On Sun, 26 Feb 2012, Ben Hutchings wrote:
On Thu, 2012-02-23 at 17:59 -0800, Shannon Dealy wrote: [...]some point, outside software MUST be providing bogus information to the driver. I say this because after the deep sleep bug occcurs and the hardware has been power cycled (through hibernate), the device driver and firmware have been reloaded and right from the start they show the "deep sleep" state again.For a complete power-cycle, you may need to remove both the power cord and the battery. I don't think the Intel wireless cards have any support for wake-on-WAN, but in general devices may still be partly powered as long as any power source is connected to the system.
True enough, and I don't remember if I pulled the battery back when I was going after this problem (it has been almost a year since I did all my tests). There is also a switch on the side that kills all RF devices (WLAN and Bluetooth) but I can't verify that it completely kills power to the devices.
The log messages you sent are indicative of a total failure of communication with the card. My suspicion would be that the hardware has developed a fault, but it could be as simple as the card being slightly loose in its slot.
Hardware failure might make sense except that it has always behaved this way, others have similar symptoms, and most importantly, regardless of what behaviors are occurring (there are a number of problems this card exhibits), while hibernate won't clear the "deep sleep" problem, rebooting the system (without power cycling) clears the problems EVERY time. Whatever the trigger (intermittent hardware fault or driver bug), it appears the kernel is getting into a state from which it is incapable of recovering.
Having said that, are you setting the pcie_aspm kernel parameter?
No, wasn't even aware of its existance. System shows it is currently at "default" unfortunately, the debug kernel I am currently running has the pcie_aspm regression which prevents it from being changed without rebooting which is rather time consuming with my setup. There doesn't seem to be an obvious way to tell what the "default" is.
I should note that I have not seen the "deep sleep" problem in many months, though it is hard to say when or if it has stopped since it often took weeks to show up and my usage patterns have changed considerably. In particular I noticed that hibernate for my system seemed to be a significant contributor to triggering the bug since the problem often showed up in the first few minutes after bringing the computer out of hibernation. This is not a requirement to trigger the problem, just significantly increased the odds.
NOTE: I stopped using hibernate in large part because of the "deep sleep" bug. Unfortunately, the other problems of this card continue to plague me (chronic random communication failures being the big one).
Shannon C. Dealy | DeaTech Research Inc. dealy@deatech.com | - Custom Software Development - Phone: (800) 467-5820 | - Natural Building Instruction - or: (541) 929-4089 | www.deatech.com