[FFmpeg-devel] [PATCH v2] avcodec/arm: Fix SIGBUS on ARM when compiled with binutils 2.29

[James Cowgill]

In binutils 2.29, the behavior of the ADR instruction changed so that 1 is
added to the address of a Thumb function (previously nothing was added). This
allows the loaded address to be passed to a BLX instruction and the correct
mode change will occur.

So that the behavior matches in binutils 2.29 and pre-2.29, use .eqv to
pre-calculate the function address without the automatic +1 fixup. Then use
these new symbols as the function addresses to be loaded.

Fixes ticket 6571.
Related binutils bug: https://sourceware.org/bugzilla/show_bug.cgi?id=21458

Signed-off-by: James Cowgill <jcowgill at debian.org>
---
v2:
 Forgot to include the "avcodec/arm" commit message prefix.

 libavcodec/arm/h264idct_neon.S | 28 ++++++++++++++++++++--------
 1 file changed, 20 insertions(+), 8 deletions(-)

diff --git a/libavcodec/arm/h264idct_neon.S b/libavcodec/arm/h264idct_neon.S
index 4f68bdb9f5..04b1ea583b 100644
--- a/libavcodec/arm/h264idct_neon.S
+++ b/libavcodec/arm/h264idct_neon.S
@@ -20,6 +20,18 @@
 
 #include "libavutil/arm/asm.S"
 
+# In binutils 2.29, the behavior of the ADR instruction changed so that 1 is
+# added to the address of a Thumb function (previously nothing was added).
+#
+# These .eqv are used to pre-calculate the correct address with +CONFIG_THUMB so
+# that ADR will work with both old and new versions binutils.
+#
+# See: https://sourceware.org/bugzilla/show_bug.cgi?id=21458
+.eqv eqv_ff_h264_idct_add_neon,     X(ff_h264_idct_add_neon) + CONFIG_THUMB
+.eqv eqv_ff_h264_idct_dc_add_neon,  X(ff_h264_idct_dc_add_neon) + CONFIG_THUMB
+.eqv eqv_ff_h264_idct8_add_neon,    X(ff_h264_idct8_add_neon) + CONFIG_THUMB
+.eqv eqv_ff_h264_idct8_dc_add_neon, X(ff_h264_idct8_dc_add_neon) + CONFIG_THUMB
+
 function ff_h264_idct_add_neon, export=1
         vld1.64         {d0-d3},  [r1,:128]
         vmov.i16        q15, #0
@@ -113,8 +125,8 @@ function ff_h264_idct_add16_neon, export=1
         movne           lr,  #0
         cmp             lr,  #0
         ite             ne
-        adrne           lr,  X(ff_h264_idct_dc_add_neon) + CONFIG_THUMB
-        adreq           lr,  X(ff_h264_idct_add_neon)    + CONFIG_THUMB
+        adrne           lr,  eqv_ff_h264_idct_dc_add_neon
+        adreq           lr,  eqv_ff_h264_idct_add_neon
         blx             lr
 2:      subs            ip,  ip,  #1
         add             r1,  r1,  #32
@@ -138,8 +150,8 @@ function ff_h264_idct_add16intra_neon, export=1
         cmp             r8,  #0
         ldrsh           r8,  [r1]
         iteet           ne
-        adrne           lr,  X(ff_h264_idct_add_neon)    + CONFIG_THUMB
-        adreq           lr,  X(ff_h264_idct_dc_add_neon) + CONFIG_THUMB
+        adrne           lr,  eqv_ff_h264_idct_add_neon
+        adreq           lr,  eqv_ff_h264_idct_dc_add_neon
         cmpeq           r8,  #0
         blxne           lr
         subs            ip,  ip,  #1
@@ -166,8 +178,8 @@ function ff_h264_idct_add8_neon, export=1
         cmp             r8,  #0
         ldrsh           r8,  [r1]
         iteet           ne
-        adrne           lr,  X(ff_h264_idct_add_neon)    + CONFIG_THUMB
-        adreq           lr,  X(ff_h264_idct_dc_add_neon) + CONFIG_THUMB
+        adrne           lr,  eqv_ff_h264_idct_add_neon
+        adreq           lr,  eqv_ff_h264_idct_dc_add_neon
         cmpeq           r8,  #0
         blxne           lr
         add             r12, r12, #1
@@ -388,8 +400,8 @@ function ff_h264_idct8_add4_neon, export=1
         movne           lr,  #0
         cmp             lr,  #0
         ite             ne
-        adrne           lr,  X(ff_h264_idct8_dc_add_neon) + CONFIG_THUMB
-        adreq           lr,  X(ff_h264_idct8_add_neon)    + CONFIG_THUMB
+        adrne           lr,  eqv_ff_h264_idct8_dc_add_neon
+        adreq           lr,  eqv_ff_h264_idct8_add_neon
         blx             lr
 2:      subs            r12, r12, #4
         add             r1,  r1,  #128
-- 
2.14.1