LKML Archive on
help / color / mirror / Atom feed
From: Doug Thompson <>
To: Orion Poplawski <>,
Subject: Re: EDAC chipkill messages
Date: Fri, 19 Jan 2007 11:34:52 -0800 (PST)	[thread overview]
Message-ID: <> (raw)
In-Reply-To: <>

--- Orion Poplawski <> wrote:

> Robert Hancock wrote:
> > Orion Poplawski wrote:
> >> Can someone please explain to me what these mean?
> >>
> >> EDAC k8 MC1: general bus error: participating processor(local node
> >> origin), time-out(no timeout) memory transaction type(generic
> read), 
> >> mem or i/o(mem access), cache level(generic)
> >> EDAC MC1: CE page 0xfbf6f, offset 0x4d0, grain 8, syndrome 0xc8f4,
> row 
> >> 1, channel 0, label "": k8_edac
> >> EDAC MC1: CE - no information available: k8_edac Error Overflow
> set
> >> EDAC k8 MC1: extended error code: ECC chipkill x4 error
> >>
> >> Thanks!
> >>
> > 
> > Sounds like you're having some memory ECC errors.. some Memtest86,
> etc. 
> > runs may be in order. You may be able to figure out from this info
> what 
> > DIMM is having the problem.
> > 
> That was my assumption as well, but was hoping someone could decode
> the 
> above information and point me to the problem chip.  I ran Memtest86 
> overnight but found no problems, but don't know if it needs to run in
> a 
> particular ECC mode.
> This is a dual proc 275 system with 4 1GB DIMMs.  Guessing that MC1
> is 
> the controller on the second CPU.  Would row 1 be the second DIMM?

No that would be the FIRST DIMM, on Channel 0

Each DIMM has 2 ChipSelect Rows (CSROW)

Each csrow covers two channels across, therefore on a 4 socket memory
array, there are CSROWS 0 and 1 on the first DIMM row and CSROWS 2 and
3 on the second DIMM row.


The W and the Y DIMMs are channel 0
The X and the Z DIMMs are channel 1

csrows 0 and 1 would cross over Y and Z DIMMs
csrows 2 and 3 would cross over W and X DIMMs

The mapping problem occurs in then identifying each of the above goes
to which silk screen labeled sockets on the mobo.

Usually they are labeled:


where A is channel 0 and B is channel 1 and
the "DIMM1" would indicate the CSROWs 0 and 1
and "DIMM2" would indicate the CSROWs 2 and 3

The string 'label ""' can be filled in by a userspace script to
properly identify the DIMM silk screen according to the motherboard

The lines with "EDAC MC1:" are EDAC CORE output messages, while the
"EDAC K8:" lines are EDAC Memory Controller driver messages.
"CE" is correctable error 
MC1 is memory controller 1 (0 based)

ECC ChipKill x4 was what found the error and corrected it.

The FRU, (field replaceable unit) is the DIMM located at socket
H1_DIMM1A, according to the labeling I mentioned above.

caveat: the detector is not 100% perfect but gives a general area to
look at, the DIMM specification. Sometimes other errors can cause what
looks like a memory error, but usually a bad memory DIMM is the root
cause of the vast majority of such errors.

In addition, memtest86+ doesn't find all the bad memory in all cases,
but it is still a VERY useful tool

doug thompson

  reply	other threads:[~2007-01-19 19:34 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <fa.CZkzgccjpiJW6cRrbtRvg/>
2007-01-19  0:12 ` Robert Hancock
2007-01-19 16:45   ` Orion Poplawski
2007-01-19 19:34     ` Doug Thompson [this message]
2007-01-18 22:55 Orion Poplawski

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \ \ \ \ \
    --subject='Re: EDAC chipkill messages' \

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).