Re: Fwd: Bug#654329: libdrm: FTBFS when atomic operations are only available through libatomic-ops-dev
Thorsten Glaser wrote:
> Without words. (The patch in question is a one-liner adding a
> Build-Depends line with [m68k] appended to it, so it’d absolutely
> do no harm to other architectures.)
Are there 68k machines that can use the relevant radeon cards? If
not, a better fix might be to use --disable-radeon (but it does looks
like whatever fix is taken will probably not be at ftp.debian.org
until the m68k revival is further along).
Untested, of course.
diff --git i/debian/rules w/debian/rules
index 4c296ea3..4565b2d3 100755
--- i/debian/rules
+++ w/debian/rules
@@ -16,8 +16,6 @@ ifeq (linux, $(DEB_HOST_ARCH_OS))
confflags += --enable-vmwgfx-experimental-api
confflags += --enable-nouveau-experimental-api
NOUVEAU = yes
- confflags += --enable-radeon
- RADEON = yes
else
confflags += --disable-udev
confflags += --disable-libkms
@@ -25,11 +23,10 @@ else
confflags += --disable-vmwgfx-experimental-api
confflags += --disable-nouveau-experimental-api
NOUVEAU = no
- confflags += --disable-radeon
- RADEON = no
endif
# Intel is only on x86:
+INTEL = no
ifneq (,$(filter amd64 i386,$(DEB_HOST_ARCH_CPU)))
ifneq (,$(filter linux kfreebsd,$(DEB_HOST_ARCH_OS)))
INTEL = yes
@@ -41,6 +38,19 @@ else
confflags += --disable-intel
endif
+# No Radeon on m68k:
+RADEON = no
+ifeq (,$(filter m68k,$(DEB_HOST_ARCH_CPU))
+ifneq (,$(filter linux,$(DEB_HOST_ARCH_OS)))
+ RADEON = yes
+endif
+endif
+ifeq ($(RADEON), yes)
+ confflags += --enable-radeon
+else
+ confflags += --disable-radeon
+endif
+
###
### Actual build
###
--
Reply to: