From: Yinghai Lu <yinghai@kernel.org>
To: Andi Kleen <andi@firstfloor.org>
Cc: Jack Steiner <steiner@sgi.com>, "H. Peter Anvin" <hpa@zytor.com>,
Ingo Molnar <mingo@elte.hu>, Thomas Gleixner <tglx@linutronix.de>,
Andrew Morton <akpm@linux-foundation.org>,
David Rientjes <rientjes@google.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
Rusty Russell <rusty@rustcorp.com.au>,
Mike Travis <travis@sgi.com>
Subject: Re: [PATCH] x86: fix system without memory on node0
Date: Wed, 13 May 2009 08:58:49 -0700 [thread overview]
Message-ID: <4A0AEE39.8030704@kernel.org> (raw)
In-Reply-To: <20090513080010.GN19296@one.firstfloor.org>
Andi Kleen wrote:
>> +#if defined(CONFIG_X86_64) && defined(CONFIG_NUMA)
>> + /*
>> + * make sure boot cpu node_number is right, when boot cpu is on the
>> + * node that doesn't have mem installed
>> + */
>> + per_cpu(node_number, boot_cpu_id) = cpu_to_node(boot_cpu_id);
>> +#endif
>
> Seems like a quite crappy hac^wpatch. Why is it ever set to the wrong
> value? And why is that only the case on NUMA and 64bit?
two places touched that per_cpu(node_number,),
1. in cpu/common.c::cpu_init() and it is not for BP
#ifdef CONFIG_NUMA
if (cpu != 0 && percpu_read(node_number) == 0 &&
cpu_to_node(cpu) != NUMA_NO_NODE)
percpu_write(node_number, cpu_to_node(cpu));
#endif
for BP is traps_init ==> cpu_init
for AP is start_secondary ==> cpu_init
2. cpu/intel.c or amd.c::srat_detect_node via numa_set_node() and they are called via identify_cpu
for BP:
check_bugs ==> identify_boot_cpu ==> identify_cpu() that is rather later before numa_node_id() is used for BP...
for AP:
start_secondary=>smp_callin=>smp_store_cpu_info()=>identify_secondary_cpu ==> identify_cpu()
so only try to set that for BP more early in setup_per_cpu_areas, and don't bother set that for APs there.
(and don't want to mess the 0 before the copying BP per_cpu to APs)
or you check set the per_cpu(node_number) is early enough with setup_per_cpu_areas();
setup_percpu.c in arch/x86/kernel/ is used on lot of conf.
and in arch/x86/include/asm/topology.h, we have
#ifdef CONFIG_NUMA
#include <linux/cpumask.h>
#include <asm/mpspec.h>
#ifdef CONFIG_X86_32
/* Mappings between logical cpu number and node number */
extern int cpu_to_node_map[];
/* Returns the number of the node containing CPU 'cpu' */
static inline int cpu_to_node(int cpu)
{
return cpu_to_node_map[cpu];
}
#define early_cpu_to_node(cpu) cpu_to_node(cpu)
#else /* CONFIG_X86_64 */
/* Mappings between logical cpu number and node number */
DECLARE_EARLY_PER_CPU(int, x86_cpu_to_node_map);
/* Returns the number of the current Node. */
DECLARE_PER_CPU(int, node_number);
#define numa_node_id() percpu_read(node_number)
...
so we need to for NUMA and 64 bit.
YH
next prev parent reply other threads:[~2009-05-13 16:01 UTC|newest]
Thread overview: 90+ messages / expand[flat|nested] mbox.gz Atom feed top
2009-05-09 6:45 [PATCH 1/3] x86: remove MEMORY_HOTPLUG_RESERVE related code Yinghai Lu
2009-05-09 6:48 ` [PATCH 2/3] x86: add numa_move_cpus_to_node Yinghai Lu
2009-05-09 7:05 ` Justin P. Mattock
2009-05-12 1:27 ` Christoph Lameter
2009-05-11 21:53 ` Yinghai Lu
2009-05-12 20:59 ` Christoph Lameter
2009-05-12 17:16 ` Yinghai Lu
2009-05-12 21:21 ` Christoph Lameter
2009-05-13 5:39 ` Yinghai Lu
2009-05-14 19:34 ` Christoph Lameter
2009-05-14 20:58 ` Yinghai Lu
2009-05-09 6:50 ` [PATCH 3/3] x86: fix node_possible_map logic -v2 Yinghai Lu
2009-05-11 17:53 ` Jack Steiner
2009-05-11 19:15 ` Yinghai Lu
2009-05-11 19:36 ` Yinghai Lu
2009-05-11 19:27 ` David Rientjes
2009-05-11 21:12 ` H. Peter Anvin
2009-05-11 21:26 ` Alan Cox
2009-05-11 22:25 ` David Rientjes
2009-05-12 15:06 ` Jack Steiner
2009-05-12 15:10 ` Yinghai Lu
2009-05-12 16:16 ` Jack Steiner
2009-05-12 16:40 ` Yinghai Lu
