Hello,
syzbot found the following issue on:
HEAD commit: 325d0eab Merge branch 'akpm' (patches from Andrew)
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=13ec44d3900000
kernel config: https://syzkaller.appspot.com/x/.config?x=a1f3c5052e8097e9
dashboard link: https://syzkaller.appspot.com/bug?extid=128f4dd6e796c98b3760
compiler: gcc (GCC) 10.1.0-syz 20200507
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=101a0e9b900000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=17ae18d3900000
IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: [email protected]
BUG: memory leak
unreferenced object 0xffff88811a1d0a00 (size 512):
comm "syz-executor299", pid 6519, jiffies 4294943224 (age 15.090s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<000000009ea2686e>] kmalloc include/linux/slab.h:554 [inline]
[<000000009ea2686e>] kmalloc_array include/linux/slab.h:593 [inline]
[<000000009ea2686e>] kcalloc include/linux/slab.h:605 [inline]
[<000000009ea2686e>] udf_process_sequence+0x7e/0x1080 fs/udf/super.c:1676
[<00000000d9f60715>] udf_load_sequence fs/udf/super.c:1795 [inline]
[<00000000d9f60715>] udf_check_anchor_block+0xdc/0x1a0 fs/udf/super.c:1835
[<0000000067395456>] udf_scan_anchors+0x9e/0x240 fs/udf/super.c:1868
[<00000000966a1b37>] udf_find_anchor fs/udf/super.c:1925 [inline]
[<00000000966a1b37>] udf_load_vrs+0x1be/0x3b0 fs/udf/super.c:1990
[<0000000004bba192>] udf_fill_super+0x286/0x7a5 fs/udf/super.c:2183
[<00000000d48efb9b>] mount_bdev+0x1d3/0x210 fs/super.c:1417
[<00000000280e173c>] legacy_get_tree+0x26/0x70 fs/fs_context.c:592
[<0000000059fde270>] vfs_get_tree+0x28/0xe0 fs/super.c:1547
[<00000000904c79e7>] do_new_mount fs/namespace.c:2875 [inline]
[<00000000904c79e7>] path_mount+0x90e/0xda0 fs/namespace.c:3192
[<000000008e0f8bcd>] do_mount fs/namespace.c:3205 [inline]
[<000000008e0f8bcd>] __do_sys_mount fs/namespace.c:3413 [inline]
[<000000008e0f8bcd>] __se_sys_mount fs/namespace.c:3390 [inline]
[<000000008e0f8bcd>] __x64_sys_mount+0x140/0x190 fs/namespace.c:3390
[<00000000e981acac>] do_syscall_64+0x2d/0x70 arch/x86/entry/common.c:46
[<000000006322386a>] entry_SYSCALL_64_after_hwframe+0x44/0xa9
---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at [email protected].
syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
syzbot can test patches for this issue, for details see:
https://goo.gl/tpsmEJ#testing-patches
udf_process_sequence() is leaking memory. Free `data.part_descs_loc`
before returning.
Cc: [email protected]
Fixes: 7b78fd02fb19 ("udf: Fix handling of Partition Descriptors")
Reported-and-tested-by: [email protected]
Link: https://syzkaller.appspot.com/bug?id=c5ec4e6f5d818f3c4afd4d59342468eec08a38da
Signed-off-by: Peilin Ye <[email protected]>
---
fs/udf/super.c | 20 +++++++++++++-------
1 file changed, 13 insertions(+), 7 deletions(-)
diff --git a/fs/udf/super.c b/fs/udf/super.c
index 1c42f544096d..b0d862ab3024 100644
--- a/fs/udf/super.c
+++ b/fs/udf/super.c
@@ -1698,7 +1698,8 @@ static noinline int udf_process_sequence(
"Pointers (max %u supported)\n",
UDF_MAX_TD_NESTING);
brelse(bh);
- return -EIO;
+ ret = -EIO;
+ goto out;
}
vdp = (struct volDescPtr *)bh->b_data;
@@ -1718,7 +1719,8 @@ static noinline int udf_process_sequence(
curr = get_volume_descriptor_record(ident, bh, &data);
if (IS_ERR(curr)) {
brelse(bh);
- return PTR_ERR(curr);
+ ret = PTR_ERR(curr);
+ goto out;
}
/* Descriptor we don't care about? */
if (!curr)
@@ -1740,28 +1742,32 @@ static noinline int udf_process_sequence(
*/
if (!data.vds[VDS_POS_PRIMARY_VOL_DESC].block) {
udf_err(sb, "Primary Volume Descriptor not found!\n");
- return -EAGAIN;
+ ret = -EAGAIN;
+ goto out;
}
ret = udf_load_pvoldesc(sb, data.vds[VDS_POS_PRIMARY_VOL_DESC].block);
if (ret < 0)
- return ret;
+ goto out;
if (data.vds[VDS_POS_LOGICAL_VOL_DESC].block) {
ret = udf_load_logicalvol(sb,
data.vds[VDS_POS_LOGICAL_VOL_DESC].block,
fileset);
if (ret < 0)
- return ret;
+ goto out;
}
/* Now handle prevailing Partition Descriptors */
for (i = 0; i < data.num_part_descs; i++) {
ret = udf_load_partdesc(sb, data.part_descs_loc[i].rec.block);
if (ret < 0)
- return ret;
+ goto out;
}
- return 0;
+ ret = 0;
+out:
+ kfree(data.part_descs_loc);
+ return ret;
}
/*
--
2.25.1
On Tue 22-09-20 11:45:31, Peilin Ye wrote:
> udf_process_sequence() is leaking memory. Free `data.part_descs_loc`
> before returning.
