1. 21 12月, 2011 1 次提交
  2. 14 12月, 2011 1 次提交
  3. 01 11月, 2011 3 次提交
    • M
      EDAC: Fix incorrect edac mode reporting in sb_edac · c6e13b52
      Mark A. Grondona 提交于
      The edac driver for Sandy Bridge was found to be reporting "FPM"
      for edac_mode, which clearly doesn't make sense. It was found that
      sb_edac.c:get_dimm_config was reusing a variable for both mem_type
      and edac_type, and thus was overwriting the value after setting
      it correctly. This patch fixes that issue.
      
      Before the patch:
      /sys/devices/system/edac/mc/mc0/csrow0/edac_mode:FPM
      /sys/devices/system/edac/mc/mc0/csrow1/edac_mode:FPM
      /sys/devices/system/edac/mc/mc0/csrow2/edac_mode:FPM
      /sys/devices/system/edac/mc/mc0/csrow3/edac_mode:FPM
      
      After:
      /sys/devices/system/edac/mc/mc0/csrow0/edac_mode:S4ECD4ED
      /sys/devices/system/edac/mc/mc0/csrow1/edac_mode:S4ECD4ED
      /sys/devices/system/edac/mc/mc0/csrow2/edac_mode:S4ECD4ED
      /sys/devices/system/edac/mc/mc0/csrow3/edac_mode:S4ECD4ED
      Signed-off-by: NMark A. Grondona <mgrondona@llnl.gov>
      Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>
      c6e13b52
    • M
      edac: sb_edac: Add it to the building system · 3d78c9af
      Mauro Carvalho Chehab 提交于
      Some changes on it were required due to changeset cd90cc84c6bf0, that
      changed the glue with the MCE logic.
      Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>
      3d78c9af
    • M
      edac: Add an experimental new driver to support Sandy Bridge CPU's · eebf11a0
      Mauro Carvalho Chehab 提交于
      This driver is known to work on mine and Tony's test environments,
      using software error injection, and a partial hardware/software
      error injection tool.
      
      There's no broader range test yet to double check if the error decoding
      logic will actually point to the right DIMM, so use it with care.
      More tests are required to be sure that the driver will work on all
      different types of memory configurations.
      
      If you're willing to risk using it, I suggest you to enable EDAC debugs
      for your test machines, as the debug logs helps to track what's going
      inside the driver.
      
      Please feed me with bug reports, if you notice that the driver
      is miss-behaving.
      Tested-by: NTony Luck <tony.luck@intel.com>
      Signed-off-by: NMauro Carvalho Chehab <mchehab@redhat.com>
      eebf11a0