On Tue, Jul 05, 2011 at 11:24:13AM -0600, Grant Likely wrote: > On Tue, Jul 5, 2011 at 10:56 AM, Sascha Hauer <[email protected]> wrote: > > On Tue, Jul 05, 2011 at 11:19:24PM +0800, Shawn Guo wrote: > >> Instead of assigning the gpio range based on pdev->id, the patch > >> makes changes to get the range from gpio core that is dynamically > >> allocated. > >> > >> As a result, the uses of pdev->id can be removed from the driver. > >> This will make dt migration of the driver easier. > >> > >> Signed-off-by: Shawn Guo <[email protected]> > >> Cc: Grant Likely <[email protected]> > >> Cc: Sascha Hauer <[email protected]> > >> --- > >> arch/arm/plat-mxc/include/mach/gpio.h | 13 +++++++++---- > >> arch/arm/plat-mxc/include/mach/irqs.h | 21 +++------------------ > >> drivers/gpio/gpio-mxc.c | 18 ++++++++++-------- > >> 3 files changed, 22 insertions(+), 30 deletions(-) > >> > >> diff --git a/arch/arm/plat-mxc/include/mach/gpio.h > >> b/arch/arm/plat-mxc/include/mach/gpio.h > >> index 31c820c..abdf5d7 100644 > >> --- a/arch/arm/plat-mxc/include/mach/gpio.h > >> +++ b/arch/arm/plat-mxc/include/mach/gpio.h > >> @@ -23,10 +23,15 @@ > >> #include <mach/hardware.h> > >> #include <asm-generic/gpio.h> > >> > >> - > >> -/* There's a off-by-one betweem the gpio bank number and the gpiochip */ > >> -/* range e.g. GPIO_1_5 is gpio 5 under linux */ > >> -#define IMX_GPIO_NR(bank, nr) (((bank) - 1) * 32 + (nr)) > >> +/* > >> + * There's a off-by-one betweem the gpio bank number and the gpiochip > >> + * range e.g. GPIO_1_5 is gpio 5 under linux. > >> + * > >> + * When gpio core allocates gpio range for a bank, it starts from the > >> + * end of the total range. That is to say, bank 0 will get a higher > >> + * gpio range than bank 1. > >> + */ > >> +#define IMX_GPIO_NR(bank, nr) (ARCH_NR_GPIOS - (bank) * 32 + (nr)) > > > > That is not a good idea. First of all not all boards use this macro. > > This could be fixed, but it is a no go to allocate the gpios dynamically > > and add a macro which makes assumptions on the range which gets > > allocated. > > Plus it makes the assumption that the imx gpio controllers will be the > first ones registered, and that the allocation scheme will not change > some time in the future. You don't actually need to change this > anyway since existing static platform_device registrations can still > use the pdev->id value. It is only the DT use case that should > dynamically assign the gpio number because all uses in that case can > use a DT gpio specifier instead of a hard coded number. > Yeah, right. Will drop all the changes in this patch but except the following lines.
-----8<------------- -/* these are ordered by size to support multi-SoC kernels */ -#if defined CONFIG_SOC_IMX53 -#define MXC_GPIO_IRQS (32 * 7) -#elif defined CONFIG_ARCH_MX2 -#define MXC_GPIO_IRQS (32 * 6) -#elif defined CONFIG_SOC_IMX50 -#define MXC_GPIO_IRQS (32 * 6) -#elif defined CONFIG_ARCH_MX1 -#define MXC_GPIO_IRQS (32 * 4) -#elif defined CONFIG_ARCH_MX25 -#define MXC_GPIO_IRQS (32 * 4) -#elif defined CONFIG_SOC_IMX51 -#define MXC_GPIO_IRQS (32 * 4) -#elif defined CONFIG_ARCH_MX3 -#define MXC_GPIO_IRQS (32 * 3) -#endif - /* * The next 16 interrupts are for board specific purposes. Since * the kernel can only run on one machine at a time, we can re-use * these. If you need more, increase MXC_BOARD_IRQS, but keep it * within sensible limits. */ -#define MXC_BOARD_IRQ_START (MXC_INTERNAL_IRQS + MXC_GPIO_IRQS) +#define MXC_BOARD_IRQ_START (MXC_INTERNAL_IRQS + ARCH_NR_GPIOS) --------------------- I have to keep above changes not only because it makes code a little bit cleaner, but also it's needed to get dt irq work, as dt code gets gpio range from gpio core which dynamically allocates number from ARCH_NR_GPIOS to 0. -- Regards, Shawn _______________________________________________ devicetree-discuss mailing list [email protected] https://lists.ozlabs.org/listinfo/devicetree-discuss
