Comment 74 for bug 425411

#47 works on my Samsung NC-10 as well. I don't think it's a "proper fix" though -- but it does indicate where the problem lies. In particular, the difference between a failing situation and an expected situation is mediated by the presence of a battery event immediately after resume. And apparently SOMETHING is reacting to that battery event in an incorrect way. Perhaps some low-battery auto-suspend functionality that triggers while battery capacity is not yet correctly known? Does anybody have any hints on how to determine what is responding to this "battery event"?