linux-kernel - Re: [PATCH v3 2/2] regmap: spi: Check raw_[read|write] against max message size

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]

Message-ID: <20210915125140.GE5389@sirena.org.uk>
Date:   Wed, 15 Sep 2021 13:51:41 +0100
From:   Mark Brown <broonie@...nel.org>
To:     Lucas Tanure <tanureal@...nsource.cirrus.com>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "Rafael J . Wysocki" <rafael@...nel.org>,
        linux-kernel@...r.kernel.org, patches@...nsource.cirrus.com
Subject: Re: [PATCH v3 2/2] regmap: spi: Check raw_[read|write] against max
 message size

On Wed, Sep 15, 2021 at 01:09:51PM +0100, Lucas Tanure wrote:
> regmap-spi will split data and address between two transfers
> in the same message, so max_[read|write] must include space
> for the address and padding
> 
> Signed-off-by: Lucas Tanure <tanureal@...nsource.cirrus.com>
> ---
> 
> Changes v3:
> New series

The changelog in patch 1 claimed that v2 was the first version of a new
series.  Which is it?  If you're going to include an inter-version
changelog it should be meaningful, just writing something for the sake
of providing a changelog tends to only cause confusion.

> +		if (spi_max_message_size(spi) != SIZE_MAX)
> +			bus->max_combined_rw = spi_max_message_size(spi);
> +

Why is this conditional?  The whole point in returning SIZE_MAX is that
users don't need to care if there's a limit or not, any client logic
checking sizes will always work.

> +			if (map->max_raw_read  < map->format.buf_size ||
> +			    map->max_raw_write < map->format.buf_size) {
> +				ret = -EINVAL;
> +				goto err_hwlock;
> +			}
> +		}

This is at best an assert that the logic in _init() is self-consistent
so just silently returning an error doesn't seem like the right thing to
do, it's going to be super obscure for users.

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)