syscalls: add openat2()

Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

configure.ac

@@ -622,7 +622,8 @@ AC_CHECK_HEADER([ifaddrs.h],
 AC_HEADER_MAJOR
 
 # Check for some syscalls functions
-AC_CHECK_FUNCS([setns pivot_root sethostname unshare rand_r confstr faccessat gettid memfd_create move_mount open_tree execveat clone3 fsopen fspick fsconfig fsmount])
+AC_CHECK_FUNCS([setns pivot_root sethostname unshare rand_r confstr faccessat gettid memfd_create move_mount open_tree execveat clone3 fsopen fspick fsconfig fsmount, openat2])
+AC_CHECK_TYPES([struct open_how], [], [], [[#include <linux/openat2.h>]])
 AC_CHECK_TYPES([struct clone_args], [], [], [[#include <linux/sched.h>]])
 AC_CHECK_MEMBERS([struct clone_args.set_tid],[],[],[[#include <linux/sched.h>]])
 AC_CHECK_MEMBERS([struct clone_args.cgroup],[],[],[[#include <linux/sched.h>]])

src/lxc/conf.c

@@ -1077,7 +1077,7 @@ static int mount_autodev(const char *name, const struct lxc_rootfs *rootfs,
 		goto reset_umask;
 	}
 
-	ret = safe_mount("none", path, "tmpfs", 0, mount_options,
+	ret = safe_mount("none", "dev", "tmpfs", 0, mount_options,
 			 rootfs->path ? rootfs->mount : NULL );
 	if (ret < 0) {
 		SYSERROR("Failed to mount tmpfs on \"%s\"", path);

src/lxc/syscall_numbers.h

@@ -640,4 +640,24 @@
 	#endif
 #endif
 
+#ifndef __NR_openat2
+	#if defined __alpha__
+		#define __NR_openat2 547
+	#elif defined _MIPS_SIM
+		#if _MIPS_SIM == _MIPS_SIM_ABI32	/* o32 */
+			#define __NR_openat2 4437
+		#endif
+		#if _MIPS_SIM == _MIPS_SIM_NABI32	/* n32 */
+			#define __NR_openat2 6437
+		#endif
+		#if _MIPS_SIM == _MIPS_SIM_ABI64	/* n64 */
+			#define __NR_openat2 5437
+		#endif
+	#elif defined __ia64__
+		#define __NR_openat2 (437 + 1024)
+	#else
+		#define __NR_openat2 437
+	#endif
+#endif
+
 #endif /* __LXC_SYSCALL_NUMBERS_H */

src/lxc/syscall_wrappers.h

@@ -16,6 +16,7 @@
 #include <unistd.h>
 
 #include "config.h"
+#include "macro.h"
 #include "syscall_numbers.h"
 
 #ifdef HAVE_LINUX_MEMFD_H
@@ -26,6 +27,10 @@
 #include <sys/signalfd.h>
 #endif
 
+#ifdef HAVE_STRUCT_OPEN_HOW
+#include <linux/openat2.h>
+#endif
+
 typedef int32_t key_serial_t;
 
 #if !HAVE_KEYCTL
@@ -203,4 +208,61 @@ static inline int fsmount_lxc(int fs_fd, unsigned int flags, unsigned int attr_f
 extern int fsmount(int fs_fd, unsigned int flags, unsigned int attr_flags);
 #endif
 
+/*
+ * Arguments for how openat2(2) should open the target path. If only @flags and
+ * @mode are non-zero, then openat2(2) operates very similarly to openat(2).
+ *
+ * However, unlike openat(2), unknown or invalid bits in @flags result in
+ * -EINVAL rather than being silently ignored. @mode must be zero unless one of
+ * {O_CREAT, O_TMPFILE} are set.
+ *
+ * @flags: O_* flags.
+ * @mode: O_CREAT/O_TMPFILE file mode.
+ * @resolve: RESOLVE_* flags.
+ */
+struct lxc_open_how {
+	__u64 flags;
+	__u64 mode;
+	__u64 resolve;
+};
+
+/* how->resolve flags for openat2(2). */
+#ifndef RESOLVE_NO_XDEV
+#define RESOLVE_NO_XDEV		0x01 /* Block mount-point crossings
+					(includes bind-mounts). */
+#endif
+
+#ifndef RESOLVE_NO_MAGICLINKS
+#define RESOLVE_NO_MAGICLINKS	0x02 /* Block traversal through procfs-style
+					"magic-links". */
+#endif
+
+#ifndef RESOLVE_NO_SYMLINKS
+#define RESOLVE_NO_SYMLINKS	0x04 /* Block traversal through all symlinks
+					(implies OEXT_NO_MAGICLINKS) */
+#endif
+
+#ifndef RESOLVE_BENEATH
+#define RESOLVE_BENEATH		0x08 /* Block "lexical" trickery like
+					"..", symlinks, and absolute
+					paths which escape the dirfd. */
+#endif
+
+#ifndef RESOLVE_IN_ROOT
+#define RESOLVE_IN_ROOT		0x10 /* Make all jumps to "/" and ".."
+					be scoped inside the dirfd
+					(similar to chroot(2)). */
+#endif
+
+#ifndef HAVE_OPENAT2
+static inline int openat2(int dfd, const char *filename, struct lxc_open_how *how, size_t size)
+{
+	/* When struct open_how is updated we should update lxc as well. */
+#ifdef HAVE_STRUCT_OPEN_HOW
+	BUILD_BUG_ON(sizeof(struct lxc_open_how) != sizeof(struct open_how));
+#endif
+	return syscall(__NR_openat2, dfd, filename, (struct open_how *)how, size);
+}
+#endif /* HAVE_OPENAT2 */
+
 #endif /* __LXC_SYSCALL_WRAPPER_H */