[meta-intel] Crystalforest: cores not responding and ethernet not found

[Ben Warren]

Hi,

I'm new to both yocto and embedded x86, so apologize in advance if I get some terminology wrong.

We have a new board based on the crystal forest/gladden chipset (E3-1125C @ 2GHz).  I've built the meta-crystalforest BSP (danny version) and am trying to get it to boot from a thumb drive.

Occasionally (maybe 1 in 10 times) this will boot all the way to the login prompt, but more often I see this, where at most 2 cores come up:

===
CPU0: Intel(R) Xeon(R) CPU  @ 2.00GHz stepping 07
Performance Events: PEBS fmt1+, 16-deep LBR, SandyBridge events, Intel PMU driver.
PEBS disabled due to CPU errata.
... version:                3
... bit width:              48
... generic registers:      4
... value mask:             0000ffffffffffff
... max period:             000000007fffffff
... fixed-purpose events:   3
... event mask:             000000070000000f
Booting Node   0, Processors  #1
CPU1: Not responding.
 #2
CPU2: Not responding.
 #3
CPU3: Not responding.
 #4 #5
CPU5: Not responding.
 #6
CPU6: Not responding.
 #7 Ok.
CPU7: Not responding.
Brought up 2 CPUs
===

and then it hangs here:

===
ALSA device list:
  No soundcards found.
===

With this BSP build, no Ethernet controllers are ever found.

If I boot Fedora core 14, I see similar issues with non-responding cores, but two ixge 10GbE controllers are found.

This board uses an AMI BIOS.  It seems a bit better with power management disabled in BIOS, and I've tried several different ACPI settings.

Does anybody have insight as to what is going on?  Should I be focusing on BIOS/UEFI?  We've verified all Ethernet controllers using UEFI applications, and at least one of the 10/100/1000 controllers should have link up, so it's surprising to see nothing found by Linux.

thanks,
Ben