On Tue, Sep 20, 2022 at 09:06:22AM +0530, Rajendra Nayak wrote:
>
>
> On 9/19/2022 9:15 PM, Johan Hovold wrote:
> > On Fri, Sep 16, 2022 at 03:54:16PM +0530, Rajendra Nayak wrote:
> >> The USB controller on sc7180 does not retain the state when
> >> the system goes into low power state and the GDSC is
> >> turned off. This results in the controller reinitializing and
> >> re-enumerating all the connected devices (resulting in additional
> >> delay while coming out of suspend)
> >> Fix this by updating the .pwrsts for the USB GDSC so it only
> >> transitions to retention state in low power.
> >>
> >> Signed-off-by: Rajendra Nayak <quic_rjendra@xxxxxxxxxxx>
> >> Reviewed-by: Matthias Kaehlcke <mka@xxxxxxxxxxxx>
> >> Tested-by: Matthias Kaehlcke <mka@xxxxxxxxxxxx>
> >> ---
> >> v2:
> >> Updated the changelog
> >>
> >> drivers/clk/qcom/gcc-sc7180.c | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/clk/qcom/gcc-sc7180.c b/drivers/clk/qcom/gcc-sc7180.c
> >> index c2ea09945c47..2d3980251e78 100644
> >> --- a/drivers/clk/qcom/gcc-sc7180.c
> >> +++ b/drivers/clk/qcom/gcc-sc7180.c
> >> @@ -2224,7 +2224,7 @@ static struct gdsc usb30_prim_gdsc = {
> >> .pd = {
> >> .name = "usb30_prim_gdsc",
> >> },
> >> - .pwrsts = PWRSTS_OFF_ON,
> >> + .pwrsts = PWRSTS_RET_ON,
> >> };
> >>
> >> static struct gdsc hlos1_vote_mmnoc_mmu_tbu_hf0_gdsc = {
> >
> > It seems like the above will not work unless you also provide the
> > registers offsets that gdsc_force_mem_on() expects.
>
> That's true, but its needed only on platforms that support complete
> CX domain power collapse. sc7280 (and sc7180) does not support
> that and hence we can achieve GDSC RET without any of the RETAIN_MEM/
> RETAIN_PERIPH bits programmed.
> I explained some of that in detail on another related thread a
> while back [1]
>
> [1] https://lore.kernel.org/all/5ff21b1e-3af9-36ef-e13e-fa33f526d0e3@xxxxxxxxxxx/
Thanks for the link, that was was very informative.
Could you please update the commit message of patch 1/3 so that it also
covers these systems and explains why you don't set the RETAIN_MEM and
RETAIN_PERIPH bits for them?
As that commit message stands now, it seems that those bits must be set
for retention to work.
> > Specifically, unless you set cxc_count for the GDSC, the above call is a
> > no-op and the retention setting (i.e. the RETAIN_MEM and RETAIN_PERIPH
> > bits) will not be updated when registering the GDSC.
Johan
