[REGRESSION] i810fb: console switch causes pixel waste since 2.6.25-rc?

Hi there

in 2.6.25-rc4 and -rc5 every 5th to 10th console switching with Alt-Fn 
causes 'pixel waste' (see [1] for a picture - don't know how to describe 
better).
It doesn't happen with 2.6.24.3.

I tried a bisect on Linus' tree, but couldn't finish yet because I detected an 
intermediate state where no fb works at all. dmesg says
	agpgart: unable to get memory for graphics translation table
which causes i810fb to say
	i810fb_alloc_fbmem: cannot acquire agpgart
So all "bads" (except the first) in this git bisect log mean this broken 
agpgart.

git-bisect start
# bad: [29e8c3c304b62f31b799565c9ee85d42bd163f80] Linux 2.6.25-rc4
git-bisect bad 29e8c3c304b62f31b799565c9ee85d42bd163f80
# good: [49914084e797530d9baaf51df9eda77babc98fa8] Linux 2.6.24
git-bisect good 49914084e797530d9baaf51df9eda77babc98fa8
git-bisect bad bd45ac0c5daae35e7c71138172e63df5cf644cf6
git-bisect good 6bd48fcf73019219495f7599028296c65b749bb4
git-bisect good 746ef0cd0c7190d570c65b8e39a4ac67550ae43a
git-bisect bad 97896d04a14669b146c17d779b81ec7a339deeb3
git-bisect good 751752789162fde69474edfa15935d0a77c0bc17
git-bisect good 0a663088cd6ff9e89f285ae7689e6eee46cfb54c
git-bisect bad 51c6cf662b4b361a09fbd324f4c67875d9bcfbea

I will continue tomorrow figuring out since when agpgart is broken and since 
when it works again. Possibly that problem also causes this regression.

Distribution: Gentoo

System (Gigabyte GA-6OXM7):
00:00.0 Host bridge: Intel Corporation 82815 815 Chipset Host Bridge and 
Memory Controller Hub (rev 02)
00:02.0 VGA compatible controller: Intel Corporation 82815 CGC [Chipset 
Graphics Controller] (rev 02)
00:1e.0 PCI bridge: Intel Corporation 82801AA PCI Bridge (rev 02)
00:1f.0 ISA bridge: Intel Corporation 82801AA ISA Bridge (LPC) (rev 02)
00:1f.1 IDE interface: Intel Corporation 82801AA IDE Controller (rev 02)

Kernel config: See [2].

Boot options:
root=/dev/hda1 lapic 
video=i810fb:hsync1:30,hsync2:72,vsync1:50,vsync2:120,accel,mtrr,640x480-16@100


	Regards,
		Stefan

[1] 131 KB
<http://www-user.tu-chemnitz.de/~stefb/i810fb/pixel_waste.jpg>

[2] 32 KB
<http://www-user.tu-chemnitz.de/~stefb/i810fb/config>

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Am Sonntag, 16. März 2008 01:33 schrieb Stefan Bauer:

I tried a bisect on Linus' tree, but couldn't finish yet because I detected
an intermediate state where no fb works at all. dmesg says
	agpgart: unable to get memory for graphics translation table
which causes i810fb to say
	i810fb_alloc_fbmem: cannot acquire agpgart
[...]
I will continue tomorrow figuring out since when agpgart is broken and
since when it works again.

commit 266b9f8727976769e2ed2dad77ac9295f37e321e
Author: Thomas Gleixner [off-list ref]
Date:   Wed Jan 30 13:34:06 2008 +0100

    x86: fix ioremap RAM check
    
    Signed-off-by: Ingo Molnar [off-list ref]
    Signed-off-by: Thomas Gleixner [off-list ref]

...broke my agpgart.

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Am Sonntag, 16. März 2008 01:33 schrieb Stefan Bauer:

I will continue tomorrow figuring out since when agpgart is broken and
since when it works again. Possibly that problem also causes this
regression.

That wasn't the case. As I wrote, agpgart was broken since 266b9f8 (x86: fix 
ioremap RAM check), but works well again since 38cb47b (x86: relax RAM check 
in ioremap()).

So I restarted bisecting, finding this one:

commit eaa0ff15c30dc9799eb4d12660edb73aeb6d32c5
Author: Alexey Dobriyan [off-list ref]
Date:   Wed Feb 6 01:36:06 2008 -0800

    fix ! versus & precedence in various places

    Fix various instances of

        if (!expr & mask)

    which should probably have been

        if (!(expr & mask))

    Signed-off-by: Alexey Dobriyan [off-list ref]
    Cc: Jens Axboe [off-list ref]
    Cc: Peter Osterlund [off-list ref]
    Cc: Karsten Keil [off-list ref]
    Cc: Mauro Carvalho Chehab [off-list ref]
    Cc: "Antonino A. Daplas" [off-list ref]
    Cc: Mark Fasheh [off-list ref]
    Cc: "David S. Miller" [off-list ref]
    Cc: Jeff Garzik [off-list ref]
    Cc: Al Viro [off-list ref]
    Signed-off-by: Andrew Morton [off-list ref]
    Signed-off-by: Linus Torvalds [off-list ref]

So the regression was in since -rc1.

	Regards,
		Stefan

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/