Created attachment 521621 [details]
some debug info collected in the error situation. All HW registers are FF or 00.
Description of problem:
After suspend/resume, sometimes the sky2 LAN chip in my Fujitsu Esprimo Mobile M9400 fails to recover. Last time this happened, all HW registers were either 00 or FF.
Version-Release number of selected component (if applicable):
126.96.36.199-27.fc15.x86_64, sky2 1.28
Sometimes (occured 2x so far, estimated frequency ~1/30 cycles)
Steps to Reproduce:
[404256.145060] sky2 0000:04:00.0: eth0: disabling interface
[404260.875063] PM: Entering mem sleep
[404260.875092] Suspending console(s) (use no_console_suspend to debug)
[404261.330055] PM: suspend of devices complete after 454.409 msecs
[404261.514114] PM: late suspend of devices complete after 184.054 msecs
[404261.514548] ACPI: Preparing to enter system sleep state S3
[404261.514698] PM: Saving platform NVS memory
[404261.515344] Disabling non-boot CPUs ...
[404261.520058] kvm: disabling virtualization on CPU1
[404261.520161] CPU 1 is now offline
[404261.520704] PM: Restoring platform NVS memory
[404261.748021] sky2 0000:04:00.0: Refused to change power state, currently in D3 (***)
[404261.748032] sky2 0000:04:00.0: restoring config space at offset 0xf (was 0xffffffff, writing 0x10a)
[404261.748038] sky2 0000:04:00.0: restoring config space at offset 0xe (was 0xffffffff, writing 0x0)
[404294.102327] sky2 0000:04:00.0: eth0: phy I/O error
[404294.102333] sky2 0000:04:00.0: eth0: phy I/O error
The problem occured apparently at (***).
188.8.131.52-27.fc15.x86_64 is fairly old at this point and F15 is going EOL. If you still see this with the 3.3 kernels in F16/F17 please open a new bug.