Bug 185944
| Summary: | Radeon driver crash on dualhead setup | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Jan "Yenya" Kasprzak <kas> | ||||||
| Component: | xorg-x11-drv-ati | Assignee: | Adam Jackson <ajax> | ||||||
| Status: | CLOSED CURRENTRELEASE | QA Contact: | |||||||
| Severity: | medium | Docs Contact: | |||||||
| Priority: | medium | ||||||||
| Version: | 5 | CC: | bernd.bartmann, bogado, chrisspen, clydekunkel7734, dath, eric.moret, fs, mcepl, ncunning, nutello, redhatbugzilla, rs, sbeam, stadler, surkum, xgl-maint | ||||||
| Target Milestone: | --- | ||||||||
| Target Release: | --- | ||||||||
| Hardware: | x86_64 | ||||||||
| OS: | Linux | ||||||||
| Whiteboard: | |||||||||
| Fixed In Version: | FC6 | Doc Type: | Bug Fix | ||||||
| Doc Text: | Story Points: | --- | |||||||
| Clone Of: | Environment: | ||||||||
| Last Closed: | 2007-07-03 16:11:54 UTC | Type: | --- | ||||||
| Regression: | --- | Mount Type: | --- | ||||||
| Documentation: | --- | CRM: | |||||||
| Verified Versions: | Category: | --- | |||||||
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||
| Cloudforms Team: | --- | Target Upstream Version: | |||||||
| Embargoed: | |||||||||
| Attachments: |
|
||||||||
|
Description
Jan "Yenya" Kasprzak
2006-03-20 10:51:11 UTC
Created attachment 126340 [details]
xorg.conf for dual-head setup (works with 6.8.2 drivers)
Just as a me too - similar hardware (X300SE, but dual dvi instead of dvi/vga). It didn't crash for me, but dual head was absolutely not working. Although the desktop obviously extended accross both monitors (the mouse traversed, windows could be placed elsewhere, etc), both displays showed the same half of the desktop. The same fix (copying the fc3 modules directory and adding as modulespath) also fixed it for me. Er - make that the fc4 modules, not fc3. Same problem but slighly diferent hardware, RV100 dual CRT, works fine with MergeFB in fc4, but after upgrade to fc5 the secondary screen shows only garbage (copy of primary) except for the cursor(hw) who looks fine when you move it between the screens. Trying other configurations hangs the machine. Similar problem here. Dell desktop (Dimension 8400) with "ATI Technologies Inc RV370 5B60 [Radeon X300 (PCIE)]" (driver in xorg.conf "radeon"). One card, two outputs (DVI and analog). Seconds monitor (analog) is always black (but recieves signal) and eventually X freezes and requires cold restart. Removing BusID "PCI:1:0:0" entry in xorg.conf for the second device simply clones desktop on both monitors. Similar problem here as comment #5 (same video card), and X freezes eventually and required a cold restart. Mostly, it works. I've made it through a half day already without a lockup. Here's the relavent parts of xorg.conf: Section "Monitor" Identifier "Monitor0" VendorName "Monitor Vendor" ModelName "Dell 2001FP (Digital)" DisplaySize 410 310 ### Comment all HorizSync and VertSync values to use DDC: # HorizSync 31.0 - 80.0 # VertRefresh 56.0 - 76.0 Option "dpms" EndSection Section "Device" Identifier "Videocard0" Driver "radeon" VendorName "Videocard vendor" BoardName "ATI Technologies Inc RV370 5B60 [Radeon X300 (PCIE)]" Option "MergedFB" "on" Option "MonitorLayout" "TMDS,TMDS" Option "CRT2Position" "RightOf" Option "MetaModes" "1600x1200-1600x1200" EndSection Section "Screen" Identifier "Screen0" Device "Videocard0" Monitor "Monitor0" DefaultDepth 24 SubSection "Display" Viewport 0 0 Depth 16 Modes "800x600" "640x480" EndSubSection SubSection "Display" Viewport 0 0 Depth 24 Modes "1600x1200" "1400x1050" "1280x960" "1152x864" "1024x768" "800x600" "640x480" EndSubSection EndSection At the risk of being yet another "me too" report, I would like to say that I
believe I am maybe running into this same problem as well. My setup is a
little different:
02:00.0 VGA compatible controller: ATI Technologies Inc RV350 AQ [Radeon 9600]
02:00.1 Display controller: ATI Technologies Inc RV350 AQ [Radeon 9600]
(Secondary)
When I tried using a dual head setup my box would seemingly lock up and I had
to physically power it off (ctrl-alt-backspace didn't help, and my capslock
quit working). I didn't try to ssh into the machine to make sure it was
completely gone, but I could do that or perhaps a little more troubleshooting
if it would help someone get to the bottom of this problem.
