Bug 9898 - The low VideoRam problem with no accelation method in use
Summary: The low VideoRam problem with no accelation method in use
Status: VERIFIED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/intel (show other bugs)
Version: git
Hardware: x86 (IA32) Linux (All)
: high major
Assignee: Wang Zhenyu
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: 10102
  Show dependency treegraph
 
Reported: 2007-02-06 19:29 UTC by xuji
Modified: 2007-03-06 23:38 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:


Attachments
xorg.conf (3.05 KB, text/plain)
2007-02-06 19:37 UTC, xuji
no flags Details
xorg.log (38.20 KB, text/plain)
2007-02-06 19:38 UTC, xuji
no flags Details
xorg.log after setting VideoRam (50.29 KB, text/plain)
2007-02-07 00:44 UTC, xuji
no flags Details

Description xuji 2007-02-06 19:29:08 UTC
--Architecture(ia32,ia32e): ia32 & ia32e 
--Xf86_video_intel modesetting(785a59ead0e8d1d681b2cb6827ee58ad2c51f8c6)
--Mesa: master(8518ddda7b5f8e43de4859042fe59f55aa0c2760)
--Xserver: server-1.2-branch(4e6872fca75b77c342a7e1e715b1c30dc2642a73)
--OS: FC5

Description:
=======================
If no accelation method is implemented, X can not start with following info:

(EE) intel(0): Failed to allocate framebuffer. Is your VideoRAM set too low?
Fatal server error:
AddScreen/ScreenInit failed for driver 0

Moreover, if the VideoRam is set, X can start but DRI is disabled:

(EE) AIGLX: Screen 0 is not DRI capable

Reproduce step:
=======================
Revise xorg.conf to:
#       Option      "AccelMethod" "exa"
        Option      "NoAccel" "TRUE"

then start X

Expected result:
=======================
X can start.
Comment 1 xuji 2007-02-06 19:37:52 UTC
Created attachment 8614 [details]
xorg.conf
Comment 2 xuji 2007-02-06 19:38:16 UTC
Created attachment 8615 [details]
xorg.log
Comment 3 Wang Zhenyu 2007-02-06 20:02:54 UTC
You may set VideoRam in your xorg.conf for no accel or other accel method. 
Current code seems failed to detect what should be the apropriate mem size, I'll take a look, you may use VideoRam for now. 
Comment 4 xuji 2007-02-06 21:33:22 UTC
As I said above, if I assigned VideoRam with a value 131072 or so
then X can be started with DRI disabled

>>Moreover, if the VideoRam is set, X can start but DRI is disabled:

>>(EE) AIGLX: Screen 0 is not DRI capable
Comment 5 Wang Zhenyu 2007-02-06 21:48:13 UTC
Then you should attach the failure log in that case. It might doesn't load your dri driver correctly, maybe an install or config issue. 
Comment 6 xuji 2007-02-07 00:44:21 UTC
Created attachment 8620 [details]
xorg.log after setting VideoRam
Comment 7 Eric Anholt 2007-02-26 12:31:47 UTC
Please try with the current modesetting branch and no VideoRam setting.
Comment 8 xuji 2007-02-26 19:01:23 UTC
with the current modesetting 2D driver, the problem remains the same. 
Comment 9 Daniel Stone 2007-02-27 01:36:14 UTC
Sorry about the phenomenal bug spam, guys.  Adding xorg-team@ to the QA contact so bugs don't get lost in future.
Comment 10 xuji 2007-02-27 19:39:00 UTC
Now the issue has some changes:
With the NoAccel option enabled, starting X will fail with following backtrace:

Backtrace:
0: X(xf86SigHandler+0x81) [0x80b3dd1]
1: [0xb7f24420]
2: X(AddScreen+0x1ee) [0x806f49e]
3: X(InitOutput+0x21e) [0x80a0b2e]
4: X(main+0x27b) [0x806fc3b]
5: /lib/libc.so.6(__libc_start_main+0xdc) [0x9eb7e4]
6: X(FontFileCompleteXLFD+0xa9) [0x806f181]

Fatal server error:
Caught signal 11.  Server aborting

Setting the VideoRam value renders the same result.
Comment 11 Eric Anholt 2007-03-06 14:23:28 UTC
Fix pushed.


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.