linux-kernel - Re: [PATCH v5] wifi: ath11k: Fix NULL pointer check in ath11k_ce_rx_post

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <f0077f4f-922d-56b1-a5b7-bb01e4462e87@quicinc.com>
Date: Wed, 27 Nov 2024 20:53:07 +0530
From: Vasanthakumar Thiagarajan <quic_vthiagar@...cinc.com>
To: Baichuan Qi <zghbqbc@...il.com>, <markus.elfring@....de>
CC: <ath11k@...ts.infradead.org>, <jjohnson@...nel.org>, <kvalo@...nel.org>,
        <linux-kernel@...r.kernel.org>, <linux-wireless@...r.kernel.org>
Subject: Re: [PATCH v5] wifi: ath11k: Fix NULL pointer check in
 ath11k_ce_rx_post_pipe()



On 11/27/2024 8:08 PM, Baichuan Qi wrote:
> Current implementation of `ath11k_ce_rx_post_pipe()` checks for
> NON-NULL of either `dest_ring` or `status_ring` using an OR (||).
> Both rings, especially `dest_ring`, should be ensured to be
> NON-NULL in this function.
> 
> If only one of the rings is valid, such as `dest_ring` is NULL
> and `status_ring` is NON-NULL, the OR (||) check would not stop
> `ath11k_ce_rx_post_pipe()`, the subsequent call to
> `ath11k_ce_rx_buf_enqueue_pipe()` will access the NULL pointer,
> resulting in a driver crash.
> 
> Fix the NON-NULL check by changing the OR (||) to AND (&&),
> and return an error code `-EIO` to indicate
> `ath11k_ce_rx_post_pipe()` is stopped with an NULL pointer
> error, ensuring that the function only proceeds when both
> `dest_ring` and `status_ring` are NON-NULL.
> 
> Link: https://lore.kernel.org/ath11k/a9ccc947-20b2-4322-84e5-c96aaa604e63@web.de
> Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices")

This does not really fix any real issue. Please check ath11k_ce_alloc_pipe()
where initialization would fail if anyone of pipe->dest_ring and
pipe->status_ring allocation fails for ce pipe used for Rx.

> Signed-off-by: Baichuan Qi <zghbqbc@...il.com>
> ---
> V4 -> V5: add err code in NULL check
> V3 -> V4: reorder describe info
> V2 -> V3: add Link URL to mailing list archives
> V1 -> V2: rewrite commit message and fix tag
> 
>   drivers/net/wireless/ath/ath11k/ce.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/net/wireless/ath/ath11k/ce.c b/drivers/net/wireless/ath/ath11k/ce.c
> index e66e86bdec20..223dab928453 100644
> --- a/drivers/net/wireless/ath/ath11k/ce.c
> +++ b/drivers/net/wireless/ath/ath11k/ce.c
> @@ -324,8 +324,10 @@ static int ath11k_ce_rx_post_pipe(struct ath11k_ce_pipe *pipe)
>   	dma_addr_t paddr;
>   	int ret = 0;
>   
> -	if (!(pipe->dest_ring || pipe->status_ring))
> -		return 0;
> +	if (!(pipe->dest_ring && pipe->status_ring)) {
> +		ret = -EIO;
> +		return ret;
> +	}

This will always fail as the caller loops through all the supported ce pipes
and ce pipes used for Tx will not have either dest_ring or status_ring.
Please ensure the patch is tested properly.

So NAK

Vasanth