On Sun, Jul 31, 2011 at 4:09 PM, Dave Chinner [off-list ref] wrote:

Hence if we can't grab the superblock here, it is simply another
case of a "new work pending" interrupt, right? And so aborting the
work is the correct thing to do? Especially as it avoids all the
ordering problems of redirtying inodes and allows the writeback work
to restart (form whatever context it is stared from next time) where
it stopped.

Ok, that does sound like a reasonable approach to me. Which probably
means that there is some oversimplification and gotcha hidden in that
argument, but I don't see it immediately.

I do agree that it would be nice if we could avoid re-ordering
problems. The whole dirty time issue and whether inodes are actually
correctly ordered on the dirty lists has always been one huge rats
nest.  Ask Andrew about it, he was espousing his love for that code at
the recent Intel Tech Days ;)

         Linus