[glib/wip/oholy/bind-mounts: 47/47] gunixmounts: Mark mounts as system internal instead of filtering out



commit 5634540d9d8ddfeb4cf7dd09977919b483c69739
Author: Ondrej Holy <oholy redhat com>
Date:   Thu Apr 26 16:22:36 2018 +0200

    gunixmounts: Mark mounts as system internal instead of filtering out
    
    mntent-based implementation filter out mounts with device path that was
    repeated. Consequently, it is not possible to show such mounts in UI even
    with x-gvfs-show, because they are not returned from g_unix_mounts_get.
    libmount-based implementation currently doesn't filter out any mounts
    which causes issues to our volume monitors. Let's rather mark mounts
    which don't point into fs root as system_internal. This approach won't be
    affected by mount order as is mntent-based implementation. It will mark
    more mounts as system_internal than it is filtered out with mntend-based
    implementation, but there will be always possibility to show them in UI
    over x-gvfs-show, which was not possible with mntend-based. We can
    probably introduce some improvements later to not mark unique mounts as
    system internal even if they don't point into fs root...
    
    https://gitlab.gnome.org/GNOME/glib/issues/1271

 gio/gunixmounts.c | 33 +++++++++++++++++++++++++++++----
 1 file changed, 29 insertions(+), 4 deletions(-)
---
diff --git a/gio/gunixmounts.c b/gio/gunixmounts.c
index 642ed3347..c993e4f20 100644
--- a/gio/gunixmounts.c
+++ b/gio/gunixmounts.c
@@ -389,8 +389,9 @@ g_unix_is_system_device_path (const char *device_path)
 
 static gboolean
 guess_system_internal (const char *mountpoint,
-                      const char *fs,
-                      const char *device)
+                       const char *fs,
+                       const char *device,
+                       const char *root)
 {
   if (g_unix_is_system_fs_type (fs))
     return TRUE;
@@ -400,7 +401,29 @@ guess_system_internal (const char *mountpoint,
 
   if (g_unix_is_mount_path_system_internal (mountpoint))
     return TRUE;
-  
+
+  /* It is not possible to reliably detect mounts which were created by bind
+   * operation. mntent-based _g_get_unix_mounts() implementation blindly skips
+   * mounts with a device path that is repeated (e.g. mounts created by bind
+   * operation, btrfs subvolumes). This usually chooses the most important
+   * mounts (i.e. which points to the root of filesystem), but it doesn't work
+   * in all cases and also it is not ideal that those mounts are completely
+   * ignored (e.g. x-gvfs-show doesn't work for them, trash backend can't handle
+   * files on btrfs subvolumes). libmount-based _g_get_unix_mounts()
+   * implementation provides a root path. So there is no need to completely
+   * ignore those mounts, because e.g. our volume monitors can use the root path
+   * to not mengle those mounts with the "regular" mounts (i.e. which points to
+   * the root). But because those mounts usually just duplicate other mounts and
+   * are completely ignored with mntend-based implementation, let's mark them as
+   * system internal. Given the different approches it doesn't mean that all
+   * mounts which were ignored will be system internal now, but this should work
+   * in most cases. For more info, see g_unix_mount_get_root_path() annotation,
+   * comment in mntent-based _g_get_unix_mounts() implementation and the
+   * https://gitlab.gnome.org/GNOME/glib/issues/1271 issue.
+   */
+  if (root != NULL && g_strcmp0 (root, "/") != 0)
+    return TRUE;
+
   return FALSE;
 }
 
@@ -427,7 +450,9 @@ create_unix_mount_entry (const char *device_path,
   mount_entry->is_system_internal =
     guess_system_internal (mount_entry->mount_path,
                            mount_entry->filesystem_type,
-                           mount_entry->device_path);
+                           mount_entry->device_path,
+                           mount_entry->root_path);
+
   return mount_entry;
 }
 


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]