linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/1] x86: fix early cpu_to_node panic from nr_free_zone_pages
       [not found] <20080121230644.752379000@sgi.com>
@ 2008-01-21 23:06 ` travis
  2008-01-22 11:22   ` Ingo Molnar
  0 siblings, 1 reply; 2+ messages in thread
From: travis @ 2008-01-21 23:06 UTC (permalink / raw)
  To: Andrew Morton, Andi Kleen, mingo
  Cc: Christoph Lameter, linux-mm, linux-kernel

[-- Attachment #1: fix-cpu_to_node-panic --]
[-- Type: text/plain, Size: 1491 bytes --]

An early call to nr_free_zone_pages() calls numa_node_id() which
needs to call early_cpu_to_node() since per_cpu(cpu_to_node_map)
might not be setup yet.

I also had to export x86_cpu_to_node_map_early_ptr because of some
calls from the network code to numa_node_id():

	net/ipv4/netfilter/arp_tables.c:
	net/ipv4/netfilter/ip_tables.c:
	net/ipv4/netfilter/ip_tables.c:

Applies to both:
	
	2.6.24-rc8-mm1
	2.6.24-rc8-mm1 + latest (08/01/21) git-x86 patch

Signed-off-by: Mike Travis <travis@sgi.com>
---
 arch/x86/mm/numa_64.c      |    2 --
 include/asm-x86/topology.h |    2 ++
 2 files changed, 2 insertions(+), 2 deletions(-)

--- a/arch/x86/mm/numa_64.c
+++ b/arch/x86/mm/numa_64.c
@@ -37,9 +37,7 @@ int x86_cpu_to_node_map_init[NR_CPUS] = 
 void *x86_cpu_to_node_map_early_ptr;
 DEFINE_PER_CPU(int, x86_cpu_to_node_map) = NUMA_NO_NODE;
 EXPORT_PER_CPU_SYMBOL(x86_cpu_to_node_map);
-#ifdef	CONFIG_DEBUG_PER_CPU_MAPS
 EXPORT_SYMBOL(x86_cpu_to_node_map_early_ptr);
-#endif
 
 s16 apicid_to_node[MAX_LOCAL_APIC] __cpuinitdata = {
 	[0 ... MAX_LOCAL_APIC-1] = NUMA_NO_NODE
--- a/include/asm-x86/topology.h
+++ b/include/asm-x86/topology.h
@@ -37,6 +37,8 @@ extern int cpu_to_node_map[];
 DECLARE_PER_CPU(int, x86_cpu_to_node_map);
 extern int x86_cpu_to_node_map_init[];
 extern void *x86_cpu_to_node_map_early_ptr;
+/* Returns the number of the current Node. */
+#define numa_node_id()		(early_cpu_to_node(raw_smp_processor_id()))
 #endif
 
 extern cpumask_t node_to_cpumask_map[];

-- 


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH 1/1] x86: fix early cpu_to_node panic from nr_free_zone_pages
  2008-01-21 23:06 ` [PATCH 1/1] x86: fix early cpu_to_node panic from nr_free_zone_pages travis
@ 2008-01-22 11:22   ` Ingo Molnar
  0 siblings, 0 replies; 2+ messages in thread
From: Ingo Molnar @ 2008-01-22 11:22 UTC (permalink / raw)
  To: travis
  Cc: Andrew Morton, Andi Kleen, Christoph Lameter, linux-mm, linux-kernel


* travis@sgi.com <travis@sgi.com> wrote:

> An early call to nr_free_zone_pages() calls numa_node_id() which
> needs to call early_cpu_to_node() since per_cpu(cpu_to_node_map)
> might not be setup yet.
> 
> I also had to export x86_cpu_to_node_map_early_ptr because of some
> calls from the network code to numa_node_id():
> 
> 	net/ipv4/netfilter/arp_tables.c:
> 	net/ipv4/netfilter/ip_tables.c:
> 	net/ipv4/netfilter/ip_tables.c:
> 
> Applies to both:
> 	
> 	2.6.24-rc8-mm1
> 	2.6.24-rc8-mm1 + latest (08/01/21) git-x86 patch

thanks, applied.

	Ingo

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2008-01-22 11:23 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <20080121230644.752379000@sgi.com>
2008-01-21 23:06 ` [PATCH 1/1] x86: fix early cpu_to_node panic from nr_free_zone_pages travis
2008-01-22 11:22   ` Ingo Molnar

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).