On 27/07/2021 19:29, Stephen Boyd wrote:
Quoting Bjorn Andersson (2021-07-07 21:03:06)
On Wed 07 Jul 20:08 CDT 2021, Bryan O'Donoghue wrote:
The current implementation omits the necessary mmcx supply, which means if
you are running the code for this block and a prior clock driver, like say
the videocc hasn't run, then a reset will be generated the first time we
touch these registers.
Fixes: 5d66ca79b58c ("clk: qcom: Add camera clock controller driver for SM8250")
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@xxxxxxxxxx>
camcc isn't enabled in the upstream dts yet and I expect that we're
going to conclude on defining these GDSCs as subdomains of the cc's
power-domain in time for v5.15.
I don't mind if Stephen picks this to make sure the driver is
functional, but I will hold off on the dts change.
Seems like this is superseded by the GDSC patches that use subdomains
instead of a fake supply?
yep
