On Sat, Dec 07, 2013 at 21:32, Vadim Zhukov wrote: > 2013/12/7 Ted Unangst <[email protected]>: >> On Sat, Dec 07, 2013 at 14:11, Vadim Zhukov wrote: >>> This patch fixes problems in KDE4, that relies on sharing (process-shared) >>> semaphores via mmap(). This feature is used in the wild, so if we claim >>> that we support process-shared semaphores, we have to implement it, too. >>> >>> This changes the sem_t definition: it becomes a synonim to struct __sem >>> instead of being a pointer to it. This way we allow apps to have sem_t in >>> shared memory via sem_init() and thus share it between processes (by, >>> e.g., mmap()'ing the same file). Of course, this forces a major library >>> version bump. >> >> This requires more thought. If the introduction of pshared support is >> breaking things, maybe just that part can be reverted for now. > > Well, it's the half-done introduction of pshared support is breaking > things. I do not see any fallout now that sharing unnamed semaphores > works. But what we'll gain with revert? Except that it's still better > than current situation.
One of the hallmarks of the original libpthread was that all data structures were opaque, and hidden via pointers. That in turn made it possible to write a binary compatible librthread. I never would have started librthread if it hadn't been for that compatibility. So I don't like turning my back on it. Exposing the size of sem_t is a *major* step.
