cgroup: fix subsys bitops
Li Zefan [Sat, 23 Feb 2008 23:24:11 +0000 (15:24 -0800)]
Cgroup uses unsigned long for subsys bitops, not unsigned long long.

Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Acked-by: Paul Menage <menage@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/cgroup.c

index 947fe3b..8412593 100644 (file)
@@ -319,7 +319,7 @@ static struct css_set *find_existing_css_set(
        /* Built the set of subsystem state objects that we want to
         * see in the new css_set */
        for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
-               if (root->subsys_bits & (1ull << i)) {
+               if (root->subsys_bits & (1UL << i)) {
                        /* Subsystem is in this hierarchy. So we want
                         * the subsystem state from the new
                         * cgroup */
@@ -689,7 +689,7 @@ static int rebind_subsystems(struct cgroupfs_root *root,
        added_bits = final_bits & ~root->actual_subsys_bits;
        /* Check that any added subsystems are currently free */
        for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) {
-               unsigned long long bit = 1ull << i;
+               unsigned long bit = 1UL << i;
                struct cgroup_subsys *ss = subsys[i];
                if (!(bit & added_bits))
                        continue;