RAID superblock....

RAID superblock....

[Rogier Wolff]

Hi,

The MD code I see doing: 


488         sb_offset = calc_dev_sboffset(rdev->dev, rdev->mddev, 1);
489         rdev->sb_offset = sb_offset;
490         fsync_dev(dev);
491         set_blocksize (dev, MD_SB_BYTES);
492         bh = bread (dev, sb_offset / MD_SB_BLOCKS, MD_SB_BYTES);


where sb_offset is calculated as: 

290         if (blk_size[MAJOR(dev)])
291                 size = blk_size[MAJOR(dev)][MINOR(dev)];

Now, for aguments sake, I have a 4k disk. I'd expect the size to be 4
(1k blocks, according to the comment near the definition of blk_size). 

Thus the "bread" would effectively try to read the block at offset 4k. 

That would be past the end of my mini-disk, right?

I would have expected a "-1" in there somewhere, to get the last block
of the dev, and not the block just past the end of the drive.

Anyway on the old machine, I still cannot find the raid superblock by
hand, but the drives now mount, so the kernel must have been able to
locate them somehow......

The machine is still running 2.4.16 + IDE patches for 48 bit
addressing.

The working machine is an 850MHz PIII w/384Mb RAM, the non-working
machine is an AMD 1800+ MP w/1G RAM (with another one of those
processors sitting idle close by)...

				Roger. 

-- 
** R.E.Wolff@BitWizard.nl ** http://www.BitWizard.nl/ ** +31-15-2137555 **
*-- BitWizard writes Linux device drivers for any device you may have! --*
* There are old pilots, and there are bold pilots. 
* There are also old, bald pilots. 

Re: RAID superblock....

[Neil Brown]

On Sunday March 10, R.E.Wolff@BitWizard.nl wrote:
> 
> Hi,
> 
> The MD code I see doing: 
> 
> 
> 488         sb_offset = calc_dev_sboffset(rdev->dev, rdev->mddev, 1);
> 489         rdev->sb_offset = sb_offset;
> 490         fsync_dev(dev);
> 491         set_blocksize (dev, MD_SB_BYTES);
> 492         bh = bread (dev, sb_offset / MD_SB_BLOCKS, MD_SB_BYTES);
> 
> 
> where sb_offset is calculated as: 
> 
> 290         if (blk_size[MAJOR(dev)])
> 291                 size = blk_size[MAJOR(dev)][MINOR(dev)];

You missed:
	if (persistent)
		size = MD_NEW_SIZE_BLOCKS(size);

where MD_NEW_SIZE_BLOCKS is
#define MD_NEW_SIZE_BLOCKS(x)		((x & ~(MD_RESERVED_BLOCKS - 1)) - MD_RESERVED_BLOCKS)

and there you have your "-1".

> Anyway on the old machine, I still cannot find the raid superblock by
> hand, but the drives now mount, so the kernel must have been able to
> locate them somehow......

The superblock should be located between 64K and 128K from the end of
the device, on a 64K boundary.
> 


NeilBrown

Re: RAID superblock

[Wakko Warner]

> > Hi,
> >  After I create a RAID setup on the drives,The
> > superblock will be generated at the end of the drives.
> > If I move these drives to other linux system, will
> > this
> >  system recognise the RAID setup without reconfiguring
> > the Linux ?
> 
> If the CHS / LBA settings are the same, and the kernel is the same : Yes.

While this subject is fresh, what would be wrong with using the entire drive
as opposed to creating a partition and adding the partition to the raid?

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: RAID superblock

[Peter Samuelson]

[Wakko Warner]
> While this subject is fresh, what would be wrong with using the
> entire drive as opposed to creating a partition and adding the
> partition to the raid?

Does it autodetect an entire drive?  The autodetect logic for
partitions looks at the 'partition type' byte, which of course doesn't
exist for a whole drive.

Just a thought .. I don't run RAID here.

Peter
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: RAID superblock

[Igmar Palsenberg]

On Sun, 29 Oct 2000, Peter Samuelson wrote:

> [Wakko Warner]
> > While this subject is fresh, what would be wrong with using the
> > entire drive as opposed to creating a partition and adding the
> > partition to the raid?
> 
> Does it autodetect an entire drive?  The autodetect logic for
> partitions looks at the 'partition type' byte, which of course doesn't
> exist for a whole drive.
> 
> Just a thought .. I don't run RAID here.

A good one. I seriously doubt that it indeed will detect drives. The're
not partitions, the're drives.

Don't think the current RAID code handles entire drives. 



		Igmar

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/

Re: RAID superblock

[Wakko Warner]

> > [Wakko Warner]
> > > While this subject is fresh, what would be wrong with using the
> > > entire drive as opposed to creating a partition and adding the
> > > partition to the raid?
> > 
> > Does it autodetect an entire drive?  The autodetect logic for
> > partitions looks at the 'partition type' byte, which of course doesn't
> > exist for a whole drive.

Actually, I don't think it does.  I've not booted into single user mode
(where the raid hasn'tbeen setup yet) to see.

> > Just a thought .. I don't run RAID here.
> 
> A good one. I seriously doubt that it indeed will detect drives. The're
> not partitions, the're drives.
> 
> Don't think the current RAID code handles entire drives. 

Autodetect, probably not.  But it doesn't seem to have any problems with it
as far as usability.

-- 
 Lab tests show that use of micro$oft causes cancer in lab animals
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
Please read the FAQ at http://www.tux.org/lkml/