From mboxrd@z Thu Jan 1 00:00:00 1970 Message-ID: <539AACB9.5080105@ti.com> Date: Fri, 13 Jun 2014 10:48:09 +0300 From: Roger Quadros MIME-Version: 1.0 To: Tony Lindgren Subject: Re: [PATCH 15/36] ARM: OMAP2+: gpmc: Allow drivers to query GPMC_CLK period References: <1402477001-31132-1-git-send-email-rogerq@ti.com> <1402477001-31132-16-git-send-email-rogerq@ti.com> <20140613072645.GN17845@atomide.com> In-Reply-To: <20140613072645.GN17845@atomide.com> Content-Type: text/plain; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit Cc: devicetree@vger.kernel.org, linux-omap@vger.kernel.org, nsekhar@ti.com, linux-kernel@vger.kernel.org, kyungmin.park@samsung.com, linux-mtd@lists.infradead.org, pekon@ti.com, ezequiel.garcia@free-electrons.com, javier@dowhile0.org, computersforpeace@gmail.com, dwmw2@infradead.org List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On 06/13/2014 10:26 AM, Tony Lindgren wrote: > * Roger Quadros [140611 01:59]: >> GPMC_CLK is the external clock output pin that is used for syncronous >> accesses. >> >> Device drivers need to know the fastest possible GPMC_CLK period in order >> to calculate the most optimal device timings. Add the function >> omap_gpmc_get_clk_period() to allow drivers to get the nearset possible >> (equal to or greater than) GPMC_CLK period given the minimum >> clock period supported by the attached device. >> >> This is especially needed by the onenand driver as it calculates >> device timings on the fly for various onenand speed grades. > > Here too this should probably still be done by the gpmc to driver > glue layer, not by the actual driver that shoud be Linux generic. Well, this is only needed by the omap-onenand driver to perform the timing calculations at run-time. Other option is to model the GPMC_CLK (external) as a clock and request the rate using the clock framework. But since this and the retime() is only used by 2 OMAP specific drivers, I'm not sure if it is worth the effort. cheers, -roger