Hi Boris, On 30/08/18 11:43, Borislav Petkov wrote: > On Wed, Aug 29, 2018 at 06:33:52PM +0000, Fan Wu wrote: >> The current ghes_edac driver does not update per-dimm error >> counters when reporting memory errors, because there is no >> platform-independent way to find DIMMs based on the error >> information provided by firmware. This patch offers a solution >> for platforms whose firmwares provide valid module handles >> (SMBIOS type 17) in error records. In this case ghes_edac will >> use the module handles to locate DIMMs and thus makes per-dimm >> error reporting possible.
> If we're going to do this, it needs to be tested on an x86 box which loads > ghes_edac. Adding Toshi to Cc. Good point, thanks. > Otherwise it must remain ARM-specific. Hmmm, that would be a shame. This should only be a problem if HPE Servers set CPER_MEM_VALID_MODULE_HANDLE, but don't actually provide module handles, or if firmware has a different idea of what they are. If firmware never sets CPER_MEM_VALID_MODULE_HANDLE, this patch shouldn't change anything. (Someone must have an x86 that sets CPER_MEM_VALID_MODULE_HANDLE, otherwise the code wouldn't be there right?!) Thanks, James