> > On 2020-08-29 00:32, Avri Altman wrote: > >> > >> The zero value Auto-Hibernate Timer is a valid setting, and it > >> indicates the Auto-Hibernate feature being disabled. Correctly > > Right. So " ufshcd_auto_hibern8_enable" is no longer an appropriate > > name. > > Maybe ufshcd_auto_hibern8_set instead? > Thanks for your comment. I am ok with the name change suggestion. > > > > Also, did you verified that no other platform relies on its non-zero > > value? > I only tested the change on Qualcomm's platform. I do not have other > platforms to do the test. > The UFS host controller spec JESD220E, Section 5.2.5 says > "Software writes “0” to disable Auto-Hibernate Idle Timer". So the spec > supports this zero value. > Some options: > - We could add a hba->caps so that we only apply the change for > Qualcomm's platforms. > This is not preferred because it is following the spec implementations. > - Or other platforms that do not support the zero value needs a caps. Yeah, I don't think another caps is required, Maybe just an ack from Stanley. Thanks, Avri