My broken xorg.conf:
# Xorg configuration created by system-config-display
Section "ServerLayout"
Identifier "Multihead layout"
Screen 0 "Screen0" LeftOf "Screen1"
Screen 1 "Screen1" 0 0
InputDevice "Mouse0" "CorePointer"
InputDevice "Keyboard0" "CoreKeyboard"
Option "Xinerama" "on"
Option "Clone" "off"
EndSection
Section "Files"
RgbPath "/usr/share/X11/rgb"
ModulePath "/usr/lib64/xorg/modules"
FontPath "/usr/share/X11/fonts/misc/"
FontPath "/usr/share/X11/fonts/TTF/"
FontPath "/usr/share/X11/fonts/OTF"
FontPath "/usr/share/X11/fonts/Type1/"
FontPath "/usr/share/X11/fonts/CID/"
FontPath "/usr/share/X11/fonts/100dpi/"
FontPath "/usr/share/X11/fonts/75dpi/"
EndSection
Section "Module"
Load "vnc"
Load "glx"
Load "record"
Load "dbe"
Load "xtrap"
Load "extmod"
Load "type1"
Load "freetype"
Load "dri"
EndSection
Section "InputDevice"
Identifier "Keyboard0"
Driver "kbd"
EndSection
Section "InputDevice"
Identifier "Mouse0"
Driver "mouse"
Option "Protocol" "auto"
Option "Device" "/dev/input/mouse0"
Option "ZAxisMapping" "4 5 6 7"
EndSection
Section "Monitor"
#DisplaySize 400 300 # mm
### Comment all HorizSync and VertSync values to use DDC:
Identifier "Monitor0"
VendorName "VSC"
ModelName "P225f-6"
### Comment all HorizSync and VertSync values to use DDC:
HorizSync 30.0 - 127.0
VertRefresh 50.0 - 160.0
Option "DPMS"
EndSection
Section "Monitor"
Identifier "Monitor1"
VendorName "Monitor Vendor"
ModelName "Dell P1110"
### Comment all HorizSync and VertSync values to use DDC:
HorizSync 30.0 - 121.0
VertRefresh 48.0 - 160.0
Option "dpms"
EndSection
Section "Device"
### Available Driver options are:-
### Values: <i>: integer, <f>: float, <bool>: "True"/"False",
### <string>: "String", <freq>: "<f> Hz/kHz/MHz"
### [arg]: arg optional
#Option "NoAccel" # [<bool>]
#Option "SWcursor" # [<bool>]
#Option "Dac6Bit" # [<bool>]
#Option "Dac8Bit" # [<bool>]
#Option "BusType" # [<str>]
#Option "CPPIOMode" # [<bool>]
#Option "CPusecTimeout" # <i>
#Option "AGPMode" # <i>
#Option "AGPFastWrite" # [<bool>]
#Option "AGPSize" # <i>
#Option "GARTSize" # <i>
#Option "RingSize" # <i>
#Option "BufferSize" # <i>
#Option "EnableDepthMoves" # [<bool>]
#Option "EnablePageFlip" # [<bool>]
#Option "NoBackBuffer" # [<bool>]
#Option "DMAForXv" # [<bool>]
#Option "PanelOff" # [<bool>]
#Option "DDCMode" # [<bool>]
#Option "MonitorLayout" # [<str>]
#Option "IgnoreEDID" # [<bool>]
#Option "UseFBDev" # [<bool>]
#Option "MergedFB" # [<bool>]
#Option "CRT2HSync" # [<str>]
#Option "CRT2VRefresh" # [<str>]
#Option "CRT2Position" # [<str>]
#Option "MetaModes" # [<str>]
#Option "MergedDPI" # [<str>]
#Option "MergedXinerama" # [<bool>]
#Option "MergedXineramaCRT2IsScreen0" # [<bool>]
#Option "MergedNonRectangular" # [<bool>]
#Option "MergedMouseRestriction" # [<bool>]
#Option "DisplayPriority" # [<str>]
#Option "PanelSize" # [<str>]
#Option "ForceMinDotClock" # <freq>
#Option "ColorTiling" # [<bool>]
#Option "VideoKey" # <i>
#Option "RageTheatreCrystal" # <i>
#Option "RageTheatreTunerPort" # <i>
#Option "RageTheatreCompositePort" # <i>
#Option "RageTheatreSVideoPort" # <i>
#Option "TunerType" # <i>
#Option "RageTheatreMicrocPath" # <str>
