capabilities: Introduce CAP_CHECKPOINT_RESTORE
Subsystems (8)
v4 cover capabilities: Introduce CAP_CHECKPOINT_RESTORE
This is v4 of the 'Introduce CAP_CHECKPOINT_RESTORE' patchset. There is only one change from v3 to address Jann's comment on patch 3/3
(That is not necessarily true in the presence of LSMs like SELinux:
You'd have to be able to FILE__EXECUTE_NO_TRANS the target executable according to the system's security policy.)
Nicolas updated the last patch (3/3). The first two patches are unchanged from v3.
Adrian Reber (2):
capabilities: Introduce CAP_CHECKPOINT_RESTORE
selftests: add clone3() CAP_CHECKPOINT_RESTORE test
Nicolas Viennot (1):
prctl: Allow ptrace capable processes to change /proc/self/exe
fs/proc/base.c | 8 +- include/linux/capability.h | 6 + include/linux/lsm_hook_defs.h | 1 + include/linux/security.h | 6 + include/uapi/linux/capability.h | 9 +- kernel/pid.c | 2 +- kernel/pid_namespace.c | 2 +- kernel/sys.c | 12 +- security/commoncap.c | 26 +++ security/security.c | 5 + security/selinux/hooks.c | 14 ++ security/selinux/include/classmap.h | 5 +- tools/testing/selftests/clone3/Makefile | 4 +- .../clone3/clone3_cap_checkpoint_restore.c | 203 ++++++++++++++++++ 14 files changed, 285 insertions(+), 18 deletions(-) create mode 100644 tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c
base-commit: f2b92b14533e646e434523abdbafddb727c23898
Signature · 1 line
2.26.2
PATCH v4 1/3 capabilities: Introduce CAP_CHECKPOINT_RESTORE
This patch introduces CAP_CHECKPOINT_RESTORE, a new capability facilitating checkpoint/restore for non-root users.
Over the last years, The CRIU (Checkpoint/Restore In Userspace) team has been asked numerous times if it is possible to checkpoint/restore a process as non-root. The answer usually was: 'almost'.
The main blocker to restore a process as non-root was to control the PID of the restored process. This feature available via the clone3 system call, or via /proc/sys/kernel/ns_last_pid is unfortunately guarded by CAP_SYS_ADMIN.
In the past two years, requests for non-root checkpoint/restore have increased due to the following use cases:
* Checkpoint/Restore in an HPC environment in combination with a resource manager distributing jobs where users are always running as non-root. There is a desire to provide a way to checkpoint and restore long running jobs.
* Container migration as non-root
* We have been in contact with JVM developers who are integrating CRIU into a Java VM to decrease the startup time. These checkpoint/restore applications are not meant to be running with CAP_SYS_ADMIN.
We have seen the following workarounds:
* Use a setuid wrapper around CRIU:
See https://github.com/FredHutch/slurm-examples/blob/master/checkpointer/lib/checkpointer/checkpointer-suid.c
* Use a setuid helper that writes to ns_last_pid.
Unfortunately, this helper delegation technique is impossible to use with
clone3, and is thus prone to races.
See https://github.com/twosigma/set_ns_last_pid
* Cycle through PIDs with fork() until the desired PID is reached:
This has been demonstrated to work with cycling rates of 100,000 PIDs/s
See https://github.com/twosigma/set_ns_last_pid
* Patch out the CAP_SYS_ADMIN check from the kernel
* Run the desired application in a new user and PID namespace to provide
a local CAP_SYS_ADMIN for controlling PIDs. This technique has limited use in
typical container environments (e.g., Kubernetes) as /proc is
typically protected with read-only layers (e.g., /proc/sys) for hardening
purposes. Read-only layers prevent additional /proc mounts (due to proc's SB_I_USERNS_VISIBLE property), making the use of new PID namespaces limited as
certain applications need access to /proc matching their PID namespace.
The introduced capability allows to:
* Control PIDs when the current user is CAP_CHECKPOINT_RESTORE capable for the corresponding PID namespace via ns_last_pid/clone3.
* Open files in /proc/pid/map_files when the current user is CAP_CHECKPOINT_RESTORE capable in the root namespace, useful for recovering files that are unreachable via the file system such as deleted files, or memfd files.
See corresponding selftest for an example with clone3().
--- fs/proc/base.c | 8 ++++---- include/linux/capability.h | 6 ++++++ include/uapi/linux/capability.h | 9 ++++++++- kernel/pid.c | 2 +- kernel/pid_namespace.c | 2 +- security/selinux/include/classmap.h | 5 +++-- 6 files changed, 23 insertions(+), 9 deletions(-)
fs/proc/base.c +4 -4
@@ -2189,16 +2189,16 @@ struct map_files_info { };/* -* Only allow CAP_SYS_ADMIN to follow the links, due to concerns about how the -* symlinks may be used to bypass permissions on ancestor directories in the -* path to the file in question. +* Only allow CAP_SYS_ADMIN and CAP_CHECKPOINT_RESTORE to follow the links, due +* to concerns about how the symlinks may be used to bypass permissions on +* ancestor directories in the path to the file in question. */ static const char *proc_map_files_get_link (struct dentry *dentry,struct inode *inode,struct delayed_call *done) { -if (!capable (CAP_SYS_ADMIN )) +if (!capable (CAP_SYS_ADMIN ) && !capable (CAP_CHECKPOINT_RESTORE ))return ERR_PTR (-EPERM );return proc_pid_get_link (dentry, inode, done);
include/linux/capability.h +6 -0
@@ -261,6 +261,12 @@ static inline bool bpf_capable(void)return capable (CAP_BPF ) ||capable (CAP_SYS_ADMIN ); } +static inline bool checkpoint_restore_ns_capable (struct user_namespace *ns) +{ +return ns_capable (ns,CAP_CHECKPOINT_RESTORE ) || +ns_capable (ns,CAP_SYS_ADMIN ); +} +/* audit system wants to get cap info from files as well */ extern int get_vfs_caps_from_disk (const struct dentry *dentry,struct cpu_vfs_cap_data *cpu_caps);
include/uapi/linux/capability.h +8 -1
@@ -408,7 +408,14 @@ struct vfs_ns_cap_data {*/ #define CAP_BPF 39 -#define CAP_LAST_CAP CAP_BPF + +/* Allow checkpoint/restore related operations */ +/* Allow PID selection during clone3() */ +/* Allow writing to ns_last_pid */ + +#define CAP_CHECKPOINT_RESTORE 40 + +#define CAP_LAST_CAP CAP_CHECKPOINT_RESTORE#define cap_valid (x) ((x) >= 0 && (x) <= CAP_LAST_CAP)
kernel/pid.c +1 -1
@@ -198,7 +198,7 @@ struct pid *alloc_pid(struct pid_namespace *ns, pid_t *set_tid,if (tid !=1 && !tmp->child_reaper) goto out_free; retval = -EPERM ; -if (!ns_capable (tmp->user_ns,CAP_SYS_ADMIN )) +if (!checkpoint_restore_ns_capable (tmp->user_ns)) goto out_free; set_tid_size--; }
kernel/pid_namespace.c +1 -1
@@ -269,7 +269,7 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write,struct ctl_table tmp = *table;int ret, next; -if (write && !ns_capable (pid_ns->user_ns,CAP_SYS_ADMIN )) +if (write && !checkpoint_restore_ns_capable (pid_ns->user_ns))return -EPERM ;/*
security/selinux/include/classmap.h +3 -2
@@ -27,9 +27,10 @@ "audit_control", "setfcap"#define COMMON_CAP2_PERMS "mac_override", "mac_admin", "syslog", \ - "wake_alarm", "block_suspend", "audit_read", "perfmon", "bpf" + "wake_alarm", "block_suspend", "audit_read", "perfmon", "bpf", \ + "checkpoint_restore" -#if CAP_LAST_CAP >CAP_BPF +#if CAP_LAST_CAP >CAP_CHECKPOINT_RESTORE #error New capability defined, please update COMMON_CAP2_PERMS.#endif
Signature · 1 line
2.26.2
PATCH v4 2/3 selftests: add clone3() CAP_CHECKPOINT_RESTORE test
In reply to Adrian ReberThis adds a test that changes its UID, uses capabilities to get CAP_CHECKPOINT_RESTORE and uses clone3() with set_tid to create a process with a given PID as non-root.
--- tools/testing/selftests/clone3/Makefile | 4 +- .../clone3/clone3_cap_checkpoint_restore.c | 203 ++++++++++++++++++ 2 files changed, 206 insertions(+), 1 deletion(-) create mode 100644 tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c
tools/testing/selftests/clone3/Makefile +3 -1
@@ -1,6 +1,8 @@ # SPDX-License-Identifier: GPL-2.0 CFLAGS += -g -I../../../../usr/include/ +LDLIBS += -lcap -TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid +TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid \ + clone3_cap_checkpoint_restore include ../lib.mk
tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c new file +203 -0
203 lines not on this page · raw message
Signature · 1 line
2.26.2
PATCH v4 3/3 prctl: Allow ptrace capable processes to change /proc/self/exe
In reply to Adrian ReberFrom: Nicolas Viennot <[email protected]>
Previously, the current process could only change the /proc/self/exe link with local CAP_SYS_ADMIN.
This commit relaxes this restriction by permitting such change with CAP_CHECKPOINT_RESTORE, and the ability to use ptrace.
With access to ptrace facilities, a process can do the following: fork a child, execve() the target executable, and have the child use ptrace() to replace the memory content of the current process. This technique makes it possible to masquerade an arbitrary program as any executable, even setuid ones.
