[regression] after recent upgrade to 6.7.192-4ubuntu1 dual-head support is broken

[Herbert V. Riedel]

Binary package hint: xserver-xorg-video-ati

after the recent xserver-xorg-video-ati package upgrade few days ago, my dual-head setup broke;
I got a

0000:00:10.0 VGA compatible controller: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10]

in my notebook (internal panel 1440x960), with an external DVI port to which I attach an external 1600x1200 panel in mergedfb configuration; this worked fine with version 6.6.193-1ubuntu1 (I assume this was the previous version, since that's the only one I could find in the apt repository) of xserver-xorg-video-ati;

but now it doesn't properly work anymore, the external display becomes black and the desktop doesn't extend to it anymore;

see attached logfiles for more

[Tormod Volden (tormodvolden)]

mergedfb has been deprecated in favour of xrandr. Please see http://www.intellinuxgraphics.org/dualhead.html for how to set it up. And yes this needs to be well documented, in the release notes in particular.

[Herbert V. Riedel (hvr)]

I'll give that guide a try and report back asap whether I can get back my dual-head functionality

[Herbert V. Riedel (hvr)]

I was able to restore dual-head operation with the new driver by using xrandr, although it misdetected the DVI attached panel as a VGA type panel, but setting

Option "MacModel" "powerbook-duallink"

helped to fix that; but now I'm experiencing other problems with compiz and also in plain 2d mode which I didn't have with the previous setup:

1. when using compiz it can freeze the Xserver and the screen doesn't update anymore and I can't switch VT, but the system is still running, as I can ssh into it and kill the Xserver manually

2. when not using compiz, I get graphics corruption when (non-opaque) moving windows, horizontal stripes from the background or elsewhere get mixed in and only get removed again when the X11 app redraws its window

3. when starting the Xserver without any session (i.e. just "/usr/bin/X :0" or even just "Xnest :1") I get a pointer icon resembling '<>' instead of the usual '><' cross..

[Bryce Harrington (bryce)]

Thanks for the testing feedback Herbert.

I see in the changelog for 6.7.193 a fix for the mac issue:

Sjoerd Simons (1):
      radeon: Fix Option "MacModel".

I wonder if that addresses the issue you saw. I'm sure we'll be getting that fix backported, although probably not before gutsy beta is released. We should verify it fixes the issue before closing this bug.

Regarding mergedFB, what Tormod said is correct - mergedfb is deprecated in favor of xrandr. I'll see if I can find how to add stuff to Gutsy release notes and add something about that.

For your additional 3 bugs, these probably should be entered as separate new bug reports. #1 sounds like other compiz issues people have reported - hard to say though without additional detail. #2 sounds like something that should be reported upstream - please include a photo. #3 sounds really weird; does that happen only with -ati or also if you use -vesa? In each of these reports, please make sure to include exact steps to reproduce, copies of your xorg.conf, Xorg.0.log after recreating the issue, output from lspci -vvnn, and a photo if appropriate.

[Tormod Volden (tormodvolden)]

(3) is a known issue, I read about it the other day but I can not remember where, in an upstream bug report or on IRC. It had to do with big-endian mismatch etc.

[Herbert V. Riedel (hvr)]

I'm still trying to get more information regarding #1 and #2, especially since they don't seem to occur deterministically

regarding #3, I can workaround it by enabling option "SWCursor"; but that causes nasty square artifacts around the pointer;
afaik, using VESA is not an option on powerbooks, but I've attached the lspci -vvnn output for the mean time

btw, while playing around with radeon options I found out that using
 Option "DepthBits" "16"
in combination with a screen depth of 24 is a bad idea... it causes major disturbances on the whole display as soon as I start glxgears... ;-)

