Det råder ju lite delade meningar om hur pass "fixad" aos-buggen verkligen är. Så här på rak arm låter det ju ganska fantastiskt om en ensam dev skulle lyckats med det som varken Google eller Samsung klarar av, utan att kapa ganska ordentligt i saker som utvecklarna tydligen anser bör vara med i koden.
"XDA member AndreiLux wrote up a fantastic summary of all his findings on the mythical Android OS bug. For those of you new to Android, or who have never been too concerned with battery usage, there is a process that normally shows under battery usage in Android>Settings called Android OS. This process normally has a high percentage value attached to it, which means if your battery is draining fast, this little guy has likely been responsible for it. AndreiLux explains that after using a few CPU monitoring programs to determine battery consumption more accurately, he found that this process may be related to the deep sleep state your device should reach after the screen goes off. What does happen seems to be an internal conflict, sending your device into sort of an infinite loop while attempting to sleep, consuming tons of battery in the process. After a bit more digging around, the problem appears related to the kernel, as many observed this after 2.6.35.14 came out.
There is no real fix for this, although many claim that some of the fixes out there make things better. According to the dev, the fixes only mask the process, hiding it from sight, but don’t actually kill it. Finally, the dev has posted a few work-arounds to try out, if you feel adventurous enough."
Läs mer på XDA
Som AndreiLux säger:
"The above SiyahKernel [min anm] "fix" has been proven to be not really a fix, but a change of behavior of the bug. It's still present and the drain still prevails. I'd like to rephrase that last sentence, it still prevails for some people, and it is in my opinion not a complete fix of the problem, but yet still appears that it changes it enough for lots of people that it doesn't trigger anymore. [14/10]As more time passes and testing is done, it seems more and more apparent that it's just a masking of the problem that's being done.
Tha change has been pinpointed to the patching up to the Linux 2.6.35.14 kernel, so any custom kernel implementing those patches will now observe the new behavior."
Längre ner i XDA-tråden kan man läsa om att de börjar misstänka att en del av problemen kan vara så enkelt som dåligt skrivna android program, som t.ex. Skype. 92 vlan_wakes på 8 minuter kan ju inte vara bra
Last edited: Oct 18, 2011