[Parted-maintainers] Bug#380226: NTFS (partition) not recreated correctly after resize:incorrect start sector

[Ben Hutchings]

reassign 380226 partman-base
thanks

This is not a bug in libparted; parted_server is simply not specifying
the correct constraint for the resize operation.

This patch might fix parted_server, but I don't know how to test it.

diff -u parted_server.c~ parted_server.
--- parted_server.c~	2007-01-11 13:57:27.000000000 +0000
+++ parted_server.c	2007-02-27 23:50:18.000000000 +0000
@@ -712,8 +712,6 @@
                         ped_file_system_close(fs);
                         fs = NULL;
                 }
-                if (NULL == fs && NULL != ped_file_system_probe(&(part->geom)))
-                        return false;
         } else {
                 fs = NULL;
         }
@@ -727,10 +725,50 @@
         log("successfully checked");
         if (part->type & PED_PARTITION_LOGICAL)
                 maximize_extended_partition(disk);
-        if (NULL != fs)
+        if (NULL != fs) {
                 constraint = ped_file_system_get_resize_constraint(fs);
-        else
-                constraint = ped_constraint_any(disk->dev);
+        } else {
+                PedFileSystemType *fs_type;
+                PedGeometry *fs_geom;
+                PedAlignment start_align;
+                PedGeometry full_dev;
+                fs_type = ped_file_system_probe(&(part->geom));
+                log("probed file system: %s", NULL != fs_type ? "yes" : "no");
+                if (NULL != fs_type) {
+                        fs_geom = ped_file_system_probe_specific(fs_type,
+                                                                 &part->geom);
+                } else {
+                        fs_geom = NULL;
+                }
+                if (NULL != fs_geom) {
+                        /* We cannot resize or move the fs but we can move the
+                         * end of the partition so long as it contains the
+                         * whole fs.
+                         */
+                        if (ped_alignment_init(&start_align,
+                                                fs_geom->start, 0)
+                            && ped_geometry_init(&full_dev, disk->dev,
+                                                 0, disk->dev->length - 1)) {
+                                constraint = ped_constraint_new(
+                                        &start_align,
+                                        ped_alignment_any,
+                                        &full_dev, &full_dev,
+                                        fs_geom->length,
+                                        disk->dev->length);
+                        } else {
+                                constraint = NULL;
+                        }
+                        ped_geometry_destroy(fs_geom);
+                } else {
+                        constraint = NULL;
+                }
+        }
+        if (open_filesystem && NULL == constraint) {
+                log("failed to get resize constraint");
+                if (NULL != fs)
+                        ped_file_system_close(fs);
+                return false;
+        }
         if (!ped_disk_set_partition_geom(disk, part, constraint, start, end))
                 result = false;
         else if (NULL == fs)
-- END --

I'm assuming that open_filesystem will be set when resizing the NTFS
partition.  This should be the case if it's not a virtual partition
(what does that mean?).

Ben.

-- 
Ben Hutchings
Power corrupts.  Absolute power is kind of neat.
                           - John Lehman, Secretary of the US Navy 1981-1987
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
Url : http://lists.alioth.debian.org/pipermail/parted-maintainers/attachments/20070228/b93a7a6e/attachment.pgp