From: Jeff Layton <jlayton@redhat.com>
To: "Steve French" <smfrench@gmail.com>
Cc: "Dave Kleikamp" <shaggy@linux.vnet.ibm.com>,
"Nick Piggin" <npiggin@suse.de>,
pbadari@us.ibm.com, linux-fsdevel <linux-fsdevel@vger.kernel.org>,
"linux-cifs-client@lists.samba.org"
<linux-cifs-client@lists.samba.org>
Subject: Re: fsx-linux failing with latest cifs-2.6 git tree
Date: Fri, 21 Nov 2008 17:50:17 -0500 [thread overview]
Message-ID: <20081121175017.4e656475@tleilax.poochiereds.net> (raw)
In-Reply-To: <524f69650811211238x3c81d899r225f46776aeb52f1@mail.gmail.com>
On Fri, 21 Nov 2008 14:38:18 -0600
"Steve French" <smfrench@gmail.com> wrote:
> Fix attached.
>
> Shaggy/Jeff/Nick etc. do you want to review/ack it since it is late in the rc?
>
Talking with Steve on IRC, we thought it might be better to optimize
away the read when possible. I think this patch should do it. We skip
the read if the write starts past the current end of the file, or if
the offset into the page of the beginning of the write is 0 and we're
writing past the current end of the file. In those situations we just
zero out the rest of the page.
Combined patch inlined below. I also took the liberty of adding a page
pointer to make the code look a little cleaner.
Thoughts?
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <smfrench@us.ibm.com>
---
fs/cifs/file.c | 42 ++++++++++++++++++++++++++++++------------
1 files changed, 30 insertions(+), 12 deletions(-)
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index b691b89..1d252b8 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -2062,26 +2062,43 @@ static int cifs_write_begin(struct file *file, struct address_space *mapping,
{
pgoff_t index = pos >> PAGE_CACHE_SHIFT;
loff_t offset = pos & (PAGE_CACHE_SIZE - 1);
+ loff_t page_start = pos & PAGE_MASK;
+ loff_t i_size;
+ struct page *page;
+ char *buf;
+ int rc = 0;
cFYI(1, ("write_begin from %lld len %d", (long long)pos, len));
- *pagep = __grab_cache_page(mapping, index);
- if (!*pagep)
- return -ENOMEM;
+ page = __grab_cache_page(mapping, index);
+ if (!page) {
+ rc = -ENOMEM;
+ goto out;
+ }
- if (PageUptodate(*pagep))
- return 0;
+ if (PageUptodate(page))
+ goto out;
/* If we are writing a full page it will be up to date,
no need to read from the server */
if (len == PAGE_CACHE_SIZE && flags & AOP_FLAG_UNINTERRUPTIBLE)
- return 0;
-
- if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
- int rc;
+ goto out;
+ i_size = i_size_read(page->mapping->host);
+ if (pos >= i_size || ((offset == 0) && ((pos + len) >= i_size))) {
+ /*
+ * optimize away the read when the unwritten part of the page
+ * isn't expected to have any existing file data. Just zero
+ * out the unused parts.
+ */
+ buf = kmap(page);
+ memset(buf, 0, offset);
+ memset(buf + offset + len, 0, PAGE_CACHE_SIZE - (offset + len));
+ kunmap(page);
+ SetPageUptodate(page);
+ } else if ((file->f_flags & O_ACCMODE) != O_WRONLY) {
/* might as well read a page, it is fast enough */
- rc = cifs_readpage_worker(file, *pagep, &offset);
+ cifs_readpage_worker(file, page, &page_start);
/* we do not need to pass errors back
e.g. if we do not have read access to the file
@@ -2093,8 +2110,9 @@ static int cifs_write_begin(struct file *file, struct address_space *mapping,
racing with this read? In any case
this will be written out by write_end so is fine */
}
-
- return 0;
+out:
+ *pagep = page;
+ return rc;
}
const struct address_space_operations cifs_addr_ops = {
--
1.5.5.1
next prev parent reply other threads:[~2008-11-21 22:50 UTC|newest]
Thread overview: 38+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20081121105613.09a8cb8e@tleilax.poochiereds.net>
[not found] ` <524f69650811210820s549de2bah3181cbc0c5633091@mail.gmail.com>
[not found] ` <20081121112249.0b408b55@tleilax.poochiereds.net>
[not found] ` <524f69650811210846q7502fd99m6f4d335bb6ac1b65@mail.gmail.com>
[not found] ` <524f69650811211109w659e5decoa34a8e0f907772a3@mail.gmail.com>
[not found] ` <524f69650811211113q4fffcc70of88cb85db531c358@mail.gmail.com>
[not found] ` <1227296476.20845.8.camel@norville.austin.ibm.com>
[not found] ` <524f69650811211218v78295682lcf6dce842327b097@mail.gmail.com>
2008-11-21 20:38 ` Fwd: fsx-linux failing with latest cifs-2.6 git tree Steve French
2008-11-21 20:41 ` Dave Kleikamp
2008-11-21 21:02 ` Steve French
2008-11-21 23:44 ` Nick Piggin
2008-11-21 20:50 ` Jeff Layton
2008-11-21 22:50 ` Jeff Layton [this message]
2008-11-21 23:02 ` Dave Kleikamp
2008-11-21 23:25 ` Jeff Layton
2008-11-22 1:04 ` Steve French
2008-11-22 1:50 ` Jeff Layton
2008-11-21 23:53 ` Nick Piggin
2008-11-22 1:51 ` Jeff Layton
2008-11-22 2:02 ` Steve French
2008-11-22 4:47 ` Dave Kleikamp
2008-11-22 15:39 ` [linux-cifs-client] " Jeff Layton
2008-11-22 20:27 ` Dave Kleikamp
2008-11-23 11:57 ` Jeff Layton
2008-11-24 2:32 ` Steve French
2008-11-24 11:19 ` [linux-cifs-client] " Jeff Layton
2008-11-26 4:04 ` Steve French
2008-11-26 11:54 ` Jeff Layton
2008-11-26 12:11 ` Jeff Layton
2008-11-26 13:09 ` [linux-cifs-client] " Nick Piggin
2008-11-26 15:08 ` Jeff Layton
2008-11-26 15:23 ` Nick Piggin
2008-11-26 16:37 ` Jeff Layton
2008-11-27 8:33 ` Nick Piggin
2008-11-28 12:18 ` Jeff Layton
2008-11-30 21:44 ` Steve French
2008-11-30 22:17 ` Jeff Layton
2008-12-01 8:44 ` Nick Piggin
2008-12-01 11:28 ` Jeff Layton
2008-12-01 11:32 ` Nick Piggin
2008-12-01 11:55 ` Jeff Layton
2008-12-01 17:43 ` Steve French
2008-11-26 19:46 ` Steve French
2008-11-24 20:00 ` Dave Kleikamp
2008-11-26 13:02 ` Nick Piggin
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=20081121175017.4e656475@tleilax.poochiereds.net \
--to=jlayton@redhat.com \
--cc=linux-cifs-client@lists.samba.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=npiggin@suse.de \
--cc=pbadari@us.ibm.com \
--cc=shaggy@linux.vnet.ibm.com \
--cc=smfrench@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).