Created attachment 28739 [details]
XorgLog.txt

Created attachment 28740 [details]
CurrentDmesg.txt

Created attachment 28741 [details]
BootDmesg.txt

Created attachment 28797 [details]
Font tearing screenshot

Added the screenshot of font tearing. With most of the options from the xorg.conf the removed this is less likely, but now different.

Section "Screen"
        Identifier      "Configured Screen Device"
        Device  "Configured Video Device"
        SubSection "Display"
                Virtual 3080 1050
        EndSubSection
EndSection

Section "Device"
        Identifier      "Configured Video Device"
        #Option         "EXANoComposite"                "true"
        #Option         "TripleBuffer"                 "true"
        #Option         "EnablePageFlip"                "true"
        #Option         "EnableDepthMoves"              "true"
        #Option         "MigrationHeuristic"            "smart"
        Option  "AccelMethod"                   "XAA"
        #Option         "RenderAccel"                   "true"
        #Option                 "AccelDFS"                      "true"
        #Option          "DynamicClocks"                "on"   #This is for laptop users, it saves energy when in battery mode.
        #Option         "AGPMode"                       "4"
        #Option         "ColorTiling"                   "1"
        #Option         "DMAForXv"                      "1" 
EndSection

Section "ServerFlags"
        Option  "DontZap"       "False"
EndSection

Does

    Option "EXANoDownloadFromScreen"

work around the problem?

EXANoDownloadFromScreen only helped when no OpenGL app was runnning at least for the time I tested. Once I started Auteria (http://www.auteria.com)game in the window, the font tearing and artifacts started appearing again. Also the text inside the game is flickering.  

(In reply to comment #7)
> EXANoDownloadFromScreen only helped when no OpenGL app was runnning at least
> for the time I tested. Once I started Auteria (http://www.auteria.com)game in
> the window, the font tearing and artifacts started appearing again.

A similar mechanism as for DownloadFromScreen is used for evicting buffer objects from VRAM, so this indicates that VRAM -> AGP transfers are generally unreliable on your system. One thing I notice in your dmesg is

[35162.662996] [drm] AGP mode requested: 4

Maybe you can try different values for the radeon.agpmode parameter, including 1, 2, not specifying any explicitly or -1 to disable AGP.

If none of that helps, maybe you can try booting the kernel or loading the radeon kernel module with radeon.test=1 and see if that gives any interesting results.

> Also the text inside the game is flickering.  

That could be a separate issue.

using radeon.agpmode=1 and EXANoDownloadFromScreen still results in font tearing. I'll try with agpmode=-1 to be sure. 

attaching dmesg with radeon.test = 1 output.

I cannot test for long as using EXA causes random hard freezes if EXANoComposite is not set.

Created attachment 28892 [details]
dmesg log

Testing now with radeon.agpmode=-1 so far no tearing/artifacts or hard freezes with EXA running for 6 hours.

If the AGP transfers are borked, how come the fglrx driver worked with AGP 4x and why are both tearing and hard freezes only showing up with ati driver ?

Created attachment 30587 [details]
dmesg log (RV250)

I can confirm this bug (with another ATI card - see below). Forcing the card into PCI mode using the driver option 'radeon.agpmode=-1' is the only solution for me (even though performance suffers); I tried AGP modes 1, 2, 4 without any results; option 'EXANoDownloadFromScreen' didn't work for me either.

I was affected by bug #444139 (https://bugs.launchpad.net/bugs/444139), so I used boot option 'radeon.modeset=1' to turn on KMS. Since then I'm affected by this bug. The only way to get rid of screen artifacts and font tearing (as shown by Gytis Raciukaitis) is boot option 'radeon.agpmode=-1'.

I attach the output of dmesg without option 'radeon.agpmode=-1'

01:00.0 VGA compatible controller [0300]: ATI Technologies Inc Radeon RV250 [Mobility FireGL 9000] [1002:4c66] (rev 01)

Created attachment 30655 [details]
dmesg with pci-mode forced

I can also confirm this bug on an Radeon 9800 PRO AGP. Running the ddx from git, I would get massive corruption all over the screen, but putting the card in AGP mode with the kernel parameter proposed here makes it all go away.

Created attachment 30939 [details]
weird fonts

Created attachment 30940 [details]
artifacts

I can also confirm this bug on a thinkpad T42 laptop running arch linux.

