WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-users

Re: [Xen-users] Dom0 seeing 2Gb, but not 4GB ram

To: Tom Brown <tbrown@xxxxxxxxxxxxx>, Mark Williamson <mark.williamson@xxxxxxxxxxxx>
Subject: Re: [Xen-users] Dom0 seeing 2Gb, but not 4GB ram
From: Mark Lassiter <mackdaddylassiter@xxxxxxxxx>
Date: Tue, 15 Nov 2005 11:47:59 -0800 (PST)
Cc: Xen Users <xen-users@xxxxxxxxxxxxxxxxxxx>
Delivery-date: Tue, 15 Nov 2005 19:48:09 +0000
Domainkey-signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=Message-ID:Received:Date:From:Subject:To:Cc:In-Reply-To:MIME-Version:Content-Type:Content-Transfer-Encoding; b=NCS6RcnyJYztyzA1mAnanBf03aVwze+qw1dRxc20acgeU1Qim7ed3XjeY4EO/5YYIz7idSjmqlkt+UurF+hpF0KdXMkh5hDODmda8P3gdJ7S2+Ja1R869e4NWW+kI1QCBkI7DfVVF6EhSjTwlenchvk0rcnEvUx0nW6DpTvmWjo= ;
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <Pine.LNX.4.44.0511112317560.8669-100000@xxxxxxxxxxxxxxxxxx>
List-help: <mailto:xen-users-request@lists.xensource.com?subject=help>
List-id: Xen user discussion <xen-users.lists.xensource.com>
List-post: <mailto:xen-users@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/cgi-bin/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
That makes sense.  I suppose the mapping of memory to
the CPUs is part of the NUMA architecture, but I'll
not speculate further as I'm outta my league there :)

At some point when our project slows down, I'm going
to revisit the 64bit FC4 and Xen.  After working
through some more problems, I'm beginning to think the
problems were all related to my kernel builds and
mistakes made in that process.  I'll post back when I
have worked it out.

Thanks,
Mark

--- Tom Brown <tbrown@xxxxxxxxxxxxx> wrote:

> >
> > CONFIG_HIGHMEM4G won't fix your problem: access to
> 4Gigs of RAM isn't enough
> > on your system.  See below...
> >
> > > Nov  9 19:02:03 localhost kernel:  BIOS-e820:
> > > 0000000000100000 - 0000000080000000 (usable)
> >
> > OK, that's about 2 gig of your memory, mapped into
> a sane place in the
> > physical address space...
> >
> > > Nov  9 19:02:03 localhost kernel:  BIOS-e820:
> > > 00000000ff780000 - 0000000100000000 (reserved)
> >
> > And this is an enormous hole, taking up the next 2
> gigs of your physical
> > address space.
> >
> > Even with 4G himem, your CPU will only be able to
> access this far: 2Gigs of
> > your real memory, 2Gigs that your BIOS (for some
> reason) has reserved for IO
> > regions.
> >
> > This means that your other 2Gig is mapped too high
> in the physical address
> > space: with 4G himem there is simply no way to
> address it.
> >
> > > Nov  9 19:02:03 localhost kernel:  BIOS-e820:
> > > 0000000100000000 - 0000000180000000 (usable)
> >
> > And here's the other 2Gig.  This is mapped from
> 4Gig-6Gig in your
> physical
> > address space, so isn't mappable without PAE
> support.  With PAE you can
> > easily address this range, so the memory becomes
> usable for you.
> >
> > I've never *seen* a memory hole 2gigs in size. 
> Maybe there's a reason for it
> > or maybe I just had a sheltered upbringing.
> 
> 
> I just booted a box like this :-)
> 
> 
>    Linux version 2.4.29 (root@xxxxxxxxxxxxxxxxxxxx)
> (gcc version 2.96
>    20000731 (Red Hat Linux 7.1 2.96-81)) #3 Wed Oct
> 12 15:11:08 PDT 2005
>    BIOS-provided physical RAM map:
>     BIOS-e820: 0000000000000000 - 000000000009fc00
> (usable)
>     BIOS-e820: 000000000009fc00 - 00000000000a0000
> (reserved)
>     BIOS-e820: 00000000000e8000 - 0000000000100000
> (reserved)
>     BIOS-e820: 0000000000100000 - 000000007fff0000
> (usable)
>     BIOS-e820: 000000007fff0000 - 000000007ffff000
> (ACPI data)
>     BIOS-e820: 000000007ffff000 - 0000000080000000
> (ACPI NVS)
>     BIOS-e820: 00000000ff780000 - 0000000100000000
> (reserved)
>     BIOS-e820: 0000000100000000 - 0000000180000000
> (usable)
>    Warning only 4GB will be used.
>    Use a PAE enabled kernel.
>    3200MB HIGHMEM available.
>    896MB LOWMEM available.
>    On node 0 totalpages: 1048576
> <snip>
>    Memory: 2044840k/4194304k available (1518k kernel
> code, 51860k reserved,
>    487k data, 284k init, 1179584k highmem)
> <snip>
> 
> so, (for the reasons Mark outlined) the box is only
> showing 2 gig
> of memory on a kernel compiled for 4 gig...  However
> it _did_ say
> there was an issue, and what one solution is.
> 
> Now, if you look at the motherboard manual for this
> dual opteron
> system, you will see that half the memory is
> connected to one
> processor, and the other half of the memory is
> connected to the
> other processor. Since I only installed the memory a
> few hours
> ago, the image is fairly fresh in my mind :)
> 
> Anyhow, I thought I would bring it up.
> The motherboard is described here:
> 
>  
>
http://www.supermicro.com/Aplus/motherboard/Opteron/8132/H8DAR-T.cfm
> 
> manual:
>  
>
http://www.supermicro.com/manuals/motherboard/A_8132/MNL-H8DAR-TE.pdf
> 
> I haven't had a chance to do much more than boot it,
> but while I can say
> that the Opteron supermicro products aren't quite up
> to the standard of
> their P4/Xeon mainstream products (they just aren't
> as polished -- the
> fans are loose, there was no system manual, and I'm
> ticked that heatsinks
> seem hard to get), I expect that we (baremetal.com)
> will focus on them...
> 
>  
>
http://www.supermicro.com/Aplus/system/1U/1020/AS-1020A-T.cfm
> 
> I have a pair of 270's and 4 gig of memory
> (hopefully 8 before I
> go into production) in this box. Will report if we
> have any
> trouble (I don't expect to).
> 
> I believe that we can "just" run a 64 bit XEN and 64
> bit kernel
> (e.g. amd64) with our normal 32 bit Linux
> distribution? I will
> know soon enough.
> 
> -Tom
> 
> 
> 



                
__________________________________ 
Yahoo! FareChase: Search multiple travel sites in one click.
http://farechase.yahoo.com

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users