public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: Jakub Kicinski <kuba@kernel.org>
To: pvalerio@redhat.com
Cc: Jakub Kicinski <kuba@kernel.org>,
	davem@davemloft.net, andrew+netdev@lunn.ch, pabeni@redhat.com,
	lorenzo@kernel.org, claudiu.beznea@tuxon.dev,
	nicolas.ferre@microchip.com, theo.lebrun@bootlin.com,
	netdev@vger.kernel.org, edumazet@google.com
Subject: Re: [net-next,5/8] cadence: macb: add XDP support for gem
Date: Mon, 19 Jan 2026 11:36:24 -0800	[thread overview]
Message-ID: <20260119193624.1255742-1-kuba@kernel.org> (raw)
In-Reply-To: <20260115222531.313002-6-pvalerio@redhat.com>

This is an AI-generated review of your patch. The human sending this
email has considered the AI review valid, or at least plausible.

For local reproduction steps see https://netdev-ai.bots.linux.dev/ai-local.html
---
cadence: macb: add XDP support for gem

This commit introduces XDP support for the macb/gem driver with XDP_PASS,
XDP_DROP, and XDP_REDIRECT verdicts.

> diff --git a/drivers/net/ethernet/cadence/macb_main.c b/drivers/net/ethernet/cadence/macb_main.c
> index 89f0c4dc3884..1f62100a4c4d 100644
> --- a/drivers/net/ethernet/cadence/macb_main.c
> +++ b/drivers/net/ethernet/cadence/macb_main.c

[ ... ]

> -static int gem_create_page_pool(struct macb_queue *queue)
> +static int gem_create_page_pool(struct macb_queue *queue, int qid)
>  {
>  	struct page_pool_params pp_params = {
>  		.order = 0,
>  		.flags = PP_FLAG_DMA_MAP | PP_FLAG_DMA_SYNC_DEV,
>  		.pool_size = queue->bp->rx_ring_size,
>  		.nid = NUMA_NO_NODE,
> -		.dma_dir = DMA_FROM_DEVICE,
> +		.dma_dir = rcu_access_pointer(queue->bp->prog)
> +				? DMA_BIDIRECTIONAL
> +				: DMA_FROM_DEVICE,

The page pool is created with DMA direction based on whether an XDP program
is attached at creation time. However, gem_xdp_setup() below allows attaching
XDP programs to a running interface without recreating the page pool.

If the interface is opened without XDP (page pool uses DMA_FROM_DEVICE), and
then an XDP program is attached via ndo_bpf, the pool will still use
DMA_FROM_DEVICE. For XDP redirect, DMA_BIDIRECTIONAL is needed for proper
cache coherency on non-cache-coherent architectures.

The stmmac driver handles this by calling stmmac_xdp_release() and
stmmac_xdp_open() when XDP program state changes on a running interface,
which recreates the page pool with the correct DMA direction.

Should gem_xdp_setup() reset the interface (or at least the page pools) when
attaching or detaching XDP programs on a running interface to ensure the DMA
direction matches?

[ ... ]

> +static int gem_xdp_setup(struct net_device *dev, struct bpf_prog *prog,
> +			 struct netlink_ext_ack *extack)
> +{
> +	struct macb *bp = netdev_priv(dev);
> +	struct bpf_prog *old_prog;
> +
> +	if (prog && !gem_xdp_valid_mtu(bp, dev->mtu)) {
> +		NL_SET_ERR_MSG_MOD(extack, "MTU too large for XDP");
> +		return -EOPNOTSUPP;
> +	}
> +
> +	old_prog = rcu_replace_pointer(bp->prog, prog, lockdep_rtnl_is_held());
> +	if (old_prog)
> +		bpf_prog_put(old_prog);
> +
> +	return 0;
> +}

This function swaps the XDP program via RCU but does not reset the interface
or recreate the page pools. This relates to the DMA direction concern above.

  reply	other threads:[~2026-01-19 19:36 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-01-15 22:25 [PATCH net-next 0/8] net: macb: Add XDP support and page pool integration Paolo Valerio
2026-01-15 22:25 ` [PATCH net-next 1/8] net: macb: move Rx buffers alloc from link up to open Paolo Valerio
2026-01-15 22:25 ` [PATCH net-next 2/8] net: macb: rename rx_skbuff into rx_buff Paolo Valerio
2026-01-15 22:25 ` [PATCH net-next 3/8] cadence: macb: Add page pool support handle multi-descriptor frame rx Paolo Valerio
2026-01-16 17:16   ` Andrew Lunn
2026-01-19 18:58     ` Paolo Valerio
2026-01-22 22:24     ` Paolo Valerio
2026-01-22 23:04       ` Andrew Lunn
2026-01-25 19:02         ` Paolo Valerio
2026-01-26 14:29           ` Andrew Lunn
2026-01-26 18:45             ` Théo Lebrun
2026-01-26 23:51               ` Paolo Valerio
2026-01-27 15:48                 ` Théo Lebrun
2026-01-26 23:34             ` Paolo Valerio
2026-01-19 19:36   ` [net-next,3/8] " Jakub Kicinski
2026-01-22 14:39     ` Théo Lebrun
2026-01-22 15:16       ` Jakub Kicinski
2026-01-26 14:55   ` [PATCH net-next 3/8] " Théo Lebrun
2026-02-20 15:45   ` Théo Lebrun
2026-01-15 22:25 ` [PATCH net-next 4/8] cadence: macb: use the current queue number for stats Paolo Valerio
2026-01-15 22:25 ` [PATCH net-next 5/8] cadence: macb: add XDP support for gem Paolo Valerio
2026-01-19 19:36   ` Jakub Kicinski [this message]
2026-01-15 22:25 ` [PATCH net-next 6/8] cadence: macb: make macb_tx_skb generic Paolo Valerio
2026-01-15 22:25 ` [PATCH net-next 7/8] cadence: macb: make tx path skb agnostic Paolo Valerio
2026-01-15 22:25 ` [PATCH net-next 8/8] cadence: macb: introduce xmit support Paolo Valerio
2026-01-19 19:36   ` [net-next,8/8] " Jakub Kicinski
2026-02-02 16:31 ` [PATCH net-next 0/8] net: macb: Add XDP support and page pool integration Théo Lebrun
2026-02-13 16:57   ` [PATCH 1/6] net: macb: rename release_buff() -> macb_tx_release_buff() Théo Lebrun
2026-02-13 16:57     ` [PATCH 2/6] net: macb: drop two labels in gem_rx() Théo Lebrun
2026-02-13 16:57     ` [PATCH 3/6] net: macb: always use DMA_BIDIRECTIONAL on page pool buffers Théo Lebrun
2026-02-13 16:57     ` [PATCH 4/6] net: macb: account for stats in Rx XDP codepaths Théo Lebrun
2026-02-13 16:57     ` [PATCH 5/6] net: macb: improve Rx refill error message Théo Lebrun
2026-02-13 16:57     ` [PATCH 6/6] net: macb: rework macb_tx_complete() processing loop Théo Lebrun
2026-02-13 16:57 ` [PATCH net-next 0/8] net: macb: Add XDP support and page pool integration Théo Lebrun
2026-02-13 17:02   ` Théo Lebrun
2026-02-14 15:37   ` Paolo Valerio
2026-02-16  9:17     ` Théo Lebrun
2026-02-19 18:05       ` Paolo Valerio
2026-02-20 15:58         ` Théo Lebrun

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260119193624.1255742-1-kuba@kernel.org \
    --to=kuba@kernel.org \
    --cc=andrew+netdev@lunn.ch \
    --cc=claudiu.beznea@tuxon.dev \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=lorenzo@kernel.org \
    --cc=netdev@vger.kernel.org \
    --cc=nicolas.ferre@microchip.com \
    --cc=pabeni@redhat.com \
    --cc=pvalerio@redhat.com \
    --cc=theo.lebrun@bootlin.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox