lsblk: use devno to check if the filesystem is mounted

The device (for example LVM logical volume) could be renamed and then
the device name from /proc/self/mountinfo does not match with reality.
So, we also need to check devno. Unfortunately we cannot completely
rely on devno, because for example btrfs uses psudo device numbers.

References: https://bugzilla.redhat.com/show_bug.cgi?id=980463
Signed-off-by: Karel Zak <kzak@redhat.com>

diff --git a/libmount/src/tab.c b/libmount/src/tab.c
index 5ac252ce1994ff898e869b59dcd03961e56c2d4f..20028a641844dcaebea094d2f5a65d348b2b187c 100644 (file)
--- a/libmount/src/tab.c
+++ b/libmount/src/tab.c
@@ -1117,6 +1117,7 @@ int mnt_table_is_fs_mounted(struct libmnt_table *tb, struct libmnt_fs *fstab_fs)
        const char *src = NULL, *tgt = NULL;
        char *xtgt = NULL;
        int rc = 0;
+       dev_t devno = 0;
 
        assert(tb);
        assert(fstab_fs);
@@ -1148,6 +1149,14 @@ int mnt_table_is_fs_mounted(struct libmnt_table *tb, struct libmnt_fs *fstab_fs)
        if (src && tb->cache && !mnt_fs_is_pseudofs(fstab_fs))
                src = mnt_resolve_spec(src, tb->cache);
 
+       if (src && root) {
+               struct stat st;
+
+               devno = mnt_fs_get_devno(fstab_fs);
+               if (!devno && stat(src, &st) == 0 && S_ISBLK(st.st_mode))
+                       devno = st.st_rdev;
+       }
+
        tgt = mnt_fs_get_target(fstab_fs);
 
        if (!tgt || !src) {
@@ -1158,7 +1167,12 @@ int mnt_table_is_fs_mounted(struct libmnt_table *tb, struct libmnt_fs *fstab_fs)
 
        while (mnt_table_next_fs(tb, &itr, &fs) == 0) {
 
-               if (!mnt_fs_streq_srcpath(fs, src)) {
+               int eq = mnt_fs_streq_srcpath(fs, src);
+
+               if (!eq && devno && mnt_fs_get_devno(fs) == devno)
+                       eq = 1;
+
+               if (!eq) {
                        /* The source does not match. Maybe the source is a loop
                         * device backing file.
                         */