| Message ID | 1464233085-19165-1-git-send-email-adas@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
----- Original Message ----- > From: "Abhi Das" <adas@redhat.com> > To: cluster-devel@redhat.com, linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org > Cc: "Abhi Das" <adas@redhat.com>, "Miklos Szeredi" <mszeredi@redhat.com>, "Jens Axboe" <axboe@fb.com>, "Al Viro" > <viro@zeniv.linux.org.uk> > Sent: Wednesday, May 25, 2016 10:24:45 PM > Subject: [PATCH] fs: Do not check for valid page->mapping in page_cache_pipe_buf_confirm > > If the page is truncated after being spliced into the pipe, it's > probably not invalid. > > For filesystems that invalidate pages, we used to return -ENODATA > even though the data is there, it's just possibly different from > what was spliced into the pipe. We shouldn't have to throw away > the buffer or return error in this case. > > Signed-off-by: Abhi Das <adas@redhat.com> > CC: Miklos Szeredi <mszeredi@redhat.com> > CC: Jens Axboe <axboe@fb.com> > CC: Al Viro <viro@zeniv.linux.org.uk> > --- > fs/splice.c | 9 --------- > 1 file changed, 9 deletions(-) > > diff --git a/fs/splice.c b/fs/splice.c > index dd9bf7e..b9899b99 100644 > --- a/fs/splice.c > +++ b/fs/splice.c > @@ -106,15 +106,6 @@ static int page_cache_pipe_buf_confirm(struct > pipe_inode_info *pipe, > lock_page(page); > > /* > - * Page got truncated/unhashed. This will cause a 0-byte > - * splice, if this is the first page. > - */ > - if (!page->mapping) { > - err = -ENODATA; > - goto error; > - } > - > - /* > * Uh oh, read-error from disk. > */ > if (!PageUptodate(page)) { > -- > 2.4.3 > > Hi Jens/Al, Any comments on this patch? FWIW, I've done some testing on GFS2 w/ this patch and it seems to be holding up ok. The test is designed to verify writes using splice reads and even if the pages were truncated, the data read in from the spliced pipe looks consistent with what was expected. Cheers! --Abhi -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/splice.c b/fs/splice.c index dd9bf7e..b9899b99 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -106,15 +106,6 @@ static int page_cache_pipe_buf_confirm(struct pipe_inode_info *pipe, lock_page(page); /* - * Page got truncated/unhashed. This will cause a 0-byte - * splice, if this is the first page. - */ - if (!page->mapping) { - err = -ENODATA; - goto error; - } - - /* * Uh oh, read-error from disk. */ if (!PageUptodate(page)) {
If the page is truncated after being spliced into the pipe, it's probably not invalid. For filesystems that invalidate pages, we used to return -ENODATA even though the data is there, it's just possibly different from what was spliced into the pipe. We shouldn't have to throw away the buffer or return error in this case. Signed-off-by: Abhi Das <adas@redhat.com> CC: Miklos Szeredi <mszeredi@redhat.com> CC: Jens Axboe <axboe@fb.com> CC: Al Viro <viro@zeniv.linux.org.uk> --- fs/splice.c | 9 --------- 1 file changed, 9 deletions(-)