[R423] gamma bug, display too bright when using DVI-VGA adapter [patch]

[raido357]

Binary package hint: xserver-xorg-video-ati

Ubuntu Lucid Lynx radeon driver has a bug, which causes gamma/contrast/brightness settings to be messed up.

http://ubuntuforums.org/showthread.php?t=1438810

Display is okay with Karmic Koala kernel 2.6.31-20.

Using xgamma, to lower gamma makes it worse.

ProblemType: Bug
Architecture: i386
Date: Fri Mar 26 12:50:34 2010
DistroRelease: Ubuntu 10.04
DkmsStatus: Error: [Errno 2] No such file or directory
Package: xserver-xorg-video-ati 1:6.12.191-1ubuntu2
ProcCmdLine: BOOT_IMAGE=/vmlinuz-2.6.32-17-generic root=UUID=f3621de2-2969-49d4-bdd3-882dee6a5df0 ro quiet splash
ProcEnviron:
 LANG=et_EE.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.32-17.26-generic 2.6.32.10+drm33.1
SourcePackage: xserver-xorg-video-ati
Uname: Linux 2.6.32-17-generic i686
dmi.bios.date: 06/05/2008
dmi.bios.vendor: Award Software International, Inc.
dmi.bios.version: F7c
dmi.board.name: 945P-S3
dmi.board.vendor: Gigabyte Technology Co., Ltd.
dmi.board.version: x.x
dmi.chassis.type: 3
dmi.modalias: dmi:bvnAwardSoftwareInternational,Inc.:bvrF7c:bd06/05/2008:svn:pn:pvr:rvnGigabyteTechnologyCo.,Ltd.:rn945P-S3:rvrx.x:cvn:ct3:cvr:
system:
 distro: Ubuntu
 codename: lucid
 architecture: i686
 kernel: 2.6.32-17-generic

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation 82945G/GZ/P/PL Memory Controller Hub [8086:2770] (rev 02)
 Subsystem: Giga-byte Technology Device [1458:5000]
01:00.0 VGA compatible controller [0300]: ATI Technologies Inc R423 UK [Radeon X800SE (PCIE)] [1002:554b]
 Subsystem: Club-3D BV Device [196d:1085]

[Bryce Harrington (bryce)]

Please explain this in more detail. What are the exact steps you're using to reproduce this issue? Saying it's "messed up" isn't very clear. HOW is it messed up?

[xzero1 (kdogn)]

I am the one who started the original post. To clarify the bug, I have added camera photos using identical manual settings of the screens from Lucid (radeon) and karmic (flgrx). Note that all images are of the same bootchart file. The images are as follows:
2149: Lucid with default settings (1.0)
2150: Lucid with xgamma set to .6
2151: Karmic with default settings (1.0)
2152: karmic with xgamma set to .6

Notice the loss of shading and detail in the Lucid shots easily seen on the init and collector rows. Although the bug may seem trivial it does affect the usability of Lucid a good deal. Please consider this a serious bug.
For additional information I use a desktop pc with an xfx radeon hd 4770 and additional info can be seen in the bootchart photos.

[raido357 (raido357)]

Please explain this in more detail. What are the exact steps you're using to reproduce this issue? Saying it's "messed up" isn't very clear. HOW is it messed up?

---

Too bright, text boxs disappear on forms, it's like sun shining directly on display. Just hard to see things.
Menu bars or all melted to one, because of brightness.

2149 picture of previous post, is exactly the same problem that i have.

There are no steps to reproduce, just upgraded Karmic Koala -> Lucid Lynx.

[Ed Coles (edicoles)]

I have the same problem. The screen has just got really bright for no reason, and there seems to be no way to control it using software. It started happening after I updated recently. I'm using a Radeon HD 4350 on an AMD64 installation

[Ed Coles (edicoles)]

I've tried adjusting my monitor settings but it doesn't change anything. This is probably something to do with the software

[raido357 (raido357)]

I've found something.

My display has DVI and VGA inputs. I am using DVI->VGA adapter for ANALOG input for my monitor. The ANALOG input is too bright, like described above.

If switch to DVI, then colors/brightness/gamma are fine.

So the problem is with ANALOG output from VGA.

I tried with different DVI->VGA converter, same issue - too bright display.

I can't switch permanently to DVI, because I need DVI for other machine with HDMI->DVI adapter.

[Tim Champion (timchampion)]

Me too.

I agree that this issue should be considered serious. It 100% affects the usability of Ubuntu. I can (barely) cope with the issue for a short time, but it * IS * a deal-breaker if its not fixed in the Beta. I was also using 10.04 Alpha (1 or 2, don't recall), and I was NOT having this issue on the same computer & exact same hardware.

lspci says my video card is: ATI Technologies Inc Radeon RV100 QY [Radeon 7000/VE] [1002:5159]

I am also using the analog (d-sub) connector on the monitor.
My "exact steps to reproduce the issue" are to install 10.04 Beta1 on a desktop computer with an ATI video card that connects to an LCD monitor with a DVI -> VGA(D-sub) cable. Load any web page, and notice that the page is completely washed out. html color "lightgray" becomes pure white. Any other light-ish pastel color become white. Text box borders are not visible. No amount of adjustments to the display brightness and/or contrast help at all. "xgamma -gamma .6" makes it more bearable, but only marginally.

