For example, any potential contention issue where two execution threads want to read the same data structures and don't want to write. A mutex serialises these threads whereas a rw-semaphore would allow them to execute concurrently.
The only example I can think of off the top of my head is multiple simultaneous readers of the /proc filesystem, which is contrived but would be easy to arrange, even for non-privileged users. Off the top of my head, I can't see why developers would choose mutexes over rw-semaphores. If you take the write-lock then you have exclusive access to the resource, exactly as with a mutex. There's nothing that says you can't read the resource as well, and so I don't really understand your objection. I suppose you're really recoiling from the prospect of auditing the all ALSA code. However, I'm not suggesting that anyone tear the whole code-base apart. Some of the mutexes are declared static, which means that they are file-local. Someone (e.g. me?) could therefore modify them one-by-one and submit *small* patches, provided I knew that I wouldn't be wasting my time. Cheers, Chris --- Paul Davis <[EMAIL PROTECTED]> wrote: > >I have noticed that ALSA tends to use mutexes a lot > >for its internal locking. Has anyone ever > considered > >using rw-semaphores instead? The biggest impact > would > >probably be to the /proc filesystem since it would > >allow multiple concurrent readers, but I suppose > that > >it might also reduce other contention issues. > > what contention issues? almost all the locks in ALSA > are per-device > (e.g. per PCM device), so unless you have 2 or more > threads accessing > the device at the same time, there is no contention. > i haven't seen > any source code that has such a design, and it would > be quite odd > given the synchronous nature of audio interfaces. i > have seen very, > very little locking that happens at the "top level" > of ALSA. am i > missing them? > > in addition, its quite a tricky task to work through > the code and > decide whether a r-lock, w-lock or mutex is needed > at any point in > time, but i suspect that in most cases, a mutex is > called for (i.e. we > read and then write the value of a variable). > > comments? > > --p > __________________________________________________ Do You Yahoo!? NEW from Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month. http://geocities.yahoo.com/ps/info1 _______________________________________________ Alsa-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/alsa-devel
