Re: [PATCH] bcache: Fix a NULL or wild pointer dereference in btree_gc_rewrite_node()

[Coly Li <colyli@suse.de>]

On 1/25/22 12:47 AM, Zhou Qingyang wrote:
> In btree_gc_rewrite_node(), btree_node_alloc_replacement() is assigned to
> n and return error code or NULL on failure. n is passed to
> bch_btree_node_write_sync() and there is a dereference of it in
> bch_btree_node_write_sync() without checks, which may lead to wild
> pointer dereference or NULL pointer dereference depending on n.
>
> Fix this bug by adding IS_ERR_OR_NULL check of n.
>
> This bug was found by a static analyzer.
>
> Builds with 'make allyesconfig' show no new warnings,
> and our static analyzer no longer warns about this code.
>
> Fixes: ("bcache: Rework btree cache reserve handling")
> Signed-off-by: Zhou Qingyang <zhou1615@umn.edu>
> ---
> The analysis employs differential checking to identify inconsistent
> security operations (e.g., checks or kfrees) between two code paths
> and confirms that the inconsistent operations are not recovered in the
> current function or the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>
>   drivers/md/bcache/btree.c | 2 ++
>   1 file changed, 2 insertions(+)
>
> diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
> index 88c573eeb598..06d42292e86c 100644
> --- a/drivers/md/bcache/btree.c
> +++ b/drivers/md/bcache/btree.c
> @@ -1504,6 +1504,8 @@ static int btree_gc_rewrite_node(struct btree *b, struct btree_op *op,
>   		return 0;
>   
>   	n = btree_node_alloc_replacement(replace, NULL);
> +	if (IS_ERR_OR_NULL(n))
> +		return 0;
>   

Hi Qingyang,

This is a valid fix with my first glance. I add this patch into my 
testing queue.

Thanks.

Coly Li


>   	/* recheck reserve after allocating replacement node */
>   	if (btree_check_reserve(b, NULL)) {