[Tim Champion (timchampion)]

I have attached a photo of 2 monitors, from 2 separate computers with the same HTML color chart, side-by-side. On the left is a Windows computer, and the right is the Ubuntu 10.04 Beta1 computer that is afflicted with this bug.

The colors on the Ubuntu page that shouldn't be white, but are displaying as such are:
Gainsboro
Lavender
LavenderBlush
LemonChiffon
LightCyan
LightGoldenRodYellow
LightGray

Notice that Lime and LimeGreen appear to be almost identical on the right, and very different on the left.

[Ed Coles (edicoles)]

I've found something.

My display has DVI and VGA inputs. I am using DVI->VGA adapter for ANALOG input for my monitor. The ANALOG input is too bright, like described above.

If switch to DVI, then colors/brightness/gamma are fine.

So the problem is with ANALOG output from VGA.

I tried with different DVI->VGA converter, same issue - too bright display.

I can't switch permanently to DVI, because I need DVI for other machine with HDMI->DVI adapter.
--------

That's pretty much the same for me. Only I have to use the DVI - VGA adaptor since my card only has DVI output and my monitor is VGA

[Bryce Harrington (bryce)]

No, it does not qualify as "serious" in comparison with other X bugs. It sounds like more than a cosmetic issue, though, so setting to medium.

[Madko (madko)]

Same problem here on HD3650. Just to mention that it's not a "cosmetic" issue, Xorg display is really unusable, black text on white background is unreadable, text on button too, so you can't use the computer...

[Carl Hetherington (cth-carlh)]

I think I'm seeing this same problem on a Radeon HD3450. I have two identical monitors, both with VGA inputs. One is connected to the card's VGA output, and looks fine. The other is connected to the card's DVI output via an adapter, and looks washed out in the sense described by other reporters. This has happened during the upgrade to Lucid.

[Carl Hetherington (cth-carlh)]

Problems disappear when using fglrx, but then you lose compiz (and free-ness!)

[Bryce Harrington (bryce)]

Forwarding this bug from Ubuntu reporter raido357:
http://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/548709

[Problem]
Several reporters confirm that the screen is too bright after upgrading from karmic.

[Original Description]
Ubuntu Lucid Lynx radeon driver has a bug, which causes gamma/contrast/brightness settings to be messed up. Too bright, text boxs disappear on forms, it's like sun shining directly on display. Just hard to see things. Menu bars or all melted to one, because of brightness.

http://ubuntuforums.org/showthread.php?t=1438810

Display is okay with Karmic Koala kernel 2.6.31-20.

Using xgamma, to lower gamma makes it worse.

My display has DVI and VGA inputs. I am using DVI->VGA adapter for ANALOG input for my monitor. The ANALOG input is too bright, like described above.

If switch to DVI, then colors/brightness/gamma are fine.

So the problem is with ANALOG output from VGA.

I tried with different DVI->VGA converter, same issue - too bright display.

I can't switch permanently to DVI, because I need DVI for other machine with HDMI->DVI adapter.

 To clarify the bug, I have added camera photos using identical manual settings of the screens from Lucid (radeon) and karmic (flgrx). Note that all images are of the same bootchart file. The images are as follows:
2149: Lucid with default settings (1.0)
2150: Lucid with xgamma set to .6
2151: Karmic with default settings (1.0)
2152: karmic with xgamma set to .6

Notice the loss of shading and detail in the Lucid shots easily seen on the init and collector rows. Although the bug may seem trivial it does affect the usability of Lucid a good deal. Please consider this a serious bug.
For additional information I use a desktop pc with an xfx radeon hd 4770 and additional info can be seen in the bootchart photos.

Another user with HD 4350 reports seeing the same behavior, also using a DVI-VGA adaptor. Another on RV100 says they see similar symptoms and are using the analog (d-sub) connector. A user on HD3450 sees the same, also with a DVI-VGA adapter.

