svn commit: r195003 - head/sys/ufs/ufs

[Sean Nicholas Barkas]

Author: snb
Date: Thu Jun 25 20:40:13 2009
New Revision: 195003
URL: http://svn.freebsd.org/changeset/base/195003

Log:
  Fix a bug reported by pho@ where one can induce a panic by decreasing
  vfs.ufs.dirhash_maxmem below the current amount of memory used by dirhash. When
  ufsdirhash_build() is called with the memory in use greater than dirhash_maxmem,
  it attempts to free up memory by calling ufsdirhash_recycle(). If successful in
  freeing enough memory, ufsdirhash_recycle() leaves the dirhash list locked. But
  at this point in ufsdirhash_build(), the list is not explicitly unlocked after
  the call(s) to ufsdirhash_recycle(). When we next attempt to lock the dirhash
  list, we will get a "panic: _mtx_lock_sleep: recursed on non-recursive mutex
  dirhash list".
  
  Tested by:	pho
  Approved by:	dwmalone (mentor)
  MFC after:	3 weeks

Modified:
  head/sys/ufs/ufs/ufs_dirhash.c

Modified: head/sys/ufs/ufs/ufs_dirhash.c
==============================================================================
--- head/sys/ufs/ufs/ufs_dirhash.c	Thu Jun 25 20:35:46 2009	(r195002)
+++ head/sys/ufs/ufs/ufs_dirhash.c	Thu Jun 25 20:40:13 2009	(r195003)
@@ -348,9 +348,12 @@ ufsdirhash_build(struct inode *ip)
 	int dirblocks, i, j, memreqd, nblocks, narrays, nslots, slot;
 
 	/* Take care of a decreased sysctl value. */
-	while (ufs_dirhashmem > ufs_dirhashmaxmem)
+	while (ufs_dirhashmem > ufs_dirhashmaxmem) {
 		if (ufsdirhash_recycle(0) != 0)
 			return (-1);
+		/* Recycled enough memory, so unlock the list. */
+		DIRHASHLIST_UNLOCK();
+	}
 
 	/* Check if we can/should use dirhash. */
 	if (ip->i_size < ufs_mindirhashsize || OFSFMT(ip->i_vnode) ||