> I have a similar issue with the X220, the problem is a watchdog > timer, > that I suspect is in the Intel ME. It expires without being reset > and > forces the machine to restart. Or at least that is the cause of > that > happening on my X230's. I've ripped a few of them apart and > analyzed > their guts and found only the CPU and a few other chips are active > during suspend. I've probed all the buses of those other chips and > none > make a peep when the machine reboots, the only chip left active is > the > Intel ME chunk of the CPU, and for obvious reasons, I have no idea > what > it is doing, so I suspect it is the culprit.
I think there is at least some aspect of software at play here however. I did not experience these issues while running Debian 9 on the machine. It could be that Linux uses some horrible hack to make suspend work reliably, but it does nevertheless work. > I gave up on the work a few months ago since it seemed easier to > just > accept that suspend isn't going to work and just use suspend-to-disk > or > just shut the machine down completely. I had intended to use suspend-to-disk with this machine, but I found that applications that use hardware acceleration (namely Firefox) do not function after resuming from suspend to disk. The specific symptom is that the application's window is just black with no visible contents. Restarting it does nothing. This is very likely a problem with inteldrm. Disabling hardware acceleration in FF fixes the problem, but makes it almost unusably slow. > If you want to do more, and have > access to a Windows machine, you can try pulling apart the Lenovo > drivers to see what the Lenovo-specific ACPI driver is doing when > the > machine goes into suspend. I don't, but I had planned to throw Windows on a spare disk and see if updating the firmware / BIOS / playing with the proprietary driver helps or yields any useful information. ... Maybe I should look at running coreboot on the T430, since it's supported now. Thanks for your detailed response!
