Re: [PATCH net-next v9 4/6] net: tn40xx: add basic Rx handling

["Russell King (Oracle)" <linux@armlinux.org.uk>]

On Thu, Jun 06, 2024 at 08:26:06AM +0900, FUJITA Tomonori wrote:
> +static int tn40_rxdb_alloc_elem(struct tn40_rxdb *db)
> +{
> +	return db->stack[--(db->top)];

Parens are unnecessary here.

> +static void tn40_rxdb_free_elem(struct tn40_rxdb *db, unsigned int n)
> +{
> +	db->stack[(db->top)++] = n;

Same here.

> +	dno = tn40_rxdb_available(db) - 1;
> +	i = dno;
> +	while (i > 0) {
> +		page = page_pool_dev_alloc_pages(priv->page_pool);
> +		if (!page)
> +			break;
> +
> +		idx = tn40_rxdb_alloc_elem(db);
> +		tn40_set_rx_desc(priv, idx, page_pool_get_dma_addr(page));
> +		dm = tn40_rxdb_addr_elem(db, idx);
> +		dm->page = page;
> +
> +		i--;
> +	}

While reviewing the rxdb stack, I came across this - this while() loop
is an open-coded for() loop:

	for (i = dno; i > 0; i--) {
		page = page_pool_dev_alloc_pages(priv->page_pool);
		...
		dm->page = page;
	}

Is there any reason not to use a for() loop here?

> +	if (i != dno)
> +		tn40_write_reg(priv, f->m.reg_wptr,
> +			       f->m.wptr & TN40_TXF_WPTR_WR_PTR);

...

>+struct tn40_rxdb {
> +	int *stack;
> +	struct tn40_rx_map *elems;
> +	int nelem;
> +	int top;

I assume neither of these should ever be negative, so should these be
"unsigned int" ?

Thanks.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!