On Wed, Jul 07, 2004 at 04:52:23PM -0700, Pete Zaitcev wrote:
Not typically. I argued against gratious use of rwlocks for years. See http://www.advogato.org/person/Zaitcev/diary.html?start=153
I agree with Pete here.
Me too -- for "most driver usage", as in that URL. A spinlock and/or semaphore should suffice for each device. (And the scope of a spinlock often can't be just a few lines, although sometimes that works with fine grained locking.)
But shared/unshared locks do have legitimate uses, usually in code that necessarily exceeds those "simple" cases. And not just as a performance optimization either ... a lock that only needs exclusive access in specific low-volume codepaths can be a huge simplification over multiple fine grain locks. (And that might easily describe much of khubd...)
Now as for reentrant locks ... those are worth avoiding!
- Dave
Why are we trying to implement a bastard BKL implementation here?
With this implementation, it sounds like we can starve out legitimate users.
Alan, why would we try to take a lock twice for the same codepath?
thanks,
greg k-h
-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - digital self defense, top technical experts, no vendor pitches, unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