as for the "MacModel" fix, in case you are referring to

$ git-diff-tree -p d49a60bc939c9f3e9d57c23e44263f5bb52f09fb
d49a60bc939c9f3e9d57c23e44263f5bb52f09fb
diff --git a/src/radeon_driver.c b/src/radeon_driver.c
index 0d12168..9472f93 100644
--- a/src/radeon_driver.c
+++ b/src/radeon_driver.c
@@ -189,7 +189,7 @@ static const OptionInfoRec RADEONOptions[] = {
     { OPTION_DEFAULT_CONNECTOR_TABLE, "DefaultConnectorTable", OPTV_BOOLEAN, {0}, FALSE },
     { OPTION_DEFAULT_TMDS_PLL, "DefaultTMDSPLL", OPTV_BOOLEAN, {0}, FALSE },
 #if defined(__powerpc__)
- { OPTION_CONNECTORTABLE, "MacModel", OPTV_STRING, {0}, FALSE },
+ { OPTION_MAC_MODEL, "MacModel", OPTV_STRING, {0}, FALSE },
 #endif
     { -1, NULL, OPTV_NONE, {0}, FALSE }
 };

that's already part of the current gutsy xorg-xserver-video-ati 6.7.193-1ubuntu1; alas the radeon driver still cannot initialize the external TMDS, thus I need to have the DVI panel attached while booting up, so that openfirmware can initialize it for the radeon driver...

[Herbert V. Riedel (hvr)]

as for the big-endian hint; I just stumbled over

Commit: 771c37fcac8104b894120cc3fc0154b31b0318d1
Author: Alex Deucher <email address hidden> Mon, 25 Jun 2007 17:53:26 -0400

    RADEON: attempt to fix cursor on big endian (only tested on x86)

[Bryce Harrington (bryce)]

6.7.194 has been added to Gutsy - can you update and retest, and report if the issue still exists? I assume the aforementioned commit has been included in this version.

[Herbert V. Riedel (hvr)]

jfyi,

(1) is still an issue, I'm trying to hunt it down over the next weeks as it's really annoying... and I'm wondering if I'm the only having problems with compiz+RV350M10...

(2) seems to have gone

(3) is still an issue, I've gone upstream with it, and there's a fix for it:
https://bugs.freedesktop.org/show_bug.cgi?id=12672
https://bugs.freedesktop.org/show_bug.cgi?id=11796

[Dragonmaster Lou (lou-techhouse)]

Does XRandR support separate desktops on separate monitors, or does it create only one big screen that spans both monitors?

[Bryce Harrington (bryce)]

one big screen

[Tormod Volden (tormodvolden)]

> (3) is a known issue, I read about it the other day but I can not remember where, in an upstream bug report or on IRC. It had to do with big-endian mismatch etc.

That one is https://bugs.freedesktop.org/show_bug.cgi?id=11796

[Bryce Harrington (bryce)]

Herbert, I assume this is still a problem with Gutsy-final?

Hardy Alpha-1 will be released later this week with the latest -ati. Once that's out, if you could, please test that. We might be able to provide a backport of that driver if it solves this issue.

[Tormod Volden (tormodvolden)]

Please also try:
 Option "BusType" "PCI"

You can also try reducing the Virtual size to the minimum you need.

[Henrik Nilsen Omma (henrik)]

This bug was nominated for Gutsy but does currently not qualify for a 7.10 stable release update (SRU) and the nomination is therefore declined. According the the SRU policy, the fix should already be deployed and tested in the current development version before an update to the stable releases will be considered. With 7.10 now released, that policy applies to this bug. See: https://wiki.ubuntu.com/StableReleaseUpdates .

The bug is not being closed as work will continue on fixing it for the next release, Hardy Heron (8.04). To help improve the state of this bug see: https://wiki.ubuntu.com/Bugs/HowToTriage and https://wiki.ubuntu.com/DebuggingProcedures .

[Herbert V. Riedel (hvr)]

jfyi, I've tried gutsy from todays apt repository, and made the following observations so far (i.e. might be incomplete):

the xserver-xorg-video-ati 1:6.7.197-1ubuntu1 on a powerpc mac mini (i.e. not my powerbook, which is my production environment I can't sacrifice right now yet for testing :-), which has the following ati card

0000:00:10.0 VGA compatible controller: ATI Technologies Inc RV280 [Radeon 9200] (rev 01) (prog-if 00 [VGA controller])
        Subsystem: ATI Technologies Inc RV280 [Radeon 9200]
        Flags: bus master, 66MHz, medium devsel, latency 255, IRQ 48
        Memory at 98000000 (32-bit, prefetchable) [size=128M]
        I/O ports at 0400 [size=256]
        Memory at 90000000 (32-bit, non-prefetchable) [size=64K]
        Expansion ROM at f1000000 [size=128K]
        Capabilities: [58] AGP version 2.0
        Capabilities: [50] Power Management version 2

1) freezes the machine hard when starting the X server; compiling my own ati driver from git works

2) the big-endian mismatch caused 1-bpp cursor issue is still there

[Herbert V. Riedel (hvr)]

jfyi, the 1-bpp cursor issue was filed as bug #141118

[Tormod Volden (tormodvolden)]

> freezes the machine hard when starting the X server; compiling my own ati driver from git works

ati driver 6.8.0 is in Hardy now, and I posted a Gutsy build on https://wiki.ubuntu.com/XorgOnTheEdge. Does it work ok?

[Herbert V. Riedel (hvr)]

On Tue, 2008-02-26 at 11:20 +0000, Tormod Volden wrote:
> > freezes the machine hard when starting the X server; compiling my own
> > ati driver from git works

> ati driver 6.8.0 is in Hardy now, and I posted a Gutsy build on
> https://wiki.ubuntu.com/XorgOnTheEdge. Does it work ok?

...is it only me, or are there really no powerpc builds to test
with? :-)

[Herbert V. Riedel (hvr)]

however, I can confirm that Hardy's xserver-xorg-video-ati 1:6.8.0-1
seems to work fine on the mac mini;

as for the powerbook (which still has Gutsy), these days I need to
enable the following #if-conditional (on current git HEAD), in order to
have a more or less usable DVI-out:

diff --git a/src/legacy_output.c b/src/legacy_output.c
index 0de13df..bdac24d 100644
--- a/src/legacy_output.c
+++ b/src/legacy_output.c
@@ -278,7 +278,7 @@ RADEONRestoreDVOChip(ScrnInfoPtr pScrn, xf86OutputPtr output)
   RADEONDVOWriteByte(radeon_output->DVOChip, 0x0c, 0x89);
   RADEONDVOWriteByte(radeon_output->DVOChip, 0x08, 0x3b);
   break;
-#if 0
+#if 1
   /* needs work see bug 10418 */
      case RADEON_SIL_1178:
   RADEONDVOWriteByte(radeon_output->DVOChip, 0x0f, 0x44);

[Tormod Volden (tormodvolden)]

> ...is it only me, or are there really no powerpc builds to test with? :-)

You're right, the PPA doesn't build for powerpc unfortunately (due to powerpc being degraded to only community-supported, I guess).

> as for the powerbook (which still has Gutsy), these days I need to
> enable the following #if-conditional (on current git HEAD), in order to
> have a more or less usable DVI-out:

If they fix it upstream we'll get it into Ubuntu is probably all we can promise here. As long as it is not straightened out in upstream git yet, I am not sure we'll have Ubuntu resources to maintain it on our own. Thanks for all your contributions upstream in all cases.

[Brian Murray (brian-murray)]

Thanks for taking the time to report this bug and helping to make Ubuntu better. In the development cycle for Intrepid there have been some vast improvements in the open source ati video driver and we could use your help testing them. Could you please download the latest Alpha CD image of Intrepid and test this particular bug just using the Live CD? You can find the latest image at http://www.ubuntu.com/testing . Your testing can help make Ubuntu and the open source ati driver even better! Thanks in advance.

[Andrew Janke (a.janke)]

Hi Brian,

Thanks for the tip.

I just downloaded and had a bash.. No joy from the live CD, (tried
fglrx as well) so I went for broke, shrunk an existing partition and
installed intrepid.

After a bit of stuffing about (mainly looking at all the new
"pretties" in intrepid!) I can report resounding success! thanks and
kudos to all who have been working on the ati/radeon driver.

