Re: locking question: do_mmap(), do_munmap()

Manfred Spraul (manfreds@colorfullife.com)
Tue, 12 Oct 1999 00:13:37 +0200

Alexander Viro wrote:
> What does it buy you over the simple semaphore here? Do you really see a
> contention scenario?

I think you are right, I see no case where a normal semaphore would
lock-up and the rw semaphore would not lock up.

we win something _if_ vm_ops->swapout() is extremely slow:
* with lock_kernel() [ie currently], multiple threads can sleep within
vm_ops->swapout() of the same "struct mm"
* an rw-semaphore would mimic that behaviour.
* a normal semaphore would prevent that.

I'm not sure if it is worth to implement a rw-semaphore, especially
since we win something in a very obscure case, but we loose cpu-cycles
for every down_rw()/up_rw() [there is no 2 asm-instruction rw-semaphore]

--
	Manfred


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/