The 32K timer isn't available on all OMAP devices.
Signed-off-by: Alessio Igor Bogani <[email protected]>
---
arch/arm/plat-omap/Makefile | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/arch/arm/plat-omap/Makefile b/arch/arm/plat-omap/Makefile
index 8d88584..b1b321c 100644
--- a/arch/arm/plat-omap/Makefile
+++ b/arch/arm/plat-omap/Makefile
@@ -3,11 +3,13 @@
#
# Common support
-obj-y := sram.o dma.o fb.o counter_32k.o
+obj-y := sram.o dma.o fb.o
obj-m :=
obj-n :=
obj- :=
+obj-$(CONFIG_OMAP_32K_TIMER) += counter_32k.o
+
# omap_device support (OMAP2+ only at the moment)
obj-$(CONFIG_OMAP_DM_TIMER) += dmtimer.o
--
1.7.9.5
On 11/26/12 11:28, Alessio Igor Bogani wrote:
> The 32K timer isn't available on all OMAP devices.
>
> Signed-off-by: Alessio Igor Bogani <[email protected]>
> ---
> arch/arm/plat-omap/Makefile | 4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm/plat-omap/Makefile b/arch/arm/plat-omap/Makefile
> index 8d88584..b1b321c 100644
> --- a/arch/arm/plat-omap/Makefile
> +++ b/arch/arm/plat-omap/Makefile
> @@ -3,11 +3,13 @@
> #
>
> # Common support
> -obj-y := sram.o dma.o fb.o counter_32k.o
> +obj-y := sram.o dma.o fb.o
> obj-m :=
> obj-n :=
> obj- :=
>
> +obj-$(CONFIG_OMAP_32K_TIMER) += counter_32k.o
We are moving away from this config option in favor of runtime detection,
so I don't think this patch is appropriate.
Why do you need this?
> +
> # omap_device support (OMAP2+ only at the moment)
>
> obj-$(CONFIG_OMAP_DM_TIMER) += dmtimer.o
--
Regards,
Igor.
Hi Igor,
On 26/11/2012 13:02, Igor Grinberg wrote:
> On 11/26/12 11:28, Alessio Igor Bogani wrote:
[...]
>> # Common support
>> -obj-y := sram.o dma.o fb.o counter_32k.o
>> +obj-y := sram.o dma.o fb.o
>> obj-m :=
>> obj-n :=
>> obj- :=
>>
>> +obj-$(CONFIG_OMAP_32K_TIMER) += counter_32k.o
>
> We are moving away from this config option in favor of runtime detection,
Well, I'll be happy when it'll happen.
> Why do you need this?
Because until now the build system doesn't honour the config file.
Indeed it builds that source code file also when I set
CONFIG_OMAP_32K_TIMER to n.
The runtime detection isn't a good excuse for doesn't make the build
system working like users expect.
Ciao,
Alessio
On 11/26/12 15:15, Alessio Igor Bogani wrote:
> Hi Igor,
>
> On 26/11/2012 13:02, Igor Grinberg wrote:
>> On 11/26/12 11:28, Alessio Igor Bogani wrote:
> [...]
>>> # Common support
>>> -obj-y := sram.o dma.o fb.o counter_32k.o
>>> +obj-y := sram.o dma.o fb.o
>>> obj-m :=
>>> obj-n :=
>>> obj- :=
>>>
>>> +obj-$(CONFIG_OMAP_32K_TIMER) += counter_32k.o
>>
>> We are moving away from this config option in favor of runtime detection,
>
> Well, I'll be happy when it'll happen.
>
>> Why do you need this?
>
> Because until now the build system doesn't honour the config file. Indeed it builds that source code file also when I set CONFIG_OMAP_32K_TIMER to n.
>
> The runtime detection isn't a good excuse for doesn't make the build system working like users expect.
So, the problem is the users expectations...
If you look, at Tony's omap-for-v3.8/timer branch,
patch: ARM: OMAP2+: timer: remove CONFIG_OMAP_32K_TIMER
it should change the expectations (at least I tried to do this in Kconfig file).
So, to the question of honoring the config option - yes,
but it is a work in progress on removing that one.
If you have a real issue that you are trying to fix - it is totally different thing,
but if it is just config option honoring... then I don't think we should merge this patch.
--
Regards,
Igor.