From: Jeff Layton <jlayton@kernel.org>
To: ceph-devel@vger.kernel.org
Cc: ukernel@gmail.com, idryomov@gmail.com
Subject: Re: [PATCH 5/5] ceph: fold ceph_update_writeable_page into ceph_write_begin
Date: Wed, 16 Sep 2020 15:16:19 -0400 [thread overview]
Message-ID: <7817f98d3b2daafe113bf8290cc8c7adbb86fe99.camel@kernel.org> (raw)
In-Reply-To: <20200916173854.330265-6-jlayton@kernel.org>
On Wed, 2020-09-16 at 13:38 -0400, Jeff Layton wrote:
> ...and reorganize the loop for better clarity.
>
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
> fs/ceph/addr.c | 148 ++++++++++++++++++++++---------------------------
> 1 file changed, 65 insertions(+), 83 deletions(-)
>
> diff --git a/fs/ceph/addr.c b/fs/ceph/addr.c
> index a6c54cfc3fea..89c4d8a9a5eb 100644
> --- a/fs/ceph/addr.c
> +++ b/fs/ceph/addr.c
> @@ -1251,6 +1251,8 @@ static int context_is_writeable_or_written(struct inode *inode,
> * @inode: inode associated with page
> * @page: page being dirtied
> *
> + * We are only allowed to write into/dirty the page if the page is
> + * clean, or already dirty within the same snap context.
> * Returns NULL on success, negative error code on error, and a snapc ref that should be
> * waited on otherwise.
> */
> @@ -1307,104 +1309,84 @@ ceph_find_incompatible(struct inode *inode, struct page *page)
> /*
> * We are only allowed to write into/dirty the page if the page is
> * clean, or already dirty within the same snap context.
> - *
> - * called with page locked.
> - * return success with page locked,
> - * or any failure (incl -EAGAIN) with page unlocked.
> */
> -static int ceph_update_writeable_page(struct file *file,
> - loff_t pos, unsigned len,
> - struct page *page)
> +static int ceph_write_begin(struct file *file, struct address_space *mapping,
> + loff_t pos, unsigned len, unsigned flags,
> + struct page **pagep, void **fsdata)
> {
> struct inode *inode = file_inode(file);
> struct ceph_inode_info *ci = ceph_inode(inode);
> struct ceph_snap_context *snapc;
> - loff_t page_off = pos & PAGE_MASK;
> + struct page *page = NULL;
> + pgoff_t index = pos >> PAGE_SHIFT;
> int pos_in_page = pos & ~PAGE_MASK;
> - int end_in_page = pos_in_page + len;
> - loff_t i_size;
> - int r;
> + int r = 0;
>
> -retry_locked:
> - snapc = ceph_find_incompatible(inode, page);
> - if (snapc) {
> - if (IS_ERR(snapc)) {
> - r = PTR_ERR(snapc);
> - goto fail_unlock;
> + dout("write_begin file %p inode %p page %p %d~%d\n", file, inode, page, (int)pos, (int)len);
> +
> + for (;;) {
> + page = grab_cache_page_write_begin(mapping, index, 0);
> + if (!page) {
> + r = -ENOMEM;
> + break;
> }
> - unlock_page(page);
> - ceph_queue_writeback(inode);
> - r = wait_event_killable(ci->i_cap_wq,
> - context_is_writeable_or_written(inode, snapc));
> - ceph_put_snap_context(snapc);
> - return -EAGAIN;
> - }
>
> - if (PageUptodate(page)) {
> - dout(" page %p already uptodate\n", page);
> - return 0;
> - }
> + snapc = ceph_find_incompatible(inode, page);
> + if (snapc) {
> + if (IS_ERR(snapc)) {
> + r = PTR_ERR(snapc);
> + break;
> + }
> + unlock_page(page);
> + ceph_queue_writeback(inode);
> + r = wait_event_killable(ci->i_cap_wq,
> + context_is_writeable_or_written(inode, snapc));
> + ceph_put_snap_context(snapc);
> + put_page(page);
> + page = NULL;
Now that I look a bit more, there is probably no need to hold the page
reference across this wait. I'll change that to put it after unlocking
and re-test.
