Bug 4412

Summary: Test xorg against vsw4 generates failures
Product: xorg Reporter: Stew Benedict <sbenedict>
Component: Server/DDX/XvfbAssignee: Xorg Project Team <xorg-team>
Status: RESOLVED FIXED QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: high CC: fengrui
Version: git   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
short lsb-test-vsw4 journal with just cpyar tests
none
patch that corrects lsb-test-vsw4 failures none

Description Stew Benedict 2005-09-10 05:22:37 UTC 
Testing cvs as of 20050907 with LSB's lsb-test-vsw4 gives 17 errors in the cpyar
tests. (test uses Xvfb for testing)

ftp://ftp.freestandards.org/pub/lsb/test_suites/released-3.0.0/binary/runtime/lsb-test-vsw4-3.0.0-4.i486.rpm

Failures all seem to be with 16bit pixdepth and are of this nature:

520|0 6 00032583 1 155|TRACE:--- Running test with pixmap depth 16
520|0 6 00032583 1 156|TRACE:src=0, dest=0
520|0 6 00032583 1 157|CHECK:XCopyArea-6  69, line 241
520|0 6 00032583 1 158|TRACE:src=1, dest=0
520|0 6 00032583 1 159|REPORT:GXclear fail expected 0, got 1

I'll attach the full journal.

Problem code seems to be in xc/programs/Xserver/fb/fbcopy.c. If I undef USE_MMX
the problem goes away, or if I revert fbcopy.c to the 6.8.2 version without the
additional code wrapped in #ifdef USE_MMX the tests also pass. Perhaps someone
more familiar with the code can come up with a better fix. (my patch attached)
Comment 1 Stew Benedict 2005-09-10 05:24:29 UTC 
Created attachment 3219 [details]
short lsb-test-vsw4 journal with just cpyar tests
Comment 2 Stew Benedict 2005-09-10 05:25:06 UTC 
Created attachment 3220 [details] [review]
patch that corrects lsb-test-vsw4 failures
Comment 3 Stew Benedict 2005-09-10 05:41:38 UTC 
I neglected to update my email in bugzilla before posting the bug.
mandrakesoft.com is gone, current email is sbenedict@mandriva.com.
Comment 4 Daniel Stone 2007-02-27 01:27:55 UTC 
Sorry about the phenomenal bug spam, guys.  Adding xorg-team@ to the QA contact so bugs don't get lost in future.
Comment 5 Jesse Adkins 2010-11-19 15:33:33 UTC 
USE_MMX is gone in favor of FB_ACCESS_WRAPPER in the latest git. Additionally, the conditions are entirely different, as well as the method of blitting.

I also see that the condition change is a reaction to a rather nasty bug being fixed. I'm going to assume that your issue was caused by the bug that we've already fixed. I'm doing this mostly because your link mentions a .rpm file (which I can't open), no source code, and it's over five years old.

In the event that this is still an issue for you on a newer xserver, can you provide a pointer to the source code of the tests?

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.