linux-kernel - Re: 2.6.20-mm2

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 for Android: free password hash cracker in your pocket

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

Message-Id: <200702230836.37096.david-b@pacbell.net>
Date:	Fri, 23 Feb 2007 08:36:36 -0800
From:	David Brownell <david-b@...bell.net>
To:	"Rafael J. Wysocki" <rjw@...k.pl>
Cc:	Andrew Morton <akpm@...ux-foundation.org>,
	linux-kernel@...r.kernel.org
Subject: Re: 2.6.20-mm2

> > > > rtc_cmos 00:02: rtc core: registered rtc_cmos as rtc0
> > > > Unable to handle kernel NULL pointer dereference at 0000000000000030 RIP: 
> > > >  [<ffffffff804032c3>] rtc_sysfs_remove_device+0x23/0x50

The bug isn't in rtc_cmos, but that's the only driver that would currently
show the bug.  (Triggered by finding the IRQ claimed by the legacy driver,
result of a "but it should still be safe!" misconfiguration.)

The following fixes this on my 2.6.21-rc1 system; same fix should apply
to MM2.

=====================	CUT HERE
Fix an oops on the rtc_device_unregister() path by waiting until the
last moment before nulling the rtc->ops vector.  Fix some potential
oopses by having the rtc_class_open()/rtc_class_close() interface
increase the RTC's reference count while an RTC handle is available
outside the RTC framework.

Signed-off-by: David Brownell <dbrownell@...rs.sourceforge.net>

Index: g26/drivers/rtc/class.c
===================================================================
--- g26.orig/drivers/rtc/class.c	2007-02-21 12:03:20.000000000 -0800
+++ g26/drivers/rtc/class.c	2007-02-23 08:00:52.000000000 -0800
@@ -204,10 +204,16 @@ EXPORT_SYMBOL_GPL(rtc_device_register);
  */
 void rtc_device_unregister(struct rtc_device *rtc)
 {
-	mutex_lock(&rtc->ops_lock);
-	rtc->ops = NULL;
-	mutex_unlock(&rtc->ops_lock);
-	class_device_unregister(&rtc->class_dev);
+	if (class_device_get(&rtc->class_dev) != NULL) {
+		mutex_lock(&rtc->ops_lock);
+		/* remove innards of this RTC, then disable it, before
+		 * letting any rtc_class_open() users access it again
+		 */
+		class_device_unregister(&rtc->class_dev);
+		rtc->ops = NULL;
+		mutex_unlock(&rtc->ops_lock);
+		class_device_put(&rtc->class_dev);
+	}
 }
 EXPORT_SYMBOL_GPL(rtc_device_unregister);
 
Index: g26/drivers/rtc/interface.c
===================================================================
--- g26.orig/drivers/rtc/interface.c	2006-11-27 15:37:13.000000000 -0800
+++ g26/drivers/rtc/interface.c	2007-02-23 07:45:09.000000000 -0800
@@ -179,7 +179,7 @@ struct class_device *rtc_class_open(char
 	down(&rtc_class->sem);
 	list_for_each_entry(class_dev_tmp, &rtc_class->children, node) {
 		if (strncmp(class_dev_tmp->class_id, name, BUS_ID_SIZE) == 0) {
-			class_dev = class_dev_tmp;
+			class_dev = class_device_get(class_dev_tmp);
 			break;
 		}
 	}
@@ -197,6 +197,7 @@ EXPORT_SYMBOL_GPL(rtc_class_open);
 void rtc_class_close(struct class_device *class_dev)
 {
 	module_put(to_rtc_device(class_dev)->owner);
+	class_device_put(class_dev);
 }
 EXPORT_SYMBOL_GPL(rtc_class_close);
 
-
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/