From mboxrd@z Thu Jan 1 00:00:00 1970 From: Julia Lawall Date: Tue, 09 Oct 2012 12:51:33 +0000 Subject: Re: [PATCH 3/13] drivers/media/tuners/qt1010.c: use macros for i2c_msg initialization Message-Id: List-Id: References: <1349624323-15584-1-git-send-email-Julia.Lawall@lip6.fr> <1349624323-15584-5-git-send-email-Julia.Lawall@lip6.fr> <5071FA5D.30003@gmail.com> <50726110.5020901@gmail.com> <20121009141220.412c15c8@endymion.delvare> In-Reply-To: <20121009141220.412c15c8@endymion.delvare> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Jean Delvare Cc: Julia Lawall , Ryan Mallon , Antti Palosaari , kernel-janitors@vger.kernel.org, shubhrajyoti@ti.com, Mauro Carvalho Chehab , linux-media@vger.kernel.org, linux-kernel@vger.kernel.org On Tue, 9 Oct 2012, Jean Delvare wrote: > Hi Julia, > > On Mon, 8 Oct 2012 07:24:11 +0200 (CEST), Julia Lawall wrote: > > > Sorry, I mean either: > > > > > > I2C_MSG_WRITE(priv->cfg->i2c_address, ®, sizeof(reg)), > > > I2C_MSG_READ(priv->cfg->i2c_address, val, sizeof(*val)), > > > > Of course. Sorry for not having seen that. I can do that. > > Eek, no, you can't, not in the general case at least. sizeof(*val) will > return the size of the _first_ element of the destination buffer, which > has nothing to do with the length of that buffer (which in turn might > be rightfully longer than the read length for this specific message.) I was actually only going to do it when the size was 1 and the type was u8 *. But your other email suggests that converting to sizeof is just not a good idea at all. So I will drop that part of the rule. julia