Rage IIC onboard video on Compaq Proliant ML530 is non-functional with the "ati" driver in 4.2.99.2 in phoebe and rawhide. When starting up, the display timings are very screwed resulting in a display that is very distorted with randomly oscillating chunks on the screen, etc. It is totally unuseable, although it does run. You can fiddle enough to get the GNOME logout dialog and crack a guess as to where to click to logout. The "vesa" driver has been tested however and seems to work ok. If this can't be fixed in time for final release the driver should be resorted to "vesa" in order to have working video.
00:05.0 VGA compatible controller: ATI Technologies Inc 3D Rage IIC 215IIC [Mach64 GT IIC] (rev 7a) (prog-if 00 [VGA]) Subsystem: ATI Technologies Inc Rage IIC Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B- Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- Latency: 64 (2000ns min), cache line size 08 Region 0: Memory at c5000000 (32-bit, prefetchable) [size=16M] Region 1: I/O ports at 2000 [size=256] Region 2: Memory at c6bfe000 (32-bit, non-prefetchable) [size=4K] Expansion ROM at <unassigned> [disabled] [size=128K] Capabilities: [5c] Power Management version 1 Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-) Status: D0 PME-Enable- DSel=0 DScale=0 PME- 00:05.0 Class 0300: 1002:4756 (rev 7a) Subsystem: 1002:4756 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping+ SERR- FastB2B- Status: Cap+ 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- Latency: 64 (2000ns min), cache line size 08 Region 0: Memory at c5000000 (32-bit, prefetchable) [size=16M] Region 1: I/O ports at 2000 [size=256] Region 2: Memory at c6bfe000 (32-bit, non-prefetchable) [size=4K] Expansion ROM at <unassigned> [disabled] [size=128K] Capabilities: [5c] Power Management version 1 Flags: PMEClk- DSI- D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-) Status: D0 PME-Enable- DSel=0 DScale=0 PME-
This is still broken in CVS as of today. Looks like we might have to change this one to "vesa" soon.
After doing more testing, oddly enough, I've determined that this problem is also reproduceable in stock Red Hat Linux 8.0. Since nobody else has reported it to us in bugzilla, and I haven't gotten any email, nor seen anyone mention it on XFree86 mailing lists, I do not consider this a pressing issue compared to other problems. It is entirely possible that the video chip onboard my proliant is faulty also. I'm not going to debug this until I can get several other people to confirm the problem is reproduceable for them as well.
Jeremy Katz indicates that Rage XL works for him higher than 1024x768 ok with our latest tree, so this is likely specific to Rage IIC
Update: Problem still present in latest builds. 2 other people with Rage IIC are able to reproduce this.
Screenshots at: ftp://people.redhat.com/mharris/pictures/RageIIC
Nobody else has reported this problem, and I can't seem to find similar bug reports in freedesktop or xfree86 bugzilla, nor googling, so I'm wondering if this problem is a hardware bug or not. I'll need to test it on our current OS release to see if there's a difference before digging into the problem further, but that wont happen for a while as it isn't yet time to upgrade that box. Setting status to "DEFERRED" until my devel.capslock.lan box gets upgraded to FC2 or later, and I have a chance to test/debug the issue further.
Still haven't upgraded my test box to FC5, but still planning on it.
Add to FC6Destop tracker
Finally upgraded my test box to FC5 but forgot about this bug.. Will test soon and update.
I still have this machine, but no longer use it actively, and when I do it is generally exclusively used over the network so video hardware support isn't that important overall. I mentioned I'd test it soon and update above, but to be honest the machine hasn't even been powered on more than once or twice since then, still running FC5. Since it isn't a big deal to me overall, and it's easy to workaround by using lower resolution, and I haven't noticed any other public bug reports of this nature, either it is fixed now, or it is a problem with very few people affected by it. Either way, it shouldn't be a big priority to anyone overall in the big picture of this since it is ancient hardware. I'm closing the bug for now to cut down on clutter in bugzilla, however if I do happen to reproduce it again in the future, I'll log a new bug report in X.Org bugzilla instead as that's a better place to track it anyhow.