>
> Cc: [email protected]
> Fixes: 7b78fd02fb19 ("udf: Fix handling of Partition Descriptors")
> Reported-and-tested-by: [email protected]
> Link: https://syzkaller.appspot.com/bug?id=c5ec4e6f5d818f3c4afd4d59342468eec08a38da
> Signed-off-by: Peilin Ye <[email protected]>
Thanks for the patch but I've just yesterday written exactly the same patch
and merged it to my tree...
Honza
> ---
> fs/udf/super.c | 20 +++++++++++++-------
> 1 file changed, 13 insertions(+), 7 deletions(-)
>
> diff --git a/fs/udf/super.c b/fs/udf/super.c
> index 1c42f544096d..b0d862ab3024 100644
> --- a/fs/udf/super.c
> +++ b/fs/udf/super.c
> @@ -1698,7 +1698,8 @@ static noinline int udf_process_sequence(
> "Pointers (max %u supported)\n",
> UDF_MAX_TD_NESTING);
> brelse(bh);
> - return -EIO;
> + ret = -EIO;
> + goto out;
> }
>
> vdp = (struct volDescPtr *)bh->b_data;
> @@ -1718,7 +1719,8 @@ static noinline int udf_process_sequence(
> curr = get_volume_descriptor_record(ident, bh, &data);
> if (IS_ERR(curr)) {
> brelse(bh);
> - return PTR_ERR(curr);
> + ret = PTR_ERR(curr);
> + goto out;
> }
> /* Descriptor we don't care about? */
> if (!curr)
> @@ -1740,28 +1742,32 @@ static noinline int udf_process_sequence(
> */
> if (!data.vds[VDS_POS_PRIMARY_VOL_DESC].block) {
> udf_err(sb, "Primary Volume Descriptor not found!\n");
> - return -EAGAIN;
> + ret = -EAGAIN;
> + goto out;
> }
> ret = udf_load_pvoldesc(sb, data.vds[VDS_POS_PRIMARY_VOL_DESC].block);
> if (ret < 0)
> - return ret;
> + goto out;
>
> if (data.vds[VDS_POS_LOGICAL_VOL_DESC].block) {
> ret = udf_load_logicalvol(sb,
> data.vds[VDS_POS_LOGICAL_VOL_DESC].block,
> fileset);
> if (ret < 0)
> - return ret;
> + goto out;
> }
>
> /* Now handle prevailing Partition Descriptors */
> for (i = 0; i < data.num_part_descs; i++) {
> ret = udf_load_partdesc(sb, data.part_descs_loc[i].rec.block);
> if (ret < 0)
> - return ret;
> + goto out;
> }
>
> - return 0;
> + ret = 0;
> +out:
> + kfree(data.part_descs_loc);
> + return ret;
> }
>
> /*
> --
> 2.25.1
>
--
Jan Kara <[email protected]>
SUSE Labs, CR
Hi,
On Wed, Sep 23, 2020 at 12:04:05PM +0200, Jan Kara wrote:
> On Tue 22-09-20 11:45:31, Peilin Ye wrote:
> > udf_process_sequence() is leaking memory. Free `data.part_descs_loc`
> > before returning.
> >
> > Cc: [email protected]
> > Fixes: 7b78fd02fb19 ("udf: Fix handling of Partition Descriptors")
> > Reported-and-tested-by: [email protected]
> > Link: https://syzkaller.appspot.com/bug?id=c5ec4e6f5d818f3c4afd4d59342468eec08a38da
> > Signed-off-by: Peilin Ye <[email protected]>
>
> Thanks for the patch but I've just yesterday written exactly the same patch
> and merged it to my tree...
Ah, no worries, happy to see the bug gets fixed!
Peilin Ye