the amd64_edac driver is now the preferred driver for 64-bit AMD processors (over the k8_edac driver).
Created attachment 320887 [details] add the strict_st* functions needed by the amd64_edac driver The amd64_edac driver depends on these functions. They were pulled in from a later mainline tree.
Created attachment 320888 [details] add the amd64_edac driver The amd64_edac driver is the preferred driver for AMD edac. It supercedes the k8_edac driver for revF and earlier.
Created attachment 320889 [details] Allow ref F parts to work correctly This patch is required to let the rev F machines work with the latest amd64_edac driver. Basically, it reorders the driver to determine the node type earlier in the initialization process. This patch has been sent upstream to the edac developer, but has not been committed to SVN yet.
Driver available on mrg-rt.2.6.24.7-93. Tested on sun-x4600m2-01.rhts.bos.redhat.com. --------------------------------------------------------------------------- [root@sun-x4600m2-01 ~]# lsmod | grep edac k8_edac 25544 0 amd64_edac 44384 0 edac_core 52240 11 k8_edac,amd64_edac [root@sun-x4600m2-01 ~]# --------------------------------------------------------------------------- EDAC amd64: This node reports that Memory ECC is currently 'disabled'. EDAC amd64: bit 0x400000 in register 0x40 of the MISC_CONTROL device (0000:00:18.2) should be enabled EDAC amd64: bit 0x10 in MSR register 0x17b of Memory node 0 should be enabled EDAC amd64: This node reports that Memory ECC is currently 'disabled'. EDAC amd64: bit 0x400000 in register 0x40 of the MISC_CONTROL device (0000:00:19.2) should be enabled EDAC amd64: bit 0x10 in MSR register 0x17b of Memory node 1 should be enabled ................ EDAC MC: Rev F or later detected EDAC amd64: 'ecc_enable_override' parameter is active, Enabling AMD ECC hardware now: CAUTION EDAC MC0: Giving out device to 'amd64_edac' 'Athlon64/Opteron/Rev F': DEV 0000:00:18.2 EDAC MC: Rev F or later detected EDAC amd64: 'ecc_enable_override' parameter is active, Enabling AMD ECC hardware now: CAUTION EDAC MC1: Giving out device to 'amd64_edac' 'Athlon64/Opteron/Rev F': DEV 0000:00:19.2 --------------------------------------------------------------------------- Needed to use ecc_enable_override=1 when loading the module, as the ECC flag was not enabled in BIOS. The amd64_edac was attempted loaded first on boot, but due to missing ECC flag, it did not load. The k8_edac was then loaded as a second attempt. I had to unload k8_edac and reload amd64_edac to complete this test.
I have an updated ibm-prtm that does two things: 1. Adds an /etc/modprobe.d/ibm_amd file which has the ecc_emable_override=1 option 2. Blacklists k8_edac using /etc/modprobe.d/blacklist-k8_edac I'll push this package through brew today
An advisory has been issued which should help the problem described in this bug report. This report is therefore being closed with a resolution of ERRATA. For more information on therefore solution and/or where to find the updated files, please follow the link below. You may reopen this bug report if the solution does not work for you. http://rhn.redhat.com/errata/RHSA-2009-0009.html