From: Greg KH <gregkh@suse.de>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: torvalds@linux-foundation.org, akpm@linux-foundation.org,
alan@lxorguk.ukuu.org.uk, xfs@oss.sgi.com,
Alex Elder <aelder@sgi.com>, Christoph Hellwig <hch@lst.de>
Subject: [46/53] xfs: avoid direct I/O write vs buffered I/O race
Date: Tue, 22 Nov 2011 16:23:53 -0800 [thread overview]
Message-ID: <20111123002408.908319178@clark.kroah.org> (raw)
In-Reply-To: <20111123002419.GA8531@kroah.com>
3.0-stable review patch. If anyone has any objections, please let me know.
------------------
From: Christoph Hellwig <hch@infradead.org>
commit c58cb165bd44de8aaee9755a144136ae743be116 upstream.
Currently a buffered reader or writer can add pages to the pagecache
while we are waiting for the iolock in xfs_file_dio_aio_write. Prevent
this by re-checking mapping->nrpages after we got the iolock, and if
nessecary upgrade the lock to exclusive mode. To simplify this a bit
only take the ilock inside of xfs_file_aio_write_checks.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Alex Elder <aelder@sgi.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
---
fs/xfs/linux-2.6/xfs_file.c | 17 ++++++++++++++---
1 file changed, 14 insertions(+), 3 deletions(-)
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -669,6 +669,7 @@ xfs_file_aio_write_checks(
xfs_fsize_t new_size;
int error = 0;
+ xfs_rw_ilock(ip, XFS_ILOCK_EXCL);
error = generic_write_checks(file, pos, count, S_ISBLK(inode->i_mode));
if (error) {
xfs_rw_iunlock(ip, XFS_ILOCK_EXCL | *iolock);
@@ -760,14 +761,24 @@ xfs_file_dio_aio_write(
*iolock = XFS_IOLOCK_EXCL;
else
*iolock = XFS_IOLOCK_SHARED;
- xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
+ xfs_rw_ilock(ip, *iolock);
ret = xfs_file_aio_write_checks(file, &pos, &count, iolock);
if (ret)
return ret;
+ /*
+ * Recheck if there are cached pages that need invalidate after we got
+ * the iolock to protect against other threads adding new pages while
+ * we were waiting for the iolock.
+ */
+ if (mapping->nrpages && *iolock == XFS_IOLOCK_SHARED) {
+ xfs_rw_iunlock(ip, *iolock);
+ *iolock = XFS_IOLOCK_EXCL;
+ xfs_rw_ilock(ip, *iolock);
+ }
+
if (mapping->nrpages) {
- WARN_ON(*iolock != XFS_IOLOCK_EXCL);
ret = -xfs_flushinval_pages(ip, (pos & PAGE_CACHE_MASK), -1,
FI_REMAPF_LOCKED);
if (ret)
@@ -812,7 +823,7 @@ xfs_file_buffered_aio_write(
size_t count = ocount;
*iolock = XFS_IOLOCK_EXCL;
- xfs_rw_ilock(ip, XFS_ILOCK_EXCL | *iolock);
+ xfs_rw_ilock(ip, *iolock);
ret = xfs_file_aio_write_checks(file, &pos, &count, iolock);
if (ret)
next prev parent reply other threads:[~2011-11-23 0:42 UTC|newest]
Thread overview: 54+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-11-23 0:24 [00/53] 3.0.11-stable review Greg KH
2011-11-23 0:23 ` [01/53] genirq: Fix irqfixup, irqpoll regression Greg KH
2011-11-23 0:23 ` [02/53] [SCSI] fix WARNING: at drivers/scsi/scsi_lib.c:1704 Greg KH
2011-11-23 0:23 ` [03/53] [SCSI] hpsa: Disable ASPM Greg KH
2011-11-23 0:23 ` [04/53] [SCSI] aacraid: controller hangs if kernel uses non-default ASPM policy Greg KH
2011-11-23 0:23 ` [05/53] [media] saa7164: Add support for another HVR2200 hardware revision Greg KH
2011-11-23 0:23 ` [06/53] drm/i915/pch: Save/restore PCH_PORT_HOTPLUG across suspend Greg KH
2011-11-23 0:23 ` [07/53] ARM: 7150/1: Allow kernel unaligned accesses on ARMv6+ Greg KH
2011-11-23 0:23 ` [08/53] Net, libertas: Resolve memory leak in if_spi_host_to_card() Greg KH
2011-11-23 0:23 ` [09/53] rt2x00: Fix sleep-while-atomic bug in powersaving code Greg KH
2011-11-23 0:23 ` [10/53] mac80211: fix NULL dereference in radiotap code Greg KH
2011-11-23 0:23 ` [11/53] mac80211: fix bug in ieee80211_build_probe_req Greg KH
2011-11-23 0:23 ` [12/53] nl80211: fix HT capability attribute validation Greg KH