Architecture: i386
Date: Fri Mar 26 12:50:34 2010
DistroRelease: Ubuntu 10.04
DkmsStatus: Error: [Errno 2] No such file or directory
Package: xserver-xorg-video-ati 1:6.12.191-1ubuntu2
ProcCmdLine: BOOT_IMAGE=/vmlinuz-2.6.32-17-generic root=UUID=f3621de2-2969-49d4-bdd3-882dee6a5df0 ro quiet splash
ProcEnviron:
 LANG=et_EE.UTF-8ProcVersionSignature: Ubuntu 2.6.32-17.26-generic 2.6.32.10+drm33.1
SourcePackage: xserver-xorg-video-ati
Uname: Linux 2.6.32-17-generic i686
dmi.bios.date: 06/05/2008
dmi.bios.vendor: Award Software International, Inc.
dmi.bios.version: F7c
dmi.board.name: 945P-S3
dmi.board.vendor: Gigabyte Technology Co., Ltd.
dmi.board.version: x.x
dmi.chassis.type: 3
dmi.modalias: dmi:bvnAwardSoftwareInternational,Inc.:bvrF7c:bd06/05/2008:svn:pn:pvr:rvnGigabyteTechnologyCo.,Ltd.:rn945P-S3:rvrx.x:cvn:ct3:cvr:
system: codename: lucid
 architecture: i686
 kernel: 2.6.32-17-generic

[Bryce Harrington (bryce)]

Created an attachment (id=34695)
XorgLog.txt

[Bryce Harrington (bryce)]

Created an attachment (id=34696)
CurrentDmesg.txt

[Bryce Harrington (bryce)]

Created an attachment (id=34697)
BootDmesg.txt

[Bryce Harrington (bryce)]

Bugzilla won't let me attach the photos but here's a link: http://launchpadlibrarian.net/42068556/gammaPhotos.tar.gz

"""
I am the one who started the original post. To clarify the bug, I have added camera photos using identical manual settings of the screens from Lucid (radeon) and karmic (flgrx). Note that all images are of the same bootchart file. The images are as follows:
2149: Lucid with default settings (1.0)
2150: Lucid with xgamma set to .6
2151: Karmic with default settings (1.0)
2152: karmic with xgamma set to .6

Notice the loss of shading and detail in the Lucid shots easily seen on the init and collector rows. Although the bug may seem trivial it does affect the usability of Lucid a good deal. Please consider this a serious bug.
For additional information I use a desktop pc with an xfx radeon hd 4770 and additional info can be seen in the bootchart photos.
"""

[Bryce Harrington (bryce)]

I have attached a photo of 2 monitors, from 2 separate computers with the same HTML color chart, side-by-side. On the left is a Windows computer, and the right is the Ubuntu 10.04 Beta1 computer that is afflicted with this bug.

The colors on the Ubuntu page that shouldn't be white, but are displaying as such are:
Gainsboro
Lavender
LavenderBlush
LemonChiffon
LightCyan
LightGoldenRodYellow
LightGray

Notice that Lime and LimeGreen appear to be almost identical on the right, and very different on the left.

http://launchpadlibrarian.net/42412653/IMG_0016.jpg

[agd5f (agd5f)]

This is an issue with the TV DAC BG/ADJ values. Can you attach a copy of the problematic video bioses and in the case of the RV100 and R420, the output of "radeontool regmatch TV_DAC_CNTL"?

[Tim Champion (timchampion)]

Installing fglrx fixed my problem too, limits Visual Effect to 'None' by doing so, but who cares. Thanks for the work-around Carl! Don't know this post helps with the bug, except maybe that it confirms that the bug is with (or some how related to) the default Radeon driver.

[agd5f (agd5f)]

This patch:
http://marc.info/?l=dri-devel&m=127052658905865&w=2
fixes the issues with R4xx cards. For the RV100 and other older cards, I'll need to see the vbios and value of tv_dac_cntl as requested. For r6xx, I this patch should fix the issue:
http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=commitdiff;h=6a8a2d702b33c6ed5c789f21b4e89fdf221f01ca

[Bryce Harrington (bryce)]

Madko, read more carefully, I didn't say this was a cosmetic issue. There are other issues where X refuses to even start up, or crashes/locksup during use, and have no obvious way to work around it; those are serious issues. This is just a regular severity X bug. However given the number of people seeing similar issues, most of whom are using DVI-VGA adapters, that ups the priority on this a bit.

[Bryce Harrington (bryce)]

raido357 - I've forwarded this bug upstream to http://bugs.freedesktop.org/show_bug.cgi?id=27478 - please subscribe yourself to this bug, in case they need further information or wish you to test something.

Thanks everyone for providing such good debugging data on this.

[Tim Champion (timchampion)]

I have an older video card (lspci says)
ATI Technologies Inc Radeon RV100 QY [Radeon 7000/VE]

