On 21:07 23/05, Jan Kara wrote:
On Thu 23-05-19 10:27:22, Goldwyn Rodrigues wrote:
> On 16:04 23/05, Jan Kara wrote:
> > On Mon 29-04-19 12:26:47, Goldwyn Rodrigues wrote:
> > > From: Goldwyn Rodrigues <rgoldwyn(a)suse.com>
> > >
> > > Inorder to make sure mmap'd files don't change after snapshot,
> > > writeprotect the mmap pages on snapshot. This is done by performing
> > > a data writeback on the pages (which simply mark the pages are
> > > wrprotected). This way if the user process tries to access the memory
> > > we will get another fault and we can perform a CoW.
> > >
> > > In order to accomplish this, we tag all CoW pages as
> > > PAGECACHE_TAG_TOWRITE, and add the mmapd inode in delalloc_inodes.
> > > During snapshot, it starts writeback of all delalloc'd inodes and
> > > here we perform a data writeback. We don't want to keep the inodes
> > > in delalloc_inodes until it umount (WARN_ON), so we remove it
> > > during inode evictions.
> > >
> > > Signed-off-by: Goldwyn Rodrigues <rgoldwyn(a)suse.com>
> > OK, so here you use PAGECACHE_TAG_TOWRITE. But why is not
> > PAGECACHE_TAG_DIRTY enough for you? Also why isn't the same needed also
> > normal non-DAX inodes? There you also need to trigger CoW on mmap write so
> > I just don't see the difference...
> Because dax_writeback_mapping_range() writebacks pages marked
> PAGECACHE_TAG_TOWRITE and not PAGECACHE_TAG_DIRTY. Should it
> writeback pages marked as PAGECACHE_TAG_DIRTY as well?
It does writeback PAGECACHE_TAG_DIRTY pages - tag_pages_for_writeback()
moves PAGECACHE_TAG_DIRTY to PAGECACHE_TAG_TOWRITE...
Yes, I missed that. It just needs a writeback with WB_SYNC_ALL.