Re: Potential use-after-free in shrink_page_list

From: Jan Kara
Date: Tue Oct 13 2015 - 04:00:45 EST

Hello,

On Wed 07-10-15 18:43:30, Andrey Konovalov wrote:
> While fuzzing the kernel (4.3-rc4) with KASAN and Trinity I got the
> following report:
>
> ==================================================================
> BUG: KASan: use after free in shrink_page_list+0x93a/0xf10 at addr
> ffff88003487da80

Hum, looking into a trace it seems KASAN really found a page with inode
belonging to a destroyed BDI. I think this may be the same problem that was
also reported in http://lists.openwall.net/linux-ext4/2015/08/13/3.
Christoph, did you have time to look into it?

Honza

> Read of size 8 by task kswapd0/622
> =============================================================================
> BUG kmalloc-16 (Not tainted): kasan: bad access detected
> -----------------------------------------------------------------------------
>
> Disabling lock debugging due to kernel taint
> INFO: Allocated in bdi_init+0xb9/0x480 age=235622 cpu=0 pid=6
> [< none >] __slab_alloc+0x44a/0x480 mm/slub.c:2402
> [< inline >] slab_alloc mm/slub.c:2470
> [< none >] kmem_cache_alloc_trace+0x12a/0x160 mm/slub.c:2529
> [< inline >] kmalloc include/linux/slab.h:440
> [< inline >] kzalloc include/linux/slab.h:593
> [< inline >] cgwb_bdi_init mm/backing-dev.c:749
> [< none >] bdi_init+0xb9/0x480 mm/backing-dev.c:775
> [< none >] blk_alloc_queue_node+0xfc/0x380 block/blk-core.c:656
> [< none >] blk_init_queue_node+0x1f/0x60 block/blk-core.c:754
> [< none >] blk_init_queue+0xe/0x10 block/blk-core.c:745
> [< none >] __scsi_alloc_queue+0x14/0x30 drivers/scsi/scsi_lib.c:2139
> [< none >] scsi_alloc_queue+0x1c/0x80 drivers/scsi/scsi_lib.c:2151
> [< none >] scsi_alloc_sdev+0x3cd/0x5f0 scsi_scan.c:0
> [< none >] scsi_probe_and_add_lun+0xc3a/0x10a0 scsi_scan.c:0
> [< none >] __scsi_add_device+0x112/0x120 ??:0
> [< none >] ata_scsi_scan_host+0xed/0x260 ??:0
> [< none >] async_port_probe+0x61/0x80 drivers/ata/libata-core.c:6097
> [< none >] async_run_entry_fn+0x74/0x190 async.c:0
> [< none >] process_one_work+0x276/0x630 kernel/workqueue.c:2030
> [< none >] worker_thread+0x98/0x720 kernel/workqueue.c:2162
> INFO: Freed in bdi_destroy+0x1d9/0x200 age=1073 cpu=0 pid=5919
> [< none >] __slab_free+0x150/0x270 mm/slub.c:2587
> [< inline >] slab_free mm/slub.c:2736
> [< none >] kfree+0x13a/0x150 mm/slub.c:3522
> [< inline >] wb_exit include/linux/backing-dev.h:483
> [< none >] bdi_destroy+0x1d9/0x200 mm/backing-dev.c:839
> [< none >] blk_cleanup_queue+0x158/0x190 block/blk-core.c:579
> [< none >] __scsi_remove_device+0x63/0x110 ??:0
> [< none >] scsi_remove_device+0x27/0x40 ??:0
> [< none >] sdev_store_delete+0x22/0x30 scsi_sysfs.c:0
> [< none >] dev_attr_store+0x39/0x50 drivers/base/core.c:137
> [< none >] sysfs_kf_write+0x8a/0xa0 file.c:0
> [< none >] kernfs_fop_write+0x167/0x200 file.c:0
> [< none >] __vfs_write+0x57/0x170 ??:0
