[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Mon, 11 Nov 2013 12:26:38 +0100
From: Michal Nazarewicz <mina86@...a86.com>
To: Herbert Xu <herbert@...dor.apana.org.au>
Cc: "David S. Miller" <davem@...emloft.net>,
linux-crypto@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH RESEND] crypto: make sure *blkcipher_walk_init properly initialises walk
On Mon, Nov 11 2013, Herbert Xu <herbert@...dor.apana.org.au> wrote:
> Nack. The field flags is used as a bit-field and all bits other
> than those initialised that you see are used internally by the
> walker function and will be initialised on demand.
>
> Please do not just rely on tools such as coverity and actually
> read the code when submitting patches.
I have read the code which is why I concluded that it is safe to replace
the bit operations with a simple assignment. Since, as you described,
all other bit fields are initialised on demand anyway, there is no harm
in setting them to zero here. Especially since I see no advantages of
the current approach, but instead see two disadvantages: longer machine
code (load-modiy-store vs. store) and confusion of tools such as
Coverity.
But of course, if you want it as it is, I won't be bothering you.
--
Best regards, _ _
.o. | Liege of Serenely Enlightened Majesty of o' \,=./ `o
..o | Computer Science, Michał “mina86” Nazarewicz (o o)
ooo +--<mpn@...gle.com>--<xmpp:mina86@...ber.org>--ooO--(_)--Ooo--
Download attachment "signature.asc" of type "application/pgp-signature" (836 bytes)
Powered by blists - more mailing lists