Messages in this thread | | | Date | Fri, 13 Nov 1998 11:31:32 -0500 (EST) | From | "Benjamin C.R. LaHaise" <> | Subject | Re: global kernel lock |
| |
On Sat, 14 Nov 1998, Richard Gooch wrote:
> I have two simple questions: > > - is my reading of the code correct
Yeap!
> - does anyone have removing this on their ToDo for 2.3?
If I understand things correctly, the holdup on making read & write SMP aware is the lack of locking in the page cache. Stephen has patches to make the page cache smp/irq safe, and I'm sure those 'll be one of the first things to be looked at when 2.3 comes along. That way, we'll be able to get rid of the wonderful 'dd if=/dev/urandom bs=1024k' SMP DOS attack.
Another [related] big wart that needs a plan of attack, is how to deal with fget & fput in threads, and that might be a 2.2 item -- thread A does a big write(), then thread B then does a close() on the fd... Uh-oh. This is where a rwlock version of a semaphore would be *really* handy.
-ben
- 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/
| |