Sorry Sasha ,
I will rewrite and test this patch before I send it in.
Cheers Nick

On Sun, Jun 29, 2014 at 4:24 PM, Sasha Levin [off-list ref] wrote:

On 06/28/2014 11:03 PM, Nick Krause wrote:

quoted

I thought I compiled tested it maybe not. I thought M= would work for
alpha directory but maybe I have to enable in on x86.
I don't known if there is a way to do this.
Cheers Nick

On Sat, Jun 28, 2014 at 3:53 PM, Sasha Levin [off-list ref] wrote:

quoted

On 06/28/2014 12:07 AM, Nicholas Krause wrote:

quoted

This patch fixs the FIXME message in the function *cpumask_of_node
for using this function multiple times and the issue with recaluting
the cpu node mask when reusing this function.

Signed-off-by: Nicholas Krause <redacted>
---
 arch/alpha/include/asm/topology.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/arch/alpha/include/asm/topology.h b/arch/alpha/include/asm/topology.h
index 9251e13..d301f66 100644
--- a/arch/alpha/include/asm/topology.h
+++ b/arch/alpha/include/asm/topology.h

@@ -31,6 +31,9 @@ static const struct cpumask *cpumask_of_node(int node)
      if (node == -1)
              return cpu_all_mask;

+     else if (node == &node_to_cpumask_map[node])
+             return &node_to_cpumask_map[node];
+
      cpumask_clear(&node_to_cpumask_map[node]);

      for_each_online_cpu(cpu) {

Since this patch doesn't even compile, care to tell us how you tested it?


Thanks,
Sasha

Please stop top posting.

Your explanation is missing the part about actually making sure that your patch
works is missing, how did you intend to make sure that your patch doesn't
break things?


Thanks,
Sasha