similar to: [PATCH] configure: only use -mstackrealign for mingw32

Tristan Matthews wrote: > --- > configure.ac | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > diff --git a/configure.ac b/configure.ac > index eb9b0cc..e7d68c3 100644 > --- a/configure.ac > +++ b/configure.ac > @@ -399,9 +399,11 @@ if test x$ac_cv_c_compiler_gnu = xyes ; then > > if test "x$asm_optimisation$sse_os" =

On Fri, Apr 10, 2015 at 1:40 PM, lvqcl <lvqcl.mail at gmail.com> wrote: > Tristan Matthews wrote: > > > if test "x$asm_optimisation$sse_os" = "xyesyes" ; then > > XIPH_ADD_CFLAGS([-msse2]) > > - XIPH_ADD_CFLAGS([-mstackrealign]) > > + if test "$host_os" = "mingw32" ; then >

And only for i686. --- configure.ac | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/configure.ac b/configure.ac index eb9b0cc..8dd5b0d 100644 --- a/configure.ac +++ b/configure.ac @@ -399,11 +399,16 @@ if test x$ac_cv_c_compiler_gnu = xyes ; then if test "x$asm_optimisation$sse_os" = "xyesyes" ; then XIPH_ADD_CFLAGS([-msse2]) -

Thomas Zander wrote: > In any case, the disable-SSE matter is still important. People are > still using flac on x86 machines without SSE, for instance AMD Geode > CPUs seem to live forever. libFLAC detects CPU SSE support in runtime, so --disable-sse is necessary for cuch CPUs only because it disables -msse2 switch. Maybe it makes sense to add new switch, --no-force-sse2 or

Tristan Matthews wrote: > if test "x$asm_optimisation$sse_os" = "xyesyes" ; then > XIPH_ADD_CFLAGS([-msse2]) > - XIPH_ADD_CFLAGS([-mstackrealign]) > + if test "$host_os" = "mingw32" ; then > + XIPH_ADD_CFLAGS([-mstackrealign]) > + fi > fi But sse_os==no doesn't prevent libFLAC from using SSE intrinsics. So I think

Dave Yeo wrote: > Doesn't SSE support imply SSE2+ support? Not for the CPU. Just because a CPU supports SSE, does not mean it is guaranteed to support SSE2+. For OS support, I'm not sure. Didn't later version of SSE add new registers? > I have a '96 install of an OS, it has been upgraded until end of life, > and it handles SSE4+ instructions fine even though the

On 03/14/16 03:51 PM, lvqcl wrote: > With --disable-sse, FLAC__SSE_OS is undefined and FLAC__cpu_info() uses > sigemptyset/sigaction to determine OS SSE support. That's not quite right as I have to build binaries with --disable-sse (I build and distribute both with and without) so that some users on PII's don't get a sigill, even though the OS supports SSE. So it seems to

OS SSE support is now mandatory, so the description of --disable-sse option isn't correct now (actually, it wasn't fully correct even before this). *Currently all it does is disables -msse2 flag* The relevant parts of configure.ac: AC_ARG_ENABLE(sse, AC_HELP_STRING([--disable-sse], [Disable SSE if the OS does not support SSE instructions]), [case "${enableval}" in yes)

if test "x$asm_optimisation$sse_os" = "xyesyes" ; then XIPH_ADD_CFLAGS([-msse2]) fi -msse2 tells the compiler that it can generate SSE2 instructions. The resulting code will _not_ run on CPUs without SSE2. This is counter to the assembly optimizations themselves, which check at runtime if extensions like SSE etc. are available. This

