Re: Update shared mappings

[Andrea Arcangeli <andrea@e-mind.com>]

On Wed, 2 Dec 1998, Stephen C. Tweedie wrote:

>else's mm semaphore.  If you have two processes doing that to each other
>(ie. two processes mapping the same file r/w and doing msyncs), then you
>can most certainly still deadlock.

The thing would be trivially fixable if it would exists a down_trylock() 
that returns 0 if the semaphore was just held. I rejected now the
update_shared_mappings from my tree in the meantime though.

I have a question. Please consider only the UP case (as if linux would not
support SMP at all). Is it possible that while we are running inside
sys_msync() and another process has the mmap semaphore held?

Stephen I read some emails about a PG_dirty flag. Could you tell me some
more about that flag? 

Andrea Arcangeli

--
This is a majordomo managed list.  To unsubscribe, send a message with
the body 'unsubscribe linux-mm me@address' to: majordomo@kvack.org