Re: [PATCH v5 net-next 5/9] net: dsa: netc: initialize the group bitmap of ETT and ECT

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] ETT/ECT table group size calculation incorrectly uses the number of available ports rather than total physical hardware ports, leading to cross-group table entry overlap and corruption when some ports are unused.
--

commit fef50c5cec2c9c640a1d7dda90141b99f16c02ff
Author: Wei Fang <wei.fang@nxp.com>

net: dsa: netc: initialize the group bitmap of ETT and ECT

Initializes the group index bitmaps for the Egress Treatment Table (ETT) and
Egress Count Table (ECT) based on hardware capabilities. The bitmap size is
calculated by dividing total hardware entries by the number of available ports
for software allocation management.

> diff --git a/drivers/net/dsa/netc/netc_main.c b/drivers/net/dsa/netc/netc_main.c
> index fa7dd307ce130..d4475ad7ed6c2 100644
> --- a/drivers/net/dsa/netc/netc_main.c
> +++ b/drivers/net/dsa/netc/netc_main.c
[ ... ]
> +static u32 netc_num_available_ports(struct netc_switch *priv)
> +{
> +	struct dsa_port *dp;
> +	u32 num_ports = 0;
> +
> +	dsa_switch_for_each_available_port(dp, priv->ds)
> +		num_ports++;
> +
> +	return num_ports;
> +}
> +
> +static int netc_init_ntmp_bitmap_sizes(struct netc_switch *priv)
> +{
> +	u32 num_ports = netc_num_available_ports(priv);
> +	struct netc_switch_regs *regs = &priv->regs;
> +	struct ntmp_user *ntmp = &priv->ntmp;
> +	u32 val;
> +
> +	if (!num_ports)
> +		return -EINVAL;
> +
> +	val = netc_base_rd(regs, NETC_ETTCAPR);
> +	ntmp->ett_bitmap_size = NETC_GET_NUM_ENTRIES(val) / num_ports;

[Severity: High]
Does calculating the group size based on the number of available ports rather
than the total physical hardware ports create an issue if some ports are
disabled?

If a board has disabled ports, this will produce a smaller group size. Since
the hardware inherently uses the uncompressed physical port ID to index into
the ETT/ECT during packet egress, could this lead to cross-group table
overlap?

For example, if port 1 is disabled on a 4-port switch, the group size becomes
3. If software programs port 3's configuration at base + 2, but the hardware
accesses base + 3 on packet egress, won't this overlap with the start of the
next group, potentially applying incorrect egress treatments?

> +	if (!ntmp->ett_bitmap_size)
> +		return -EINVAL;
> +
> +	val = netc_base_rd(regs, NETC_ECTCAPR);
> +	ntmp->ect_bitmap_size = NETC_GET_NUM_ENTRIES(val) / num_ports;

[Severity: High]
Similarly, will using the available ports count here cause the same cross-group
boundary overlap issue for the ECT group size during hardware accesses?

> +	if (!ntmp->ect_bitmap_size)
> +		return -EINVAL;
> +
> +	return 0;
> +}

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260611021458.2629145-1-wei.fang@oss.nxp.com?part=5