]> www.infradead.org Git - users/willy/pagecache.git/commitdiff
net: stmmac: Apply new page pool parameters when SPH is enabled
authorFurong Xu <0x1207@gmail.com>
Fri, 7 Feb 2025 08:56:39 +0000 (16:56 +0800)
committerJakub Kicinski <kuba@kernel.org>
Tue, 11 Feb 2025 02:04:00 +0000 (18:04 -0800)
Commit df542f669307 ("net: stmmac: Switch to zero-copy in
non-XDP RX path") makes DMA write received frame into buffer at offset
of NET_SKB_PAD and sets page pool parameters to sync from offset of
NET_SKB_PAD. But when Header Payload Split is enabled, the header is
written at offset of NET_SKB_PAD, while the payload is written at
offset of zero. Uncorrect offset parameter for the payload breaks dma
coherence [1] since both CPU and DMA touch the page buffer from offset
of zero which is not handled by the page pool sync parameter.

And in case the DMA cannot split the received frame, for example,
a large L2 frame, pp_params.max_len should grow to match the tail
of entire frame.

[1] https://lore.kernel.org/netdev/d465f277-bac7-439f-be1d-9a47dfe2d951@nvidia.com/

Reported-by: Jon Hunter <jonathanh@nvidia.com>
Reported-by: Brad Griffis <bgriffis@nvidia.com>
Suggested-by: Ido Schimmel <idosch@idosch.org>
Fixes: df542f669307 ("net: stmmac: Switch to zero-copy in non-XDP RX path")
Signed-off-by: Furong Xu <0x1207@gmail.com>
Tested-by: Jon Hunter <jonathanh@nvidia.com>
Tested-by: Thierry Reding <treding@nvidia.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Link: https://patch.msgid.link/20250207085639.13580-1-0x1207@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
drivers/net/ethernet/stmicro/stmmac/stmmac_main.c

index b34ebb916b8989658b04a6dc44751350e7119b61..c0ae7db96f46ffc21279d273911a668b23f9c35f 100644 (file)
@@ -2094,6 +2094,11 @@ static int __alloc_dma_rx_desc_resources(struct stmmac_priv *priv,
        pp_params.offset = stmmac_rx_offset(priv);
        pp_params.max_len = dma_conf->dma_buf_sz;
 
+       if (priv->sph) {
+               pp_params.offset = 0;
+               pp_params.max_len += stmmac_rx_offset(priv);
+       }
+
        rx_q->page_pool = page_pool_create(&pp_params);
        if (IS_ERR(rx_q->page_pool)) {
                ret = PTR_ERR(rx_q->page_pool);