[GIT PULL] Samsung devel-dt for v3.3

[Kukjin Kim]

Arnd Bergmann wrote:
> 
> On Saturday 24 December 2011, Kukjin Kim wrote:
> > Hi Arnd and Olof,
> >
> > Please pull the samsung-devel-dt for v3.3.
> > It includes support DT for Samsung EXYNOS4.
> 
> Pulled into next/dt, thanks a lot!
> 
Thanks.

> Unfortunately, this got two conflicts against "ARM: exynos4: convert to
> CONFIG_MULTI_IRQ_HANDLER" from Marc Zyngier, and I suspect my temporary
> resolution is incorrect:
> 
> --- a/arch/arm/mach-exynos/cpu.c
> +++ b/arch/arm/mach-exynos/cpu.c
> @@@ -206,6 -209,24 +208,13 @@@ void __init exynos4_init_clocks(int xta
>         exynos4_setup_clocks();
>   }
> 
>  -static void exynos4_gic_irq_fix_base(struct irq_data *d)
>  -{
>  -      struct gic_chip_data *gic_data = irq_data_get_irq_chip_data(d);
>  -
>  -      gic_data->cpu_base = S5P_VA_GIC_CPU +
>  -                          (gic_bank_offset * smp_processor_id());
>  -
>  -      gic_data->dist_base = S5P_VA_GIC_DIST +
>  -                          (gic_bank_offset * smp_processor_id());
>  -}
>  -
> + #ifdef CONFIG_OF
> + static const struct of_device_id exynos4_dt_irq_match[] = {
> +       { .compatible = "arm,cortex-a9-gic", .data = gic_of_init, },
> +       {},
> + };
> + #endif
> +
>   void __init exynos4_init_irq(void)
>   {
>         int irq;
> @@@ -213,8 -233,17 +222,13 @@@
> 
>         gic_bank_offset = soc_is_exynos4412() ? 0x4000 : 0x8000;
> 
> -       gic_init_bases(0, IRQ_PPI(0), S5P_VA_GIC_DIST, S5P_VA_GIC_CPU,
> gic_bank_offset);
> +       if (!of_have_populated_dt())
>  -              gic_init(0, IRQ_PPI(0), S5P_VA_GIC_DIST, S5P_VA_GIC_CPU);
> ++              gic_init_bases(0, IRQ_PPI(0), S5P_VA_GIC_DIST,
S5P_VA_GIC_CPU,
> gic_bank_offset);
> + #ifdef CONFIG_OF
> +       else
> +               of_irq_init(exynos4_dt_irq_match);
> + #endif
> 
>  -      gic_arch_extn.irq_eoi = exynos4_gic_irq_fix_base;
>  -      gic_arch_extn.irq_unmask = exynos4_gic_irq_fix_base;
>  -      gic_arch_extn.irq_mask = exynos4_gic_irq_fix_base;
>  -
>         for (irq = 0; irq < MAX_COMBINER_NR; irq++) {
> 
>                 combiner_init(irq, (void __iomem *)S5P_VA_COMBINER(irq),
> 
> 
Looks ok to me.

> The other conflict was a one-line change from 1fb3726cf3 "ARM: SAMSUNG:
> Move timer irq
> numbers to end of linux irq space" (from Thomas Abraham) in code that was
> removed
> by Marc's patch, so I kept removing that.
> 
OK.

> Maybe you can do a merge of your branch with Russell's devel-stable branch
> and tell me
> which conflict resolution you would like.
> 
OK, if any changes, let you know.

Thanks.

Best regards,
Kgene.
--
Kukjin Kim <kgene.kim at samsung.com>, Senior Engineer,
SW Solution Development Team, Samsung Electronics Co., Ltd.