* Re: [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support
2014-12-17 9:42 [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support Magnus Damm
@ 2014-12-17 13:49 ` Geert Uytterhoeven
2014-12-17 14:17 ` Magnus Damm
` (4 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Geert Uytterhoeven @ 2014-12-17 13:49 UTC (permalink / raw)
To: linux-sh
Hi Magnus,
On Wed, Dec 17, 2014 at 10:42 AM, Magnus Damm <magnus.damm@gmail.com> wrote:
> From: Magnus Damm <damm+renesas@opensource.se>
>
> Enable sh73a0 Multiplatform support for the generic sh73a0
> machine vector. No board support is enabled, and the board
> code for KZM9G DT Reference is left by itself.
>
> Signed-off-by: Magnus Damm <damm+renesas@opensource.se>
> --- 0001/arch/arm/mach-shmobile/Makefile
> +++ work/arch/arm/mach-shmobile/Makefile 2014-12-16 19:12:19.000000000 +0900
> @@ -7,7 +7,7 @@ obj-y := timer.o console.o
>
> # CPU objects
> obj-$(CONFIG_ARCH_SH7372) += setup-sh7372.o intc-sh7372.o pm-sh7372.o
> -obj-$(CONFIG_ARCH_SH73A0) += setup-sh73a0.o intc-sh73a0.o pm-sh73a0.o
> +obj-$(CONFIG_ARCH_SH73A0) += setup-sh73a0.o pm-sh73a0.o
> obj-$(CONFIG_ARCH_R8A73A4) += setup-r8a73a4.o
> obj-$(CONFIG_ARCH_R8A7740) += setup-r8a7740.o pm-r8a7740.o
> obj-$(CONFIG_ARCH_R8A7778) += setup-r8a7778.o
> @@ -65,7 +65,7 @@ obj-$(CONFIG_MACH_BOCKW) += board-bockw.
> obj-$(CONFIG_MACH_BOCKW_REFERENCE) += board-bockw-reference.o
> obj-$(CONFIG_MACH_MARZEN) += board-marzen.o
> obj-$(CONFIG_MACH_ARMADILLO800EVA) += board-armadillo800eva.o
> -obj-$(CONFIG_MACH_KZM9G) += board-kzm9g.o
> +obj-$(CONFIG_MACH_KZM9G) += board-kzm9g.o intc-sh73a0.o
> obj-$(CONFIG_MACH_KZM9G_REFERENCE) += board-kzm9g-reference.o
Building intc-sh73a0.o for kzm9g-legacy only, means we will no longer
instantiate
the GIC from C board code on kzm9g-reference. So the issue I mentioned in
the sh_tmu thread:
| On kzm9g-reference, the TWD is not instantiated, causing the same hang.
| If I instantiate the TWD from C board code there, it fails with
|
| twd: can't register interrupt 29 (-22)
|
| which looks like a symptom of https://lkml.org/lkml/2014/12/2/339
is indeed caused by the GIC.
> --- 0011/arch/arm/mach-shmobile/setup-sh73a0.c
> +++ work/arch/arm/mach-shmobile/setup-sh73a0.c 2014-12-16 19:13:33.000000000 +0900
> @@ -764,7 +764,9 @@ void __init __weak sh73a0_register_twd(v
> void __init sh73a0_earlytimer_init(void)
> {
> shmobile_init_delay();
> +#ifndef CONFIG_COMMON_CLK
> sh73a0_clock_init();
> +#endif
> shmobile_earlytimer_init();
> sh73a0_register_twd();
> }
> @@ -780,9 +782,10 @@ void __init sh73a0_add_early_devices(voi
>
> void __init sh73a0_add_standard_devices_dt(void)
> {
> +#ifndef CONFIG_COMMON_CLK
> /* clocks are setup late during boot in the case of DT */
> sh73a0_clock_init();
> -
> +#endif
> of_platform_populate(NULL, of_default_bus_match_table, NULL, NULL);
> }
These changes conflict with "[PATCH v7 6/6] ARM: shmobile: sh73a0: disable
legacy clock initialization", on which this series claims to depend on?
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support
2014-12-17 9:42 [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support Magnus Damm
2014-12-17 13:49 ` Geert Uytterhoeven
@ 2014-12-17 14:17 ` Magnus Damm
2014-12-17 14:24 ` Geert Uytterhoeven
` (3 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Magnus Damm @ 2014-12-17 14:17 UTC (permalink / raw)
To: linux-sh
Hi Geert,
On Wed, Dec 17, 2014 at 10:49 PM, Geert Uytterhoeven
<geert@linux-m68k.org> wrote:
> Hi Magnus,
>
> On Wed, Dec 17, 2014 at 10:42 AM, Magnus Damm <magnus.damm@gmail.com> wrote:
>> From: Magnus Damm <damm+renesas@opensource.se>
>>
>> Enable sh73a0 Multiplatform support for the generic sh73a0
>> machine vector. No board support is enabled, and the board
>> code for KZM9G DT Reference is left by itself.