2011-11-23 0:23 ` [13/53] cfg80211: fix bug on regulatory core exit on access to last_request Greg KH
2011-11-23 0:23 ` [14/53] ip6_tunnel: copy parms.name after register_netdevice Greg KH
2011-11-23 0:23 ` [15/53] PM / driver core: disable devices runtime PM during shutdown Greg KH
2011-11-23 0:23 ` [16/53] pch_phub: Support new device LAPIS Semiconductor ML7831 IOH Greg KH
2011-11-23 0:23 ` [17/53] pch_phub: Fix MAC address writing issue for LAPIS ML7831 Greg KH
2011-11-23 0:23 ` [18/53] pch_uart: Fix hw-flow control issue Greg KH
2011-11-23 0:23 ` [19/53] pch_uart: Fix DMA resource leak issue Greg KH
2011-11-23 0:23 ` [20/53] pch_uart: Support new device LAPIS Semiconductor ML7831 IOH Greg KH
2011-11-23 0:23 ` [21/53] tty: hvc_dcc: Fix duplicate character inputs Greg KH
2011-11-23 0:23 ` [22/53] TTY: ldisc, allow waiting for ldisc arbitrarily long Greg KH
2011-11-23 0:23 ` [23/53] TTY: ldisc, move wait idle to caller Greg KH
2011-11-23 0:23 ` [24/53] TTY: ldisc, wait for ldisc infinitely in hangup Greg KH
2011-11-23 0:23 ` [25/53] nfs: when attempting to open a directory, fall back on normal lookup (try #5) Greg KH
2011-11-23 0:23 ` [26/53] pcie-gadget-spear: Add "platform:" prefix for platform modalias Greg KH
2011-11-23 0:23 ` [27/53] drivers/base/node.c: fix compilation error with older versions of gcc Greg KH
2011-11-23 0:23 ` [28/53] xhci: Set slot and ep0 flags for address command Greg KH
2011-11-23 0:23 ` [29/53] usb, xhci: Clear warm reset change event during init Greg KH
2011-11-23 0:23 ` [30/53] usb, xhci: fix lockdep warning on endpoint timeout Greg KH
2011-11-23 0:23 ` [31/53] USB: XHCI: resume root hubs when the controller resumes Greg KH
2011-11-23 0:23 ` [32/53] USB: option: release new PID for ZTE 3G modem Greg KH
2011-11-23 0:23 ` [33/53] USB: option: add PID of Huawei E173s " Greg KH
2011-11-23 0:23 ` [34/53] USB: serial: pl2303: rm duplicate id Greg KH
2011-11-23 0:23 ` [35/53] USB: cdc-acm: Fix disconnect() vs close() race Greg KH
2011-11-23 0:23 ` [36/53] USB: workaround for bug in old version of GCC Greg KH
2011-11-23 0:23 ` [37/53] USB: ark3116 initialisation fix Greg KH
2011-11-23 0:23 ` [38/53] USB: Fix Corruption issue in USB ftdi driver ftdi_sio.c Greg KH
2011-11-23 0:23 ` [39/53] usb-storage: Accept 8020i-protocol commands longer than 12 bytes Greg KH
2011-11-23 0:23 ` [40/53] USB: EHCI: fix HUB TT scheduling issue with iso transfer Greg KH
2011-11-23 0:23 ` [41/53] USB: add quirk for Logitech C600 web cam Greg KH
2011-11-23 0:23 ` [42/53] USB: quirks: adding more quirky webcams to avoid squeaky audio Greg KH
2011-11-23 0:23 ` [43/53] xfs: fix error handling for synchronous writes Greg KH
2011-11-23 0:23 ` [44/53] xfs: fix xfs_mark_inode_dirty during umount Greg KH
2011-11-23 0:23 ` [45/53] xfs: dont serialise direct IO reads on page cache Greg KH
2011-11-23 0:23 ` Greg KH [this message]
2011-11-23 0:23 ` [47/53] xfs: Return -EIO when xfs_vn_getattr() failed Greg KH
2011-11-23 0:23 ` [48/53] xfs: fix buffer flushing during unmount Greg KH
2011-11-23 0:23 ` [49/53] xfs: Fix possible memory corruption in xfs_readlink Greg KH
2011-11-23 0:23 ` [50/53] xfs: use doalloc flag in xfs_qm_dqattach_one() Greg KH
2011-11-23 0:23 ` [51/53] xfs: fix ->write_inode return values Greg KH
2011-11-23 0:23 ` [52/53] drm/i915: fix IVB cursor support Greg KH
2011-11-23 0:24 ` [53/53] drm/i915: always set FDI composite sync bit Greg KH
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=20111123002408.908319178@clark.kroah.org \
--to=gregkh@suse.de \
--cc=aelder@sgi.com \
--cc=akpm@linux-foundation.org \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=hch@lst.de \
--cc=linux-kernel@vger.kernel.org \
--cc=stable@vger.kernel.org \
--cc=torvalds@linux-foundation.org \
--cc=xfs@oss.sgi.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
all inboxes | Powered by JetHome®