Linux kernel -stable discussions
 help / color / mirror / Atom feed
* [PATCH 4.4 01/13] GFS2: don't set rgrp gl_object until it's inserted into rgrp tree
@ 2019-08-27 23:09 Ben Hutchings
  2019-08-27 23:10 ` [PATCH 4.4 02/13] net: arc_emac: fix koops caused by sk_buff free Ben Hutchings
                   ` (12 more replies)
  0 siblings, 13 replies; 15+ messages in thread
From: Ben Hutchings @ 2019-08-27 23:09 UTC (permalink / raw)
  To: Greg Kroah-Hartman, Sasha Levin; +Cc: stable

From: Bob Peterson <rpeterso@redhat.com>

commit 36e4ad0316c017d5b271378ed9a1c9a4b77fab5f upstream.

Before this patch, function read_rindex_entry would set a rgrp
glock's gl_object pointer to itself before inserting the rgrp into
the rgrp rbtree. The problem is: if another process was also reading
the rgrp in, and had already inserted its newly created rgrp, then
the second call to read_rindex_entry would overwrite that value,
then return a bad return code to the caller. Later, other functions
would reference the now-freed rgrp memory by way of gl_object.
In some cases, that could result in gfs2_rgrp_brelse being called
twice for the same rgrp: once for the failed attempt and once for
the "real" rgrp release. Eventually the kernel would panic.
There are also a number of other things that could go wrong when
a kernel module is accessing freed storage. For example, this could
result in rgrp corruption because the fake rgrp would point to a
fake bitmap in memory too, causing gfs2_inplace_reserve to search
some random memory for free blocks, and find some, since we were
never setting rgd->rd_bits to NULL before freeing it.

This patch fixes the problem by not setting gl_object until we
have successfully inserted the rgrp into the rbtree. Also, it sets
rd_bits to NULL as it frees them, which will ensure any accidental
access to the wrong rgrp will result in a kernel panic rather than
file system corruption, which is preferred.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
[bwh: Backported to 4.4: adjust context]
Signed-off-by: Ben Hutchings <ben.hutchings@codethink.co.uk>
---
 fs/gfs2/rgrp.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/fs/gfs2/rgrp.c b/fs/gfs2/rgrp.c
index ef24894edecc..9c159e6ad116 100644
--- a/fs/gfs2/rgrp.c
+++ b/fs/gfs2/rgrp.c
@@ -739,6 +739,7 @@ void gfs2_clear_rgrpd(struct gfs2_sbd *sdp)
 
 		gfs2_free_clones(rgd);
 		kfree(rgd->rd_bits);
+		rgd->rd_bits = NULL;
 		return_all_reservations(rgd);
 		kmem_cache_free(gfs2_rgrpd_cachep, rgd);
 	}
@@ -933,10 +934,6 @@ static int read_rindex_entry(struct gfs2_inode *ip)
 	if (error)
 		goto fail;
 
-	rgd->rd_gl->gl_object = rgd;
-	rgd->rd_gl->gl_vm.start = (rgd->rd_addr * bsize) & PAGE_CACHE_MASK;
-	rgd->rd_gl->gl_vm.end = PAGE_CACHE_ALIGN((rgd->rd_addr +
-						  rgd->rd_length) * bsize) - 1;
 	rgd->rd_rgl = (struct gfs2_rgrp_lvb *)rgd->rd_gl->gl_lksb.sb_lvbptr;
 	rgd->rd_flags &= ~(GFS2_RDF_UPTODATE | GFS2_RDF_PREFERRED);
 	if (rgd->rd_data > sdp->sd_max_rg_data)
@@ -944,14 +941,20 @@ static int read_rindex_entry(struct gfs2_inode *ip)
 	spin_lock(&sdp->sd_rindex_spin);
 	error = rgd_insert(rgd);
 	spin_unlock(&sdp->sd_rindex_spin);
-	if (!error)
+	if (!error) {
+		rgd->rd_gl->gl_object = rgd;
+		rgd->rd_gl->gl_vm.start = (rgd->rd_addr * bsize) & PAGE_MASK;
+		rgd->rd_gl->gl_vm.end = PAGE_ALIGN((rgd->rd_addr +
+						    rgd->rd_length) * bsize) - 1;
 		return 0;
+	}
 
 	error = 0; /* someone else read in the rgrp; free it and ignore it */
 	gfs2_glock_put(rgd->rd_gl);
 
 fail:
 	kfree(rgd->rd_bits);
+	rgd->rd_bits = NULL;
 	kmem_cache_free(gfs2_rgrpd_cachep, rgd);
 	return error;
 }
-- 
Ben Hutchings, Software Developer                         Codethink Ltd
https://www.codethink.co.uk/                 Dale House, 35 Dale Street
                                     Manchester, M1 2HF, United Kingdom



^ permalink raw reply related	[flat|nested] 15+ messages in thread
end of thread, other threads:[~2019-08-28  8:39 UTC | newest]

Thread overview: 15+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-08-27 23:09 [PATCH 4.4 01/13] GFS2: don't set rgrp gl_object until it's inserted into rgrp tree Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 02/13] net: arc_emac: fix koops caused by sk_buff free Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 03/13] vhost-net: set packet weight of tx polling to 2 * vq size Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 04/13] vhost_net: use packet weight for rx handler, too Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 05/13] vhost_net: introduce vhost_exceeds_weight() Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 06/13] vhost: " Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 07/13] vhost_net: fix possible infinite loop Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 08/13] vhost: scsi: add weight support Ben Hutchings
2019-08-27 23:10 ` [PATCH 4.4 09/13] siphash: add cryptographically secure PRF Ben Hutchings
2019-08-27 23:11 ` [PATCH 4.4 10/13] siphash: implement HalfSipHash1-3 for hash tables Ben Hutchings
2019-08-28  8:38   ` Greg Kroah-Hartman
2019-08-27 23:11 ` [PATCH 4.4 11/13] inet: switch IP ID generator to siphash Ben Hutchings
2019-08-27 23:11 ` [PATCH 4.4 12/13] netfilter: ctnetlink: don't use conntrack/expect object addresses as id Ben Hutchings
2019-08-27 23:11 ` [PATCH 4.4 13/13] netfilter: conntrack: Use consistent ct id hash calculation Ben Hutchings
2019-08-28  2:57 ` [PATCH 4.4 01/13] GFS2: don't set rgrp gl_object until it's inserted into rgrp tree Sasha Levin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox