linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Ben Hutchings <ben@decadent.org.uk>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk,
	Darren Hart <dvhart@linux.intel.com>,
	Dave Jones <davej@redat.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@elte.hu>, John Kacur <jkacur@redhat.com>,
	Linus Torvalds <torvalds@linux-foundation.org>
Subject: [ 83/89] futex: avoid wake_futex() for a PI futex_q
Date: Mon, 03 Dec 2012 14:33:09 +0000	[thread overview]
Message-ID: <20121203143159.298836370@decadent.org.uk> (raw)
In-Reply-To: <20121203143146.549859007@decadent.org.uk>

3.2-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Darren Hart <dvhart@linux.intel.com>

commit aa10990e028cac3d5e255711fb9fb47e00700e35 upstream.

Dave Jones reported a bug with futex_lock_pi() that his trinity test
exposed.  Sometime between queue_me() and taking the q.lock_ptr, the
lock_ptr became NULL, resulting in a crash.

While futex_wake() is careful to not call wake_futex() on futex_q's with
a pi_state or an rt_waiter (which are either waiting for a
futex_unlock_pi() or a PI futex_requeue()), futex_wake_op() and
futex_requeue() do not perform the same test.

Update futex_wake_op() and futex_requeue() to test for q.pi_state and
q.rt_waiter and abort with -EINVAL if detected.  To ensure any future
breakage is caught, add a WARN() to wake_futex() if the same condition
is true.

This fix has seen 3 hours of testing with "trinity -c futex" on an
x86_64 VM with 4 CPUS.

[akpm@linux-foundation.org: tidy up the WARN()]
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Reported-by: Dave Jones <davej@redat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: John Kacur <jkacur@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
---
 kernel/futex.c |   18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

diff --git a/kernel/futex.c b/kernel/futex.c
index 20ef219..19eb089 100644
--- a/kernel/futex.c
+++ b/kernel/futex.c
@@ -843,6 +843,9 @@ static void wake_futex(struct futex_q *q)
 {
 	struct task_struct *p = q->task;
 
+	if (WARN(q->pi_state || q->rt_waiter, "refusing to wake PI futex\n"))
+		return;
+
 	/*
 	 * We set q->lock_ptr = NULL _before_ we wake up the task. If
 	 * a non-futex wake up happens on another CPU then the task
@@ -1078,6 +1081,10 @@ retry_private:
 
 	plist_for_each_entry_safe(this, next, head, list) {
 		if (match_futex (&this->key, &key1)) {
+			if (this->pi_state || this->rt_waiter) {
+				ret = -EINVAL;
+				goto out_unlock;
+			}
 			wake_futex(this);
 			if (++ret >= nr_wake)
 				break;
@@ -1090,6 +1097,10 @@ retry_private:
 		op_ret = 0;
 		plist_for_each_entry_safe(this, next, head, list) {
 			if (match_futex (&this->key, &key2)) {
+				if (this->pi_state || this->rt_waiter) {
+					ret = -EINVAL;
+					goto out_unlock;
+				}
 				wake_futex(this);
 				if (++op_ret >= nr_wake2)
 					break;
@@ -1098,6 +1109,7 @@ retry_private:
 		ret += op_ret;
 	}
 
+out_unlock:
 	double_unlock_hb(hb1, hb2);
 out_put_keys:
 	put_futex_key(&key2);
@@ -1387,9 +1399,13 @@ retry_private:
 		/*
 		 * FUTEX_WAIT_REQEUE_PI and FUTEX_CMP_REQUEUE_PI should always
 		 * be paired with each other and no other futex ops.
+		 *
+		 * We should never be requeueing a futex_q with a pi_state,
+		 * which is awaiting a futex_unlock_pi().
 		 */
 		if ((requeue_pi && !this->rt_waiter) ||
-		    (!requeue_pi && this->rt_waiter)) {
+		    (!requeue_pi && this->rt_waiter) ||
+		    this->pi_state) {
 			ret = -EINVAL;
 			break;
 		}



  parent reply	other threads:[~2012-12-03 14:36 UTC|newest]

