Skip to contentSkip to conversation

xfs: fix an undefined behaviour in _da3_path_shift

PATCH Superseded Qian Cai 3 participants 6 messages Feb 25, 2020 XFS filesystem

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Qian Cai

author
#0 lore raw

PATCH 1/1 xfs: fix an undefined behaviour in _da3_path_shift

state->path.active could be 1 in xfs_da3_node_lookup_int() and then in xfs_da3_path_shift() could see state->path.blk[-1].

 UBSAN: Undefined behaviour in fs/xfs/libxfs/xfs_da_btree.c:1989:14
 index -1 is out of range for type 'xfs_da_state_blk_t [5]'
 Call trace:
  dump_backtrace+0x0/0x2c8
  show_stack+0x20/0x2c
  dump_stack+0xe8/0x150
  __ubsan_handle_out_of_bounds+0xe4/0xfc
  xfs_da3_path_shift+0x860/0x86c [xfs]
  xfs_da3_node_lookup_int+0x7c8/0x934 [xfs]
  xfs_dir2_node_addname+0x2c8/0xcd0 [xfs]
  xfs_dir_createname+0x348/0x38c [xfs]
  xfs_create+0x6b0/0x8b4 [xfs]
  xfs_generic_create+0x12c/0x1f8 [xfs]
  xfs_vn_mknod+0x3c/0x4c [xfs]
  xfs_vn_create+0x34/0x44 [xfs]
  do_last+0xd4c/0x10c8
  path_openat+0xbc/0x2f4
  do_filp_open+0x74/0xf4
  do_sys_openat2+0x98/0x180
  __arm64_sys_openat+0xf8/0x170
  do_el0_svc+0x170/0x240
  el0_sync_handler+0x150/0x250
  el0_sync+0x164/0x180
---
 fs/xfs/libxfs/xfs_da_btree.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
fs/xfs/libxfs/xfs_da_btree.c +5 -1
index 875e04f82541..0906b7748a3f 100644
@@ -1986,7 +1986,11 @@ static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork)
 	ASSERT(path != NULL);
 	ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	level = (path->active-1) - 1;	/* skip bottom layer in path */
-	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
+
+	if (level >= 0)
+		blk = &path->blk[level];
+
+	for (; level >= 0; blk--, level--) {
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 					   blk->bp->b_addr);
 
Signature · 1 line
1.8.3.1
#1 lore raw
Quoted 2 lines · Qian Cai

state->path.active could be 1 in xfs_da3_node_lookup_int() and then in xfs_da3_path_shift() could see state->path.blk[-1].

Under what circumstancs can it be 1? Is this a longstanding bug in XFS? A corrupted filesystem? A deliberately corrupted filesystem?

Quoted 35 lines
 UBSAN: Undefined behaviour in fs/xfs/libxfs/xfs_da_btree.c:1989:14
 index -1 is out of range for type 'xfs_da_state_blk_t [5]'
 Call trace:
  dump_backtrace+0x0/0x2c8
  show_stack+0x20/0x2c
  dump_stack+0xe8/0x150
  __ubsan_handle_out_of_bounds+0xe4/0xfc
  xfs_da3_path_shift+0x860/0x86c [xfs]
  xfs_da3_node_lookup_int+0x7c8/0x934 [xfs]
  xfs_dir2_node_addname+0x2c8/0xcd0 [xfs]
  xfs_dir_createname+0x348/0x38c [xfs]
  xfs_create+0x6b0/0x8b4 [xfs]
  xfs_generic_create+0x12c/0x1f8 [xfs]
  xfs_vn_mknod+0x3c/0x4c [xfs]
  xfs_vn_create+0x34/0x44 [xfs]
  do_last+0xd4c/0x10c8
  path_openat+0xbc/0x2f4
  do_filp_open+0x74/0xf4
  do_sys_openat2+0x98/0x180
  __arm64_sys_openat+0xf8/0x170
  do_el0_svc+0x170/0x240
  el0_sync_handler+0x150/0x250
  el0_sync+0x164/0x180
---
 fs/xfs/libxfs/xfs_da_btree.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
fs/xfs/libxfs/xfs_da_btree.c +0 -0
index 875e04f82541..0906b7748a3f 100644

fs/xfs/libxfs/xfs_da_btree.c static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork) · line 1991

Quoted 7 lines
fs/xfs/libxfs/xfs_da_btree.c +3 -1
@@ -1986,7 +1986,11 @@ static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork)
 	ASSERT(path != NULL);
 	ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	level = (path->active-1) - 1;	/* skip bottom layer in path */
-	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
+
+	if (level >= 0)
+		blk = &path->blk[level];

...because if the reason is "corrupt metadata" then perhaps this should return -EFSCORRUPTED? But I don't know enough about the context to know the answer to that question.

