| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CAEAjamu1FRhz6StCe_55XY5s389ZP_xmCF69k987En+1z53=eg@mail.gmail.com>
Date: Tue, 18 Sep 2018 13:40:49 -0400
From: Kyungtae Kim <kt0755@...il.com>
To: jikos@...nel.org, axboe@...nel.dk
Cc: DaeRyong Jeong <threeearcat@...il.com>,
Byoungyoung Lee <lifeasageek@...il.com>,
linux-block@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: KASAN: use-after-free Read in raw_cmd_done
We report the crash:
KASAN: use-after-free in raw_cmd_done
This crash was found in v4.19-rc2 using Alexkidd-fuzzer (a modified
version of Syzkaller).
Our analysis shows that the race arose between a kernel thread (kworker) and
the invoked ioctl syscall.
kernel config: https://kt0755.github.io/etc/config-4.19-rc2
repro: https://kt0755.github.io/etc/repro.92596.c
Analysis:
When there is a race between "raw_cmd_done"
(driver/block/floppy.c:3054) and "raw_cmd_free" (driver/block/floppy.c:3115),
a pointer variable "raw_cmd" in raw_cmd_done may point to the
freed memory area. When raw_cmd_done dereferences this
pointer, use-after-free occurs.
<Thread-0 (raw_cmd_ioctl)> <Thread-1 (floppy_shutdown)>
1. raw_cmd_copyin(&my_raw_cmd)
// allocate my_raw_cmd
2. raw_cmd = my_raw_cmd
3. cont = &raw_cmd_cont
4. wait_til_done(floppy_start)
-------------------------------------------------------------------------------------------
5. cont->done(0)
//
call raw_cmd_done()
------------------------------------------------------------------------------------------
6. raw_cmd_free(&my_raw_cmd)
// kfree(my_raw_cmd)
------------------------------------------------------------------------------------------
7. rmd_cmd->flag
// Use-after-free
# Call Sequence
<Thread-0>
fd_ioctl
fd_locked_ioctl
raw_cmd_ioctl
raw_cmd_free
<Thread-1>
worker_thread
process_one_work
floppy_shutdown
raw_cmd_done
We observed that raw_cmd_ioctl (say T0) is invoked by fd_locked_ioctl
(FDRAWCMD).
It first allocates "raw_cmd" (driver/block/floppy.c:3212), and then
wait until floppy_start is done. In the meantime, floppy_shutdown scheduled
by floppy_start is called (say T1) and invokes raw_cmd_done.
Unfortunately, before raw_cmd_done accesses raw_cmd, T0 waked up and
deallocated raw_cmd in raw_cmd_free (driver/block/floppy.c:3225).
So the access to raw_cmd (by T1) fails.
We suspect that the insufficient locking between raw_cmd_done and
raw_cnd_free, is the cause of the crash.
Crash log
==================================================================
BUG: KASAN: use-after-free in raw_cmd_done+0x58a/0x5e0
drivers/block/floppy.c:3054
Read of size 4 at addr ffff880119b35cc0 by task kworker/u4:0/7
CPU: 0 PID: 7 Comm: kworker/u4:0 Not tainted 4.19.0-rc2 #2
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
Workqueue: floppy floppy_shutdown
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0xd2/0x148 lib/dump_stack.c:113
print_address_description+0x7f/0x2c0 mm/kasan/report.c:256
kasan_report_error mm/kasan/report.c:354 [inline]
kasan_report+0x271/0x3a0 mm/kasan/report.c:412
__asan_report_load4_noabort+0x19/0x20 mm/kasan/report.c:432
raw_cmd_done+0x58a/0x5e0 drivers/block/floppy.c:3054
floppy_shutdown+0x127/0x1f0 drivers/block/floppy.c:1851