>>
>> Signed-off-by: Magnus Damm <damm+renesas@opensource.se>
>
>> --- 0001/arch/arm/mach-shmobile/Makefile
>> +++ work/arch/arm/mach-shmobile/Makefile 2014-12-16 19:12:19.000000000 +0900
>> @@ -7,7 +7,7 @@ obj-y := timer.o console.o
>>
>> # CPU objects
>> obj-$(CONFIG_ARCH_SH7372) += setup-sh7372.o intc-sh7372.o pm-sh7372.o
>> -obj-$(CONFIG_ARCH_SH73A0) += setup-sh73a0.o intc-sh73a0.o pm-sh73a0.o
>> +obj-$(CONFIG_ARCH_SH73A0) += setup-sh73a0.o pm-sh73a0.o
>> obj-$(CONFIG_ARCH_R8A73A4) += setup-r8a73a4.o
>> obj-$(CONFIG_ARCH_R8A7740) += setup-r8a7740.o pm-r8a7740.o
>> obj-$(CONFIG_ARCH_R8A7778) += setup-r8a7778.o
>> @@ -65,7 +65,7 @@ obj-$(CONFIG_MACH_BOCKW) += board-bockw.
>> obj-$(CONFIG_MACH_BOCKW_REFERENCE) += board-bockw-reference.o
>> obj-$(CONFIG_MACH_MARZEN) += board-marzen.o
>> obj-$(CONFIG_MACH_ARMADILLO800EVA) += board-armadillo800eva.o
>> -obj-$(CONFIG_MACH_KZM9G) += board-kzm9g.o
>> +obj-$(CONFIG_MACH_KZM9G) += board-kzm9g.o intc-sh73a0.o
>> obj-$(CONFIG_MACH_KZM9G_REFERENCE) += board-kzm9g-reference.o
>
> Building intc-sh73a0.o for kzm9g-legacy only, means we will no longer
> instantiate
> the GIC from C board code on kzm9g-reference. So the issue I mentioned in
> the sh_tmu thread:
>
> | On kzm9g-reference, the TWD is not instantiated, causing the same hang.
> | If I instantiate the TWD from C board code there, it fails with
> |
> | twd: can't register interrupt 29 (-22)
> |
> | which looks like a symptom of https://lkml.org/lkml/2014/12/2/339
>
> is indeed caused by the GIC.
Oh, thanks for bringing this to my attention. So sh73a0 legacy DT
reference needs some care too?
>> --- 0011/arch/arm/mach-shmobile/setup-sh73a0.c
>> +++ work/arch/arm/mach-shmobile/setup-sh73a0.c 2014-12-16 19:13:33.000000000 +0900
>> @@ -764,7 +764,9 @@ void __init __weak sh73a0_register_twd(v
>> void __init sh73a0_earlytimer_init(void)
>> {
>> shmobile_init_delay();
>> +#ifndef CONFIG_COMMON_CLK
>> sh73a0_clock_init();
>> +#endif
>> shmobile_earlytimer_init();
>> sh73a0_register_twd();
>> }
>> @@ -780,9 +782,10 @@ void __init sh73a0_add_early_devices(voi
>>
>> void __init sh73a0_add_standard_devices_dt(void)
>> {
>> +#ifndef CONFIG_COMMON_CLK
>> /* clocks are setup late during boot in the case of DT */
>> sh73a0_clock_init();
>> -
>> +#endif
>> of_platform_populate(NULL, of_default_bus_match_table, NULL, NULL);
>> }
>
> These changes conflict with "[PATCH v7 6/6] ARM: shmobile: sh73a0: disable
> legacy clock initialization", on which this series claims to depend on?
That's right. I have a different version of this patch that is
compatible with the one above.
I was hoping on getting this series merged somewhat soonish -
independently of the CCF patches - and without those #ifdef hunks the
code will not compile. But the drawback is that the patch above needs
to be dropped.
I'm happy to adjust, but I do think in general that merging stuff
early must be better than postponing and increasing the burden...
