On Thu, 29 May 2014, Reinhard Tartler wrote:
On Thu, May 29, 2014 at 8:51 AM, Martin Storsjö <[email protected]> wrote:
From: Thierry Fauck <[email protected]>
---
Patch taken from
https://bugs.launchpad.net/ubuntu/+source/libav/+bug/1263802,
with some minor fixups.
This fixes compilation for this target (tested with a cross compiler),
but it's completely untested in practice so far (qemu doesn't support
it yet, and I don't have any such hardware to test it on).
---
libavcodec/ppc/asm.S | 8 ++++++++
1 file changed, 8 insertions(+)
diff --git a/libavcodec/ppc/asm.S b/libavcodec/ppc/asm.S
index 4d4285b..141dee9 100644
--- a/libavcodec/ppc/asm.S
+++ b/libavcodec/ppc/asm.S
@@ -36,12 +36,20 @@
.macro extfunc name
.global X(\name)
+#if _CALL_ELF == 2
+ .text
+X(\name):
+ addis %r2, %r12, .TOC.-X(\name)@ha
+ addi %r2, %r2, .TOC.-X(\name)@l
+ .localentry X(\name), .-X(\name)
+#else
.section .opd, "aw"
X(\name):
.quad L(\name), .TOC.@tocbase, 0
.previous
.type X(\name), STT_FUNC
L(\name):
+#endif
.endm
.macro movrel rd, sym, gp
--
1.8.5.2 (Apple Git-48)
Patch LGTM and builds successfully in Ubuntu/utopic:
https://launchpad.net/ubuntu/+source/libav/6:10.1-1ubuntu1/+build/6052839
Please CC: stable and push this patch to release/10.
Is there any way we could get this patch tested with a "make fate" on such
a machine? Since this deals with handwritten assembly (which this patch
enables building) there's a big chance of it not working unless the
contrary has been proven, and shipping such builds is a bit adventurous.
// Martin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
