bug-parted
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Why parted shouldn't write to random fields in disk labels


From: Andrew Clausen
Subject: Re: Why parted shouldn't write to random fields in disk labels
Date: Mon, 19 Jul 2004 01:46:45 +1000
User-agent: Mutt/1.3.28i

On Sat, Jul 17, 2004 at 07:45:29PM -0700, John Gilmore wrote:
> I've seen this problem discussed on this list before, where installing
> Fedora Core 2 makes a separate Windows partition stop booting because
> the CHS fields in the MBR got clobbered:
> 
>   http://lwn.net/Articles/86835/
> 
> The proposal to "fix" this involves writing some kind of random
> numbers into every MBR so we or the kernel can tell one disk from
> another, as I understand it.

Our proposal, not lwm's... correct.  However, I think the suggestion
was only for new partition tables created by Parted.

> Solving a problem that was caused by writing a field that we never use
> and didn't need to write (in a structure shared by all OS's that
> access the disk),

It isn't in a structure.  It's free space that the boot loader is free
to use.  If you create a new partition table, there is no boot loader
yet.  Boot loaders are free to overwrite, but are unlikely to.

> by writing and then using a different field that we
> never before needed to read or write, seems to me likely to compound,
> rather than fix, the existing problem.

We wouldn't be reading the field in particular... just the entire block.
It would make it more likely that Linux can confidently identify the
hard disks' BIOS mappings.

I would like to emphasize: write, not overwrite.

Cheers,
Andrew





reply via email to

[Prev in Thread] Current Thread [Next in Thread]