Unable to align partition
Gene C.
czar at czarc.net
Sun Mar 28 23:47:43 UTC 2004
On Sunday 28 March 2004 14:49, Alan Cox wrote:
> On Sun, Mar 28, 2004 at 01:14:19PM -0500, Gene C. wrote:
> > > clearly a bug in parted/disk druid as the partition numbers can be (and
> > > should have been) taken from the table itself.
> >
> > Any suggestions on how to get this fixed? ... or at least looked at?
>
> The fix for the general case is "if there is a partition table, validate
> it and if sane believe the geometry of it"
OK, I did some research (searching the kernel archives for references to
parted as well as looking at the bug-parted mailing list archives). The
kernel list archives have a later 2003 discussion of the problem (lots of
works including by Andrew Clausen) but no conclusion ... or perhaps the
conclusion is that each side thinks it is the other side's responsibility.
I do not really care if either LBA or chs partition table parameters are used
so long as:
1. It works.
2. It does not screw up existing partition table definitions.
It is not clear to me that the this is the current situation!
If something is not done to fix this then something needs to be put into
RELEASE-NOTES telling users about the hdx=c,h,s kernel boot parameter with
c,h,s being LBA values (all problems that I have seen are where the disk is
partition with LBA values but then are interpreted as physical c,h,s values
with the 2.6 kernel plus parted.
I also found some reference to an LBA flag described in the parted "manual"
(see the "set" command) but have not idea where this is stored or how (or if)
this information is used.
--
Gene
More information about the fedora-test-list
mailing list