Bug 92761 - [NVF0] dmesg full of nouveau E[ PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
Summary: [NVF0] dmesg full of nouveau E[ PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1...
Status: NEW
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/nouveau (show other bugs)
Version: unspecified
Hardware: Other Linux (All)
: medium normal
Assignee: Nouveau Project
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-10-31 23:48 UTC by ascanio.alba7
Modified: 2015-11-23 12:41 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
dmesg output (4.13 KB, text/plain)
2015-10-31 23:48 UTC, ascanio.alba7
no flags Details
lspci.txt (192 bytes, text/plain)
2015-11-19 12:43 UTC, ascanio.alba7
no flags Details
xrandr (1.32 KB, text/plain)
2015-11-19 12:43 UTC, ascanio.alba7
no flags Details
Xorg.0.log (26.66 KB, text/plain)
2015-11-19 12:44 UTC, ascanio.alba7
no flags Details
mmiotrace output (7.37 MB, application/octet-stream)
2015-11-19 12:54 UTC, ascanio.alba7
no flags Details
test patch (632 bytes, patch)
2015-11-19 22:48 UTC, Ben Skeggs
no flags Details | Splinter Review
take 2 (1.50 KB, patch)
2015-11-19 23:31 UTC, Ben Skeggs
no flags Details | Splinter Review
dmesg kernel 4.4 + nouveau patch (take2) (189.60 KB, text/plain)
2015-11-21 12:51 UTC, ascanio.alba7
no flags Details
take 3 (33.34 KB, patch)
2015-11-22 19:51 UTC, Ben Skeggs
no flags Details | Splinter Review

Note You need to log in before you can comment on or make changes to this bug.
Description ascanio.alba7 2015-10-31 23:48:38 UTC
Created attachment 119321 [details]
dmesg output

1. nouveau on Fedora 23
2. dmesg is full of 
[68759.141891] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.161134] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.172614] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.172695] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.279172] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.293372] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.305836] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.305903] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.470585] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
[68759.474667] nouveau E[   PIBUS][0000:01:00.0] GPC4: 0x5233e4 0xbadf1301 (0x01024215)
3. kernel 4.2.3-300.fc23.x86_64
4. xorg-x11-drv-nouveau-1.0.12-0.3.fc23.x86_64
5. 01:00.0 VGA compatible controller: NVIDIA Corporation GK110 [GeForce GTX 780] (rev a1)
Comment 1 Ben Skeggs 2015-11-08 22:15:06 UTC
I don't suppose you're able to grab me a mmiotrace[1] of this board from the latest NVIDIA binary driver?

[1] http://nouveau.freedesktop.org/wiki/MmioTrace/
Comment 2 ascanio.alba7 2015-11-19 12:43:19 UTC
Created attachment 119933 [details]
lspci.txt
Comment 3 ascanio.alba7 2015-11-19 12:43:41 UTC
Created attachment 119934 [details]
xrandr
Comment 4 ascanio.alba7 2015-11-19 12:44:08 UTC
Created attachment 119935 [details]
Xorg.0.log
Comment 5 ascanio.alba7 2015-11-19 12:54:42 UTC
Created attachment 119937 [details]
mmiotrace output

mmiotrace output

Thanks for your patience.

Waited for nvidia Xorg 1.18 ABI 20 compatiblity
Comment 6 ascanio.alba7 2015-11-19 12:56:38 UTC
1. Using nvidia 352.63 to generate mmiotrace.

2. This card locks up with gdm -> desktop (wayland or X)
a. gdm starts desktop and immediately returns to gdm
b. 2nd attempt to start desktop usually results in a hard lockup - alt-sysrq works

3. lightdm -> X is ok
Comment 7 Ben Skeggs 2015-11-19 22:48:35 UTC
Created attachment 119955 [details] [review]
test patch

Thanks for the trace.

Are you able to give the attached patch a try and see if it helps the situation at all?  A more complete solution is probably required in general, but this might help in your case.

Thanks,
Ben.
Comment 8 Ben Skeggs 2015-11-19 22:55:55 UTC
Actually, don't test that patch.  I've not finished my first coffee of the morning and it contains a rather large thinko on my part :P
Comment 9 Ben Skeggs 2015-11-19 23:31:14 UTC
Created attachment 119957 [details] [review]
take 2

Coffee has been drunk, so here's a somewhat better attempt at what I was trying to achieve ;)
Comment 10 ascanio.alba7 2015-11-21 10:19:31 UTC
I'm trying to match Fedora 23 kernels approx 4.2.25 - 4.2.26 couldn't patch the second diff:


# grep nvkm_rd32 drivers/gpu/drm/nouveau/nvkm/engine/gr/*.c
[root@localhost linux]# 



Any recommendation for kernel: 4.4rc?
Comment 11 ascanio.alba7 2015-11-21 10:22:01 UTC
Actually, just found that section in master...will try from there...
Comment 12 ascanio.alba7 2015-11-21 12:50:10 UTC
Unfortunately the patch didn't work.

Built kernel 4.4 rc1 with nouveau patch still get:

[   58.996163] nouveau 0000:01:00.0: priv: GPC4: 5233e4 badf1301 (01024215)
[   59.331297] nouveau 0000:01:00.0: priv: GPC4: 5233e4 badf1301 (01024215)
[   63.246550] nouveau 0000:01:00.0: priv: GPC4: 5233e4 badf1301 (01024215)
Comment 13 ascanio.alba7 2015-11-21 12:51:14 UTC
Created attachment 120004 [details]
dmesg kernel 4.4 + nouveau patch (take2)
Comment 14 Ben Skeggs 2015-11-22 19:51:43 UTC
Created attachment 120025 [details] [review]
take 3

Looks like the ucode needs some fixes too, new patch attached.
Comment 15 ascanio.alba7 2015-11-23 12:41:19 UTC
Two thumbs up for take 3 patch.

1. No more log spamming with PIBUS
2. lightdm to X desktop works
3. gdm/X to X desktop works
4. gdm/Wayland to X or Wayland desktop works.

Gnome terminal has an extra white border(Wayland) but I think that is a separate problem.

The first working nouveau on this card yet.

Hope you can get this in for 4.4 so Fedora will pick it up.

Thanks!


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.