[lkml]   [2014]   [Jul]   [24]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
Subject[PATCH v4 0/5] Simplify hotplug/suspend handling
Series of patchs to simplify policy/sysfs/kobj/locking handling across

The following have been tested so far on a 2x2 cluster environment:
- Boot with 2 cpus and no cpufreq driver.
- mod probe driver and see cpufreq sysfs files show up only for the 1st cluster.
- Online the rest of the 2 CPUs and have files show up correctly.
- rmmod the driver and see the files go away.
- modprobe again (or back and forth multiples times) and see it work.
- suspend/resume works as expected.
- When a cluster is offline, all read/writes to its sysfs files return an error

- Split it up into smaller patches
- Will handle physical CPU removal correctly
- Fixed earlier mistake of deleting code under !recover_policy
- Dropped some code refactor that reuses a lot of code between add/remove
- Dropped fix for exiting hotplug race with cpufreq driver probe/rmmod
- Dropped changes will come later once this series is acked.

Saravana Kannan (5):
cpufreq: Don't wait for CPU to going offline to restart governor
cpufreq: Keep track of which CPU owns the kobj/sysfs nodes separately
cpufreq: Don't destroy/realloc policy/sysfs on hotplug/suspend
cpufreq: Properly handle physical CPU hot-add/hot-remove
cpufreq: Delete dead code related to policy save/restore

drivers/cpufreq/cpufreq.c | 238 ++++++++++++++++++----------------------------
include/linux/cpufreq.h | 1 +
2 files changed, 93 insertions(+), 146 deletions(-)


The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
hosted by The Linux Foundation

 \ /
  Last update: 2014-07-25 03:41    [W:0.377 / U:0.236 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site