On Thu, May 13, 2004 at 04:32:23PM +1000, Ian Maclaine-cross wrote: > On Mon, May 10, 2004 at 03:50:57PM -0500, Branden Robinson wrote: > > tag 233001 + upstream > > thanks > > > > On Fri, Apr 30, 2004 at 03:07:46PM +1000, Ian Maclaine-cross wrote: > > > On Tue, Apr 20, 2004 at 08:20:39PM -0500, Branden Robinson wrote: > > > > On Mon, Feb 16, 2004 at 08:32:25PM +1100, Ian Maclaine-cross wrote: > > > > > Xfree86 3.3.6 worked with my Tseng ET4000/W32p graphics card but > > > > > 4.2 does not except with a low resolution vesa or vga mode. > > > > > > > > Did you *try* the "tseng" driver instead? > > > > > > Yes indeed and it did nothing. > > > > Hrmph. > > > > You appear to have a motherboard that the PCI bus handling code in > > XFree86 4.3.0 doesn't deal with well; a lot of "phantom" buses show up, > > and this *may* be confusing the X server into not knowing where to find > > your video card. > > My mainboard is a Gigabyte GA-5AA and CPU is AMD-K6-3/450. > /proc/pci has the video card OK: > > Bus 0, device 8, function 0: > VGA compatible controller: Tseng Labs Inc ET4000/W32p rev C (rev 0). > IRQ 11. > Non-prefetchable 32 bit memory at 0xef000000 [0xefffffff]. > > Works OK with xfree86 3.3.6. Yeah, the PCI bus logic was extensively overhauled in 4.x. > I have not built xfree86 before. I read your instructions and I think > I should be able to follow them sometime in June. Since I last mailed you, I have committed the following to the Debian XFree86 SVN trunk: r1373 | branden | 2004-05-10 13:39:21 -0500 (Mon, 10 May 2004) | 17 lines Update (from XFree86 CVS, 2003-12-18) the XFree86 X server's PCI bus handling code and device recognition databases from CVS. Resync patch #452. Delete superseded patch #453. + Simplify internal interfaces in the PCI code and remove the Xserver's interference with normal PCMCIA operation (Marc La France). + Fixed typo in getPciBiosTypes() (Bugzilla #382, Jakub Bogusz). + Change DEVID macro to work around glitch in SCO's C compiler (Marc La France). + Fixed support for 64bit PCI bus on 32bit systems (Egbert Eich). + Be a little more precise about differentiating between active and inactive non-video PCI resources (Marc La France). + Fix bug in detection of multi-function PCI devices (Marc La France, in partial resolution of Bugzilla #574). + Rip out incorrect limits on the number of PCI buses an ix86 chipset can handle and implement an improved solution for avoiding "phantom" PCI buses (Marc La France, Bugzilla #604). That last item may hold out some hope for your motherboard. -- G. Branden Robinson | Imagination was given man to Debian GNU/Linux | compensate for what he is not, and branden@debian.org | a sense of humor to console him for http://people.debian.org/~branden/ | what he is.
Attachment:
signature.asc
Description: Digital signature