Hello Heiko and Doug,

On Wed, Feb 15, 2017 at 2:01 PM, Heiko St?bner [off-list ref] wrote:

Am Mittwoch, 15. Februar 2017, 12:27:59 CET schrieb Javier Martinez Canillas:

quoted

Hello Doug,

On Tue, Feb 14, 2017 at 6:01 PM, Douglas Anderson [off-list ref]

wrote:

quoted

quoted

The PMU Cortex M0 on rk3399 is intended to be used for things like
DDRFreq transitions, suspend/resume, and other things that are the
purview of ARM Trusted Firmware and not the kernel.  As such, the
kernel shouldn't be messing with the clocks.  Add CLK_IGNORE_UNUSED to
these clocks.

Isn't CLK_IS_CRITICAL a more suitable flag for this case?

From the patch description it looks like the clock is expected to be
controlled from firmware in most cases as I guess the Cortex M0 will be used
for that all the time now. And the clock is not expected to run all the time.

So I'd think clk_ignore_unused is the correct one. The whole clock-subtree for
these clocks also does not get affected by other clocks, as it is
independendly coming from PLLs.

I see. Thanks a lot for your explanations. I just asked because
CLK_IGNORE_UNUSED only prevents the clock to be disabled by
clk_disable_unused() but still can be disabled if a driver explicitly
looks it up and calls clk_disable() or if the clock is affected by
other clocks.

But I understand now that both scenarios are not possible. So yes,
CLK_IGNORE_UNUSED is more suitable in this case.

Heiko

Best regards,
Javier