Re: [PATCH iwl-next 03/10] ixgbevf: use libeth in Rx processing

[Simon Horman <horms@kernel.org>]

On Wed, Feb 25, 2026 at 06:19:48PM +0100, Larysa Zaremba wrote:
> On Tue, Feb 24, 2026 at 05:45:50PM +0000, Simon Horman wrote:
> > On Mon, Feb 23, 2026 at 10:52:10AM +0100, Larysa Zaremba wrote:
> > > Use page_pool buffers by the means of libeth in the Rx queues, this
> > > significantly reduces code complexity of the driver itself.
> > > 
> > > Suggested-by: Alexander Lobakin <aleksander.lobakin@intel.com>
> > > Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
> > > Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
> > > Signed-off-by: Larysa Zaremba <larysa.zaremba@intel.com>
> > 
> > ...
> > 
> > > @@ -3257,12 +3133,26 @@ static int ixgbevf_setup_all_tx_resources(struct ixgbevf_adapter *adapter)
> > >  int ixgbevf_setup_rx_resources(struct ixgbevf_adapter *adapter,
> > >  			       struct ixgbevf_ring *rx_ring)
> > >  {
> > > -	int size;
> > > +	struct libeth_fq fq = {
> > > +		.count		= rx_ring->count,
> > > +		.nid		= NUMA_NO_NODE,
> > > +		.type		= LIBETH_FQE_MTU,
> > > +		.xdp		= !!rx_ring->xdp_prog,
> > > +		.idx		= rx_ring->queue_index,
> > > +		.buf_len	= IXGBEVF_RX_PAGE_LEN(rx_ring->xdp_prog ?
> > > +						      LIBETH_XDP_HEADROOM :
> > > +						      LIBETH_SKB_HEADROOM),
> > > +	};
> > > +	int ret;
> > >  
> > > -	size = sizeof(struct ixgbevf_rx_buffer) * rx_ring->count;
> > > -	rx_ring->rx_buffer_info = vmalloc(size);
> > > -	if (!rx_ring->rx_buffer_info)
> > > -		goto err;
> > > +	ret = libeth_rx_fq_create(&fq, &rx_ring->q_vector->napi);
> > > +	if (ret)
> > > +		return ret;
> > > +
> > > +	rx_ring->pp = fq.pp;
> > > +	rx_ring->rx_fqes = fq.fqes;
> > > +	rx_ring->truesize = fq.truesize;
> > > +	rx_ring->rx_buf_len = fq.buf_len;
> > >  
> > >  	u64_stats_init(&rx_ring->syncp);
> > >  
> > > @@ -3270,25 +3160,29 @@ int ixgbevf_setup_rx_resources(struct ixgbevf_adapter *adapter,
> > >  	rx_ring->size = rx_ring->count * sizeof(union ixgbe_adv_rx_desc);
> > >  	rx_ring->size = ALIGN(rx_ring->size, 4096);
> > >  
> > > -	rx_ring->desc = dma_alloc_coherent(rx_ring->dev, rx_ring->size,
> > > +	rx_ring->desc = dma_alloc_coherent(fq.pp->p.dev, rx_ring->size,
> > >  					   &rx_ring->dma, GFP_KERNEL);
> > >  
> > >  	if (!rx_ring->desc)
> > 
> > Hi Larysa,
> > 
> > Prior to this patch, if this error condition was met,
> > then function would return -ENOMEM. But now it will return 0.
> > This does not seem intentional.
> > 
> > Flagged by Smatch.
> 
> No, it is not intentional, though the problem is fixed in a later patch.
> Will correct this one in v2.

Thanks, and sorry for not noticing that it is fixed later in the series.