Re: [PATCH v3 ath-next 2/2] wifi: ath11k: fix HTC rx insufficient length

[Johan Hovold <johan@kernel.org>]

On Fri, Mar 14, 2025 at 02:13:53PM +0800, Miaoqing Pan wrote:
> A relatively unusual race condition occurs between host software
> and hardware, where the host sees the updated destination ring head
> pointer before the hardware updates the corresponding descriptor.
> When this situation occurs, the length of the descriptor returns 0.
> 
> The current error handling method is to increment descriptor tail
> pointer by 1, but 'sw_index' is not updated, causing descriptor and
> skb to not correspond one-to-one, resulting in the following error:
> 
> ath11k_pci 0006:01:00.0: HTC Rx: insufficient length, got 1488, expected 1492
> ath11k_pci 0006:01:00.0: HTC Rx: insufficient length, got 1460, expected 1484
> 
> To address this problem and work around the broken hardware,
> temporarily skip processing the current descriptor and handle it
> again next time. However, to prevent this descriptor from
> continuously returning 0, use the skb control block (cb) to set
> a flag. If the length returns 0 again, this descriptor will be
> discarded.
> 
> Tested-on: QCA6698AQ hw2.1 PCI WLAN.HSP.1.1-04546-QCAHSPSWPL_V1_V2_SILICONZ_IOE-1
> 
> Reported-by: Johan Hovold <johan+linaro@kernel.org>
> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218623
> Signed-off-by: Miaoqing Pan <quic_miaoqing@quicinc.com>

> @@ -387,18 +387,36 @@ static int ath11k_ce_completed_recv_next(struct ath11k_ce_pipe *pipe,
>  
>  	ath11k_hal_srng_access_begin(ab, srng);
>  
> -	desc = ath11k_hal_srng_dst_get_next_entry(ab, srng);
> +	desc = ath11k_hal_srng_dst_peek(ab, srng);
>  	if (!desc) {
>  		ret = -EIO;
>  		goto err;
>  	}
>  
>  	*nbytes = ath11k_hal_ce_dst_status_get_length(desc);

As I mentioned elsewhere, this function also sets the length field in
the descriptor to zero. So if there's a racing update, you may never see
the updated length.

> -	if (*nbytes == 0) {
> -		ret = -EIO;
> -		goto err;
> +	if (unlikely(*nbytes == 0)) {
> +		struct ath11k_skb_rxcb *rxcb =
> +			ATH11K_SKB_RXCB(pipe->dest_ring->skb[sw_index]);
> +
> +		/* A relatively unusual race condition occurs between host
> +		 * software and hardware, where the host sees the updated
> +		 * destination ring head pointer before the hardware updates
> +		 * the corresponding descriptor.
> +		 *
> +		 * Temporarily skip processing the current descriptor and handle
> +		 * it again next time. However, to prevent this descriptor from
> +		 * continuously returning 0, set 'is_desc_len0' flag. If the
> +		 * length returns 0 again, this descriptor will be discarded.
> +		 */
> +		if (!rxcb->is_desc_len0) {
> +			rxcb->is_desc_len0 = true;
> +			ret = -EIO;
> +			goto err;
> +		}

If you add the memory barrier and make sure not to clear the length
field above, do you still see the length sometimes always reading zero
if you retry more than once (i.e. drop the is_desc_len0 flag)?

Perhaps the device is really passing you a zero-length descriptor that
can simply be discarded straight away?

Johan