From: Al Viro <viro@ZenIV.linux.org.uk>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: George Spelvin <linux@horizon.com>,
linux-fsdevel <linux-fsdevel@vger.kernel.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: Update of file offset on write() etc. is non-atomic with I/O
Date: Tue, 4 Mar 2014 00:23:43 +0000 [thread overview]
Message-ID: <20140304002342.GS18016@ZenIV.linux.org.uk> (raw)
In-Reply-To: <CA+55aFyWorcX0qYAcD2bj=bOqa-X6Gmb_Ph4g25nqwSPxo4XSg@mail.gmail.com>
On Mon, Mar 03, 2014 at 03:59:36PM -0800, Linus Torvalds wrote:
> I doubt it's worth caring about. Even when passing things in memory,
> the end result isn't that much worse than the fget_light() model that
> passes just one of the two fields in memory.
I'm not sure if that's the right approach, TBH. I wonder if something
like
static inline struct fd fdget(int fd)
{
unsigned long v = __fdget(fd);
return (struct fd){(struct file *)(v & ~1), v & 1};
}
would not be a better starting point, with __fdget(fd) being
{
struct files_struct *files = current->files;
struct file *file;
if (atomic_read(&files->count) == 1) {
file = __fcheck_files(files, fd);
if (file && (file->f_mode & FMODE_PATH))
return 0;
return (unsigned long)file;
} else {
file = __fget(fd, FMODE_PATH);
return file ? 1 | (unsigned long)file : 0;
}
}
next prev parent reply other threads:[~2014-03-04 0:23 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-03-03 21:03 Update of file offset on write() etc. is non-atomic with I/O George Spelvin
2014-03-03 21:26 ` Al Viro
2014-03-03 21:52 ` Linus Torvalds
2014-03-03 22:01 ` Al Viro
2014-03-03 22:17 ` Linus Torvalds
2014-03-03 23:28 ` Al Viro
2014-03-03 23:34 ` Linus Torvalds
2014-03-03 23:42 ` Al Viro
2014-03-03 23:59 ` Linus Torvalds
2014-03-04 0:23 ` Al Viro [this message]
2014-03-04 0:42 ` Linus Torvalds
2014-03-04 1:05 ` Al Viro
2014-03-04 20:00 ` Al Viro
2014-03-04 21:17 ` Linus Torvalds
2014-03-05 0:04 ` Al Viro
2014-03-10 15:55 ` Al Viro
2014-03-03 22:55 ` Linus Torvalds
2014-03-03 23:23 ` Linus Torvalds
2014-03-03 23:39 ` Al Viro
2014-03-03 23:54 ` Linus Torvalds
2014-03-03 23:54 ` Al Viro
2014-03-04 20:11 ` Cedric Blancher
2014-03-04 0:07 ` George Spelvin
2014-05-04 7:04 ` Michael Kerrisk
[not found] <a8df285f-de7f-4a3a-9a19-e0ad07ab3a5c@blur>
2014-02-20 18:15 ` Zuckerman, Boris
2014-02-20 18:29 ` Al Viro
2014-02-21 6:01 ` Michael Kerrisk (man-pages)
2014-02-23 1:18 ` Kevin Easton
2014-02-23 7:38 ` Michael Kerrisk (man-pages)
-- strict thread matches above, loose matches on Subject: below --
2014-02-17 15:41 Michael Kerrisk (man-pages)
2014-02-18 13:00 ` Michael Kerrisk
2014-02-20 17:14 ` Linus Torvalds
2014-03-03 17:36 ` Linus Torvalds
2014-03-03 21:45 ` Al Viro
2014-03-03 21:56 ` Linus Torvalds
2014-03-03 22:09 ` Al Viro
2014-03-03 22:20 ` Linus Torvalds
2014-03-03 22:01 ` Linus Torvalds
2014-03-03 22:10 ` Al Viro
2014-03-03 22:22 ` Linus Torvalds
2014-03-06 15:03 ` Michael Kerrisk (man-pages)
2014-03-07 3:38 ` Yongzhi Pan
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=20140304002342.GS18016@ZenIV.linux.org.uk \
--to=viro@zeniv.linux.org.uk \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux@horizon.com \
--cc=torvalds@linux-foundation.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).