From: Matthew Wilcox <willy@infradead.org>
To: Christoph Hellwig <hch@infradead.org>
Cc: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-mm@kvack.org, Jeff Layton <jlayton@kernel.org>,
Chris Mason <clm@fb.com>
Subject: Re: [PATCH v2 6/9] iomap,xfs: Convert from readpages to readahead
Date: Tue, 14 Jan 2020 23:42:43 -0800 [thread overview]
Message-ID: <20200115074243.GA31744@bombadil.infradead.org> (raw)
In-Reply-To: <20200115071628.GA3460@infradead.org>
On Tue, Jan 14, 2020 at 11:16:28PM -0800, Christoph Hellwig wrote:
> On Tue, Jan 14, 2020 at 06:38:40PM -0800, Matthew Wilcox wrote:
> > static loff_t
> > +iomap_readahead_actor(struct inode *inode, loff_t pos, loff_t length,
> > void *data, struct iomap *iomap, struct iomap *srcmap)
> > {
> > struct iomap_readpage_ctx *ctx = data;
> > @@ -410,10 +381,8 @@ iomap_readpages_actor(struct inode *inode, loff_t pos, loff_t length,
> > ctx->cur_page = NULL;
> > }
> > if (!ctx->cur_page) {
> > - ctx->cur_page = iomap_next_page(inode, ctx->pages,
> > - pos, length, &done);
> > - if (!ctx->cur_page)
> > - break;
> > + ctx->cur_page = readahead_page(inode->i_mapping,
> > + pos / PAGE_SIZE);
>
> Don't we at least need a sanity check for a NULL cur_page here?
I don't think so. The caller has already put the locked page into the
page cache at that index. If the page has gone away, that's a bug, and
I don't think BUG_ON is all that much better than a NULL pointer derefence.
Indeed, readahead_page() checks PageLocked, so it can't return NULL.
> Also the readahead_page version in your previous patch seems to expect
> a byte offset, so the division above would not be required.
Oops. I had intended to make readahead_pages() look like this:
struct page *readahead_page(struct address_space *mapping, pgoff_t index)
{
struct page *page = xa_load(&mapping->i_pages, index);
VM_BUG_ON_PAGE(!PageLocked(page), page);
return page;
}
If only our tools could warn about these kinds of mistakes.
> (and should
> probably be replaced with a right shift anyway no matter where it ends
> up)
If the compiler can't tell that x / 4096 and x >> 12 are precisely the same
and choose the more efficient of the two, we have big problems.
> > +unsigned
> > +iomap_readahead(struct address_space *mapping, pgoff_t start,
> > unsigned nr_pages, const struct iomap_ops *ops)
> > {
> > struct iomap_readpage_ctx ctx = {
> > - .pages = pages,
> > .is_readahead = true,
> > };
> > - loff_t pos = page_offset(list_entry(pages->prev, struct page, lru));
> > - loff_t last = page_offset(list_entry(pages->next, struct page, lru));
> > - loff_t length = last - pos + PAGE_SIZE, ret = 0;
> > + loff_t pos = start * PAGE_SIZE;
> > + loff_t length = nr_pages * PAGE_SIZE;
>
> Any good reason not to pass byte offsets for start and length?
>
> > + return length / PAGE_SIZE;
>
> Same for the return value?
>
> For the file systems that would usually be a more natural interface than
> a page index and number of pages.
That seems to depend on the filesystem. iomap definitely would be happier
with loff_t, but cifs prefers pgoff_t. I should probably survey a few
more filesystems and see if there's a strong lean in one direction or
the other.
next prev parent reply other threads:[~2020-01-15 7:42 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-15 2:38 [RFC v2 0/9] Replacing the readpages a_op Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 1/9] mm: Fix the return type of __do_page_cache_readahead Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 2/9] readahead: Ignore return value of ->readpages Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 3/9] XArray: Add xarray_for_each_range Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 4/9] readahead: Put pages in cache earlier Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 5/9] mm: Add readahead address space operation Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 6/9] iomap,xfs: Convert from readpages to readahead Matthew Wilcox
2020-01-15 7:16 ` Christoph Hellwig
2020-01-15 7:42 ` Matthew Wilcox [this message]
2020-01-24 22:53 ` Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 7/9] cifs: " Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 8/9] mm: Remove add_to_page_cache_locked Matthew Wilcox
2020-01-15 2:38 ` [PATCH v2 9/9] mm: Unify all add_to_page_cache variants Matthew Wilcox
2020-01-15 7:20 ` Christoph Hellwig
2020-01-15 7:44 ` Matthew Wilcox
2020-01-18 23:13 ` [RFC v2 0/9] Replacing the readpages a_op Matthew Wilcox
2020-01-21 11:36 ` Jan Kara
2020-01-21 21:48 ` Matthew Wilcox
2020-01-22 9:44 ` Jan Kara
2020-01-23 10:31 ` Jan Kara
2020-01-22 23:47 ` Dave Chinner
2020-01-23 10:21 ` Jan Kara
2020-01-23 22:29 ` Dave Chinner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200115074243.GA31744@bombadil.infradead.org \
--to=willy@infradead.org \
--cc=clm@fb.com \
--cc=hch@infradead.org \
--cc=jlayton@kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-xfs@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).