--- include/linux/lsm_hook_defs.h | 1 + include/linux/security.h | 6 ++++++ kernel/sys.c | 12 ++++-------- security/commoncap.c | 26 ++++++++++++++++++++++++++ security/security.c | 5 +++++ security/selinux/hooks.c | 14 ++++++++++++++ 6 files changed, 56 insertions(+), 8 deletions(-)
include/linux/lsm_hook_defs.h +1 -0
@@ -211,6 +211,7 @@ LSM_HOOK(int, 0, task_kill, struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred) LSM_HOOK(int, -ENOSYS, task_prctl, int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5) +LSM_HOOK(int, 0, prctl_set_mm_exe_file, struct file *exe_file) LSM_HOOK(void, LSM_RET_VOID, task_to_inode, struct task_struct *p, struct inode *inode) LSM_HOOK(int, 0, ipc_permission, struct kern_ipc_perm *ipcp, short flag)
include/linux/security.h +6 -0
@@ -412,6 +412,7 @@ int security_task_kill(struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred);int security_task_prctl (int option,unsigned long arg2,unsigned long arg3,unsigned long arg4,unsigned long arg5); +int security_prctl_set_mm_exe_file (struct file *exe_file);void security_task_to_inode (struct task_struct *p,struct inode *inode);int security_ipc_permission (struct kern_ipc_perm *ipcp,short flag);void security_ipc_getsecid (struct kern_ipc_perm *ipcp,u32 *secid);
@@ -1124,6 +1125,11 @@ static inline int security_task_prctl(int option, unsigned long arg2,return cap_task_prctl (option, arg2, arg3, arg4, arg5); } +static inline int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return cap_prctl_set_mm_exe_file (exe_file); +} +static inline void security_task_to_inode (struct task_struct *p,struct inode *inode) { }
kernel/sys.c +4 -8
@@ -1851,6 +1851,10 @@ static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd)if (err) goto exit; + err =security_prctl_set_mm_exe_file (exe.file); +if (err) + goto exit; +/* * Forbid mm->exe_file change if old file still mapped. */
@@ -2006,14 +2010,6 @@ static int prctl_set_mm_map(int opt, const void __user *addr, unsigned long data }if (prctl_map.exe_fd != (u32)-1 ) { -/* -* Make sure the caller has the rights to -* change /proc/pid/exe link: only local sys admin should -* be allowed to. -*/ -if (!ns_capable (current_user_ns (),CAP_SYS_ADMIN )) -return -EINVAL ; - error =prctl_set_mm_exe_file (mm, prctl_map.exe_fd);if (error)return error;
security/commoncap.c +26 -0
@@ -1291,6 +1291,31 @@ int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, } } +/** +* cap_prctl_set_mm_exe_file - Determine whether /proc/self/exe can be changed +* by the current process. +* @exe_file: The new exe file +* Returns 0 if permission is granted, -ve if denied. +* +* The current process is permitted to change its /proc/self/exe link via two policies: +* 1) The current user can do checkpoint/restore. At the time of this writing, +* this means CAP_SYS_ADMIN or CAP_CHECKPOINT_RESTORE capable. +* 2) The current user can use ptrace. +* +* With access to ptrace facilities, a process can do the following: +* fork a child, execve() the target executable, and have the child use +* ptrace() to replace the memory content of the current process. +* This technique makes it possible to masquerade an arbitrary program as the +* target executable, even if it is setuid. +*/ +int cap_prctl_set_mm_exe_file (struct file *exe_file) +{ +if (checkpoint_restore_ns_capable (current_user_ns ())) +return 0 ; + +return security_ptrace_access_check (current,PTRACE_MODE_ATTACH_REALCREDS ); +} +/** * cap_vm_enough_memory - Determine whether a new virtual mapping is permitted * @mm: The VM space in which the new mapping is to be made
@@ -1356,6 +1381,7 @@ static struct security_hook_list capability_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (mmap_file, cap_mmap_file),LSM_HOOK_INIT (task_fix_setuid, cap_task_fix_setuid),LSM_HOOK_INIT (task_prctl, cap_task_prctl), +LSM_HOOK_INIT (prctl_set_mm_exe_file, cap_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_setscheduler, cap_task_setscheduler),LSM_HOOK_INIT (task_setioprio, cap_task_setioprio), LSM_HOOK_INIT(task_setnice, cap_task_setnice),
security/security.c +5 -0
@@ -1790,6 +1790,11 @@ int security_task_prctl(int option, unsigned long arg2, unsigned long arg3,return rc; } +int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return call_int_hook (prctl_set_mm_exe_file,0 , exe_file); +} +void security_task_to_inode (struct task_struct *p,struct inode *inode) {call_void_hook (task_to_inode, p, inode);
security/selinux/hooks.c +14 -0
@@ -4156,6 +4156,19 @@ static int selinux_task_kill(struct task_struct *p, struct kernel_siginfo *info, secid, task_sid(p), SECCLASS_PROCESS, perm, NULL); } +static int selinux_prctl_set_mm_exe_file (struct file *exe_file) +{ +u32 sid =current_sid (); + +struct common_audit_data ad = { + .type =LSM_AUDIT_DATA_FILE , + .u.file = exe_file, + }; + +return avc_has_perm (&selinux_state, sid, sid, +SECCLASS_FILE ,FILE__EXECUTE_NO_TRANS , &ad); +} +static void selinux_task_to_inode (struct task_struct *p,struct inode *inode) {
@@ -7057,6 +7070,7 @@ static struct security_hook_list selinux_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (task_getscheduler, selinux_task_getscheduler),LSM_HOOK_INIT (task_movememory, selinux_task_movememory),LSM_HOOK_INIT (task_kill, selinux_task_kill), +LSM_HOOK_INIT (prctl_set_mm_exe_file, selinux_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_to_inode, selinux_task_to_inode), LSM_HOOK_INIT(ipc_permission, selinux_ipc_permission),
Signature · 1 line
2.26.2
kernel test robot
build botRe: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
Thank you for the patch! Yet something to improve:
Hi Adrian,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on pcmoore-selinux/next]
[also build test ERROR on kselftest/next linus/master v5.8-rc3 next-20200701] [cannot apply to security/next-testing]
[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Adrian-Reber/capabilities-Introduce-CAP_CHECKPOINT_RESTORE/20200701-145119 base: https://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux.git next config: i386-tinyconfig (attached as .config) compiler: gcc-9 (Debian 9.3.0-14) 9.3.0 reproduce (this is a W=1 build): # save the attached .config to linux build tree make W=1 ARCH=i386
If you fix the issue, kindly add following tag as appropriate
All errors (new ones prefixed by >>):
make[2]: *** No rule to make target 'fs//fs_parser.o', needed by 'fs//built-in.a'. In file included from include/linux/fs_context.h:14,
from include/linux/fs_parser.h:11,
from include/linux/ramfs.h:5,
from fs//ramfs/file-mmu.c:29:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file
cc1: some warnings being treated as errors
make[3]: *** [scripts/Makefile.build:281: fs//ramfs/file-mmu.o] Error 1
In file included from fs//attr.c:17:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file
In file included from include/linux/fs_context.h:14,
from include/linux/fs_parser.h:11,
from include/linux/ramfs.h:5,
from fs//ramfs/inode.c:33:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file cc1: some warnings being treated as errors
make[2]: *** [scripts/Makefile.build:281: fs//xattr.o] Error 1
cc1: some warnings being treated as errors
make[2]: *** [scripts/Makefile.build:497: fs//ramfs] Error 2 make[2]: Target '__build' not remade because of errors.
make[1]: *** [Makefile:1764: fs/] Error 2
make[1]: Target '__all' not remade because of errors.
make[3]: *** [scripts/Makefile.build:281: fs//ramfs/inode.o] Error 1 make[3]: Target '__build' not remade because of errors.
--
make[2]: *** No rule to make target 'fs//fs_parser.o', needed by 'fs//built-in.a'. In file included from include/linux/fs_context.h:14,
from include/linux/fs_parser.h:11,
from include/linux/ramfs.h:5,
from fs//ramfs/file-mmu.c:29:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file
cc1: some warnings being treated as errors
make[3]: *** [scripts/Makefile.build:281: fs//ramfs/file-mmu.o] Error 1
In file included from fs//attr.c:17:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file
In file included from include/linux/fs_context.h:14,
from include/linux/fs_parser.h:11,
from include/linux/ramfs.h:5,
from fs//ramfs/inode.c:33:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file cc1: some warnings being treated as errors
make[2]: *** [scripts/Makefile.build:281: fs//xattr.o] Error 1
cc1: some warnings being treated as errors
make[2]: *** [scripts/Makefile.build:497: fs//ramfs] Error 2 make[2]: Target '__build' not remade because of errors. make[3]: *** [scripts/Makefile.build:281: fs//ramfs/inode.o] Error 1 make[3]: Target '__build' not remade because of errors.
--
In file included from include/linux/fs_context.h:14,
from include/linux/fs_parser.h:11,
from include/linux/ramfs.h:5,
from fs//ramfs/file-mmu.c:29:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file
cc1: some warnings being treated as errors
make[3]: *** [scripts/Makefile.build:281: fs//ramfs/file-mmu.o] Error 1
In file included from include/linux/fs_context.h:14,
from include/linux/fs_parser.h:11,
from include/linux/ramfs.h:5,
from fs//ramfs/inode.c:33:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_file cc1: some warnings being treated as errors
make[3]: *** [scripts/Makefile.build:281: fs//ramfs/inode.o] Error 1 make[3]: Target '__build' not remade because of errors.
--
In file included from fs//attr.c:17:
include/linux/security.h: In function 'security_prctl_set_mm_exe_file':
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file'; did you mean 'security_prctl_set_mm_exe_file'? [-Werror=implicit-function-declaration]
1130 | return cap_prctl_set_mm_exe_file(exe_file);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| security_prctl_set_mm_exe_filecc1: some warnings being treated as errors
vim +1130 include/linux/security.h
1127
1128 static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
1129 {
> 1130 return cap_prctl_set_mm_exe_file(exe_file);
1131 }
1132
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
Quoted 49 lines · Adrian Reber
This patch introduces
CAP_CHECKPOINT_RESTORE, a new capability facilitating checkpoint/restore for non-root users.Over the last years, The CRIU (Checkpoint/Restore In Userspace) team has been asked numerous times if it is possible to checkpoint/restore a process as non-root. The answer usually was: 'almost'.
The main blocker to restore a process as non-root was to control the PID of the restored process. This feature available via the clone3 system call, or via
/proc/sys/kernel/ns_last_pidis unfortunately guarded byCAP_SYS_ADMIN.In the past two years, requests for non-root checkpoint/restore have increased due to the following use cases:
* Checkpoint/Restore in an HPC environment in combination with a resource manager distributing jobs where users are always running as non-root. There is a desire to provide a way to checkpoint and restore long running jobs.
* Container migration as non-root
* We have been in contact with JVM developers who are integrating CRIU into a Java VM to decrease the startup time. These checkpoint/restore applications are not meant to be running withCAP_SYS_ADMIN.We have seen the following workarounds:
* Use a setuid wrapper around CRIU:
See https://github.com/FredHutch/slurm-examples/blob/master/checkpointer/lib/checkpointer/checkpointer-suid.c
* Use a setuid helper that writes tons_last_pid.
Unfortunately, this helper delegation technique is impossible to use with
clone3, and is thus prone to races.
See https://github.com/twosigma/set_ns_last_pid
* Cycle through PIDs withfork()until the desired PID is reached:
This has been demonstrated to work with cycling rates of 100,000 PIDs/s
See https://github.com/twosigma/set_ns_last_pid
* Patch out theCAP_SYS_ADMINcheck from the kernel
* Run the desired application in a new user and PID namespace to provide
a localCAP_SYS_ADMINfor controlling PIDs. This technique has limited use in
typical container environments (e.g., Kubernetes) as /proc is
typically protected with read-only layers (e.g.,/proc/sys) for hardening
purposes. Read-only layers prevent additional /proc mounts (due to proc'sSB_I_USERNS_VISIBLEproperty), making the use of new PID namespaces limited as
certain applications need access to /proc matching their PID namespace.The introduced capability allows to:
* Control PIDs when the current user isCAP_CHECKPOINT_RESTOREcapable for the corresponding PID namespace via ns_last_pid/clone3.
* Open files in/proc/pid/map_fileswhen the current user isCAP_CHECKPOINT_RESTOREcapable in the root namespace, useful for recovering files that are unreachable via the file system such as deleted files, or memfd files.
Quoted 5 lines
See corresponding selftest for an example with
clone3().Signed-off-by Adrian Reber <[email protected]>Signed-off-by Nicolas Viennot <[email protected]>---
I think that now looks reasonable. A few comments.
Before we proceed, please split the addition of checkpoint_restore_ns_capable() out into a separate patch.
In fact, I think the cleanest way of doing this would be:
- 0/n capability: add CAP_CHECKPOINT_RESTORE
- 1/n pid: use checkpoint_restore_ns_capable() for set_tid
- 2/n pid_namespace: use checkpoint_restore_ns_capable() for ns_last_pid
- 3/n: proc: require checkpoint_restore_ns_capable() in init userns for map_files
(commit subjects up to you of course) and a nice commit message for each time we relax a permissions on something so we have a clear separate track record for each change in case we need to revert something. Then the rest of the patches in this series. Testing patches probably last.
