summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorLuke Shumaker <lukeshu@lukeshu.com>2017-06-01 13:59:20 -0400
committerLuke Shumaker <lukeshu@lukeshu.com>2017-06-16 17:06:55 -0400
commit2fbe69d3ffd1be92eda13ea782337349d63af14b (patch)
treed4eff39a8337341d49c1f185ff97b3c268a52fa8
parentdc87440fa9ce11de578895cdf5326bb288cd78dc (diff)
nspawn: mount_sysfs(): Reword the comment about /sys/fs/cgroup
The comment explains the obvious, but doesn't even mention the tricky part. Of course we need do set things up before we remount read-only! That's the general theme of the function! What was totally non-obvious is why we only need to create it if cg_ns_supported(), as the directory needs to exist no matter what. From reading the code, I was convinced that it was broken on pre-cgns kernels (pre-4.6, unless a distro backported it). So explain that skippint creating if !cg_ns_supported() is an optimization.
-rw-r--r--src/nspawn/nspawn-mount.c15
1 files changed, 11 insertions, 4 deletions
diff --git a/src/nspawn/nspawn-mount.c b/src/nspawn/nspawn-mount.c
index 2dabe2ae5b..1814ea6ca5 100644
--- a/src/nspawn/nspawn-mount.c
+++ b/src/nspawn/nspawn-mount.c
@@ -288,10 +288,17 @@ int mount_sysfs(const char *dest) {
x = prefix_roota(top, "/fs/kdbus");
(void) mkdir_p(x, 0755);
- /* Create mountpoint for cgroups. Otherwise we are not allowed since we
- * remount /sys read-only.
- */
- if (cg_ns_supported()) {
+ /* We need to ensure that /sys/fs/cgroup exists before we remount /sys read-only.
+ *
+ * If !use_cgns, then this was already done by the outer child; so we only need to do it here it if use_cgns.
+ * This function doesn't know whether use_cgns, but !cg_ns_supported()⇒!use_cgns, so we can "optimize" the case
+ * where we _know_ !use_cgns, and deal with a no-op mkdir_p() in the false-positive where cgns_supported() but
+ * !use_cgns.
+ *
+ * But is it really much of an optimization? We're potentially spending an access(2) (cg_ns_supported() could
+ * be cached from a previous call) to potentially save an lstat(2) and mkdir(2); and all of them are on virtual
+ * fileystems, so they should all be pretty cheap. */
+ if (cg_ns_supported()) { /* if (use_cgns) { */
x = prefix_roota(top, "/fs/cgroup");
(void) mkdir_p(x, 0755);
}