Re: [PATCH RFC net-next 6/6] cadence: macb/gem: introduce xmit support

["Théo Lebrun" <theo.lebrun@bootlin.com>]

On Tue Dec 2, 2025 at 6:34 PM CET, Paolo Valerio wrote:
> On 27 Nov 2025 at 04:07:52 PM, Théo Lebrun <theo.lebrun@bootlin.com> wrote:
>> On Wed Nov 19, 2025 at 2:53 PM CET, Paolo Valerio wrote:
>>> diff --git a/drivers/net/ethernet/cadence/macb_main.c b/drivers/net/ethernet/cadence/macb_main.c
>>> index eeda1a3871a6..bd62d3febeb1 100644
>>> --- a/drivers/net/ethernet/cadence/macb_main.c
>>> +++ b/drivers/net/ethernet/cadence/macb_main.c
>>> +static int macb_xdp_submit_frame(struct macb *bp, struct xdp_frame *xdpf,
>>> +				 struct net_device *dev, dma_addr_t addr)
>>> +{
>>> +	enum macb_tx_buff_type buff_type;
>>> +	struct macb_tx_buff *tx_buff;
>>> +	int cpu = smp_processor_id();
>>> +	struct macb_dma_desc *desc;
>>> +	struct macb_queue *queue;
>>> +	unsigned long flags;
>>> +	dma_addr_t mapping;
>>> +	u16 queue_index;
>>> +	int err = 0;
>>> +	u32 ctrl;
>>> +
>>> +	queue_index = cpu % bp->num_queues;
>>> +	queue = &bp->queues[queue_index];
>>> +	buff_type = !addr ? MACB_TYPE_XDP_NDO : MACB_TYPE_XDP_TX;
>>
>> I am not the biggest fan of piggy-backing on !!addr to know which
>> codepath called us. If the macb_xdp_submit_frame() call in gem_xdp_run()
>> ever gives an addr=0 coming from macb_get_addr(bp, desc), then we will
>> be submitting NDO typed frames and creating additional DMA mappings
>> which would be a really hard to debug bug.
>
> I guess we can add a separate boolean, WDYT?

I agree!

Thanks,

--
Théo Lebrun, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com