Quoted 21 lines
fs/proc/base.c | 8 ++++---- include/linux/capability.h | 6 ++++++ include/uapi/linux/capability.h | 9 ++++++++- kernel/pid.c | 2 +- kernel/pid_namespace.c | 2 +- security/selinux/include/classmap.h | 5 +++-- 6 files changed, 23 insertions(+), 9 deletions(-)fs/proc/base.c +2
-3@@ -2189,16 +2189,16 @@ struct map_files_info { };/* -* Only allow CAP_SYS_ADMIN to follow the links, due to concerns about how the -* symlinks may be used to bypass permissions on ancestor directories in the -* path to the file in question. +* Only allow CAP_SYS_ADMIN and CAP_CHECKPOINT_RESTORE to follow the links, due +* to concerns about how the symlinks may be used to bypass permissions on
fs/proc/base.c struct map_files_info { · line 2202
Quoted 10 lines
fs/proc/base.c +2
-1+* ancestor directories in the path to the file in question. */ static const char *proc_map_files_get_link (struct dentry *dentry,struct inode *inode,struct delayed_call *done) { -if (!capable (CAP_SYS_ADMIN )) +if (!capable (CAP_SYS_ADMIN ) && !capable (CAP_CHECKPOINT_RESTORE ))return ERR_PTR (-EPERM );
I think it's clearer if you just use:checkpoint_restore_ns_capable(&init_user_ns)
Quoted 87 lines
fs/proc/base.c +1
-0+static inline bool checkpoint_restore_ns_capable(struct user_namespace *ns) return proc_pid_get_link(dentry, inode, done);include/linux/capability.h +6
-0@@ -261,6 +261,12 @@ static inline bool bpf_capable(void)return capable (CAP_BPF ) ||capable (CAP_SYS_ADMIN ); } +static inline bool checkpoint_restore_ns_capable (struct user_namespace *ns) +{ +return ns_capable (ns,CAP_CHECKPOINT_RESTORE ) || +ns_capable (ns,CAP_SYS_ADMIN ); +} +/* audit system wants to get cap info from files as well */ extern int get_vfs_caps_from_disk (const struct dentry *dentry,struct cpu_vfs_cap_data *cpu_caps);include/uapi/linux/capability.h +8
-1@@ -408,7 +408,14 @@ struct vfs_ns_cap_data {*/ #define CAP_BPF 39 -#define CAP_LAST_CAP CAP_BPF + +/* Allow checkpoint/restore related operations */ +/* Allow PID selection during clone3() */ +/* Allow writing to ns_last_pid */ + +#define CAP_CHECKPOINT_RESTORE 40 + +#define CAP_LAST_CAP CAP_CHECKPOINT_RESTORE#define cap_valid (x) ((x) >= 0 && (x) <= CAP_LAST_CAP)kernel/pid.c +1
-1@@ -198,7 +198,7 @@ struct pid *alloc_pid(struct pid_namespace *ns, pid_t *set_tid,if (tid !=1 && !tmp->child_reaper) goto out_free; retval = -EPERM ; -if (!ns_capable (tmp->user_ns,CAP_SYS_ADMIN )) +if (!checkpoint_restore_ns_capable (tmp->user_ns)) goto out_free; set_tid_size--; }kernel/pid_namespace.c +1
-1@@ -269,7 +269,7 @@ static int pid_ns_ctl_handler(struct ctl_table *table, int write,struct ctl_table tmp = *table;int ret, next; -if (write && !ns_capable (pid_ns->user_ns,CAP_SYS_ADMIN )) +if (write && !checkpoint_restore_ns_capable (pid_ns->user_ns))return -EPERM ;/* security/selinux/include/classmap.h +3
-2@@ -27,9 +27,10 @@ "audit_control", "setfcap"#define COMMON_CAP2_PERMS "mac_override", "mac_admin", "syslog", \ - "wake_alarm", "block_suspend", "audit_read", "perfmon", "bpf" + "wake_alarm", "block_suspend", "audit_read", "perfmon", "bpf", \ + "checkpoint_restore" -#if CAP_LAST_CAP >CAP_BPF +#if CAP_LAST_CAP >CAP_CHECKPOINT_RESTORE #error New capability defined, please update COMMON_CAP2_PERMS.#endif Signature · 1 line
2.26.2
Re: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
In reply to Adrian ReberQuoted 112 lines · Adrian Reber
From: Nicolas Viennot <[email protected]>
Previously, the current process could only change the
/proc/self/exelink with localCAP_SYS_ADMIN.
This commit relaxes this restriction by permitting such change withCAP_CHECKPOINT_RESTORE, and the ability to use ptrace.With access to ptrace facilities, a process can do the following: fork a child,
execve()the target executable, and have the child useptrace()to replace the memory content of the current process. This technique makes it possible to masquerade an arbitrary program as any executable, even setuid ones.Signed-off-by Nicolas Viennot <[email protected]>Signed-off-by Adrian Reber <[email protected]>--- include/linux/lsm_hook_defs.h | 1 + include/linux/security.h | 6 ++++++ kernel/sys.c | 12 ++++-------- security/commoncap.c | 26 ++++++++++++++++++++++++++ security/security.c | 5 +++++ security/selinux/hooks.c | 14 ++++++++++++++ 6 files changed, 56 insertions(+), 8 deletions(-)include/linux/lsm_hook_defs.h +1
-0@@ -211,6 +211,7 @@ LSM_HOOK(int, 0, task_kill, struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred) LSM_HOOK(int, -ENOSYS, task_prctl, int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5) +LSM_HOOK(int, 0, prctl_set_mm_exe_file, struct file *exe_file) LSM_HOOK(void, LSM_RET_VOID, task_to_inode, struct task_struct *p, struct inode *inode) LSM_HOOK(int, 0, ipc_permission, struct kern_ipc_perm *ipcp, short flag)include/linux/security.h +6
-0@@ -412,6 +412,7 @@ int security_task_kill(struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred);int security_task_prctl (int option,unsigned long arg2,unsigned long arg3,unsigned long arg4,unsigned long arg5); +int security_prctl_set_mm_exe_file (struct file *exe_file);void security_task_to_inode (struct task_struct *p,struct inode *inode);int security_ipc_permission (struct kern_ipc_perm *ipcp,short flag);void security_ipc_getsecid (struct kern_ipc_perm *ipcp,u32 *secid);@@ -1124,6 +1125,11 @@ static inline int security_task_prctl(int option, unsigned long arg2,return cap_task_prctl (option, arg2, arg3, arg4, arg5); } +static inline int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return cap_prctl_set_mm_exe_file (exe_file); +} +static inline void security_task_to_inode (struct task_struct *p,struct inode *inode) { }kernel/sys.c +4
-8@@ -1851,6 +1851,10 @@ static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd)if (err) goto exit; + err =security_prctl_set_mm_exe_file (exe.file); +if (err) + goto exit; +/* * Forbid mm->exe_file change if old file still mapped. */ @@ -2006,14 +2010,6 @@ static int prctl_set_mm_map(int opt, const void __user *addr, unsigned long data }if (prctl_map.exe_fd != (u32)-1 ) { -/* -* Make sure the caller has the rights to -* change /proc/pid/exe link: only local sys admin should -* be allowed to. -*/ -if (!ns_capable (current_user_ns (),CAP_SYS_ADMIN )) -return -EINVAL ; - error =prctl_set_mm_exe_file (mm, prctl_map.exe_fd);if (error)return error;security/commoncap.c +14
-0@@ -1291,6 +1291,31 @@ int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, } } +/** +* cap_prctl_set_mm_exe_file - Determine whether /proc/self/exe can be changed +* by the current process. +* @exe_file: The new exe file +* Returns 0 if permission is granted, -ve if denied. +* +* The current process is permitted to change its /proc/self/exe link via two policies: +* 1) The current user can do checkpoint/restore. At the time of this writing, +* this means CAP_SYS_ADMIN or CAP_CHECKPOINT_RESTORE capable. +* 2) The current user can use ptrace. +* +* With access to ptrace facilities, a process can do the following: +* fork a child, execve() the target executable, and have the child use +* ptrace() to replace the memory content of the current process.
security/commoncap.c int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, · line 1317
Quoted 10 lines
security/commoncap.c +10
-0+* This technique makes it possible to masquerade an arbitrary program as the +* target executable, even if it is setuid. +*/ +int cap_prctl_set_mm_exe_file (struct file *exe_file) +{ +if (checkpoint_restore_ns_capable (current_user_ns ())) +return 0 ; + +return security_ptrace_access_check (current,PTRACE_MODE_ATTACH_REALCREDS ); +}
What is the reason for having this be a new security hook? Doesn't look like it needs to be unless I'm missing something. This just seems more complex than it needs to be.
I might be wrong here but if you look at the callsites for security_ptrace_access_check() right now, you'll see that it's only called from kernel/ptrace.c in __ptrace_may_access() and that function checks right at the top:
/* Don't let security modules deny introspection */ if (same_thread_group(task, current)) return 0;
since you're passing in same_thread_group(current, current) you're passing this check and never even hitting security_ptrace_access_check(). So the contract seems to be (as is obvious from the comment) that a task can't be denied ptrace introspection. But if you're using security_ptrace_access_check(current) here and _if_ there would be any lsm that would deny ptrace introspection to current you'd suddenly introduce a callsite where ptrace introspection is denied. That seems wrong. So either you meant to do something else here or you really just want:
checkpoint_restore_ns_capable(current_user_ns())
and none of the rest. But I might be missing the big picture in this patch.
Quoted 65 lines
security/commoncap.c +4
-0+if (checkpoint_restore_ns_capable (current_user_ns ())) +return 0 ; +/** * cap_vm_enough_memory - Determine whether a new virtual mapping is permitted * @mm: The VM space in which the new mapping is to be made @@ -1356,6 +1381,7 @@ static struct security_hook_list capability_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (mmap_file, cap_mmap_file),LSM_HOOK_INIT (task_fix_setuid, cap_task_fix_setuid),LSM_HOOK_INIT (task_prctl, cap_task_prctl), +LSM_HOOK_INIT (prctl_set_mm_exe_file, cap_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_setscheduler, cap_task_setscheduler),LSM_HOOK_INIT (task_setioprio, cap_task_setioprio), LSM_HOOK_INIT(task_setnice, cap_task_setnice),security/security.c +5
-0@@ -1790,6 +1790,11 @@ int security_task_prctl(int option, unsigned long arg2, unsigned long arg3,return rc; } +int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return call_int_hook (prctl_set_mm_exe_file,0 , exe_file); +} +void security_task_to_inode (struct task_struct *p,struct inode *inode) {call_void_hook (task_to_inode, p, inode);security/selinux/hooks.c +14
-0@@ -4156,6 +4156,19 @@ static int selinux_task_kill(struct task_struct *p, struct kernel_siginfo *info, secid, task_sid(p), SECCLASS_PROCESS, perm, NULL); } +static int selinux_prctl_set_mm_exe_file (struct file *exe_file) +{ +u32 sid =current_sid (); + +struct common_audit_data ad = { + .type =LSM_AUDIT_DATA_FILE , + .u.file = exe_file, + }; + +return avc_has_perm (&selinux_state, sid, sid, +SECCLASS_FILE ,FILE__EXECUTE_NO_TRANS , &ad); +} +static void selinux_task_to_inode (struct task_struct *p,struct inode *inode) {@@ -7057,6 +7070,7 @@ static struct security_hook_list selinux_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (task_getscheduler, selinux_task_getscheduler),LSM_HOOK_INIT (task_movememory, selinux_task_movememory),LSM_HOOK_INIT (task_kill, selinux_task_kill), +LSM_HOOK_INIT (prctl_set_mm_exe_file, selinux_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_to_inode, selinux_task_to_inode), LSM_HOOK_INIT(ipc_permission, selinux_ipc_permission),Signature · 1 line
2.26.2
kernel test robot
build botRe: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
In reply to Adrian ReberThank you for the patch! Yet something to improve:
Hi Adrian,
Thank you for the patch! Yet something to improve:
[auto build test ERROR on pcmoore-selinux/next]
[also build test ERROR on kselftest/next linus/master v5.8-rc3 next-20200701] [cannot apply to security/next-testing]
[If your patch is applied to the wrong git tree, kindly drop us a note. And when submitting patch, we suggest to use as documented in
https://git-scm.com/docs/git-format-patch]
url: https://github.com/0day-ci/linux/commits/Adrian-Reber/capabilities-Introduce-CAP_CHECKPOINT_RESTORE/20200701-145119 base: https://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/selinux.git next config: x86_64-allnoconfig (attached as .config) compiler: clang version 11.0.0 (https://github.com/llvm/llvm-project c8f1d442d0858f66fd4128fde6f67eb5202fa2b1)
reproduce (this is a W=1 build):
wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross chmod +x ~/bin/make.cross
# install x86_64 cross compiling tool for clang build
# apt-get install binutils-x86-64-linux-gnu
# save the attached .config to linux build tree
COMPILER_INSTALL_PATH=$HOME/0day COMPILER=clang make.cross ARCH=x86_64
If you fix the issue, kindly add following tag as appropriate
All errors (new ones prefixed by >>):
In file included from kernel/fork.c:50:
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file' [-Werror,-Wimplicit-function-declaration]
return cap_prctl_set_mm_exe_file(exe_file);
^
include/linux/security.h:1130:9: note: did you mean 'security_prctl_set_mm_exe_file'?
include/linux/security.h:1128:19: note: 'security_prctl_set_mm_exe_file' declared here
static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
^ 1 error generated.