Thread overview: 100+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-12-03 14:31 [ 00/89] 3.2.35-stable review Ben Hutchings
2012-12-03 14:31 ` [ 01/89] UBIFS: introduce categorized lprops counter Ben Hutchings
2012-12-03 14:31 ` [ 02/89] UBIFS: fix mounting problems after power cuts Ben Hutchings
2012-12-03 14:31 ` [ 03/89] futex: Handle futex_pi OWNER_DIED take over correctly Ben Hutchings
2012-12-03 14:31 ` [ 04/89] mac80211: sync acccess to tx_filtered/ps_tx_buf queues Ben Hutchings
2012-12-03 14:31 ` [ 05/89] ASoC: wm8978: pll incorrectly configured when codec is master Ben Hutchings
2012-12-03 14:31 ` [ 06/89] device_cgroup: fix RCU usage Ben Hutchings
2012-12-06 19:36   ` Herton Ronaldo Krzesinski
2012-12-07  1:41     ` Ben Hutchings
2012-12-03 14:31 ` [ 07/89] ASoC: dapm: Use card_list during DAPM shutdown Ben Hutchings
2012-12-03 14:31 ` [ 08/89] s390/signal: set correct address space control Ben Hutchings
2012-12-03 14:31 ` [ 09/89] wireless: allow 40 MHz on world roaming channels 12/13 Ben Hutchings
2012-12-03 14:31 ` [ 10/89] drm/i915/sdvo: clean up connectors on intel_sdvo_init() failures Ben Hutchings
2012-12-03 14:31 ` [ 11/89] s390/gup: add missing TASK_SIZE check to get_user_pages_fast() Ben Hutchings
2012-12-03 14:31 ` [ 12/89] USB: option: add Novatel E362 and Dell Wireless 5800 USB IDs Ben Hutchings
2012-12-03 14:31 ` [ 13/89] USB: option: add Alcatel X220/X500D " Ben Hutchings
2012-12-03 14:32 ` [ 14/89] drm/radeon: fix logic error in atombios_encoders.c Ben Hutchings
2012-12-03 14:32 ` [ 15/89] ttm: Clear the ttm page allocated from high memory zone correctly Ben Hutchings
2012-12-03 14:32 ` [ 16/89] memcg: oom: fix totalpages calculation for memory.swappiness==0 Ben Hutchings
2012-12-03 14:32 ` [ 17/89] tmpfs: change final i_blocks BUG to WARNING Ben Hutchings
2012-12-03 14:32 ` [ 18/89] x86: Exclude E820_RESERVED regions and memory holes above 4 GB from direct mapping Ben Hutchings
2012-12-03 14:32 ` [ 19/89] x86, mm: Find_early_table_space based on ranges that are actually being mapped Ben Hutchings
2012-12-03 14:32 ` [ 20/89] x86, mm: Undo incorrect revert in arch/x86/mm/init.c Ben Hutchings
2012-12-03 14:32 ` [ 21/89] netfilter: Mark SYN/ACK packets as invalid from original direction Ben Hutchings
2012-12-03 14:32 ` [ 22/89] netfilter: Validate the sequence number of dataless ACK packets as well Ben Hutchings
2012-12-03 14:32 ` [ 23/89] netfilter: nf_nat: dont check for port change on ICMP tuples Ben Hutchings
2012-12-03 14:32 ` [ 24/89] ipv4: avoid undefined behavior in do_ip_setsockopt() Ben Hutchings
2012-12-03 14:32 ` [ 25/89] ipv6: setsockopt(IPIPPROTO_IPV6, IPV6_MINHOPCOUNT) forgot to set return value Ben Hutchings
2012-12-03 14:32 ` [ 26/89] net: correct check in dev_addr_del() Ben Hutchings
2012-12-03 14:32 ` [ 27/89] net-rps: Fix brokeness causing OOO packets Ben Hutchings
2012-12-03 14:32 ` [ 28/89] usb: use usb_serial_put in usb_serial_probe errors Ben Hutchings
2012-12-03 14:32 ` [ 29/89] PCI : Calculate right add_size Ben Hutchings
2012-12-03 14:32 ` [ 30/89] Input: i8042 - also perform controller reset when suspending Ben Hutchings