Grapics card (lspci):
"01:00.0 VGA compatible controller: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10]"

Some random characters start to look very weird after some random time (2 minutes to several hours) of operation (see screenshots above)

(In reply to comment #16)
> I can also confirm this bug on a thinkpad T42 laptop running arch linux.
> 
> Grapics card (lspci):
> "01:00.0 VGA compatible controller: ATI Technologies Inc RV350 [Mobility Radeon
> 9600 M10]"
> 
> Some random characters start to look very weird after some random time (2
> minutes to several hours) of operation (see screenshots above)
> 

me too, running Linux oddysseus 2.6.32-rc6 #1 Mon Nov 9 15:34:54 CET 2009 i686 GNU/Linux from git

xorg also from git 1.7.99 till 07.11.2009 does not fix it

libdrm PACKAGE_VERSION = 2.4.15 (git)
ddx_xf86-video-ati PACKAGE_VERSION = 6.12.99
mesa VERSION=7.7-devel

01:00.0 VGA compatible controller: ATI Technologies Inc Radeon Mobility M7 LW [Radeon Mobility 7500] (prog-if 00 [VGA controller])
        Subsystem: IBM ThinkPad T30

X.Org X Server 1.7.99.2
Release Date: (unreleased)
X Protocol Version 11, Revision 0
Build Operating System: Linux 2.6.32-rc6 i686 
Current Operating System: Linux oddysseus 2.6.32-rc7 #1 Mon Nov 9 15:34:54 CET 2009 i686
Kernel command line: BOOT_IMAGE=/boot/vmlinuz-2.6.32-rc7 root=UUID=cc527f28-a768-4379-accb-56069aee2d0a ro CONSOLE=/dev/tty1 radeon.modeset=1 gfxpayload=1024x768x24 quiet splash drm.debug=0xf lapic resume=/dev/hda5
Build Date: 06 November 2009  02:07:21PM

<snap from actual Xorg.log>
(II) [KMS] Kernel modesetting enabled.
(**) RADEON(0): Depth 24, (--) framebuffer bpp 32
(II) RADEON(0): Pixel depth = 24 bits stored in 4 bytes (32 bpp pixmaps)
(==) RADEON(0): Default visual is TrueColor
(**) RADEON(0): Option "AccelDFS"
(**) RADEON(0): Option "AccelMethod" "XXA"
(**) RADEON(0): Option "DRI" "true"
(==) RADEON(0): RGB weight 888
(II) RADEON(0): Using 8 bits per RGB (8 bit DAC)
(--) RADEON(0): Chipset: "ATI Radeon Mobility M7 LW (AGP)" (ChipID = 0x4c57)
(II) RADEON(0): AGP card detected
(II) RADEON(0): KMS Color Tiling: disabled
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 11, (OK)
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 11, (OK)
drmOpenByBusid: Searching for BusID pci:0000:01:00.0
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: open result is 11, (OK)
drmOpenByBusid: drmOpenMinor returns 11
drmOpenByBusid: drmGetBusid reports pci:0000:01:00.0
(II) [drm] DRM interface version 1.3
(II) [drm] DRM open master succeeded.
(II) RADEON(0): Output VGA-0 using monitor section Standardbildschirm
(II) RADEON(0): Output DVI-0 has no monitor section
(II) RADEON(0): Output LVDS has no monitor section
(II) RADEON(0): Output S-video has no monitor section
(II) RADEON(0): Output VGA-0 disconnected
(II) RADEON(0): Output DVI-0 disconnected
(II) RADEON(0): Output LVDS connected
(II) RADEON(0): Output S-video disconnected
(II) RADEON(0): Using exact sizes for initial modes
(II) RADEON(0): Output LVDS using initial mode 1024x768
(II) RADEON(0): Using default gamma of (1.0, 1.0, 1.0) unless otherwise stated.
(II) RADEON(0): mem size init: gart size :4000000 vram size: s:1000000 visible:c00000
(==) RADEON(0): DPI set to (96, 96)
(II) Loading sub module "fb"
(II) LoadModule: "fb"
(II) Loading /opt/xorg/lib/xorg/modules/libfb.so
(II) Module fb: vendor="X.Org Foundation"
        compiled for 1.7.99.2, module version = 1.0.0
        ABI class: X.Org ANSI C Emulation, version 0.4
(II) RADEON(0): [DRI2] Setup complete
(II) RADEON(0): Front buffer size: 3072K
(II) RADEON(0): Remaining VRAM size (used for pixmaps): 9184K
(==) RADEON(0): Backing store disabled
(II) RADEON(0): Direct rendering enabled
(II) RADEON(0): Render acceleration enabled for R100 type cards.
(II) RADEON(0): Setting EXA maxPitchBytes
(II) EXA(0): Driver allocated offscreen pixmaps
(II) EXA(0): Driver registered support for the following operations:
(II)         Solid
(II)         Copy
(II)         Composite (RENDER acceleration)
(II)         UploadToScreen
(II)         DownloadFromScreen
...
(II) RADEON(0): Acceleration enabled
(**) RADEON(0): DPMS enabled
(==) RADEON(0): Silken mouse enabled
(II) RADEON(0): Set up textured video
(II) RADEON(0): RandR 1.2 enabled, ignore the following RandR disabled message.
(WW) RADEON(0): Option "DynamicClocks" is not used
(WW) RADEON(0): Option "BIOSHotkeys" is not used
(WW) RADEON(0): Option "MergedFB" is not used
(WW) RADEON(0): Option "CRT2HSync" is not used
(WW) RADEON(0): Option "CRT2VRefresh" is not used
(WW) RADEON(0): Option "MetaModes" is not used
(WW) RADEON(0): Option "XAANoOffscreenPixmaps" is not used
(WW) RADEON(0): Option "GARTSize" is not used
(WW) RADEON(0): Option "AGPMode" is not used
(WW) RADEON(0): Option "AGPSize" is not used
...
(II) AIGLX: enabled GLX_MESA_copy_sub_buffer
(II) AIGLX: enabled GLX_SGI_make_current_read
(II) AIGLX: enabled GLX_SGI_swap_control and GLX_MESA_swap_control
(II) AIGLX: GLX_EXT_texture_from_pixmap backed by buffer objects
(II) AIGLX: Loaded and initialized /opt/xorg/lib/dri/radeon_dri.so
(II) GLX: Initialized DRI2 GL provider for screen 0
(II) RADEON(0): Setting screen physical size to 270 x 203
...

Created attachment 31079 [details]
the location and the navgator bar are corrupted

The fonts used inside the epiphany are here okay yet. after quite some working with apps, the fonts get weird in xterm/gnome-terminal/browser. 
But changing the fontsize (ctrl++) in browser or in the gnome-terminal dies help
until these font(sets) are also corrupted.

Are things any better with the radeon drm modules from drm-next?

(In reply to comment #19)
> Are things any better with the radeon drm modules from drm-next?
> 

I have compiled today the drm-2.6/drm-next branch (2.6.31-rc9) with my configs and I noticed no chars looks weird anymore (until now) xD

Than I have dumped all radeon regs,lspci, dmesg and suspended to disk from Xorg.
The resume fails here after the message "Loading Console Fonts [OK]"

When I s2disk at the console with a running xorg and resume, the resume and teh boot are okay, but I cannot get into a working Xorg as with the 2.6.32-rc6 (resume comes up with different IB schedule errors, restarting Xorg does not help, the schedule fails -> https://bugs.freedesktop.org/show_bug.cgi?id=23103 will post there my logs in a few minutes).

Since today I have also merged linux-2.6 
commit 6b46362c0ea472b174c336786fd406c504326ad4
Merge: 961767b eeba575
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Wed Nov 11 11:32:04 2009 -0800
Merge branch 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6


And one good message from today: I have to mention that I do not detect the weird chars in 2.6.32-rc6 anymore, too. I will stress testing some apps on the desktop now. 

thx airlie

Created attachment 50323 [details]
qayyng dmesg output

Created attachment 50324 [details]
qayyng xorg.conf

Created attachment 50326 [details]
qayyng Xorg.0.log

Created attachment 50327 [details]
qayyng screenshot 01

Created attachment 50329 [details]
qayyng screenshot 02

I appear to be having the same or a similar problem.  I'm attaching my Xorg log, Xorg conf, and dmesg output.

As you'll see, I've tried a number of the options suggested here and elsewhere.  I've had varying degrees of success.  Before I started playing with it, I was getting a minimal amount of artifacts (see the first screenshot), and while I've managed to get it back like that a couple times since, I can't reproduce it now that I'm keeping track of what I've changed.