Here is the output of radeontool regmatch TV_DAC_CNTL:
TV_DAC_CNTL (088c) 0x07480043 (122159171)
MC: 0x00 0x00000000
MC: 0x01 0x00000000
MC: 0x02 0x00000000
MC: 0x03 0x00000000
MC: 0x04 0x00000000
MC: 0x05 0x00000000
MC: 0x06 0x00000000
MC: 0x07 0x00000000
MC: 0x08 0x00000000
MC: 0x09 0x00000000
MC: 0x0a 0x00000000
MC: 0x0b 0x00000000
MC: 0x0c 0x00000000
MC: 0x0d 0x00000000
MC: 0x0e 0x00000000
MC: 0x0f 0x00000000
MC: 0x10 0x00000000
MC: 0x11 0x00000000
MC: 0x12 0x00000000
MC: 0x13 0x00000000
MC: 0x14 0x00000000
MC: 0x15 0x00000000
MC: 0x16 0x00000000
MC: 0x17 0x00000000
MC: 0x18 0x00000000
MC: 0x19 0x00000000
MC: 0x1a 0x00000000
MC: 0x1b 0x00000000
MC: 0x1c 0x00000000
MC: 0x1d 0x00000000
MC: 0x1e 0x00000000
MC: 0x1f 0x00000000
MC: 0x20 0x00000000
MC: 0x21 0x00000000
MC: 0x22 0x00000000
MC: 0x23 0x00000000
MC: 0x24 0x00000000
MC: 0x25 0x00000000
MC: 0x26 0x00000000
MC: 0x27 0x00000000
MC: 0x28 0x00000000
MC: 0x29 0x00000000
MC: 0x2a 0x00000000
MC: 0x2b 0x00000000
MC: 0x2c 0x00000000
MC: 0x2d 0x00000000
MC: 0x2e 0x00000000

You also said you wanted to see the vbios, but I can figure out how to get that information. Is that something I need to look at during boot time, or is that something I can get any time?

Please let me know if I can do anything else to help. I am by no means an expert, but I can test stuff out if I'm given good enough instruction.

[agd5f (agd5f)]

(In reply to comment #8)
> I have an older video card (lspci says)
> ATI Technologies Inc Radeon RV100 QY [Radeon 7000/VE]
>
> Here is the output of radeontool regmatch TV_DAC_CNTL:
> TV_DAC_CNTL (088c) 0x07480043 (122159171)

that looks correct. Does:
radeontool regset TV_DAC_CNTL 0x07280043
help?

>
> You also said you wanted to see the vbios, but I can figure out how to get that
> information. Is that something I need to look at during boot time, or is that
> something I can get any time?
>

Any time.

> Please let me know if I can do anything else to help. I am by no means an
> expert, but I can test stuff out if I'm given good enough instruction.

You'll need the pci bus id from lspci, e.g.,
01:05.0 VGA compatible controller: ATI Technologies Inc RS690
^^^^^^^
It will have some leading zeros in sysfs.

cd /sys/bus/pci/devices/<pci bus id>
echo 1 > rom
cat rom > /tmp/vbios.rom
echo 0 > rom

[Tim Champion (timchampion)]

>that looks correct. Does:
>radeontool regset TV_DAC_CNTL 0x07280043
>help?

Nope, has no effect. - unless I need to restart X in order for it to take effect. (I did not restart X)

>You'll need the pci bus id from lspci, e.g.,
>01:05.0 VGA compatible controller: ATI Technologies Inc RS690
>^^^^^^^
>It will have some leading zeros in sysfs.
>
>cd /sys/bus/pci/devices/<pci bus id>
>echo 1 > rom
>cat rom > /tmp/vbios.rom
>echo 0 > rom

Ok, I feel inept here. I preformed the above as root, and I'm not sure what I should be posting as a result. do you want the contents of /sys/bus/pci/devices/0000:01:00.0/rom ?

I displayed /tmp/vbios.rom, and it appeared to be a binary file. As a shot-in-the-dark, I'll add /tmp/vbios.rom as an attachment in just a moment here.

[Tim Champion (timchampion)]

Created an attachment (id=34714)
vbios of rv100

[agd5f (agd5f)]

(In reply to comment #10)
> >that looks correct. Does:
> >radeontool regset TV_DAC_CNTL 0x07280043
> >help?
>
> Nope, has no effect. - unless I need to restart X in order for it to take
> effect. (I did not restart X)
>

No need to restart X. what port are you using for your problematic monitor? VGA? DVI?

> >You'll need the pci bus id from lspci, e.g.,
> >01:05.0 VGA compatible controller: ATI Technologies Inc RS690
> >^^^^^^^
> >It will have some leading zeros in sysfs.
> >
> >cd /sys/bus/pci/devices/<pci bus id>
> >echo 1 > rom
> >cat rom > /tmp/vbios.rom
> >echo 0 > rom
>
> Ok, I feel inept here. I preformed the above as root, and I'm not sure what I
> should be posting as a result. do you want the contents of
> /sys/bus/pci/devices/0000:01:00.0/rom ?

yes. It's a binary file.

[agd5f (agd5f)]

for the rv100, what connectors does your card actually have? Also, can you attach your xorg log and dmesg output? Perhaps it's the primary dac that's causing problems. Can you get the output of:
radeontool regmatch DAC_MACRO_CNTL

[Tim Champion (timchampion)]

Created an attachment (id=34715)
rv100_xorg_dmesg_vbios.tgz

My video card has a single connector that looks to be a dual link DVI M1-DA. This image matches the pins on my adapter: http://en.wikipedia.org/wiki/File:M1-DA.svg
I am using a cable that goes from that dual link above, to 2x VGA. I'm only using VGA #1 on this computer.

