Re: [PATCH net-next v5 7/8] net: bcmgenet: allow the MTU to be changed

[Nicolai Buchwitz]

On 17.10.2026 03:24, Jakub Kicinski wrote:
> On Tue, 13 Oct 2026 16:20:36 +0200 Nicolai Buchwitz wrote:
> > +	ret = bcmgenet_init_dma(priv, true);
> > +	if (ret) {
> > +		/* Retry the size that was allocated a moment ago */
> > +		WRITE_ONCE(dev->mtu, old_mtu);
> > +		priv->rx_buf_len = bcmgenet_rx_buf_len(old_mtu);
> > +		bcmgenet_set_mtu_regs(priv, old_mtu);
> > +		if (bcmgenet_init_dma(priv, true)) {
>
> This will fail under memory pressure leaving the device inoperable.

Aye, I'll allocate the new rings before freeing the old ones, then a
failed MTU change just returns -ENOMEM and the rollback can go.

> > +	priv->datapath_up = true;
>
> Hallmark of a bad design

Agreed, and redundant anyway once fini_dma() NULLs rx_cbs/tx_cbs. I have
a net series for that, on which I will rebuild this patch.

Thanks,
Nicolai