lkml.org 
[lkml]   [2016]   [Dec]   [7]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Subject[PATCH 00/12] PM / OPP: Use kref and move away from RCU locking
Date
Hi Rafael,

You can pretty much ignore this series until all other OPP cleanup/fixes
get merged. I am posting these to get early reviews from Stephen as
these patches have been lying with me for almost a week now. And I am
also _not_ pushing these for 4.10-rc1. It all depends on how the reviews
go.

The RCU locking isn't well suited for the OPP core. The RCU locking fits
better for reader heavy stuff, while the OPP core have at max one or two
readers only at a time.

Over that, it was getting very confusing the way RCU locking was used
within the OPP core. The individual OPPs are mostly well handled, i.e.
for an update a new structure was created and then that replaced the
older one. But the OPP tables were updated directly all the time from
various parts of the core. Though they were mostly used from within RCU
locked region, they didn't had much to do with RCU and were governed by
the mutex instead.

And that mixed with the 'opp_table_lock' has made the core even more
confusing.

Similar concerns were shared by Stephen Boyd earlier [1].

This patchset simplifies the locking in OPP core to great extent using
Kernel reference counting mechanism along with per OPP table mutex.
And finally it gets rid of RCU locking as well.

Each and every patch of this series is individually:
- build tested
- boot tested with cpufreq-dt.ko module. Insmod and rmmod to make sure
the OPPs and the OPP tables are getting freed.

More testing is also done by various build and boot bots for last few
days. And they reported lots of issues (both build and boot time) that
helped making this series more robust:
- Kernel CI (Linaro)
- Fengguang Wu's bot (Intel)

This series has few dependencies though. It is rebased over:
pm/bleeding-edge
+ OPP cleanup series [2]
+ few devfreq fixes [3], [4], and [5].
+ A recent revert [6]

Though all of those shall get merged before we end up reviewing this
series.

--
viresh

[1] https://marc.info/?l=linux-kernel&m=147742717527548&w=2
[2] https://marc.info/?l=linux-pm&m=148108573618896&w=2
[3] https://patchwork.kernel.org/patch/9455789/
[4] https://patchwork.kernel.org/patch/9455757/
[5] https://marc.info/?l=linux-pm&m=148090824301852&w=2
[6] https://marc.info/?l=linux-kernel&m=148110674223377&w=2

Viresh Kumar (12):
PM / OPP: Add per OPP table mutex
PM / OPP: Add 'struct kref' to OPP table
PM / OPP: Return opp_table from dev_pm_opp_set_*() routines
PM / OPP: Take reference of the OPP table while adding/removing OPPs
PM / OPP: Use dev_pm_opp_get_opp_table() instead of _add_opp_table()
PM / OPP: Add 'struct kref' to struct dev_pm_opp
PM / OPP: Update OPP users to put reference
PM / OPP: Take kref from _find_opp_table()
PM / OPP: Move away from RCU locking
PM / OPP: Simplify _opp_set_availability()
PM / OPP: Simplify dev_pm_opp_get_max_volt_latency()
PM / OPP: Update Documentation to remove RCU specific bits

Documentation/power/opp.txt | 47 +-
arch/arm/mach-omap2/pm.c | 5 +-
drivers/base/power/opp/core.c | 888 +++++++++++------------------------
drivers/base/power/opp/cpu.c | 66 +--
drivers/base/power/opp/of.c | 94 +---
drivers/base/power/opp/opp.h | 31 +-
drivers/clk/tegra/clk-dfll.c | 17 +-
drivers/cpufreq/exynos5440-cpufreq.c | 5 +-
drivers/cpufreq/imx6q-cpufreq.c | 10 +-
drivers/cpufreq/mt8173-cpufreq.c | 8 +-
drivers/cpufreq/omap-cpufreq.c | 4 +-
drivers/cpufreq/sti-cpufreq.c | 13 +-
drivers/devfreq/devfreq.c | 14 +-
drivers/devfreq/exynos-bus.c | 14 +-
drivers/devfreq/governor_passive.c | 4 +-
drivers/devfreq/rk3399_dmc.c | 16 +-
drivers/devfreq/tegra-devfreq.c | 4 +-
drivers/thermal/cpu_cooling.c | 11 +-
drivers/thermal/devfreq_cooling.c | 14 +-
include/linux/pm_opp.h | 48 +-
20 files changed, 429 insertions(+), 884 deletions(-)

--
2.7.1.410.g6faf27b

\
 
 \ /
  Last update: 2016-12-07 11:39    [W:0.362 / U:5.504 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site