Re: [PATCH v2 4/5] clk: amlogic: Optimize PLL enable timing

[Chuan Liu <chuan.liu@amlogic.com>]

Hi Jerome,

On 10/30/2025 4:38 PM, Jerome Brunet wrote:
> [ EXTERNAL EMAIL ]
> 
> On Thu 30 Oct 2025 at 13:24, Chuan Liu via B4 Relay <devnull+chuan.liu.amlogic.com@kernel.org> wrote:
> 
>> From: Chuan Liu <chuan.liu@amlogic.com>
>>
>> l_detect controls the enablement of the PLL lock detection module.
>> It should remain disabled while the internal PLL circuits are
>> reaching a steady state; otherwise, the lock signal may be falsely
>> triggered high.
>>
>> Before enabling the internal power supply of the PLL, l_detect should
>> be disabled. After the PLL’s internal circuits have stabilized,
>> l_detect should be enabled to prevent false lock signal triggers.
> 
> You to reformat this description. It feel that both paragraph are saying
> the same thing.
> 

Ok, drop it.

>>
>> Currently, only A1 supports both l_detect and current_en, so this
>> patch will only affect A1.
>>
>> Signed-off-by: Chuan Liu <chuan.liu@amlogic.com>
>> ---
>>   drivers/clk/meson/clk-pll.c | 28 +++++++++++++++-------------
>>   1 file changed, 15 insertions(+), 13 deletions(-)
>>
>> diff --git a/drivers/clk/meson/clk-pll.c b/drivers/clk/meson/clk-pll.c
>> index 6c794adb8ccd..c6eebde1f516 100644
>> --- a/drivers/clk/meson/clk-pll.c
>> +++ b/drivers/clk/meson/clk-pll.c
>> @@ -383,36 +383,38 @@ static int meson_clk_pll_enable(struct clk_hw *hw)
>>        if (MESON_PARM_APPLICABLE(&pll->rst))
>>                meson_parm_write(clk->map, &pll->rst, 1);
>>
>> +     /* Disable the PLL lock-detect module */
>> +     if (MESON_PARM_APPLICABLE(&pll->l_detect))
>> +             meson_parm_write(clk->map, &pll->l_detect, 1);
>> +
>>        /* Enable the pll */
>>        meson_parm_write(clk->map, &pll->en, 1);
>>        /* Wait for Bandgap and LDO to power up and stabilize */
>>        udelay(20);
>>
>> -     /* Take the pll out reset */
>> -     if (MESON_PARM_APPLICABLE(&pll->rst))
>> -             meson_parm_write(clk->map, &pll->rst, 0);
> 
> Why is the reset moving around ? nothing is said in the description about
> that
> 

The function of current_en in the PLL is similar to en. It's more
reliable to assert the reset signal after current_en is set high.

I'll update the commit description accordingly in the next version.

>> -
>> -     /* Wait for PLL loop stabilization */
>> -     udelay(20);
>> -
>>        /*
>>         * Compared with the previous SoCs, self-adaption current module
>>         * is newly added for A1, keep the new power-on sequence to enable the
>>         * PLL. The sequence is:
>> -      * 1. enable the pll, delay for 10us
>> +      * 1. enable the pll, delay for 20us
>>         * 2. enable the pll self-adaption current module, delay for 40us
>>         * 3. enable the lock detect module
>>         */
>>        if (MESON_PARM_APPLICABLE(&pll->current_en)) {
>> -             udelay(10);
>>                meson_parm_write(clk->map, &pll->current_en, 1);
>> -             udelay(40);
>> +             udelay(20);
>>        }
>>
>> -     if (MESON_PARM_APPLICABLE(&pll->l_detect)) {
>> -             meson_parm_write(clk->map, &pll->l_detect, 1);
>> +     /* Take the pll out reset */
>> +     if (MESON_PARM_APPLICABLE(&pll->rst))
>> +             meson_parm_write(clk->map, &pll->rst, 0);
>> +
>> +     /* Wait for PLL loop stabilization */
>> +     udelay(20);
>> +
>> +     /* Enable the lock-detect module */
>> +     if (MESON_PARM_APPLICABLE(&pll->l_detect))
>>                meson_parm_write(clk->map, &pll->l_detect, 0);
>> -     }
>>
>>        if (meson_clk_pll_wait_lock(hw)) {
>>                /* disable PLL when PLL lock failed. */
> 
> --
> Jerome