Re: kern/48474: disk enumeration at boot is off

[John Nemeth <jnemeth%cue.bc.ca@localhost>]

On Dec 23,  6:40pm, htodd%twofifty.com@localhost wrote:
}
} >Number:         48474
} >Synopsis:       wd1 at boot is wd2 in running kernel
} >Arrival-Date:   Mon Dec 23 18:40:00 +0000 2013
} >Originator:     Hisashi Todd Fujinaka
} >Release:        NetBSD 6.99.28
} >Organization:
} None
} >Environment:
}       
} System: NetBSD mara.i8u.org 6.99.28 NetBSD 6.99.28 (GENERIC) #0: Sun Dec 22 
14:23:09 PST 2013 
htodd%mara.i8u.org@localhost:/usr/obj/amd64-s/sys/arch/amd64/compile/GENERIC 
amd64
} Architecture: x86_64
} Machine: amd64
} >Description:
} On boot, I see disks wd0 and wd1. From dmesg:
} wd0 at atabus3 drive 0
} wd0: <WDC WD5000AAKS-00TMA0>
} wd0: drive supports 16-sector PIO transfers, LBA48 addressing
} wd0: 465 GB, 969021 cyl, 16 head, 63 sec, 512 bytes/sect x 976773168 sectors
} wd0: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133)
} wd0(ahcisata1:0:0): using PIO mode 4, DMA mode 2, Ultra-DMA mode 6 
(Ultra/133) (using DMA)
} wd1 at atabus4 drive 0
} wd1: <ST3250620AS>
} wd1: drive supports 16-sector PIO transfers, LBA48 addressing
} wd1: 232 GB, 484521 cyl, 16 head, 63 sec, 512 bytes/sect x 488397168 sectors
} wd1: drive supports PIO mode 4, DMA mode 2, Ultra-DMA mode 6 (Ultra/133)
} wd1(ahcisata1:1:0): using PIO mode 4, DMA mode 2, Ultra-DMA mode 6 
(Ultra/133) (using DMA)
} boot device: wd0
} root on wd0a dumps on wd0b
} 
} But on the running kernel, trying to access wd1a gives errors:
} :~ > sudo mount /dev/wd1a /mnt
} mount: cannot open `/dev/wd1a': Device not configured
} Exit 1
} 
} Turns out that the running kernel thinks it's wd2. This is causing problems 
booting from the "second" disk as well.
} /dev/wd2a: file system not clean (fs_clean=0x10); please fsck(8)
} /dev/wd2a: lost blocks 0 files 0
} 
}       
} >How-To-Repeat:
} The bootblocks might be wrong; I used NetBSD-6 first and then

     The bootblocks has nothing to do with it.

} upgraded this disk using build.sh.

     You need to be more clear about exactly what you did as this
should not happen.

     I.e. exactly what version of NetBSD and for what architecture
was your starting point; how did you invoke build.sh; did you run
postinstall or etcupdate afterwards; etc.

     This sounds an awful lot like going from i386 to amd64, or at
least using the wrong MAKEDEV script.  The device minor numbers
for the two platforms are incompatible, and this problem is what
you will see if you use i386 devices with an amd64 kernel.

} But basically, install two disks and reboot the system with NetBSD-6 or 
NetBSD-current.
}       
} >Fix:
} Yes, please.
}       
}-- End of excerpt from htodd%twofifty.com@localhost