[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5c4c63ff-2191-c57b-50b9-20a7b9a60d86@oracle.com>
Date: Tue, 18 Jul 2023 11:17:01 -0500
From: Dave Kleikamp <dave.kleikamp@...cle.com>
To: Andrew Kanner <andrew.kanner@...il.com>,
jfs-discussion@...ts.sourceforge.net, linux-kernel@...r.kernel.org
Cc: linux-kernel-mentees@...ts.linuxfoundation.org
Subject: Re: [PATCH 1/1] fs/jfs: prevent double-free in dbUnmount() after
failed jfs_remount()
On 7/18/23 10:39AM, Andrew Kanner wrote:
> On Sat, Jul 01, 2023 at 05:05:42PM +0300, Andrew Kanner wrote:
>> Syzkaller reported the following issue:
>> ==================================================================
>> BUG: KASAN: double-free in slab_free mm/slub.c:3787 [inline]
>> BUG: KASAN: double-free in __kmem_cache_free+0x71/0x110 mm/slub.c:3800
>> Free of addr ffff888086408000 by task syz-executor.4/12750
>> [...]
>> Call Trace:
>> <TASK>
>> [...]
>> kasan_report_invalid_free+0xac/0xd0 mm/kasan/report.c:482
>> ____kasan_slab_free+0xfb/0x120
>> kasan_slab_free include/linux/kasan.h:177 [inline]
>> slab_free_hook mm/slub.c:1781 [inline]
>> slab_free_freelist_hook+0x12e/0x1a0 mm/slub.c:1807
>> slab_free mm/slub.c:3787 [inline]
>> __kmem_cache_free+0x71/0x110 mm/slub.c:3800
>> dbUnmount+0xf4/0x110 fs/jfs/jfs_dmap.c:264
>> jfs_umount+0x248/0x3b0 fs/jfs/jfs_umount.c:87
>> jfs_put_super+0x86/0x190 fs/jfs/super.c:194
>> generic_shutdown_super+0x130/0x310 fs/super.c:492
>> kill_block_super+0x79/0xd0 fs/super.c:1386
>> deactivate_locked_super+0xa7/0xf0 fs/super.c:332
>> cleanup_mnt+0x494/0x520 fs/namespace.c:1291
>> task_work_run+0x243/0x300 kernel/task_work.c:179
>> resume_user_mode_work include/linux/resume_user_mode.h:49 [inline]
>> exit_to_user_mode_loop+0x124/0x150 kernel/entry/common.c:171
>> exit_to_user_mode_prepare+0xb2/0x140 kernel/entry/common.c:203
>> __syscall_exit_to_user_mode_work kernel/entry/common.c:285 [inline]
>> syscall_exit_to_user_mode+0x26/0x60 kernel/entry/common.c:296
>> do_syscall_64+0x49/0xb0 arch/x86/entry/common.c:86
>> entry_SYSCALL_64_after_hwframe+0x63/0xcd
>> [...]
