* [PATCH 5.4 1/3] ext4: add two helper functions extent_logical_end() and pa_logical_end()
@ 2023-10-28 6:53 Baokun Li
2023-10-28 6:53 ` [PATCH 5.4 2/3] ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow Baokun Li
` (2 more replies)
0 siblings, 3 replies; 6+ messages in thread
From: Baokun Li @ 2023-10-28 6:53 UTC (permalink / raw)
To: stable
Cc: gregkh, sashal, tytso, jack, ritesh.list, patches, yangerkun,
libaokun1
commit 43bbddc067883d94de7a43d5756a295439fbe37d upstream.
When we use lstart + len to calculate the end of free extent or prealloc
space, it may exceed the maximum value of 4294967295(0xffffffff) supported
by ext4_lblk_t and cause overflow, which may lead to various problems.
Therefore, we add two helper functions, extent_logical_end() and
pa_logical_end(), to limit the type of end to loff_t, and also convert
lstart to loff_t for calculation to avoid overflow.
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/r/20230724121059.11834-2-libaokun1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Conflicts:
fs/ext4/mballoc.c
Signed-off-by: Baokun Li <libaokun1@huawei.com>
---
fs/ext4/mballoc.c | 7 +++----
fs/ext4/mballoc.h | 14 ++++++++++++++
2 files changed, 17 insertions(+), 4 deletions(-)
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 9099e112fda5..e07420c6ddee 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -3124,7 +3124,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
/* first, let's learn actual file size
* given current request is allocated */
- size = ac->ac_o_ex.fe_logical + EXT4_C2B(sbi, ac->ac_o_ex.fe_len);
+ size = extent_logical_end(sbi, &ac->ac_o_ex);
size = size << bsbits;
if (size < i_size_read(ac->ac_inode))
size = i_size_read(ac->ac_inode);
@@ -3462,8 +3462,7 @@ ext4_mb_use_preallocated(struct ext4_allocation_context *ac)
/* all fields in this condition don't change,
* so we can skip locking for them */
if (ac->ac_o_ex.fe_logical < pa->pa_lstart ||
- ac->ac_o_ex.fe_logical >= (pa->pa_lstart +
- EXT4_C2B(sbi, pa->pa_len)))
+ ac->ac_o_ex.fe_logical >= pa_logical_end(sbi, pa))
continue;
/* non-extent files can't have physical blocks past 2^32 */
@@ -4234,7 +4233,7 @@ static void ext4_mb_group_or_file(struct ext4_allocation_context *ac)
if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY))
return;
- size = ac->ac_o_ex.fe_logical + EXT4_C2B(sbi, ac->ac_o_ex.fe_len);
+ size = extent_logical_end(sbi, &ac->ac_o_ex);
isize = (i_size_read(ac->ac_inode) + ac->ac_sb->s_blocksize - 1)
>> bsbits;
diff --git a/fs/ext4/mballoc.h b/fs/ext4/mballoc.h
index 88c98f17e3d9..4e442bad6d73 100644
--- a/fs/ext4/mballoc.h
+++ b/fs/ext4/mballoc.h
@@ -199,6 +199,20 @@ static inline ext4_fsblk_t ext4_grp_offs_to_block(struct super_block *sb,
(fex->fe_start << EXT4_SB(sb)->s_cluster_bits);
}
+static inline loff_t extent_logical_end(struct ext4_sb_info *sbi,
+ struct ext4_free_extent *fex)
+{
+ /* Use loff_t to avoid end exceeding ext4_lblk_t max. */
+ return (loff_t)fex->fe_logical + EXT4_C2B(sbi, fex->fe_len);
+}
+
+static inline loff_t pa_logical_end(struct ext4_sb_info *sbi,
+ struct ext4_prealloc_space *pa)
+{
+ /* Use loff_t to avoid end exceeding ext4_lblk_t max. */
+ return (loff_t)pa->pa_lstart + EXT4_C2B(sbi, pa->pa_len);
+}
+
typedef int (*ext4_mballoc_query_range_fn)(
struct super_block *sb,
ext4_group_t agno,
--
2.31.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 5.4 2/3] ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow
2023-10-28 6:53 [PATCH 5.4 1/3] ext4: add two helper functions extent_logical_end() and pa_logical_end() Baokun Li
@ 2023-10-28 6:53 ` Baokun Li
2023-10-31 14:13 ` Patch "ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow" has been added to the 5.4-stable tree gregkh
2023-10-28 6:53 ` [PATCH 5.4 3/3] ext4: avoid overlapping preallocations due to overflow Baokun Li
2023-10-31 14:13 ` Patch "ext4: add two helper functions extent_logical_end() and pa_logical_end()" " gregkh
2 siblings, 1 reply; 6+ messages in thread
From: Baokun Li @ 2023-10-28 6:53 UTC (permalink / raw)
To: stable
Cc: gregkh, sashal, tytso, jack, ritesh.list, patches, yangerkun,
libaokun1, stable
commit bc056e7163ac7db945366de219745cf94f32a3e6 upstream.