2012-12-03 14:32 ` [ 31/89] ixgbe: add support for new 82599 device id Ben Hutchings
2012-12-03 14:32 ` [ 32/89] ixgbe: add support for X540-AT1 Ben Hutchings
2012-12-03 14:32 ` [ 33/89] drm/i915: Check VBIOS value for determining LVDS dual channel mode, too Ben Hutchings
2012-12-03 14:32 ` [ 34/89] get_dvb_firmware: fix download site for tda10046 firmware Ben Hutchings
2012-12-03 14:32 ` [ 35/89] m68k: fix sigset_t accessor functions Ben Hutchings
2012-12-03 14:32 ` [ 36/89] HID: add quirk for Freescale i.MX28 ROM recovery Ben Hutchings
2012-12-03 14:32 ` [ 37/89] brcm80211: smac: only print block-ack timeout message at trace level Ben Hutchings
2012-12-03 14:32 ` [ 38/89] bas_gigaset: fix pre_reset handling Ben Hutchings
2012-12-03 14:32 ` [ 39/89] GFS2: Test bufdata with buffer locked and gfs2_log_lock held Ben Hutchings
2012-12-03 14:32 ` [ 40/89] ptp: update adjfreq callback description Ben Hutchings
2012-12-03 14:32 ` [ 41/89] watchdog: iTCO_wdt: add Intel Lynx Point DeviceIDs Ben Hutchings
2012-12-03 14:32 ` [ 42/89] acer-wmi: support for P key on TM8372 Ben Hutchings
2012-12-03 14:32 ` [ 43/89] xhci: Remove warnings about MSI and MSI-X capabilities Ben Hutchings
2012-12-03 14:32 ` [ 44/89] xhci: Remove scary warnings about transfer issues Ben Hutchings
2012-12-03 14:32 ` [ 45/89] x86, mce, therm_throt: Dont report power limit and package level thermal throttle events in mcelog Ben Hutchings
2012-12-03 14:32 ` [ 46/89] Input: bcm5974 - set BUTTONPAD property Ben Hutchings
2012-12-03 14:32 ` [ 47/89] watchdog: using u64 in get_sample_period() Ben Hutchings
2012-12-03 14:32 ` [ 48/89] x86, amd: Disable way access filter on Piledriver CPUs Ben Hutchings
2012-12-03 14:32 ` [ 49/89] mtd: ofpart: Fix incorrect NULL check in parse_ofoldpart_partitions() Ben Hutchings
2012-12-03 14:32 ` [ 50/89] mtd: slram: invalid checking of absolute end address Ben Hutchings
2012-12-03 14:32 ` [ 51/89] jffs2: Fix lock acquisition order bug in jffs2_write_begin Ben Hutchings
2012-12-03 14:32 ` [ 52/89] [SCSI] isci: copy fis 0x34 response into proper buffer Ben Hutchings
2012-12-03 14:32 ` [ 53/89] mac80211: deinitialize ibss-internals after emptiness check Ben Hutchings
2012-12-03 14:32 ` [ 54/89] [PARISC] fix virtual aliasing issue in get_shared_area() Ben Hutchings
2012-12-03 14:32 ` [ 55/89] rtlwifi: rtl8192cu: Add new USB ID Ben Hutchings
2012-12-03 14:32 ` [ 56/89] mwifiex: fix system hang issue in cmd timeout error case Ben Hutchings
2012-12-03 14:32 ` [ 57/89] mwifiex: report error to MMC core if we cannot suspend Ben Hutchings
2012-12-03 14:32 ` [ 58/89] xfs: drop buffer io reference when a bad bio is built Ben Hutchings
2012-12-03 14:32 ` [ 59/89] ALSA: ua101, usx2y: fix broken MIDI output Ben Hutchings
2012-12-03 14:32 ` [ 60/89] sparc64: not any error from do_sigaltstack() should fail rt_sigreturn() Ben Hutchings
2012-12-03 14:32 ` [ 61/89] reiserfs: Fix lock ordering during remount Ben Hutchings
2012-12-03 14:32 ` [ 62/89] reiserfs: Protect reiserfs_quota_on() with write lock Ben Hutchings
2012-12-03 14:32 ` [ 63/89] reiserfs: Protect reiserfs_quota_write() " Ben Hutchings
2012-12-03 14:32 ` [ 64/89] reiserfs: Move quota calls out of " Ben Hutchings
