Weekly Engineering News 36/2008

Andy Green andy at openmoko.com
Thu Sep 11 08:07:04 CEST 2008

Hash: SHA1

Somebody in the thread at some point said:
| Ok, I guess I'll not debate this any further; we'll agree to disagree on
| the chip.
| The real question behind the PMU debate is this -- is the current state
| of power management and suspend in the kernel for the GTA02 the best
| that we can hope to get?  If not, are there any planned efforts to
| resolve the suspend/resume issues, or at least identify them?  (I'm at a
| standstill on my kernel efforts; I lack the knowledge, equipment, and
| documentation to track down these remaining problems.)
| Please don't say that this is going to be like the GTA01 -- left
| partially working, and handed over to the community to sort out while Om
|  focuses on the successor.

I have done plenty of work on this on andy-2.6.26, but the outcome of it
so far is improvements but to push the spotlight on to nasty Glamo problems.

In the end the PMU itself is disconnected from suspend - resume issues,
it does that job pretty well down in hardware (if you discount losing
CPU core power while we are still running :-) ).

Suspend - resume trouble is able to survive us poking at it because it
is very dark in there during this time, we don't really have a clear
idea of what happens when and how many asynchronous actions like
interrupts are flying about.  And if we try to serialize them by
printk(), it changes the behaviour we are looking at.  This is the
thorniest issue left for us.

There is a time when we stop doing real work on GTA02, I'm not going to
kid you otherwise, but that time is not close while we still ship them:
I am highly interested in this issue myself and there is still stuff
that is going to get done in coming weeks for sure.

- -Andy
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org


More information about the devel mailing list