When we calculate the end position of ext4_free_extent, this position may
be exactly where ext4_lblk_t (i.e. uint) overflows. For example, if
ac_g_ex.fe_logical is 4294965248 and ac_orig_goal_len is 2048, then the
computed end is 0x100000000, which is 0. If ac->ac_o_ex.fe_logical is not
the first case of adjusting the best extent, that is, new_bex_end > 0, the
following BUG_ON will be triggered:
=========================================================
kernel BUG at fs/ext4/mballoc.c:5116!
invalid opcode: 0000 [#1] PREEMPT SMP PTI
CPU: 3 PID: 673 Comm: xfs_io Tainted: G E 6.5.0-rc1+ #279
RIP: 0010:ext4_mb_new_inode_pa+0xc5/0x430
Call Trace:
<TASK>
ext4_mb_use_best_found+0x203/0x2f0
ext4_mb_try_best_found+0x163/0x240
ext4_mb_regular_allocator+0x158/0x1550
ext4_mb_new_blocks+0x86a/0xe10
ext4_ext_map_blocks+0xb0c/0x13a0
ext4_map_blocks+0x2cd/0x8f0
ext4_iomap_begin+0x27b/0x400
iomap_iter+0x222/0x3d0
__iomap_dio_rw+0x243/0xcb0
iomap_dio_rw+0x16/0x80
=========================================================
A simple reproducer demonstrating the problem:
mkfs.ext4 -F /dev/sda -b 4096 100M
mount /dev/sda /tmp/test
fallocate -l1M /tmp/test/tmp
fallocate -l10M /tmp/test/file
fallocate -i -o 1M -l16777203M /tmp/test/file
fsstress -d /tmp/test -l 0 -n 100000 -p 8 &
sleep 10 && killall -9 fsstress
rm -f /tmp/test/tmp
xfs_io -c "open -ad /tmp/test/file" -c "pwrite -S 0xff 0 8192"
We simply refactor the logic for adjusting the best extent by adding
a temporary ext4_free_extent ex and use extent_logical_end() to avoid
overflow, which also simplifies the code.
Cc: stable@kernel.org # 6.4
Fixes: 93cdf49f6eca ("ext4: Fix best extent lstart adjustment logic in ext4_mb_new_inode_pa()")
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/r/20230724121059.11834-3-libaokun1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Conflicts:
fs/ext4/mballoc.c
Signed-off-by: Baokun Li <libaokun1@huawei.com>
---
fs/ext4/mballoc.c | 31 ++++++++++++++-----------------
1 file changed, 14 insertions(+), 17 deletions(-)
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index e07420c6ddee..901302735f1c 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -3683,8 +3683,11 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
return -ENOMEM;
if (ac->ac_b_ex.fe_len < ac->ac_g_ex.fe_len) {
- int new_bex_start;
- int new_bex_end;
+ struct ext4_free_extent ex = {
+ .fe_logical = ac->ac_g_ex.fe_logical,
+ .fe_len = ac->ac_g_ex.fe_len,
+ };
+ loff_t orig_goal_end = extent_logical_end(sbi, &ex);
/* we can't allocate as much as normalizer wants.
