aboutsummaryrefslogtreecommitdiff
path: root/arch
diff options
context:
space:
mode:
authorNathan Lynch <ntl@pobox.com>2005-06-28 20:55:25 +1000
committerPaul Mackerras <paulus@samba.org>2005-06-28 20:55:25 +1000
commitf5f1cc5437961a4bd93d615099f26780819e72d3 (patch)
tree7daa443dbcb5cda9e057a818168102abe48d3b8c /arch
parenta341ad97245d01c923995cfe7deacd0c8aee6e16 (diff)
[PATCH] ppc64: don't create spurious symlinks under node0 sysdev
On partitioned systems we can wind up creating spurious symlinks in /sys/devices/system/node/node0 to non-present cpus. The symlinks are not broken; the problem is that we're potentially misinforming userspace that there is a relationship between node0 and cpus which are to be added later. There's no guarantee at all that a cpu which is added later will belong to node 0. Signed-off-by: Nathan Lynch <ntl@pobox.com> Signed-off-by: Paul Mackerras <paulus@samba.org>
Diffstat (limited to 'arch')
-rw-r--r--arch/ppc64/kernel/sysfs.c7
1 files changed, 6 insertions, 1 deletions
diff --git a/arch/ppc64/kernel/sysfs.c b/arch/ppc64/kernel/sysfs.c
index c8fa6569b2f..2f704a2cafb 100644
--- a/arch/ppc64/kernel/sysfs.c
+++ b/arch/ppc64/kernel/sysfs.c
@@ -400,7 +400,12 @@ static int __init topology_init(void)
struct cpu *c = &per_cpu(cpu_devices, cpu);
#ifdef CONFIG_NUMA
- parent = &node_devices[cpu_to_node(cpu)];
+ /* The node to which a cpu belongs can't be known
+ * until the cpu is made present.
+ */
+ parent = NULL;
+ if (cpu_present(cpu))
+ parent = &node_devices[cpu_to_node(cpu)];
#endif
/*
* For now, we just see if the system supports making