On Sat, 2004-04-24 at 00:14, S�rgio Monteiro Basto wrote: > now debug.txt > > On Fri, 2004-04-23 at 15:06, Brian Paul wrote: > > Felix K�hling wrote: > > > I can reproduce this here too. I got a backtrace (see below) and took a > > > look around s_linetemp.h. There is no INTERP_W for lines so the > > > assertion must fail for textured lines. > > > > > > Don't worry about the segfault, that's my own assert macro. Without that > > > I can't get a backtrace with my gdb+glibc combination. For the same > > > reason line numbers are off by 7 in s_span.c (sorry for the > > > inconvenience). > > > > > > s_span.c(572): interpolate_texcoords: Assertion 'span->interpMask & SPAN_W' > > > failed. > > > > I've checked in the fix. It was some code I mistakenly left in place > > from testing. > > > > -Brian well my point is this: First did you change this code in last two weeks ? if yes, ok could be a little problem in Mesa. if no, maybe is a problem with drm and new drm_pciids.h well, makes sense or not ? :) in drm/linux we should run: sh ../scripts/create_linux_pci_lists.sh < ../shared/drm_pciids.txt and this creates drm_pciids.h
after do this we can apply on drm dir: cd drm scripts/create_lk_drm.sh /usr/src/linux-2.4.26 2.4 This script create_lk_drm.sh they forgot to put in Config.in the S3 savage (my video card). lk.diff correct this issue about lk.diff don't know if one dependence to $CONFIG_AGP is correct but is copied from kernel-2.22-ac4. -- S�rgio M. B.
--- drivers/char/drm/Config.in 2004-04-23 00:17:30.000000000 +0100 +++ drm/linux/Config.in 2004-04-23 00:16:25.000000000 +0100 @@ -12,6 +12,7 @@ dep_tristate ' Intel I810' CONFIG_DRM_I810 $CONFIG_AGP dep_tristate ' Intel 830M/845G/852GM/855GM/865G' CONFIG_DRM_I830 $CONFIG_AGP dep_tristate ' Matrox g200/g400' CONFIG_DRM_MGA $CONFIG_AGP -tristate ' SiS' CONFIG_DRM_SIS -tristate ' Via Unichrome' CONFIG_DRM_VIA +dep_tristate ' SiS' CONFIG_DRM_SIS $CONFIG_AGP +dep_tristate ' S3 Savage' CONFIG_DRM_SAVAGE $CONFIG_AGP +tristate ' Via Unichrome or VIA CLE266' CONFIG_DRM_VIA
