Comment 137 for bug 422536

What I find incomprehensible in this «kernel oops» is the fact that it revolves around ECC RAM. Why would the OS be looking to load the kernel module for ECC RAM? My Bios does not contain any visible assignment for ECC RAM, though I believe that the Phenom II memory controller has the capacity to operate ECC RAM if this type of memory is installed. The only setting I find in the bios which might affect the operation of ECC RAM is this «ganged» and «unganged» setting under DRAM Configuration, where as far as I know «ganged» offers ECC RAM better functionality but this is not the same thing as ECC RAM enabled (for this there is no setting and I have the latest Bios for the 790FXT motherboard from Gigabyte.

Still the question remains, why does the os check the bios to see if ECC RAM is enabled? This strikes me as wrong, unless of course the RAM were actually ECC. I know that I am using non-ECC RAM so even if my bios had a setting for ECC RAM this setting would not be enabled and of course this would mean that the module should not be loaded—but then why does this trigger a kernel alert? It doesn't make any sense unless the Kernel thinks that the RAM is ECC and that therefore the kernel module for ECC RAM should be loaded but can't be loaded because the BIOS setting is not enabled—so where would the information that the RAM is ECC be coming from?