public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: "Dr. David Alan Gilbert" <linux@spamtrap.treblig.org>
To: linux-kernel@vger.kernel.org
Subject: kernel ram testing
Date: Sun, 9 Jun 2002 02:10:38 +0100	[thread overview]
Message-ID: <20020609011038.GA1078@gallifrey> (raw)

Hi,
  After a particularly odd experience with a duff dimm a few weeks ago*  I
started thinking again about building continuous RAM testing into the kernel.

I've so far put together a little piece of code that regularly checksums
the kernel text (from _text to _etext).  But now I have some questions:

 1) _text and _etext seem to be defined differently on different
 architectures (some don't define both).  Ideally I think I'd like just
 the kernel text in a way that can be extracted with objdump at build
 time (so I can calculate the sum then for comparison). Suggestions on
 whether _text and _etext are appropriate would be useful.

 2) I want to test the rest of RAM as well - as much as possible,
 repeatedly while the system is running.  I could repeatedly allocate,
 test and deallocate a page. But will I keep getting the same page
 rather than circling through the rest of RAM? Can I ask for particular
 physical pages portably?

 3) I also want to test pages that are being used by the buffers and
 processes; I'd really like to kick a page onto a different physical
 page so I can test it - is this easily possible? (I could lock the
 whole machine down, copy the page somewhere else and then test it, but
 I fear this would take too long).


At the moment my code lives in a kernel thread (kintegrityd) that runs
and then sleeps for a while before running again (at lowest priority
above idle). (I'm presently fiddling with this under user mode linux
2.4.18).

Thoughts and suggestions welcome.

Dave

(*) The DIMM appears, from memtest86, to have had an approximately 4MB
range where one bit in 32bits (or was it perhaps wider) was stuck low.
First noticed by a mismatch on md5sums of a downloaded debian package
and by ' characters appearing in an od I did of the corrupted package on
a machine that otherwise appeared stable.

 ---------------- Have a happy GNU millennium! ----------------------   
/ Dr. David Alan Gilbert    | Running GNU/Linux on Alpha,68K| Happy  \ 
\ gro.gilbert @ treblig.org | MIPS,x86,ARM, SPARC and HP-PA | In Hex /
 \ _________________________|_____ http://www.treblig.org   |_______/

                 reply	other threads:[~2002-06-09  1:10 UTC|newest]

Thread overview: [no followups] expand[flat|nested]  mbox.gz  Atom feed

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:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

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

  git send-email \
    --in-reply-to=20020609011038.GA1078@gallifrey \
    --to=linux@spamtrap.treblig.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox