From: Konstantin Khorenko <khore...@virtuozzo.com>
Recent versions of containerd (as a part of k3s-1.19.5)
started to apply strict rules when parsing the contents of
'devices.list' files located in the devices cgroup.
Namely, the access token is allowed to contain only those values [rwm],
that are described in
https://www.kernel.org/doc/Documentation/cgroup-v1/devices.txt
In vzkernel we do have an extra permission in device cgroup to allow
mount of a block device inside a Container ('M'), so this upsets
containerd.
Let's leave 'devices.{allow,deny}' files to be able to handle vz
specific "M" permission, but 'devices.list' to show only [rwm]
permissions suppressing possible "M" presence.
Let's introduce another file 'devices.extra_list' to show all
permissions, including possible "M".
$ echo "b 253:3182 rmM" > devices.allow
$ cat devices.list
...
b 253:3182 rm
$ cat devices.extra_list
...
b 253:3182 rmM
https://jira.sw.ru/browse/PSBM-123743
Signed-off-by: Konstantin Khorenko <khore...@virtuozzo.com>
Reviewed-by: Pavel Tikhomirov <ptikhomi...@virtuozzo.com>
Changes:
v2: fix DEVCG_DEFAULT_ALLOW branch in devcgroup_seq_show() to show 'M'
as well in devices::devices.extra_list.
(cherry-picked from vz8 commit 3d5a8118a9b6 ("ve/device_cgroup: Introduce
"devices.extra_list" cgroup file"))
Signed-off-by: Nikita Yushchenko <nikita.yushche...@virtuozzo.com>
---
include/linux/device_cgroup.h | 4 ++--
security/device_cgroup.c | 15 +++++++++++++--
2 files changed, 15 insertions(+), 4 deletions(-)
diff --git a/include/linux/device_cgroup.h b/include/linux/device_cgroup.h
index 365ffc2c516e..eb6a060176da 100644
--- a/include/linux/device_cgroup.h
+++ b/include/linux/device_cgroup.h
@@ -5,8 +5,8 @@
#define DEVCG_ACC_READ 2
#define DEVCG_ACC_WRITE 4
#define DEVCG_ACC_MOUNT 64
-#define DEVCG_ACC_MASK (DEVCG_ACC_MKNOD | DEVCG_ACC_READ | DEVCG_ACC_WRITE | \
- DEVCG_ACC_MOUNT)
+#define DEVCG_ACC_MASK (DEVCG_ACC_MKNOD | DEVCG_ACC_READ | DEVCG_ACC_WRITE)
+#define DEVCG_ACC_EXTRA_MASK (DEVCG_ACC_MASK | DEVCG_ACC_MOUNT)
#define DEVCG_DEV_BLOCK 1
#define DEVCG_DEV_CHAR 2
diff --git a/security/device_cgroup.c b/security/device_cgroup.c
index ee76745795b8..448c5bef0996 100644
--- a/security/device_cgroup.c
+++ b/security/device_cgroup.c
@@ -232,6 +232,7 @@ static void devcgroup_css_free(struct cgroup_subsys_state
*css)
#define DEVCG_ALLOW 1
#define DEVCG_DENY 2
#define DEVCG_LIST 3
+#define DEVCG_EXTRA_LIST 32
#define MAJMINLEN 13
#define ACCLEN 5
@@ -274,6 +275,11 @@ static int devcgroup_seq_show(struct seq_file *m, void *v)
struct dev_cgroup *devcgroup = css_to_devcgroup(seq_css(m));
struct dev_exception_item *ex;
char maj[MAJMINLEN], min[MAJMINLEN], acc[ACCLEN];
+ short type, mask;
+
+ type = (short)seq_cft(m)->private;
+ mask = (type == DEVCG_EXTRA_LIST) ?
+ DEVCG_ACC_EXTRA_MASK : DEVCG_ACC_MASK;
rcu_read_lock();
/*
@@ -283,14 +289,14 @@ static int devcgroup_seq_show(struct seq_file *m, void *v)
* This way, the file remains as a "whitelist of devices"
*/
if (devcgroup->behavior == DEVCG_DEFAULT_ALLOW) {
- set_access(acc, DEVCG_ACC_MASK);
+ set_access(acc, mask);
set_majmin(maj, ~0);
set_majmin(min, ~0);
seq_printf(m, "%c %s:%s %s\n", type_to_char(DEVCG_DEV_ALL),
maj, min, acc);
} else {
list_for_each_entry_rcu(ex, &devcgroup->exceptions, list) {
- set_access(acc, ex->access);
+ set_access(acc, ex->access & mask);
set_majmin(maj, ex->major);
set_majmin(min, ex->minor);
seq_printf(m, "%c %s:%s %s\n", type_to_char(ex->type),
@@ -802,6 +808,11 @@ static struct cftype dev_cgroup_files[] = {
.seq_show = devcgroup_seq_show,
.private = DEVCG_LIST,
},
+ {
+ .name = "extra_list",
+ .seq_show = devcgroup_seq_show,
+ .private = DEVCG_EXTRA_LIST,
+ },
{ } /* terminate */
};
--
2.30.2
_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel
