Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1738417

[PATCH 4.9 61/77] net/netfilter/nf_conntrack_core: Fix net_conntrack_lock()

From Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Newsgroups linux.kernel
Subject [PATCH 4.9 61/77] net/netfilter/nf_conntrack_core: Fix net_conntrack_lock()
Date 2017-09-24 22:50 +0200
Message-ID <utm0J-4Xd-105@gated-at.bofh.it> (permalink)
References <utlQZ-4Ti-7@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


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

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

From: Manfred Spraul <manfred@colorfullife.com>

commit 3ef0c7a730de0bae03d86c19570af764fa3c4445 upstream.

As we want to remove spin_unlock_wait() and replace it with explicit
spin_lock()/spin_unlock() calls, we can use this to simplify the
locking.

In addition:
- Reading nf_conntrack_locks_all needs ACQUIRE memory ordering.
- The new code avoids the backwards loop.

Only slightly tested, I did not manage to trigger calls to
nf_conntrack_all_lock().

V2: With improved comments, to clearly show how the barriers
    pair.

Fixes: b16c29191dc8 ("netfilter: nf_conntrack: use safer way to lock all buckets")
Signed-off-by: Manfred Spraul <manfred@colorfullife.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: Sasha Levin <sasha.levin@oracle.com>
Cc: Pablo Neira Ayuso <pablo@netfilter.org>
Cc: netfilter-devel@vger.kernel.org
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 net/netfilter/nf_conntrack_core.c |   52 +++++++++++++++++++++-----------------
 1 file changed, 29 insertions(+), 23 deletions(-)

--- a/net/netfilter/nf_conntrack_core.c
+++ b/net/netfilter/nf_conntrack_core.c
@@ -95,19 +95,26 @@ static struct conntrack_gc_work conntrac
 
 void nf_conntrack_lock(spinlock_t *lock) __acquires(lock)
 {
+	/* 1) Acquire the lock */
 	spin_lock(lock);
-	while (unlikely(nf_conntrack_locks_all)) {
-		spin_unlock(lock);
 
-		/*
-		 * Order the 'nf_conntrack_locks_all' load vs. the
-		 * spin_unlock_wait() loads below, to ensure
-		 * that 'nf_conntrack_locks_all_lock' is indeed held:
-		 */
-		smp_rmb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
-		spin_unlock_wait(&nf_conntrack_locks_all_lock);
-		spin_lock(lock);
-	}
+	/* 2) read nf_conntrack_locks_all, with ACQUIRE semantics
+	 * It pairs with the smp_store_release() in nf_conntrack_all_unlock()
+	 */
+	if (likely(smp_load_acquire(&nf_conntrack_locks_all) == false))
+		return;
+
+	/* fast path failed, unlock */
+	spin_unlock(lock);
+
+	/* Slow path 1) get global lock */
+	spin_lock(&nf_conntrack_locks_all_lock);
+
+	/* Slow path 2) get the lock we want */
+	spin_lock(lock);
+
+	/* Slow path 3) release the global lock */
+	spin_unlock(&nf_conntrack_locks_all_lock);
 }
 EXPORT_SYMBOL_GPL(nf_conntrack_lock);
 
@@ -148,28 +155,27 @@ static void nf_conntrack_all_lock(void)
 	int i;
 
 	spin_lock(&nf_conntrack_locks_all_lock);
-	nf_conntrack_locks_all = true;
 
-	/*
-	 * Order the above store of 'nf_conntrack_locks_all' against
-	 * the spin_unlock_wait() loads below, such that if
-	 * nf_conntrack_lock() observes 'nf_conntrack_locks_all'
-	 * we must observe nf_conntrack_locks[] held:
-	 */
-	smp_mb(); /* spin_lock(&nf_conntrack_locks_all_lock) */
+	nf_conntrack_locks_all = true;
 
 	for (i = 0; i < CONNTRACK_LOCKS; i++) {
-		spin_unlock_wait(&nf_conntrack_locks[i]);
+		spin_lock(&nf_conntrack_locks[i]);
+
+		/* This spin_unlock provides the "release" to ensure that
+		 * nf_conntrack_locks_all==true is visible to everyone that
+		 * acquired spin_lock(&nf_conntrack_locks[]).
+		 */
+		spin_unlock(&nf_conntrack_locks[i]);
 	}
 }
 
 static void nf_conntrack_all_unlock(void)
 {
-	/*
-	 * All prior stores must be complete before we clear
+	/* All prior stores must be complete before we clear
 	 * 'nf_conntrack_locks_all'. Otherwise nf_conntrack_lock()
 	 * might observe the false value but not the entire
-	 * critical section:
+	 * critical section.
+	 * It pairs with the smp_load_acquire() in nf_conntrack_lock()
 	 */
 	smp_store_release(&nf_conntrack_locks_all, false);
 	spin_unlock(&nf_conntrack_locks_all_lock);

