Re: [PATCH net v2] net: airoha: Move entries to queue head in case of DMA mapping failure in airoha_dev_xmit()

[Lorenzo Bianconi <lorenzo@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 2338 bytes --]

> In order to respect the original descriptor order and avoid any
> potential IOMMU fault or memory corruption, move pending queue entries
> to the head of hw queue tx_list if the DMA mapping of current inflight
> packet fails in airoha_dev_xmit routine.
> 
> Fixes: 3f47e67dff1f7 ("net: airoha: Add the capability to consume out-of-order DMA tx descriptors")
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> ---
> Changes in v2:
> - Remove leftover airoha_queue_entry initialization inside the loop from
>   the previous version.
> - Link to v1: https://lore.kernel.org/r/20260428-airoha-xmit-unmap-error-path-v1-1-aec44c6765c1@kernel.org
> ---
>  drivers/net/ethernet/airoha/airoha_eth.c | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
> index 5effb4a4ae84..fb4f51d67dd7 100644
> --- a/drivers/net/ethernet/airoha/airoha_eth.c
> +++ b/drivers/net/ethernet/airoha/airoha_eth.c
> @@ -2123,14 +2123,12 @@ static netdev_tx_t airoha_dev_xmit(struct sk_buff *skb,
>  	return NETDEV_TX_OK;
>  
>  error_unmap:
> -	while (!list_empty(&tx_list)) {
> -		e = list_first_entry(&tx_list, struct airoha_queue_entry,
> -				     list);
> +	list_for_each_entry(e, &tx_list, list) {
>  		dma_unmap_single(dev->dev.parent, e->dma_addr, e->dma_len,
>  				 DMA_TO_DEVICE);
>  		e->dma_addr = 0;
> -		list_move_tail(&e->list, &q->tx_list);
>  	}
> +	list_splice(&tx_list, &q->tx_list);
>  
>  	spin_unlock_bh(&q->lock);
>  error:

commenting on Sashiko's reported issues:
https://sashiko.dev/#/patchset/20260429-airoha-xmit-unmap-error-path-v2-1-32e43b7c6d25%40kernel.org

- Should the driver be updated to use skb_frag_dma_map() for mapping the SKB
  fragments, and correspondingly use dma_unmap_page() for those fragments in
  this error handling path?
  As we already pointed out here [0], we decided to skip this request and keep
  using dma_map_single() even for skb fragments.

Regards,
Lorenzo

[0] https://lore.kernel.org/netdev/20260427173935.6a1e36ee@kernel.org/

> 
> ---
> base-commit: 0c7a5ba011d336df4fcd1f667fcc16ea5549be12
> change-id: 20260428-airoha-xmit-unmap-error-path-8125594efb23
> 
> Best regards,
> -- 
> Lorenzo Bianconi <lorenzo@kernel.org>
> 

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]