From mboxrd@z Thu Jan 1 00:00:00 1970 From: computersforpeace@gmail.com (Brian Norris) Date: Tue, 21 Oct 2014 23:10:57 -0700 Subject: [PATCH] ARM: don't set unused name in struct flash_platform_data In-Reply-To: References: <1411993853-6309-1-git-send-email-zajec5@gmail.com> <20141022055326.GB16128@brian-ubuntu> Message-ID: <20141022061057.GC16128@brian-ubuntu> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Oct 22, 2014 at 08:03:45AM +0200, Rafa? Mi?ecki wrote: > On 22 October 2014 07:53, Brian Norris wrote: > > On Mon, Sep 29, 2014 at 02:30:53PM +0200, Rafa? Mi?ecki wrote: > >> Loading correct SPI driver (m25p80) is handled using modalias from the > >> struct spi_board_info. There is no point of setting name in the > >> platform_data, m25p80 ignores it anyway. > > > > Wait, is 'name' actually ignored? It looks to me like it sets the MTD > > name. See the comments in include/linux/spi/flash.h and > > arch/arm/include/asm/mach/flash.h (BTW, why do we have two definitions > > for this??): > > > > @name: optional flash device name (eg, as used with mtdparts=) > > > > And I think it's used for exactly that in m25p80.c: > > > > if (data && data->name) > > flash->mtd.name = data->name; > > And very few lines later in the m25p80.c you have this: > if (data && data->type) > id = spi_nor_match_id(data->type); How is that relevant to the value of 'flash->mtd.name'? > Code I touched in my patch was using both: name and type. So what you > got it "type" took a precedence and "name" value was ignored. It seems > like ppl used to use "name" to trigger "m25p80" probe, which isn't > needed (it's done by SPI layer). > > Also see my comment about data->type usage I added in: > http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=32f1b7c8352fd33d41bcec3cfb054ccdcfd40a42 But I'm not talking about probing / driver matching. The 'name' field is used for assigning the MTD name deterministically. This name is used for things like 'mtdparts'. Brian