View Single Post
olf's Avatar
Posts: 304 | Thanked: 1,246 times | Joined on Aug 2015
#10
Originally Posted by marmistrz View Post
Yes, https://build.merproject.org/package...r:core/libiphb
Thanks for the pointer.

[...]

mcetool -sdisabled

Originally Posted by marmistrz View Post
[...]
And there was `mcetool -searly`. This should only enable the early-suspend. Do you have any references on what that actually does?
As of mcetool v1.90.0.1 (© 2011; per mcetool -V) on SailfishOS 2.1.3.7:
mcetool -s, --set-suspend-policy=<enabled|disabled|early|disable_on_charger>

The webpage you found, ...
Originally Posted by marmistrz View Post
[...]
And this post, claiming that early-suspend was removed: https://plus.google.com/111524780435...ts/RCV8EP3hFEm
... provides the best description I have read yet (while saying Goodbye to this feature), better than https://together.jolla.com/question/...#post-id-55138 and all the linked webpages there.

But by practical experience, -senabled and -searly do not differ significantly in battery runtime, if anything at all.
Also the specific AlienDalvik issue, I tested -searly for, was unaltered: "Android apps running in the background (e.g. K9-Mail, Conversations) loose their network connection permanently, after SailfishOS automatically switching from mobile network (3G) to WLAN (but not vice versa). A restart of AlienDalvik resolves this, a restart of the SailfishOS network services does not!".

AFAIU, in theory -searly should keep CPU and I/O active in its lowest power state / sleep mode, while -senabled may suspend CPU and peripherals completely (basically like a power off), only guaranteeing RAM content to be retained ("Suspend to RAM/ StR").

Well, as Jolla 1 phones run Linux 3.4.108 (specifically such an "Android kernel" as discussed in the GooglePlus article), this feature is probably also not present in the SailfishOS kernel.
Still I wonder, if it causes any difference in the Power Management framework?

One may try to observe suspend actions statistically via mcetool --get-suspend-stats (with a rather slim output IMO and unfortunately -v does not seem to make a difference).

Last edited by olf; 2018-01-14 at 04:33.
 

The Following 2 Users Say Thank You to olf For This Useful Post: