[PATCH 05/16] ARM: S5PC100: new clocks definition

[Jassi Brar]

On Wed, May 12, 2010 at 5:37 PM, Marek Szyprowski
<m.szyprowski at samsung.com> wrote:
> Hello,
>
> On Wednesday, May 12, 2010 9:34 AM Kukjin Kim wrote:
>
>> ...
>> > >
>> > > Marek,
>> > >
>> > > Actually, need to modify many part to fit new clock scheme.
>> > > For example, to use clksrc_clk is easily to handle clock.
>> >
>> > I thought I covered most of the changes. Probably only the armclk
>> > related functions and a cleanup in s5pc100/mach/regs-clock.h are
>> > left and these can be easily fixed.
>> >
>> > > S.LSI already made the patch for C100 clock changing which is using
>> > > clksrc_clk structure. So may I submit it?
>> >
>> > Could you compare your version with the final version that is created
>> > by my patch series? I don't think that there will be many differences
>> > there.
>> >
>>
>> There are defined clock source of C100 available clock in SYS.LSI patch
>> like
>> clocks defined by CLK_SRC1 register.
>> SYS.LSI has covered all possible C100 clocks in our patch.
>
> So this is rather a question of general policy about the clocks. Do we
> really need to define them even if no driver would use it? I'm not
> convinced. On the other hand defining all clocks now may reduce the driver
> development time in the future.
It is always better for a code to support as many cases as possible.
After all, we wouldn't want to wait for weeks until the clock support
for a device to
appear upstream for a Maintainer to accept the device driver in his
tree. Would we?