Date: Sun, 16 Oct 2016 01:01:48 +0200 From: Bernd Walter <ticso@cicely7.cicely.de> To: Mark Millard <markmi@dsl-only.net> Cc: Bernd Walter <ticso@cicely7.cicely.de>, freebsd-arm <freebsd-arm@freebsd.org>, jbeich@freebsd.org, ticso@cicely.de Subject: Re: libx264 fails to build on a raspberry pi2 Message-ID: <20161015230148.GR99217@cicely7.cicely.de> In-Reply-To: <C51085A4-78BB-45D3-9EB0-4BE7B921C7DE@dsl-only.net> References: <20161014165021.GN99217@cicely7.cicely.de> <C51085A4-78BB-45D3-9EB0-4BE7B921C7DE@dsl-only.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Oct 14, 2016 at 11:48:40PM -0700, Mark Millard wrote:
> On 2016-Oct-14, at 9:50 AM, Bernd Walter <ticso at cicely7.cicely.de> wrote:
>
> > This seems to be a new problem, as I've installed the port before
> > on another Pi2 System (11-RC1):
> > libx264-0.144.2533_1 H.264/MPEG-4 AVC Video Encoding (Library)
> >
> > System is the official FreeBSD 11-RELEASE image for Pi2.
> > FreeBSD rpi2 11.0-RELEASE-p1 FreeBSD 11.0-RELEASE-p1 #0 r306420: Thu Sep 29 13:01:42 UTC 2016 root@releng2.nyi.freebsd.org:/usr/obj/arm.armv6/usr/src/sys/RPI2 arm
> >
> > [61]cicely1# svn info /usr/ports/multimedia/libx264
> > Path: multimedia/libx264
> > Working Copy Root Path: /usr/ports
> > URL: file:///data/freebsd-svn/ports/head/multimedia/libx264
> > Relative URL: ^/head/multimedia/libx264
> > Repository Root: file:///data/freebsd-svn/ports
> > Repository UUID: 35697150-7ecd-e111-bb59-0022644237b5
> > Revision: 423971
> > Node Kind: directory
> > Schedule: normal
> > Last Changed Author: jbeich
> > Last Changed Rev: 422675
> > Last Changed Date: 2016-09-23 15:15:06 +0200 (Fri, 23 Sep 2016)
> >
> > ...
> > cc -Wshadow -O3 -ffast-math -O -pipe -isystem /usr/local/include -fno-strict-aliasing -mcpu=arm1136jf-s -Wall -I. -I. -isystem /usr/local/include -O -pipe -isystem /usr/local/include -fno-strict-aliasing -mcpu=arm1136jf-s -std=gnu99 -D_GNU_SOURCE -fPIC -fomit-frame-pointer -fno-tree-vectorize -isystem /usr/local/include -c -o encoder/encoder.o encoder/encoder.c
>
> Above (and later below) -mcpu=arm1136jf-s is odd for an rpi2b. Is "36jf" right for any raspberry pi version? "76jzf" I'm aware of for older rpi's (if I remember right).
>
> -mcpu=cortex-a7 would match an rpi2b as I understand.
>
> For rpi2b I by default use:
>
> # more /etc/make.conf
> DEFAULT_VERSIONS+=perl5=5.22
> WRKDIRPREFIX=/usr/obj/portswork
> WITH_DEBUG=
> WITH_DEBUG_FILES=
> MALLOC_PRODUCTION=
> CFLAGS+= -mcpu=cortex-a7
> CXXFLAGS+= -mcpu=cortex-a7
> CPPFLAGS+= -mcpu=cortex-a7
>
> (and similarly for buildworld and buildkernel as well, not just ports).
>
>
> But figuring out why -mcpu=arm1136jf-s is in use for your rpi2b targeted build would seem to be a direction to go for what you were building.
I havn't touched make.conf beside adding WRKDIRPREFIX.
