I am currently working on porting a copy of Android using the 2.6.37 kernel and the 2.3.2_r1 userland.
During testing we note it goes into a repeated suspend-resume loop where it resumes and then immediately goes back to sleep again. Debugging using a simple shell (rebooting each time is taking too long) shows the same problem, with the system going into an infinite suspend-resume loop and requiring a reset. Looking into it, the resume causes the mmc system to schedule work to detect the card on resume, which causes the tasklet fired off to release the wakelock on the mmc. This release of the wakelock then causes the wakelock driver to decide to send the system back to sleep. Is there any known way to stop this, or is there something else we should be looking into to fix this issue? -- Ben Dooks <[email protected]> http://www.fluff.org/ben/ -- unsubscribe: [email protected] website: http://groups.google.com/group/android-porting