> [< none >] vfs_write+0xeb/0x250 ??:0
> [< none >] SyS_write+0x53/0xb0 ??:0
> [< none >] entry_SYSCALL_64_fastpath+0x12/0x71
> arch/x86/entry/entry_64.S:185
>
> INFO: Slab 0xffffea0000d21f40 objects=12 used=4 fp=0xffff88003487d3f0
> flags=0x100000000000080
> INFO: Object 0xffff88003487da80 @offset=2688 fp=0x (null)
>
> Bytes b4 ffff88003487da70: a3 08 ff ff 00 00 00 00 aa 18 57 81 ff ff
> ff ff ..........W.....
> Object ffff88003487da80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
> 00 ................
> Redzone ffff88003487da90: bb bb bb bb bb bb bb bb
> ........
> Padding ffff88003487dbc8: 84 d5 76 81 ff ff ff ff
> ..v.....
> CPU: 0 PID: 622 Comm: kswapd0 Tainted: G B 4.3.0-rc4-kasan #16
> Hardware name: Red Hat KVM, BIOS 0.5.1 01/01/2007
> ffff88003487da80 ffff880034c0f878 ffffffff814a3e7c ffff880036003b00
> ffff880034c0f8a8 ffffffff812090b8 ffff880036003b00 ffffea0000d21f40
> ffff88003487da80 ffffea0000ceaf98 ffff880034c0f8d0 ffffffff8120def1
> Call Trace:
> [<ffffffff814a3e7c>] dump_stack+0x44/0x58 lib/dump_stack.c:15
> [<ffffffff812090b8>] print_trailer+0xf8/0x150 mm/slub.c:650
> [<ffffffff8120def1>] object_err+0x31/0x40 mm/slub.c:657
> [<ffffffff81210215>] kasan_report_error+0x1e5/0x3f0 ??:0
> [<ffffffff81210804>] kasan_report+0x34/0x40 ??:0
> [< inline >] ? inode_write_congested include/linux/backing-dev.h:193
> [<ffffffff811bcc1a>] ? shrink_page_list+0x93a/0xf10 mm/vmscan.c:957
> [<ffffffff8120f564>] __asan_load8+0x64/0xa0 ??:0
> [<ffffffff811c814f>] ? page_mapping+0x2f/0x70 ??:0
> [< inline >] inode_write_congested include/linux/backing-dev.h:193
> [<ffffffff811bcc1a>] shrink_page_list+0x93a/0xf10 mm/vmscan.c:957
> [<ffffffff811bda74>] shrink_inactive_list+0x2f4/0x5f0 mm/vmscan.c:1610
> [< inline >] shrink_list mm/vmscan.c:1945
> [<ffffffff811bea0a>] shrink_lruvec+0x87a/0xa50 mm/vmscan.c:2229
> [<ffffffff811beca0>] shrink_zone+0xc0/0x2c0 mm/vmscan.c:2413
> [< inline >] kswapd_shrink_zone mm/vmscan.c:3116
> [< inline >] balance_pgdat mm/vmscan.c:3291
> [<ffffffff811bffad>] kswapd+0x64d/0xb70 mm/vmscan.c:3499
> [<ffffffff811bf960>] ? zone_reclaim+0x2a0/0x2a0 mm/vmscan.c:3820
> [<ffffffff810a491f>] kthread+0x10f/0x130 kthread.c:0
> [<ffffffff810a4810>] ? kthread_park+0x70/0x70 ??:0
> [<ffffffff81d4fb5f>] ret_from_fork+0x3f/0x70 arch/x86/entry/entry_64.S:472
> [<ffffffff810a4810>] ? kthread_park+0x70/0x70 ??:0
> Memory state around the buggy address:
> ffff88003487d980: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> ffff88003487da00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> >ffff88003487da80: fb fb fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> ^
> ffff88003487db00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
> ffff88003487db80: fc fc fc fc fc fc fc fc fc fc 00 00 fc fc fc fc
> ==================================================================
>
> Thanks!
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/