Re: entry for boot partition in protective MBR of GPT disklabel?

[K.G.]

On Sat, 12 Nov 2005 10:54:32 -0500 "Ristuccia, Brian" <address@hidden> wrote:

> Any thoughts on whether or not this is a practical idea before I go spending 
> time on it? 
> 
> It looks like the kernel is happy so long as there's at least one GPT (is it 
> type 0xEE?) partition table entry. Are there any other things I should 
> consider?
> 

My only thoughts is that you should limit the number of
"real" (!= GPT) partitions you want in MBR to 2, so you
can put a GPT partition before and after if needed, and
only need primary partitions to do that. You should also
make sure that your bootloader uses LBA because CHS
makes things an incredible mess 50% of the time...

I also think the best way to implement that kind of
stuff is to create and use a flag (see partition_set_flag,
partition_get_flag and partition_is_flag_available
functions in disk_*.c, and also PedPartitionFlag enum
in disk.h) to select the partition(s) you want to put
in the MBR. Then you'll just have to modify the function
called by gpt_write that create the protective MBR.

Good luck :)
Guillaume Knispel



> -----Original Message-----
> From: K.G. [mailto:address@hidden
> Sent: Fri 11/11/2005 6:35 PM
> To:   Ristuccia, Brian
> Cc:   
> Subject:      Re: entry for boot partition in protective MBR of GPT disklabel?
> On Fri, 11 Nov 2005 17:24:55 -0500 Brian Ristuccia <address@hidden> wrote:
> > Is there any way to cause parted to make an entry in the protective MBR
> > for a partition which exists entirely below 0xFFFFFFFF? I'd like to use
> > this feature to make the boot partition at the beginning of my large
> > block device accessible to a bootloader which only understands MSDOS
> > style partition tables.
> 
> There is nothing in the current code to do that...
> 
> Guillaume Knispel