#Option "RageTheatreMicrocType" # <str>
#Option "RenderAccel" # [<bool>]
#Option "SubPixelOrder" # [<str>]
#Option "ShowCache" # [<bool>]
#Option "DynamicClocks" # [<bool>]
#Option "BIOSHotkeys" # [<bool>]
#Option "VGAAccess" # [<bool>]
#Option "ReverseDDC" # [<bool>]
#Option "LVDSProbePLL" # [<bool>]
#Option "AccelMethod" # <str>
Identifier "Card0"
Driver "ati"
VendorName "ATI Technologies Inc"
BoardName "RV350 AQ [Radeon 9600]"
BusID "PCI:2:0:0"
EndSection
Section "Device"
Identifier "Videocard1"
Driver "radeon"
VendorName "Videocard Vendor"
BoardName "ATI Technologies Inc RV350 AQ [Radeon 9600]"
BusID "PCI:2:0:0"
Screen 1
EndSection
Section "Screen"
Identifier "Screen0"
Device "Card0"
Monitor "Monitor0"
DefaultDepth 16
SubSection "Display"
Viewport 0 0
Depth 16
Modes "1600x1200" "1400x1050" "1280x960" "1152x864"
"1024x768" "800x600" "640x480"
EndSubSection
EndSection
Section "Screen"
Identifier "Screen1"
Device "Videocard1"
Monitor "Monitor1"
DefaultDepth 16
SubSection "Display"
Viewport 0 0
Depth 16
Modes "1600x1200"
EndSubSection
EndSection
This appears to be generic dualhead brokenness in the X11R7.0 ATI radeon driver. There have been various other bug reports of this as well. I want to add "me too" here, I also have the same problem with eventual lockups in my laptop. I will be ading my email to the CC list and I am eager to have this bug squashed so I am offering my help to test stuff out and help with the solution. :-) I've make dualhead work using xorg-x11-drv-ati-6.5.7.3-4.cvs20060404 from rawhide on my system which uses an "ATI Technologies Inc Radeon RV100 QY [Radeon 7000/VE]" with a dual CRT config, spanning screens. Juan are there any RPMs I download? I have a Radeon X300 PCI-X card with one analog and one digital output. I was having the crashing problem, and this driver worked to cause it not to crash. Also, the system seems to think it has a dual head desktop in terms of space available. However, the analog monitor doesn't receive any out put once I enter X (it is there during bios, boot, etc.). Any thoughts? Juan, would you post your xorg.conf so I can use it as a reference? What do you have for the driver entry? I've tried "ati" (blank secondary screen) "radeon" (blank secondary) and "fglrx" (scrambled first screen or system crash) but can't get xinerama to work. I tried with the updated driver (primary display is a 1400x1050 laptop screen and the secondary is a 1280x1024 LCD) but the primary is scrambled and the secondary seems to be set to 1400x1050 so it scrolls around the screen. Even with Xinerama enabled the secondary is still a clone of the primary. This is the same result as with the stock driver. Also, what kernel are you using? I have both 2.6.15-1.2054 and 2.6.16-1.2080 and the new driver locks the system (hard reboot required) under 2.6.16. Created attachment 127530 [details]
xorg.conf file working dualhead mergedfb
Another "me too" to report. After initially upgrading to FC5 I experienced some
X-freezing, but now I can't reproduce that. Instead, I've narrowed my dual-head
setups into three categories:
Setup:
No dual head.