> + if (r != 0)
> + break;
> + continue;
> + }
>
> - /* full page? */
> - if (pos_in_page == 0 && len == PAGE_SIZE)
> - return 0;
> + if (PageUptodate(page)) {
> + dout(" page %p already uptodate\n", page);
> + break;
> + }
>
> - /* past end of file? */
> - i_size = i_size_read(inode);
> -
> - if (page_off >= i_size ||
> - (pos_in_page == 0 && (pos+len) >= i_size &&
> - end_in_page - pos_in_page != PAGE_SIZE)) {
> - dout(" zeroing %p 0 - %d and %d - %d\n",
> - page, pos_in_page, end_in_page, (int)PAGE_SIZE);
> - zero_user_segments(page,
> - 0, pos_in_page,
> - end_in_page, PAGE_SIZE);
> - return 0;
> - }
> + /*
> + * In some cases we don't need to read at all:
> + * - full page write
> + * - write that lies completely beyond EOF
> + * - write that covers the the page from start to EOF or beyond it
> + */
> + if ((pos_in_page == 0 && len == PAGE_SIZE) ||
> + (pos >= i_size_read(inode)) ||
> + (pos_in_page == 0 && (pos + len) >= i_size_read(inode))) {
> + zero_user_segments(page, 0, pos_in_page,
> + pos_in_page + len, PAGE_SIZE);
> + break;
> + }
>
> - /* we need to read it. */
> - r = ceph_do_readpage(file, page);
> - if (r < 0) {
> - if (r == -EINPROGRESS)
> - return -EAGAIN;
> - goto fail_unlock;
> + /*
> + * We need to read it. If we get back -EINPROGRESS, then the page was
> + * handed off to fscache and it will be unlocked when the read completes.
> + * Refind the page in that case so we can reacquire the page lock. Otherwise
> + * we got a hard error or the read was completed synchronously.
> + */
> + r = ceph_do_readpage(file, page);
> + if (r != -EINPROGRESS)
> + break;
> }
> - goto retry_locked;
> -fail_unlock:
> - unlock_page(page);
> - return r;
> -}
>
> -/*
> - * We are only allowed to write into/dirty the page if the page is
> - * clean, or already dirty within the same snap context.
> - */
> -static int ceph_write_begin(struct file *file, struct address_space *mapping,
> - loff_t pos, unsigned len, unsigned flags,
> - struct page **pagep, void **fsdata)
> -{
> - struct inode *inode = file_inode(file);
> - struct page *page;
> - pgoff_t index = pos >> PAGE_SHIFT;
> - int r;
> -
> - do {
> - /* get a page */
> - page = grab_cache_page_write_begin(mapping, index, 0);
> - if (!page)
> - return -ENOMEM;
> -
> - dout("write_begin file %p inode %p page %p %d~%d\n", file,
> - inode, page, (int)pos, (int)len);
> -
> - r = ceph_update_writeable_page(file, pos, len, page);
> - if (r < 0)
> + if (r < 0) {
> + if (page) {
> + unlock_page(page);
> put_page(page);
> - else
> - *pagep = page;
> - } while (r == -EAGAIN);
> -
> + }
> + } else {
> + *pagep = page;
> + }
> return r;
> }
>
--
Jeff Layton <jlayton@kernel.org>
next prev parent reply other threads:[~2020-09-16 19:17 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-16 17:38 [PATCH 0/5] ceph: addr.c cleanups Jeff Layton
2020-09-16 17:38 ` [PATCH 1/5] ceph: break out writeback of incompatible snap context to separate function Jeff Layton
2020-09-16 17:38 ` [PATCH 2/5] ceph: don't call ceph_update_writeable_page from page_mkwrite Jeff Layton
2020-09-16 17:38 ` [PATCH 3/5] ceph: fold ceph_sync_readpages into ceph_readpage Jeff Layton
2020-09-16 17:38 ` [PATCH 4/5] ceph: fold ceph_sync_writepages into writepage_nounlock Jeff Layton
2020-09-16 17:38 ` [PATCH 5/5] ceph: fold ceph_update_writeable_page into ceph_write_begin Jeff Layton
2020-09-16 19:16 ` Jeff Layton [this message]
2021-06-11 14:14 ` Andrew W Elble
2021-06-11 14:52 ` Jeff Layton
2021-06-11 15:11 ` David Howells
2021-06-11 15:20 ` Matthew Wilcox
2021-06-11 15:38 ` Jeff Layton
2021-06-11 15:35 ` David Howells
2021-06-11 15:59 ` Andrew W Elble
2021-06-11 17:56 ` Matthew Wilcox
2021-06-11 21:47 ` David Howells
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=7817f98d3b2daafe113bf8290cc8c7adbb86fe99.camel@kernel.org \
--to=jlayton@kernel.org \
--cc=ceph-devel@vger.kernel.org \
--cc=idryomov@gmail.com \
--cc=ukernel@gmail.com \
/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).