--
In file included from kernel/events/core.c:34:
In file included from include/linux/syscalls.h:85:
In file included from include/trace/syscall.h:7:
In file included from include/linux/trace_events.h:10:
In file included from include/linux/perf_event.h:59:
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file' [-Werror,-Wimplicit-function-declaration]
return cap_prctl_set_mm_exe_file(exe_file);
^
include/linux/security.h:1130:9: note: did you mean 'security_prctl_set_mm_exe_file'?
include/linux/security.h:1128:19: note: 'security_prctl_set_mm_exe_file' declared here
static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
^
kernel/events/core.c:6490:6: warning: no previous prototype for function 'perf_pmu_snapshot_aux' [-Wmissing-prototypes]
long perf_pmu_snapshot_aux(struct perf_buffer *rb,
^ kernel/events/core.c:6490:1: note: declare 'static' if the function is not intended to be used outside of this translation unit long perf_pmu_snapshot_aux(struct perf_buffer *rb,
^
static
1 warning and 1 error generated.
--
In file included from kernel/printk/printk.c:34:
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file' [-Werror,-Wimplicit-function-declaration]
return cap_prctl_set_mm_exe_file(exe_file);
^
include/linux/security.h:1130:9: note: did you mean 'security_prctl_set_mm_exe_file'?
include/linux/security.h:1128:19: note: 'security_prctl_set_mm_exe_file' declared here
static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
^
In file included from kernel/printk/printk.c:60:
kernel/printk/internal.h:59:20: warning: no previous prototype for function 'vprintk_func' [-Wmissing-prototypes]
__printf(1, 0) int vprintk_func(const char *fmt, va_list args) { return 0; }
^
kernel/printk/internal.h:59:16: note: declare 'static' if the function is not intended to be used outside of this translation unit
__printf(1, 0) int vprintk_func(const char *fmt, va_list args) { return 0; }
^
static
kernel/printk/printk.c:174:5: warning: no previous prototype for function 'devkmsg_sysctl_set_loglvl' [-Wmissing-prototypes]
int devkmsg_sysctl_set_loglvl(struct ctl_table *table, int write,
^ kernel/printk/printk.c:174:1: note: declare 'static' if the function is not intended to be used outside of this translation unit int devkmsg_sysctl_set_loglvl(struct ctl_table *table, int write,
^
static
2 warnings and 1 error generated.
--
In file included from kernel/sched/fair.c:23:
In file included from kernel/sched/sched.h:61:
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file' [-Werror,-Wimplicit-function-declaration]
return cap_prctl_set_mm_exe_file(exe_file);
^
include/linux/security.h:1130:9: note: did you mean 'security_prctl_set_mm_exe_file'?
include/linux/security.h:1128:19: note: 'security_prctl_set_mm_exe_file' declared here
static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
^
kernel/sched/fair.c:5352:6: warning: no previous prototype for function 'init_cfs_bandwidth' [-Wmissing-prototypes]
void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
^
kernel/sched/fair.c:5352:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
void init_cfs_bandwidth(struct cfs_bandwidth *cfs_b) {}
^
static
kernel/sched/fair.c:11078:6: warning: no previous prototype for function 'free_fair_sched_group' [-Wmissing-prototypes]
void free_fair_sched_group(struct task_group *tg) { }
^
kernel/sched/fair.c:11078:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
void free_fair_sched_group(struct task_group *tg) { }
^
static
kernel/sched/fair.c:11080:5: warning: no previous prototype for function 'alloc_fair_sched_group' [-Wmissing-prototypes]
int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
^
kernel/sched/fair.c:11080:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
int alloc_fair_sched_group(struct task_group *tg, struct task_group *parent)
^
static
kernel/sched/fair.c:11085:6: warning: no previous prototype for function 'online_fair_sched_group' [-Wmissing-prototypes]
void online_fair_sched_group(struct task_group *tg) { }
^
kernel/sched/fair.c:11085:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
void online_fair_sched_group(struct task_group *tg) { }
^
static
kernel/sched/fair.c:11087:6: warning: no previous prototype for function 'unregister_fair_sched_group' [-Wmissing-prototypes]
void unregister_fair_sched_group(struct task_group *tg) { }
^ kernel/sched/fair.c:11087:1: note: declare 'static' if the function is not intended to be used outside of this translation unit void unregister_fair_sched_group(struct task_group *tg) { }
^
static
5 warnings and 1 error generated.
--
In file included from kernel/sched/rt.c:6:
In file included from kernel/sched/sched.h:61:
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file' [-Werror,-Wimplicit-function-declaration]
return cap_prctl_set_mm_exe_file(exe_file);
^
include/linux/security.h:1130:9: note: did you mean 'security_prctl_set_mm_exe_file'?
include/linux/security.h:1128:19: note: 'security_prctl_set_mm_exe_file' declared here
static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
^
kernel/sched/rt.c:253:6: warning: no previous prototype for function 'free_rt_sched_group' [-Wmissing-prototypes]
void free_rt_sched_group(struct task_group *tg) { }
^
kernel/sched/rt.c:253:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
void free_rt_sched_group(struct task_group *tg) { }
^
static
kernel/sched/rt.c:255:5: warning: no previous prototype for function 'alloc_rt_sched_group' [-Wmissing-prototypes]
int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent)
^
kernel/sched/rt.c:255:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
int alloc_rt_sched_group(struct task_group *tg, struct task_group *parent)
^
static
kernel/sched/rt.c:668:6: warning: no previous prototype for function 'sched_rt_bandwidth_account' [-Wmissing-prototypes]
bool sched_rt_bandwidth_account(struct rt_rq *rt_rq)
^ kernel/sched/rt.c:668:1: note: declare 'static' if the function is not intended to be used outside of this translation unit bool sched_rt_bandwidth_account(struct rt_rq *rt_rq)
^
static
3 warnings and 1 error generated.
--
In file included from kernel/time/hrtimer.c:30:
In file included from include/linux/syscalls.h:85:
In file included from include/trace/syscall.h:7:
In file included from include/linux/trace_events.h:10:
In file included from include/linux/perf_event.h:59:
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file' [-Werror,-Wimplicit-function-declaration]
return cap_prctl_set_mm_exe_file(exe_file);
^
include/linux/security.h:1130:9: note: did you mean 'security_prctl_set_mm_exe_file'?
include/linux/security.h:1128:19: note: 'security_prctl_set_mm_exe_file' declared here
static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
^
kernel/time/hrtimer.c:120:21: warning: initializer overrides prior initialization of this subobject [-Winitializer-overrides]
[CLOCK_REALTIME] = HRTIMER_BASE_REALTIME,
^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:118:27: note: previous initialization is here
[0 ... MAX_CLOCKS - 1] = HRTIMER_MAX_CLOCK_BASES,
^~~~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:121:22: warning: initializer overrides prior initialization of this subobject [-Winitializer-overrides]
[CLOCK_MONOTONIC] = HRTIMER_BASE_MONOTONIC,
^~~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:118:27: note: previous initialization is here
[0 ... MAX_CLOCKS - 1] = HRTIMER_MAX_CLOCK_BASES,
^~~~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:122:21: warning: initializer overrides prior initialization of this subobject [-Winitializer-overrides]
[CLOCK_BOOTTIME] = HRTIMER_BASE_BOOTTIME,
^~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:118:27: note: previous initialization is here
[0 ... MAX_CLOCKS - 1] = HRTIMER_MAX_CLOCK_BASES,
^~~~~~~~~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:123:17: warning: initializer overrides prior initialization of this subobject [-Winitializer-overrides]
[CLOCK_TAI] = HRTIMER_BASE_TAI,
^~~~~~~~~~~~~~~~
kernel/time/hrtimer.c:118:27: note: previous initialization is here
[0 ... MAX_CLOCKS - 1] = HRTIMER_MAX_CLOCK_BASES,
^~~~~~~~~~~~~~~~~~~~~~~ 4 warnings and 1 error generated.
--
In file included from kernel/time/posix-stubs.c:13:
In file included from include/linux/syscalls.h:85:
In file included from include/trace/syscall.h:7:
In file included from include/linux/trace_events.h:10:
In file included from include/linux/perf_event.h:59:
>> include/linux/security.h:1130:9: error: implicit declaration of function 'cap_prctl_set_mm_exe_file' [-Werror,-Wimplicit-function-declaration]
return cap_prctl_set_mm_exe_file(exe_file);
^
include/linux/security.h:1130:9: note: did you mean 'security_prctl_set_mm_exe_file'?
include/linux/security.h:1128:19: note: 'security_prctl_set_mm_exe_file' declared here
static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
^
kernel/time/posix-stubs.c:25:17: warning: no previous prototype for function 'sys_ni_posix_timers' [-Wmissing-prototypes]
asmlinkage long sys_ni_posix_timers(void)
^
kernel/time/posix-stubs.c:25:12: note: declare 'static' if the function is not intended to be used outside of this translation unit
asmlinkage long sys_ni_posix_timers(void)
^
static
kernel/time/posix-stubs.c:73:5: warning: no previous prototype for function 'do_clock_gettime' [-Wmissing-prototypes]
int do_clock_gettime(clockid_t which_clock, struct timespec64 *tp)
^ kernel/time/posix-stubs.c:73:1: note: declare 'static' if the function is not intended to be used outside of this translation unit int do_clock_gettime(clockid_t which_clock, struct timespec64 *tp)
^
static
2 warnings and 1 error generated.
vim +/cap_prctl_set_mm_exe_file +1130 include/linux/security.h
1127
1128 static inline int security_prctl_set_mm_exe_file(struct file *exe_file)
1129 {
> 1130 return cap_prctl_set_mm_exe_file(exe_file);
1131 }
1132
---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org
Serge E. Hallyn
maintainerRe: [PATCH v4 2/3] selftests: add clone3() CAP_CHECKPOINT_RESTORE test
In reply to Adrian ReberQuoted 3 lines · Adrian Reber
This adds a test that changes its UID, uses capabilities to get
CAP_CHECKPOINT_RESTOREand usesclone3()withset_tidto create a process with a given PID as non-root.
Seems worth also verifying that it fails if you have no capabilities. I don't see that in the existing clone3/ test dir.
Quoted 232 lines
Signed-off-by Adrian Reber <[email protected]>--- tools/testing/selftests/clone3/Makefile | 4 +- .../clone3/clone3_cap_checkpoint_restore.c | 203 ++++++++++++++++++ 2 files changed, 206 insertions(+), 1 deletion(-) create mode 100644 tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.ctools/testing/selftests/clone3/Makefile +3
-1@@ -1,6 +1,8 @@ # SPDX-License-Identifier: GPL-2.0 CFLAGS += -g -I../../../../usr/include/ +LDLIBS += -lcap -TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid +TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid \ + clone3_cap_checkpoint_restore include ../lib.mktools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c new file +203
-0@@ -0,0 +1,203 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* +* Based on Christian Brauner's clone3() example. +* These tests are assuming to be running in the host's +* PID namespace. +*/ + +/* capabilities related code based on selftests/bpf/test_verifier.c */ + +#define _GNU_SOURCE +#include <errno.h> +#include <linux/types.h> +#include <linux/sched.h> +#include <stdio.h> +#include <stdlib.h> +#include <stdbool.h> +#include <sys/capability.h> +#include <sys/prctl.h> +#include <sys/syscall.h> +#include <sys/types.h> +#include <sys/un.h> +#include <sys/wait.h> +#include <unistd.h> +#include <sched.h> + +#include "../kselftest.h" +#include "clone3_selftests.h" + +#ifndef MAX_PID_NS_LEVEL +#define MAX_PID_NS_LEVEL 32 +#endif + +static void child_exit (int ret) +{ +fflush (stdout); +fflush (stderr); +_exit (ret); +} + +static int call_clone3_set_tid (pid_t * set_tid,size_t set_tid_size) +{ +int status; +pid_t pid =-1 ; + +struct clone_args args = { + .exit_signal =SIGCHLD , + .set_tid =ptr_to_u64 (set_tid), + .set_tid_size = set_tid_size, + }; + + pid =sys_clone3 (&args,sizeof (struct clone_args )); +if (pid <0 ) { +ksft_print_msg ("%s - Failed to create new process\n" , +strerror (errno)); +return -errno; + } + +if (pid ==0 ) { +int ret; +char tmp =0 ; + +ksft_print_msg + ("I am the child, my PID is %d (expected %d)\n" , +getpid (), set_tid[0 ]); + +if (set_tid[0 ] !=getpid ()) +child_exit (EXIT_FAILURE ); +child_exit (EXIT_SUCCESS ); + } + +ksft_print_msg ("I am the parent (%d). My child's pid is %d\n" , +getpid (), pid); + +if (waitpid (pid, &status,0 ) <0 ) { +ksft_print_msg ("Child returned %s\n" ,strerror (errno)); +return -errno; + } + +if (!WIFEXITED (status)) +return -1 ; + +return WEXITSTATUS (status); +} + +static int test_clone3_set_tid (pid_t * set_tid, +size_t set_tid_size,int expected) +{ +int ret; + +ksft_print_msg ("[%d] Trying clone3() with CLONE_SET_TID to %d\n" , +getpid (), set_tid[0 ]); + ret =call_clone3_set_tid (set_tid, set_tid_size); + +ksft_print_msg + ("[%d] clone3() with CLONE_SET_TID %d says :%d - expected %d\n" , +getpid (), set_tid[0 ], ret, expected); +if (ret != expected) { +ksft_test_result_fail + ("[%d] Result (%d) is different than expected (%d)\n" , +getpid (), ret, expected); +return -1 ; + } +ksft_test_result_pass + ("[%d] Result (%d) matches expectation (%d)\n" ,getpid (), ret, + expected); + +return 0 ; +} + +struct libcap { +struct __user_cap_header_struct hdr; +struct __user_cap_data_struct data[2 ]; +}; + +static int set_capability () +{ +cap_value_t cap_values[] = {CAP_SETUID ,CAP_SETGID }; +struct libcap *cap; +int ret =-1 ; +cap_t caps; + + caps =cap_get_proc (); +if (!caps) { +perror ("cap_get_proc" ); +return -1 ; + } + +/* Drop all capabilities */ +if (cap_clear (caps)) { +perror ("cap_clear" ); + goto out; + } + +cap_set_flag (caps,CAP_EFFECTIVE ,2 , cap_values,CAP_SET ); +cap_set_flag (caps,CAP_PERMITTED ,2 , cap_values,CAP_SET ); + + cap = (struct libcap *) caps; + +/* 40 -> CAP_CHECKPOINT_RESTORE */ + cap->data[1 ].effective |=1 << (40 -32 ); + cap->data[1 ].permitted |=1 << (40 -32 ); + +if (cap_set_proc (caps)) { +perror ("cap_set_proc" ); + goto out; + } + ret =0 ; +out: +if (cap_free (caps)) +perror ("cap_free" ); +return ret; +} + +int main (int argc,char *argv[]) +{ +pid_t pid; +int status; +int ret =0 ; +pid_t set_tid[1 ]; +uid_t uid =getuid (); + +ksft_print_header (); +test_clone3_supported (); +ksft_set_plan (2 ); + +if (uid !=0 ) { + ksft_cnt.ksft_xskip = ksft_plan; +ksft_print_msg ("Skipping all tests as non-root\n" ); +return ksft_exit_pass (); + } + +memset (&set_tid,0 ,sizeof (set_tid)); + +/* Find the current active PID */ + pid =fork (); +if (pid ==0 ) { +ksft_print_msg ("Child has PID %d\n" ,getpid ()); +child_exit (EXIT_SUCCESS ); + } +if (waitpid (pid, &status,0 ) <0 ) +ksft_exit_fail_msg ("Waiting for child %d failed" , pid); + +/* After the child has finished, its PID should be free. */ + set_tid[0 ] = pid; + +if (set_capability ()) +ksft_test_result_fail + ("Could not set CAP_CHECKPOINT_RESTORE\n" ); +prctl (PR_SET_KEEPCAPS ,1 ,0 ,0 ,0 ); +/* This would fail without CAP_CHECKPOINT_RESTORE */ +setgid (1000 ); +setuid (1000 ); + set_tid[0 ] = pid; + ret |=test_clone3_set_tid (set_tid,1 , -EPERM ); +if (set_capability ()) +ksft_test_result_fail + ("Could not set CAP_CHECKPOINT_RESTORE\n" ); +/* This should work as we have CAP_CHECKPOINT_RESTORE as non-root */ + ret |=test_clone3_set_tid (set_tid,1 ,0 ); + +return !ret ?ksft_exit_pass () :ksft_exit_fail (); +}Signature · 1 line
2.26.2
Re: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
In reply to Adrian ReberQuoted 7 lines · Adrian Reber
From: Nicolas Viennot <[email protected]>
Previously, the current process could only change the
/proc/self/exelink with localCAP_SYS_ADMIN.
This commit relaxes this restriction by permitting such change withCAP_CHECKPOINT_RESTORE, and the ability to use ptrace.
Quoted 8 lines
With access to ptrace facilities, a process can do the following: fork a child,
execve()the target executable, and have the child useptrace()to replace the memory content of the current process. This technique makes it possible to masquerade an arbitrary program as any executable, even setuid ones.Signed-off-by Nicolas Viennot <[email protected]>Signed-off-by Adrian Reber <[email protected]>
This is scary. But I believe it is safe.
I am a bit curious about the implications of the selinux patch. IIUC you are using the permission of the tracing process to execute the file without transition, so this is a way to work around the policy which might prevent the tracee from doing so. Given that SELinux wants to be MAC, I'm not *quite* sure that's considered kosher. You also are skipping the PROCESS__PTRACE to SECCLASS_PROCESS check which selinux_bprm_set_creds does later on. Again I'm just not quite sure what's considered normal there these days.
Paul, do you have input there?
Quoted 169 lines
--- include/linux/lsm_hook_defs.h | 1 + include/linux/security.h | 6 ++++++ kernel/sys.c | 12 ++++-------- security/commoncap.c | 26 ++++++++++++++++++++++++++ security/security.c | 5 +++++ security/selinux/hooks.c | 14 ++++++++++++++ 6 files changed, 56 insertions(+), 8 deletions(-)include/linux/lsm_hook_defs.h +1
-0@@ -211,6 +211,7 @@ LSM_HOOK(int, 0, task_kill, struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred) LSM_HOOK(int, -ENOSYS, task_prctl, int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5) +LSM_HOOK(int, 0, prctl_set_mm_exe_file, struct file *exe_file) LSM_HOOK(void, LSM_RET_VOID, task_to_inode, struct task_struct *p, struct inode *inode) LSM_HOOK(int, 0, ipc_permission, struct kern_ipc_perm *ipcp, short flag)include/linux/security.h +6
-0@@ -412,6 +412,7 @@ int security_task_kill(struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred);int security_task_prctl (int option,unsigned long arg2,unsigned long arg3,unsigned long arg4,unsigned long arg5); +int security_prctl_set_mm_exe_file (struct file *exe_file);void security_task_to_inode (struct task_struct *p,struct inode *inode);int security_ipc_permission (struct kern_ipc_perm *ipcp,short flag);void security_ipc_getsecid (struct kern_ipc_perm *ipcp,u32 *secid);@@ -1124,6 +1125,11 @@ static inline int security_task_prctl(int option, unsigned long arg2,return cap_task_prctl (option, arg2, arg3, arg4, arg5); } +static inline int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return cap_prctl_set_mm_exe_file (exe_file); +} +static inline void security_task_to_inode (struct task_struct *p,struct inode *inode) { }kernel/sys.c +4
-8@@ -1851,6 +1851,10 @@ static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd)if (err) goto exit; + err =security_prctl_set_mm_exe_file (exe.file); +if (err) + goto exit; +/* * Forbid mm->exe_file change if old file still mapped. */ @@ -2006,14 +2010,6 @@ static int prctl_set_mm_map(int opt, const void __user *addr, unsigned long data }if (prctl_map.exe_fd != (u32)-1 ) { -/* -* Make sure the caller has the rights to -* change /proc/pid/exe link: only local sys admin should -* be allowed to. -*/ -if (!ns_capable (current_user_ns (),CAP_SYS_ADMIN )) -return -EINVAL ; - error =prctl_set_mm_exe_file (mm, prctl_map.exe_fd);if (error)return error;security/commoncap.c +26
-0@@ -1291,6 +1291,31 @@ int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, } } +/** +* cap_prctl_set_mm_exe_file - Determine whether /proc/self/exe can be changed +* by the current process. +* @exe_file: The new exe file +* Returns 0 if permission is granted, -ve if denied. +* +* The current process is permitted to change its /proc/self/exe link via two policies: +* 1) The current user can do checkpoint/restore. At the time of this writing, +* this means CAP_SYS_ADMIN or CAP_CHECKPOINT_RESTORE capable. +* 2) The current user can use ptrace. +* +* With access to ptrace facilities, a process can do the following: +* fork a child, execve() the target executable, and have the child use +* ptrace() to replace the memory content of the current process. +* This technique makes it possible to masquerade an arbitrary program as the +* target executable, even if it is setuid. +*/ +int cap_prctl_set_mm_exe_file (struct file *exe_file) +{ +if (checkpoint_restore_ns_capable (current_user_ns ())) +return 0 ; + +return security_ptrace_access_check (current,PTRACE_MODE_ATTACH_REALCREDS ); +} +/** * cap_vm_enough_memory - Determine whether a new virtual mapping is permitted * @mm: The VM space in which the new mapping is to be made @@ -1356,6 +1381,7 @@ static struct security_hook_list capability_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (mmap_file, cap_mmap_file),LSM_HOOK_INIT (task_fix_setuid, cap_task_fix_setuid),LSM_HOOK_INIT (task_prctl, cap_task_prctl), +LSM_HOOK_INIT (prctl_set_mm_exe_file, cap_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_setscheduler, cap_task_setscheduler),LSM_HOOK_INIT (task_setioprio, cap_task_setioprio), LSM_HOOK_INIT(task_setnice, cap_task_setnice),security/security.c +5
-0@@ -1790,6 +1790,11 @@ int security_task_prctl(int option, unsigned long arg2, unsigned long arg3,return rc; } +int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return call_int_hook (prctl_set_mm_exe_file,0 , exe_file); +} +void security_task_to_inode (struct task_struct *p,struct inode *inode) {call_void_hook (task_to_inode, p, inode);security/selinux/hooks.c +14
-0@@ -4156,6 +4156,19 @@ static int selinux_task_kill(struct task_struct *p, struct kernel_siginfo *info, secid, task_sid(p), SECCLASS_PROCESS, perm, NULL); } +static int selinux_prctl_set_mm_exe_file (struct file *exe_file) +{ +u32 sid =current_sid (); + +struct common_audit_data ad = { + .type =LSM_AUDIT_DATA_FILE , + .u.file = exe_file, + }; + +return avc_has_perm (&selinux_state, sid, sid, +SECCLASS_FILE ,FILE__EXECUTE_NO_TRANS , &ad); +} +static void selinux_task_to_inode (struct task_struct *p,struct inode *inode) {@@ -7057,6 +7070,7 @@ static struct security_hook_list selinux_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (task_getscheduler, selinux_task_getscheduler),LSM_HOOK_INIT (task_movememory, selinux_task_movememory),LSM_HOOK_INIT (task_kill, selinux_task_kill), +LSM_HOOK_INIT (prctl_set_mm_exe_file, selinux_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_to_inode, selinux_task_to_inode), LSM_HOOK_INIT(ipc_permission, selinux_ipc_permission),Signature · 1 line
2.26.2
Serge E. Hallyn
maintainerRe: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
In reply to Christian BraunerQuoted 119 lines · Christian Brauner
Quoted 118 lines · Adrian Reber
From: Nicolas Viennot <[email protected]>
Previously, the current process could only change the
/proc/self/exelink with localCAP_SYS_ADMIN.
This commit relaxes this restriction by permitting such change withCAP_CHECKPOINT_RESTORE, and the ability to use ptrace.With access to ptrace facilities, a process can do the following: fork a child,
execve()the target executable, and have the child useptrace()to replace the memory content of the current process. This technique makes it possible to masquerade an arbitrary program as any executable, even setuid ones.Signed-off-by Nicolas Viennot <[email protected]>Signed-off-by Adrian Reber <[email protected]>--- include/linux/lsm_hook_defs.h | 1 + include/linux/security.h | 6 ++++++ kernel/sys.c | 12 ++++-------- security/commoncap.c | 26 ++++++++++++++++++++++++++ security/security.c | 5 +++++ security/selinux/hooks.c | 14 ++++++++++++++ 6 files changed, 56 insertions(+), 8 deletions(-)include/linux/lsm_hook_defs.h +1
-0@@ -211,6 +211,7 @@ LSM_HOOK(int, 0, task_kill, struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred) LSM_HOOK(int, -ENOSYS, task_prctl, int option, unsigned long arg2, unsigned long arg3, unsigned long arg4, unsigned long arg5) +LSM_HOOK(int, 0, prctl_set_mm_exe_file, struct file *exe_file) LSM_HOOK(void, LSM_RET_VOID, task_to_inode, struct task_struct *p, struct inode *inode) LSM_HOOK(int, 0, ipc_permission, struct kern_ipc_perm *ipcp, short flag)include/linux/security.h +6
-0@@ -412,6 +412,7 @@ int security_task_kill(struct task_struct *p, struct kernel_siginfo *info, int sig, const struct cred *cred);int security_task_prctl (int option,unsigned long arg2,unsigned long arg3,unsigned long arg4,unsigned long arg5); +int security_prctl_set_mm_exe_file (struct file *exe_file);void security_task_to_inode (struct task_struct *p,struct inode *inode);int security_ipc_permission (struct kern_ipc_perm *ipcp,short flag);void security_ipc_getsecid (struct kern_ipc_perm *ipcp,u32 *secid);@@ -1124,6 +1125,11 @@ static inline int security_task_prctl(int option, unsigned long arg2,return cap_task_prctl (option, arg2, arg3, arg4, arg5); } +static inline int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return cap_prctl_set_mm_exe_file (exe_file); +} +static inline void security_task_to_inode (struct task_struct *p,struct inode *inode) { }kernel/sys.c +4
-8@@ -1851,6 +1851,10 @@ static int prctl_set_mm_exe_file(struct mm_struct *mm, unsigned int fd)if (err) goto exit; + err =security_prctl_set_mm_exe_file (exe.file); +if (err) + goto exit; +/* * Forbid mm->exe_file change if old file still mapped. */ @@ -2006,14 +2010,6 @@ static int prctl_set_mm_map(int opt, const void __user *addr, unsigned long data }if (prctl_map.exe_fd != (u32)-1 ) { -/* -* Make sure the caller has the rights to -* change /proc/pid/exe link: only local sys admin should -* be allowed to. -*/ -if (!ns_capable (current_user_ns (),CAP_SYS_ADMIN )) -return -EINVAL ; - error =prctl_set_mm_exe_file (mm, prctl_map.exe_fd);if (error)return error;security/commoncap.c +20
-0@@ -1291,6 +1291,31 @@ int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, } } +/** +* cap_prctl_set_mm_exe_file - Determine whether /proc/self/exe can be changed +* by the current process. +* @exe_file: The new exe file +* Returns 0 if permission is granted, -ve if denied. +* +* The current process is permitted to change its /proc/self/exe link via two policies: +* 1) The current user can do checkpoint/restore. At the time of this writing, +* this means CAP_SYS_ADMIN or CAP_CHECKPOINT_RESTORE capable. +* 2) The current user can use ptrace. +* +* With access to ptrace facilities, a process can do the following: +* fork a child, execve() the target executable, and have the child use +* ptrace() to replace the memory content of the current process. +* This technique makes it possible to masquerade an arbitrary program as the +* target executable, even if it is setuid. +*/ +int cap_prctl_set_mm_exe_file (struct file *exe_file) +{ +if (checkpoint_restore_ns_capable (current_user_ns ()))
Quoted 8 lines
security/commoncap.c int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, · line 1317
Quoted 4 lines
security/commoncap.c +4
-0+return 0 ; + +return security_ptrace_access_check (current,PTRACE_MODE_ATTACH_REALCREDS ); +}What is the reason for having this be a new security hook? Doesn't look like it needs to be unless I'm missing something. This just seems more complex than it needs to be.
Yeah, agreed, and that in turn risks actually making it less safe. Is there a good reason not to do what Christian suggests?
Quoted 89 lines
I might be wrong here but if you look at the callsites for
security_ptrace_access_check()right now, you'll see that it's only called fromkernel/ptrace.cin__ptrace_may_access()and that function checks right at the top:/* Don't let security modules deny introspection */ if (same_thread_group(task, current)) return 0;since you're passing in
same_thread_group(current, current)you're passing this check and never even hittingsecurity_ptrace_access_check(). So the contract seems to be (as is obvious from the comment) that a task can't be denied ptrace introspection. But if you're usingsecurity_ptrace_access_check(current)here and _if_ there would be any lsm that would deny ptrace introspection to current you'd suddenly introduce a callsite where ptrace introspection is denied. That seems wrong. So either you meant to do something else here or you really just want:checkpoint_restore_ns_capable(current_user_ns())
and none of the rest. But I might be missing the big picture in this patch.
Quoted 65 lines
security/commoncap.c +4
-0+if (checkpoint_restore_ns_capable (current_user_ns ())) +return 0 ; +/** * cap_vm_enough_memory - Determine whether a new virtual mapping is permitted * @mm: The VM space in which the new mapping is to be made @@ -1356,6 +1381,7 @@ static struct security_hook_list capability_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (mmap_file, cap_mmap_file),LSM_HOOK_INIT (task_fix_setuid, cap_task_fix_setuid),LSM_HOOK_INIT (task_prctl, cap_task_prctl), +LSM_HOOK_INIT (prctl_set_mm_exe_file, cap_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_setscheduler, cap_task_setscheduler),LSM_HOOK_INIT (task_setioprio, cap_task_setioprio), LSM_HOOK_INIT(task_setnice, cap_task_setnice),security/security.c +5
-0@@ -1790,6 +1790,11 @@ int security_task_prctl(int option, unsigned long arg2, unsigned long arg3,return rc; } +int security_prctl_set_mm_exe_file (struct file *exe_file) +{ +return call_int_hook (prctl_set_mm_exe_file,0 , exe_file); +} +void security_task_to_inode (struct task_struct *p,struct inode *inode) {call_void_hook (task_to_inode, p, inode);security/selinux/hooks.c +14
-0@@ -4156,6 +4156,19 @@ static int selinux_task_kill(struct task_struct *p, struct kernel_siginfo *info, secid, task_sid(p), SECCLASS_PROCESS, perm, NULL); } +static int selinux_prctl_set_mm_exe_file (struct file *exe_file) +{ +u32 sid =current_sid (); + +struct common_audit_data ad = { + .type =LSM_AUDIT_DATA_FILE , + .u.file = exe_file, + }; + +return avc_has_perm (&selinux_state, sid, sid, +SECCLASS_FILE ,FILE__EXECUTE_NO_TRANS , &ad); +} +static void selinux_task_to_inode (struct task_struct *p,struct inode *inode) {@@ -7057,6 +7070,7 @@ static struct security_hook_list selinux_hooks[] __lsm_ro_after_init = {LSM_HOOK_INIT (task_getscheduler, selinux_task_getscheduler),LSM_HOOK_INIT (task_movememory, selinux_task_movememory),LSM_HOOK_INIT (task_kill, selinux_task_kill), +LSM_HOOK_INIT (prctl_set_mm_exe_file, selinux_prctl_set_mm_exe_file),LSM_HOOK_INIT (task_to_inode, selinux_task_to_inode), LSM_HOOK_INIT(ipc_permission, selinux_ipc_permission),Signature · 1 line
2.26.2
Paul Moore
maintainerRe: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
In reply to Serge E. HallynQuoted 31 lines · Serge E. Hallyn
Quoted 7 lines · Adrian Reber
From: Nicolas Viennot <[email protected]>
Previously, the current process could only change the
/proc/self/exelink with localCAP_SYS_ADMIN.
This commit relaxes this restriction by permitting such change withCAP_CHECKPOINT_RESTORE, and the ability to use ptrace.Quoted 8 lines
With access to ptrace facilities, a process can do the following: fork a child,
execve()the target executable, and have the child useptrace()to replace the memory content of the current process. This technique makes it possible to masquerade an arbitrary program as any executable, even setuid ones.Signed-off-by Nicolas Viennot <[email protected]>Signed-off-by Adrian Reber <[email protected]>This is scary. But I believe it is safe.
Reviewed-by Serge Hallyn <[email protected]>I am a bit curious about the implications of the selinux patch. IIUC you are using the permission of the tracing process to execute the file without transition, so this is a way to work around the policy which might prevent the tracee from doing so. Given that SELinux wants to be MAC, I'm not *quite* sure that's considered kosher. You also are skipping the
PROCESS__PTRACEtoSECCLASS_PROCESScheck whichselinux_bprm_set_credsdoes later on. Again I'm just not quite sure what's considered normal there these days.
Quoted 1 line
Paul, do you have input there?
I agree, the SELinux hook looks wrong. Building on what Christian said, this looks more like a ptrace operation than an exec operation.
Signature · 2 lines
paul moore www.paul-moore.com
Quoted 66 lines · Christian Brauner
Quoted 49 lines · Adrian Reber
This patch introduces
CAP_CHECKPOINT_RESTORE, a new capability facilitating checkpoint/restore for non-root users.Over the last years, The CRIU (Checkpoint/Restore In Userspace) team has been asked numerous times if it is possible to checkpoint/restore a process as non-root. The answer usually was: 'almost'.
The main blocker to restore a process as non-root was to control the PID of the restored process. This feature available via the clone3 system call, or via
/proc/sys/kernel/ns_last_pidis unfortunately guarded byCAP_SYS_ADMIN.In the past two years, requests for non-root checkpoint/restore have increased due to the following use cases:
* Checkpoint/Restore in an HPC environment in combination with a resource manager distributing jobs where users are always running as non-root. There is a desire to provide a way to checkpoint and restore long running jobs.
* Container migration as non-root
* We have been in contact with JVM developers who are integrating CRIU into a Java VM to decrease the startup time. These checkpoint/restore applications are not meant to be running withCAP_SYS_ADMIN.We have seen the following workarounds:
* Use a setuid wrapper around CRIU:
See https://github.com/FredHutch/slurm-examples/blob/master/checkpointer/lib/checkpointer/checkpointer-suid.c
* Use a setuid helper that writes tons_last_pid.
Unfortunately, this helper delegation technique is impossible to use with
clone3, and is thus prone to races.
See https://github.com/twosigma/set_ns_last_pid
* Cycle through PIDs withfork()until the desired PID is reached:
This has been demonstrated to work with cycling rates of 100,000 PIDs/s
See https://github.com/twosigma/set_ns_last_pid
* Patch out theCAP_SYS_ADMINcheck from the kernel
* Run the desired application in a new user and PID namespace to provide
a localCAP_SYS_ADMINfor controlling PIDs. This technique has limited use in
typical container environments (e.g., Kubernetes) as /proc is
typically protected with read-only layers (e.g.,/proc/sys) for hardening
purposes. Read-only layers prevent additional /proc mounts (due to proc'sSB_I_USERNS_VISIBLEproperty), making the use of new PID namespaces limited as
certain applications need access to /proc matching their PID namespace.The introduced capability allows to:
* Control PIDs when the current user isCAP_CHECKPOINT_RESTOREcapable for the corresponding PID namespace via ns_last_pid/clone3.
* Open files in/proc/pid/map_fileswhen the current user isCAP_CHECKPOINT_RESTOREcapable in the root namespace, useful for recovering files that are unreachable via the file system such as deleted files, or memfd files.Quoted 5 lines
See corresponding selftest for an example with
clone3().Signed-off-by Adrian Reber <[email protected]>Signed-off-by Nicolas Viennot <[email protected]>---
I think that now looks reasonable. A few comments.
Before we proceed, please split the addition of
checkpoint_restore_ns_capable()out into a separate patch.
In fact, I think the cleanest way of doing this would be:
- 0/n capability: addCAP_CHECKPOINT_RESTORE
- 1/n pid: usecheckpoint_restore_ns_capable()forset_tid
- 2/npid_namespace: usecheckpoint_restore_ns_capable()forns_last_pid
- 3/n: proc: requirecheckpoint_restore_ns_capable()in init userns formap_files
Quoted 4 lines
(commit subjects up to you of course) and a nice commit message for each time we relax a permissions on something so we have a clear separate track record for each change in case we need to revert something. Then the rest of the patches in this series. Testing patches probably last.