2009-05-12 18:03 ` Jack Steiner
2009-05-12 21:31 ` Yinghai Lu
2009-05-12 21:58 ` Jack Steiner
2009-05-12 23:13 ` Yinghai Lu
2009-05-12 23:26 ` Yinghai Lu
2009-05-12 15:43 ` Andi Kleen
2009-05-13 1:34 ` [PATCH] x86: fix system without memory on node0 Yinghai Lu
2009-05-13 8:00 ` Andi Kleen
2009-05-13 15:58 ` Yinghai Lu [this message]
2009-05-13 13:35 ` Ingo Molnar
2009-05-13 16:52 ` Jack Steiner
2009-05-13 17:43 ` Yinghai Lu
2009-05-13 18:08 ` Yinghai Lu
2009-05-12 7:15 ` [PATCH 3/3] x86: fix node_possible_map logic -v2 Andi Kleen
2009-05-11 21:33 ` Jack Steiner
2009-05-11 22:56 ` David Rientjes
2009-05-11 23:00 ` Yinghai Lu
2009-05-12 7:09 ` Andi Kleen
2009-05-12 1:02 ` [PATCH 1/3] x86: remove MEMORY_HOTPLUG_RESERVE related code Christoph Lameter
2009-05-12 11:16 ` Mel Gorman
2009-05-13 5:29 ` Yinghai Lu
2009-05-13 9:55 ` Mel Gorman
2009-05-13 6:13 ` [PATCH] x86: remove MEMORY_HOTPLUG_RESERVE related code -v2 Yinghai Lu
2009-05-13 14:59 ` Mel Gorman
2009-05-14 16:38 ` [PATCH 1/5] " Yinghai Lu
2009-05-14 16:40 ` [PATCH 2/5] x86: add numa_move_cpus_to_node Yinghai Lu
2009-05-14 16:41 ` [PATCH 3/5] x86: fix node_possible_map logic -v2 Yinghai Lu
2009-05-18 7:40 ` [tip:x86/mm] x86, mm: Fix node_possible_map logic tip-bot for Yinghai Lu
2009-05-14 16:42 ` [PATCH 4/5] x86: fix system without memory on node0 -v2 Yinghai Lu
2009-05-18 7:40 ` [tip:x86/mm] x86: fix system without memory on node0 tip-bot for Yinghai Lu
2009-05-14 16:43 ` [PATCH 5/5] mm: clear N_HIGH_MEMORY map before se set it again -v2 Yinghai Lu
2009-05-14 16:54 ` Andrew Morton
2009-05-14 17:05 ` Yinghai Lu
2009-05-14 17:25 ` Andrew Morton
2009-05-14 17:34 ` Yinghai Lu
2009-05-14 19:44 ` Christoph Lameter
2009-06-04 5:16 ` [RESEND PATCH] " Yinghai Lu
2009-06-04 16:38 ` Christoph Lameter
2009-06-04 16:48 ` Yinghai Lu
2009-06-04 17:11 ` Christoph Lameter
2009-06-04 17:26 ` [PATCH] mm: clear N_HIGH_MEMORY map before se set it again -v4 Yinghai Lu
2009-06-19 6:42 ` Nathan Lynch
2009-06-19 8:18 ` Yinghai Lu
2009-06-19 8:43 ` Nathan Lynch
2009-06-19 16:16 ` Yinghai Lu
2009-06-20 23:43 ` Yinghai Lu
2009-06-22 4:39 ` Nathan Lynch
2009-06-22 15:38 ` [PATCH] x86: only clear node_states for 64bit Yinghai Lu
2009-06-26 20:54 ` Andrew Morton
2009-06-26 21:09 ` Yinghai Lu
2009-06-27 17:17 ` Ingo Molnar
2009-06-27 20:40 ` Yinghai Lu
2009-06-29 7:39 ` Yinghai Lu
2009-05-18 7:39 ` [tip:x86/mm] mm, x86: remove MEMORY_HOTPLUG_RESERVE related code tip-bot for Yinghai Lu
[not found] ` <20090511095022.GA23121@elte.hu>
[not found] ` <20090511163158.c4e4d334.akpm@linux-foundation.org>
[not found] ` <20090512090704.GC18004@elte.hu>
[not found] ` <4A0A6700.3070100@kernel.org>
[not found] ` <20090513133635.GB7384@elte.hu>
[not found] ` <4A0AFA6E.5050200@kernel.org>
[not found] ` <20090515173521.GA29647@elte.hu>
2009-05-15 21:38 ` tip: patches in git for irq and numa Yinghai Lu
2009-05-18 7:29 ` Ingo Molnar
2009-05-18 13:50 ` Peter Zijlstra
2009-05-18 13:56 ` Ingo Molnar
2009-05-18 15:03 ` Yinghai Lu
2009-05-18 15:09 ` Ingo Molnar
2009-05-18 15:11 ` Peter Zijlstra
2009-05-18 17:23 ` Yinghai Lu
2009-05-19 9:37 ` Ingo Molnar
2009-05-19 10:31 ` Peter Zijlstra
2009-05-19 12:26 ` Ingo Molnar
2009-05-19 9:39 ` [tip:irq/numa] x86, io-apic: Don't mark pin_programmed early tip-bot for Yinghai Lu
2009-05-19 12:30 ` tip-bot for Yinghai Lu
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=4A0AEE39.8030704@kernel.org \
--to=yinghai@kernel.org \
--cc=akpm@linux-foundation.org \
--cc=andi@firstfloor.org \
--cc=hpa@zytor.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=rientjes@google.com \
--cc=rusty@rustcorp.com.au \
--cc=steiner@sgi.com \
--cc=tglx@linutronix.de \
--cc=travis@sgi.com \
/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
Powered by JetHome