Re: fb layer & ioremap_wc

On Monday, May 05, 2008 11:59 am Frans Pop wrote:

On Monday 05 May 2008, you wrote:

quoted

On Monday, May 05, 2008 10:32 am Frans Pop wrote:

quoted

I suspect it could be vesafb/fbcon related instead. Normally I boot my
system with 'quiet vga=791', i.e. with vesafb. I then see the
artifacts.

When I boot without 'vga=791', I hit another, unrelated regression
(which I'll report separately) [1].

When I boot with 'video=vfb:off', I do _not_ get the artifacts when X
exits.

Ahhh, I missed that part of your config.  That could definitely have an
effect on things...  You'll probably want something like the attached

Not sure what to make of this.
With your patch I still get the artifacts, but they are displayed a lot
shorter. I get only a flash instead of a-2 seconds.

Hm, so I guess some other user is still using UC on the region.  [Note to fb 
list:  the patch just made vesafb use ioremap_wc instead of ioremap).

quoted

(there are other places in the fb layer that want similar treatment,
iirc, maybe fb_pgprotect?).

Do you want to try a more complete review/patch or should this be punted
over to the framebuffer folks?

Yeah, those bits have changed since I last looked at them, hopefully the fb 
guys can help.

Jesse

FYI

This is from a thread started on lkml about a regression with 2.6.26-rc1 
where artifacts are shown when X is exited if X86_PAT is enabled. 

The full thread can be seen here:
http://www.gossamer-threads.com/lists/linux/kernel/914812?do=post_view_threaded

The most relevant part starts here:
http://www.gossamer-threads.com/lists/linux/kernel/915956?do=post_view_threaded

On Monday 05 May 2008, Jesse Barnes wrote:

On Monday, May 05, 2008 11:59 am Frans Pop wrote:

quoted

On Monday 05 May 2008, you wrote:

quoted

On Monday, May 05, 2008 10:32 am Frans Pop wrote:

quoted

I suspect it could be vesafb/fbcon related instead. Normally I boot
my system with 'quiet vga=791', i.e. with vesafb. I then see the
artifacts.

When I boot with 'video=vfb:off', I do _not_ get the artifacts when
X exits.

Ahhh, I missed that part of your config.  That could definitely have
an effect on things...  You'll probably want something like the
attached

Not sure what to make of this.
With your patch I still get the artifacts, but they are displayed a lot
shorter. I get only a flash instead of a-2 seconds.

Hm, so I guess some other user is still using UC on the region.  [Note to
fb list:  the patch just made vesafb use ioremap_wc instead of ioremap).

quoted

quoted

(there are other places in the fb layer that want similar treatment,
iirc, maybe fb_pgprotect?).

Do you want to try a more complete review/patch or should this be
punted over to the framebuffer folks?

Yeah, those bits have changed since I last looked at them, hopefully the
fb guys can help.

-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
Don't miss this year's exciting event. There's still time to save $100. 
Use priority code J8TL2D2. 
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone

On Monday 05 May 2008, Jesse Barnes wrote:

On Monday, May 05, 2008 11:59 am Frans Pop wrote:

quoted

On Monday 05 May 2008, you wrote:

quoted

On Monday, May 05, 2008 10:32 am Frans Pop wrote:

quoted

I suspect it could be vesafb/fbcon related instead. Normally I
boot my system with 'quiet vga=791', i.e. with vesafb. I then see
the artifacts.

When I boot with 'video=vfb:off', I do _not_ get the artifacts
when X exits.

Ahhh, I missed that part of your config.  That could definitely
have an effect on things...  You'll probably want something like
the attached

Not sure what to make of this.
With your patch I still get the artifacts, but they are displayed a
lot shorter. I get only a flash instead of a-2 seconds.

Hm, so I guess some other user is still using UC on the region.  [Note
to fb list:  the patch just made vesafb use ioremap_wc instead of
ioremap).

AFAIK there has been no progress on this issue, which is now listed on the 
regression list for .26 [1]. As there has been no response from any fb 
devs, is someone else maybe willing to have a closer look at this?

I'm still seeing the artifacts with -rc6 and the severity seems to depend 
on what was last displayed: sometimes the whole display is covered with 
colored nonsense, artistic maybe but not desired...

Cheers,
FJP

[1] http://bugzilla.kernel.org/show_bug.cgi?id=10843

-------------------------------------------------------------------------
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
http://sourceforge.net/services/buy/index.php