Hi Mark, On 28.05.2020 15:43, Mark Brown wrote: > On Thu, May 28, 2020 at 03:11:30PM +0200, Marek Szyprowski wrote: >> Balancing coupled regulators must wait until the clients for all of the >> coupled regualtors set their constraints, otherwise the balancing code >> might change the voltage of the not-yet-constrained regulator to the >> value below the bootloader-configured operation point, what might cause a >> system crash. > This forces every supply to have something which explicitly manages > voltages which means that if one of the coupled supplies doesn't really > care about the voltage (perhaps doesn't even have any explicit > consumers) and just needs to be within a certain range of another supply > then it'll end up restricting things needlessly. Frankly, that's exactly what we need for Exynos5422 case. If devfreq driver is not enabled/compiled, we want to keep the "vdd_int" volatage unchanged. This confirms me that we really need to have a custom coupler for Exynos5422 case. It will solve such issues without adding hacks to regulator core. > Saravana was trying to do some stuff with sync_state() which might be > interesting here although I have concerns with that approach too: > > https://lore.kernel.org/lkml/20200527074057.246606-1-saravanak@xxxxxxxxxx/ This still doesn't solve the above mentioned case. Best regards -- Marek Szyprowski, PhD Samsung R&D Institute Poland
