On Fri, 2012-04-13 at 00:26 -0700, Andrew Morton wrote: > On Fri, 13 Apr 2012 09:38:28 +0300 Artem Bityutskiy <dedekind1@xxxxxxxxx> wrote: > > > > That implies that we retain ->write_super, probably in a modified form. > > > Modified to permit the VFS to determine whether the superblock needs > > > treatment, if ->s_dirt doesn't suffice. > > > > I tried this approach and it was vetoed by Al Viro. Although it is > > simpler to me to resurrect my old patches, I agree with Al that killing > > '->write_super()' is a better approach. > > Well, it can be done without a super_operation vector - pass the > library code a superblock* and a function address. But the difference > is pointless fluff. May be, let see how many FSes will actually can share things. Per-FS implementation is better because you do not have to worry about refcounting and the FS gone by the time a timer expires. Also, when you know the FS specifics, you can make a decision about whether the timer can be made deferrable. Sorry, I did not understand what you meant by "the difference is pointless fluff" - difference between what and what? > > Also, if you look at this from the angle that only few old FSes will > > have this, it becomes not that bad. I assume I will change this > > patch-set and won't use delayed works here. > > I don't think I understand that. You intend to alter this patchset? Yeah, I think I'll be able to implement one of the two ideas I described in the previous e-mail, test, and send version two of this patch-set. Thanks! -- Best Regards, Artem Bityutskiy
Attachment:
signature.asc
Description: This is a digitally signed message part