Yes, makes sense. I was thinking about this already, but I was not sure if it I should do it or not. But I had the same idea already.
Adrian
Re: [PATCH v4 2/3] selftests: add clone3() CAP_CHECKPOINT_RESTORE test
In reply to Serge E. HallynQuoted 7 lines · Serge E. Hallyn
Quoted 3 lines · Adrian Reber
This adds a test that changes its UID, uses capabilities to get
CAP_CHECKPOINT_RESTOREand usesclone3()withset_tidto create a process with a given PID as non-root.Seems worth also verifying that it fails if you have no capabilities. I don't see that in the existing clone3/ test dir.
Bit confused about what you mean. This test does:
* switch UID to 1000
* run clone3() with set_tid set and expect EPERM
* set CAP_CHECKPOINT_RESTORE capability
* run clone3() with set_tid set and expect success
So it already does what I think you are asking for. Did I misunderstand your comment?
Adrian
Quoted 233 lines
Quoted 232 lines
Signed-off-by Adrian Reber <[email protected]>--- tools/testing/selftests/clone3/Makefile | 4 +- .../clone3/clone3_cap_checkpoint_restore.c | 203 ++++++++++++++++++ 2 files changed, 206 insertions(+), 1 deletion(-) create mode 100644 tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.ctools/testing/selftests/clone3/Makefile +3
-1@@ -1,6 +1,8 @@ # SPDX-License-Identifier: GPL-2.0 CFLAGS += -g -I../../../../usr/include/ +LDLIBS += -lcap -TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid +TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid \ + clone3_cap_checkpoint_restore include ../lib.mktools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c new file +203
-0@@ -0,0 +1,203 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* +* Based on Christian Brauner's clone3() example. +* These tests are assuming to be running in the host's +* PID namespace. +*/ + +/* capabilities related code based on selftests/bpf/test_verifier.c */ + +#define _GNU_SOURCE +#include <errno.h> +#include <linux/types.h> +#include <linux/sched.h> +#include <stdio.h> +#include <stdlib.h> +#include <stdbool.h> +#include <sys/capability.h> +#include <sys/prctl.h> +#include <sys/syscall.h> +#include <sys/types.h> +#include <sys/un.h> +#include <sys/wait.h> +#include <unistd.h> +#include <sched.h> + +#include "../kselftest.h" +#include "clone3_selftests.h" + +#ifndef MAX_PID_NS_LEVEL +#define MAX_PID_NS_LEVEL 32 +#endif + +static void child_exit (int ret) +{ +fflush (stdout); +fflush (stderr); +_exit (ret); +} + +static int call_clone3_set_tid (pid_t * set_tid,size_t set_tid_size) +{ +int status; +pid_t pid =-1 ; + +struct clone_args args = { + .exit_signal =SIGCHLD , + .set_tid =ptr_to_u64 (set_tid), + .set_tid_size = set_tid_size, + }; + + pid =sys_clone3 (&args,sizeof (struct clone_args )); +if (pid <0 ) { +ksft_print_msg ("%s - Failed to create new process\n" , +strerror (errno)); +return -errno; + } + +if (pid ==0 ) { +int ret; +char tmp =0 ; + +ksft_print_msg + ("I am the child, my PID is %d (expected %d)\n" , +getpid (), set_tid[0 ]); + +if (set_tid[0 ] !=getpid ()) +child_exit (EXIT_FAILURE ); +child_exit (EXIT_SUCCESS ); + } + +ksft_print_msg ("I am the parent (%d). My child's pid is %d\n" , +getpid (), pid); + +if (waitpid (pid, &status,0 ) <0 ) { +ksft_print_msg ("Child returned %s\n" ,strerror (errno)); +return -errno; + } + +if (!WIFEXITED (status)) +return -1 ; + +return WEXITSTATUS (status); +} + +static int test_clone3_set_tid (pid_t * set_tid, +size_t set_tid_size,int expected) +{ +int ret; + +ksft_print_msg ("[%d] Trying clone3() with CLONE_SET_TID to %d\n" , +getpid (), set_tid[0 ]); + ret =call_clone3_set_tid (set_tid, set_tid_size); + +ksft_print_msg + ("[%d] clone3() with CLONE_SET_TID %d says :%d - expected %d\n" , +getpid (), set_tid[0 ], ret, expected); +if (ret != expected) { +ksft_test_result_fail + ("[%d] Result (%d) is different than expected (%d)\n" , +getpid (), ret, expected); +return -1 ; + } +ksft_test_result_pass + ("[%d] Result (%d) matches expectation (%d)\n" ,getpid (), ret, + expected); + +return 0 ; +} + +struct libcap { +struct __user_cap_header_struct hdr; +struct __user_cap_data_struct data[2 ]; +}; + +static int set_capability () +{ +cap_value_t cap_values[] = {CAP_SETUID ,CAP_SETGID }; +struct libcap *cap; +int ret =-1 ; +cap_t caps; + + caps =cap_get_proc (); +if (!caps) { +perror ("cap_get_proc" ); +return -1 ; + } + +/* Drop all capabilities */ +if (cap_clear (caps)) { +perror ("cap_clear" ); + goto out; + } + +cap_set_flag (caps,CAP_EFFECTIVE ,2 , cap_values,CAP_SET ); +cap_set_flag (caps,CAP_PERMITTED ,2 , cap_values,CAP_SET ); + + cap = (struct libcap *) caps; + +/* 40 -> CAP_CHECKPOINT_RESTORE */ + cap->data[1 ].effective |=1 << (40 -32 ); + cap->data[1 ].permitted |=1 << (40 -32 ); + +if (cap_set_proc (caps)) { +perror ("cap_set_proc" ); + goto out; + } + ret =0 ; +out: +if (cap_free (caps)) +perror ("cap_free" ); +return ret; +} + +int main (int argc,char *argv[]) +{ +pid_t pid; +int status; +int ret =0 ; +pid_t set_tid[1 ]; +uid_t uid =getuid (); + +ksft_print_header (); +test_clone3_supported (); +ksft_set_plan (2 ); + +if (uid !=0 ) { + ksft_cnt.ksft_xskip = ksft_plan; +ksft_print_msg ("Skipping all tests as non-root\n" ); +return ksft_exit_pass (); + } + +memset (&set_tid,0 ,sizeof (set_tid)); + +/* Find the current active PID */ + pid =fork (); +if (pid ==0 ) { +ksft_print_msg ("Child has PID %d\n" ,getpid ()); +child_exit (EXIT_SUCCESS ); + } +if (waitpid (pid, &status,0 ) <0 ) +ksft_exit_fail_msg ("Waiting for child %d failed" , pid); + +/* After the child has finished, its PID should be free. */ + set_tid[0 ] = pid; + +if (set_capability ()) +ksft_test_result_fail + ("Could not set CAP_CHECKPOINT_RESTORE\n" ); +prctl (PR_SET_KEEPCAPS ,1 ,0 ,0 ,0 ); +/* This would fail without CAP_CHECKPOINT_RESTORE */ +setgid (1000 ); +setuid (1000 ); + set_tid[0 ] = pid; + ret |=test_clone3_set_tid (set_tid,1 , -EPERM ); +if (set_capability ()) +ksft_test_result_fail + ("Could not set CAP_CHECKPOINT_RESTORE\n" ); +/* This should work as we have CAP_CHECKPOINT_RESTORE as non-root */ + ret |=test_clone3_set_tid (set_tid,1 ,0 ); + +return !ret ?ksft_exit_pass () :ksft_exit_fail (); +}Signature · 1 line
2.26.2
Re: [PATCH v4 2/3] selftests: add clone3() CAP_CHECKPOINT_RESTORE test
Quoted 11 lines · Adrian Reber
Quoted 7 lines · Serge E. Hallyn
Quoted 3 lines · Adrian Reber
This adds a test that changes its UID, uses capabilities to get
CAP_CHECKPOINT_RESTOREand usesclone3()withset_tidto create a process with a given PID as non-root.Seems worth also verifying that it fails if you have no capabilities. I don't see that in the existing clone3/ test dir.
Bit confused about what you mean. This test does:
Quoted 7 lines
* switch UID to 1000
* runclone3()withset_tidset and expect EPERM
* setCAP_CHECKPOINT_RESTOREcapability
* runclone3()withset_tidset and expect successSo it already does what I think you are asking for. Did I misunderstand your comment?
Ah, no, I missed that line doing the call with -EPERM. Thanks!
Quoted 235 lines
Adrian
Quoted 233 lines
Quoted 232 lines
Signed-off-by Adrian Reber <[email protected]>--- tools/testing/selftests/clone3/Makefile | 4 +- .../clone3/clone3_cap_checkpoint_restore.c | 203 ++++++++++++++++++ 2 files changed, 206 insertions(+), 1 deletion(-) create mode 100644 tools/testing/selftests/clone3/clone3_cap_checkpoint_restore.ctools/testing/selftests/clone3/Makefile +3
-1@@ -1,6 +1,8 @@ # SPDX-License-Identifier: GPL-2.0 CFLAGS += -g -I../../../../usr/include/ +LDLIBS += -lcap -TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid +TEST_GEN_PROGS := clone3 clone3_clear_sighand clone3_set_tid \ + clone3_cap_checkpoint_restore include ../lib.mktools/testing/selftests/clone3/clone3_cap_checkpoint_restore.c new file +203
-0@@ -0,0 +1,203 @@ +// SPDX-License-Identifier: GPL-2.0 + +/* +* Based on Christian Brauner's clone3() example. +* These tests are assuming to be running in the host's +* PID namespace. +*/ + +/* capabilities related code based on selftests/bpf/test_verifier.c */ + +#define _GNU_SOURCE +#include <errno.h> +#include <linux/types.h> +#include <linux/sched.h> +#include <stdio.h> +#include <stdlib.h> +#include <stdbool.h> +#include <sys/capability.h> +#include <sys/prctl.h> +#include <sys/syscall.h> +#include <sys/types.h> +#include <sys/un.h> +#include <sys/wait.h> +#include <unistd.h> +#include <sched.h> + +#include "../kselftest.h" +#include "clone3_selftests.h" + +#ifndef MAX_PID_NS_LEVEL +#define MAX_PID_NS_LEVEL 32 +#endif + +static void child_exit (int ret) +{ +fflush (stdout); +fflush (stderr); +_exit (ret); +} + +static int call_clone3_set_tid (pid_t * set_tid,size_t set_tid_size) +{ +int status; +pid_t pid =-1 ; + +struct clone_args args = { + .exit_signal =SIGCHLD , + .set_tid =ptr_to_u64 (set_tid), + .set_tid_size = set_tid_size, + }; + + pid =sys_clone3 (&args,sizeof (struct clone_args )); +if (pid <0 ) { +ksft_print_msg ("%s - Failed to create new process\n" , +strerror (errno)); +return -errno; + } + +if (pid ==0 ) { +int ret; +char tmp =0 ; + +ksft_print_msg + ("I am the child, my PID is %d (expected %d)\n" , +getpid (), set_tid[0 ]); + +if (set_tid[0 ] !=getpid ()) +child_exit (EXIT_FAILURE ); +child_exit (EXIT_SUCCESS ); + } + +ksft_print_msg ("I am the parent (%d). My child's pid is %d\n" , +getpid (), pid); + +if (waitpid (pid, &status,0 ) <0 ) { +ksft_print_msg ("Child returned %s\n" ,strerror (errno)); +return -errno; + } + +if (!WIFEXITED (status)) +return -1 ; + +return WEXITSTATUS (status); +} + +static int test_clone3_set_tid (pid_t * set_tid, +size_t set_tid_size,int expected) +{ +int ret; + +ksft_print_msg ("[%d] Trying clone3() with CLONE_SET_TID to %d\n" , +getpid (), set_tid[0 ]); + ret =call_clone3_set_tid (set_tid, set_tid_size); + +ksft_print_msg + ("[%d] clone3() with CLONE_SET_TID %d says :%d - expected %d\n" , +getpid (), set_tid[0 ], ret, expected); +if (ret != expected) { +ksft_test_result_fail + ("[%d] Result (%d) is different than expected (%d)\n" , +getpid (), ret, expected); +return -1 ; + } +ksft_test_result_pass + ("[%d] Result (%d) matches expectation (%d)\n" ,getpid (), ret, + expected); + +return 0 ; +} + +struct libcap { +struct __user_cap_header_struct hdr; +struct __user_cap_data_struct data[2 ]; +}; + +static int set_capability () +{ +cap_value_t cap_values[] = {CAP_SETUID ,CAP_SETGID }; +struct libcap *cap; +int ret =-1 ; +cap_t caps; + + caps =cap_get_proc (); +if (!caps) { +perror ("cap_get_proc" ); +return -1 ; + } + +/* Drop all capabilities */ +if (cap_clear (caps)) { +perror ("cap_clear" ); + goto out; + } + +cap_set_flag (caps,CAP_EFFECTIVE ,2 , cap_values,CAP_SET ); +cap_set_flag (caps,CAP_PERMITTED ,2 , cap_values,CAP_SET ); + + cap = (struct libcap *) caps; + +/* 40 -> CAP_CHECKPOINT_RESTORE */ + cap->data[1 ].effective |=1 << (40 -32 ); + cap->data[1 ].permitted |=1 << (40 -32 ); + +if (cap_set_proc (caps)) { +perror ("cap_set_proc" ); + goto out; + } + ret =0 ; +out: +if (cap_free (caps)) +perror ("cap_free" ); +return ret; +} + +int main (int argc,char *argv[]) +{ +pid_t pid; +int status; +int ret =0 ; +pid_t set_tid[1 ]; +uid_t uid =getuid (); + +ksft_print_header (); +test_clone3_supported (); +ksft_set_plan (2 ); + +if (uid !=0 ) { + ksft_cnt.ksft_xskip = ksft_plan; +ksft_print_msg ("Skipping all tests as non-root\n" ); +return ksft_exit_pass (); + } + +memset (&set_tid,0 ,sizeof (set_tid)); + +/* Find the current active PID */ + pid =fork (); +if (pid ==0 ) { +ksft_print_msg ("Child has PID %d\n" ,getpid ()); +child_exit (EXIT_SUCCESS ); + } +if (waitpid (pid, &status,0 ) <0 ) +ksft_exit_fail_msg ("Waiting for child %d failed" , pid); + +/* After the child has finished, its PID should be free. */ + set_tid[0 ] = pid; + +if (set_capability ()) +ksft_test_result_fail + ("Could not set CAP_CHECKPOINT_RESTORE\n" ); +prctl (PR_SET_KEEPCAPS ,1 ,0 ,0 ,0 ); +/* This would fail without CAP_CHECKPOINT_RESTORE */ +setgid (1000 ); +setuid (1000 ); + set_tid[0 ] = pid; + ret |=test_clone3_set_tid (set_tid,1 , -EPERM ); +if (set_capability ()) +ksft_test_result_fail + ("Could not set CAP_CHECKPOINT_RESTORE\n" ); +/* This should work as we have CAP_CHECKPOINT_RESTORE as non-root */ + ret |=test_clone3_set_tid (set_tid,1 ,0 ); + +return !ret ?ksft_exit_pass () :ksft_exit_fail (); +}Signature · 1 line
2.26.2
RE: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
In reply to Paul MooreQuoted 14 lines
Quoted 14 lines
This is scary. But I believe it is safe.