Currently the only way to compile FLAC using GCC w/o SSE support is to disable asm optimizations (see configure.ac): if test "x$asm_optimisation" = "xyes" ; then XIPH_ADD_CFLAGS([-msse2]) fi Also it's not possible to enable SSE4.1 intrinsic functions even with -msse4.1 option. The patch fixes both problems. --------------- BTW: I'm not sure that share/compat.h

lvqcl wrote: > lvqcl wrote: > > > 1) in config.h FLAC__HAS_X86INTRIN macro is always defined and empty, > > even if x86intrin.h is not available. > > > > 2) sse_os is defined as 'yes' or 'no', but AM_CONDITIONAL tests it for 'true': > > The patch is attached. Please check it. Looks good. I need to do a little testing. > > 3)

flac configury does a XIPH_ADD_CFLAGS([-Wextra]) around line 426, however it unconditionally adds -Wextra some lines before that too. The attached patch removes that unconditional addition of -Wextra which (i) removes duplicate addition, and (ii) allows older gcc versions to compile the tree peacefully. Please consider applying. -- O.S. -------------- next part -------------- diff --git

The attached small configury patch replaces -Wextra with -W in the main CFLAGS and CXXFLAGS assignments. A few lines after them, -Wextra is conditionally added already with XIPH_ADD_CFLAGS([-Wextra]), so this allows build using older compilers. Regards. -- O.S. -------------- next part -------------- diff --git a/configure.ac b/configure.ac index 6d0fa00..cc5a247 100644 --- a/configure.ac +++

This reverts commit 2860f1780ca92c779ee0a2c545ae1b9c4818dc53. Conflicts: configure.ac Do not use -fgnu89-inline as it can emit duplicate symbols for inline functions that are declared in multiple object files being linked together. For example on clang 5.0 targetting iOS (clang advertises itself as gcc 4.2, the last GPL version of gcc) --- configure.ac | 12 ++++-------- 1 file changed, 4

Olivier Tristan <o.tristan at uvi.net> ?????(?) ? ????? ?????? Fri, 21 Mar 2014 22:41:00 +0400: > Check with -mfpmath=387 to be sure that x87 FPU code is used and not some > SSE optim made by GCC I added "XIPH_ADD_CFLAGS([-mfpmath=387])" into configure.ac Still the result is different from SSE version. --------------- MSVS adds two instructions to the generated code after

1) in config.h FLAC__HAS_X86INTRIN macro is always defined and empty, even if x86intrin.h is not available. 2) sse_os is defined as 'yes' or 'no', but AM_CONDITIONAL tests it for 'true': AM_CONDITIONAL(FLaC__SSE_OS, test "x$sse_os" = xtrue) It seems that it should be changed to AM_CONDITIONAL(FLaC__SSE_OS, test "x$sse_os" = xyes) 3) configure

The attached small configury patch extends visibility attributes usage to darwin (osx) builds. Tested by cross-compiling on a linux host using gcc-.4.0.2 and gcc-4.2.1 against 10.4 and 10.6 SDKs. Regards. -- O.S. -------------- next part -------------- diff --git a/configure.ac b/configure.ac index 6d0fa00..d3c302a 100644 --- a/configure.ac +++ b/configure.ac @@ -395,7 +395,6 @@ if test

Hi all, Second and hopefully final pre-release is here: http://downloads.xiph.org/releases/flac/beta/ I have personally tested this code on: x86-linux x86_64-linux powerpc-linux armhf-linux i386-freebsd9.1 i386-openbsd5.2 I also cross-compiled from Linux to 32 bit Windows and the compile ran to completion (the test suite requires a bunch of hacking before it can

Tristan Matthews wrote: > > One workaround is to use -mstackrealign where ever -msse2 is used > > Yes, that was one of the options mentioned in the article I linked > (http://www.peterstock.co.uk/games/mingw_sse) Can someone confirm that -mstackrealign fixes this issue? If it works, that seems like the sanest solution. Erik --

On 03/09/15 03:59 PM, lvqcl wrote: > Tristan Matthews wrote: > >> >Unable to test on win32 at the moment, please give this a try. Feedback welcome. > > I wonder why misaligned stack is a problem only for the FLAC library. > There are many other libraries inside VLC... > > BTW: FLAC 1.3.1 uses -msse2 option by default. Do you remove it? > Otherwise libFLAC may crash