* so, found space must get proper lstart
@@ -3703,29 +3706,23 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
* still cover original start
* 3. Else, keep the best ex at start of original request.
*/
- new_bex_end = ac->ac_g_ex.fe_logical +
- EXT4_C2B(sbi, ac->ac_g_ex.fe_len);
- new_bex_start = new_bex_end - EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
- if (ac->ac_o_ex.fe_logical >= new_bex_start)
- goto adjust_bex;
+ ex.fe_len = ac->ac_b_ex.fe_len;
- new_bex_start = ac->ac_g_ex.fe_logical;
- new_bex_end =
- new_bex_start + EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
- if (ac->ac_o_ex.fe_logical < new_bex_end)
+ ex.fe_logical = orig_goal_end - EXT4_C2B(sbi, ex.fe_len);
+ if (ac->ac_o_ex.fe_logical >= ex.fe_logical)
goto adjust_bex;
- new_bex_start = ac->ac_o_ex.fe_logical;
- new_bex_end =
- new_bex_start + EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
+ ex.fe_logical = ac->ac_g_ex.fe_logical;
+ if (ac->ac_o_ex.fe_logical < extent_logical_end(sbi, &ex))
+ goto adjust_bex;
+ ex.fe_logical = ac->ac_o_ex.fe_logical;
adjust_bex:
- ac->ac_b_ex.fe_logical = new_bex_start;
+ ac->ac_b_ex.fe_logical = ex.fe_logical;
BUG_ON(ac->ac_o_ex.fe_logical < ac->ac_b_ex.fe_logical);
BUG_ON(ac->ac_o_ex.fe_len > ac->ac_b_ex.fe_len);
- BUG_ON(new_bex_end > (ac->ac_g_ex.fe_logical +
- EXT4_C2B(sbi, ac->ac_g_ex.fe_len)));
+ BUG_ON(extent_logical_end(sbi, &ex) > orig_goal_end);
}
/* preallocation can change ac_b_ex, thus we store actually
--
2.31.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 5.4 3/3] ext4: avoid overlapping preallocations due to overflow
2023-10-28 6:53 [PATCH 5.4 1/3] ext4: add two helper functions extent_logical_end() and pa_logical_end() Baokun Li
2023-10-28 6:53 ` [PATCH 5.4 2/3] ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow Baokun Li
@ 2023-10-28 6:53 ` Baokun Li
2023-10-31 14:13 ` Patch "ext4: avoid overlapping preallocations due to overflow" has been added to the 5.4-stable tree gregkh
2023-10-31 14:13 ` Patch "ext4: add two helper functions extent_logical_end() and pa_logical_end()" " gregkh
2 siblings, 1 reply; 6+ messages in thread
From: Baokun Li @ 2023-10-28 6:53 UTC (permalink / raw)
To: stable
Cc: gregkh, sashal, tytso, jack, ritesh.list, patches, yangerkun,
libaokun1
commit bedc5d34632c21b5adb8ca7143d4c1f794507e4c upstream.
Let's say we want to allocate 2 blocks starting from 4294966386, after
predicting the file size, start is aligned to 4294965248, len is changed
to 2048, then end = start + size = 0x100000000. Since end is of
type ext4_lblk_t, i.e. uint, end is truncated to 0.
This causes (pa->pa_lstart >= end) to always hold when checking if the
current extent to be allocated crosses already preallocated blocks, so the
resulting ac_g_ex may cross already preallocated blocks. Hence we convert
the end type to loff_t and use pa_logical_end() to avoid overflow.
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/r/20230724121059.11834-4-libaokun1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Conflicts:
fs/ext4/mballoc.c
Signed-off-by: Baokun Li <libaokun1@huawei.com>
---
fs/ext4/mballoc.c | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 901302735f1c..a0ec645bd7ed 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -3094,8 +3094,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb);
struct ext4_super_block *es = sbi->s_es;
int bsbits, max;
- ext4_lblk_t end;
- loff_t size, start_off;
+ loff_t size, start_off, end;
loff_t orig_size __maybe_unused;
ext4_lblk_t start;
struct ext4_inode_info *ei = EXT4_I(ac->ac_inode);
@@ -3203,7 +3202,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
/* check we don't cross already preallocated blocks */
rcu_read_lock();
list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) {
- ext4_lblk_t pa_end;
+ loff_t pa_end;
if (pa->pa_deleted)
continue;
@@ -3213,8 +3212,7 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
continue;
}
- pa_end = pa->pa_lstart + EXT4_C2B(EXT4_SB(ac->ac_sb),
- pa->pa_len);
+ pa_end = pa_logical_end(EXT4_SB(ac->ac_sb), pa);
/* PA must not overlap original request */
BUG_ON(!(ac->ac_o_ex.fe_logical >= pa_end ||
@@ -3243,12 +3241,11 @@ ext4_mb_normalize_request(struct ext4_allocation_context *ac,
/* XXX: extra loop to check we really don't overlap preallocations */
rcu_read_lock();
list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) {
- ext4_lblk_t pa_end;
+ loff_t pa_end;
spin_lock(&pa->pa_lock);
if (pa->pa_deleted == 0) {
- pa_end = pa->pa_lstart + EXT4_C2B(EXT4_SB(ac->ac_sb),
- pa->pa_len);
+ pa_end = pa_logical_end(EXT4_SB(ac->ac_sb), pa);
BUG_ON(!(start >= pa_end || end <= pa->pa_lstart));
}
spin_unlock(&pa->pa_lock);
--
2.31.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Patch "ext4: add two helper functions extent_logical_end() and pa_logical_end()" has been added to the 5.4-stable tree
2023-10-28 6:53 [PATCH 5.4 1/3] ext4: add two helper functions extent_logical_end() and pa_logical_end() Baokun Li
2023-10-28 6:53 ` [PATCH 5.4 2/3] ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow Baokun Li
2023-10-28 6:53 ` [PATCH 5.4 3/3] ext4: avoid overlapping preallocations due to overflow Baokun Li
@ 2023-10-31 14:13 ` gregkh
2 siblings, 0 replies; 6+ messages in thread
From: gregkh @ 2023-10-31 14:13 UTC (permalink / raw)
To: gregkh, jack, libaokun1, patches, ritesh.list, sashal, tytso,
yangerkun
Cc: stable-commits
This is a note to let you know that I've just added the patch titled
ext4: add two helper functions extent_logical_end() and pa_logical_end()
to the 5.4-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
ext4-add-two-helper-functions-extent_logical_end-and-pa_logical_end.patch
and it can be found in the queue-5.4 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@vger.kernel.org> know about it.
From stable-owner@vger.kernel.org Sat Oct 28 08:49:06 2023
From: Baokun Li <libaokun1@huawei.com>
Date: Sat, 28 Oct 2023 14:53:29 +0800
Subject: ext4: add two helper functions extent_logical_end() and pa_logical_end()
To: <stable@vger.kernel.org>
Cc: <gregkh@linuxfoundation.org>, <sashal@kernel.org>, <tytso@mit.edu>, <jack@suse.cz>, <ritesh.list@gmail.com>, <patches@lists.linux.dev>, <yangerkun@huawei.com>, <libaokun1@huawei.com>
Message-ID: <20231028065331.948339-1-libaokun1@huawei.com>
From: Baokun Li <libaokun1@huawei.com>
commit 43bbddc067883d94de7a43d5756a295439fbe37d upstream.
When we use lstart + len to calculate the end of free extent or prealloc
space, it may exceed the maximum value of 4294967295(0xffffffff) supported
by ext4_lblk_t and cause overflow, which may lead to various problems.
Therefore, we add two helper functions, extent_logical_end() and
pa_logical_end(), to limit the type of end to loff_t, and also convert
lstart to loff_t for calculation to avoid overflow.
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/r/20230724121059.11834-2-libaokun1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
fs/ext4/mballoc.c | 7 +++----
fs/ext4/mballoc.h | 14 ++++++++++++++
2 files changed, 17 insertions(+), 4 deletions(-)
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -3124,7 +3124,7 @@ ext4_mb_normalize_request(struct ext4_al
/* first, let's learn actual file size
* given current request is allocated */
- size = ac->ac_o_ex.fe_logical + EXT4_C2B(sbi, ac->ac_o_ex.fe_len);
+ size = extent_logical_end(sbi, &ac->ac_o_ex);
size = size << bsbits;
if (size < i_size_read(ac->ac_inode))
size = i_size_read(ac->ac_inode);
@@ -3462,8 +3462,7 @@ ext4_mb_use_preallocated(struct ext4_all
/* all fields in this condition don't change,
* so we can skip locking for them */
if (ac->ac_o_ex.fe_logical < pa->pa_lstart ||
- ac->ac_o_ex.fe_logical >= (pa->pa_lstart +
- EXT4_C2B(sbi, pa->pa_len)))
+ ac->ac_o_ex.fe_logical >= pa_logical_end(sbi, pa))
continue;
/* non-extent files can't have physical blocks past 2^32 */
@@ -4234,7 +4233,7 @@ static void ext4_mb_group_or_file(struct
if (unlikely(ac->ac_flags & EXT4_MB_HINT_GOAL_ONLY))
return;
- size = ac->ac_o_ex.fe_logical + EXT4_C2B(sbi, ac->ac_o_ex.fe_len);
+ size = extent_logical_end(sbi, &ac->ac_o_ex);
isize = (i_size_read(ac->ac_inode) + ac->ac_sb->s_blocksize - 1)
>> bsbits;
--- a/fs/ext4/mballoc.h
+++ b/fs/ext4/mballoc.h
@@ -199,6 +199,20 @@ static inline ext4_fsblk_t ext4_grp_offs
(fex->fe_start << EXT4_SB(sb)->s_cluster_bits);
}
+static inline loff_t extent_logical_end(struct ext4_sb_info *sbi,
+ struct ext4_free_extent *fex)
+{
+ /* Use loff_t to avoid end exceeding ext4_lblk_t max. */
+ return (loff_t)fex->fe_logical + EXT4_C2B(sbi, fex->fe_len);
+}
+
+static inline loff_t pa_logical_end(struct ext4_sb_info *sbi,
+ struct ext4_prealloc_space *pa)
+{
+ /* Use loff_t to avoid end exceeding ext4_lblk_t max. */
+ return (loff_t)pa->pa_lstart + EXT4_C2B(sbi, pa->pa_len);
+}
+
typedef int (*ext4_mballoc_query_range_fn)(
struct super_block *sb,
ext4_group_t agno,
Patches currently in stable-queue which might be from stable-owner@vger.kernel.org are
queue-5.4/ext4-avoid-overlapping-preallocations-due-to-overflow.patch
queue-5.4/ext4-fix-bug-in-ext4_mb_new_inode_pa-due-to-overflow.patch
queue-5.4/ext4-add-two-helper-functions-extent_logical_end-and-pa_logical_end.patch
^ permalink raw reply [flat|nested] 6+ messages in thread
* Patch "ext4: avoid overlapping preallocations due to overflow" has been added to the 5.4-stable tree
2023-10-28 6:53 ` [PATCH 5.4 3/3] ext4: avoid overlapping preallocations due to overflow Baokun Li
@ 2023-10-31 14:13 ` gregkh
0 siblings, 0 replies; 6+ messages in thread
From: gregkh @ 2023-10-31 14:13 UTC (permalink / raw)
To: gregkh, jack, libaokun1, patches, ritesh.list, sashal, tytso,
yangerkun
Cc: stable-commits
This is a note to let you know that I've just added the patch titled
ext4: avoid overlapping preallocations due to overflow
to the 5.4-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
ext4-avoid-overlapping-preallocations-due-to-overflow.patch
and it can be found in the queue-5.4 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@vger.kernel.org> know about it.
