From: cbouatmailru@gmail.com (Anton Vorontsov) Date: Thu, 30 Sep 2010 13:42:02 +0400 Subject: [PATCH 4/6] mmc: sdhci_pltfm: pass more data on custom init-call In-Reply-To: <19620.19198.88841.623775@ipc1.ka-ro> References: <1285790884-3516-1-git-send-email-w.sang@pengutronix.de> <1285790884-3516-5-git-send-email-w.sang@pengutronix.de> <19620.19198.88841.623775@ipc1.ka-ro> Message-ID: <20100930094202.GA581@oksana.dev.rtsoft.ru> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thu, Sep 30, 2010 at 10:31:58AM +0200, Lothar Wa?mann wrote: > > The custom init call may need more data to perform its job, so we pass > > it a pointer to pdata, too. Also, always use the platform_id specific > > data even if platform_data is present. Doing that, platform_data can > > additionally be parsed by init() for board-specific information (via > > sdhci->mmc->parent). > > > Why not a pointer to the platform_device itself? Then the callback > function would have even more information (e.g. platform_device ID) > and could easily get at the platform_data too. dev->platform_data may not be there. There are two ways to pass the "platform data" to the sdhci-pltfm driver: a) via dev->platform_data b) via platform_device_id->driver_data Technically, we can do something like this in the probe(): if (!pdev->dev.platform_data) pdev->dev.platform_data = platform_device_id->driver_data; pdata->init(dev); Note, in this case we must also NULLify platform_data it at remove(), as platform_device_release() will try to free statically allocated memory. Thanks, -- Anton Vorontsov email: cbouatmailru at gmail.com irc://irc.freenode.net/bd2