The output of radeontool regmatch DAC_MACRO_CNTL:
DAC_MACRO_CNTL (0d04) 0x00000000 (0)
(all the repetitive zeros that follow are excluded)

I am attaching an archive of
dmesg.txt
Xorg.0.log
vbios.rom

I not going to be able to respond here for a few hours, hope that is enough information for a while.

[agd5f (agd5f)]

(In reply to comment #14)
> Created an attachment (id=34715) [details]
> rv100_xorg_dmesg_vbios.tgz
>
> My video card has a single connector that looks to be a dual link DVI M1-DA.
> This image matches the pins on my adapter:
> http://en.wikipedia.org/wiki/File:M1-DA.svg
> I am using a cable that goes from that dual link above, to 2x VGA. I'm only
> using VGA #1 on this computer.

perfect. thanks!

>
> The output of radeontool regmatch DAC_MACRO_CNTL:
> DAC_MACRO_CNTL (0d04) 0x00000000 (0)
> (all the repetitive zeros that follow are excluded)

Does:
radeontool regset DAC_MACRO_CNTL 0x00000808
fix the issue?

[Tim Champion (timchampion)]

Ok, I lied, I'm still responding.

radeontool regset DAC_MACRO_CNTL 0x00000808
fixed my problem! It might look a little on the blue side, but that also could be just because I was so used to looking at it all washed out for so long that my eyes are playing tricks on me. HTML colors appear to be looking right, and html lightgray is no longer displaying as white.

[agd5f (agd5f)]

Created an attachment (id=34718)
dac adj fix for legacy chips

This patch should fix your rv100 (an other r1xx-r3xx cards with similar problems) issues.

[Dodgy (dodgy)]

I am using:

01:00.0 VGA compatible controller: ATI Technologies Inc RV620 LE [Radeon HD 3450]

and am still having these problems.

The patch mentioned above for R6xxx cards appears already to have been applied to the kernel that is shipping with Ubuntu Lucid.

I just tried to copy the video bios, but I get:

carl@houllier:/sys/bus/pci/devices/0000:01:00.0$ sudo echo 1 > rom
bash: rom: Permission denied

Let me know if there's any other information I can provide.

[agd5f (agd5f)]

(In reply to comment #18)
> I am using:
>
> 01:00.0 VGA compatible controller: ATI Technologies Inc RV620 LE [Radeon HD
> 3450]
>
> and am still having these problems.
>
> The patch mentioned above for R6xxx cards appears already to have been applied
> to the kernel that is shipping with Ubuntu Lucid.
>

Ok.

> I just tried to copy the video bios, but I get:
>
> carl@houllier:/sys/bus/pci/devices/0000:01:00.0$ sudo echo 1 > rom
> bash: rom: Permission denied
>
> Let me know if there's any other information I can provide.

You need to be root.

[Dodgy (dodgy)]

Created an attachment (id=34734)
ROM from RV620 card

[Dodgy (dodgy)]

> You need to be root.

Gah! Sorry. I thought sudo would be sufficient.

ROM attached.

[Bryce Harrington (bryce)]

I've verified that these two patches were already included in the kernel 2.6.32-17.26 this was reported against:
http://git.kernel.org/?p=linux/kernel/git/airlied/drm-2.6.git;a=commitdiff;h=6a8a2d702b33c6ed5c789f21b4e89fdf221f01ca
http://marc.info/?l=dri-devel&m=127052658905865&w=2

This patch was not present in that kernel, though:
https://bugs.freedesktop.org/attachment.cgi?id=34718

Thanks, I'll forward this bug to our kernel team for further handling.

[Kbkang (kbkang)]

(In reply to comment #44)
checked out but atom.c still is not there, by searching the whole filesystem there is no sign of this file so patch cannot be applied.

[agd5f (agd5f)]

You need to install the kernel source package. I don't know whether LL has this patch or not.

[Kbkang (kbkang)]

(In reply to comment #46)
tried to sudo apt-get install kernel-sources but i get an error, translated from italian it says:
"The package kernel-source doens't have available versions, but it is named by another package. This could mean that the package is missing, obsolete or it is only available inside another source."

[Kbkang (kbkang)]

(In reply to comment #46)
installed kernel source from package manager, still having bright colors and missing atom.c ....

[Kbkang (kbkang)]

(In reply to comment #48)
don't know if this is relevant:
- even as root i can't obtain the rom of my video card with instruction posted before.
- Atom.c does not appear anywhere in my system.
- i did any previous step using proprietary drivers but even now that i uninstalled them there is nothing different.
- radeontool instruction give always "mapping ctrl region" response.
- avivotool regmatch 0x7ef4 and avivotool regmatch 0x7ff4 give segmentation fault error.

looks like i got many strange things going on here !?!?