From stable-owner@vger.kernel.org Sat Oct 28 08:49:06 2023
From: Baokun Li <libaokun1@huawei.com>
Date: Sat, 28 Oct 2023 14:53:31 +0800
Subject: ext4: avoid overlapping preallocations due to overflow
To: <stable@vger.kernel.org>
Cc: <gregkh@linuxfoundation.org>, <sashal@kernel.org>, <tytso@mit.edu>, <jack@suse.cz>, <ritesh.list@gmail.com>, <patches@lists.linux.dev>, <yangerkun@huawei.com>, <libaokun1@huawei.com>
Message-ID: <20231028065331.948339-3-libaokun1@huawei.com>
From: Baokun Li <libaokun1@huawei.com>
commit bedc5d34632c21b5adb8ca7143d4c1f794507e4c upstream.
Let's say we want to allocate 2 blocks starting from 4294966386, after
predicting the file size, start is aligned to 4294965248, len is changed
to 2048, then end = start + size = 0x100000000. Since end is of
type ext4_lblk_t, i.e. uint, end is truncated to 0.
This causes (pa->pa_lstart >= end) to always hold when checking if the
current extent to be allocated crosses already preallocated blocks, so the
resulting ac_g_ex may cross already preallocated blocks. Hence we convert
the end type to loff_t and use pa_logical_end() to avoid overflow.
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/r/20230724121059.11834-4-libaokun1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
fs/ext4/mballoc.c | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -3094,8 +3094,7 @@ ext4_mb_normalize_request(struct ext4_al
struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb);
struct ext4_super_block *es = sbi->s_es;
int bsbits, max;
- ext4_lblk_t end;
- loff_t size, start_off;
+ loff_t size, start_off, end;
loff_t orig_size __maybe_unused;
ext4_lblk_t start;
struct ext4_inode_info *ei = EXT4_I(ac->ac_inode);
@@ -3203,7 +3202,7 @@ ext4_mb_normalize_request(struct ext4_al
/* check we don't cross already preallocated blocks */
rcu_read_lock();
list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) {
- ext4_lblk_t pa_end;
+ loff_t pa_end;
if (pa->pa_deleted)
continue;
@@ -3213,8 +3212,7 @@ ext4_mb_normalize_request(struct ext4_al
continue;
}
- pa_end = pa->pa_lstart + EXT4_C2B(EXT4_SB(ac->ac_sb),
- pa->pa_len);
+ pa_end = pa_logical_end(EXT4_SB(ac->ac_sb), pa);
/* PA must not overlap original request */
BUG_ON(!(ac->ac_o_ex.fe_logical >= pa_end ||
@@ -3243,12 +3241,11 @@ ext4_mb_normalize_request(struct ext4_al
/* XXX: extra loop to check we really don't overlap preallocations */
rcu_read_lock();
list_for_each_entry_rcu(pa, &ei->i_prealloc_list, pa_inode_list) {
- ext4_lblk_t pa_end;
+ loff_t pa_end;
spin_lock(&pa->pa_lock);
if (pa->pa_deleted == 0) {
- pa_end = pa->pa_lstart + EXT4_C2B(EXT4_SB(ac->ac_sb),
- pa->pa_len);
+ pa_end = pa_logical_end(EXT4_SB(ac->ac_sb), pa);
BUG_ON(!(start >= pa_end || end <= pa->pa_lstart));
}
spin_unlock(&pa->pa_lock);
Patches currently in stable-queue which might be from stable-owner@vger.kernel.org are
queue-5.4/ext4-avoid-overlapping-preallocations-due-to-overflow.patch
queue-5.4/ext4-fix-bug-in-ext4_mb_new_inode_pa-due-to-overflow.patch
queue-5.4/ext4-add-two-helper-functions-extent_logical_end-and-pa_logical_end.patch
^ permalink raw reply [flat|nested] 6+ messages in thread
* Patch "ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow" has been added to the 5.4-stable tree
2023-10-28 6:53 ` [PATCH 5.4 2/3] ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow Baokun Li
@ 2023-10-31 14:13 ` gregkh
0 siblings, 0 replies; 6+ messages in thread
From: gregkh @ 2023-10-31 14:13 UTC (permalink / raw)
To: gregkh, jack, libaokun1, patches, ritesh.list, sashal, stable,
tytso, yangerkun
Cc: stable-commits
This is a note to let you know that I've just added the patch titled
ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow
to the 5.4-stable tree which can be found at:
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
The filename of the patch is:
ext4-fix-bug-in-ext4_mb_new_inode_pa-due-to-overflow.patch
and it can be found in the queue-5.4 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree,
please let <stable@vger.kernel.org> know about it.