--D

Quoted 8 lines
fs/xfs/libxfs/xfs_da_btree.c +2 -0
+
+	for (; level >= 0; blk--, level--) {
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 					   blk->bp->b_addr);
 
Signature · 1 line
1.8.3.1
author
#2 lore raw
Quoted 6 lines · Darrick J. Wong
Quoted 2 lines · Qian Cai

state->path.active could be 1 in xfs_da3_node_lookup_int() and then in xfs_da3_path_shift() could see state->path.blk[-1].

Under what circumstancs can it be 1? Is this a longstanding bug in XFS? A corrupted filesystem? A deliberately corrupted filesystem?

in xfs_da3_node_lookup_int(),

	for (blk = &state->path.blk[0], state->path.active = 1;
			 state->path.active <= XFS_DA_NODE_MAXDEPTH;
			 blk++, state->path.active++) {
<snip>
		if (magic == XFS_ATTR_LEAF_MAGIC ||
		    magic == XFS_ATTR3_LEAF_MAGIC) {
			blk->magic = XFS_ATTR_LEAF_MAGIC;
			blk->hashval = xfs_attr_leaf_lasthash(blk->bp, NULL);
			break;
		}

		if (magic == XFS_DIR2_LEAFN_MAGIC ||
		    magic == XFS_DIR3_LEAFN_MAGIC) {
			blk->magic = XFS_DIR2_LEAFN_MAGIC;
			blk->hashval = xfs_dir2_leaf_lasthash(args->dp,
							      blk->bp, NULL);
			break;

Isn't that if the first iteration in the loop calls any of those "break", it will have state->path.active = 1 ?

I suppose this is a long-standing bug that need UBSAN (no obvious harm could be done later because it will bail out immediately in xfs_da3_path_shift()) and a set of specific conditions to met to trigger.

Quoted 58 lines
Quoted 36 lines

 UBSAN: Undefined behaviour in fs/xfs/libxfs/xfs_da_btree.c:1989:14
 index -1 is out of range for type 'xfs_da_state_blk_t [5]'
 Call trace:
  dump_backtrace+0x0/0x2c8
  show_stack+0x20/0x2c
  dump_stack+0xe8/0x150
  __ubsan_handle_out_of_bounds+0xe4/0xfc
  xfs_da3_path_shift+0x860/0x86c [xfs]
  xfs_da3_node_lookup_int+0x7c8/0x934 [xfs]
  xfs_dir2_node_addname+0x2c8/0xcd0 [xfs]
  xfs_dir_createname+0x348/0x38c [xfs]
  xfs_create+0x6b0/0x8b4 [xfs]
  xfs_generic_create+0x12c/0x1f8 [xfs]
  xfs_vn_mknod+0x3c/0x4c [xfs]
  xfs_vn_create+0x34/0x44 [xfs]
  do_last+0xd4c/0x10c8
  path_openat+0xbc/0x2f4
  do_filp_open+0x74/0xf4
  do_sys_openat2+0x98/0x180
  __arm64_sys_openat+0xf8/0x170
  do_el0_svc+0x170/0x240
  el0_sync_handler+0x150/0x250
  el0_sync+0x164/0x180
---
 fs/xfs/libxfs/xfs_da_btree.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
fs/xfs/libxfs/xfs_da_btree.c +0 -0
index 875e04f82541..0906b7748a3f 100644

fs/xfs/libxfs/xfs_da_btree.c static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork) · line 1991

Quoted 7 lines
fs/xfs/libxfs/xfs_da_btree.c +3 -1
@@ -1986,7 +1986,11 @@ static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork)
 	ASSERT(path != NULL);
 	ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	level = (path->active-1) - 1;	/* skip bottom layer in path */
-	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
+
+	if (level >= 0)
+		blk = &path->blk[level];

...because if the reason is "corrupt metadata" then perhaps this should return -EFSCORRUPTED? But I don't know enough about the context to know the answer to that question.

--D

Quoted 8 lines
fs/xfs/libxfs/xfs_da_btree.c +2 -0
+
+	for (; level >= 0; blk--, level--) {
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 					   blk->bp->b_addr);
 
Signature · 1 line
1.8.3.1
#3 lore raw
Quoted 28 lines · Qian Cai
Quoted 6 lines · Darrick J. Wong
Quoted 2 lines · Qian Cai

state->path.active could be 1 in xfs_da3_node_lookup_int() and then in xfs_da3_path_shift() could see state->path.blk[-1].

Under what circumstancs can it be 1? Is this a longstanding bug in XFS? A corrupted filesystem? A deliberately corrupted filesystem?

in xfs_da3_node_lookup_int(),

	for (blk = &state->path.blk[0], state->path.active = 1;
			 state->path.active <= XFS_DA_NODE_MAXDEPTH;
			 blk++, state->path.active++) {
<snip>
		if (magic == XFS_ATTR_LEAF_MAGIC ||
		    magic == XFS_ATTR3_LEAF_MAGIC) {
			blk->magic = XFS_ATTR_LEAF_MAGIC;
			blk->hashval = xfs_attr_leaf_lasthash(blk->bp, NULL);
			break;
		}

		if (magic == XFS_DIR2_LEAFN_MAGIC ||
		    magic == XFS_DIR3_LEAFN_MAGIC) {
			blk->magic = XFS_DIR2_LEAFN_MAGIC;
			blk->hashval = xfs_dir2_leaf_lasthash(args->dp,
							      blk->bp, NULL);
			break;
Quoted 2 lines

Isn't that if the first iteration in the loop calls any of those "break", it will have state->path.active = 1 ?