Effect:
Both screens show identical image.
Setup:
Second Video Card: ATI Technologies Inc RV370 5B60 [Radeon (PCIE)]
Layout: Individual or Spanning Desktops
Effect:
Both screens show identical image, except for the mouse, which can travel
between the two screens and appears as a roughly 200x200 pixel blue square.
Setup:
Second Video Card: ATI Technologies Inc RV370 [Radeon X300SE]
Layout: Individual or Spaning Desktops
Effect:
Both screens show identical image.
Second Video Card switches back to "ATI Technologies Inc RV370 5B60 [Radeon
(PCIE)]"
Aside from not having functioning dual head in any case, I haven't experienced
any crashes or lockups.
Same problem here. This effectively means that dual head on FC5 doesn't work at all with BOTH major graphic cards brands (NVIDIA and ATI) and the stock drivers. Is there anything we can test out to help getting this working again ASAP? I already tried the latest rpms from rawhide, but with the same results. Due to this problem FC5 is completely unusable for me :-(( Me Too! Why no news on this one? I just installed FC5 on a lab full of Dell Optiplex 280 PC's with ATI X300 dual video displays. Under FC4, there was never any trouble with the dual head display and MergedFB. Now machines are constantly locking up at random, no error logs. Does the devel version of the ati driver fix the dual head problem? Dual head worked fine for me on FC4, upgrading to FC5 brought random lockups. These seemed to be most easily triggered by looking at the Fedora People blog aggregator in any of the available web browsers. Hardware is a Dell PC with a: 01:00.0 VGA compatible controller: ATI Technologies Inc RV370 5B60 [Radeon X300 (PCIE)] With two Iiyama flat screens (not sure of the model number). Upgrading from the default xorg-x11-drv-ati to xorg-x11-drv-ati-6.6.0-1 from development has fixed the problem for me. Me too. Had dual head Xinerama working on Fc3. Clean wipe and upgrade to FC5 broke it. Have the following cards (lspci output): 01:00.0 VGA compatible controller: ATI Technologies Inc Rage 128 PF/PRO AGP 4x TMDS 02:0b.0 VGA compatible controller: ATI Technologies Inc Radeon RV100 QY [Radeon 7000/VE] System freezes with some configurations, others only one screen works. Have not been able to get the monitor attached to Screen 2 (the R128) to do anything other than wake from suspend. /var/log/Xorg.0.log reports that (WW) R128: No matching Device section for instance (BusID PCI:2:11:0) found and then: (II) LoadModule: "radeon" (II) Reloading /usr/lib/xorg/modules/drivers/radeon_drv.so (EE) Screen 1 deleted because of no matching config section. the first WW is weird because it is clearly the Radeon attached to that bus, and is set up correctly in xorg.conf. No idea why R128 seems to be complaining about it. Me too. Clean install of FC5 fails. 01:00.0 is an ATI Radeon X300SE 01:00.1 is the second aperture on this card 02:00.0 is another ATI Radeon X300SE 02:00.1 is the second aperture on this card The second card is never demonstrably recognized. When I add a BusID 2:0:0 for my second head, the system freezes and requires a hard reset. When I use the second aperture on the first card, it clones the first monitor no matter what I try (although I have not yet tried MergedFB), and then eventually hangs, requiring a hard reset. (This is on a P4 3GHz HT 32-bit CPU, with an SMP kernel.) When I do not supply any BusIDs (and thereby avoid the immediate hang) I get the exact same messages in Xorg.0.log as SZ BEAM, modulo BusIDs: (WW) radeon: No matching Device section for instance (BusID PCI:1:0:1) found (WW) radeon: No matching Device section for instance (BusID PCI:2:0:0) found (WW) radeon: No matching Device section for instance (BusID PCI:2:0:1) found (II) LoadModule: "radeon" (II) Reloading /usr/lib/xorg/modules/drivers/radeon_drv.so (EE) Screen 1 deleted because of no matching config section. I have found that just before the eventual hang with one head, there is a comment in the Xorg.0.log (discovered when it is backed up the next time X launches, after a reset) that says that the Radeon FIFO has