From stable-owner@vger.kernel.org Sat Oct 28 08:49:06 2023
From: Baokun Li <libaokun1@huawei.com>
Date: Sat, 28 Oct 2023 14:53:30 +0800
Subject: ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow
To: <stable@vger.kernel.org>
Cc: <gregkh@linuxfoundation.org>, <sashal@kernel.org>, <tytso@mit.edu>, <jack@suse.cz>, <ritesh.list@gmail.com>, <patches@lists.linux.dev>, <yangerkun@huawei.com>, <libaokun1@huawei.com>, <stable@kernel.org>
Message-ID: <20231028065331.948339-2-libaokun1@huawei.com>
From: Baokun Li <libaokun1@huawei.com>
commit bc056e7163ac7db945366de219745cf94f32a3e6 upstream.
When we calculate the end position of ext4_free_extent, this position may
be exactly where ext4_lblk_t (i.e. uint) overflows. For example, if
ac_g_ex.fe_logical is 4294965248 and ac_orig_goal_len is 2048, then the
computed end is 0x100000000, which is 0. If ac->ac_o_ex.fe_logical is not
the first case of adjusting the best extent, that is, new_bex_end > 0, the
following BUG_ON will be triggered:
=========================================================
kernel BUG at fs/ext4/mballoc.c:5116!
invalid opcode: 0000 [#1] PREEMPT SMP PTI
CPU: 3 PID: 673 Comm: xfs_io Tainted: G E 6.5.0-rc1+ #279
RIP: 0010:ext4_mb_new_inode_pa+0xc5/0x430
Call Trace:
<TASK>
ext4_mb_use_best_found+0x203/0x2f0
ext4_mb_try_best_found+0x163/0x240
ext4_mb_regular_allocator+0x158/0x1550
ext4_mb_new_blocks+0x86a/0xe10
ext4_ext_map_blocks+0xb0c/0x13a0
ext4_map_blocks+0x2cd/0x8f0
ext4_iomap_begin+0x27b/0x400
iomap_iter+0x222/0x3d0
__iomap_dio_rw+0x243/0xcb0
iomap_dio_rw+0x16/0x80
=========================================================
A simple reproducer demonstrating the problem:
mkfs.ext4 -F /dev/sda -b 4096 100M
mount /dev/sda /tmp/test
fallocate -l1M /tmp/test/tmp
fallocate -l10M /tmp/test/file
fallocate -i -o 1M -l16777203M /tmp/test/file
fsstress -d /tmp/test -l 0 -n 100000 -p 8 &
sleep 10 && killall -9 fsstress
rm -f /tmp/test/tmp
xfs_io -c "open -ad /tmp/test/file" -c "pwrite -S 0xff 0 8192"
We simply refactor the logic for adjusting the best extent by adding
a temporary ext4_free_extent ex and use extent_logical_end() to avoid
overflow, which also simplifies the code.
