Red Hat Bugzilla – Bug 138480
[PATCH] GPT PMBR disk >2TB fix
Last modified: 2007-11-30 17:07:14 EST
Description of problem:
On Mon, 2004-11-08 at 10:45 -0600, Matt Domsch wrote:
(patch submitted to firstname.lastname@example.org for inclusion after 1.6.16)
Patch below (applies against 1.6.11 and 1.6.16 at least) sets the
SizeInLBA field in the Protective MBR on GPT disks to be 0xFFFFFFFF if
the disk size is really larger than that. This is to be consistent
with the EFI Specification.
Necessary to prevent overflowing this field on really large disks.
The 2.6.10 Linux kernel partitioning code will start checking
this value for sanity, and 0xFFFFFFFF is a special value there now.
Note, though no single disk can be larger than 2TB yet, a RAID
controller could project such, likewise the new partitionable md_d0pX
software RAID volumes could be larger than 2TB and want to put a GPT
partition table on that. (yes, I've verified that this works).
Version-Release number of selected component (if applicable):
parted <= 1.6.16-0
Created attachment 106345 [details]
Patch as submitted to email@example.com
Patch accepted by upstream maintainer Andrew Clausen, so I expect it
in parted-1.6.17 when released.
merged into the -devel package. Will merge over to the RHEL4 branch
once I see how the new geometry probing works out. So leaving open
for now and putting it on a tracker so I don't forget about it.
Refreshing parted package to 1.6.19 to pull in the slew of geometry
related fixes Andrew has added recently and also gets this.