* Re: Remaining kautobuild errors with mainline
2008-09-09 9:33 ` Russell King - ARM Linux
@ 2008-09-09 15:22 ` Tony Lindgren
2008-09-11 10:30 ` Pandita, Vikram
1 sibling, 0 replies; 4+ messages in thread
From: Tony Lindgren @ 2008-09-09 15:22 UTC (permalink / raw)
To: Russell King - ARM Linux; +Cc: linux-omap
* Russell King - ARM Linux <linux@arm.linux.org.uk> [080909 02:34]:
> On Tue, Sep 09, 2008 at 08:59:29AM +0100, Russell King - ARM Linux wrote:
> > As of last nights merge of the ARM tree, the only remaining build error
> > is:
> >
> > arch/arm/mach-omap1/built-in.o: In function `sx1_mmc_init':
> > board-sx1-mmc.c:(.init.text+0xd64): undefined reference to `omap_set_mmc_info'
> > arch/arm/mach-omap1/built-in.o: In function `h2_mmc_init':
> > arch/arm/mach-omap1/board-h2-mmc.c:98: undefined reference to `omap_set_mmc_info'
> >
> > This seems to be because of arch/arm/plat-omap/devices.c not being up to
> > date; it seems to pass a struct omap_mmc_conf to the OMAP MMC driver,
> > but the MMC driver expects a struct omap_mmc_platform_data instead.
> >
> > Can I lift just the MMC changes for arch/arm/plat-omap/devices.c from
> > the OMAP tree to fix this, or is there something more required for this?
Sure, go for it!
> I couldn't pull the commits from the omapzoom tree because there seem to
> be differences between mainline and omapzoom which aren't identifyable in
> the omapzoom history.
Yeah some patches just do not apply directly because of layers of fixes,
so redoing and splitting is needed.
> So... what I have is:
>
> From c37e0062150668dd260a6ce0664697f0fbcf2239 Mon Sep 17 00:00:00 2001
> From: Russell King <rmk@dyn-67.arm.linux.org.uk>
> Date: Tue, 9 Sep 2008 10:16:22 +0100
> Subject: [PATCH] [ARM] OMAP: Fix MMC device data
>
> OMAPs MMC device data was passing the wrong structure via the platform
> device. Moreover, a missing function means that both sx1_defconfig
> and omap_h2_1610_defconfig builds failed with
>
> undefined reference to `omap_set_mmc_info'
>
> errors. Fix this by updating the MMC support from the omapzoom tree.
>
> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Acked-by: Tony Lindgren <tony@atomide.com>
> ---
> arch/arm/plat-omap/devices.c | 88 +++++++++++++++++++++++++++++++++++-------
> 1 files changed, 74 insertions(+), 14 deletions(-)
>
> diff --git a/arch/arm/plat-omap/devices.c b/arch/arm/plat-omap/devices.c
> index 187e3d8..370c2cf 100644
> --- a/arch/arm/plat-omap/devices.c
> +++ b/arch/arm/plat-omap/devices.c
> @@ -21,6 +21,7 @@
>
> #include <mach/tc.h>
> #include <mach/board.h>
> +#include <mach/mmc.h>
> #include <mach/mux.h>
> #include <mach/gpio.h>
> #include <mach/menelaus.h>
> @@ -194,25 +195,38 @@ void omap_mcbsp_register_board_cfg(struct omap_mcbsp_platform_data *config,
>
> /*-------------------------------------------------------------------------*/
>
> -#if defined(CONFIG_MMC_OMAP) || defined(CONFIG_MMC_OMAP_MODULE)
> +#if defined(CONFIG_MMC_OMAP) || defined(CONFIG_MMC_OMAP_MODULE) || \
> + defined(CONFIG_MMC_OMAP_HS) || defined(CONFIG_MMC_OMAP_HS_MODULE)
>
> -#ifdef CONFIG_ARCH_OMAP24XX
> +#if defined(CONFIG_ARCH_OMAP24XX) || defined(CONFIG_ARCH_OMAP34XX)
> #define OMAP_MMC1_BASE 0x4809c000
> -#define OMAP_MMC1_INT INT_24XX_MMC_IRQ
> +#define OMAP_MMC1_END (OMAP_MMC1_BASE + 0x1fc)
> +#define OMAP_MMC1_INT INT_24XX_MMC_IRQ
> +
> +#define OMAP_MMC2_BASE 0x480b4000
