Several users reported this crash of NULL pointer or general protection,
the story is that we add a rbtree for speedup ulist iteration, and we
use krealloc() to address ulist growth, and krealloc() use memcpy to copy
old data to new memory area, so it''s OK for an array as it
doesn''t use
pointers while it''s not OK for a rbtree as it uses pointers.
So krealloc() will mess up our rbtree and it ends up with crash.
Reviewed-by: Wang Shilong <wangsl-fnst@cn.fujitsu.com>
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
---
v3: fix a return value problem(Thanks Wang Shilong).
v2: fix an use-after-free bug and a finger error(Thanks Zach and Josef).
fs/btrfs/ulist.c | 15 +++++++++++++++
1 files changed, 15 insertions(+), 0 deletions(-)
diff --git a/fs/btrfs/ulist.c b/fs/btrfs/ulist.c
index 7b417e2..b0a523b2 100644
--- a/fs/btrfs/ulist.c
+++ b/fs/btrfs/ulist.c
@@ -205,6 +205,10 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
u64 new_alloced = ulist->nodes_alloced + 128;
struct ulist_node *new_nodes;
void *old = NULL;
+ int i;
+
+ for (i = 0; i < ulist->nnodes; i++)
+ rb_erase(&ulist->nodes[i].rb_node, &ulist->root);
/*
* if nodes_alloced == ULIST_SIZE no memory has been allocated
@@ -224,6 +228,17 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
ulist->nodes = new_nodes;
ulist->nodes_alloced = new_alloced;
+
+ /*
+ * krealloc actually uses memcpy, which does not copy rb_node
+ * pointers, so we have to do it ourselves. Otherwise we may
+ * be bitten by crashes.
+ */
+ for (i = 0; i < ulist->nnodes; i++) {
+ ret = ulist_rbtree_insert(ulist, &ulist->nodes[i]);
+ if (ret < 0)
+ return ret;
+ }
}
ulist->nodes[ulist->nnodes].val = val;
ulist->nodes[ulist->nnodes].aux = aux;
--
1.7.7
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs"
in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Zach Brown
2013-Jun-28 19:43 UTC
Re: [PATCH v3] Btrfs: fix crash regarding to ulist_add_merge
On Fri, Jun 28, 2013 at 12:37:45PM +0800, Liu Bo wrote:> Several users reported this crash of NULL pointer or general protection, > the story is that we add a rbtree for speedup ulist iteration, and we > use krealloc() to address ulist growth, and krealloc() use memcpy to copy > old data to new memory area, so it''s OK for an array as it doesn''t use > pointers while it''s not OK for a rbtree as it uses pointers. > > So krealloc() will mess up our rbtree and it ends up with crash. > > Reviewed-by: Wang Shilong <wangsl-fnst@cn.fujitsu.com> > Signed-off-by: Liu Bo <bo.li.liu@oracle.com>Yeah, this should fix the probem. Thanks for being persistent. Reviewed-by: Zach Brown <zab@redhat.com>> + for (i = 0; i < ulist->nnodes; i++) > + rb_erase(&ulist->nodes[i].rb_node, &ulist->root);(still twitching over here because this is a bunch of work that achieves nothing :)) - z -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Jun 28, 2013 at 12:43:14PM -0700, Zach Brown wrote:> On Fri, Jun 28, 2013 at 12:37:45PM +0800, Liu Bo wrote: > > Several users reported this crash of NULL pointer or general protection, > > the story is that we add a rbtree for speedup ulist iteration, and we > > use krealloc() to address ulist growth, and krealloc() use memcpy to copy > > old data to new memory area, so it''s OK for an array as it doesn''t use > > pointers while it''s not OK for a rbtree as it uses pointers. > > > > So krealloc() will mess up our rbtree and it ends up with crash. > > > > Reviewed-by: Wang Shilong <wangsl-fnst@cn.fujitsu.com> > > Signed-off-by: Liu Bo <bo.li.liu@oracle.com> > > Yeah, this should fix the probem. Thanks for being persistent. > > Reviewed-by: Zach Brown <zab@redhat.com> > > > + for (i = 0; i < ulist->nnodes; i++) > > + rb_erase(&ulist->nodes[i].rb_node, &ulist->root); > > (still twitching over here because this is a bunch of work that achieves > nothing :))Hmm, I think that this is necessary for the inline array inside ulist, so I keep it :) - liubo -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Jan Schmidt
2013-Jul-29 08:23 UTC
Re: [PATCH v3] Btrfs: fix crash regarding to ulist_add_merge
On Fri, June 28, 2013 at 06:37 (+0200), Liu Bo wrote:> Several users reported this crash of NULL pointer or general protection, > the story is that we add a rbtree for speedup ulist iteration, and we > use krealloc() to address ulist growth, and krealloc() use memcpy to copy > old data to new memory area, so it''s OK for an array as it doesn''t use > pointers while it''s not OK for a rbtree as it uses pointers. > > So krealloc() will mess up our rbtree and it ends up with crash. > > Reviewed-by: Wang Shilong <wangsl-fnst@cn.fujitsu.com> > Signed-off-by: Liu Bo <bo.li.liu@oracle.com> > --- > v3: fix a return value problem(Thanks Wang Shilong). > v2: fix an use-after-free bug and a finger error(Thanks Zach and Josef). > > fs/btrfs/ulist.c | 15 +++++++++++++++ > 1 files changed, 15 insertions(+), 0 deletions(-) > > diff --git a/fs/btrfs/ulist.c b/fs/btrfs/ulist.c > index 7b417e2..b0a523b2 100644 > --- a/fs/btrfs/ulist.c > +++ b/fs/btrfs/ulist.c > @@ -205,6 +205,10 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux, > u64 new_alloced = ulist->nodes_alloced + 128; > struct ulist_node *new_nodes; > void *old = NULL; > + int i; > + > + for (i = 0; i < ulist->nnodes; i++) > + rb_erase(&ulist->nodes[i].rb_node, &ulist->root); > > /* > * if nodes_alloced == ULIST_SIZE no memory has been allocated > @@ -224,6 +228,17 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux, > > ulist->nodes = new_nodes; > ulist->nodes_alloced = new_alloced; > + > + /* > + * krealloc actually uses memcpy, which does not copy rb_node > + * pointers, so we have to do it ourselves. Otherwise we may > + * be bitten by crashes. > + */ > + for (i = 0; i < ulist->nnodes; i++) { > + ret = ulist_rbtree_insert(ulist, &ulist->nodes[i]); > + if (ret < 0) > + return ret; > + } > } > ulist->nodes[ulist->nnodes].val = val; > ulist->nodes[ulist->nnodes].aux = aux; >Reviewed-by: Jan Schmidt <list.btrfs@jan-o-sch.net> Josef, how about sending this one for the next 3.11 rc and to 3.10 stable? Any objections? -Jan -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Jul 29, 2013 at 10:23:34AM +0200, Jan Schmidt wrote:> On Fri, June 28, 2013 at 06:37 (+0200), Liu Bo wrote: > > Several users reported this crash of NULL pointer or general protection, > > the story is that we add a rbtree for speedup ulist iteration, and we > > use krealloc() to address ulist growth, and krealloc() use memcpy to copy > > old data to new memory area, so it''s OK for an array as it doesn''t use > > pointers while it''s not OK for a rbtree as it uses pointers. > > > > So krealloc() will mess up our rbtree and it ends up with crash. > > > > Reviewed-by: Wang Shilong <wangsl-fnst@cn.fujitsu.com> > > Signed-off-by: Liu Bo <bo.li.liu@oracle.com> > > --- > > v3: fix a return value problem(Thanks Wang Shilong). > > v2: fix an use-after-free bug and a finger error(Thanks Zach and Josef). > > > > fs/btrfs/ulist.c | 15 +++++++++++++++ > > 1 files changed, 15 insertions(+), 0 deletions(-) > > > > diff --git a/fs/btrfs/ulist.c b/fs/btrfs/ulist.c > > index 7b417e2..b0a523b2 100644 > > --- a/fs/btrfs/ulist.c > > +++ b/fs/btrfs/ulist.c > > @@ -205,6 +205,10 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux, > > u64 new_alloced = ulist->nodes_alloced + 128; > > struct ulist_node *new_nodes; > > void *old = NULL; > > + int i; > > + > > + for (i = 0; i < ulist->nnodes; i++) > > + rb_erase(&ulist->nodes[i].rb_node, &ulist->root); > > > > /* > > * if nodes_alloced == ULIST_SIZE no memory has been allocated > > @@ -224,6 +228,17 @@ int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux, > > > > ulist->nodes = new_nodes; > > ulist->nodes_alloced = new_alloced; > > + > > + /* > > + * krealloc actually uses memcpy, which does not copy rb_node > > + * pointers, so we have to do it ourselves. Otherwise we may > > + * be bitten by crashes. > > + */ > > + for (i = 0; i < ulist->nnodes; i++) { > > + ret = ulist_rbtree_insert(ulist, &ulist->nodes[i]); > > + if (ret < 0) > > + return ret; > > + } > > } > > ulist->nodes[ulist->nnodes].val = val; > > ulist->nodes[ulist->nnodes].aux = aux; > > > > Reviewed-by: Jan Schmidt <list.btrfs@jan-o-sch.net> > > Josef, how about sending this one for the next 3.11 rc and to 3.10 stable? Any > objections?A good candidate for -stable, along with josef''s ''last ref'' fixes about __tree_mod_log_rewind(), thanks, -liubo -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html