Message ID | 20201027180713.7642-1-ionela.voinescu@arm.com (mailing list archive) |
---|---|
Headers | show |
Series | condition EAS enablement on FI support | expand |
Hi guys, On Tuesday 27 Oct 2020 at 18:07:10 (+0000), Ionela Voinescu wrote: > Given the maturity gained by cpufreq-based Frequency Invariance (FI) > support following the patches at [1], this series conditions Energy > Aware Scheduling (EAS) enablement on a frequency invariant system. > > Currently, EAS can be enabled on a system without FI support, leading > to incorrect (energy-wise) task placements. As no warning is emitted, > it could take some debugging effort to track the behavior back to the > lack of FI support; this series changes that by disabling EAS > (and advertising it) when FI support is missing. > > The series is structured as follows: > - 1/3 - create function that can rebuild the scheduling and EAS' > performance domains if EAS' initial conditions change > - 2/3 - arm64: rebuild scheduling and performance domains in the > case of late, counter-driven FI initialisation. > - 3/3 - condition EAS enablement on FI support > > RESEND v2: rebase and retest on v5.10-rc1 > This still applies nicely on v5.10-rc4. Peter, given that 1/3 and 2/3 are acked, do you think there's anything else that needs to be done for this? Thank you, Ionela. > v1 -> v2: > - v1 can be found at [2] > - 1/2 - moved now "static inline" rebuild_sched_domains_energy() stub > in sched/topology.h header, as suggested by Quentin. > - Changed the order of 2/3 and 3/3 to maintain bisection as suggested > by Quentin. > - 3/3 - place the warning under sched_debug() as per Quentin and > Dietmar's recommendations. > > [1] Most recent version at: > https://lore.kernel.org/lkml/20200901205549.30096-1-ionela.voinescu@arm.com/ > [2] https://lore.kernel.org/lkml/20200924123937.20938-1-ionela.voinescu@arm.com/ > > Many thanks, > Ionela. > > Ionela Voinescu (3): > sched/topology,schedutil: wrap sched domains rebuild > arm64: rebuild sched domains on invariance status changes > sched/topology: condition EAS enablement on FIE support > > arch/arm64/kernel/topology.c | 10 ++++++++++ > include/linux/sched/topology.h | 8 ++++++++ > kernel/sched/cpufreq_schedutil.c | 9 +-------- > kernel/sched/topology.c | 27 ++++++++++++++++++++------- > 4 files changed, 39 insertions(+), 15 deletions(-) > > > base-commit: 3650b228f83adda7e5ee532e2b90429c03f7b9ec > -- > 2.17.1 >
On Wed, Nov 18, 2020 at 11:42:34AM +0000, Ionela Voinescu wrote: > Hi guys, > > On Tuesday 27 Oct 2020 at 18:07:10 (+0000), Ionela Voinescu wrote: > > Given the maturity gained by cpufreq-based Frequency Invariance (FI) > > support following the patches at [1], this series conditions Energy > > Aware Scheduling (EAS) enablement on a frequency invariant system. > > > > Currently, EAS can be enabled on a system without FI support, leading > > to incorrect (energy-wise) task placements. As no warning is emitted, > > it could take some debugging effort to track the behavior back to the > > lack of FI support; this series changes that by disabling EAS > > (and advertising it) when FI support is missing. > > > > The series is structured as follows: > > - 1/3 - create function that can rebuild the scheduling and EAS' > > performance domains if EAS' initial conditions change > > - 2/3 - arm64: rebuild scheduling and performance domains in the > > case of late, counter-driven FI initialisation. > > - 3/3 - condition EAS enablement on FI support > > > > RESEND v2: rebase and retest on v5.10-rc1 > > > > This still applies nicely on v5.10-rc4. > > Peter, given that 1/3 and 2/3 are acked, do you think there's anything > else that needs to be done for this? I'll take it, thanks!