Messages in this thread | | | From | "Stephen C. Tweedie" <> | Date | Wed, 26 Jan 2000 18:19:15 +0000 (GMT) | Subject | Re: [RFC] change in /proc/devices |
| |
Hi,
On Tue, 25 Jan 2000 23:59:56 +0000 (GMT), Alan Cox <alan@lxorguk.ukuu.org.uk> said:
>> immaturity >> in comparision with ... for example System V - which solved those >> problems already decades >> ago.
> SYS5 didnt have modules decades ago. The problem isnt actually that big nor > an interface issue, its a locking thing in part tickled by our finer grained > scheduling. It just means figuring out how to put the new locks in as few > places as possible
This is one place where r/w semaphores might make it a _lot_ easier to get the locking right in a fairly general manner. Ben's semaphores are _really_ lean to take, and using scheduler-safe locks will make it much easier to delegate the locking to the mid-layers instead of forcing every driver to do it internally.
--Stephen
- 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/
| |