disappeared. ------- Additional Comments From james.antill 2006-06-09 14:51 EST ------- Note that the http://download.fedora.redhat.com/pub/fedora/linux/core/updates/testing/5/i386/xorg-x11-drv-ati-6.5.8.0-1.i386.rpm works for me. I've installed the said RPM and it seems to be working okay to me also. that RPM was installed here. But it does NOT work. symptoms as above. did a # yum --enablerepo=development upgrade xorg-x11-drv-ati to get the latest ati driver xorg-x11-drv-ati-6.6.0-3 and xorg-x11-server-Xorg-1.1.0-13. Still no change except that my one working screen now is too bright and flickers every 30 seconds... Our bugzilla database had a catastrophic failure on June 13, 2006 which lost all bug changes from Thurs June 9th through 13th. I have bugzilla email records of the time period however, and will be restoring as many changes as I can manually. I also tried the newer rpms from FC5/updates/testing and Rawhide, but the problem still exists. This is the only problem that stops me from using FC5. Again the question: What can we do to help get this going again? (In reply to comment #17) > Same problem here. This effectively means that dual head on FC5 doesn't work at > all with BOTH major graphic cards brands (NVIDIA and ATI) and the stock drivers. > > Is there anything we can test out to help getting this working again ASAP? I > already tried the latest rpms from rawhide, but with the same results. > > Due to this problem FC5 is completely unusable for me :-(( (In reply to comment #28) > I also tried the newer rpms from FC5/updates/testing and Rawhide, but the > problem still exists. This is the only problem that stops me from using FC5. > Again the question: What can we do to help get this going again? The best thing anyone can do, who is seriously interested in seeing this (or these, as it is not 100% clear there is one single problem in this bug report) issue resolved, is to file a bug report in X.Org bugzilla at http://bugs.freedesktop.org in the "xorg" component against the "radeon" driver, and include all relevant details in the upstream report. This is not a Red Hat specific bug or problem. We ship the stock X.Org supplied "radeon" driver, so any bugs present in it, are X.Org bugs that will affect all distributions out there. By having the issue logged directly in X.Org's bugzilla, _all_ of the developers who work on the "radeon" driver upstream are more likely to see the bug report, and may be able to investigate it directly and possibly fix it. As it stands now, the problem is highly dependent on exact specific video hardware, and each person seems to have similar or related issues, but not 100% identical issues, so it is very hard to just look at the bug report comments, etc. and conclude what the exact problem is, other than a vague general "the radeon driver's output display logic is horribly broken, and evolves breaking for various hardware each release". The reason it breaks, is because nobody actually has every single radeon card, and there's no simple way to test every single driver change on every single card out there. Not enough people beta test upstream X.Org releases before they're released as "stable" and so things regress over time. Always report X bugs directly to Xorg, and these problems get fixed faster. No guarantees of course, but at least more than 1 or 2 overworked developers ends up seeing the generic bugs when people report them to the source of the problem - X.Org bugzilla. The radeon driver's display detection logic seems to need a complete total rip-out and rewrite from scratch, as it appears to be horrendously broken for the majority of people out there. It does not appear to be a trivial minor bug fix type of thing, but a complete "rewrite the entire driver" type of thing. That's definitely something that _must_ happen at the X.Org level, and not at the distribution level. Paste the bug URL here once it's filed, and we'll track the issue in X.Org bugzilla. It looks like there are several bugs already open on freedesktop.org for this or similar problems: https://bugs.freedesktop.org/show_bug.cgi?id=2597 (lots of activity here) https://bugs.freedesktop.org/show_bug.cgi?id=5571 https://bugs.freedesktop.org/show_bug.cgi?id=7104 The first one should cover the problems we're seeing here, so I'll add my configuration to it (Dell Latitude 600 - xinerama working fine with FC4, won't work with FC5 using radeon, ati and fglrx drivers). (In reply to comment #30) > It looks like there are several bugs already open on freedesktop.org for this or > similar problems: > > https://bugs.freedesktop.org/show_bug.cgi?id=2597 (lots of activity here) Except for the fact that that is an Intel/nvidia related bug, and this bug is a Radeon bug, and this is a driver specific problem. People see someone else report a problem on a different card that uses a totally different driver, and think "Hey, I'm having the same problem on this totally other card" or "hey, I'm having a similar problem on a different card", and they add all kinds of "me too" comments for a completely unrelated problem. That makes a bug report actually be 2 to 10 bug reports in 1 actual report, of completely unrelated issues that "sound similar", and makes life on developers very difficult. > https://bugs.freedesktop.org/show_bug.cgi?id=5571 > https://bugs.freedesktop.org/show_bug.cgi?id=7104 > > The first one should cover the problems we're seeing here, so I'll add my > configuration to it (Dell Latitude 600 - xinerama working fine with FC4, won't > work with FC5 using radeon, ati and fglrx drivers). To be clear, it is very likely that the problem reported in this bug report is _specific_ to the "radeon" driver, which is quite buggy at this point with regard to output display autodetection, and separately to dualhead configurations, although the problems are almost certainly interrelated. Driver problems on any other non-ATI hardware are almost certainly completely unrelated problems in other video drivers. Even in cases where bugs on different hardware could get tracked down to a common generic server problem, and fixing the server fixes all hardware, it is best to track each hardware problem in its own bug report and close them as duplicates _after_ the connection is firmly established. Lumping multiple similar problems into one bug report, makes them not possible to isolate into individual reports once the original issue reported does get resolved for the reporter. All in one "me too" bug reports that tonnes of people pile on to are frustrating to say the least. At the risk of murkying the waters further, here are two mode freedesktop bugs that track this (or related) hangs. https://bugs.freedesktop.org/show_bug.cgi?id=5781 https://bugs.freedesktop.org/show_bug.cgi?id=7019 To be clear: These are both related to hangs on Radeon 7500's and with xorg-x11-drv-ati-6.5.8.0-1. Add to FC6Destop tracker created an upstream bug: https://bugs.freedesktop.org/show_bug.cgi?id=8461 as noted above, this is for MY symptoms and if yours are at all different then it may be best to create a new bug rather than "me too". Fedora Core 5 is no longer supported, please, could you reproduce this bug with the updated version of the currently supported distribution (Fedora Core 6, or Fedora 7, or Rawhide)? If this issue turns out to still be reproducible, please let us know in this bug report. If after a month's time we have not heard back from you, we will have to close this bug as CANTFIX/INSUFFICIENT_DATA. Setting status to NEEDINFO, and awaiting information from the reporter. Thanks in advance. I'm not the original report but here is an update from me. I would support the proposal to close this bug. Once I got a working xorg.conf arranged, dual-head has been stable for me. The driver no longer appears to be crashing. (getting a working xorg.conf arranged was another story entirely, but this has to do with issues between the driver, X, and system-config-display). I'm using FC6, xorg-x11-drv-ati-6.6.3-1.fc6, system-config-display-1.0.45-1. I agree - this bug can be closed, the problem is not present in FC6 or F7 (which is what I am running now). Closing per reporter's request. |