Re: [PATCH v3] iomap: support tail packing inline read

From: Gao Xiang
Date: Mon Jul 19 2021 - 11:48:16 EST


On Mon, Jul 19, 2021 at 04:02:30PM +0100, Matthew Wilcox wrote:
> On Mon, Jul 19, 2021 at 10:47:47PM +0800, Gao Xiang wrote:
> > @@ -246,18 +245,19 @@ iomap_readpage_actor(struct inode *inode, loff_t pos, loff_t length, void *data,
> > unsigned poff, plen;
> > sector_t sector;
> >
> > - if (iomap->type == IOMAP_INLINE) {
> > - WARN_ON_ONCE(pos);
> > - iomap_read_inline_data(inode, page, iomap);
> > - return PAGE_SIZE;
> > - }
> > -
> > - /* zero post-eof blocks as the page may be mapped */
> > iop = iomap_page_create(inode, page);
> > + /* needs to skip some leading uptodated blocks */
> > iomap_adjust_read_range(inode, iop, &pos, length, &poff, &plen);
> > if (plen == 0)
> > goto done;
> >
> > + if (iomap->type == IOMAP_INLINE) {
> > + iomap_read_inline_data(inode, page, iomap, pos);
> > + plen = PAGE_SIZE - poff;
> > + goto done;
> > + }
>
> This is going to break Andreas' case that he just patched to work.
> GFS2 needs for there to _not_ be an iop for inline data. That's
> why I said we need to sort out when to create an iop before moving
> the IOMAP_INLINE case below the creation of the iop.

I have no idea how it breaks Andreas' case from the previous commit
message: "
iomap: Don't create iomap_page objects for inline files
In iomap_readpage_actor, don't create iop objects for inline inodes.
Otherwise, iomap_read_inline_data will set PageUptodate without setting
iop->uptodate, and iomap_page_release will eventually complain.

To prevent this kind of bug from occurring in the future, make sure the
page doesn't have private data attached in iomap_read_inline_data.
"

After this patch, iomap_read_inline_data() will set iop->uptodate with
iomap_set_range_uptodate() rather than set PageUptodate() directly,
so iomap_page_release won't complain.

Am I missing something?

Thanks,
Gao Xiang

>
> If we're not going to do that first, then I recommend leaving the
> IOMAP_INLINE case where it is and changing it to ...
>
> if (iomap->type == IOMAP_INLINE)
> return iomap_read_inline_data(inode, page, iomap, pos);
>
> ... and have iomap_read_inline_data() return the number of bytes that
> it copied + zeroed (ie PAGE_SIZE - poff).