115452 - Graphic glitches on scrollbar with S3 Savage (IBM T22) + cursor/bitmap accel

Status: RESOLVED FIXED

(Core :: XUL, defect)

Description

[Olivier Cahagne]

Build ID: 2001121508 on Win2k.

This bug has happened since I switched my laptop for an IBM laptop T22 w/ S3
Savage/IX.
Steps to reproduce bug:
1. Load any URL,
2. Scroll window with keypad or mouse,
3. graphical glitches appear on scrollbar,
4. same glitches for every scrollbar, for eg. inside a textarea.
(5. the same vertical stripes also appear on the location bar but it's not 100%
reproducible).

These glitches only show on Mozilla/GTK/XUL elements, not in the rendering
window, that's why I chose XUL component, I don't know if it's the right one though.
I've been experiencing this bug ever since I had this IBM laptop, beginning of
November. The bug is consistent with every build.

I first thought of driver issues, but upgrading to the latest stable drivers
from S3 gives me the same issue.

Here're more facts from my setup:
Win2k SP2
S3 Savage/IX 8MB
Bios 4.02.50.64
Driver 5.12.1.7062

Problem is fixed if I set acceleration to the 4th step in the Display Properties
-> Settings -> Advanced -> Troubleshooting -> set accel to 4th (one before full
accel) which disables cursor and bitmap accelerations.

Didn't find dupes, there're already ATI driver issues though in some bug
reports: bug 90657, bug 108070 (where the screenshot shows nearly the same
problem, but not exactly).

Comment 1

[Olivier Cahagne]

Attachment: Graphic glitches on scrollbar and personal toolbar

Comment 2

[Warner Young]

Olivier, I work at S3.  I ran across another bug from someone else reporting
this issue before, but was unable to duplicate it.  I'll forward this bug report
to the testing department at S3, though, because your comment about graphics
acceleration makes this sound like a driver problem.  Having the screenshot
helps.  I don't know how soon I'll get a response on this, but wanted to let you
know someone at S3 is aware of it now.

Comment 3

[Olivier Cahagne]

Warner, thanks for your support. I was quite surprised with this bug because my
previous laptop (Toshiba Tecra 8100) had a S3 Savage/MX chipset and never showed
this problem.
I'll try to see if other people in my company have this problem with their IBM
T22 or T23.

Comment 4

[Olivier Cahagne]

Bug does not occur on IBM T23 with S3 SuperSavage/IX (IXC ?), 16MB, Bios
5.2E.50.12 and driver from 7/18/2001 (no version number) on Win2k.

Comment 5

[Mike Kaply [:mkaply]]

I definitely have this problem on my T20.

Comment 6

[Olivier Cahagne]

Michael, is this with OS/2 or some Windows flavour ?

Comment 7

[Mike Kaply [:mkaply]]

Sorry. This is Win2K. 

Driver is S3 Savage/IX

version 5.12.1.7065 - 5/11/2001

Comment 8

[Olivier Cahagne]

*** Bug 138445 has been marked as a duplicate of this bug. ***

Comment 9

[Olivier Cahagne]

*** Bug 101907 has been marked as a duplicate of this bug. ***

Comment 10

[Olivier Cahagne]

Warner, any news about this, I've identified two other reproducible graphic
glitches with a S3 Savage IX and various web sites that disappear when
desactivating cursor & bitmap acceleration, see bug 101907 & bug 138445.

Comment 11

[Olivier Cahagne]

Testcases are bug 101907 and bug 138445.

Comment 12

[Warner Young]

Olivier, sorry, but neither our SQA team nor I can reproduce this issue for some
reason, and I've really tried.  There is something different between our test
environments and yours, but I can't think what it could be.  I'm trying to see
if we can get our hands on a Thinkpad T22, but so far no luck.

Comment 13

[Mike Kaply [:mkaply]]

I'm not sure what having a Thinkpad would do. It's obviously a video driver 
issue. Has anyone reported it to S3?

Comment 14

[Warner Young]

Michael, I work for S3.  I've been unofficially (because I'm not in SQA or
driver team) trying to get this bug resolved.  I've tried to reproduce this on
several of our SuperSavage boards, but can't.  SQA has tried to reproduce using
one of the other Thinkpads, and can't.  So, yes, I agree it's a driver problem,
but it would really help if we could find a way to reproduce.

Comment 15

[Olivier Cahagne]

Warner, thanks for your help anyway.
Another testcase found seems to be bug 111176 comment 14. This person has it
solved when desactivating cursor & bitmap acceleration. javier, can you detail
your configuration: are you running a Savage 3D on an IBM Thinkpad ? What driver
version  and OS ?

Comment 16

[Mike Kaply [:mkaply]]

I am attempting to get ThinkPad development involved so we can get a loaner 
thinkpad sent.

If that doesn't work, I'll see if my hardware team can send a loaner thinkpad 
to S3.

Comment 17

[Olivier Cahagne]

If there's any way I can show this somewhere in EMEA to S3 (do you have any
office in EMEA ?), I'd be glad to. I happen to travel a lot in EMEA and have
several IBM thinkpads (for demo/event purposes) but cannot involve more on a
business level as Mozilla is not business-related in my case (I do not work for
IBM nor S3).

Comment 18

[javier]

Hello all,
No Olivier, I haven´t an IBM Thinkpad.I have a desktop computer, sorry.
This is my configuration:

(Card APAC/ATREND)
Chip type:      Savage3D/MV (Engineering Release), rev B 
Bios version:   2.1B.17
Driver version: 4.11.01.4005-6.13.28 (s3_6.drv)
ICD version:    4.11.01.4005-6.13.25 (s3gicd.dll)
Metal version:  1,0,2,3 (metal.dll)

Cpu  type : AuthenticAMD K6-2
OS version: Windows 98(4.10A build 2222)
DirectX version: 4.08.01.0881 (DirectX 8.1)
DxMedia version: 6.03.01.0881

Comment 19

[Warner Young]

Javier, I must have missed a previous comment.  Are you having the same type of
problem?  One problem is that the Savage3D is *old*, and there is no more
support for it.  I believe 6.13.28 is already the newest driver there is.

Comment 20

[Olivier Cahagne]

Marking FIXED as driver 7.40.54 solved the issue. They are soon to be available
on S3 web site https://archive:[e-mail]/290-298drv/ (and
then on ibm.com).

I'll reopen the other bug reports: bug 101907 and bug 138445 as this new driver
didn't solve those issues.

Thanks to Warner and S3 for the support !

Comment 21

[Robin Monks]

*** Bug 139271 has been marked as a duplicate of this bug. ***