[agd5f (agd5f)]

(In reply to comment #49)
> (In reply to comment #48)
> don't know if this is relevant:
> - even as root i can't obtain the rom of my video card with instruction posted
> before.

replace <pci bus id> with the bus id of your video card (look it up with lspci).

> - Atom.c does not appear anywhere in my system.

You apparently don't have the kernel source installed then. Also it's atom.c (note the case).

> - i did any previous step using proprietary drivers but even now that i
> uninstalled them there is nothing different.
> - radeontool instruction give always "mapping ctrl region" response.
> - avivotool regmatch 0x7ef4 and avivotool regmatch 0x7ff4 give segmentation
> fault error.

You need to run avivotool as root. radeontool isn't useful for your card.

[Kbkang (kbkang)]

Created an attachment (id=35360)
my vcard rom

(In reply to comment #50)
here is the rom finally! thanks for the help, hope this is useful.

maybe this is not the right place to ask for it, but how can i install the kernel sources then? i did it from the package manager but apparently it didn't work.

and here is what happens with avivotool:

kbkang@kbuntu:~$ sudo -s
[sudo] password for kbkang:
root@kbuntu:~# avivotool regmatch 0x7ef4
mapping fb region
Segmentation fault
root@kbuntu:~# avivotool regmatch 0x7ff4
mapping fb region
Segmentation fault
root@kbuntu:~#

[Kbkang (kbkang)]

(In reply to comment #51)
well my pc got stuck and i dont know if it's becouse of something i did while trying to compile and install the new kernel, i reached the step where i had to make modules_install, and all was going just perfect. then i moved the bzimage in the boot directory and waited till this morning to edit the grub config file but as i woke up and turned on the pc: surprise! i will post as soon as i get my pc back to life!
thanks alex for all your patience!
thanks to everyone that posted here!
see you soon!

[Kbkang (kbkang)]

(In reply to comment #52)
well i'm back, don't know why but bios was upset with my usb keyboard, now i moved it to an external hub and everything's fine again.
as of now i'm running this kernel: 2.6.32.11+drm33.2, and i applyed the patch to the atom.c file before compiling and installing the sources. but now the driver is set to the opensource ATI radeon instead of the fglrx, trying to install the wich fails always and resulting in no change in colours but no acceleration at all. i'm running out of hope...

[Kbkang (kbkang)]

(In reply to comment #53)
back again. as i had no news since my last post maybe i was misunderstood. the problem is not solved and i still have very bright colours. is there anything else i can do? i will try once more reinstalling ubuntu, the kernel and the patch, but i doubt it will work... i'll give some feedback in a few days.

Thanks!

[Philip Nelson (philip-whatsbeef)]

The above kernel has fixed this issue for me.

My issue seemed to be slightly different, it was mostly very bright but depended on what was displayed on the screen. It seemed to be doing some sort of dynamic contrast - if what was being displayed was mostly white, the contrast would lower, if what was on the screen was darker it would go very bright. I have two monitors, both vga, one with a dvi-vga adaptor. It was only happening on the DVI-0 output, the VGA-0 output was fine.

[Emmanuel Touzery (emmanuel-touzery)]

Just curious.. How come this fix was not commited yet to the repository, and the bug closed? My understanding is that many users have confirmed that the fix is correct and does fix the problem, or am I missing something?

[for me i can't confirm, i didn't test the patch but many others have]

[Kbkang (kbkang)]

(In reply to comment #55)
I don't know, i keep retrying but the problem is still there...

[agd5f (agd5f)]

(In reply to comment #55)
> Just curious.. How come this fix was not commited yet to the repository, and
> the bug closed? My understanding is that many users have confirmed that the fix
> is correct and does fix the problem, or am I missing something?

The patches have been committed to the kernel.

[Matthew Caron (matt-mattcaron)]

The above kernel fixes this for me as well.

[Emmanuel Touzery (emmanuel-touzery)]

if i understand correctly, this update wasn't pushed to lucid-updates yet, right?
if it was, with which package version?

currently i turned KMS off to work around this problem, however i'd like to go back to KMS when it becomes possible (to have a nicer boot splash, not 640x480 but full resolution).

[Dj-jankins (dj-jankins)]

Maybe this bug old/obsolete, but I recently got new Radeon HD 4350, and can confirm this bug. Display is too bright with gamma = 1, lowering gamma do not solve problem because window decorations become dark, red becomes tomato, buttons on forms in browser or other apps are white (using gtk theme 'Mist'). How I can fix this?

[agd5f (agd5f)]

(In reply to comment #58)
> Maybe this bug old/obsolete, but I recently got new Radeon HD 4350, and can
> confirm this bug. Display is too bright with gamma = 1, lowering gamma do not
> solve problem because window decorations become dark, red becomes tomato,
> buttons on forms in browser or other apps are white (using gtk theme 'Mist').
> How I can fix this?

Make sure you have a new enough kernel with the patches mentioned in this bug 2.6.34 or newer should be ok.

