On Tue, 24 Feb 2009 22:16:10 +0100 Matthieu Herrb <mhe...@gmail.com> wrote:
> On Tue, Feb 24, 2009 at 9:20 PM, J.C. Roberts > <list-...@designtools.org> wrote: > > I'm testing out the v2.6.1 intel(4) driver requested here: > > http://marc.info/?l=openbsd-tech&m=123307709522306&w=2 > > > > In my Xorg.0.log I'm getting the mysterious error: > > > > (WW) intel(0): libpciaccess reported 0 rom size, guessing > > 64kB (EE) intel(0): Bad VBT signature > > (WW) intel(0): VBIOS initialization failed. > > This can probably be ignored. (I'm not 100% sure how the intel cards > setup their BIOS, but it seems normal that it doesn't get listed in > PCI config space and thus that libpiaccess gets back to the legacy VGA > space). > > > > > > On current we're running libpciaccess 10.2, so I figured I'd give > > the more recent 10.5 a try: > > > > http://cgit.freedesktop.org/xorg/lib/libpciaccess/ > > http://cgit.freedesktop.org/xorg/lib/libpciaccess/snapshot/libpciaccess-0.10. 5.tar.gz > > > > The existing xenocara/lib/libpciaccess/configure script shows it was > > created with autoconf-2.61 but everything I've tried to create a new > > one with the above source has resulted in a non-working script. > > > > What is the autoconf incantation necessary to generate a proper > > configure script that will work correctly within the xenocara tree? > > > > Copy Makefile.bsd-wrapper from the xenocara version and run > XENOCARA_RERUN_AUTOCONF=-Yes make -f Makefile.bsd-wrapper build > > > But a more recent libpciaccess will not change anything... > Thank you. I also copied over the libpicaccess/src/openbsd_pci.c since the one we have in the xenocara tree is newer than the one from xorg. As you expected, the new version made no difference with the error, but I've got this (wild, unsubstantiated) hunch the failures I'm seeing with EXA acceleration on the 845G chipset are rooted in the inability to read the VBIOS... --still tinkering with it. -- J.C. Roberts