Back to linux.kernel | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCH 4.9 00/77] 4.9.52-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 60/77] PCI: pciehp: Report power fault only once until we clear it Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 30/77] powerpc: Fix DAR reporting when alignment handler faults Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 55/77] ftrace: Fix memleak when unregistering dynamic ops when tracing disabled Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 42/77] scsi: zfcp: trace high part of "new" 64 bit SCSI LUN Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 54/77] ftrace: Fix selftest goto location on error Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 75/77] mac80211_hwsim: Use proper TX power Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 21/77] MIPS: math-emu: <MADDF|MSUBF>.<D|S>: Fix some cases of infinite inputs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 27/77] [PATCH - RESEND] crypto: AF_ALG - remove SGL terminator indicator when  chaining Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 04/77] orangefs: Dont clear SGID when inheriting ACLs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 06/77] drm/sun4i: Implement drm_driver lastclose to restore fbdev console Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 44/77] scsi: megaraid_sas: Check valid aen class range to avoid kernel panic Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:40 +0200
  [PATCH 4.9 63/77] s390/mm: fix race on mm->context.flush_mm Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:50 +0200
  [PATCH 4.9 65/77] media: uvcvideo: Prevent heap overflow when accessing mapped controls Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:50 +0200
  [PATCH 4.9 67/77] bcache: initialize dirty stripes in flash_dev_run() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:50 +0200
  [PATCH 4.9 61/77] net/netfilter/nf_conntrack_core: Fix net_conntrack_lock() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 22:50 +0200
  [PATCH 4.9 56/77] tracing: Add barrier to trace_printk() buffer nesting modification Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 68/77] bcache: Fix leak of bdev reference Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 24/77] MIPS: math-emu: <MADDF|MSUBF>.S: Fix accuracy (32-bit case) Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 74/77] mac80211: fix VLAN handling with TXQs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 59/77] PCI: shpchp: Enable bridge bus mastering if MSI is enabled Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 58/77] ARC: Re-enable MMU upon Machine Check exception Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 26/77] crypto: ccp - Fix XTS-AES-128 support on v5 CCPs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 66/77] PM / devfreq: Fix memory leak when fail to register device Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 64/77] media: v4l2-compat-ioctl32: Fix timespec conversion Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 69/77] bcache: do not subtract sectors_to_gc for bypassed IO Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 73/77] bcache: fix bch_hprint crash and improve output Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 62/77] s390/mm: fix local TLB flushing vs. detach of an mm address space Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 57/77] tracing: Apply trace_clock changes to instance max buffer Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 18/77] MIPS: math-emu: MINA.<D|S>: Fix some cases of infinity and zero inputs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 76/77] mac80211: flush hw_roc_start work before cancelling the ROC Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:10 +0200
  [PATCH 4.9 39/77] scsi: zfcp: fix missing trace records for early returns in TMF eh handlers Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 40/77] scsi: zfcp: fix payload with full FCP_RSP IU in SCSI trace records Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 41/77] scsi: zfcp: trace HBA FSF response by default on dismiss or timedout late response Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 34/77] skd: Submit requests to firmware before triggering the doorbell Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 33/77] skd: Avoid that module unloading triggers a use-after-free Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 50/77] scsi: sg: factor out sg_fill_request_table() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 11/77] pinctrl/amd: save pin registers over suspend/resume Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 48/77] scsi: sg: use standard lists for sg_requests Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 03/77] mm: prevent double decrease of nr_reserved_highatomic Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 35/77] scsi: zfcp: fix queuecommand for scsi_eh commands when DIX enabled Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 47/77] scsi: sg: remove save_scat_len Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 17/77] MIPS: math-emu: <MAXA|MINA>.<D|S>: Fix cases of both infinite inputs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 45/77] scsi: megaraid_sas: Return pended IOCTLs with cmd_status MFI_STAT_WRONG_STATE in case adapter is dead Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 31/77] block: Relax a check in blk_start_queue() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 53/77] scsi: qla2xxx: Fix an integer overflow in sysfs code Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 10/77] tty: fix __tty_insert_flip_char regression Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 29/77] ext4: fix quota inconsistency during orphan cleanup for read-only mounts Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 43/77] scsi: megaraid_sas: set minimum value of resetwaittime to be 1 secs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 12/77] Input: i8042 - add Gigabyte P57 to the keyboard reset table Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 16/77] MIPS: math-emu: <MAXA|MINA>.<D|S>: Fix cases of input values with opposite signs Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 15/77] MIPS: math-emu: <MAX|MIN>.<D|S>: Fix cases of both inputs negative Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 07/77] IB/addr: Fix setting source address in addr6_resolve() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 09/77] tty: improve tty_insert_flip_char() slow path Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 19/77] MIPS: math-emu: Handle zero accumulator case in MADDF and MSUBF separately Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 51/77] scsi: sg: fixup infoleak when using SG_GET_REQUEST_TABLE Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 28/77] ext4: fix incorrect quotaoff if the quota feature is enabled Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 05/77] IB/{qib, hfi1}: Avoid flow control testing for RDMA write operation Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  [PATCH 4.9 20/77] MIPS: math-emu: <MADDF|MSUBF>.<D|S>: Fix NaN propagation Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-24 23:20 +0200
  Re: [PATCH 4.9 00/77] 4.9.52-stable review Guenter Roeck <linux@roeck-us.net> - 2017-09-25 03:10 +0200
  Re: [PATCH 4.9 00/77] 4.9.52-stable review Tom Gall <tom.gall@linaro.org> - 2017-09-25 06:40 +0200
    Re: [PATCH 4.9 00/77] 4.9.52-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-09-25 08:30 +0200
  Re: [PATCH 4.9 00/77] 4.9.52-stable review Shuah Khan <shuahkh@osg.samsung.com> - 2017-09-26 01:20 +0200

csiph-web