After adding all *FLAGS:
cc -Wshadow -O3 -ffast-math -O -pipe -mcpu=cortex-a7 -isystem /usr/local/include -fno-strict-aliasing -mcpu=arm1136jf-s -Wall -I. -I. -mcpu=cortex-a7 -isystem /usr/local/include -O -pipe -mcpu=cortex-a7 -isystem /usr/local/include -fno-strict-aliasing -mcpu=arm1136jf-s -std=gnu99 -D_GNU_SOURCE -fPIC -fomit-frame-pointer -fno-tree-vectorize -mcpu=cortex-a7 -isystem /usr/local/include -c -o common/opencl.o common/opencl.c
cc -Wshadow -O3 -ffast-math -O -pipe -mcpu=cortex-a7 -isystem /usr/local/include -fno-strict-aliasing -mcpu=arm1136jf-s -Wall -I. -I. -mcpu=cortex-a7 -isystem /usr/local/include -O -pipe -mcpu=cortex-a7 -isystem /usr/local/include -fno-strict-aliasing -mcpu=arm1136jf-s -std=gnu99 -D_GNU_SOURCE -fPIC -fomit-frame-pointer -fno-tree-vectorize -mcpu=cortex-a7 -isystem /usr/local/include -c -o encoder/slicetype-cl.o encoder/slicetype-cl.c
cc -I. -I. -c -DSTACK_ALIGNMENT=4 -DPIC -DHIGH_BIT_DEPTH=0 -DBIT_DEPTH=8 -o common/arm/cpu-a.o common/arm/cpu-a.S
cc -I. -I. -c -DSTACK_ALIGNMENT=4 -DPIC -DHIGH_BIT_DEPTH=0 -DBIT_DEPTH=8 -o common/arm/pixel-a.o common/arm/pixel-a.S
cc -I. -I. -c -DSTACK_ALIGNMENT=4 -DPIC -DHIGH_BIT_DEPTH=0 -DBIT_DEPTH=8 -o common/arm/mc-a.o common/arm/mc-a.S
cc -I. -I. -c -DSTACK_ALIGNMENT=4 -DPIC -DHIGH_BIT_DEPTH=0 -DBIT_DEPTH=8 -o common/arm/dct-a.o common/arm/dct-a.S
cc -I. -I. -c -DSTACK_ALIGNMENT=4 -DPIC -DHIGH_BIT_DEPTH=0 -DBIT_DEPTH=8 -o common/arm/quant-a.o common/arm/quant-a.S
cc -I. -I. -c -DSTACK_ALIGNMENT=4 -DPIC -DHIGH_BIT_DEPTH=0 -DBIT_DEPTH=8 -o common/arm/deblock-a.o common/arm/deblock-a.S
<instantiation>:29:5: error: instruction requires: armv6t2
rbit r1, r1
^
common/arm/quant-a.S:362:1: note: while in macro instantiation
decimate_score_1x 15
^
<instantiation>:29:5: error: instruction requires: armv6t2
rbit r1, r1
^
/tmp/quant-a-967f0b.s:515:1: note: while in macro instantiation
decimate_score_1x 16
^
gmake[2]: *** [Makefile:217: common/arm/quant-a.o] Error 1
gmake[2]: *** Waiting for unfinished jobs....
gmake[2]: Leaving directory '/usr/obj/usr/ports/multimedia/libx264/work/x264-snapshot-20160921-2245-stable'
===> Compilation failed unexpectedly.
That -mcpu=arm1136jf-s is something from the multimedia/x264/Makefile,
on which the libx264 port is based on.
413179 jbeich CFLAGS_armv6+= -mcpu=arm1136jf-s
However the problem is at least with one of the assembly files, which
don't get those args at all.
--
B.Walter <bernd@bwct.de> http://www.bwct.de
Modbus/TCP Ethernet I/O Baugruppen, ARM basierte FreeBSD Rechner uvm.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20161015230148.GR99217>