On Mon, Feb 21, 2022 at 9:14 PM Limonciello, Mario <Mario.Limonciello@xxxxxxx> wrote: > > [AMD Official Use Only] > > > > > OK > > https://lore.kernel.org/platform-driver-x86/20220221200728.2323469-1-mario.limonciello@xxxxxxx/T/#u The comment in amd_pmc_resume() doesn't match the code, because the QoS request is updated to "don't care" in there. Also, I'd move that update to the end of the function in case what happens in amd_pmc_resume() interferes with idle states. Moreover, since the "no idle states" period covers the entire suspend time now AFAICS, I would use a small int (eg. 3) instead of 0 in the cpu_latency_qos_update_request() call in amd_pmc_verify_czn_rtc() to allow C1 to be entered via cpuidle during that period (which shouldn't really hurt). And the comment would need to be updated accordingly. > > > > > Long term - are you opposed to drivers/acpi/x86/s2idle.c moving to > > drivers/platform/x86/? > > > > It is tied to the code in sleep.c, so I'd rather not move it. > > OK if we're keeping it there, would you be open to a patch exporting a symbol to > let other drivers register another callback after LPS0? It's somewhat similar to what my > test hack did and might be a better long term solution if so. Well, please send the patch and I will tell you if it looks good to me.
