From: Alexander Viro <viro@math.psu.edu>
To: Linus Torvalds <torvalds@transmeta.com>
Cc: David Woodhouse <dwmw2@infradead.org>, linux-kernel@vger.kernel.org
Subject: [PATCH] Re: kernel BUG at buffer.c:827 in test12-pre6 and 7
Date: Fri, 8 Dec 2000 13:11:54 -0500 (EST) [thread overview]
Message-ID: <Pine.GSO.4.21.0012081301160.27010-100000@weyl.math.psu.edu> (raw)
In-Reply-To: <Pine.GSO.4.21.0012080410180.24770-100000@weyl.math.psu.edu>
Folks, see if the following patch helps. AFAICS it closes a pretty real
race - we could call block_write_full_page() for a page that has sync
IO in progress and blindly change ->b_end_io callbacks on the bh with
pending requests. With a little bit of bad luck they would complete before
we got to ll_rw_block(), thus leading to extra UnlockPage(). All it takes
is a pageout on a page that got partial write recently - if some fragments
were still unmapped we get to call get_block() on them and it can easily
block, providing a decent window for that race.
Fix: postpone changing ->b_end_io until the call of ll_rw_block(); if by
the time of ll_rw_block() some fragments will still have IO in progress -
wait on them.
Comments?
Cheers,
Al
--- buffer.c Fri Dec 8 16:19:53 2000
+++ buffer.c.new Fri Dec 8 16:26:44 2000
@@ -1577,6 +1577,26 @@
* "Dirty" is valid only with the last case (mapped+uptodate).
*/
+static void write_array_async(struct page *page, struct buffer_head **p, int n)
+{
+ int i;
+ if (!n) {
+ UnlockPage(page);
+ return;
+ }
+ /*
+ * If there are pending requests on these guys - wait before changing
+ * ->b_end_io.
+ */
+ for (i=0; i<n; i++) {
+ wait_on_buffer(p[i]);
+ set_bit(BH_Uptodate, &p[i]->b_state);
+ set_bit(BH_Dirty, &p[i]->b_state);
+ p[i]->b_end_io = end_buffer_io_async;
+ }
+ ll_rw_block(WRITE, n, p);
+}
+
/*
* block_write_full_page() is SMP-safe - currently it's still
* being called with the kernel lock held, but the code is ready.
@@ -1616,28 +1636,17 @@
if (buffer_new(bh))
unmap_underlying_metadata(bh);
}
- set_bit(BH_Uptodate, &bh->b_state);
- set_bit(BH_Dirty, &bh->b_state);
- bh->b_end_io = end_buffer_io_async;
atomic_inc(&bh->b_count);
arr[nr++] = bh;
bh = bh->b_this_page;
block++;
} while (bh != head);
- if (nr) {
- ll_rw_block(WRITE, nr, arr);
- } else {
- UnlockPage(page);
- }
+ write_array_async(page, arr, nr);
SetPageUptodate(page);
return 0;
out:
- if (nr) {
- ll_rw_block(WRITE, nr, arr);
- } else {
- UnlockPage(page);
- }
+ write_array_async(page, arr, nr);
ClearPageUptodate(page);
return err;
}
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/
next prev parent reply other threads:[~2000-12-08 18:42 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2000-12-07 22:42 kernel BUG at buffer.c:827 in test12-pre6 and 7 Joseph Cheek
2000-12-07 23:14 ` Keith Owens
2000-12-08 0:22 ` Linus Torvalds
2000-12-08 1:23 ` Joseph Cheek
2000-12-08 3:03 ` Keith Owens
2000-12-08 2:16 ` Joseph Cheek
2000-12-08 7:27 ` Tom Leete
2000-12-08 9:07 ` David Woodhouse
2000-12-08 9:58 ` [found?] " Alexander Viro
2000-12-08 18:11 ` Alexander Viro [this message]
2000-12-08 18:48 ` [PATCH] " Linus Torvalds
2000-12-08 19:13 ` Alexander Viro
2000-12-08 19:39 ` Linus Torvalds
2000-12-08 19:50 ` Daniel Phillips
2000-12-08 21:17 ` Linus Torvalds
2000-12-08 22:30 ` Alexander Viro
2000-12-08 22:42 ` Linus Torvalds
2000-12-09 4:59 ` Alexander Viro
2000-12-09 8:45 ` Linus Torvalds
2000-12-09 8:56 ` Linus Torvalds
2000-12-09 10:40 ` Alexander Viro
2000-12-09 12:56 ` Andries Brouwer
2000-12-09 13:11 ` Alexander Viro
2000-12-09 21:25 ` Mikulas Patocka
2000-12-10 1:11 ` Linus Torvalds
2000-12-09 17:28 ` Linus Torvalds
2000-12-09 18:43 ` Andi Kleen
2000-12-09 14:00 ` David S. Miller
2000-12-09 15:37 ` David Woodhouse
2000-12-12 0:54 ` [FIXED!] " Joseph Cheek
[not found] <90rjbg$8eso1$1@fido.engr.sgi.com>
2000-12-08 22:22 ` [PATCH] " Rajagopal Ananthanarayanan
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=Pine.GSO.4.21.0012081301160.27010-100000@weyl.math.psu.edu \
--to=viro@math.psu.edu \
--cc=dwmw2@infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=torvalds@transmeta.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).