process_one_work+0x923/0x1690 kernel/workqueue.c:2153
worker_thread+0x8f/0xcf0 kernel/workqueue.c:2296
kthread+0x3a3/0x470 kernel/kthread.c:246
ret_from_fork+0x3a/0x50 arch/x86/entry/entry_64.S:413
Allocated by task 16211:
save_stack+0x43/0xd0 mm/kasan/kasan.c:448
set_track mm/kasan/kasan.c:460 [inline]
kasan_kmalloc+0xc5/0xf0 mm/kasan/kasan.c:553
kmem_cache_alloc_trace+0x14a/0x2b0 mm/slub.c:2733
kmalloc include/linux/slab.h:513 [inline]
raw_cmd_copyin drivers/block/floppy.c:3130 [inline]
raw_cmd_ioctl drivers/block/floppy.c:3206 [inline]
fd_locked_ioctl+0x608/0x21b0 drivers/block/floppy.c:3544
fd_ioctl+0x40/0x60 drivers/block/floppy.c:3571
__blkdev_driver_ioctl block/ioctl.c:303 [inline]
blkdev_ioctl+0xb48/0x1a50 block/ioctl.c:601
block_ioctl+0x105/0x150 fs/block_dev.c:1883
vfs_ioctl fs/ioctl.c:46 [inline]
do_vfs_ioctl+0x1c0/0x1150 fs/ioctl.c:687
ksys_ioctl+0x9e/0xb0 fs/ioctl.c:702
__do_sys_ioctl fs/ioctl.c:709 [inline]
__se_sys_ioctl fs/ioctl.c:707 [inline]
__x64_sys_ioctl+0x7e/0xc0 fs/ioctl.c:707
do_syscall_64+0xc4/0x4e0 arch/x86/entry/common.c:290
entry_SYSCALL_64_after_hwframe+0x49/0xbe
Freed by task 16211:
save_stack+0x43/0xd0 mm/kasan/kasan.c:448
set_track mm/kasan/kasan.c:460 [inline]
__kasan_slab_free+0x14d/0x1a0 mm/kasan/kasan.c:521
kasan_slab_free+0xe/0x10 mm/kasan/kasan.c:528
slab_free_hook mm/slub.c:1371 [inline]
slab_free_freelist_hook mm/slub.c:1398 [inline]
slab_free mm/slub.c:2953 [inline]
kfree+0x109/0x2f0 mm/slub.c:3906
raw_cmd_free+0x82/0x170 drivers/block/floppy.c:3115
raw_cmd_ioctl drivers/block/floppy.c:3225 [inline]
fd_locked_ioctl+0x1d37/0x21b0 drivers/block/floppy.c:3544
fd_ioctl+0x40/0x60 drivers/block/floppy.c:3571
__blkdev_driver_ioctl block/ioctl.c:303 [inline]
blkdev_ioctl+0xb48/0x1a50 block/ioctl.c:601
block_ioctl+0x105/0x150 fs/block_dev.c:1883
vfs_ioctl fs/ioctl.c:46 [inline]
do_vfs_ioctl+0x1c0/0x1150 fs/ioctl.c:687
ksys_ioctl+0x9e/0xb0 fs/ioctl.c:702
__do_sys_ioctl fs/ioctl.c:709 [inline]
__se_sys_ioctl fs/ioctl.c:707 [inline]
__x64_sys_ioctl+0x7e/0xc0 fs/ioctl.c:707
do_syscall_64+0xc4/0x4e0 arch/x86/entry/common.c:290
entry_SYSCALL_64_after_hwframe+0x49/0xbe
The buggy address belongs to the object at ffff880119b35cc0
which belongs to the cache kmalloc-128 of size 128
The buggy address is located 0 bytes inside of
128-byte region [ffff880119b35cc0, ffff880119b35d40)
The buggy address belongs to the page:
page:ffffea000466cd40 count:1 mapcount:0 mapping:ffff88011a403500 index:0x0
flags: 0x200000000000100(slab)
raw: 0200000000000100 ffffea0004042700 0000000700000007 ffff88011a403500
raw: 0000000000000000 0000000000150015 00000001ffffffff 0000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
ffff880119b35b80: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
ffff880119b35c00: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
>ffff880119b35c80: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
^
ffff880119b35d00: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
ffff880119b35d80: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
==================================================================
Powered by blists - more mailing lists