Summary: | Make xf86-video-geode fail more gracefully on non-x86 | ||
---|---|---|---|
Product: | xorg | Reporter: | Jeremy Huddleston Sequoia <jeremyhu> |
Component: | Driver/geode | Assignee: | X.org Geode Mailing List <xorg-driver-geode> |
Status: | RESOLVED MOVED | QA Contact: | X.org Geode Mailing List <xorg-driver-geode> |
Severity: | normal | ||
Priority: | medium | CC: | jon.turney, matthieu.herrb |
Version: | git | ||
Hardware: | PowerPC | ||
OS: | Linux (All) | ||
Whiteboard: | 2011BRB_Reviewed | ||
i915 platform: | i915 features: |
Description
Jeremy Huddleston Sequoia
2011-03-24 23:48:10 UTC
(In reply to comment #0) > http://tinderbox.x.org/builds/2011-03-25-0000/logs/xf86-video-geode/#build > > > /bin/bash ../libtool --silent --tag=CC --mode=compile gcc -DHAVE_CONFIG_H -I. > -I.. -I../src/cim -I../src/gfx -I../src/panel -I../linux_v26 > -I/home/jeremy/src/freedesktop/jhbuild/build/include > -I/home/jeremy/src/freedesktop/jhbuild/external/build/include > -fvisibility=hidden -I/home/jeremy/src/freedesktop/jhbuild/build/include/xorg > -I/home/jeremy/src/freedesktop/jhbuild/build/include/pixman-1 -DHAVE_GX > -DHAVE_LX -DAMD_V4L2_VIDEO -DOPT_ACCEL -DLINUX_2_6 -Wall -Wpointer-arith > -Wstrict-prototypes -Wmissing-prototypes -Wmissing-declarations > -Wnested-externs -fno-strict-aliasing -DPNL_SUP -O0 -pipe -Wall -Wformat=2 -c > -o geode_ddc.lo geode_ddc.c > geode_common.c: In function 'geode_memory_to_screen_blt': > geode_common.c:86:6: error: impossible constraint in 'asm' > geode_common.c:93:6: error: impossible constraint in 'asm' > geode_common.c:100:6: error: impossible constraint in 'asm' > geode_common.c:107:6: error: impossible constraint in 'asm' > make[2]: Leaving directory > `/home/jeremy/src/freedesktop/jhbuild/src/xorg/driver/xf86-video-geode/src' I don't think it makes sense to build the geode driver on powerpc (or any other non x86 architecture) since afaik the geode chipset in integrated in AMD geode cpus. I figured as much, but wanted to bring it up. I'll just add it to the skip list for the tinderbox. If it ever lands in an addon package, this should be addressed. That's expected, since this driver is x86-specific. Actually, someone (memsize?) was pondering whether it would be appropriate to have autoconf scripts generate a nearly-empty Makefile whose only purpose is to echo "x86-specific driver. Skipping this one." on non-x86 architectures. I think it would be more appropriate for configure to fail with: AC_ERROR_MSG("Your architecture is not supported by this driver") And we can update jhbuildrc and build.sh to skip it on other architectures. Another somewhat related item was about whether we should implement enforcement of -m32 on x86-64 hosts to allow cross-compiling. As I recall, we never came to a definitive conclusion on that one either. (In reply to comment #5) > I think it would be more appropriate for configure to fail with: > > AC_ERROR_MSG("Your architecture is not supported by this driver") This makes sense > And we can update jhbuildrc and build.sh to skip it on other architectures. But this is also needed, so tinderbox knows not to try to build it :-) kibi proposed a patch which creates a standard python fragment for jhbuild which adds a target-specific skip list. See http://lists.x.org/archives/xorg-devel/2010-November/015514.html Does commit b9afafec0a26eaec49e7f0ff08297c2bb795a9c2 produce a suitable fix for this issue? If yes, let's close it. If not, what is missing? Any news on this? -- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/xorg/driver/xf86-video-geode/issues/3. |
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.