Re: [PATCH] blktrace: Fix uaf in blk_trace access after removing by sysfs

[Zhihao Cheng]

在 2021/9/10 17:21, Zhihao Cheng 写道:
friendly ping
> There is an use-after-free problem triggered by following process:
>
>        P1(sda)				P2(sdb)
> 			echo 0 > /sys/block/sdb/trace/enable
> 			  blk_trace_remove_queue
> 			    synchronize_rcu
> 			    blk_trace_free
> 			      relay_close
> rcu_read_lock
> __blk_add_trace
>    trace_note_tsk
>    (Iterate running_trace_list)
> 			        relay_close_buf
> 				  relay_destroy_buf
> 				    kfree(buf)
>      trace_note(sdb's bt)
>        relay_reserve
>          buf->offset <- nullptr deference (use-after-free) !!!
> rcu_read_unlock
>
> [  502.714379] BUG: kernel NULL pointer dereference, address:
> 0000000000000010
> [  502.715260] #PF: supervisor read access in kernel mode
> [  502.715903] #PF: error_code(0x0000) - not-present page
> [  502.716546] PGD 103984067 P4D 103984067 PUD 17592b067 PMD 0
> [  502.717252] Oops: 0000 [#1] SMP
> [  502.720308] RIP: 0010:trace_note.isra.0+0x86/0x360
> [  502.732872] Call Trace:
> [  502.733193]  __blk_add_trace.cold+0x137/0x1a3
> [  502.733734]  blk_add_trace_rq+0x7b/0xd0
> [  502.734207]  blk_add_trace_rq_issue+0x54/0xa0
> [  502.734755]  blk_mq_start_request+0xde/0x1b0
> [  502.735287]  scsi_queue_rq+0x528/0x1140
> ...
> [  502.742704]  sg_new_write.isra.0+0x16e/0x3e0
> [  502.747501]  sg_ioctl+0x466/0x1100
>
> Reproduce method:
>    ioctl(/dev/sda, BLKTRACESETUP, blk_user_trace_setup[buf_size=127])
>    ioctl(/dev/sda, BLKTRACESTART)
>    ioctl(/dev/sdb, BLKTRACESETUP, blk_user_trace_setup[buf_size=127])
>    ioctl(/dev/sdb, BLKTRACESTART)
>
>    echo 0 > /sys/block/sdb/trace/enable &
>    // Add delay(mdelay/msleep) before kernel enters blk_trace_free()
>
>    ioctl$SG_IO(/dev/sda, SG_IO, ...)
>    // Enters trace_note_tsk() after blk_trace_free() returned
>    // Use mdelay in rcu region rather than msleep(which may schedule out)
>
> Don't remove blk_trace by sysfs when blk_trace is at Blktrace_running
> state, just like function __blk_trace_remove() does. The state change
> process and blk_trace_remove_queue() are protected and by mutex lock
> 'q->debugfs_mutex', so the sequence of stopping blk_trace first and
> then removing it will be ensured.
>
> Fixes: c71a896154119f ("blktrace: add ftrace plugin")
> Signed-off-by: Zhihao Cheng <chengzhihao1@xxxxxxxxxx>
> ---
>   kernel/trace/blktrace.c | 13 +++++++++++--
>   1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c
> index c221e4c3f625..7fe29bb9746f 100644
> --- a/kernel/trace/blktrace.c
> +++ b/kernel/trace/blktrace.c
> @@ -1821,8 +1821,17 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev,
>   		}
>   		if (value)
>   			ret = blk_trace_setup_queue(q, bdev);
> -		else
> -			ret = blk_trace_remove_queue(q);
> +		else {
> +			/*
> +			 * Don't remove blk_trace under running state, in
> +			 * case triggering use-after-free in function
> +			 * __blk_add_trace().
> +			 */
> +			if (bt->trace_state != Blktrace_running)
> +				ret = blk_trace_remove_queue(q);
> +			else
> +				ret = -EBUSY;
> +		}
>   		goto out_unlock_bdev;
>   	}
>