linux-kernel - Re: [PATCH net 4/9] net: stmmac: selftests: Must remove UC/MC addresses to prevent false positives

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] [day] [month] [year] [list]

Message-Id: <20191030.145234.1629187794527849559.davem@davemloft.net>
Date:   Wed, 30 Oct 2019 14:52:34 -0700 (PDT)
From:   David Miller <davem@...emloft.net>
To:     Jose.Abreu@...opsys.com
Cc:     netdev@...r.kernel.org, Joao.Pinto@...opsys.com,
        peppe.cavallaro@...com, alexandre.torgue@...com,
        mcoquelin.stm32@...il.com,
        linux-stm32@...md-mailman.stormreply.com,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH net 4/9] net: stmmac: selftests: Must remove UC/MC
 addresses to prevent false positives

From: Jose Abreu <Jose.Abreu@...opsys.com>
Date: Tue, 29 Oct 2019 15:14:48 +0100

> @@ -499,9 +501,18 @@ static int stmmac_test_hfilt(struct stmmac_priv *priv)
>  	if (netdev_mc_count(priv->dev) >= priv->hw->multicast_filter_bins)
>  		return -EOPNOTSUPP;

This test above...

> +	dummy_dev = alloc_etherdev(0);
> +	if (!dummy_dev)
> +		return -ENOMEM;
> +
> +	/* Remove all MC addresses */
> +	netdev_for_each_mc_addr(ha, priv->dev)
> +		dev_mc_add(dummy_dev, ha->addr);
> +	dev_mc_flush(priv->dev);

No longer makes any sense now that you're removing all of the MC
addresses.

Also I know it seems that it should be guaranteed that re-adding all of
the previously configured MC addresses should succeed.  But I am always
wary when I see error codes ignored like this.

This test makes destructure changes to the device's configuration,
perhaps in a non-restorable fashion if errors occur re-adding the MC
list entries.

Running a test should never even remotely introduce a change in the
device state like that.

I really don't like this, to be honest.  I'd hate to be the user who
had this somehow trigger on them and then have to diagnose it. :-/