Em 29-04-2014 10:21, Pierre Labastie escreveu:
> Le 29/04/2014 13:21, Fernando de Oliveira a écrit :
>> Em 28-04-2014 20:22, Bruce Dubbs escreveu:
>>> Fernando de Oliveira wrote:
>>>> $ cat /etc/lfs-release
>>>> SVN-20140425
>>>>
>>>> Log is a bit long.
>>>>
>>>> I am thinking the problem is with gcc-4.9.0, but have seen reference to
>>>> older problems with ncurses.
>>>>
>>>> Please, I need to help with that.
>>>
>>> Did you see: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56145
>>>
>>> It's a little dated, but it does not appear to be fixed.
>>>
>>> -- Bruce
>>
>> Yes, I saw it before, thanks.
>>
>> Tried
>>
>> -O1 instead of O3 (CFLAGS="$CFLAGS -O1")
>>
>> Then tried:
>>
>> CFLAGS="$CFLAGS -fwrapv"
>>
>> because they say "When adding -fwrapv to the command line it stops gcc
>> from produce the ice".
>>
>> No success at all.
>>
>> I don't have enough knowledge to understand all said in the page, so I
>> am not sure of exactly what could be done.
>>
> I have just built lame (not tested) with gcc-4.9.0 I installed for updating,
> on an LFS 7.5 basis, using instructions in the BFLS book, and only NASM as
> optional dependency. It builds fine, but I have not tested it.
>
> The difference I notice with your log is the -maccumulate-outgoing-args
> switch, which I do not have. But I see in the doc that this flag is x86 only,
> so it might not appear on a 64 bit machine, which I use.
>
> The warning about the v_s2 variable is here too, but not the SSE one, nor the
> error, of course.
>
> I have a 64 bit virtual machine under qemu. I see that you have a 32 bit one.
> The warning about SSE is weird, actually. What do you have for:
> cat /proc/cpuinfo | grep flags
> ?
Thanks for the reply.
Before my answer: I have just solved the problem and it seems indeed to
be only for 32bit, will describe further below.
{{{
$ cat /proc/cpuinfo | grep flags
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush dts mmx fxsr sse sse2 ss nx rdtscp lm constant_tsc
arch_perfmon pebs bts xtopology nonstop_tsc aperfmperf pni ssse3 cx16
sse4_1 sse4_2 x2apic popcnt hypervisor lahf_lm ida dtherm
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush dts mmx fxsr sse sse2 ss nx rdtscp lm constant_tsc
arch_perfmon pebs bts xtopology nonstop_tsc aperfmperf pni ssse3 cx16
sse4_1 sse4_2 x2apic popcnt hypervisor lahf_lm ida dtherm
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush dts mmx fxsr sse sse2 ss nx rdtscp lm constant_tsc
arch_perfmon pebs bts xtopology nonstop_tsc aperfmperf pni ssse3 cx16
sse4_1 sse4_2 x2apic popcnt hypervisor lahf_lm ida dtherm
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov
pat pse36 clflush dts mmx fxsr sse sse2 ss nx rdtscp lm constant_tsc
arch_perfmon pebs bts xtopology nonstop_tsc aperfmperf pni ssse3 cx16
sse4_1 sse4_2 x2apic popcnt hypervisor lahf_lm ida dtherm
}}}
Now, how I solved from a completely different machine/archtecture.
Re:
https://github.com/HandBrake/HandBrake/blob/master/contrib/lame/P00-darwin-xmmintrin.patch
I adapted that patch to lame-3.99.5-gcc_4_9_0_i686_fix-1.patch:
{{{
--- lame-3.99.5.orig/configure 2012-02-28 15:54:37.000000000 -0300
+++ lame-3.99.5/configure 2014-04-29 11:15:10.118691341 -0300
@@ -11913,6 +11913,15 @@
fi
+case $host_cpu in
+i686*)
+ check_xmmintrin=
+ ;;
+*)
+ check_xmmintrin="xmmintrin.h"
+ ;;
+esac
+
for ac_header in \
errno.h \
fcntl.h \
@@ -11922,7 +11931,7 @@
sys/soundcard.h \
sys/time.h \
unistd.h \
- xmmintrin.h \
+ $check_xmmintrin \
linux/soundcard.h
do :
as_ac_Header=`$as_echo "ac_cv_header_$ac_header" | $as_tr_sh`
}}}
Then it finished make successfully. Not yet tested.
Is that patch fine to be included, or would be dangerous for other
architectures?
--
[]s,
Fernando
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page
