Benjamin Herrenschmidt wrote:

quoted

I don't know quite the detail of the new cpumask stuff ... It could be
as simple as passing a pointer instead of the value in the
cpumask_scnprintf call though...

Actually, I'll do more tests and if that ends up being the only needed
change, I'll push your patch with that small change out to powerpc next
tonight.

Sorry about that, here's an incremental... let me know if you want the
whole thing re-posted.

Thanks.

diff --git a/arch/powerpc/kernel/cacheinfo.c b/arch/powerpc/kernel/cacheinfo.c
index f3e3ae3..b33f041 100644
--- a/arch/powerpc/kernel/cacheinfo.c
+++ b/arch/powerpc/kernel/cacheinfo.c

@@ -611,7 +611,7 @@ static ssize_t shared_cpu_map_show(struct kobject *k, struct kobj_attribute *att
 	len = PAGE_SIZE - 2;
 
 	if (len > 1) {
-		n = cpumask_scnprintf(buf, len, cache->shared_cpu_map);
+		n = cpumask_scnprintf(buf, len, &cache->shared_cpu_map);
 		buf[n++] = '\n';
 		buf[n] = '\0';
 	}