1 Reply Latest reply on Oct 30, 2015 11:03 AM by ricardog_intel

    Tylersburg 55X0 chipset revision


      I recently acquired a new (to me) HP Z800 workstation, with a later model revision motherboard, specifically the updated motherboard (460838-003) that supports the Westmere-EP series processors (56XX):



      I removed the heatsink for the dual IOH to replace the thermal compound, and marked down the chip markings which were as such

      One was "AC5520 / SLH3P / J032C550 / (m)(c)'08"

      The other was "AC5520 / SLH3P / J032C550 / (m)(c)'08"

      (i.e. they were the same)


      I also updated the BIOS to the latest available from HP (3.57, 07/15/2013)


      Now, according to the "Intel 5520 and Intel 5500 Chipset Specification Update" (http://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/5520-and-5500-chipset-ioh-specificatio…), these do in fact correspond to the C-2 stepping (that supports interrupt remapping without errata 47 and 53.)


      However, when I boot the system, the Host bridge (PCI ID 8086:3406) comes up with a revision ID of 0x13h, which corresponds to the revision of the B-3 stepping. The Linux kernel, seeing a 8086:3406 rev 0x13h, assumes this is chipset stepping B-3, notes that interrupt remapping is enabled, warns and then disables it.


      Now I'm really confused - I would think the chip marking is authoritative, but the PCI revision is clearly that of the B-3 stepping. Is the PCI configuration space part of a block provided by Intel, and HP just never included the revised configuration space? Or is something else going on? Should I just defeat the check in the kernel?