Re: [patch] NetBSD disklabel support for geom_bsd

Hello!

On Fri, 17 Mar 2006, Jos Backus wrote:
>>   My patch modifies disklabel only for NetBSD partition (type 169),
>> and only in memory, not on disk. So it definitely won't affect
>> moving any partition using a non-FreeBSD aware tools.
>
> Thanks for clarifying that. I guess I phrased this poorly. I meant to say that
> if FreeBSD were to move to the NetBSD model, that would seem  to be one of the
> downsides as it must be for NetBSD today. When moving a NetBSD partition, the
> label needs to be updated whereas with a FreeBSD partition it does not.

  Well, as I wrote in my previuos letter, it's my mistake that FreeBSD
uses slice-relative offsets while NetBSD uses absolute offsets. Actually,
both systems use absolute offsets, and offset conversion isn't required
(but just accidently happens harmless).

  Actually, I can tell what has created my confusion. Did anyone analyze
how disklabel offsets look under RELENG_4? It's very interesting:

# fdisk ad4
******* Working on device /dev/ad4 *******
parameters extracted from in-core disklabel are:
cylinders=788 heads=255 sectors/track=63 (16065 blks/cyl)

parameters to be used for BIOS calculations are:
cylinders=788 heads=255 sectors/track=63 (16065 blks/cyl)

Media sector size is 512
Warning: BIOS sector numbering starts with sector 1
Information from DOS bootblock is:
The data for partition 1 is:
sysid 165,(FreeBSD/NetBSD/386BSD)
     start 63, size 10474317 (5114 Meg), flag 0
 	beg: cyl 0/ head 1/ sector 1;
 	end: cyl 651/ head 254/ sector 63
The data for partition 2 is:
sysid 165,(FreeBSD/NetBSD/386BSD)
     start 10474380, size 2184840 (1066 Meg), flag 80 (active)
 	beg: cyl 652/ head 0/ sector 1;
 	end: cyl 787/ head 254/ sector 63
The data for partition 3 is:
<UNUSED>
The data for partition 4 is:
<UNUSED>

So 1st slice starts at usual offset 63.

# dd bs=512 if=/dev/ad4s1 iseek=1 | hd | more
00000000  57 45 56 82 05 00 00 00  61 64 34 73 31 00 00 00  |WEV.....ad4s1...|
00000010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000020  00 00 00 00 00 00 00 00  00 02 00 00 3f 00 00 00  |............?...|
00000030  ff 00 00 00 8b 02 00 00  c1 3e 00 00 4d d3 9f 00  |.........>..M...|
00000040  00 00 00 00 00 00 00 00  10 0e 01 00 00 00 00 00  |................|
00000050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000080  00 00 00 00 57 45 56 82  31 f4 08 00 00 20 00 00  |....WEV.1.... ..|
00000090  00 20 00 00 00 00 08 00  00 00 00 00 00 00 00 00  |. ..............|
000000a0  07 00 00 00 00 00 10 00  00 00 08 00 00 00 00 00  |................|
000000b0  01 00 00 00 4d d3 9f 00  00 00 00 00 00 00 00 00  |....M...........|
000000c0  00 00 00 00 4d d3 87 00  00 00 18 00 00 00 00 00  |....M...........|

Partition table starts at offset 0x94, and media offset (starting sector)
goes at offset 4 in every entry. As you can see, offset of 'a' partition
is zero (bytes 98-9b of dump). So here it IS slice-relative.

But now dump this sector using base unit (ad4) entry instead of slice (ad4s1):

# dd bs=512 if=/dev/ad4 iseek=64 | hd | more
00000000  57 45 56 82 05 00 00 00  61 64 34 73 31 00 00 00  |WEV.....ad4s1...|
00000010  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
00000020  00 00 00 00 00 00 00 00  00 02 00 00 3f 00 00 00  |............?...|
00000030  ff 00 00 00 14 03 00 00  c1 3e 00 00 c2 5d c1 00  |.........>...]..|
00000040  00 00 00 00 00 00 00 00  10 0e 01 00 00 00 00 00  |................|
00000050  00 00 00 00 00 00 00 00  00 00 00 00 00 00 00 00  |................|
*
00000080  00 00 00 00 57 45 56 82  7f 7b 08 00 00 20 00 00  |....WEV..{... ..|
00000090  00 20 00 00 00 00 08 00  3f 00 00 00 00 00 00 00  |. ......?.......|
000000a0  07 00 00 00 00 00 10 00  3f 00 08 00 00 00 00 00  |........?.......|
000000b0  01 00 00 00 4d d3 9f 00  3f 00 00 00 00 00 00 00  |....M...?.......|
000000c0  00 00 00 00 4d d3 87 00  3f 00 18 00 00 00 00 00  |....M...?.......|

As you can see, now offsets are absolute. This behaviour (label 
auto-conversion "on the fly") has created so much confusion for me.
There is no such conversion in modern (e.g., RELENG_6) FreeBSD, I just see 
absolute labels in both cases (and I like it).

Sincerely, Dmitry
-- 
Atlantis ISP, System Administrator
e-mail:  dmitry_at_atlantis.dp.ua
nic-hdl: LYNX-RIPE