Bug 121266
Summary: | autopartition fails on Mac | ||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Product: | [Fedora] Fedora | Reporter: | Guy Streeter <streeter> | ||||||||||
Component: | parted | Assignee: | Jeremy Katz <katzj> | ||||||||||
Status: | CLOSED RAWHIDE | QA Contact: | |||||||||||
Severity: | medium | Docs Contact: | |||||||||||
Priority: | medium | ||||||||||||
Version: | rawhide | CC: | aldyh, bnocera, byte, clumens, dwmw2, memeyou, nobody+pnasrat, nobody | ||||||||||
Target Milestone: | --- | ||||||||||||
Target Release: | --- | ||||||||||||
Hardware: | powerpc | ||||||||||||
OS: | Linux | ||||||||||||
Whiteboard: | |||||||||||||
Fixed In Version: | 1.6.21-2 | Doc Type: | Bug Fix | ||||||||||
Doc Text: | Story Points: | --- | |||||||||||
Clone Of: | Environment: | ||||||||||||
Last Closed: | 2005-02-21 04:28:50 UTC | Type: | --- | ||||||||||
Regression: | --- | Mount Type: | --- | ||||||||||
Documentation: | --- | CRM: | |||||||||||
Verified Versions: | Category: | --- | |||||||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||||||
Embargoed: | |||||||||||||
Bug Depends On: | |||||||||||||
Bug Blocks: | 121179 | ||||||||||||
Attachments: |
|
Description
Guy Streeter
2004-04-19 20:32:14 UTC
Yes, Mac stuff is only partially implemented atm I tried manually adding an Apple Bootstrap partition. It appeared to let me do that, but the partition doesn't show up in the list. I found that when adding the Apple Bootstrap partition by hand, it would be added at the end of the list (ended up being hda4), even though I create that partition first (should have been hda2, hda1 being the partition table). bug #125699. Checked in a modification to do this onto HEAD. My understanding is that the Apple_Bootstrap partition type is not required to boot. I believe that the Apple_Bootstrap type was arbitrarily chosen because Mac OS X will not try to mount it automatically. In this manner, the partition will remain "hidden" to Mac OS X. colin->nasrat: did this in anaconda HEAD make it into fc3? I think this still doesn't happen with auto partitioning... Created attachment 111217 [details]
Partial fix.
This patch goes part of the way to fixing the problem.
- Fix the check for PMac in getAutopartitionBoot().
- Don't use LVM on PMac (should we do this based on the disk's partition type
instead of the platform)?
- Fix autoCreatePartitionRequests to work again, but do no LVM.
- Attempt to put Apple Bootstrap first, like PReP partitions
- Attempt to make isBootable() true for Apple Bootstrap partitions
The latter doesn't seem to have the desired effect, so the penultimate one
isn't happening either. I'm not sure why.
Other than that, there are a couple of remaining problems.
- When we delete all existing partitions for autopartitioning, anaconda doesn't
realise that /dev/hda1 is actually the partition table itself, in the Mac
scheme. It attempts to create a real partition as /dev/hda1, when in fact the
first partition must be /dev/hda2.
- When we select 'delete only Linux partitions' for autopartitioning, it
doesn't remove the previous Apple Bootstrap partition. Might be nice if we
could detect that it's a yaboot partition.
*** Bug 121384 has been marked as a duplicate of this bug. *** Created attachment 111228 [details]
Fixed patch
This patch should be complete. It does the following:
- Put Apple Bootstrap partitions first, like we do for PReP in fitSized()
- Fix typo in BOOTALPHA_NOT_BSD failure in doPartitioning()
- Don't let doClearPartAction() remove the partition table itself
- Fix autoCreatePartitionRequests() to accept (but ignore) asvol element
- Fix PMac detection in getAutopartitionBoot() and remove separate /boot
- Consider Apple Bootstrap partitions 'native' so they can be replaced
- Use autoCreatePartitionRequests() instead of the LVM version on PMac
- Make get_partition_file_system_type() detect Apple Bootstrap partitions
- Make getBootableRequest() treat Apple Bootstrap as bootable.
- Require Apple Bootstrap on PMac.
This seems to do the right thing in all cases I tested -- both with and without
an existing partition table, with automatic and manual partitioning.
Thanks for the patch commited Created attachment 111244 [details]
Support for lvm (LVM2 or lvm1 with no 0x8E checking)
Liased with Jeremy and have reverted some of the above.
Fix parted to support LVM flags on mac partition tables. I'm doing an install
with a libparted built from this in RHupdates right now and it seems to be
fine.
Caveat - LVM2 only unless partition type (0x8E) checking is not performed on
LVM1.
Linux localhost.localdomain 2.6.10-1.1146_FC4 #1 Sat Feb 19 00:28:29 EST 2005 ppc ppc ppc GNU/Linux /dev/mapper/VolGroup00-LogVol00 on / type ext3 (rw) /dev/hda10 on /boot type ext3 (rw) Disk geometry for /dev/hda: 0.000-9736.945 megabytes Disk label type: mac Minor Start End Filesystem Name Flags 1 0.000 0.031 Apple 2 0.031 0.057 Macintosh 3 0.058 0.093 Macintosh 4 0.094 0.120 Macintosh 5 0.120 0.156 Macintosh 6 0.156 0.253 Macintosh 7 0.254 0.503 Macintosh 8 0.504 0.753 Patch Partition 9 0.754 1.753 hfs untitled boot 10 1.754 101.753 ext3 untitled 11 101.754 9736.945 lvm Information: Don't forget to update /etc/fstab, if necessary. flags should be getting set to include lvm for the partition with a volume name of 'lvm'. Created attachment 111246 [details]
Updated mac lvm patch to fix get_flag
Disk geometry for /dev/hda: 0.000-9736.945 megabytes
Disk label type: mac
Minor Start End Filesystem Name Flags
1 0.000 0.031 Apple
2 0.031 0.057 Macintosh
3 0.058 0.093 Macintosh
4 0.094 0.120 Macintosh
5 0.120 0.156 Macintosh
6 0.156 0.253 Macintosh
7 0.254 0.503 Macintosh
8 0.504 0.753 Patch Partition
9 0.754 1.753 hfs untitled boot
10 1.754 101.753 ext3 untitled
11 101.754 9736.945 lvm lvm
|