[U-Boot] [PATCH v4 3/5] OMAP: spl: call timer_inti() from SPL

[Tom Rini]

On Mon, Aug 6, 2012 at 8:02 AM, Ilya Yanok
<ilya.yanok at cogentembedded.com> wrote:
> Hi Tom,
>
> On Mon, Aug 6, 2012 at 2:35 AM, Tom Rini <trini at ti.com> wrote:
>
>> On Sun, Aug 5, 2012 at 2:21 PM, Ilya Yanok
>> <ilya.yanok at cogentembedded.com> wrote:
>> > We need to initialize timer properly, otherwise all delays
>> > inside SPL will be wrong.
>> >
>> > Signed-off-by: Ilya Yanok <ilya.yanok at cogentembedded.com>
>> > ---
>> >
>> >  arch/arm/cpu/armv7/omap-common/spl.c |    2 ++
>> >  1 file changed, 2 insertions(+)
>> >
>> > diff --git a/arch/arm/cpu/armv7/omap-common/spl.c
>> b/arch/arm/cpu/armv7/omap-common/spl.c
>> > index 4d1ac85..f0d766c 100644
>> > --- a/arch/arm/cpu/armv7/omap-common/spl.c
>> > +++ b/arch/arm/cpu/armv7/omap-common/spl.c
>> > @@ -152,6 +152,8 @@ void board_init_r(gd_t *id, ulong dummy)
>> >         mem_malloc_init(CONFIG_SYS_SPL_MALLOC_START,
>> >                         CONFIG_SYS_SPL_MALLOC_SIZE);
>> >
>> > +       timer_init();
>> > +
>> >  #ifdef CONFIG_SPL_BOARD_INIT
>> >         spl_board_init();
>> >  #endif
>>
>> Calling it twice has other bad side-effects so there should be a
>> timer_init removal somewhere too.
>
>
> Hm, I can see any so far. I think you are mixing it with init_timer()
> function you fixed in commit 2ab2810 (BTW, these two functions deal with
> the same piece of hardware. do we really need both?).
> Speaking about 2ab2810, I think the problem was not init_timer() being
> called twice but init_timer() called _after_ timer_init() (timer_init()
> programs the timer to get correct delays but init_timer() resets the timer
> so delays are wrong again).

Ug, that is a mess, your patch is fine, I'll go figure out what's
going on in the am33xx-specific portion.

-- 
Tom