> +#define OMAP_MMC2_END (OMAP_MMC2_BASE + 0x1fc)
> +#define OMAP_MMC2_INT INT_24XX_MMC2_IRQ
> +
> #else
> +
> #define OMAP_MMC1_BASE 0xfffb7800
> +#define OMAP_MMC1_END (OMAP_MMC1_BASE + 0x7f)
> #define OMAP_MMC1_INT INT_MMC
> -#endif
> +
> #define OMAP_MMC2_BASE 0xfffb7c00 /* omap16xx only */
> +#define OMAP_MMC2_END (OMAP_MMC2_BASE + 0x7f)
> +#define OMAP_MMC2_INT INT_1610_MMC2
>
> -static struct omap_mmc_conf mmc1_conf;
> +#endif
> +
> +static struct omap_mmc_platform_data mmc1_data;
>
> static u64 mmc1_dmamask = 0xffffffff;
>
> static struct resource mmc1_resources[] = {
> {
> .start = OMAP_MMC1_BASE,
> - .end = OMAP_MMC1_BASE + 0x7f,
> + .end = OMAP_MMC1_END,
> .flags = IORESOURCE_MEM,
> },
> {
> @@ -226,26 +240,27 @@ static struct platform_device mmc_omap_device1 = {
> .id = 1,
> .dev = {
> .dma_mask = &mmc1_dmamask,
> - .platform_data = &mmc1_conf,
> + .platform_data = &mmc1_data,
> },
> .num_resources = ARRAY_SIZE(mmc1_resources),
> .resource = mmc1_resources,
> };
>
> -#ifdef CONFIG_ARCH_OMAP16XX
> +#if defined(CONFIG_ARCH_OMAP16XX) || defined(CONFIG_ARCH_OMAP2430) || \
> + defined(CONFIG_ARCH_OMAP34XX)
>
> -static struct omap_mmc_conf mmc2_conf;
> +static struct omap_mmc_platform_data mmc2_data;
>
> static u64 mmc2_dmamask = 0xffffffff;
>
> static struct resource mmc2_resources[] = {
> {
> .start = OMAP_MMC2_BASE,
> - .end = OMAP_MMC2_BASE + 0x7f,
> + .end = OMAP_MMC2_END,
> .flags = IORESOURCE_MEM,
> },
> {
> - .start = INT_1610_MMC2,
> + .start = OMAP_MMC2_INT,
> .flags = IORESOURCE_IRQ,
> },
> };
> @@ -255,7 +270,7 @@ static struct platform_device mmc_omap_device2 = {
> .id = 2,
> .dev = {
> .dma_mask = &mmc2_dmamask,
> - .platform_data = &mmc2_conf,
> + .platform_data = &mmc2_data,
> },
> .num_resources = ARRAY_SIZE(mmc2_resources),
> .resource = mmc2_resources,
> @@ -274,6 +289,19 @@ static void __init omap_init_mmc(void)
>
> /* block 1 is always available and has just one pinout option */
> mmc = &mmc_conf->mmc[0];
> +
> + if (cpu_is_omap2430() || cpu_is_omap34xx()) {
> + if (mmc->enabled)
> + (void) platform_device_register(&mmc_omap_device1);
> +
> +#if defined(CONFIG_ARCH_OMAP2430) || defined(CONFIG_ARCH_OMAP34XX)
> + mmc = &mmc_conf->mmc[1];
> + if (mmc->enabled)
> + (void) platform_device_register(&mmc_omap_device2);
> +#endif
> + return;
> + }
> +
> if (mmc->enabled) {
> if (cpu_is_omap24xx()) {
> omap_cfg_reg(H18_24XX_MMC_CMD);
> @@ -308,7 +336,20 @@ static void __init omap_init_mmc(void)
> omap_cfg_reg(MMC_DAT3);
> }
> }
> - mmc1_conf = *mmc;
> +#if defined(CONFIG_ARCH_OMAP2420)
> + if (mmc->internal_clock) {
> + /*
> + * Use internal loop-back in MMC/SDIO
> + * Module Input Clock selection
> + */
> + if (cpu_is_omap24xx()) {
> + u32 v = omap_ctrl_readl(OMAP2_CONTROL_DEVCONF0);
> + v |= (1 << 24); /* not used in 243x */
> + omap_ctrl_writel(v, OMAP2_CONTROL_DEVCONF0);
> + }
> + }
> +#endif
> + mmc1_data.conf = *mmc;
> (void) platform_device_register(&mmc_omap_device1);
> }
>
> @@ -337,14 +378,33 @@ static void __init omap_init_mmc(void)
> if (cpu_is_omap1710())
> omap_writel(omap_readl(MOD_CONF_CTRL_1) | (1 << 24),
> MOD_CONF_CTRL_1);
> - mmc2_conf = *mmc;
> + mmc2_data.conf = *mmc;
> (void) platform_device_register(&mmc_omap_device2);
> }
> #endif
> return;
> }
> +
> +void omap_set_mmc_info(int host, const struct omap_mmc_platform_data *info)
> +{
> + switch (host) {
> + case 1:
> + mmc1_data = *info;
> + break;