>> </TASK>
>>
>> Allocated by task 13352:
>> kasan_save_stack mm/kasan/common.c:45 [inline]
>> kasan_set_track+0x3d/0x60 mm/kasan/common.c:52
>> ____kasan_kmalloc mm/kasan/common.c:371 [inline]
>> __kasan_kmalloc+0x97/0xb0 mm/kasan/common.c:380
>> kmalloc include/linux/slab.h:580 [inline]
>> dbMount+0x54/0x980 fs/jfs/jfs_dmap.c:164
>> jfs_mount+0x1dd/0x830 fs/jfs/jfs_mount.c:121
>> jfs_fill_super+0x590/0xc50 fs/jfs/super.c:556
>> mount_bdev+0x26c/0x3a0 fs/super.c:1359
>> legacy_get_tree+0xea/0x180 fs/fs_context.c:610
>> vfs_get_tree+0x88/0x270 fs/super.c:1489
>> do_new_mount+0x289/0xad0 fs/namespace.c:3145
>> do_mount fs/namespace.c:3488 [inline]
>> __do_sys_mount fs/namespace.c:3697 [inline]
>> __se_sys_mount+0x2d3/0x3c0 fs/namespace.c:3674
>> do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>> do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
>> entry_SYSCALL_64_after_hwframe+0x63/0xcd
>>
>> Freed by task 13352:
>> kasan_save_stack mm/kasan/common.c:45 [inline]
>> kasan_set_track+0x3d/0x60 mm/kasan/common.c:52
>> kasan_save_free_info+0x27/0x40 mm/kasan/generic.c:518
>> ____kasan_slab_free+0xd6/0x120 mm/kasan/common.c:236
>> kasan_slab_free include/linux/kasan.h:177 [inline]
>> slab_free_hook mm/slub.c:1781 [inline]
>> slab_free_freelist_hook+0x12e/0x1a0 mm/slub.c:1807
>> slab_free mm/slub.c:3787 [inline]
>> __kmem_cache_free+0x71/0x110 mm/slub.c:3800
>> dbUnmount+0xf4/0x110 fs/jfs/jfs_dmap.c:264
>> jfs_mount_rw+0x545/0x740 fs/jfs/jfs_mount.c:247
>> jfs_remount+0x3db/0x710 fs/jfs/super.c:454
>> reconfigure_super+0x3bc/0x7b0 fs/super.c:935
>> vfs_fsconfig_locked fs/fsopen.c:254 [inline]
>> __do_sys_fsconfig fs/fsopen.c:439 [inline]
>> __se_sys_fsconfig+0xad5/0x1060 fs/fsopen.c:314
>> do_syscall_x64 arch/x86/entry/common.c:50 [inline]
>> do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
>> entry_SYSCALL_64_after_hwframe+0x63/0xcd
>> [...]
>>
>> JFS_SBI(ipbmap->i_sb)->bmap wasn't set to NULL after kfree() in
>> dbUnmount().
>>
>> Syzkaller uses faultinject to reproduce this KASAN double-free
>> warning. The issue is triggered if either diMount() or dbMount() fail
>> in jfs_remount(), since diUnmount() or dbUnmount() already happened in
>> such a case - they will do double-free on next execution: jfs_umount
>> or jfs_remount.
>>
>> Tested on both upstream and jfs-next by syzkaller.
>>
>> Reported-and-tested-by: syzbot+6a93efb725385bc4b2e9@...kaller.appspotmail.com
>> Closes: https://lore.kernel.org/all/000000000000471f2d05f1ce8bad@google.com/T/
>> Link: https://syzkaller.appspot.com/bug?extid=6a93efb725385bc4b2e9
>> Signed-off-by: Andrew Kanner <andrew.kanner@...il.com>
>> ---
>> fs/jfs/jfs_dmap.c | 1 +
>> 1 file changed, 1 insertion(+)
>>
>> diff --git a/fs/jfs/jfs_dmap.c b/fs/jfs/jfs_dmap.c
>> index a14a0f18a4c4..88afd108c2dd 100644
>> --- a/fs/jfs/jfs_dmap.c
>> +++ b/fs/jfs/jfs_dmap.c
>> @@ -269,6 +269,7 @@ int dbUnmount(struct inode *ipbmap, int mounterror)
>>
>> /* free the memory for the in-memory bmap. */
>> kfree(bmp);
>> + JFS_SBI(ipbmap->i_sb)->bmap = NULL;
>>
>> return (0);
>> }
>> --
>> 2.39.3
>>
>
> Hi, Dave.
> Are there any objections?
> Or maybe suggestions on how to test both my patch and the patch from
> Liu Shixin (some corner cases)?
Sorry. I got behind. I'll try to catch up on these today (or very soon
if something more urgent pops up).
Shaggy
Powered by blists - more mailing lists