> -----Original Message----- > From: Paul Walmsley [mailto:[email protected]] > Sent: Wednesday, February 17, 2010 11:02 AM > To: Shilimkar, Santosh > Cc: [email protected]; [email protected]; [email protected]; > [email protected]; > [email protected]; Cousson, Benoit > Subject: RE: [PATCH 05/24] omap4: Add needed IRQ line into irqs.h > > On Wed, 17 Feb 2010, Shilimkar, Santosh wrote: > > > > -----Original Message----- > > > From: Paul Walmsley [mailto:[email protected]] > > > Sent: Wednesday, February 17, 2010 3:39 AM > > > To: Shilimkar, Santosh > > > Cc: [email protected]; [email protected]; [email protected]; > > > [email protected]; > > > [email protected] > > > Subject: Re: [PATCH 05/24] omap4: Add needed IRQ line into irqs.h > > > > > > Hello Santosh, > > > > > > On Tue, 16 Feb 2010, Santosh Shilimkar wrote: > > > > > > > This patch adds DSS, keyboard and ducati irq > > > > entries in irqs.h > > > > > > > > Ones omap4 hwmod is pushed on mainlie, all the > > > > irqs lines data would be availabe. > > > > > > Instead of this patch, why not just replace the current file with the one > > > from the OMAP4 autogeneration scripts? I don't see why the IRQ macros > > > should depend on hwmod availability, and it would be better to have the > > > correct IRQ data appear in one patch, rather than having to review and > > > approve several smaller incremental patches over time. > > We plan to do this but today the autogenrated file names and one in the > > kernel code is different. > > File names or macro names? Macros ofcourse > > > Autogenrated headers are kept in alignment > > with TRM naming as you know. > > e.g > > --- autogenaretd > > #define OMAP44XX_IRQ_MCBSP2 (22 + OMAP44XX_IRQ_GIC_START) > > > > --existing kernel > > #define INT_44XX_MCBSP2_IRQ (22 + IRQ_GIC_START) > > > > > > So instead of touching every driver file and changing it only for name > > changes , > > what we are doing > > 1. Created local 44xx_dma.h and 44xx_irq.h mach-omap2 directory > > 2. All the drivers getting converted in hwmods would use defines from above > > headers > > 3. At one point of time all the drivers would use defines from above > > headers and then > > from omap4 point of view there won't be dependency on the irqs.h and > > dma.h > > > > While this is ongoing, we would like the mainline kernel to have most of the > > drivers supported. And Hence such a patch. > > I grepped through the code base for any instances of INT_44XX: > > [p...@twilight current]$ fgrep -r INT_44XX . | fgrep -v > ./arch/arm/plat-omap/include/plat/irqs.h > ./arch/arm/mach-omap2/devices.c: .start = > INT_44XX_MAIL_U0_MPU, > ./arch/arm/mach-omap2/devices.c: irq = INT_44XX_MMC4_IRQ; > ./arch/arm/mach-omap2/devices.c: irq = INT_44XX_MMC5_IRQ; > ./arch/arm/mach-omap2/timer-mpu.c: evt->irq = INT_44XX_LOCALTIMER_IRQ; > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x4a318000, .irq = > INT_44XX_GPTIMER1 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x48032000, .irq = > INT_44XX_GPTIMER2 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x48034000, .irq = > INT_44XX_GPTIMER3 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x48036000, .irq = > INT_44XX_GPTIMER4 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x40138000, .irq = > INT_44XX_GPTIMER5 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x4013a000, .irq = > INT_44XX_GPTIMER6 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x4013a000, .irq = > INT_44XX_GPTIMER7 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x4013e000, .irq = > INT_44XX_GPTIMER8 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x4803e000, .irq = > INT_44XX_GPTIMER9 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x48086000, .irq = > INT_44XX_GPTIMER10 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x48088000, .irq = > INT_44XX_GPTIMER11 }, > ./arch/arm/plat-omap/dmtimer.c: { .phys_base = 0x4a320000, .irq = > INT_44XX_GPTIMER12 }, > ./arch/arm/plat-omap/gpio.c: { OMAP44XX_GPIO1_BASE, NULL, > INT_44XX_GPIO_BANK1, IH_GPIO_BASE, > ./arch/arm/plat-omap/gpio.c: { OMAP44XX_GPIO2_BASE, NULL, > INT_44XX_GPIO_BANK2, IH_GPIO_BASE + 32, > ./arch/arm/plat-omap/gpio.c: { OMAP44XX_GPIO3_BASE, NULL, > INT_44XX_GPIO_BANK3, IH_GPIO_BASE + 64, > ./arch/arm/plat-omap/gpio.c: { OMAP44XX_GPIO4_BASE, NULL, > INT_44XX_GPIO_BANK4, IH_GPIO_BASE + 96, > ./arch/arm/plat-omap/gpio.c: { OMAP44XX_GPIO5_BASE, NULL, > INT_44XX_GPIO_BANK5, IH_GPIO_BASE + 128, > ./arch/arm/plat-omap/gpio.c: { OMAP44XX_GPIO6_BASE, NULL, > INT_44XX_GPIO_BANK6, IH_GPIO_BASE + 160, > ./arch/arm/plat-omap/dma.c: irq = INT_44XX_SDMA_IRQ0; > [p...@twilight current]$ > > > There are only five files to change. None of them are device > driver files, so they can all be merged through linux-omap. The > changes could easily be done in a single patch. So I see no > reason to not use the autogenerated IRQ file. > I think on you tree, you won't have DSS, AUDIO-ALSA, BRIDGE, USB etc. All these drivers needs to be updated as well
> > > The same reasoning should apply the DMA data also, I'd think. > > DMA file will get replaced soon with auto generated since naming > > is aligned there. > > > - Paul -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
