From: ben-linux@fluff.org (Ben Dooks) Date: Sun, 30 May 2010 05:51:51 +0100 Subject: [PATCH v2 1/5] ARM: SAMSUNG: Add keypad device support In-Reply-To: References: <1275188784-23395-1-git-send-email-jy0922.shim@samsung.com> <201005300542.37715.marek.vasut@gmail.com> Message-ID: <20100530045151.GE4720@trinity.fluff.org> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Sun, May 30, 2010 at 01:46:03PM +0900, Jassi Brar wrote: > On Sun, May 30, 2010 at 12:42 PM, Marek Vasut wrote: > > Dne Ne 30. kv??tna 2010 05:06:20 Joonyoung Shim napsal(a): > >> +struct samsung_kp_platdata { > >> + ? ? const struct matrix_keymap_data *keymap_data; > >> + ? ? unsigned int ? ? ? ? ? ?rows; > >> + ? ? unsigned int ? ? ? ? ? ?cols; > >> + ? ? unsigned int ? ? ? ? ? ?rep; > > > > I don't know, maybe using uint32_t here? On ARM, it doesn't matter so far as int > > will be always 32bit, but maybe we should just type the variables well ? > > I thought int was 32bits on all archs No, the C standard doesn't make any guarantees about the size of types, it is up to the implementation of the compiler. If I rember correctly the only guarantee in the lanugage definition is that char->short->int->long be that the next up the line be at-least as a big as the one before. For this, unsigned int should be fine. unsigned short would probably be fine too. I'm not going to worry too much about this. -- Ben Q: What's a light-year? A: One-third less calories than a regular year.