Reviewed-by Serge Hallyn <[email protected]>I am a bit curious about the implications of the selinux patch. IIUC you are using the permission of the tracing process to execute the file without transition, so this is a way to work around the policy which might prevent the tracee from doing so. Given that SELinux wants to be MAC, I'm not *quite* sure that's considered kosher. You also are skipping the
PROCESS__PTRACEtoSECCLASS_PROCESScheck whichselinux_bprm_set_credsdoes later on. Again I'm just not quite sure what's considered normal there these days.
Quoted 3 lines
Quoted 1 line
Paul, do you have input there?
I agree, the SELinux hook looks wrong. Building on what Christian said, this looks more like a ptrace operation than an exec operation.
Serge, Paul, Christian,
I made a PoC to demonstrate the change of /proc/self/exe without CAP_SYS_ADMIN using only ptrace and execve. You may find it here: https://github.com/nviennot/run_as_exe
What do you recommend to relax the security checks in the kernel when it comes to changing the exe link?
Nico
Re: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
Quoted 16 lines · Nicolas Viennot
Quoted 16 lines
Quoted 15 lines
This is scary. But I believe it is safe.
Reviewed-by Serge Hallyn <[email protected]>I am a bit curious about the implications of the selinux patch. IIUC you are using the permission of the tracing process to execute the file without transition, so this is a way to work around the policy which might prevent the tracee from doing so. Given that SELinux wants to be MAC, I'm not *quite* sure that's considered kosher. You also are skipping the
PROCESS__PTRACEtoSECCLASS_PROCESScheck whichselinux_bprm_set_credsdoes later on. Again I'm just not quite sure what's considered normal there these days.Paul, do you have input there?
Quoted 8 lines
Quoted 1 line
I agree, the SELinux hook looks wrong. Building on what Christian said, this looks more like a ptrace operation than an exec operation.
Serge, Paul, Christian,
I made a PoC to demonstrate the change of
/proc/self/exewithoutCAP_SYS_ADMINusing only ptrace and execve. You may find it here: https://github.com/nviennot/run_as_exeWhat do you recommend to relax the security checks in the kernel when it comes to changing the exe link?
Looks fun! Yeah, so that this is possible is known afaict. But you're not really circumventing the kernel check but are mucking with the EFL by changing the auxv, right?
Originally, you needed to be userns root, i.e. only uid 0 could change the /proc/self/exe link (cf. [1]). This was changed to ns_capable(CAP_SYS_ADMIN) in [2].
The original reasoning in [1] is interesting as it basically already points to your poc:
"Still note that updating exe-file link now doesn't require sys-resource
capability anymore, after all there is no much profit in preventing setup own file link (there are a number of ways to execute own code -- ptrace, ld-preload, so that the only reliable way to find which exactly code is executed is to inspect running program memory). Still we require the caller to be at least user-namespace root user."
There were arguments being made that /proc/<pid>/exe needs to be sm that userspace can have a decent amount of trust in but I believe that that's not a great argument.
But let me dig a little into the original discussion and see what the thread-model was.
At this point I'm starting to believe that it was people being cautios but better be sure.
[1]: f606b77f1a9e ("prctl: PR_SET_MM -- introduce PR_SET_MM_MAP operation") [2]: 4d28df6152aa ("prctl: Allow local CAP_SYS_ADMIN changing exe_file") [3]: https://lore.kernel.org/patchwork/patch/697304/
Christian
Re: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
Quoted 44 lines · Christian Brauner
Quoted 16 lines · Nicolas Viennot
Quoted 16 lines
Quoted 15 lines
This is scary. But I believe it is safe.
Reviewed-by Serge Hallyn <[email protected]>I am a bit curious about the implications of the selinux patch. IIUC you are using the permission of the tracing process to execute the file without transition, so this is a way to work around the policy which might prevent the tracee from doing so. Given that SELinux wants to be MAC, I'm not *quite* sure that's considered kosher. You also are skipping the
PROCESS__PTRACEtoSECCLASS_PROCESScheck whichselinux_bprm_set_credsdoes later on. Again I'm just not quite sure what's considered normal there these days.Paul, do you have input there?
Quoted 8 lines
Quoted 1 line
I agree, the SELinux hook looks wrong. Building on what Christian said, this looks more like a ptrace operation than an exec operation.
Serge, Paul, Christian,
I made a PoC to demonstrate the change of
/proc/self/exewithoutCAP_SYS_ADMINusing only ptrace and execve. You may find it here: https://github.com/nviennot/run_as_exeWhat do you recommend to relax the security checks in the kernel when it comes to changing the exe link?
Looks fun! Yeah, so that this is possible is known afaict. But you're not really circumventing the kernel check but are mucking with the EFL by changing the auxv, right?
Originally, you needed to be userns root, i.e. only uid 0 could change the
/proc/self/exelink (cf. [1]). This was changed tons_capable(CAP_SYS_ADMIN)in [2].The original reasoning in [1] is interesting as it basically already points to your poc:
"Still note that updating exe-file link now doesn't require sys-resource
capability anymore, after all there is no much profit in preventing setup own file link (there are a number of ways to execute own code -- ptrace, ld-preload, so that the only reliable way to find which exactly code is executed is to inspect running program memory). Still we require the caller to be at least user-namespace root user."
Quoted 8 lines
There were arguments being made that /proc/<pid>/exe needs to be sm that userspace can have a decent amount of trust in but I believe that that's not a great argument.
But let me dig a little into the original discussion and see what the thread-model was.
At this point I'm starting to believe that it was people being cautios but better be sure.
Ok, so the original patch proposal was presented in [4] in 2014. The final version of that patch added the PR_SET_MM_MAP we know today. The initial version presented in [4] did not require _any_ privilege.
So the reasoning for only placing the /proc/<pid>/exe link under ns_capable(CAP_SYS_ADMIN) is very thin. to quote from [5]:
"Controlling exe_fd without privileges may turn out to be dangerous. At
least things like tomoyo examine it for making policy decisions (see tomoyo_manager())."
So yes, tomoyo_get_exe() is what this was retained for apparently:
const char *tomoyo_get_exe(void)
{
struct file *exe_file;
const char *cp;
struct mm_struct *mm = current->mm;
if (!mm)
return NULL;
exe_file = get_mm_exe_file(mm);
if (!exe_file)
return NULL;
cp = tomoyo_realpath_from_path(&exe_file->f_path);
fput(exe_file);
return cp;
}The exe path is literally used in tomoyo_manager() to verify that you are allowed to change policy. That seems like a bad idea to me but then again, I don't know enough about Tomoyo. In any case, I think that means we can't remove CAP_SYS_ADMIN because that would make things worse than they are right now for Tomoyo but I also don't see why placing this under ns_capable(CAP_SYS_ADMIN) || ns_capable(CAP_CHECKPOINT_RESTORE) would make this any worse.
And Cyrill (and later in that thread Andrei) already mentioned it in [6]:
"@exe_fd is just a hint and as I mentioned if we have ptrace/preload
rights there damn a lot of ways to inject own code into any program so that a user won't even notice ;)"
Another place where the exe file is relevant is for the coredump with the -E option. But it only uses the path when generating the coredump pattern and if that's a security issue than your poc shows that this can already be achieved today.
Christian
Quoted 4 lines
[1]:
f606b77f1a9e("prctl:PR_SET_MM-- introducePR_SET_MM_MAPoperation") [2]:4d28df6152aa("prctl: Allow localCAP_SYS_ADMINchangingexe_file") [3]: https://lore.kernel.org/patchwork/patch/697304/
[4]: https://lore.kernel.org/lkml/[email protected]/
[5]: https://lore.kernel.org/lkml/CAGXu5jL3exT4j+8rjMv1O54uJWQ5UHL69Z-24b61rhXROqZamQ@mail.gmail.com/
[6]: https://lore.kernel.org/lkml/20140722203614.GF838@moon/
Re: [PATCH v4 3/3] prctl: Allow ptrace capable processes to change /proc/self/exe
On Tue, Jul 07, 2020 at 05:45:04PM +0200, Christian Brauner wrote:
...
Quoted 5 lines
Ok, so the original patch proposal was presented in [4] in 2014. The final version of that patch added the
PR_SET_MM_MAPwe know today. The initial version presented in [4] did not require _any_ privilege.
True. I still think that relyng on /proc/<pid>/exe being immutable (or guarded by caps) in a sake of security is a bit misleading, this link only a hint without any guarantees of what code is being executed once we pass cs:rip to userspace right after exec is completed. Nowadays I rather think we might need to call audit_log() here or something similar to point that exe link is changed (by criu or someone else) and simply notify node's administrator, that's all. But as you pointed tomoyo may be affected if we simply drops all caps from here. Thus I agree that the new cap won't make situation worse.
Still I'm not in touch with kernel code for a couple of years already and might be missing something obvious here.