Cheers,
/ magnus
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support
2014-12-17 9:42 [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support Magnus Damm
2014-12-17 13:49 ` Geert Uytterhoeven
2014-12-17 14:17 ` Magnus Damm
@ 2014-12-17 14:24 ` Geert Uytterhoeven
2014-12-22 10:10 ` Geert Uytterhoeven
` (2 subsequent siblings)
5 siblings, 0 replies; 7+ messages in thread
From: Geert Uytterhoeven @ 2014-12-17 14:24 UTC (permalink / raw)
To: linux-sh
Hi Magnus,
On Wed, Dec 17, 2014 at 3:17 PM, Magnus Damm <magnus.damm@gmail.com> wrote:
>>> --- 0001/arch/arm/mach-shmobile/Makefile
>>> +++ work/arch/arm/mach-shmobile/Makefile 2014-12-16 19:12:19.000000000 +0900
>>> @@ -7,7 +7,7 @@ obj-y := timer.o console.o
>>> # CPU objects
>>> obj-$(CONFIG_ARCH_SH7372) += setup-sh7372.o intc-sh7372.o pm-sh7372.o
>>> -obj-$(CONFIG_ARCH_SH73A0) += setup-sh73a0.o intc-sh73a0.o pm-sh73a0.o
>>> +obj-$(CONFIG_ARCH_SH73A0) += setup-sh73a0.o pm-sh73a0.o
>>> obj-$(CONFIG_ARCH_R8A73A4) += setup-r8a73a4.o
>>> obj-$(CONFIG_ARCH_R8A7740) += setup-r8a7740.o pm-r8a7740.o
>>> obj-$(CONFIG_ARCH_R8A7778) += setup-r8a7778.o
>>> @@ -65,7 +65,7 @@ obj-$(CONFIG_MACH_BOCKW) += board-bockw.
>>> obj-$(CONFIG_MACH_BOCKW_REFERENCE) += board-bockw-reference.o
>>> obj-$(CONFIG_MACH_MARZEN) += board-marzen.o
>>> obj-$(CONFIG_MACH_ARMADILLO800EVA) += board-armadillo800eva.o
>>> -obj-$(CONFIG_MACH_KZM9G) += board-kzm9g.o
>>> +obj-$(CONFIG_MACH_KZM9G) += board-kzm9g.o intc-sh73a0.o
>>> obj-$(CONFIG_MACH_KZM9G_REFERENCE) += board-kzm9g-reference.o
>>
>> Building intc-sh73a0.o for kzm9g-legacy only, means we will no longer
>> instantiate
>> the GIC from C board code on kzm9g-reference. So the issue I mentioned in
>> the sh_tmu thread:
>>
>> | On kzm9g-reference, the TWD is not instantiated, causing the same hang.
>> | If I instantiate the TWD from C board code there, it fails with
>> |
>> | twd: can't register interrupt 29 (-22)
>> |
>> | which looks like a symptom of https://lkml.org/lkml/2014/12/2/339
>>
>> is indeed caused by the GIC.
>
> Oh, thanks for bringing this to my attention. So sh73a0 legacy DT
> reference needs some care too?
I think it needs a call to gic_init() somewhere else.
>>> --- 0011/arch/arm/mach-shmobile/setup-sh73a0.c
>>> +++ work/arch/arm/mach-shmobile/setup-sh73a0.c 2014-12-16 19:13:33.000000000 +0900
>>> @@ -764,7 +764,9 @@ void __init __weak sh73a0_register_twd(v
>>> void __init sh73a0_earlytimer_init(void)
>>> {
>>> shmobile_init_delay();
>>> +#ifndef CONFIG_COMMON_CLK
>>> sh73a0_clock_init();
>>> +#endif
>>> shmobile_earlytimer_init();
>>> sh73a0_register_twd();
>>> }
>>> @@ -780,9 +782,10 @@ void __init sh73a0_add_early_devices(voi
>>>
>>> void __init sh73a0_add_standard_devices_dt(void)
>>> {
>>> +#ifndef CONFIG_COMMON_CLK
>>> /* clocks are setup late during boot in the case of DT */
>>> sh73a0_clock_init();
>>> -
>>> +#endif
>>> of_platform_populate(NULL, of_default_bus_match_table, NULL, NULL);
>>> }
>>
>> These changes conflict with "[PATCH v7 6/6] ARM: shmobile: sh73a0: disable
>> legacy clock initialization", on which this series claims to depend on?
>
> That's right. I have a different version of this patch that is
> compatible with the one above.
>
> I was hoping on getting this series merged somewhat soonish -
> independently of the CCF patches - and without those #ifdef hunks the
> code will not compile. But the drawback is that the patch above needs
> to be dropped.
>
> I'm happy to adjust, but I do think in general that merging stuff
> early must be better than postponing and increasing the burden...
I 100% agree.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support
2014-12-17 9:42 [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support Magnus Damm
` (2 preceding siblings ...)