> +#if defined(CONFIG_ARCH_OMAP16XX) || defined(CONFIG_ARCH_OMAP2430) || \
> + defined(CONFIG_ARCH_OMAP34XX)
> + case 2:
> + mmc2_data = *info;
> + break;
> +#endif
> + default:
> + BUG();
> + }
> +}
> +
> #else
> static inline void omap_init_mmc(void) {}
> +void omap_set_mmc_info(int host, const struct omap_mmc_platform_data *info) {}
> #endif
>
> /*-------------------------------------------------------------------------*/
> --
> 1.5.4.5
>
^ permalink raw reply [flat|nested] 4+ messages in thread* RE: Remaining kautobuild errors with mainline
2008-09-09 9:33 ` Russell King - ARM Linux
2008-09-09 15:22 ` Tony Lindgren
@ 2008-09-11 10:30 ` Pandita, Vikram
1 sibling, 0 replies; 4+ messages in thread
From: Pandita, Vikram @ 2008-09-11 10:30 UTC (permalink / raw)
To: Russell King - ARM Linux, Tony Lindgren; +Cc: linux-omap@vger.kernel.org
Russell
> -----Original Message-----
> From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-owner@vger.kernel.org] On Behalf Of Russell
> On Tue, Sep 09, 2008 at 08:59:29AM +0100, Russell King - ARM Linux wrote:
> > As of last nights merge of the ARM tree, the only remaining build error
> > is:
> >
> > arch/arm/mach-omap1/built-in.o: In function `sx1_mmc_init':
> > board-sx1-mmc.c:(.init.text+0xd64): undefined reference to `omap_set_mmc_info'
> > arch/arm/mach-omap1/built-in.o: In function `h2_mmc_init':
> > arch/arm/mach-omap1/board-h2-mmc.c:98: undefined reference to `omap_set_mmc_info'
> >
> > This seems to be because of arch/arm/plat-omap/devices.c not being up to
> > date; it seems to pass a struct omap_mmc_conf to the OMAP MMC driver,
> > but the MMC driver expects a struct omap_mmc_platform_data instead.
> >
> > Can I lift just the MMC changes for arch/arm/plat-omap/devices.c from
> > the OMAP tree to fix this, or is there something more required for this?
>
> I couldn't pull the commits from the omapzoom tree because there seem to
> be differences between mainline and omapzoom which aren't identifyable in
> the omapzoom history.
For omapzoom git the current strategy is to base the tree of l-o tree.
Every two weeks I rebase of the l-o tree and apply TI-PM and TI-extra-functionality on top of l-o tree and that's what the omapzoom tree is today.
I believe this might not be the smartest way, but with parallel development happening on l-o as well as omapzoom, sometimes by same developers, either tree has to be in sync so that none of the fixes is lost.
Other way for zoom tree would be to diverge by taking linus upmerges and selectively take l-o bug-fixes. The issue here would be the trees would diverge over time.
>
> So... what I have is:
>
> From c37e0062150668dd260a6ce0664697f0fbcf2239 Mon Sep 17 00:00:00 2001
> From: Russell King <rmk@dyn-67.arm.linux.org.uk>
> Date: Tue, 9 Sep 2008 10:16:22 +0100
> Subject: [PATCH] [ARM] OMAP: Fix MMC device data
>
> OMAPs MMC device data was passing the wrong structure via the platform
> device. Moreover, a missing function means that both sx1_defconfig
> and omap_h2_1610_defconfig builds failed with
>
> undefined reference to `omap_set_mmc_info'
>
> errors. Fix this by updating the MMC support from the omapzoom tree.
>
> Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
> ---
> arch/arm/plat-omap/devices.c | 88 +++++++++++++++++++++++++++++++++++-------
> 1 files changed, 74 insertions(+), 14 deletions(-)
>
^ permalink raw reply [flat|nested] 4+ messages in thread