From: Jyri Sarha <jyri.sarha@xxxxxxxxx> Damn, I always need to read my own patches from the mailing list before I realize that there is something wrong with them. I think the earlier version of this same patch had a weakness that in theory the contents of a stack element may not have been propagated to memory for a popping thread to read if the thread that pushed the element is scheduled out before the push function has finished (or the two threads are running on parallel CPUs). Even after heavy testing with the original patch on both ARM and x86 I have not seen this happening, but this new version should make it sure it does not happen. Cheers, Jyri Jyri Sarha (1): core: New LIFO style flist implementation v2.0 src/pulsecore/flist.c | 215 ++++++++++++++----------------------------------- 1 files changed, 61 insertions(+), 154 deletions(-)
