cgroup, doc: make cgroup_disable doc more accurate
authorQiang Huang <h.huangqiang@huawei.com>
Wed, 6 Nov 2013 21:18:09 +0000 (13:18 -0800)
committerJiri Kosina <jkosina@suse.cz>
Mon, 2 Dec 2013 13:50:46 +0000 (14:50 +0100)
In doc, it said that 'Currently supported controllers - "memory"',
but actually we can use cgroup_disable=cpu,cpuset and all other
controllers, so this is confusing for cgroup users without much
cgroup knowledge. We need to make it clear.

[some comments copied from Paul Menage's original patch 8bab8dded]

Signed-off-by: Qiang Huang <h.huangqiang@huawei.com>
Acked-by: Rob Landley <rob@landley.net>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Documentation/kernel-parameters.txt

index 479eeaf440248b8d56dca90775be614b7cbc9e9e..5589f4dbc12f1ae3108b9cd517f2aaadbef3348a 100644 (file)
@@ -511,7 +511,14 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 
        cgroup_disable= [KNL] Disable a particular controller
                        Format: {name of the controller(s) to disable}
-                               {Currently supported controllers - "memory"}
+                       The effects of cgroup_disable=foo are:
+                       - foo isn't auto-mounted if you mount all cgroups in
+                         a single hierarchy
+                       - foo isn't visible as an individually mountable
+                         subsystem
+                       {Currently only "memory" controller deal with this and
+                       cut the overhead, others just disable the usage. So
+                       only cgroup_disable=memory is actually worthy}
 
        checkreqprot    [SELINUX] Set initial checkreqprot flag value.
                        Format: { "0" | "1" }