Yes. The commit message ought to state that active == 1 is a valid state when we're trying to add an entry to a single dir leaf block and are trying to shift forward to see if there's a sibling block that would be a better place to put the new entry.

This is to build confidence in future readers that we actually understood the circumstances of the UBSAN error and aren't just monkeypatching the code to shut up the automated checks.

--D

Quoted 62 lines

I suppose this is a long-standing bug that need UBSAN (no obvious harm could be done later because it will bail out immediately in xfs_da3_path_shift()) and a set of specific conditions to met to trigger.

Quoted 58 lines
Quoted 36 lines

 UBSAN: Undefined behaviour in fs/xfs/libxfs/xfs_da_btree.c:1989:14
 index -1 is out of range for type 'xfs_da_state_blk_t [5]'
 Call trace:
  dump_backtrace+0x0/0x2c8
  show_stack+0x20/0x2c
  dump_stack+0xe8/0x150
  __ubsan_handle_out_of_bounds+0xe4/0xfc
  xfs_da3_path_shift+0x860/0x86c [xfs]
  xfs_da3_node_lookup_int+0x7c8/0x934 [xfs]
  xfs_dir2_node_addname+0x2c8/0xcd0 [xfs]
  xfs_dir_createname+0x348/0x38c [xfs]
  xfs_create+0x6b0/0x8b4 [xfs]
  xfs_generic_create+0x12c/0x1f8 [xfs]
  xfs_vn_mknod+0x3c/0x4c [xfs]
  xfs_vn_create+0x34/0x44 [xfs]
  do_last+0xd4c/0x10c8
  path_openat+0xbc/0x2f4
  do_filp_open+0x74/0xf4
  do_sys_openat2+0x98/0x180
  __arm64_sys_openat+0xf8/0x170
  do_el0_svc+0x170/0x240
  el0_sync_handler+0x150/0x250
  el0_sync+0x164/0x180
---
 fs/xfs/libxfs/xfs_da_btree.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
fs/xfs/libxfs/xfs_da_btree.c +0 -0
index 875e04f82541..0906b7748a3f 100644

fs/xfs/libxfs/xfs_da_btree.c static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork) · line 1991

Quoted 7 lines
fs/xfs/libxfs/xfs_da_btree.c +3 -1
@@ -1986,7 +1986,11 @@ static inline int xfs_dabuf_nfsb(struct xfs_mount *mp, int whichfork)
 	ASSERT(path != NULL);
 	ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	level = (path->active-1) - 1;	/* skip bottom layer in path */
-	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
+
+	if (level >= 0)
+		blk = &path->blk[level];

...because if the reason is "corrupt metadata" then perhaps this should return -EFSCORRUPTED? But I don't know enough about the context to know the answer to that question.

--D

Quoted 8 lines
fs/xfs/libxfs/xfs_da_btree.c +2 -0
+
+	for (; level >= 0; blk--, level--) {
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 					   blk->bp->b_addr);
 
Signature · 1 line
1.8.3.1
#4 lore raw
In reply to Qian Cai

I think we code do this a tad more cleaner, something like:

fs/xfs/libxfs/xfs_da_btree.c +2 -1
index 875e04f82541..542a4edfcf54 100644
@@ -1986,7 +1986,8 @@ xfs_da3_path_shift(
 	ASSERT(path != NULL);
 	ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	level = (path->active-1) - 1;	/* skip bottom layer in path */
-	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
+	for ( ; level >= 0; level--) {
+		blk = &path->blk[level];
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 					   blk->bp->b_addr);
 
author
#5 lore raw
Quoted 7 lines · Christoph Hellwig

I think we code do this a tad more cleaner, something like:

fs/xfs/libxfs/xfs_da_btree.c +0 -0
index 875e04f82541..542a4edfcf54 100644

fs/xfs/libxfs/xfs_da_btree.c xfs_da3_path_shift( · line 1993

Quoted 9 lines
fs/xfs/libxfs/xfs_da_btree.c +2 -1
@@ -1986,7 +1986,8 @@ xfs_da3_path_shift(
 	ASSERT(path != NULL);
 	ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	level = (path->active-1) - 1;	/* skip bottom layer in path */
-	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
+	for ( ; level >= 0; level--) {
+		blk = &path->blk[level];
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 					   blk->bp->b_addr);
 

Yes, indeed. I'll send a v2 until Darrick is still not convinced that

"path->active == 1" could reach here?