linux-kernel - Re: [PATCH] fix a race between /proc/lock

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20150602150104.GA19282@twins.programming.kicks-ass.net>
Date:	Tue, 2 Jun 2015 17:01:04 +0200
From:	Peter Zijlstra <peterz@...radead.org>
To:	Jerome Marchand <jmarchan@...hat.com>
Cc:	Ingo Molnar <mingo@...hat.com>, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] fix a race between /proc/lock_stat and module unloading

On Tue, Jun 02, 2015 at 04:15:24PM +0200, Jerome Marchand wrote:
> Yes, I can't reproduce the issue anymore.

Great; queued the below. Slight changes

 - s/WRITE_ONCE/RCU_INIT_POINTER/ which should be similar but more descriptive
 - const char *cname to avoid a compile warn.

---
Subject: lockdep: Fix a race between /proc/lock_stat and module unload
From: Peter Zijlstra <peterz@...radead.org>
Date: Tue, 2 Jun 2015 12:50:13 +0200

The lock_class iteration of /proc/lock_stat is not serialized against
the lockdep_free_key_range() call from module unload.

Therefore it can happen that we find a class of which ->name/->key are
no longer valid.

There is a further bug in zap_class() that left ->name dangling. Cure
this. Use RCU_INIT_POINTER() because NULL.

Since lockdep_free_key_range() is rcu_sched serialized, we can read
both ->name and ->key under rcu_read_lock_sched() (preempt-disable)
and be assured that if we observe a !NULL value it stays safe to use
for as long as we hold that lock.

If we observe both NULL, skip the entry.

Cc: Ingo Molnar <mingo@...hat.com>
Reported-by: Jerome Marchand <jmarchan@...hat.com>
Tested-by: Jerome Marchand <jmarchan@...hat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
Link: http://lkml.kernel.org/r/20150602105013.GS3644@twins.programming.kicks-ass.net
---
 kernel/locking/lockdep.c      |    3 ++-
 kernel/locking/lockdep_proc.c |   22 +++++++++++++++++-----
 2 files changed, 19 insertions(+), 6 deletions(-)

--- a/kernel/locking/lockdep.c
+++ b/kernel/locking/lockdep.c
@@ -3900,7 +3900,8 @@ static void zap_class(struct lock_class
 	list_del_rcu(&class->hash_entry);
 	list_del_rcu(&class->lock_entry);
 
-	class->key = NULL;
+	RCU_INIT_POINTER(class->key, NULL);
+	RCU_INIT_POINTER(class->name, NULL);
 }
 
 static inline int within(const void *addr, void *start, unsigned long size)
--- a/kernel/locking/lockdep_proc.c
+++ b/kernel/locking/lockdep_proc.c
@@ -426,10 +426,12 @@ static void seq_lock_time(struct seq_fil
 
 static void seq_stats(struct seq_file *m, struct lock_stat_data *data)
 {
-	char name[39];
-	struct lock_class *class;
+	struct lockdep_subclass_key *ckey;
 	struct lock_class_stats *stats;
+	struct lock_class *class;
+	const char *cname;
 	int i, namelen;
+	char name[39];
 
 	class = data->class;
 	stats = &data->stats;
@@ -440,15 +442,25 @@ static void seq_stats(struct seq_file *m
 	if (class->subclass)
 		namelen -= 2;
 
-	if (!class->name) {
+	rcu_read_lock_sched();
+	cname = rcu_dereference_sched(class->name);
+	ckey  = rcu_dereference_sched(class->key);
+
+	if (!cname && !ckey) {
+		rcu_read_unlock_sched();
+		return;
+
+	} else if (!cname) {
 		char str[KSYM_NAME_LEN];
 		const char *key_name;
 
-		key_name = __get_key_name(class->key, str);
+		key_name = __get_key_name(ckey, str);
 		snprintf(name, namelen, "%s", key_name);
 	} else {
-		snprintf(name, namelen, "%s", class->name);
+		snprintf(name, namelen, "%s", cname);
 	}
+	rcu_read_unlock_sched();
+
 	namelen = strlen(name);
 	if (class->name_version > 1) {
 		snprintf(name+namelen, 3, "#%d", class->name_version);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/