Created attachment 117657 [details] Xorg.0.log for 6.10.0 I wanted to follow up on some issues I reported a while back, and tried to see if there were any improvements/changes with 6.10.0 by using the package in Debian sid (xserver-xorg-video-r128 6.1.0-1; should be only 1 commit behind master). Using the same xorg.conf that worked in 6.9.2, X fails to start with 6.10.0 for an iMac G3. The first error in Xorg.0.log is "(EE) Screen(s) found, but none have a usable configuration." So far I haven't located a more specific error, since troubleshooting resources online show it usually being preceded by other errors in Xorg.0.log. There's also quite a bit more output in the log file for 6.9.2 (mostly from "(EE) R128(0): FBIOPUT_VSCREENINFO: Invalid argument"). I don't yet know if something else in the X stack is causing this, since the rest of it was updated as well: X Server 1.16.4 -> 1.17.2 xf86-input-wacom 0.26.0 -> 0.30.0 xf86-video-ati 6.9.4 -> 6.9.5
Created attachment 117658 [details] Xorg.0.log for 6.9.2 (This is similar to attachment 112671 [details] from bug 88767. In particular I modified HorizSycn and VertRefresh in xorg.conf so it would no longer say "(**) R128(0): *Default mode "1024x768"…" immediately after saying "(II) R128(0): Not using default mode "1024x768" (unknown reason)".)
Created attachment 117659 [details] xorg.conf This is based on 'imac 350' on http://oswaldkelso.blogspot.com/2009/11/imac-xorgconf-ppc-all-versions.html
This is what happens when no monitor is detected by the RandR hooks. The old driver worked because it lied to X that a monitor was detected whether it was or not. RandR documentation encourages drivers not to do this, but due to documentation inconsistencies, it seems like it's best to keep lying. At first it looked like there was one set of pins for DDC. Then I learned that there is another set of pins if you're on a VGA card instead of DVI. Then I learned that there is yet another set if you're on a Mobility card and not a desktop. Could it be that there is yet another set if you're on PPC instead of x86? Anyway, please try the patch https://bugs.freedesktop.org/attachment.cgi?id=117325 if you have time.
(In reply to Connor Behan from comment #3) > At first it looked like there was one set of pins for DDC. Then I learned > that there is another set of pins if you're on a VGA card instead of DVI. > Then I learned that there is yet another set if you're on a Mobility card > and not a desktop. Could it be that there is yet another set if you're on > PPC instead of x86? I'm inclined to think there is no special chip or pin differences between PPC and x86, that is, the ATI chip in slot-load an iMac G3 should be the same as one on an AGP card (some Power Macs could use PC video cards once Mac ROM was flashed in place of the PC video BIOS, though I haven't come across reports of this for a Rage 128). The CRT would indeed be connected via VGA, although I never checked if there's any DDC/EDID present; there indeed may not be any since Mac OS would know what to do by other means, and Linux users historically relied on xorg.conf modelines. Any iMac G3s with a VGA port could only "mirror", implying that it only output the same signal going to the internal CRT, so the DDC lines likely aren't even connected (though I never tested this). http://www.everymac.com/systems/apple/imac/faq/imac-g3-video-processor-external-display-support.html Special pin assignments or other clever reuse might be another matter though, given there being different firmware present…
(In reply to Christopher Chavez from comment #4) > (In reply to Connor Behan from comment #3) > > The CRT would indeed be connected via VGA, although I never checked if > there's any DDC/EDID present; there indeed may not be any since Mac OS would > know what to do by other means, and Linux users historically relied on > xorg.conf modelines. Any iMac G3s with a VGA port could only "mirror", > implying that it only output the same signal going to the internal CRT, so > the DDC lines likely aren't even connected (though I never tested this). > cf. https://68kmla.org/forums/index.php?/topic/13257-mac-mini-hardware-into-an-imac-g3/?p=136733 : "…[the internal CRT monitors] do not have standard VESA DDC lines to communicate their strangeness to a standard video card."
The patch is in git now. I'll make a release if someone can confirm that it works.
(In reply to Connor Behan from comment #6) > The patch is in git now. I'll make a release if someone can confirm that it > works. "It works!" (http://www.ibm.com/developerworks/library/x-db2mblog3/fig04.jpg) Any details/logs you'd like forwarded? Tested-by: Christopher Chavez <chrischavez@gmx.us>
If it works, the logs must be similar to the ones on my box. I just pushed the release. Thanks.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.