From: Ravikiran G Thirumalai <kiran@scalex86.org>
To: Andi Kleen <ak@suse.de>
Cc: Andrew Morton <akpm@osdl.org>,
linux-kernel@vger.kernel.org, discuss@x86-64.org,
shai@scalex86.org
Subject: Re: [patch 3/3] x86_64: Node local PDA -- allocate node local memory for pda
Date: Fri, 2 Dec 2005 12:02:34 -0800 [thread overview]
Message-ID: <20051202200234.GB3727@localhost.localdomain> (raw)
In-Reply-To: <20051202114708.GM997@wotan.suse.de>
On Fri, Dec 02, 2005 at 12:47:09PM +0100, Andi Kleen wrote:
> On Fri, Dec 02, 2005 at 12:23:09AM -0800, Ravikiran G Thirumalai wrote:
> > Patch uses a static PDA array early at boot and reallocates processor PDA
> > with node local memory when kmalloc is ready, just before pda_init.
> > The boot_cpu_pda is needed sice the cpu_pda is used even before pda_init for
> > that cpu is called (to set the static per-cpu areas offset table etc)
>
> Where is it needed? Perhaps it should be just allocated in the
> CPU triggering the other CPU start instead. Then you could avoid that
> or rather only define a __initdata boot_pda for the BP.
>
setup_per_cpu_areas() is invoked quite early in the boot process and it
writes into the cpu_pda.data_offset field for all the cpus. I'd even tried
storing the offset table for cpus in a temporary table (which can be marked
__initdata and discarded later), but there were references
to the static per cpu areas through per_cpu macros (which need to use the
cpu_pda) even before the BP boots up and starts the secondary cpus,
resulting in early exceptions.
> >
> > Index: linux-2.6.15-rc3/arch/x86_64/kernel/head64.c
> > ===================================================================
> > --- linux-2.6.15-rc3.orig/arch/x86_64/kernel/head64.c 2005-11-30 17:01:18.000000000 -0800
> > +++ linux-2.6.15-rc3/arch/x86_64/kernel/head64.c 2005-11-30 17:07:14.000000000 -0800
> > @@ -80,6 +80,7 @@
> > {
> > char *s;
> > int i;
> > + extern struct x8664_pda boot_cpu_pda[];
>
> externs only belong in include files.
Yes, I will change this and resubmit
Thanks,
Kiran
next prev parent reply other threads:[~2005-12-02 20:02 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2005-12-02 8:10 [patch 1/3] x86_64: Node local PDA -- early cpu_to_node Ravikiran G Thirumalai
2005-12-02 8:16 ` [patch 2/3] x86_64: Node local PDA -- Use macros to access cpu_pda Ravikiran G Thirumalai
2005-12-02 8:23 ` [patch 3/3] x86_64: Node local PDA -- allocate node local memory for pda Ravikiran G Thirumalai
2005-12-02 8:54 ` Eric Dumazet
2005-12-02 18:24 ` [discuss] " Ravikiran G Thirumalai
2005-12-02 9:05 ` Andrew Morton
2005-12-02 10:36 ` [RFC] NUMA aware kthread_create() ? Eric Dumazet
2005-12-02 12:32 ` [discuss] " Andi Kleen
2005-12-02 11:47 ` [patch 3/3] x86_64: Node local PDA -- allocate node local memory for pda Andi Kleen
2005-12-02 20:02 ` Ravikiran G Thirumalai [this message]
2005-12-02 22:41 ` Andi Kleen
2005-12-02 11:43 ` [patch 1/3] x86_64: Node local PDA -- early cpu_to_node Andi Kleen
2005-12-02 22:51 ` [discuss] " Ravikiran G Thirumalai
2005-12-02 23:02 ` Andi Kleen
2005-12-02 23:43 ` Ravikiran G Thirumalai
2005-12-02 23:48 ` Andi Kleen
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=20051202200234.GB3727@localhost.localdomain \
--to=kiran@scalex86.org \
--cc=ak@suse.de \
--cc=akpm@osdl.org \
--cc=discuss@x86-64.org \
--cc=linux-kernel@vger.kernel.org \
--cc=shai@scalex86.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