Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1411818 > unrolled thread
| Started by | dbasehore@chromium.org |
|---|---|
| First post | 2016-06-02 06:40 +0200 |
| Last post | 2016-06-11 22:40 +0200 |
| Articles | 4 — 3 participants |
Back to article view | Back to linux.kernel
[PATCH 0/5] Add suspend-to-idle validation for Intel SoCs dbasehore@chromium.org - 2016-06-02 06:40 +0200
Re: [PATCH 0/5] Add suspend-to-idle validation for Intel SoCs Pavel Machek <pavel@ucw.cz> - 2016-06-07 09:50 +0200
Re: [PATCH 0/5] Add suspend-to-idle validation for Intel SoCs "dbasehore ." <dbasehore@chromium.org> - 2016-06-08 02:10 +0200
Re: [PATCH 0/5] Add suspend-to-idle validation for Intel SoCs Pavel Machek <pavel@ucw.cz> - 2016-06-11 22:40 +0200
| From | dbasehore@chromium.org |
|---|---|
| Date | 2016-06-02 06:40 +0200 |
| Subject | [PATCH 0/5] Add suspend-to-idle validation for Intel SoCs |
| Message-ID | <rFsAh-1gP-5@gated-at.bofh.it> |
From: Derek Basehore <dbasehore@chromium.org> This patch set adds support for catching errors when entering freeze on Intel Skylake SoCs. Support for this can be added to newer SoCs in later patches. Verification is done by waking up the CPU once every X (default 10) seconds to check the residency of S0ix. This can't be verified before attempting to enter S0ix through mwait, so we have to repeatedly verify entry into that state. Successfully entering S0ix is no guarantee that it will be entered on the next attempt, so we have to schedule another check. This has a minimal power impact of <1% of the total system power on our systems. This relies on the recently added patch "platform/x86: Add PMC Driver for Intel Core SoC" Derek Basehore (5): x86: stub out pmc function clockevents: Add timed freeze x86, apic: Add timed freeze support freeze: Add error reporting intel_idle: Add S0ix validation arch/x86/include/asm/pmc_core.h | 6 +- arch/x86/kernel/apic/apic.c | 25 +++++- drivers/acpi/processor_idle.c | 10 ++- drivers/cpuidle/cpuidle.c | 31 +++++-- drivers/idle/intel_idle.c | 185 +++++++++++++++++++++++++++++++++++++--- include/linux/clockchips.h | 10 +++ include/linux/cpuidle.h | 10 ++- include/linux/suspend.h | 10 +++ kernel/power/suspend.c | 11 ++- kernel/time/clockevents.c | 117 +++++++++++++++++++++++++ 10 files changed, 385 insertions(+), 30 deletions(-) -- 2.8.0.rc3.226.g39d4020
[toc] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-06-07 09:50 +0200 |
| Message-ID | <rHjVU-8cE-25@gated-at.bofh.it> |
| In reply to | #1411818 |
On Wed 2016-06-01 21:33:24, dbasehore@chromium.org wrote: > From: Derek Basehore <dbasehore@chromium.org> > > This patch set adds support for catching errors when entering freeze > on Intel Skylake SoCs. Support for this can be added to newer SoCs in > later patches. > > Verification is done by waking up the CPU once every X (default 10) > seconds to check the residency of S0ix. This can't be verified before > attempting to enter S0ix through mwait, so we have to repeatedly > verify entry into that state. Successfully entering S0ix is no > guarantee that it will be entered on the next attempt, so we have to > schedule another check. This has a minimal power impact of <1% of the > total system power on our systems. Dunno. Should this be protected with something like CONFIG_TEST_SLEEP? People probably don't want this for production... -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | "dbasehore ." <dbasehore@chromium.org> |
|---|---|
| Date | 2016-06-08 02:10 +0200 |
| Message-ID | <rHzeh-17I-1@gated-at.bofh.it> |
| In reply to | #1415809 |
On Tue, Jun 7, 2016 at 12:46 AM, Pavel Machek <pavel@ucw.cz> wrote: > On Wed 2016-06-01 21:33:24, dbasehore@chromium.org wrote: >> From: Derek Basehore <dbasehore@chromium.org> >> >> This patch set adds support for catching errors when entering freeze >> on Intel Skylake SoCs. Support for this can be added to newer SoCs in >> later patches. >> >> Verification is done by waking up the CPU once every X (default 10) >> seconds to check the residency of S0ix. This can't be verified before >> attempting to enter S0ix through mwait, so we have to repeatedly >> verify entry into that state. Successfully entering S0ix is no >> guarantee that it will be entered on the next attempt, so we have to >> schedule another check. This has a minimal power impact of <1% of the >> total system power on our systems. > > Dunno. Should this be protected with something like CONFIG_TEST_SLEEP? > People probably don't want this for production... > That depends, if you switch to using suspend to idle instead of suspend to RAM, would you rather not catch power bugs due to misconfigured hardware in production? I agree that it shouldn't be on by default since freeze shouldn't fail because some IP on the SoC doesn't have firmware loaded (this happens with i915), but I was just going to leave it off by default instead of adding yet another config option for a small feature. > -- > (english) http://www.livejournal.com/~pavelmachek > (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [next] | [standalone]
| From | Pavel Machek <pavel@ucw.cz> |
|---|---|
| Date | 2016-06-11 22:40 +0200 |
| Message-ID | <rIXRf-7IN-5@gated-at.bofh.it> |
| In reply to | #1416764 |
On Tue 2016-06-07 17:07:21, dbasehore . wrote: > On Tue, Jun 7, 2016 at 12:46 AM, Pavel Machek <pavel@ucw.cz> wrote: > > On Wed 2016-06-01 21:33:24, dbasehore@chromium.org wrote: > >> From: Derek Basehore <dbasehore@chromium.org> > >> > >> This patch set adds support for catching errors when entering freeze > >> on Intel Skylake SoCs. Support for this can be added to newer SoCs in > >> later patches. > >> > >> Verification is done by waking up the CPU once every X (default 10) > >> seconds to check the residency of S0ix. This can't be verified before > >> attempting to enter S0ix through mwait, so we have to repeatedly > >> verify entry into that state. Successfully entering S0ix is no > >> guarantee that it will be entered on the next attempt, so we have to > >> schedule another check. This has a minimal power impact of <1% of the > >> total system power on our systems. > > > > Dunno. Should this be protected with something like CONFIG_TEST_SLEEP? > > People probably don't want this for production... > > > > That depends, if you switch to using suspend to idle instead of > suspend to RAM, would you rather not catch power bugs due to > misconfigured hardware in production? > I agree that it shouldn't be on by default since freeze shouldn't fail > because some IP on the SoC doesn't have firmware loaded (this happens > with i915), but I was just going to leave it off by default instead of > adding yet another config option for a small feature. I'd rather have testing features optional. In 2 years, hopefully the drivers are debugged, and people can set this to off... Pavel -- (english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web