netdev - RE: [PATCH v4 net-next 4/5] net: fec: add change_mtu to support dynamic buffer allocation

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID:
 <PAXPR04MB8510EA7CD915DE9340303B0F8807A@PAXPR04MB8510.eurprd04.prod.outlook.com>
Date: Mon, 1 Sep 2025 06:52:32 +0000
From: Wei Fang <wei.fang@....com>
To: Shenwei Wang <shenwei.wang@....com>
CC: Clark Wang <xiaoning.wang@....com>, Stanislav Fomichev <sdf@...ichev.me>,
	"imx@...ts.linux.dev" <imx@...ts.linux.dev>, "netdev@...r.kernel.org"
	<netdev@...r.kernel.org>, "linux-kernel@...r.kernel.org"
	<linux-kernel@...r.kernel.org>, dl-linux-imx <linux-imx@....com>, Andrew Lunn
	<andrew+netdev@...n.ch>, "David S. Miller" <davem@...emloft.net>, Eric
 Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>, Paolo Abeni
	<pabeni@...hat.com>, Alexei Starovoitov <ast@...nel.org>, Daniel Borkmann
	<daniel@...earbox.net>, Jesper Dangaard Brouer <hawk@...nel.org>, John
 Fastabend <john.fastabend@...il.com>
Subject: RE: [PATCH v4 net-next 4/5] net: fec: add change_mtu to support
 dynamic buffer allocation

>   * 2048 byte skbufs are allocated. However, alignment requirements
>   * varies between FEC variants. Worst case is 64, so round down by 64.
>   */
> +#define FEC_DRV_RESERVE_SPACE \
> +	(XDP_PACKET_HEADROOM + \
> +	 SKB_DATA_ALIGN((unsigned int)sizeof(struct skb_shared_info)))

"unsigned int" can be removed. And please move FEC_DRV_RESERVE_SPACE
to fec.h, FEC_ENET_RX_FRSIZE can be updated to
(PAGE_SIZE - FEC_DRV_RESERVE_SPACE).

> +static int fec_change_mtu(struct net_device *ndev, int new_mtu)
> +{
> +	struct fec_enet_private *fep = netdev_priv(ndev);
> +	int old_mtu, old_order, old_size, order, done;
> +	int ret = 0;
> +
> +	order = get_order(new_mtu + ETH_HLEN + ETH_FCS_LEN +
> FEC_DRV_RESERVE_SPACE);
> +	old_order = fep->pagepool_order;
> +	old_size = fep->rx_frame_size;
> +	old_mtu = READ_ONCE(ndev->mtu);
> +	fep->pagepool_order = order;
> +	fep->rx_frame_size = (PAGE_SIZE << order) - FEC_DRV_RESERVE_SPACE;
> +
> +	if (!netif_running(ndev)) {
> +		WRITE_ONCE(ndev->mtu, new_mtu);
> +		return 0;
> +	}
> +
> +	/* Stop TX/RX and free the buffers */
> +	napi_disable(&fep->napi);
> +	netif_tx_disable(ndev);
> +	read_poll_timeout(fec_enet_rx_napi, done, (done == 0),
> +			  10, 1000, false, &fep->napi, 10);
> +	fec_stop(ndev);
> +
> +	WRITE_ONCE(ndev->mtu, new_mtu);
> +
> +	if (fep->pagepool_order != old_order) {

If fep->pagepool_order is not changed, why need to stop TX/RX?

> +		fec_enet_free_buffers(ndev);
> +
> +		/* Create the pagepool based on the new mtu.
> +		 * Revert to the original settings if buffer
> +		 * allocation fails.
> +		 */
> +		if (fec_enet_alloc_buffers(ndev) < 0) {
> +			fep->pagepool_order = old_order;
> +			fep->rx_frame_size = old_size;
> +			WRITE_ONCE(ndev->mtu, old_mtu);
> +			fec_enet_alloc_buffers(ndev);

fec_enet_alloc_buffers() may still fail here, the best approach is to add a helper
function that allocates new buffers before freeing the old ones. However, this is
a complex change, so I think returning an error is sufficient for now, and we can
consider improving it later.