1
0
mirror of https://github.com/systemd/systemd.git synced 2024-12-22 17:35:35 +03:00

nspawn: split out seccomp call into nspawn-seccomp.[ch]

Split seccomp into nspawn-seccomp.[ch]. Currently there are no changes,
but this will make it easy in the future to share or use the seccomp logic
from systemd core.
This commit is contained in:
Djalal Harouni 2016-05-26 22:42:29 +02:00
parent 231bfb1b02
commit f011b0b87a
4 changed files with 171 additions and 100 deletions

View File

@ -3016,6 +3016,8 @@ systemd_nspawn_SOURCES = \
src/nspawn/nspawn-expose-ports.h \
src/nspawn/nspawn-cgroup.c \
src/nspawn/nspawn-cgroup.h \
src/nspawn/nspawn-seccomp.c \
src/nspawn/nspawn-seccomp.h \
src/nspawn/nspawn-register.c \
src/nspawn/nspawn-register.h \
src/nspawn/nspawn-setuid.c \

143
src/nspawn/nspawn-seccomp.c Normal file
View File

@ -0,0 +1,143 @@
/***
This file is part of systemd.
Copyright 2016 Lennart Poettering
systemd is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or
(at your option) any later version.
systemd is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with systemd; If not, see <http://www.gnu.org/licenses/>.
***/
#include <errno.h>
#include <linux/netlink.h>
#include <sys/capability.h>
#include <sys/types.h>
#ifdef HAVE_SECCOMP
#include <seccomp.h>
#endif
#include "log.h"
#ifdef HAVE_SECCOMP
#include "seccomp-util.h"
#endif
#include "nspawn-seccomp.h"
#ifdef HAVE_SECCOMP
static int seccomp_add_default_syscall_filter(scmp_filter_ctx ctx,
uint64_t cap_list_retain) {
unsigned i;
int r;
static const struct {
uint64_t capability;
int syscall_num;
} blacklist[] = {
{ CAP_SYS_RAWIO, SCMP_SYS(iopl) },
{ CAP_SYS_RAWIO, SCMP_SYS(ioperm) },
{ CAP_SYS_BOOT, SCMP_SYS(kexec_load) },
{ CAP_SYS_ADMIN, SCMP_SYS(swapon) },
{ CAP_SYS_ADMIN, SCMP_SYS(swapoff) },
{ CAP_SYS_ADMIN, SCMP_SYS(open_by_handle_at) },
{ CAP_SYS_MODULE, SCMP_SYS(init_module) },
{ CAP_SYS_MODULE, SCMP_SYS(finit_module) },
{ CAP_SYS_MODULE, SCMP_SYS(delete_module) },
{ CAP_SYSLOG, SCMP_SYS(syslog) },
};
for (i = 0; i < ELEMENTSOF(blacklist); i++) {
if (cap_list_retain & (1ULL << blacklist[i].capability))
continue;
r = seccomp_rule_add(ctx, SCMP_ACT_ERRNO(EPERM), blacklist[i].syscall_num, 0);
if (r == -EFAULT)
continue; /* unknown syscall */
if (r < 0) {
log_error_errno(r, "Failed to block syscall: %m");
return r;
}
}
return 0;
}
int setup_seccomp(uint64_t cap_list_retain) {
scmp_filter_ctx seccomp;
int r;
seccomp = seccomp_init(SCMP_ACT_ALLOW);
if (!seccomp)
return log_oom();
r = seccomp_add_secondary_archs(seccomp);
if (r < 0) {
log_error_errno(r, "Failed to add secondary archs to seccomp filter: %m");
goto finish;
}
r = seccomp_add_default_syscall_filter(seccomp, cap_list_retain);
if (r < 0)
goto finish;
/*
Audit is broken in containers, much of the userspace audit
hookup will fail if running inside a container. We don't
care and just turn off creation of audit sockets.
This will make socket(AF_NETLINK, *, NETLINK_AUDIT) fail
with EAFNOSUPPORT which audit userspace uses as indication
that audit is disabled in the kernel.
*/
r = seccomp_rule_add(
seccomp,
SCMP_ACT_ERRNO(EAFNOSUPPORT),
SCMP_SYS(socket),
2,
SCMP_A0(SCMP_CMP_EQ, AF_NETLINK),
SCMP_A2(SCMP_CMP_EQ, NETLINK_AUDIT));
if (r < 0) {
log_error_errno(r, "Failed to add audit seccomp rule: %m");
goto finish;
}
r = seccomp_attr_set(seccomp, SCMP_FLTATR_CTL_NNP, 0);
if (r < 0) {
log_error_errno(r, "Failed to unset NO_NEW_PRIVS: %m");
goto finish;
}
r = seccomp_load(seccomp);
if (r == -EINVAL) {
log_debug_errno(r, "Kernel is probably not configured with CONFIG_SECCOMP. Disabling seccomp audit filter: %m");
r = 0;
goto finish;
}
if (r < 0) {
log_error_errno(r, "Failed to install seccomp audit filter: %m");
goto finish;
}
finish:
seccomp_release(seccomp);
return r;
}
#else
int setup_seccomp(uint64_t cap_list_retain) {
return 0;
}
#endif

View File

@ -0,0 +1,24 @@
#pragma once
/***
This file is part of systemd.
Copyright 2016 Lennart Poettering
systemd is free software; you can redistribute it and/or modify it
under the terms of the GNU Lesser General Public License as published by
the Free Software Foundation; either version 2.1 of the License, or
(at your option) any later version.
systemd is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License
along with systemd; If not, see <http://www.gnu.org/licenses/>.
***/
#include <sys/types.h>
int setup_seccomp(uint64_t cap_list_retain);

View File

@ -26,9 +26,6 @@
#include <linux/loop.h>
#include <pwd.h>
#include <sched.h>
#ifdef HAVE_SECCOMP
#include <seccomp.h>
#endif
#ifdef HAVE_SELINUX
#include <selinux/selinux.h>
#endif
@ -82,15 +79,13 @@
#include "nspawn-settings.h"
#include "nspawn-setuid.h"
#include "nspawn-stub-pid1.h"
#include "nspawn-seccomp.h"
#include "parse-util.h"
#include "path-util.h"
#include "process-util.h"
#include "ptyfwd.h"
#include "random-util.h"
#include "rm-rf.h"
#ifdef HAVE_SECCOMP
#include "seccomp-util.h"
#endif
#include "selinux-util.h"
#include "signal-util.h"
#include "socket-util.h"
@ -1667,99 +1662,6 @@ static int reset_audit_loginuid(void) {
return 0;
}
static int setup_seccomp(void) {
#ifdef HAVE_SECCOMP
static const struct {
uint64_t capability;
int syscall_num;
} blacklist[] = {
{ CAP_SYS_RAWIO, SCMP_SYS(iopl) },
{ CAP_SYS_RAWIO, SCMP_SYS(ioperm) },
{ CAP_SYS_BOOT, SCMP_SYS(kexec_load) },
{ CAP_SYS_ADMIN, SCMP_SYS(swapon) },
{ CAP_SYS_ADMIN, SCMP_SYS(swapoff) },
{ CAP_SYS_ADMIN, SCMP_SYS(open_by_handle_at) },
{ CAP_SYS_MODULE, SCMP_SYS(init_module) },
{ CAP_SYS_MODULE, SCMP_SYS(finit_module) },
{ CAP_SYS_MODULE, SCMP_SYS(delete_module) },
{ CAP_SYSLOG, SCMP_SYS(syslog) },
};
scmp_filter_ctx seccomp;
unsigned i;
int r;
seccomp = seccomp_init(SCMP_ACT_ALLOW);
if (!seccomp)
return log_oom();
r = seccomp_add_secondary_archs(seccomp);
if (r < 0) {
log_error_errno(r, "Failed to add secondary archs to seccomp filter: %m");
goto finish;
}
for (i = 0; i < ELEMENTSOF(blacklist); i++) {
if (arg_retain & (1ULL << blacklist[i].capability))
continue;
r = seccomp_rule_add(seccomp, SCMP_ACT_ERRNO(EPERM), blacklist[i].syscall_num, 0);
if (r == -EFAULT)
continue; /* unknown syscall */
if (r < 0) {
log_error_errno(r, "Failed to block syscall: %m");
goto finish;
}
}
/*
Audit is broken in containers, much of the userspace audit
hookup will fail if running inside a container. We don't
care and just turn off creation of audit sockets.
This will make socket(AF_NETLINK, *, NETLINK_AUDIT) fail
with EAFNOSUPPORT which audit userspace uses as indication
that audit is disabled in the kernel.
*/
r = seccomp_rule_add(
seccomp,
SCMP_ACT_ERRNO(EAFNOSUPPORT),
SCMP_SYS(socket),
2,
SCMP_A0(SCMP_CMP_EQ, AF_NETLINK),
SCMP_A2(SCMP_CMP_EQ, NETLINK_AUDIT));
if (r < 0) {
log_error_errno(r, "Failed to add audit seccomp rule: %m");
goto finish;
}
r = seccomp_attr_set(seccomp, SCMP_FLTATR_CTL_NNP, 0);
if (r < 0) {
log_error_errno(r, "Failed to unset NO_NEW_PRIVS: %m");
goto finish;
}
r = seccomp_load(seccomp);
if (r == -EINVAL) {
log_debug_errno(r, "Kernel is probably not configured with CONFIG_SECCOMP. Disabling seccomp audit filter: %m");
r = 0;
goto finish;
}
if (r < 0) {
log_error_errno(r, "Failed to install seccomp audit filter: %m");
goto finish;
}
finish:
seccomp_release(seccomp);
return r;
#else
return 0;
#endif
}
static int setup_propagate(const char *root) {
const char *p, *q;
@ -2988,7 +2890,7 @@ static int outer_child(
if (r < 0)
return r;
r = setup_seccomp();
r = setup_seccomp(arg_retain);
if (r < 0)
return r;