[<prev] [next>] [day] [month] [year] [list]
Message-ID: <497DFEA4.4040101@vlnb.net>
Date:	Mon, 26 Jan 2009 21:19:16 +0300
From:	Vladislav Bolkhovitin <vst@...b.net>
To:	linux-scsi@...r.kernel.org
CC:	linux-kernel@...r.kernel.org, scst-devel@...ts.sourceforge.net
Subject: [PATCH][SCST]: Fix for minor scst_user race
This patch fixed race in scst_user module, which can lead to usage of already freed
command. A misbehaving user space handler can make ucmd to get dead immediately
after the lock is released and we should catch it.
Signed-off-by: Vladislav Bolkhovitin <vst@...b.net>
 scst_user.c |    8 ++++++--
 1 file changed, 5 insertions(+), 1 deletion(-)
Index: scst/src/dev_handlers/scst_user.c
===================================================================
--- scst/src/dev_handlers/scst_user.c	(revision 641)
+++ scst/src/dev_handlers/scst_user.c	(working copy)
@@ -1721,6 +1729,7 @@ static int dev_user_reply_get_cmd(struct
 	kmem_cache_free(user_get_cmd_cachep, cmd);
 
 	spin_lock_irq(&dev->cmd_lists.cmd_list_lock);
+again:
 	res = dev_user_get_next_cmd(dev, &ucmd);
 	if (res == 0) {
 		int len;
@@ -1730,12 +1739,15 @@ static int dev_user_reply_get_cmd(struct
 		 * copy of dead data to the user space, which can lead to a
 		 * leak of sensitive information.
 		 */
-		ucmd_get(ucmd);
+		if (unlikely(ucmd_get_check(ucmd))) {
+			/* Oops, this ucmd is already being destroyed. Retry. */
+			goto again;
+		}
 		spin_unlock_irq(&dev->cmd_lists.cmd_list_lock);
 
 		EXTRACHECKS_BUG_ON(ucmd->user_cmd_payload_len == 0);
--
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/
Powered by blists - more mailing lists
 
