On 1/12/15, Landry Breuil <lan...@rhaalovely.net> wrote:
> On Mon, Jan 12, 2015 at 12:48:25PM -0800, patrick keshishian wrote:
>> On 1/12/15, patrick keshishian <pkesh...@gmail.com> wrote:
>> > On 1/12/15, Landry Breuil <lan...@rhaalovely.net> wrote:
>> >> On Mon, Jan 12, 2015 at 12:29:00PM -0800, patrick keshishian wrote:
>> >>> On 1/10/15, Landry Breuil <lan...@rhaalovely.net> wrote:
>> >>> > Hi,
>> >>> >
>> >>> > here's the wip update to ffx 35.0rc3, ETA next wednesday. Been
>> >>> > using
>> >>> > 35.0betas on amd64 for the past 6 weeks without issues, also tested
>> >>> > on
>> >>> > i386 and powerpc.
>> >>> >
>> >>> > http://cgit.rhaalovely.net/mozilla-firefox/?h=release
>> >>>
>> >>> very possible this is my fault but ... the build fails for me on:
>> >>>
>> >>> kern.version=OpenBSD 5.7-beta (GENERIC) #701: Sat Jan 10 07:52:06 MST
>> >>> 2015
>> >>> dera...@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC
>> >>>
>> >>> with:
>> >>> ...
>> >>> INPUT("../../gfx/skia/SkLayer.o")
>> >>> INPUT("../../gfx/skia/SkMatrix22.o")
>> >>> INPUT("../../gfx/skia/SkMatrix44.o")
>> >>> INPUT("../../gfx/skia/SkMeshUtils.o")
>> >>> INPUT("../../gfx/skia/SkNWayCanvas.o")
>> >>> INPUT("../../gfx/skia/SkNinePatch.o")
>> >>> INPUT("../../gfx/skia/SkNullCanvas.o")
>> >>> INPUT("../../gfx/skia/SkOSFile.o")
>> >>> INPUT("../../gfx/skia/SkParseColor.o")
>> >>> INPUT("../../gfx/skia/SkParsePath.o")
>> >>> INPUT("../../gfx/skia/SkPathUtils.o")
>> >>> INPUT("../../gfx/skia/SkPictureUtils.o")
>> >>> INPUT("../../gfx/skia/SkProxyCanvas.o")
>> >>> INPUT("../../gfx/skia/SkRTConf.o")
>> >>> INPUT("../../gfx/skia/SkTextureCompressor.o")
>> >>> INPUT("../../gfx/skia/SkTextureCompressor_ASTC.o")
>> >>> INPUT("../../gfx/skia/SkTextureCompressor_LATC.o")
>> >>> INPUT("../../gfx/skia/SkTextureCompressor_R11EAC.o")
>> >>> INPUT("../../gfx/skia/SkThreadUtils_pthread.o")
>> >>> INPUT("StaticXULComponentsEnd/StaticXULComponentsEnd.o")
>> >>>
>> >>> /usr/bin/ld: Warning: gc-sections option ignored
>> >>> ../../gfx/skia/opts_check_x86.o(.data.rel.ro.local._ZL22platform_32_procs_SSE4+0x10):
>> >>> In function `SkBoxBlurGetPlatformProcs(void (**)(unsigned int const*,
>> >>> int, unsigned int*, int, int, int, int, int), void (**)(unsigned int
>> >>> const*, int, unsigned int*, int, int, int, int, int), void
>> >>> (**)(unsigned int const*, int, unsigned int*, int, int, int, int,
>> >>> int), void (**)(unsigned int const*, int, unsigned int*, int, int,
>> >>> int, int, int))':
>> >>> /usr/build/ports/pobj/firefox-35.0rc3/mozilla-release/gfx/skia/trunk/src/opts/opts_check_x86.cpp:212:
>> >>> undefined reference to `S32A_Opaque_BlitRow32_SSE4_asm'
>> >>> /usr/bin/ld: libxul.so.53.0: hidden symbol
>> >>> `S32A_Opaque_BlitRow32_SSE4_asm' isn't defined
>> >>
>> >> What does 'grep HAVE_TOOLCHAIN
>> >> /usr/build/ports/pobj/firefox-35.0rc3/build-amd64/config.status'
>> >> says ?
>> >
>> > (''' HAVE_TOOLCHAIN_SUPPORT_MSSSE3 ''', r''' 1 '''),
>> > (''' HAVE_TOOLCHAIN_SUPPORT_MSSE4_1 ''', r''' 1 '''),
>>
>> full dmesg in case it is of more help:
>> OpenBSD 5.7-beta (GENERIC) #701: Sat Jan 10 07:52:06 MST 2015
>> dera...@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC
>> real mem = 1861025792 (1774MB)
>> avail mem = 1807708160 (1723MB)
>> mpath0 at root
>> scsibus0 at mpath0: 256 targets
>> mainbus0 at root
>> bios0 at mainbus0: SMBIOS rev. 2.4 @ 0xf1010 (17 entries)
>> bios0: vendor Phoenix Technologies LTD version "v1.3307" date 05/31/2010
>> bios0: Gateway LT31
>> acpi0 at bios0: rev 2
>> acpi0: sleep states S0 S3 S4 S5
>> acpi0: tables DSDT FACP APIC MCFG HPET BOOT SLIC
>> acpi0: wakeup devices PB5_(S5) OHC1(S3) OHC2(S3) EHCI(S3) HDAU(S3)
>> acpitimer0 at acpi0: 3579545 Hz, 32 bits
>> acpimadt0 at acpi0 addr 0xfee00000: PC-AT compat
>> cpu0 at mainbus0: apid 0 (boot processor)
>> cpu0: AMD Athlon(tm) Processor L110, 1197.20 MHz
>> cpu0:
>> FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CFLUSH,MMX,FXSR,SSE,SSE2,SSE3,CX16,NXE,MMXX,FFXSR,LONG,3DNOW2,3DNOW,LAHF,SVM,EAPICSP,AMCR8,3DNOWP
>
> Interesting, your cpu doesnt have SSSE3 nor SSE4.1, while binutils/the
> configure script detects so.. that might explain why it built here and
> not on your machine. That doesnt explain why configure things they're
> here though...
I don't understand all this configure scaffolding jazz, but if I'm
looking at the right bits, the configure script is only testing
to see if the the compile command succeeds:
configure:7455: checking if toolchain supports -mssse3 option
configure:7467: cc -c -O2 -pipe -std=gnu99 -fgnu89-inline
-fno-strict-aliasing -mssse3 -Qunused-arguments conftest.c 1>&5
Now, i can't quite figure out where cc is aliased to clang,
but /usr/bin/cc fails with "unrecognized option" -Qunused-arguments
and -mssse3, but clang runs w/o complain:
$ cat moo.c
asm ("pmaddubsw %xmm2,%xmm3");
int main() {
; return 0; }
$ clang -c -O2 -pipe -std=gnu99 -fgnu89-inline -fno-strict-aliasing
-mssse3 -Qunused-arguments moo.c && echo Woo!
Woo!
The C file is a snippet from configure, w/o the #line or confdefs.h
(which I don't know how to regenerate).
thoughts?
--patrick
>
> Landry
>
