kerndat: check for statmount with STATMOUNT_BY_FD

For supporting files on "unmounted" mounts (mounts umounted with
MNT_DETACH), we need support for statmount on fds. Add kerndat check to
check support for statmount on fd.

Signed-off-by: Bhavik Sachdev <b.sachdev1904@gmail.com>

criu/cr-check.c

@@ -1400,6 +1400,17 @@ static int check_timer_cr_ids(void)
 	return 0;
 }
 
+static int check_statmount_fd(void)
+{
+	if (!kdat.has_statmount_fd) {
+		pr_warn("statmount syscall with STATMOUNT_BY_FD is unavailable,"
+				" files on unmounted mounts will not be supported\n");
+		return -1;
+	}
+
+	return 0;
+}
+
 /* musl doesn't have a statx wrapper... */
 struct staty {
 	__u32 stx_dev_major;
@@ -1731,6 +1742,7 @@ int cr_check(void)
 		ret |= check_overlayfs_maps();
 		ret |= check_timer_cr_ids();
 		ret |= check_pagemap_scan_guard_pages();
+		ret |= check_statmount_fd();
 
 		if (kdat.lsm == LSMTYPE__APPARMOR)
 			ret |= check_apparmor_stacking();

criu/include/kerndat.h

@@ -93,6 +93,7 @@ struct kerndat_s {
 	bool has_breakpoints;
 	bool has_madv_guard;
 	bool has_pagemap_scan_guard_pages;
+	bool has_statmount_fd;
 };
 
 extern struct kerndat_s kdat;

criu/include/mount-v2.h

@@ -52,6 +52,118 @@ static inline int sys_open_tree(int dfd, const char *filename, unsigned int flag
 #define RESOLVE_NO_XDEV 0x01 /* Block mount-point crossings (includes bind-mounts). */
 #endif
 
+#ifndef STATMOUNT_SB_BASIC
+#define STATMOUNT_SB_BASIC		0x00000001U     /* Want/got sb_... */
+#endif
+
+#ifndef STATMOUNT_MNT_BASIC
+#define STATMOUNT_MNT_BASIC		0x00000002U	/* Want/got mnt_... */
+#endif
+
+#ifndef STATMOUNT_PROPAGATE_FROM
+#define STATMOUNT_PROPAGATE_FROM	0x00000004U	/* Want/got propagate_from */
+#endif
+
+#ifndef STATMOUNT_MNT_ROOT
+#define STATMOUNT_MNT_ROOT		0x00000008U	/* Want/got mnt_root  */
+#endif
+
+#ifndef STATMOUNT_MNT_POINT
+#define STATMOUNT_MNT_POINT		0x00000010U	/* Want/got mnt_point */
+#endif
+
+#ifndef STATMOUNT_FS_TYPE
+#define STATMOUNT_FS_TYPE		0x00000020U	/* Want/got fs_type */
+#endif
+
+#ifndef STATMOUNT_MNT_NS_ID
+#define STATMOUNT_MNT_NS_ID		0x00000040U	/* Want/got mnt_ns_id */
+#endif
+
+#ifndef STATMOUNT_MNT_OPTS
+#define STATMOUNT_MNT_OPTS		0x00000080U	/* Want/got mnt_opts */
+#endif
+
+#ifndef STATMOUNT_FS_SUBTYPE
+#define STATMOUNT_FS_SUBTYPE		0x00000100U	/* Want/got fs_subtype */
+#endif
+
+#ifndef STATMOUNT_SB_SOURCE
+#define STATMOUNT_SB_SOURCE		0x00000200U	/* Want/got sb_source */
+#endif
+
+#ifndef STATMOUNT_OPT_ARRAY
+#define STATMOUNT_OPT_ARRAY		0x00000400U	/* Want/got opt_... */
+#endif
+
+#ifndef STATMOUNT_OPT_SEC_ARRAY
+#define STATMOUNT_OPT_SEC_ARRAY		0x00000800U	/* Want/got opt_sec... */
+#endif
+
+#ifndef STATMOUNT_SUPPORTED_MASK
+#define STATMOUNT_SUPPORTED_MASK	0x00001000U	/* Want/got supported mask flags */
+#endif
+
+#ifndef STATMOUNT_MNT_UIDMAP
+#define STATMOUNT_MNT_UIDMAP		0x00002000U	/* Want/got uidmap... */
+#endif
+
+#ifndef STATMOUNT_MNT_GIDMAP
+#define STATMOUNT_MNT_GIDMAP		0x00004000U	/* Want/got gidmap... */
+#endif
+
+#ifndef STATMOUNT_BY_FD
+#define STATMOUNT_BY_FD		0x0000001U /* want mountinfo for given fd */
+#endif
+
+#ifndef MNT_ID_REQ_SIZE_VER1
+#define MNT_ID_REQ_SIZE_VER1	32 /* sizeof second published struct */
+#endif
+
+struct mnt_id_req {
+	__u32 size;
+	__u32 fd;
+	__u64 mnt_id;
+	__u64 param;
+	__u64 mnt_ns_id;
+};
+
+struct statmount {
+	__u32 size;		/* Total size, including strings */
+	__u32 mnt_opts;		/* [str] Options (comma separated, escaped) */
+	__u64 mask;		/* What results were written */
+	__u32 sb_dev_major;	/* Device ID */
+	__u32 sb_dev_minor;
+	__u64 sb_magic;		/* ..._SUPER_MAGIC */
+	__u32 sb_flags;		/* SB_{RDONLY,SYNCHRONOUS,DIRSYNC,LAZYTIME} */
+	__u32 fs_type;		/* [str] Filesystem type */
+	__u64 mnt_id;		/* Unique ID of mount */
+	__u64 mnt_parent_id;	/* Unique ID of parent (for root == mnt_id) */
+	__u32 mnt_id_old;	/* Reused IDs used in proc/.../mountinfo */
+	__u32 mnt_parent_id_old;
+	__u64 mnt_attr;		/* MOUNT_ATTR_... */
+	__u64 mnt_propagation;	/* MS_{SHARED,SLAVE,PRIVATE,UNBINDABLE} */
+	__u64 mnt_peer_group;	/* ID of shared peer group */
+	__u64 mnt_master;	/* Mount receives propagation from this ID */
+	__u64 propagate_from;	/* Propagation from in current namespace */
+	__u32 mnt_root;		/* [str] Root of mount relative to root of fs */
+	__u32 mnt_point;	/* [str] Mountpoint relative to current root */
+	__u64 mnt_ns_id;	/* ID of the mount namespace */
+	__u32 fs_subtype;	/* [str] Subtype of fs_type (if any) */
+	__u32 sb_source;	/* [str] Source string of the mount */
+	__u32 opt_num;		/* Number of fs options */
+	__u32 opt_array;	/* [str] Array of nul terminated fs options */
+	__u32 opt_sec_num;	/* Number of security options */
+	__u32 opt_sec_array;	/* [str] Array of nul terminated security options */
+	__u64 supported_mask;	/* Mask flags that this kernel supports */
+	__u32 mnt_uidmap_num;	/* Number of uid mappings */
+	__u32 mnt_uidmap;	/* [str] Array of uid mappings (as seen from callers namespace) */
+	__u32 mnt_gidmap_num;	/* Number of gid mappings */
+	__u32 mnt_gidmap;	/* [str] Array of gid mappings (as seen from callers namespace) */
+	__u64 __spare2[43];
+	char str[];		/* Variable size part containing strings */
+};
+
 static inline long sys_openat2(int dirfd, const char *pathname, struct open_how *how, size_t size)
 {
 	return syscall(__NR_openat2, dirfd, pathname, how, size);

criu/include/util.h

@@ -67,6 +67,7 @@ extern int set_proc_fd(int fd);
 
 extern pid_t sys_clone_unified(unsigned long flags, void *child_stack, void *parent_tid, void *child_tid,
 			       unsigned long newtls);
+struct statmount *do_statmount_fd(int fd, u64 mask);
 
 /*
  * Values for pid argument of the proc opening routines below.

criu/kerndat.c

@@ -1,3 +1,4 @@
+#include <linux/stat.h>
 #include <unistd.h>
 #include <fcntl.h>
 #include <stdio.h>
@@ -1741,6 +1742,18 @@ static int kerndat_has_timer_cr_ids(void)
 	return 0;
 }
 
+static int kerndat_has_statmount_fd(void)
+{
+	struct statmount *statmnt = do_statmount_fd(STDIN_FILENO, STATMOUNT_MNT_BASIC);
+	if (!statmnt && (errno == ENOSYS || errno == EINVAL)) {
+		pr_info("statmount with STATMOUNT_BY_FD flag isn't supported\n");
+		kdat.has_statmount_fd = false;
+	} else {
+		kdat.has_statmount_fd = true;
+	}
+	return 0;
+}
+
 static void breakpoint_func(void)
 {
 	if (raise(SIGSTOP))
@@ -2125,6 +2138,10 @@ int kerndat_init(void)
 		pr_err("kerndat_has_madv_guard has failed when initializing kerndat.\n");
 		ret = -1;
 	}
+	if (!ret && kerndat_has_statmount_fd()) {
+		pr_err("kerndat_has_statmount_fd failed when initializing kerndat.\n");
+		ret = -1;
+	}
 
 	kerndat_lsm();
 	kerndat_mmap_min_addr();

criu/util.c

@@ -1425,6 +1425,48 @@ static int epoll_hangup_event(int epollfd, struct epoll_rfd *rfd)
 	return ret;
 }
 
+static int __statmount(struct mnt_id_req *req, struct statmount *stmnt,
+		       size_t bufsize, int flags)
+{
+	return syscall(__NR_statmount, req, stmnt, bufsize, flags);
+}
+
+struct statmount *do_statmount_fd(int fd, u64 mask)
+{
+	size_t bufsize = 1 << 15;
+	struct statmount *stmnt = NULL, *tmp = NULL;
+	int ret;
+
+	struct mnt_id_req req = {
+		.size	= MNT_ID_REQ_SIZE_VER1,
+		.fd	= fd,
+		.param	= mask
+	};
+
+	for (;;) {
+		tmp = xrealloc(stmnt, bufsize);
+		if (!tmp)
+			goto out;
+
+		stmnt = tmp;
+		ret = __statmount(&req, stmnt, bufsize, STATMOUNT_BY_FD);
+		if (!ret) {
+			return stmnt;
+		}
+
+		if (errno != EOVERFLOW)
+			goto out;
+
+		bufsize <<= 1;
+		if (bufsize >= UINT_MAX / 2)
+			goto out;
+	}
+
+out:
+	free(stmnt);
+	return NULL;
+}
+
 int epoll_run_rfds(int epollfd, struct epoll_event *evs, int nr_fds, int timeout)
 {
 	int ret, i, nr_events;