On Wed, 31 May 2000, Rik van Riel wrote:
> On Wed, 31 May 2000, Alexander Viro wrote:
> > On Tue, 30 May 2000, Kenneth C. Arnold wrote:
> >
> > > Rik van Riel:
> > > Threaded dcache
> >
> > Oh? Rik, if you've also done that - post the URL, I'ld really
> > like to compare our variants.
>
> I haven't. I've only seen your patch and I think we'll want
> something like that ;)
Hmm... ObPatchInQuestion: I'm sorely tempted to shift the big lock in
sys_msync(). It's definitely shiftable to msync_interval(), but there's
more to that - the only non-NULL ->sync in the tree (filemap_sync()) seems
to be almost SMP-safe, possible exception being swap_free().
Questions:
a) am I right wrt the rest of filemap_sync()?
b) does swap_free() actually need the big lock? Even if it does,
shifting the lock down there will be a big win (we'll need to protect
->fsync() too, indeed, but that still clears the path).
Cheers,
Al
-
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/
This archive was generated by hypermail 2b29 : Wed Jun 07 2000 - 21:00:12 EST