mount: support the sizelimit and offset option for loop devices

Starting with linux kernel v5.4 squashfs has a more strict parameter
checking implemented. Unlike util-linux mount, busybox never supported
the sizelimit option but simply forwards it to the kernel.
Since v5.4 mounting will fail with

    squashfs: Unknown parameter 'sizelimit'

Support the sizelimit parameter by setting it in the LOOP_SET_STATUS64
structure before handing it to the kernel.

While at it also add support for the offset option, which currently will
always be set to 0.

function                                             old     new   delta
cut_out_ull_opt                                        -     167    +167
singlemount                                         1230    1266     +36
set_loop                                             834     862     +28
losetup_main                                         479     483      +4
------------------------------------------------------------------------------
(add/remove: 1/0 grow/shrink: 3/0 up/down: 235/0)             Total: 235 bytes

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Denys Vlasenko <vda.linux@googlemail.com>

2 files changed, 62 insertions, 2 deletions

diff --git a/util-linux/losetup.c b/util-linux/losetup.c
index ac8b795..24f7a23 100644
--- a/util-linux/losetup.c
+++ b/util-linux/losetup.c
@@ -150,7 +150,7 @@ int losetup_main(int argc UNUSED_PARAM, char **argv)
 			if (opt & OPT_P) {
 				flags |= BB_LO_FLAGS_PARTSCAN;
 			}
-			if (set_loop(&d, argv[0], offset, flags) < 0)
+			if (set_loop(&d, argv[0], offset, 0, flags) < 0)
 				bb_simple_perror_msg_and_die(argv[0]);
 			return EXIT_SUCCESS;
 		}
diff --git a/util-linux/mount.c b/util-linux/mount.c
index b92e2c2..19ac139 100644
--- a/util-linux/mount.c
+++ b/util-linux/mount.c
@@ -1886,6 +1886,58 @@ static int nfsmount(struct mntent *mp, unsigned long vfsflags, char *filteropts)
 
 #endif // !ENABLE_FEATURE_MOUNT_NFS
 
+// Find "...,NAME=NUM,..." in the option string, remove "NAME=NUM" option
+// and return NUM.
+// Return 0 if not found.
+// All instances must be parsed and removed (for example, since kernel 5.4
+//      squashfs: Unknown parameter 'sizelimit'
+// will result if loopback mount option "sizelimit=NNN" is not removed
+// and squashfs sees it in option string).
+static unsigned long long cut_out_ull_opt(char *opts, const char *name_eq)
+{
+	unsigned long long ret = 0;
+
+	if (!opts) // allow NULL opts (simplifies callers' work)
+		return ret;
+
+	for (;;) {
+		char *end;
+		char *opt;
+
+		// Find comma-delimited "NAME="
+		for (;;) {
+			opt = strstr(opts, name_eq);
+			if (!opt)
+				return ret;
+			if (opt == opts)
+				break; // found it (it's first opt)
+			if (opt[-1] == ',') {
+				opts = opt - 1;
+				break; // found it (it's not a first opt)
+			}
+			// False positive like "VNAME=", we are at "N".
+			// - skip it, loop back to searching
+			opts = opt + 1;
+		}
+
+		ret = bb_strtoull(opt + strlen(name_eq), &end, 0);
+		if (errno && errno != EINVAL) {
+ err:
+			bb_error_msg_and_die("bad option '%s'", opt);
+		}
+		if (*end == '\0') {
+			// It is "[,]NAME=NUM\0" - truncate it and return
+			*opts = '\0';
+			return ret;
+		}
+		if (*end != ',')
+			goto err;
+		// We are at trailing comma
+		// Remove "NAME=NUM," and loop back to check for duplicate opts
+		overlapping_strcpy(opt, end + 1);
+	}
+}
+
 // Mount one directory.  Handles CIFS, NFS, loopback, autobind, and filesystem
 // type detection.  Returns 0 for success, nonzero for failure.
 // NB: mp->xxx fields may be trashed on exit
@@ -2029,9 +2081,16 @@ static int singlemount(struct mntent *mp, int ignore_busy)
 	) {
 		// Do we need to allocate a loopback device for it?
 		if (ENABLE_FEATURE_MOUNT_LOOP && S_ISREG(st.st_mode)) {
+			unsigned long long offset;
+			unsigned long long sizelimit;
+
 			loopFile = bb_simplify_path(mp->mnt_fsname);
 			mp->mnt_fsname = NULL; // will receive malloced loop dev name
 
+			// Parse and remove loopback options
+			offset = cut_out_ull_opt(filteropts, "offset=");
+			sizelimit = cut_out_ull_opt(filteropts, "sizelimit=");
+
 			// mount always creates AUTOCLEARed loopdevs, so that umounting
 			// drops them without any code in the userspace.
 			// This happens since circa linux-2.6.25:
@@ -2040,7 +2099,8 @@ static int singlemount(struct mntent *mp, int ignore_busy)
 			// Subject: Allow auto-destruction of loop devices
 			loopfd = set_loop(&mp->mnt_fsname,
 					loopFile,
-					0,
+					offset,
+					sizelimit,
 					((vfsflags & MS_RDONLY) ? BB_LO_FLAGS_READ_ONLY : 0)
 						| BB_LO_FLAGS_AUTOCLEAR
 			);