Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 4 May 2009 10:34:45 +0400
From:      Eygene Ryabinkin <rea-fbsd@codelabs.ru>
To:        Robert Noland <rnoland@FreeBSD.org>
Cc:        freebsd-current@freebsd.org
Subject:   Re: Patch for "device_attach: estX attach returned 6" on half of the cores
Message-ID:  <z3eD9vm%2BPfdrvO13WMzgYPrBP6U@zJPT6v7ej2A2kHQ7pL5rJQfsxXs>
In-Reply-To: <1241418372.78715.1.camel@balrog.2hip.net>
References:  <o4xm64iOwjCQABT%2BLyYf4J6bdds@psUvbB26hX94GoQd3C5ThNBxOp8> <1241418372.78715.1.camel@balrog.2hip.net>

next in thread | previous in thread | raw e-mail | index | archive | help
Robert, good day.

Mon, May 04, 2009 at 01:26:12AM -0500, Robert Noland wrote:
> On Mon, 2009-05-04 at 09:22 +0400, Eygene Ryabinkin wrote:
> > Good day.
> > 
> > Recently I had filed a PR,
> >   http://www.freebsd.org/cgi/query-pr.cgi?pr=134192
> > that should heal the situation when estX isn't attached properly
> > on the half of processor cores (the odd ones).  I am seeing this
> > only on Asus MBs, but this could show up on other hardware as well.
> > 
> > If anyone sees such symptoms, I encourage them to test the patch.
> > The patch itself contained in the PR, but for ease I had put it
> > there,
> >   http://codelabs.ru/fbsd/patches/acpi/attach-children-without-aliases.diff
> > and will sync PR's one and this one in the case of any changes.
> 
> I'm seeing this on both an ASUS and an Intel board that I have.  Both
> have core2duo E7400's in them.  I have applied this patch to the Intel
> board so far, but unfortunately now both cores fail equally.
> 
> cpu0: <ACPI CPU> on acpi0
> est0: <Enhanced SpeedStep Frequency Control> on cpu0
> est: CPU supports Enhanced Speedstep, but is not recognized.
> est: cpu_vendor GenuineIntel, msr 6164a2306004a23
> device_attach: est0 attach returned 6
  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

This means that most likely you don't have _PSS entries in your DSDT
table at all.  The patch helps in the case when some processor names are
aliased _and_ half of processors are attached properly even without this
patch.

One could check if the patch will help by examining the output of
'apicdump -d' and looking for the Alias () directives for the Processor
objects.  Here's what I have for my laptop:
-----
    Scope (_PR)
    {
        Processor (P001, 0x01, 0x00000810, 0x06) {}
        Alias (P001, CPU1)
    }

    Scope (_PR)
    {
        Processor (P002, 0x02, 0x00000810, 0x06) {}
        Alias (P002, CPU2)
    }
-----
So in this case we essentially have 4 Processor objects under _PR,
but only two objects are real ones and only they should be attached.

For the completeness, could you, please, show the output of 'acpidump
-dt 2>&1' for both of your machines?

Thanks!
-- 
Eygene
 _                ___       _.--.   #
 \`.|\..----...-'`   `-._.-'_.-'`   #  Remember that it is hard
 /  ' `         ,       __.--'      #  to read the on-line manual
 )/' _/     \   `-_,   /            #  while single-stepping the kernel.
 `-'" `"\_  ,_.-;_.-\_ ',  fsc/as   #
     _.-'_./   {_.'   ; /           #    -- FreeBSD Developers handbook
    {_.-``-'         {_/            #



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?z3eD9vm%2BPfdrvO13WMzgYPrBP6U>