With KMS disabled, there's no acceleration of any kind (I read somewhere that this is by design), but there're also no artifacts.

(In reply to comment #26)
> I appear to be having the same or a similar problem.

If there's any doubt (as there is in your case), it's better to file a separate report, to avoid cluttering up existing reports with unrelated problems.


> As you'll see, I've tried a number of the options suggested here and elsewhere.
>  I've had varying degrees of success.  Before I started playing with it, I was
> getting a minimal amount of artifacts (see the first screenshot),

That actually looks more like a Unity / X server issue (not rendering the desktop background properly) than a driver one.


> With KMS disabled, there's no acceleration of any kind (I read somewhere that
> this is by design), [...]

There should be at least 2D acceleration, but the r300g 3D driver indeed requires KMS.


I wonder if there's any point in keeping this report open, as it seems to basically boil down to "AGP is broken, and there's little we can do about it"...

(In reply to comment #27)
> (In reply to comment #26)
> > I appear to be having the same or a similar problem.
> 
> If there's any doubt (as there is in your case), it's better to file a separate
> report, to avoid cluttering up existing reports with unrelated problems.

I was wondering if that was the case.  I found a series of bugs here that seemed similar, but this one seemed to be affecting roughly a set of graphics cards mine fell under in roughly the same way.  I'd be happy to open a new bug for this issue if you think it's worth it at this point.

> 
> > As you'll see, I've tried a number of the options suggested here and elsewhere.
> >  I've had varying degrees of success.  Before I started playing with it, I was
> > getting a minimal amount of artifacts (see the first screenshot),
> 
> That actually looks more like a Unity / X server issue (not rendering the
> desktop background properly) than a driver one.

Well, background rendering is by far not the only problem.  I'll upload a screenshot where that's "clear", though I don't know that it's necessarily useful given what else you said.

> > With KMS disabled, there's no acceleration of any kind (I read somewhere that
> > this is by design), [...]
> 
> There should be at least 2D acceleration, but the r300g 3D driver indeed
> requires KMS.

Yes, 2d acceleration seems to be working okay.

> I wonder if there's any point in keeping this report open, as it seems to
> basically boil down to "AGP is broken, and there's little we can do about
> it"...

Hm, could you elaborate on how AGP is broken and how that affects this problem?  If it's something that a bug could be reported for (e.g., a kernel driver), would it be worth my time to submit one?

Created attachment 50389 [details]
qayyng screenshot 03

Created attachment 50390 [details]
qayyng screenshot 04

(In reply to comment #28)
> > If there's any doubt (as there is in your case), [...]
> 
> I was wondering if that was the case.  I found a series of bugs here that
> seemed similar, but this one seemed to be affecting roughly a set of graphics
> cards mine fell under in roughly the same way.

That's not good enough I'm afraid. It could be true for any number of different bugs.

> I'd be happy to open a new bug for this issue if you think it's worth it at
> this point.

Assuming it doesn't turn out to be a Unity issue, yes.


> > > As you'll see, I've tried a number of the options suggested here and elsewhere.
> > >  I've had varying degrees of success.  Before I started playing with it, I was
> > > getting a minimal amount of artifacts (see the first screenshot),
> > 
> > That actually looks more like a Unity / X server issue (not rendering the
> > desktop background properly) than a driver one.
> 
> Well, background rendering is by far not the only problem.

I was only referring to the first screenshot, as the above sounds like the additional problems only started after tweaking options.


> > I wonder if there's any point in keeping this report open, as it seems to
> > basically boil down to "AGP is broken, and there's little we can do about
> > it"...
> 
> Hm, could you elaborate on how AGP is broken and how that affects this problem?

Note that the above was referring to this bug report, not necessarily your problem(s).

> If it's something that a bug could be reported for (e.g., a kernel driver),
> would it be worth my time to submit one?

Probably not, lots of AGP bridges had various correctness / stability issues which could result in all kinds of funny behaviour.

I believe this is actually fixed.  I had a similar setup and had ugly tearing 
with my radeon card under Ubunto 10.10 and 11.04.  When I replaced their Mesa 
7.10 Gallium r600 DRI driver with one from mesa master (now 7.11) and their X 
driver with ati master, the issue went away.

Please retest with current git master of mesa, xserver, and xf86-video-ati and 
reopen if you still have issues.