It turned out, that the user namespace is released during the do_exit() in
exit_task_namespaces(), but the struct user_struct is released only during the
put_task_struct(), i.e. MUCH later.
On debug kernels with poisoned slabs this will cause the oops in
uid_hash_remove() because the head of the chain, which resides inside the
struct user_namespace, will be already freed and poisoned.
Since the uid hash itself is required only when someone can search it, i.e.
when the namespace is alive, we can safely unhash all the user_struct-s from
it during the namespace exiting. The subsequent free_uid() will complete the
user_struct destruction.
For example simple program
#include <sched.h>
char stack[2 * 1024 * 1024];
int f(void *foo)
{
return 0;
}
int main(void)
{
clone(f, stack + 1 * 1024 * 1024, 0x10000000, 0);
return 0;
}
run on kernel with CONFIG_USER_NS turned on will oops the
kernel immediately.
This was spotted during OpenVZ kernel testing.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: Alexey Dobriyan <[email protected]>
Acked-by: "Serge E. Hallyn" <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Surprisingly, but (spotted by Alexey Dobriyan) the uid hash still uses
list_heads, thus occupying twice as much place as it could. Convert it to
hlist_heads.
Signed-off-by: Pavel Emelyanov <[email protected]>
Signed-off-by: Alexey Dobriyan <[email protected]>
Acked-by: Serge Hallyn <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
dup_mnt_ns() and clone_uts_ns() return NULL on failure. This is wrong,
create_new_namespaces() uses ERR_PTR() to catch an error. This means that the
subsequent create_new_namespaces() will hit BUG_ON() in copy_mnt_ns() or
copy_utsname().
Modify create_new_namespaces() to also use the errors returned by the
copy_*_ns routines and not to systematically return ENOMEM.
[[email protected]: better changelog]
Signed-off-by: Cedric Le Goater <[email protected]>
Cc: Serge E. Hallyn <[email protected]>
Cc: Badari Pulavarty <[email protected]>
Cc: Pavel Emelianov <[email protected]>
Cc: Herbert Poetzl <[email protected]>
Cc: Eric W. Biederman <[email protected]>
Cc: Oleg Nesterov <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
This patch enables the unshare of user namespaces.
It adds a new clone flag CLONE_NEWUSER and implements copy_user_ns() which
resets the current user_struct and adds a new root user (uid == 0)
For now, unsharing the user namespace allows a process to reset its
user_struct accounting and uid 0 in the new user namespace should be contained
using appropriate means, for instance selinux
The plan, when the full support is complete (all uid checks covered), is to
keep the original user's rights in the original namespace, and let a process
become uid 0 in the new namespace, with full capabilities to the new
namespace.
Signed-off-by: Serge E. Hallyn <[email protected]>
Signed-off-by: Cedric Le Goater <[email protected]>
Acked-by: Pavel Emelianov <[email protected]>
Cc: Herbert Poetzl <[email protected]>
Cc: Kirill Korotaev <[email protected]>
Cc: Eric W. Biederman <[email protected]>
Cc: Chris Wright <[email protected]>
Cc: Stephen Smalley <[email protected]>
Cc: James Morris <[email protected]>
Cc: Andrew Morgan <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Basically, it will allow a process to unshare its user_struct table,
resetting at the same time its own user_struct and all the associated
accounting.
A new root user (uid == 0) is added to the user namespace upon creation.
Such root users have full privileges and it seems that theses privileges
should be controlled through some means (process capabilities ?)
The unshare is not included in this patch.
Changes since [try #4]:
- Updated get_user_ns and put_user_ns to accept NULL, and
get_user_ns to return the namespace.
Changes since [try #3]:
- moved struct user_namespace to files user_namespace.{c,h}
Changes since [try #2]:
- removed struct user_namespace* argument from find_user()
Changes since [try #1]:
- removed struct user_namespace* argument from find_user()
- added a root_user per user namespace
Signed-off-by: Cedric Le Goater <[email protected]>
Signed-off-by: Serge E. Hallyn <[email protected]>
Acked-by: Pavel Emelianov <[email protected]>
Cc: Herbert Poetzl <[email protected]>
Cc: Kirill Korotaev <[email protected]>
Cc: Eric W. Biederman <[email protected]>
Cc: Chris Wright <[email protected]>
Cc: Stephen Smalley <[email protected]>
Cc: James Morris <[email protected]>
Cc: Andrew Morgan <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Pavel Emelyanov