| Message ID | 20170606171346.21413-1-raj.khem@gmail.com |
|---|---|
| State | Accepted |
| Commit | 6866ad53c5d49781002470195b4aae1ad1afde5c |
| Headers | show |
On 6 June 2017 at 18:13, Khem Raj <raj.khem@gmail.com> wrote: > This is wrongly detected by configure since it only checks for > __sync_add_and_fetch and not __atomic_fetch_add, it decides > to let compiler emit references to atomic functions which > it expects to be provided by libgcc, but thats not the > case for armv5 arch > Is the true fix to adjust configure to look for the right operations? Ross -- _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
On Tue, Jun 6, 2017 at 10:32 AM, Burton, Ross <ross.burton@intel.com> wrote: > > On 6 June 2017 at 18:13, Khem Raj <raj.khem@gmail.com> wrote: >> >> This is wrongly detected by configure since it only checks for >> __sync_add_and_fetch and not __atomic_fetch_add, it decides >> to let compiler emit references to atomic functions which >> it expects to be provided by libgcc, but thats not the >> case for armv5 arch > > > Is the true fix to adjust configure to look for the right operations? > yes. and there is such a patch proposed as well to mesa but it has got some feedback so I did not bring that in. This patch instead is surgical. > Ross -- _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
On 6 June 2017 at 18:33, Khem Raj <raj.khem@gmail.com> wrote: > yes. and there is such a patch proposed as well to mesa but > it has got some feedback so I did not bring that in. This patch > instead is surgical. > Cool. Just got to remember to back this out when upstream is fixed. Thanks, Ross -- _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
On Tue, Jun 6, 2017 at 11:02 AM, Burton, Ross <ross.burton@intel.com> wrote: > > On 6 June 2017 at 18:33, Khem Raj <raj.khem@gmail.com> wrote: >> >> yes. and there is such a patch proposed as well to mesa but >> it has got some feedback so I did not bring that in. This patch >> instead is surgical. > > > Cool. Just got to remember to back this out when upstream is fixed. > yes, although effects will be same. > Thanks, > Ross -- _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
diff --git a/meta/recipes-graphics/mesa/mesa.inc b/meta/recipes-graphics/mesa/mesa.inc index 67039a9962..014044f381 100644 --- a/meta/recipes-graphics/mesa/mesa.inc +++ b/meta/recipes-graphics/mesa/mesa.inc @@ -79,6 +79,8 @@ PACKAGECONFIG[unwind] = "--enable-libunwind,--disable-libunwind,libunwind" # llvmpipe is slow if compiled with -fomit-frame-pointer (e.g. -O2) FULL_OPTIMIZATION_append = " -fno-omit-frame-pointer" +CFLAGS_append_armv5 = " -DMISSING_64BIT_ATOMICS" + # Multiple virtual/gl providers being built breaks staging EXCLUDE_FROM_WORLD = "1"
This is wrongly detected by configure since it only checks for __sync_add_and_fetch and not __atomic_fetch_add, it decides to let compiler emit references to atomic functions which it expects to be provided by libgcc, but thats not the case for armv5 arch [YOCTO #11616] Signed-off-by: Khem Raj <raj.khem@gmail.com> --- meta/recipes-graphics/mesa/mesa.inc | 2 ++ 1 file changed, 2 insertions(+) -- 2.13.0 -- _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core