Filesystems with subvolumes still take exclusive ownership of the block device when mounted. If the mount command is later called for another subvolume, it's possible that the do_mount_by_pattern() loop will attempt a different filesystem first, which will fail immediately with EBUSY. I ran into this while working with an APFS driver, using util-linux 2.32.1. The problem is gone in more recent releases because mount can now recognize APFS by the superblock, but allow EBUSY anyway in case that fails. Signed-off-by: Ernesto A. Fernández <ernesto@xxxxxxxxxxxxx> --- libmount/src/context_mount.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/libmount/src/context_mount.c b/libmount/src/context_mount.c index 55ebf7945..69dc97465 100644 --- a/libmount/src/context_mount.c +++ b/libmount/src/context_mount.c @@ -957,7 +957,8 @@ static int do_mount_by_pattern(struct libmnt_context *cxt, const char *pattern) if (is_success_status(cxt)) break; if (mnt_context_get_syscall_errno(cxt) != EINVAL && - mnt_context_get_syscall_errno(cxt) != ENODEV) + mnt_context_get_syscall_errno(cxt) != ENODEV && + mnt_context_get_syscall_errno(cxt) != EBUSY) break; } mnt_free_filesystems(filesystems); -- 2.27.0