2012-12-03 14:32 ` [ 65/89] md: Reassigned the parameters if read_seqretry returned true in func md_is_badblock Ben Hutchings
2012-12-03 14:32 ` [ 66/89] md: Avoid write invalid address if read_seqretry returned true Ben Hutchings
2012-12-03 14:32 ` [ 67/89] drm/radeon: properly track the crtc not_enabled case evergreen_mc_stop() Ben Hutchings
2012-12-03 15:24   ` Josh Boyer
2012-12-03 15:35     ` Deucher, Alexander
2012-12-03 23:26       ` Josh Boyer
2012-12-03 23:40         ` Deucher, Alexander
2012-12-04 14:35           ` Josh Boyer
2012-12-06 18:14             ` Greg KH
2012-12-09 23:25       ` Ben Hutchings
2012-12-03 14:32 ` [ 68/89] radeon: add AGPMode 1 quirk for RV250 Ben Hutchings
2012-12-03 14:32 ` [ 69/89] x86-32: Fix invalid stack address while in softirq Ben Hutchings
2012-12-03 14:32 ` [ 70/89] x86-32: Export kernel_stack_pointer() for modules Ben Hutchings
2012-12-03 14:32 ` [ 71/89] x86, microcode, AMD: Add support for family 16h processors Ben Hutchings
2012-12-03 14:32 ` [ 72/89] ALSA: hda - Add new codec ALC283 ALC290 support Ben Hutchings
2012-12-03 14:32 ` [ 73/89] ALSA: hda - Add support for Realtek ALC292 Ben Hutchings
2012-12-03 14:33 ` [ 74/89] selinux: fix sel_netnode_insert() suspicious rcu dereference Ben Hutchings
2012-12-03 14:33 ` [ 75/89] Dove: Attempt to fix PMU/RTC interrupts Ben Hutchings
2012-12-03 14:33 ` [ 76/89] Dove: Fix irq_to_pmu() Ben Hutchings
2012-12-03 14:33 ` [ 77/89] ARM: Kirkwood: Update PCI-E fixup Ben Hutchings
2012-12-03 14:33 ` [ 78/89] [PARISC] fix user-triggerable panic on parisc Ben Hutchings
2012-12-03 14:33 ` [ 79/89] dm: fix deadlock with request based dm and queue request_fn recursion Ben Hutchings
2012-12-03 14:33 ` [ 80/89] block: Dont access request after it might be freed Ben Hutchings
2012-12-03 14:33 ` [ 81/89] jbd: Fix lock ordering bug in journal_unmap_buffer() Ben Hutchings
2012-12-03 14:33 ` [ 82/89] can: bcm: initialize ifindex for timeouts without previous frame reception Ben Hutchings
2012-12-03 14:33 ` Ben Hutchings [this message]
2012-12-03 14:33 ` [ 84/89] mm/vmemmap: fix wrong use of virt_to_page Ben Hutchings
2012-12-03 14:33 ` [ 85/89] mm: vmscan: fix endless loop in kswapd balancing Ben Hutchings
2012-12-03 14:33 ` [ 86/89] mm: soft offline: split thp at the beginning of soft_offline_page() Ben Hutchings
2012-12-03 14:33 ` [ 87/89] workqueue: exit rescuer_thread() as TASK_RUNNING Ben Hutchings
2012-12-03 14:33 ` [ 88/89] intel_idle: initial IVB support Ben Hutchings
2012-12-03 14:33 ` [ 89/89] intel_idle: enable IVB Xeon support Ben Hutchings
2012-12-03 15:09 ` [ 00/89] 3.2.35-stable review Ben Hutchings

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=20121203143159.298836370@decadent.org.uk \
    --to=ben@decadent.org.uk \
    --cc=akpm@linux-foundation.org \
    --cc=alan@lxorguk.ukuu.org.uk \
    --cc=davej@redat.com \
    --cc=dvhart@linux.intel.com \
    --cc=jkacur@redhat.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=peterz@infradead.org \
    --cc=stable@vger.kernel.org \
    --cc=tglx@linutronix.de \
    --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).