All of lore.kernel.org
 help / color / mirror / Atom feed
* [Buildroot] [PATCH 1/1] ffmpeg: fix neon support for aarch64
@ 2017-04-14  7:08 Bernd Kuhls
  2017-04-14  8:54 ` Thomas Petazzoni
  0 siblings, 1 reply; 2+ messages in thread
From: Bernd Kuhls @ 2017-04-14  7:08 UTC (permalink / raw)
  To: buildroot

ffmpeg ported arm 32-bit NEON code to aarch64 as hard dependency:
http://ffmpeg.org/pipermail/ffmpeg-devel/2016-November/202733.html

Buildroot option BR2_ARM_CPU_HAS_NEON handles only the 32bit case:
https://git.buildroot.net/buildroot/commit/?id=0976cd6cd64a128a2ac921e4f35f0e7cbe306692

For aarch64 we need to make sure to pass --enable-neon in order to
prevent a link error, fixes

http://autobuild.buildroot.net/results/fa4/fa46003249664b1f97f6c18253c84cd3d6fb9cdb/

Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
---
 package/ffmpeg/ffmpeg.mk | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/package/ffmpeg/ffmpeg.mk b/package/ffmpeg/ffmpeg.mk
index 059f8320e..c0c6705e1 100644
--- a/package/ffmpeg/ffmpeg.mk
+++ b/package/ffmpeg/ffmpeg.mk
@@ -432,9 +432,12 @@ FFMPEG_CONF_OPTS += --disable-vfp
 endif
 ifeq ($(BR2_ARM_CPU_HAS_NEON),y)
 FFMPEG_CONF_OPTS += --enable-neon
+ifeq ($(BR2_aarch64),y)
+FFMPEG_CONF_OPTS += --enable-neon
 else
 FFMPEG_CONF_OPTS += --disable-neon
 endif
+endif
 
 ifeq ($(BR2_mips)$(BR2_mipsel)$(BR2_mips64)$(BR2_mips64el),y)
 ifeq ($(BR2_MIPS_SOFT_FLOAT),y)
-- 
2.11.0

^ permalink raw reply related	[flat|nested] 2+ messages in thread

* [Buildroot] [PATCH 1/1] ffmpeg: fix neon support for aarch64
  2017-04-14  7:08 [Buildroot] [PATCH 1/1] ffmpeg: fix neon support for aarch64 Bernd Kuhls
@ 2017-04-14  8:54 ` Thomas Petazzoni
  0 siblings, 0 replies; 2+ messages in thread
From: Thomas Petazzoni @ 2017-04-14  8:54 UTC (permalink / raw)
  To: buildroot

Hello,

On Fri, 14 Apr 2017 09:08:48 +0200, Bernd Kuhls wrote:
> ffmpeg ported arm 32-bit NEON code to aarch64 as hard dependency:
> http://ffmpeg.org/pipermail/ffmpeg-devel/2016-November/202733.html
> 
> Buildroot option BR2_ARM_CPU_HAS_NEON handles only the 32bit case:
> https://git.buildroot.net/buildroot/commit/?id=0976cd6cd64a128a2ac921e4f35f0e7cbe306692
> 
> For aarch64 we need to make sure to pass --enable-neon in order to
> prevent a link error, fixes
> 
> http://autobuild.buildroot.net/results/fa4/fa46003249664b1f97f6c18253c84cd3d6fb9cdb/
> 
> Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
> ---
>  package/ffmpeg/ffmpeg.mk | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/package/ffmpeg/ffmpeg.mk b/package/ffmpeg/ffmpeg.mk
> index 059f8320e..c0c6705e1 100644
> --- a/package/ffmpeg/ffmpeg.mk
> +++ b/package/ffmpeg/ffmpeg.mk
> @@ -432,9 +432,12 @@ FFMPEG_CONF_OPTS += --disable-vfp
>  endif
>  ifeq ($(BR2_ARM_CPU_HAS_NEON),y)
>  FFMPEG_CONF_OPTS += --enable-neon
> +ifeq ($(BR2_aarch64),y)
> +FFMPEG_CONF_OPTS += --enable-neon

I don't understand this patch. How can it work? We're already in a
BR2_ARM_CPU_HAS_NEON condition, and if it's true we're passing
--enable-neon.

Are you sure you didn't want to use an "else ifeq" here ? Are you sure
your patch fixes anything ?

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2017-04-14  8:54 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-04-14  7:08 [Buildroot] [PATCH 1/1] ffmpeg: fix neon support for aarch64 Bernd Kuhls
2017-04-14  8:54 ` Thomas Petazzoni

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.