On Sat, 23.08.08 19:14, Thomas Wagner ([EMAIL PROTECTED]) wrote: Hmm, could you please post requests like this on the PA ML?
> As maintainer of the build spec form the SFE repository I have > kind of a request for help with atomic ops. > > SFEpulseaudio.spec uses version 0.9.5 which is pre-libatomic-ops. > It runs on Solaris but the atomic-ops implementation of later pulsaudio > releases prevents upgrading SFEpulseaudio (stalled at 0.9.5). > > If anyone want's to look into extending the atomic-ops code in > pulseaudio I would like to share my first coding steps. Uh -- 0.9.5 is very old. Because libatomic_ops is unfortunately out-of-date and unmaintained I'd prefer moving more atomic op implementation into PA's sources itself. We now have native implementations for x86-64 and ARM, I'd be happy to merge further implementations. See src/pulsecore/atomic.h. The preferred way of doing atomic operations however is via gcc's __sync intrinsics. gcc supports them now on most archs: http://gcc.gnu.org/onlinedocs/gcc-4.1.0/gcc/Atomic-Builtins.html Maybe the sun compiler has similar intrinsics? As far as I know Sun libc also provides membar_xxx() and atomic_xxx()? Please note that the SunAudio driver might need some updating before PA on Solaris is fun again. Also note that stuff like 'glitch-free' is only available on ALSA for now. As long as the underlying audio API supports rewinding of playback it should be possible to port g-f onto other APIs as well. Lennart -- Lennart Poettering Red Hat, Inc. lennart [at] poettering [dot] net ICQ# 11060553 http://0pointer.net/lennart/ GnuPG 0x1A015CC4 _______________________________________________ libcanberra-discuss mailing list [email protected] https://tango.0pointer.de/mailman/listinfo/libcanberra-discuss