I first tried the Monitor Resolution Settings thingo but I couldn't
figure out how to make it detect things so I went back to my old
tricks fiddling with xrandr with a bog-stock xorg.conf.

My path to nirvana:

Add this to the "Screen" section in xorg.conf: (I run a 19" VGA + 21"
DVI from a docking station on a T42 Thinkpad)

        SubSection "Display"
  Modes "1400x1050"
  Virtual 2880 1200
 EndSubSection

# xrandr
<lots of stuff>

<think think>

# xrandr --output LVDS --off
# xrandr --output VGA-0 --mode 1280x1024
# xrandr --output DVI-0 --mode 1600x1200 --pos 1280x0

And it WORKED! I am most most most chuffed, no more Hardy with
backported feisty ati build for this little black duck.
Of course desktop effects dont work and there is still the small
annoying green flicker on very dark images on the DVI monitor but from
all I have read this is a hardware problem with this card. Something
to do with the DVI port on the dock not really being supposed to be
used for games like this.

I then went back to the "Monitor Resolution Settings" thingo and was
astounded to see the attached. I would now suspect that I could have
done this point and click but am not quite game to try just yet.. :)

Let me know if you want me to add all this info to the bugtracker
thing but suffice to say I would count this one as stomped, squished,
mashed up and obliterated. The only problem of course now is that I
have no reason to buy a T61 with the nvidia card in it...

Andrew

2008/8/15 Brian Murray <email address hidden>:
> Thanks for taking the time to report this bug and helping to make Ubuntu
> better. In the development cycle for Intrepid there have been some vast
> improvements in the open source ati video driver and we could use your
> help testing them. Could you please download the latest Alpha CD image
> of Intrepid and test this particular bug just using the Live CD? You
> can find the latest image at http://www.ubuntu.com/testing . Your
> testing can help make Ubuntu and the open source ati driver even better!
> Thanks in advance.
>
> ** Changed in: xserver-xorg-video-ati (Ubuntu)
> Status: Triaged => Incomplete
>
> --
> [regression] after recent upgrade to 6.7.192-4ubuntu1 dual-head support is broken
> https://bugs.launchpad.net/bugs/141161
> You received this bug notification because you are a direct subscriber
> of a duplicate bug.
>

--
Andrew Janke - <email address hidden>
Department of Geriatric Medicine, ANU
(<email address hidden> || http://a.janke.googlepages.com/)
Canberra->Australia [phone]

[Bryce Harrington (bryce)]

Andrew, great news - however your hardware isn't the same as Herbert's so whatever issue you were seeing is probably an unrelated one. So we'll leave this bug awaiting testing from Herbert.

[Bryce Harrington (bryce)]

Hi, it's been some time since we last heard from you - are you still interested in this bug? If so, it'd be great if you could re-test it against intrepid, and provide the info mentioned above. Otherwise, we'll close this as expired.

[Dragonmaster Lou (lou-techhouse)]

Closing it should be fine.

Bryce Harrington wrote:
> Hi, it's been some time since we last heard from you - are you still
> interested in this bug? If so, it'd be great if you could re-test it
> against intrepid, and provide the info mentioned above. Otherwise,
> we'll close this as expired.
>

--
<email address hidden> | "Dragonmaster Lou" | Tech House Alum
   "Searching for a distant star, heading off to Iscandar, leaving all
    we love behind, who knows what dangers we'll find..."

[Andrew Janke (a.janke)]

On Wed, Oct 1, 2008 at 4:15 AM, Bryce Harrington
<email address hidden> wrote:
> Hi, it's been some time since we last heard from you - are you still
> interested in this bug? If so, it'd be great if you could re-test it
> against intrepid, and provide the info mentioned above. Otherwise,
> we'll close this as expired.

Still works fine for me. (with xrandr)

--
Andrew Janke
(<email address hidden> || http://a.janke.googlepages.com/)
Canberra->Australia [phone]

[Bryce Harrington (bryce)]

Great, thanks for letting us know the issue is resolved.