Re: writepage fs corruption fixes

[Andrew Morton <akpm@osdl.org>]

Andrea Arcangeli <andrea@suse.de> wrote:
>
> 1) The major one (the only one I believe that was triggering [only on
>  ext2 due the fact mpage is working only there]) is the marking of the bh
>  clean despite we could still run into the "confused" path. After that
>  the confused path really becomes confused and it writes nothing and fs
>  corruption triggers silenty (the reugular writepage only writes bh that
>  are marked dirty, it never attempts to submit_bh anything marked clean).
>  The mpage-writepage code must never mark the bh clean as far as it
>  wants to still fallback in the regular writepage which depends on the bh
>  to be dirty (i.e. the "goto confused" path). This could only triggers
>  with memory pressure (it also needs buffer_heads_over_limit == 0, and
>  that is frequent under mm pressure).

ooh, nasty, yes.  You must be testing the crap out if it.

>  3) Third bug is in the regular writepage, the nr_underway == 0 code was
>  walking buffers on an unlocked page without keeping the bh pinned, and
>  in turn the bh could be released under it by the VM. Fix is to delay the
>  put_bh loop. (this might have triggered but it's not certain)

PG_writeback protects the page from truncate, from invalidate and from page
reclaim.  pagevec_strip() won't touch the buffers due to the
PageWriteback() test in try_to_release_page().  So I think we're OK in
there.  I can add a couple more coment fixes for this.