2014-12-17 14:24 ` Geert Uytterhoeven
@ 2014-12-22 10:10 ` Geert Uytterhoeven
2014-12-23 0:12 ` Simon Horman
2014-12-23 0:15 ` Simon Horman
5 siblings, 0 replies; 7+ messages in thread
From: Geert Uytterhoeven @ 2014-12-22 10:10 UTC (permalink / raw)
To: linux-sh
On Wed, Dec 17, 2014 at 10:42 AM, Magnus Damm <magnus.damm@gmail.com> wrote:
>
> --- 0001/arch/arm/mach-shmobile/Kconfig
> +++ work/arch/arm/mach-shmobile/Kconfig 2014-12-16 19:11:36.000000000 +0900
> @@ -52,6 +52,11 @@ config ARCH_R7S72100
> bool "RZ/A1H (R7S72100)"
> select SYS_SUPPORTS_SH_MTU2
>
> +config ARCH_SH73A0
> + bool "SH-Mobile AG5 (R8A73A00)"
> + select ARCH_RMOBILE
> + select RENESAS_INTC_IRQPIN
> +
> config ARCH_R8A7740
> bool "R-Mobile A1 (R8A77400)"
> select ARCH_RMOBILE
Shall we keep this sorted?
That does mean the entries for the old SH73xx are at the end of the
list, though.
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support
2014-12-17 9:42 [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support Magnus Damm
` (3 preceding siblings ...)
2014-12-22 10:10 ` Geert Uytterhoeven
@ 2014-12-23 0:12 ` Simon Horman
2014-12-23 0:15 ` Simon Horman
5 siblings, 0 replies; 7+ messages in thread
From: Simon Horman @ 2014-12-23 0:12 UTC (permalink / raw)
To: linux-sh
On Mon, Dec 22, 2014 at 11:10:26AM +0100, Geert Uytterhoeven wrote:
> On Wed, Dec 17, 2014 at 10:42 AM, Magnus Damm <magnus.damm@gmail.com> wrote:
> >
> > --- 0001/arch/arm/mach-shmobile/Kconfig
> > +++ work/arch/arm/mach-shmobile/Kconfig 2014-12-16 19:11:36.000000000 +0900
> > @@ -52,6 +52,11 @@ config ARCH_R7S72100
> > bool "RZ/A1H (R7S72100)"
> > select SYS_SUPPORTS_SH_MTU2
> >
> > +config ARCH_SH73A0
> > + bool "SH-Mobile AG5 (R8A73A00)"
> > + select ARCH_RMOBILE
> > + select RENESAS_INTC_IRQPIN
> > +
> > config ARCH_R8A7740
> > bool "R-Mobile A1 (R8A77400)"
> > select ARCH_RMOBILE
>
> Shall we keep this sorted?
> That does mean the entries for the old SH73xx are at the end of the
> list, though.
Sorting it is nice. But are other entries out of order?
In either case, feel free to submit a cleanup.
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support
2014-12-17 9:42 [PATCH v2 03/05] ARM: shmobile: sh73a0: Add Multiplatform support Magnus Damm
` (4 preceding siblings ...)
2014-12-23 0:12 ` Simon Horman
@ 2014-12-23 0:15 ` Simon Horman
5 siblings, 0 replies; 7+ messages in thread
From: Simon Horman @ 2014-12-23 0:15 UTC (permalink / raw)
To: linux-sh
On Tue, Dec 23, 2014 at 09:12:34AM +0900, Simon Horman wrote:
> On Mon, Dec 22, 2014 at 11:10:26AM +0100, Geert Uytterhoeven wrote:
> > On Wed, Dec 17, 2014 at 10:42 AM, Magnus Damm <magnus.damm@gmail.com> wrote:
> > >
> > > --- 0001/arch/arm/mach-shmobile/Kconfig
> > > +++ work/arch/arm/mach-shmobile/Kconfig 2014-12-16 19:11:36.000000000 +0900
> > > @@ -52,6 +52,11 @@ config ARCH_R7S72100
> > > bool "RZ/A1H (R7S72100)"
> > > select SYS_SUPPORTS_SH_MTU2
> > >
> > > +config ARCH_SH73A0
> > > + bool "SH-Mobile AG5 (R8A73A00)"
> > > + select ARCH_RMOBILE
> > > + select RENESAS_INTC_IRQPIN
> > > +
> > > config ARCH_R8A7740
> > > bool "R-Mobile A1 (R8A77400)"
> > > select ARCH_RMOBILE
> >
> > Shall we keep this sorted?
> > That does mean the entries for the old SH73xx are at the end of the
> > list, though.
>
> Sorting it is nice. But are other entries out of order?
> In either case, feel free to submit a cleanup.
Sorry, for some reason I thought I had queued up the above patch.
I think sorting makes sense. And it looks like the other entries are already
sorted. So adjusting this patch seems like a good idea to me. But I'll
let Magnus decide.
^ permalink raw reply [flat|nested] 7+ messages in thread