On Jun 10, Eric Dumazet wrote:

On Tue, Jun 10, 2025 at 2:12 AM Lorenzo Bianconi [off-list ref] wrote:

quoted

[...]

quoted

@@ -767,7 +887,7 @@ static int airoha_qdma_init_rx_queue(struct airoha_queue *q,
        int qid = q - &qdma->q_rx[0], thr;
        dma_addr_t dma_addr;

-       q->buf_size = PAGE_SIZE / 2;
+       q->buf_size = pp_params.max_len / (2 * (1 + lro_queue));

Tell us more... It seems small LRO packets will consume a lot of
space, incurring a small skb->len/skb->truesize ratio, and bad TCP WAN
performance.

I think the main idea is forward to hw LRO queues (queues 24-31 in this
case) just specific protocols with mostly big packets but I completely
agree we have an issue for small packets. One possible approach would be
to define a threshold (e.g. 256B) and allocate a buffer or page from the
page allocator for small packets (something similar to what mt7601u driver
is doing[0]).  What do you think?

And order-5 pages are unlikely to be available in the long run anyway.

I agree. I guess we can reduce the order to ~ 2 (something similar to
mtk_eth_soc hw LRO implementation [1]).

LRO support would only make sense if the NIC is able to use multiple
order-0 pages to store the payload.

The hw supports splitting big packets over multiple order-0 pages if we
increase the MTU over one page size, but according to my understanding
hw LRO requires contiguous memory to work.

Regards,
Lorenzo

[0] https://github.com/torvalds/linux/blob/master/drivers/net/wireless/mediatek/mt7601u/dma.c#L146
[1] https://github.com/torvalds/linux/blob/master/drivers/net/ethernet/mediatek/mtk_eth_soc.c#L2258