Re: [RFC PATCH] OMAP3:PM: Fix OPP scale logic

[Nishanth Menon <nm@ti.com>]

Roger Quadros had written, on 08/04/2009 03:00 AM, the following:
> ext Nishanth Menon wrote:
>> diff --git a/arch/arm/mach-omap2/resource34xx.c b/arch/arm/mach-omap2/resource34xx.c
>> index 25535a3..1ceaed8 100644
>> --- a/arch/arm/mach-omap2/resource34xx.c
>> +++ b/arch/arm/mach-omap2/resource34xx.c
>> @@ -240,13 +240,23 @@ static int program_opp_freq(int res, int target_level, int current_level)
>>  	lock_scratchpad_sem();
>>  	if (res == VDD1_OPP) {
>>  		curr_opp = &curr_vdd1_opp;
>> -		clk_set_rate(dpll1_clk, mpu_opps[target_level].rate);
>> -		clk_set_rate(dpll2_clk, dsp_opps[target_level].rate);
>> +		ret = clk_set_rate(dpll1_clk, mpu_opps[target_level].rate);
>> +		if (unlikely(ret))
>> +			return ret;
> 
> if we return here we're not calling unlock_scratchpad_sem(). if you remove the 
> return statement the expected functionality will be achieved by the next if(ret) 
> statement.
yep. my bad.. thanks.. will fix and resend.
> 
> looks like you are changing the return behaviour from opp level to 0/1. You 
> should explain this in a function header comment.
This function does not have a function header comment. this information 
is already part of the patch header. Quote:"
This changes program_freq_opp return type in the process
for program_opp to handle error in a consistent
manner.
"

Will send out a rev2 based on comment 1 (unlock_scratchpad_sem).

-- 
Regards,
Nishanth Menon