[agd5f (agd5f)]

(In reply to comment #58)
> Maybe this bug old/obsolete, but I recently got new Radeon HD 4350, and can
> confirm this bug. Display is too bright with gamma = 1, lowering gamma do not
> solve problem because window decorations become dark, red becomes tomato,
> buttons on forms in browser or other apps are white (using gtk theme 'Mist').
> How I can fix this?

Make sure you have a new enough kernel with the patches mentioned in this bug. 2.6.34 or newer should be ok.

[Dj-jankins (dj-jankins)]

(In reply to comment #60)
> Make sure you have a new enough kernel with the patches mentioned in this bug.
> 2.6.34 or newer should be ok.

Yes, I have installed latest stable kernel 2.6.34. I use VGA adapter and UMS (KMS strangely freezes my X server with kernel panic when I shutdown/reboot my box. I will fix this later if possible).

[Dj-jankins (dj-jankins)]

(In reply to comment #61)

> Yes, I have installed latest stable kernel 2.6.34. I use VGA adapter and UMS
> (KMS strangely freezes my X server with kernel panic when I shutdown/reboot my
> box. I will fix this later if possible).

I fixed issue with KMS, but bug still visible, even with KMS.

[agd5f (agd5f)]

(In reply to comment #62)
> (In reply to comment #61)
>
> > Yes, I have installed latest stable kernel 2.6.34. I use VGA adapter and UMS
> > (KMS strangely freezes my X server with kernel panic when I shutdown/reboot my
> > box. I will fix this later if possible).
>
> I fixed issue with KMS, but bug still visible, even with KMS.

Please attach a copy of your vbios as per the instructions in comment 9. And attach the output of the following commands:
avivotool regmatch 0x7ef4
avivotool regmatch 0x7ff4

You can get avivotool here (part of the radeontool package):
http://cgit.freedesktop.org/~airlied/radeontool/

All commands must be run as root.

[Dj-jankins (dj-jankins)]

(In reply to comment #63)
>
> Please attach a copy of your vbios as per the instructions in comment 9. And
> attach the output of the following commands:
> avivotool regmatch 0x7ef4
> avivotool regmatch 0x7ff4
>
> You can get avivotool here (part of the radeontool package):
> http://cgit.freedesktop.org/~airlied/radeontool/
>
> All commands must be run as root.

Output of avivotool:
# avivotool regmatch 0x7ef4
0x7ef4 0x00221f00 (2236160)
# avivotool regmatch 0x7ff4
0x7ff4 0x00221f02 (2236162)

[Dj-jankins (dj-jankins)]

Created an attachment (id=36589)
vbios from Radeon HD 4350

rom vbios from Radeon HD 4350

[Volodymyr Shcherbyna (volodymyr)]

Just want to share my expirience about this problem.

I also met the same issue in Kubuntu. After installing fglrx the "bright-display" issue is fixed, however a new issues are introduced by fglrx. The whole UI is much slower with fglrx. Just as an example, when I click "Exit-Shutdown" the shutdown confirmation window appears after 5-6 seconds and xorg process is eating 100% of CPU. If my visual effects are enabled, and windows are transparent moving any window becomes a hell slow.

In fact, Kubuntu on virtual machine at my work place in vmware is running faster than the Kubuntu at home. And at home I have Core 2 Quad 2.5 Ghz with 4 Gb of RAM ...

[Brownout (brownout)]

@Volodymyr M. Shcherbyna you might be interested in https://bugs.launchpad.net/ubuntu/+source/fglrx-installer/+bug/351186

[Volodymyr Shcherbyna (volodymyr)]

Thanks, but I've noticed that the recent kernel fixes this bug. I am running 2.6.32-23-generic and I have no problem with default kubuntu video driver anymore. Moreover, the default video driver works perfectly: nothing is slow, all effects are working like they should be.

So my suggestion to those who met this problem right after installing Ubuntu or Kubuntu CD - do not install fglrx, just upgrade the system and the bug will disappear by itself. For the newbies: click on "I" icon on the very right bottom corner and agree to install updates.

To admins: you can post this suggestion in the head of the bug as a workaround so that people would not waste their time trying to find the solution?

Thanks.

[Nikolay Rysev (mad-f3ka)]

Created an attachment (id=36687)
ROM of RV530

I have the same isuue with 06:00.0 VGA compatible controller: ATI Technologies Inc RV530 [Radeon X1600] on kernel 2.6.34

[Dj-jankins (dj-jankins)]

(In reply to comment #66)
> Created an attachment (id=36687) [details]
> ROM of RV530
>
> I have the same isuue with 06:00.0 VGA compatible controller: ATI Technologies
> Inc RV530 [Radeon X1600] on kernel 2.6.34

I'm using same model now and it works fine.
lspci:
02:00.0 VGA compatible controller: ATI Technologies Inc RV530 [Radeon X1600]

[Dj-jankins (dj-jankins)]

> Please attach a copy of your vbios as per the instructions in comment 9. And
> attach the output of the following commands:
> avivotool regmatch 0x7ef4
> avivotool regmatch 0x7ff4
>
> You can get avivotool here (part of the radeontool package):
> http://cgit.freedesktop.org/~airlied/radeontool/
>
> All commands must be run as root.

