GRUB_IEEE1275_FLAG_NO_PARTITION_0

GRUB_IEEE1275_FLAG_NO_PARTITION_0

[Robert Millan]

What's the point of GRUB_IEEE1275_FLAG_NO_PARTITION_0 ?  We have code that
checks this flag, but nobody initialises it:

  disk/ieee1275/ofdisk.c:  if (! grub_ieee1275_test_flag (GRUB_IEEE1275_FLAG_NO_PARTITION_0))
  include/grub/ieee1275/ieee1275.h:  GRUB_IEEE1275_FLAG_NO_PARTITION_0,

-- 
Robert Millan

My spam trap is honeypot@aybabtu.com.  Note: this address is only intended
for spam harvesters.  Writing to it will get you added to my black list.

Re: GRUB_IEEE1275_FLAG_NO_PARTITION_0

[Hollis Blanchard]

On Mon, 2007-07-09 at 17:05 +0200, Robert Millan wrote:
> What's the point of GRUB_IEEE1275_FLAG_NO_PARTITION_0 ?  We have code that
> checks this flag, but nobody initialises it:
> 
>   disk/ieee1275/ofdisk.c:  if (! grub_ieee1275_test_flag (GRUB_IEEE1275_FLAG_NO_PARTITION_0))
>   include/grub/ieee1275/ieee1275.h:  GRUB_IEEE1275_FLAG_NO_PARTITION_0,

In IEEE1275, partition 0 (as in "disk:0") means "the whole disk".
However, CodeGen-based firmware (that means Genesi) have a bug where
partition 0 is actually disk partition 1. In that case, to access the
whole disk, GRUB must open "disk" instead of "disk:0".

I have no idea how the initialization got lost; it should be initialized
in the same place that the other flags are. Checking cvs history may be
in order.

-Hollis

Re: GRUB_IEEE1275_FLAG_NO_PARTITION_0

[Robert Millan]

On Tue, Jul 10, 2007 at 03:26:43PM -0500, Hollis Blanchard wrote:
> On Mon, 2007-07-09 at 17:05 +0200, Robert Millan wrote:
> > What's the point of GRUB_IEEE1275_FLAG_NO_PARTITION_0 ?  We have code that
> > checks this flag, but nobody initialises it:
> > 
> >   disk/ieee1275/ofdisk.c:  if (! grub_ieee1275_test_flag (GRUB_IEEE1275_FLAG_NO_PARTITION_0))
> >   include/grub/ieee1275/ieee1275.h:  GRUB_IEEE1275_FLAG_NO_PARTITION_0,
> 
> In IEEE1275, partition 0 (as in "disk:0") means "the whole disk".
> However, CodeGen-based firmware (that means Genesi) have a bug where
> partition 0 is actually disk partition 1. In that case, to access the
> whole disk, GRUB must open "disk" instead of "disk:0".

Yes, but don't we already have GRUB_IEEE1275_FLAG_0_BASED_PARTITIONS for that ?

> I have no idea how the initialization got lost; it should be initialized
> in the same place that the other flags are. Checking cvs history may be
> in order.

It was this commit:

2005-04-30  Hollis Blanchard  <hollis@penguinppc.org>

	* boot/powerpc/ieee1275/cmain.c: [...]
	[...]. Set GRUB_IEEE1275_FLAG_REAL_MODE and
	GRUB_IEEE1275_FLAG_0_BASED_PARTITIONS.

doesn't mention it, but it also added GRUB_IEEE1275_FLAG_NO_PARTITION_0.

Should we merge these flags?  They appear to mean the same to me.

-- 
Robert Millan

My spam trap is honeypot@aybabtu.com.  Note: this address is only intended
for spam harvesters.  Writing to it will get you added to my black list.

Re: GRUB_IEEE1275_FLAG_NO_PARTITION_0

[Robert Millan]

On Sat, Jul 14, 2007 at 12:31:02AM +0200, Robert Millan wrote:
> On Tue, Jul 10, 2007 at 03:26:43PM -0500, Hollis Blanchard wrote:
> > On Mon, 2007-07-09 at 17:05 +0200, Robert Millan wrote:
> > > What's the point of GRUB_IEEE1275_FLAG_NO_PARTITION_0 ?  We have code that
> > > checks this flag, but nobody initialises it:
> > > 
> > >   disk/ieee1275/ofdisk.c:  if (! grub_ieee1275_test_flag (GRUB_IEEE1275_FLAG_NO_PARTITION_0))
> > >   include/grub/ieee1275/ieee1275.h:  GRUB_IEEE1275_FLAG_NO_PARTITION_0,
> > 
> > In IEEE1275, partition 0 (as in "disk:0") means "the whole disk".
> > However, CodeGen-based firmware (that means Genesi) have a bug where
> > partition 0 is actually disk partition 1. In that case, to access the
> > whole disk, GRUB must open "disk" instead of "disk:0".
> 
> Yes, but don't we already have GRUB_IEEE1275_FLAG_0_BASED_PARTITIONS for that ?

Ah, I understand now.  GRUB_IEEE1275_FLAG_0_BASED_PARTITIONS implies
GRUB_IEEE1275_FLAG_NO_PARTITION_0 since partition 0 is the first partition
and cannot represent whole disk.  However, GRUB_IEEE1275_FLAG_NO_PARTITION_0
doesn't imply GRUB_IEEE1275_FLAG_0_BASED_PARTITIONS.

I think it would be clearer if GRUB_IEEE1275_FLAG_NO_PARTITION_0 was renamed to
GRUB_IEEE1275_FLAG_PARTITION_0_IS_NOT_WHOLE_DISK.  What do you think?

-- 
Robert Millan

My spam trap is honeypot@aybabtu.com.  Note: this address is only intended
for spam harvesters.  Writing to it will get you added to my black list.