Re: [PATCH 2/2] clk: bcm2835: Skip PLLC clocks when deciding on a new clock parent

[Martin Sperl]

> On 02.05.2016, at 17:29, Eric Anholt <eric@xxxxxxxxxx> wrote:
> 
> Martin Sperl <kernel@xxxxxxxxxxxxxxxx> writes:
> 
>>> On 26.04.2016, at 21:39, Eric Anholt <eric@xxxxxxxxxx> wrote:
>>> 
>>> If the firmware had set up a clock to source from PLLC, go along with
>>> it.  But if we're looking for a new parent, we don't want to switch it
>>> to PLLC because the firmware will force PLLC (and thus the AXI bus
>>> clock) to different frequencies during over-temp/under-voltage,
>>> without notification to Linux.
>>> 
>>> On my system, this moves the Linux-enabled HDMI state machine and DSI1
>>> escape clock over to plld_per from pllc_per.  EMMC still ends up on
>>> pllc_per, because the firmware had set it up to use that.
>>> 
>>> Signed-off-by: Eric Anholt <eric@xxxxxxxxxx>
>>> Fixes: 41691b8862e2 ("clk: bcm2835: Add support for programming the audio domain clocks")
>>> â
>> 
>> I guess this patch looks to me as if it is a policy inside the kernel,
>> which is AFAIK frowned upon.
> 
> Can you come up with a use for putting peripherals on PLLC ever, such
> that we need choice?

For PLLC not right now, but with clk_notifier_register drivers could
work around those clock changes (assuming we get that information
from the firmware somehow - or if we could move this decision into the
kernel: even better).

But I can come up with a scenario that would make use of the pllh_aux
under some circumstances - e.g when requesting 290039Hz on clock gp0/1/2.

Similarly: if we ever enable the testdebugX clocks these become immediate
candidates for parent-clocks as well which can result in more headache.

Being able to define which clocks to use at least give the dts author
a means also to control clock selection if he wants to enable the
testdebug clocks.

Another similar situation can occur with plla_per - under some
circumstances it may get used unexpectedly.

Martin