On 5/10/20 12:51 pm, Avri Altman wrote: >> >> >> On 5/10/20 11:02 am, Avri Altman wrote: >>> HI, >>> >>>> Drivers that wish to support DeepSleep need to set a new capability flag >>>> UFSHCD_CAP_DEEPSLEEP and provide a hardware reset via the existing >>>> ->device_reset() callback. >>> I would expect that this capability controls sending SSU 4, but it only >>> controls >> the sysfs entry? >> >> The sysfs entry is the only way to request DeepSleep. > Some chipset vendors use their own modules, or even the device tree > to set their default rpm_lvl / spm_lvl.
I would not expect them to set it wrongly but what are you suggesting?