netdev - RE: [PATCH net-next 10/11] net: stmmac: Introduce selftests support

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]

Date:   Thu, 9 May 2019 08:25:14 +0000
From:   Jose Abreu <Jose.Abreu@...opsys.com>
To:     Andrew Lunn <andrew@...n.ch>, Jose Abreu <Jose.Abreu@...opsys.com>
CC:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        Joao Pinto <Joao.Pinto@...opsys.com>,
        "David S . Miller" <davem@...emloft.net>,
        Giuseppe Cavallaro <peppe.cavallaro@...com>,
        Alexandre Torgue <alexandre.torgue@...com>
Subject: RE: [PATCH net-next 10/11] net: stmmac: Introduce selftests support

From: Andrew Lunn <andrew@...n.ch>
Date: Thu, May 09, 2019 at 03:23:30

> > +static int stmmac_test_eee(struct stmmac_priv *priv)
> > +{
> > +	struct stmmac_extra_stats *initial, *final;
> > +	int timeout = 100;
> > +	int ret;
> > +
> > +	ret = stmmac_test_loopback(priv);
> > +	if (ret)
> > +		goto out_free_final;
> > +
> > +	/* We have no traffic in the line so, sooner or later it will go LPI */
> > +	while (--timeout) {
> > +		memcpy(final, &priv->xstats, sizeof(*final));
> > +
> > +		if (final->irq_tx_path_in_lpi_mode_n >
> > +		    initial->irq_tx_path_in_lpi_mode_n)
> > +			break;
> > +		msleep(100);
> > +	}
> > +
> > +	if (!timeout) {
> > +		ret = -ETIMEDOUT;
> > +		goto out_free_final;
> > +	}
> 
> Retries would be a better name than timeout.

Ok.

> 
> Also, 100 * 100 ms seems like a long time.

Ah, yeah. I will adjust to 0.5 or maybe 1 sec max.

> 
> > +static int stmmac_filter_check(struct stmmac_priv *priv)
> > +{
> > +	if (!(priv->dev->flags & IFF_PROMISC))
> > +		return 0;
> > +
> > +	netdev_warn(priv->dev, "Test can't be run in promiscuous mode!\n");
> > +	return 1;
> 
> Maybe return EOPNOTSUPP here,

Ok.

> 
> > +}
> > +
> > +static int stmmac_test_hfilt(struct stmmac_priv *priv)
> > +{
> > +	unsigned char gd_addr[ETH_ALEN] = {0x01, 0x0c, 0xcd, 0x04, 0x00, 0x00};
> > +	unsigned char bd_addr[ETH_ALEN] = {0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b};
> 
> What does gd and bd mean?

Good and Bad :D

> 
> > +	struct stmmac_packet_attrs attr = { };
> > +	int ret;
> > +
> > +	if (stmmac_filter_check(priv))
> > +		return -EOPNOTSUPP;
> 
> and just return the error code from the call.
> 
> > +
> > +	ret = dev_mc_add(priv->dev, gd_addr);
> > +	if (ret)
> > +		return ret;
> > +
> > +	attr.dst = gd_addr;
> > +
> > +	/* Shall receive packet */
> > +	ret = __stmmac_test_loopback(priv, &attr);
> > +	if (ret)
> > +		goto cleanup;
> > +
> > +	attr.dst = bd_addr;
> > +
> > +	/* Shall NOT receive packet */
> > +	ret = __stmmac_test_loopback(priv, &attr);
> > +	ret = !ret;
> 
> What is this test testing? gd is a multicast, where as bd is not.  I
> expect the hardware treats multicast different to unicast. So it would
> make more sense to test two different multicast addresses, one which
> has been added via dev_mc_addr, and one that has not?

Hmm, yeah makes sense. I will adjust.

> 
> > +
> > +cleanup:
> > +	dev_mc_del(priv->dev, gd_addr);
> > +	return ret;
> > +}
> > +
> > +static int stmmac_test_pfilt(struct stmmac_priv *priv)
> > +{
> > +	unsigned char gd_addr[ETH_ALEN] = {0x01, 0x02, 0x03, 0x04, 0x05, 0x06};
> > +	unsigned char bd_addr[ETH_ALEN] = {0x06, 0x07, 0x08, 0x09, 0x0a, 0x0b};
> > +	struct stmmac_packet_attrs attr = { };
> > +	int ret;
> > +
> > +	if (stmmac_filter_check(priv))
> > +		return -EOPNOTSUPP;
> > +
> > +	ret = dev_uc_add(priv->dev, gd_addr);
> > +	if (ret)
> > +		return ret;
> > +
> > +	attr.dst = gd_addr;
> > +
> > +	/* Shall receive packet */
> > +	ret = __stmmac_test_loopback(priv, &attr);
> > +	if (ret)
> > +		goto cleanup;
> 
> gb is a multicast address. Does dev_uc_add() return an error? If it
> does not we should not expect it to actually work, since a multicast
> address should not match a unicast address?

It doesn't return an error and it does calls the set_filter callback in 
netdev. I will adjust to use unicast address.

> You also seem to be missing a test for adding a unicast address via
> dev_uc_add() and receiving packets for that address, but not receiving
> multicast packets.

Hmm, what if interface was already configured to receive Multicast before 
running the tests ?

> 
> > +static const struct stmmac_test {
> > +	char name[ETH_GSTRING_LEN];
> > +	int lb;
> > +	int (*fn)(struct stmmac_priv *priv);
> > +} stmmac_selftests[] = {
> > +	{
> > +		.name = "MAC Loopback         ",
> > +		.lb = STMMAC_LOOPBACK_MAC,
> > +		.fn = stmmac_test_loopback,
> 
> stmmac_test_mac_loopback might be a better name.

Ok.

Thanks for the review!

Thanks,
Jose Miguel Abreu