* Update (lib)x264 to 20190324 * Stop using GNU Autotools and use libx264's own configuration facility * Drop hardcoded CFLAGS, x264 will handle those fine on its own This will override toolchain optimizaion and set -O3 irregardless of setting. * Rework LTO and ASM optmization selection to make it more compact and readable. This drops optimization for x86 32-bit which is being deprecated in favour of x86_64 in general and the very few systems still in use that doesn't support 64-bit are too slow to be usable anyway. * Import patches to fix compilation on ARM and x86 (32-bit) from OpenEmbedded * Minor style fixes to Makefile Source: http://cgit.openembedded.org/openembedded-core/tree/meta/recipes-multimedia/x264/x264 Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
28 lines
979 B
Diff
28 lines
979 B
Diff
From a72bf499a0674fc75eedf15008b424e28f67e4bd Mon Sep 17 00:00:00 2001
|
|
From: Andrei Gherzan <andrei@gherzan.ro>
|
|
Date: Fri, 2 Feb 2018 15:10:08 +0200
|
|
Subject: [PATCH] dont default to cortex-a9 with neon
|
|
|
|
-march flag is not in CFLAGS so this will always default to
|
|
-mcpu=cortex-a8 -mfpu=neon.
|
|
|
|
Upstream-Status: Pending
|
|
|
|
Signed-off-by: Andrei Gherzan <andrei@gherzan.ro>
|
|
Signed-off-by: Maxin B. John <maxin.john@intel.com>
|
|
---
|
|
configure | 3 ---
|
|
1 file changed, 3 deletions(-)
|
|
|
|
--- a/configure
|
|
+++ b/configure
|
|
@@ -940,9 +940,6 @@ if [ $asm = auto -a \( $ARCH = X86 -o $A
|
|
fi
|
|
|
|
if [ $asm = auto -a $ARCH = ARM ] ; then
|
|
- # set flags so neon is built by default
|
|
- [ $compiler == CL ] || echo $CFLAGS | grep -Eq '(-mcpu|-march|-mfpu)' || CFLAGS="$CFLAGS -mcpu=cortex-a8 -mfpu=neon"
|
|
-
|
|
cc_check '' '' '__asm__("add r0, r1, r2");' && define HAVE_ARM_INLINE_ASM
|
|
if [ $compiler = CL ] && cpp_check '' '' 'defined(_M_ARM) && _M_ARM >= 7' ; then
|
|
define HAVE_ARMV6
|