veyepar: The Future of Power Management in Linux

Author(s):	Rafael J. Wysocki
Location	Boardwalk Gallery
Date	jan Wed 18	Days Raw Files
Start	21:15	First Raw Start	20:48
Duration	0:45:0	Offset	0:26:01
End	22:00	Last Raw End	22:18
Chapters	00:00 0:03:52 0:33:51
Total cuts_time	45 min.

http://linux.conf.au/schedule/presentation/30/
raw-playlist raw-mp4-playlist encoded-files-playlist host archive tweet mp4 svg png

assets release.pdf The_Future_of_Power_Management_in_Linux.json
logs
Admin: episode episode list cut list raw files day marks day marks day image_files

State:

Locked:

clear this to unlock

Locked by:

user/process that locked.

Start:

initially scheduled time from master, adjusted to match reality

Duration:

length in hh:mm:ss

Name:

Video Title (shows in video search results)

Emails:

email(s) of the presenter(s)

Released:

has someone authorised pubication

Normalise:

Channelcopy:

m=mono, 01=copy left to right, 10=right to left, 00=ignore.

Thumbnail:

filename.png

Description:

markdown

Power management has been an important part of Linux kernel development for the last several years. Support for many important power management features, such as system suspend and hibernation, CPU idle states management, CPU frequency management, runtime power management of I/O devices and more, has been added to the kernel during that time. However, there still are challenges facing the Linux power management developers.
The first challenge is that the majority of power management features available in the kernel today were developed in isolation and while they work reasonably well individually, they may not play so well with each other. At the same time, hardware design trends lead to more and more integration and there is a clear need for the Linux power management features to work more closely together too. For this reason, there have been efforts to integrate runtime PM of I/O devices with system suspend and hibernation for some time. There also is work in progress on integrating CPU power management with the CPU scheduler. All of that work will likely continue into the future.
Further challenges are related to new types of hardware that become available, like persistent memory. Traditional power management features, like system suspend and hibernation, will have to be rethought and modified in response to that and some other changes will likely become necessary in the power management area.
There also are challenges coming from new display technologies and networking that require relatively high power to be allocated to the transmission of data alone putting the other components of the system under power pressure and causing them to be effectively underutilized. Similar problems appear when system are thermally challenged or when their power supplies are not capable of supporting all of the system components at full capacity at the same time.
All of this means that the future of power management in Linux will certainly be interesting. We have many interesting problems to solve and quite a lot of work to do, but hopefully power management in Linux will work really well going forward.

Comment:

production notes