.cgi?id=69928)
on the same hw:

module load:
https://bugs.freedesktop.org/attachment.cgi?id=87677


simple xinit with sleep:
mmiotrace_xinit.log.xz (3.1 MB)
https://mega.co.nz/#!u4RwFLjK!Yzi4UVujupmRdEIMFZtbQBOtjyoJFR1jrXheO1RUpew

starting glxgears:
mmiotrace_glxgears.log.xz (3.4 MB)
https://mega.co.nz/#!744wlQ7b!CBnbnEYfsIwOuXKFojgy3kuecVkC6U43Bye4nZkKMUk

starting nvidia-settings - since dual screen is active already just browse through options and then quit:
mmiotrace_dualsettings.log.xz (2.7 MB)
https://mega.co.nz/#!L1ZR3ApZ!XEuYymYmvLyQSQdpwQ0HaiBEipp2CBHz9y3PHNBNn6Q


glxgears posted because: https://bugs.freedesktop.org/show_bug.cgi?id=69952

Does posting the card [1] resolve this and/or any of your other bugs ?

Cheers,
Emil

[1] Try either one of the below two
* via the kernel command line - append nouveau.config=NvForcePost=1
* s2ram before running any test - mplayer/glxgears

I enabled posting via kernel cmdline.

cat /proc/cmdline:

BOOT_IMAGE=/vmlinuz-3.12.0 root=UUID=c52e229d-be9c-4d85-b262-5085459dc2d9 ro rootflags=subvol=@ sysrq_always_enabled ignore_loglevel debug log_buf_len=5M nouveau.config=NvMSI=0 nouveau.config=NvForcePost=1 sysrq_always_enabled


The crash still occurs:
- LXDE running glxgears
- XFCE running glxgears

Running on KDE : running glxgears "freezes" / makes the Desktop very slow but it does not crash Xorg there.

System:
-Linux Mint 15 (Ubuntu 13.04)
-NVAA (Geforce 8200) onboard (Jetway JNC62K)
-dual monitor setup (HDMI+VGA)

-Kernel 3.12.0

Steps to replicate this time:
- Login to desktop (XFCE, dual screen)
- in open terminal window : run glxgears
- switch to other virtual desktop
- Xorg hangs, crash (login screen reappers)

Created attachment 89628 [details]
dmesg after crash of Xorg when running glxgears

the log contains more debug info; cmdline had "nouveau.debug=trace drm.debug=14"

[  190.676020] nouveau E[glxgears[3270]] failed to idle channel 0xcccc0000 [glxgears[3270]]
[  205.676014] nouveau E[glxgears[3270]] failed to idle channel 0xcccc0000 [glxgears[3270]]
[  205.676119] nouveau E[     PFB][0000:02:00.0] trapped read at 0x0020012020 on channel 0x0000797f [unknown] SEMAPHORE_BG/PFIFO_READ/00 reason: PAGE_NOT_PRESENT

after that many:
nouveau E[  PGRAPH][0000:02:00.0]  ILLEGAL_MTHD ILLEGAL_CLASS

Created attachment 90379 [details]
Same error on kernel 3.13-rc2

Kernel 3.13-rc2 has it too 

complete log attached with nouveau trace enabled; possibly interesting snippets (these were seen previously too):
[   47.726763] nouveau E[   PFIFO][0000:02:00.0] DMA_PUSHER - ch 2 [Xorg[2244]] get 0x0020016b30 put 0x0020016b4c ib_get 0x0000024b ib_put 0x000002a9 state 0x80000030 (err: INVALID_CMD) push 0x00400040
[   47.810523] nouveau E[  PGRAPH][0000:02:00.0] DATA_ERROR INVALID_VALUE
[   47.810545] nouveau E[  PGRAPH][0000:02:00.0]  DATA_ERROR
[   47.810565] nouveau E[  PGRAPH][0000:02:00.0] ch 2 [0x0007b45000 Xorg[2244]] subc 2 class 0x502d mthd 0x08d4 data 0x00144230
[  131.224959] nouveau E[   PFIFO][0000:02:00.0] DMA_PUSHER - ch 2 [Xorg[2244]] get 0x002002b068 put 0x002002b11c ib_get 0x0000016c ib_put 0x000001bb state 0x800048e0 (err: INVALID_CMD) push 0x00400040
[  131.392877] nouveau E[  PGRAPH][0000:02:00.0] DATA_ERROR INVALID_VALUE
[  131.392898] nouveau E[  PGRAPH][0000:02:00.0]  DATA_ERROR
[  131.392918] nouveau E[  PGRAPH][0000:02:00.0] ch 2 [0x0007b45000 Xorg[2244]] subc 2 class 0x502d mthd 0x08dc data 0x00144230


[  162.340013] nouveau E[Xorg[2244]] failed to idle channel 0xcccc0000 [Xorg[2244]]
[  177.340015] nouveau E[Xorg[2244]] failed to idle channel 0xcccc0000 [Xorg[2244]]
[  207.520133] nouveau E[     PFB][0000:02:00.0] trapped read at 0x0020012020 on channel 0x0000797f [unknown] SEMAPHORE_BG/PFIFO_READ/00 reason: PAGE_NOT_PRESENT

There was some patches specifically for NVAA/AC cards that went into kernel 3.19. They solve a different issue, but it would still be worth to test it - a lot of other patches were merged since 3.13, which might also help with it.

-- 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-nouveau/issues/61.