Re: [PATCH v3 2/3] tty: serial: msm: Remove duplicate operations on clocks in startup/shutdown

From: Stephen Boyd
Date: Thu May 07 2015 - 19:34:59 EST


On 04/29/15 08:45, Pramod Gurav wrote:
> Thanks Stephen for review.
>
> On Fri, April 10, 2015 11:33 pm, Stephen Boyd wrote:
>> On 04/10/15 05:19, Pramod Gurav wrote:
>>> @@ -683,8 +679,7 @@ static void msm_power(struct uart_port *port,
>>> unsigned int state,
>>>
>>> switch (state) {
>>> case 0:
>>> - clk_prepare_enable(msm_port->clk);
>>> - clk_prepare_enable(msm_port->pclk);
>>> + msm_init_clock(port);
>> Hm... now we would call msm_serial_set_mnd_regs() whenever we power on
>> the port? Presumably we only need to do that once when we probe (or when
>> we resume from a sleep state that resets the registers, i.e.
>> hibernation) but I guess we're getting saved by the fact that the
>> if/else if pair in msm_serial_set_mnd_regs_from_uartclk would never be
>> true after the first time we call it?
> I tried replacing msm_init_clock() call with msm_serial_set_mnd_regs() in
> msm_startup() as msm_startup gets called just after msm_power() so that
> clk_prepare_enable() is followed by mnd settings. But it does not get the
> kernel booted for some reason.

That's concerning. Did you also drop the call to msm_init_clock() from
msm_power() as is done in this patch? If that's done then it seems that
nothing would be different besides the removal of clk_prepare_enable()
in msm_startup().

>
> So, can I get a acked-by for this patch or you still think it can be done
> in a better way?

Using msm_init_clock() in the msm_power() doesn't look symmetrical so if
we can avoid it I would prefer that.

--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/