Re: [PATCH] clk: aspeed: Prevent reset if clock is enabled

From: Joel Stanley
Date: Wed Mar 07 2018 - 22:54:13 EST


On Thu, Mar 8, 2018 at 2:12 PM, Joel Stanley <joel@xxxxxxxxx> wrote:
> Hi Eddie,
>
> On Thu, Mar 8, 2018 at 3:06 AM, Eddie James <eajames@xxxxxxxxxxxxxxxxxx> wrote:
>> According to the Aspeed specification, the reset and enable sequence
>> should be done when the clock is stopped. The specification doesn't
>> define behavior if the reset is done while the clock is enabled.
>>
>> From testing on the AST2500, the LPC Controller has problems if the
>> clock is reset while enabled.
>>
>> Therefore, check whether the clock is enabled or not before performing
>> the reset and enable sequence in the Aspeed clock driver.
>>
>> Signed-off-by: Eddie James <eajames@xxxxxxxxxxxxxxxxxx>
>> ---
>> drivers/clk/clk-aspeed.c | 6 ++++++
>> 1 file changed, 6 insertions(+)
>>
>> diff --git a/drivers/clk/clk-aspeed.c b/drivers/clk/clk-aspeed.c
>> index 9f7f931..a13054d 100644
>> --- a/drivers/clk/clk-aspeed.c
>> +++ b/drivers/clk/clk-aspeed.c
>> @@ -212,6 +212,12 @@ static int aspeed_clk_enable(struct clk_hw *hw)
>> u32 clk = BIT(gate->clock_idx);
>> u32 rst = BIT(gate->reset_idx);
>> u32 enval;
>> + u32 reg;
>> +
>> + /* Only reset/enable/unreset if clock is stopped */
>> + regmap_read(gate->map, ASPEED_CLK_STOP_CTRL, &reg);
>> + if (!(reg & clk))
>> + return 0;
>
> This doesn't generalise to all of the clocks, as some clocks use set
> to disable. Perhaps we could do something like this:
>
> /* Only reset/enable/unreset if clock is stopped. The LPC clock
> on ast2500 has issues otherwise */

We've seen this on both the ast2400 and ast2500, so if you do take my
suggestion perhaps just say "The LPC clock has issues otherwise".

> enval = (gate->flags & CLK_GATE_SET_TO_DISABLE) ? 0 : clk;
> regmap_read(gate->map, ASPEED_CLK_STOP_CTRL, &reg);
> if ((reg & clk) == enval) {
> spin_unlock_irqrestore(gate->lock, flags);
> return 0;
> }
>
> I think we should also do this operation under the lock.
>
> Please cc Ryan Chen <ryan_chen@xxxxxxxxxxxxxx> so he can confirm that
> this workaround is valid, and credit Lei who spent a lot of time
> investigating this issue. Perhaps "Root-caused-by".
>
> Cheers,
>
> Joel