> > They call ->setattr() (devfs_notify_change(), in your case) and that
> > has nothing to icache (you already have the inode). Or had I
> > completely misparsed you?
>
> You parsed correctly. I know ->setattr() is called. I just wanted to
> make sure that the icache didn't have some subtle interaction I was
> missing. Such as ->write_inode() not being called.
Oh... That actually begs for another question - why the heck do you
need ->write_inode()?
Sorry - I had missed the presense of that animal. Hmm... OK.
Variant 1:
insert_inode_hash(inode). Silly, but will work
Variant 2:
kill devfs_write_inode() - do its equivalent in notify_change()
_and_ set the de->atime in ->readlink() and ->follow_link(). The latter
is due to update_atime() logics. And no, I don't like it.
Actually, it stinks - we have UPDATE_ATIME implemented by direct assignment to
->i_atime instead of the proper way (->setattr()). Linus, would you mind if
I change that? That would be the Right Thing(tm), but it might be too heavy.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/