On Thu, Mar 17, 2022 at 09:16:20PM +0000, Matthew Wilcox wrote:
> On Thu, Mar 17, 2022 at 12:26:35PM -0700, Linus Torvalds wrote:
> > On Thu, Mar 17, 2022 at 8:04 AM Matthew Wilcox <willy@xxxxxxxxxxxxx> wrote:
> > >
> > > So how about we do something like this:
> > >
> > > - Make folio_start_writeback() and set_page_writeback() return void,
> > > fixing up AFS and NFS.
> > > - Add a folio_wait_start_writeback() to use in the VFS
> > > - Remove the calls to set_page_writeback() in the filesystems
> >
> > That sounds lovely, but it does worry me a bit. Not just the odd
> > 'keepwrite' thing, but also the whole ordering between the folio bit
> > and the tagging bits. Does the ordering possibly matter?
>
> I wouldn't change the ordering of setting the xarray bits and the
> writeback flag; they'd just be set a little earlier. It'd all be done
> while the page was still locked. But you're right, there's lots of
> subtle interactions here.
>
> > That whole "xyz_writeback_keepwrite()" thing seems odd. It's used in
> > only one place (the folio version isn't used at all):
> >
> > ext4_writepage():
> >
> > ext4_walk_page_buffers() fails:
> > redirty_page_for_writepage(wbc, page);
> > keep_towrite = true;
> > ext4_bio_write_page().
> >
> > which just looks odd. Why does it even try to continue to do the
> > writepage when the page buffer thing has failed?
> >
> > In the regular write path (ie ext4_write_begin()), a
> > ext4_walk_page_buffers() failure is fatal or causes a retry). Why is
> > ext4_writepage() any different? Particularly since it wants to keep
> > the page dirty, then trying to do the writeback just seems wrong.
> >
> > So this code is all a bit odd, I suspect there are decades of "people
> > continued to do what they historically did" changes, and it is all
> > worrisome.
>
> I found the commit: 1c8349a17137 ("ext4: fix data integrity sync in
> ordered mode"). Fortunately, we have a documented test for this,
> generic/127, so we'll know if we've broken it.
Looks like a footgun. ext4 needs the keepwrite stuff for block size <
page size, in the case where a page has both written and
delalloc/unwritten buffers on it. In that case ext4_writepage tries
to write just the written blocks and leave the dealloc/unwritten
buffers alone because it can't do allocation in ->writepage context.
I say footgun, because the nested ->writepage call that needs
keepwrite comes a from journal stop context in the high level
->writepages context that is doing allocation that will allow the
entire page to be written. i.e. it seems a bit silly to be
triggering partial page writeback that skips delalloc/unwritten
extents, but then needs special awareness of higher level IO that is
in progress that is currently doing the allocation that will allow
all the delalloc/unwritten extents on the page to also be written
back...
Cheers,
Dave.
--
Dave Chinner
david@xxxxxxxxxxxxx
