Bug 14010 - TNT2 generate too much noise in logs
Summary: TNT2 generate too much noise in logs
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/nouveau (show other bugs)
Version: git
Hardware: x86 (IA32) Linux (All)
: medium normal
Assignee: Nouveau Project
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-01-10 09:30 UTC by Andrew Randrianasulu
Modified: 2008-01-10 17:51 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
/var/log/debug, last 1000 lines (79.26 KB, text/plain)
2008-01-10 09:32 UTC, Andrew Randrianasulu
no flags Details
/var/log/messages (133.32 KB, text/plain)
2008-01-10 09:32 UTC, Andrew Randrianasulu
no flags Details

Description Andrew Randrianasulu 2008-01-10 09:30:53 UTC
xserver git e070c2cbef0d6fbbafce8d417e8b29231c2fdc50
drm git 5f15f317fb304f6a2321c033d401f603b365f2d0
xf86-video-nouveau 31b1a894079ca3983d17ac7a81d6ecd420f7d377
kernel 2.6.23.13 from kernel.org
Window manager - E 0.16.8.11-001
additional software: gkrellm 2.3.0 running in background

hw:

lspci -v
01:00.0 VGA compatible controller: nVidia Corporation NV5 [RIVA TNT2/TNT2 Pro] (rev 15) (prog-if 00 [VGA])
        Subsystem: ASUSTeK Computer Inc. AGP-V3800PRO
        Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 19
        Memory at de000000 (32-bit, non-prefetchable) [size=16M]
        Memory at da000000 (32-bit, prefetchable) [size=32M]
        Expansion ROM at dfef0000 [disabled] [size=64K]
        Capabilities: [60] Power Management version 1
        Capabilities: [44] AGP version 2.0

lspci -vn
01:00.0 0300: 10de:0028 (rev 15) (prog-if 00 [VGA])
        Subsystem: 1043:4000
        Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 19
        Memory at de000000 (32-bit, non-prefetchable) [size=16M]
        Memory at da000000 (32-bit, prefetchable) [size=32M]
        Expansion ROM at dfef0000 [disabled] [size=64K]
        Capabilities: [60] Power Management version 1
        Capabilities: [44] AGP version 2.0

During normal usage my /var/log/messages grow very quickly, and can eat 300mb+ in few hours.

Log full with messages like 

[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d01da
[drm] PGRAPH_NOTIFY - nSource: STATE_INVALID, nStatus: INVALID_STATE PROTECTION_FAULT
[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d0006
[drm] PGRAPH_NOTIFY - nSource: STATE_INVALID, nStatus: INVALID_STATE PROTECTION_FAULT
[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d01da
[drm] PGRAPH_NOTIFY - nSource: STATE_INVALID, nStatus: INVALID_STATE PROTECTION_FAULT
[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d0006
[drm] PGRAPH_NOTIFY - nSource: STATE_INVALID, nStatus: INVALID_STATE PROTECTION_FAULT
[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d01da
[drm] PGRAPH_NOTIFY - nSource: STATE_INVALID, nStatus: INVALID_STATE PROTECTION_FAULT
[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d0006
[drm] PGRAPH_NOTIFY - nSource: STATE_INVALID, nStatus: INVALID_STATE PROTECTION_FAULT
[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d00de
[drm] PGRAPH_NOTIFY - nSource: STATE_INVALID, nStatus: INVALID_STATE PROTECTION_FAULT
[drm] PGRAPH_NOTIFY - Ch 1/4 Class 0x005f Mthd 0x0308 Data 0x0001e83c:0x000d0006

i will add last 1000 lines from just X start. This card doesn't work in randr12 mode, but this is separate issue.
Comment 1 Andrew Randrianasulu 2008-01-10 09:32:12 UTC
Created attachment 13644 [details]
/var/log/debug, last 1000 lines
Comment 2 Andrew Randrianasulu 2008-01-10 09:32:47 UTC
Created attachment 13645 [details]
/var/log/messages
Comment 3 Andrew Randrianasulu 2008-01-10 09:42:13 UTC
I don't know when this behaviour started, i tried compile old nouveau from 06 aug 2007 (right after drm interface changes) - but it doesn't compiled on my too new X server, and version from around 15 Oct 2007 still have this bug. Git bisect give me bogus result (first good commit == first bad commit) if i mark non-compiled versions as 'bad' ones. And trying all individual revisions one after another is too much for me. 

Comment 4 Andrew Randrianasulu 2008-01-10 17:49:05 UTC
simple fix from Ben Skeggs, works really well:

diff --git a/shared-core/nv04_graph.c b/shared-core/nv04_graph.c
index 81a6d5c..50ecc74 100644
--- a/shared-core/nv04_graph.c
+++ b/shared-core/nv04_graph.c
@@ -497,7 +497,8 @@ int nv04_graph_init(struct drm_device *d
        /*haiku same*/
 
        /*NV_WRITE(NV04_PGRAPH_DEBUG_3, 0xfad4ff31);*/
-       NV_WRITE(NV04_PGRAPH_DEBUG_3, 0x10d4ff31);
+       // NV_WRITE(NV04_PGRAPH_DEBUG_3, 0x10d4ff31);
+       NV_WRITE(NV04_PGRAPH_DEBUG_3, 0xf0d4ff31); 
        /*haiku and blob 10d4*/



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.