>Output of avivotool:
># avivotool regmatch 0x7ef4
>0x7ef4 0x00221f00 (2236160)
># avivotool regmatch 0x7ff4
>0x7ff4 0x00221f02 (2236162)

>vbios from Radeon HD 4350

Please answer anything :| Yes or no.
I can't use new adapter because it hurts my eyes.

[Nikolay Rysev (mad-f3ka)]

Hmm… Looks like i have a hardware bug :( BIOS logo on POST is green, but should be red. Sorry.

[agd5f (agd5f)]

(In reply to comment #68)
> >Output of avivotool:
> ># avivotool regmatch 0x7ef4
> >0x7ef4 0x00221f00 (2236160)
> ># avivotool regmatch 0x7ff4
> >0x7ff4 0x00221f02 (2236162)

> Please answer anything :| Yes or no.
> I can't use new adapter because it hurts my eyes.

The values there look fine. Does:
avivotool regset 0x7ff4 0x00222002
help? or
avivotool regset 0x7ff4 0x00202002

[Dj-jankins (dj-jankins)]

(In reply to comment #70)
> (In reply to comment #68)
> > >Output of avivotool:
> > ># avivotool regmatch 0x7ef4
> > >0x7ef4 0x00221f00 (2236160)
> > ># avivotool regmatch 0x7ff4
> > >0x7ff4 0x00221f02 (2236162)
>
> > Please answer anything :| Yes or no.
> > I can't use new adapter because it hurts my eyes.
>
> The values there look fine. Does:
> avivotool regset 0x7ff4 0x00222002
> help? or
> avivotool regset 0x7ff4 0x00202002

No. This don't help.
I can compare the colors with my old video card and I'm sure - the problem appears when I boot with new Radeon HD adapter.

[Dj-jankins (dj-jankins)]

(In reply to comment #71)
>
> No. This don't help.
> I can compare the colors with my old video card and I'm sure - the problem
> appears when I boot with new Radeon HD adapter.

Same behavior when I booted Ubuntu 10.04 livecd. On HD 4350 screen brighter comparing with old X1600.

[Andy Extance (aextance)]

Should Kernel 2.6.35.1-candela also fix this? It seems to have catapulted me straight into a different problem, where rather than just appearing "washed out", when I try and use my laptop through a monitor, it doesn't boot at all.

[madbiologist (me-again)]

@Andy - that sounds like a different bug. Also I'm not sure whether you are using a Radeon or not. Please use ubuntu-bug to file a new bug report.

[Julian Wiedmann (jwiedmann)]

The patch ("more atom parser fixes (v2)") landed in Lucid's 2.6.32-23.37.

[ssi2011 (ssi-delta-elevator)]

The bug is still NOT fixed!

OS: Ubuntu 10.04
Kernel Version: 2.6.32-30-generic #59-Ubuntu SMP Tue Mar 1 21:30:21 UTC 2011 i686 GNU/Linux
VGA cable was used.

This bug makes MySQL Workbench 5.2.33 (newest version to date) EER Diagram grid invisible. This becomes very annoying when using diagram where grid should be light grey color, also default new layer becomes white, while it should be lanvender.

Loaded up color chart and it looks exactly like the missing color computer to the right in Bryce Harrington's picture. Luckly, I have DVI ports in both computer and monitor. Changed VGA to DVI cable, problem disappeared right away.

So the bug in the Lucid driver is still NOT fixed!

[Andrey Barrientos M. (cpoliticas2005)]

Unfortunately this problem is affecting me in Quantal Quetzal as well. Basically the behavior is the same as the other fellow users are experiencing: A too light-saturated Desktop Display that cannot be corrected using any gamma utility.

In the past I could fix this using the radeon.modeset=0 variable in the kernel boot, however since the upgrade, if I do that I don't have any display, it logins normally but no image whatsoever. So in order to boot in the system again I should use the radeon.modeset=1 variable again, with the undesirable color side effects. Nevertheless this seems to be affecting the Desktop only, because if I use any OpenGL program, like the Open Arena game, colors shows normal.

Alas I've tried some of the solutions suggested in this thread with no luck. I've attached the dmesg and lscpci logs to this report as the documentation indicates, however if you need more information in order to fix this issue, I would gladly do whatever it takes to see this fixed.

Some basic information for reference is I'm using Kubuntu 12.10 Quantal Quetzal and using an ATI Radeon R100 video card. Using the uname -a shows me this kernel info: Linux 3.5.0-17-generic #28-Ubuntu SMP Tue Oct 9 19:32:08 UTC 2012 i686 athlon i686 GNU/Linux

I appreciate any help given with this issue and I thank you in advance for taking the time to read this report and try to solve it.

Best regards,

-Andrey