diff options
| author | Jim Meyering <meyering@redhat.com> | 2012-02-26 18:04:37 +0100 |
|---|---|---|
| committer | Jim Meyering <meyering@redhat.com> | 2012-02-26 18:04:37 +0100 |
| commit | 4559c510f02a691d90329fd646e8eefe2e73e6d2 (patch) | |
| tree | 3c8e3b3de09db0e5b11872992faa4cfc7f7a0054 /NEWS | |
| parent | 187e7cba48094505f00532698dfb04ab40575231 (diff) | |
| download | parted-4559c510f02a691d90329fd646e8eefe2e73e6d2.tar.gz | |
libparted: avoid unwarranted failed assertion during mklabel
This could happen when running "parted -s $dev mklabel $type" for
nearly any $type, assuming the underlying disk's sector size is
larger than 512 bytes. This particular failure was noticed by
running tests/t9050-partition-table-types.sh with a simulated
sector size of 1024, as is done as part of a top-level "make check",
but so far, this failure has shown up only on RHEL-6.2.
That test attempts to exercise an ugly part of partition table
creation: during creation, parted checks for an existing
partition table by iterating through its list of known types and
attempting to read the disk, looking for a preexisting table.
Knowing this, the test takes all pairs of partition table types,
<A,B> and first creates a type of type A, and then creates on top
of that (erasing the original) a table of type B.
* libparted/labels/mac.c (_disk_analyse_ghost_size): Remove bogus
assertion. If the number of the sector with a matching signature
is too large, simply reject it by returning 0.
* NEWS (Bug fixes): Mention this.
Diffstat (limited to 'NEWS')
| -rw-r--r-- | NEWS | 3 |
1 files changed, 3 insertions, 0 deletions
@@ -71,6 +71,9 @@ GNU parted NEWS -*- outline -*- of partitions the device can hold. With this change, parted now correctly informs kernel about new partitions on partitionable MD RAID devices. + libparted: creating a new partition table on a device with an existing + mac partition table could result in a failed assertion. + ** Changes in behavior Floppy drives are no longer scanned on linux: they cannot be partitioned |