Cc: stable@kernel.org # 6.4
Fixes: 93cdf49f6eca ("ext4: Fix best extent lstart adjustment logic in ext4_mb_new_inode_pa()")
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://lore.kernel.org/r/20230724121059.11834-3-libaokun1@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
fs/ext4/mballoc.c | 31 ++++++++++++++-----------------
1 file changed, 14 insertions(+), 17 deletions(-)
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -3680,8 +3680,11 @@ ext4_mb_new_inode_pa(struct ext4_allocat
return -ENOMEM;
if (ac->ac_b_ex.fe_len < ac->ac_g_ex.fe_len) {
- int new_bex_start;
- int new_bex_end;
+ struct ext4_free_extent ex = {
+ .fe_logical = ac->ac_g_ex.fe_logical,
+ .fe_len = ac->ac_g_ex.fe_len,
+ };
+ loff_t orig_goal_end = extent_logical_end(sbi, &ex);
/* we can't allocate as much as normalizer wants.
* so, found space must get proper lstart
@@ -3700,29 +3703,23 @@ ext4_mb_new_inode_pa(struct ext4_allocat
* still cover original start
* 3. Else, keep the best ex at start of original request.
*/
- new_bex_end = ac->ac_g_ex.fe_logical +
- EXT4_C2B(sbi, ac->ac_g_ex.fe_len);
- new_bex_start = new_bex_end - EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
- if (ac->ac_o_ex.fe_logical >= new_bex_start)
- goto adjust_bex;
+ ex.fe_len = ac->ac_b_ex.fe_len;
- new_bex_start = ac->ac_g_ex.fe_logical;
- new_bex_end =
- new_bex_start + EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
- if (ac->ac_o_ex.fe_logical < new_bex_end)
+ ex.fe_logical = orig_goal_end - EXT4_C2B(sbi, ex.fe_len);
+ if (ac->ac_o_ex.fe_logical >= ex.fe_logical)
goto adjust_bex;
- new_bex_start = ac->ac_o_ex.fe_logical;
- new_bex_end =
- new_bex_start + EXT4_C2B(sbi, ac->ac_b_ex.fe_len);
+ ex.fe_logical = ac->ac_g_ex.fe_logical;
+ if (ac->ac_o_ex.fe_logical < extent_logical_end(sbi, &ex))
+ goto adjust_bex;
+ ex.fe_logical = ac->ac_o_ex.fe_logical;
adjust_bex:
- ac->ac_b_ex.fe_logical = new_bex_start;
+ ac->ac_b_ex.fe_logical = ex.fe_logical;
BUG_ON(ac->ac_o_ex.fe_logical < ac->ac_b_ex.fe_logical);
BUG_ON(ac->ac_o_ex.fe_len > ac->ac_b_ex.fe_len);
- BUG_ON(new_bex_end > (ac->ac_g_ex.fe_logical +
- EXT4_C2B(sbi, ac->ac_g_ex.fe_len)));
+ BUG_ON(extent_logical_end(sbi, &ex) > orig_goal_end);
}
/* preallocation can change ac_b_ex, thus we store actually
Patches currently in stable-queue which might be from stable-owner@vger.kernel.org are
queue-5.4/ext4-avoid-overlapping-preallocations-due-to-overflow.patch
queue-5.4/ext4-fix-bug-in-ext4_mb_new_inode_pa-due-to-overflow.patch
queue-5.4/ext4-add-two-helper-functions-extent_logical_end-and-pa_logical_end.patch
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2023-10-31 14:13 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-10-28 6:53 [PATCH 5.4 1/3] ext4: add two helper functions extent_logical_end() and pa_logical_end() Baokun Li
2023-10-28 6:53 ` [PATCH 5.4 2/3] ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow Baokun Li
2023-10-31 14:13 ` Patch "ext4: fix BUG in ext4_mb_new_inode_pa() due to overflow" has been added to the 5.4-stable tree gregkh
2023-10-28 6:53 ` [PATCH 5.4 3/3] ext4: avoid overlapping preallocations due to overflow Baokun Li
2023-10-31 14:13 ` Patch "ext4: avoid overlapping preallocations due to overflow" has been added to the 5.4-stable tree gregkh
2023-10-31 14:13 ` Patch "ext4: add two helper functions extent_logical_end() and pa_logical_end()" " gregkh
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.