NanoBSD on Soekris Net4801

blueh2o · 04-14-2015, 10:10 AM

I'm trying to build nanobsd on a CF card for a Soekris Net4801. I started out with the howto at https://www.freebsd.org/doc/en_US.IS...bsd/howto.html. I can get it to build, but when I dd the image to the CF card and try to boot it, It says it can't find /boot/loader. I've tried about a million other things, and all I can get is either that message or booting just hangs with no output. What am I doing wrong?

designator · 04-14-2015, 12:51 PM

Can you boot anything else? Empty GRUB or something?

blueh2o · 04-14-2015, 01:05 PM

I can boot FreeBSD from an IDE disk.

designator · 04-14-2015, 01:28 PM

Can you grab one of prepared images and try to boot from that? It sounds more like CF card boot issue than a problem with your image, but this would confirm my guess.

pfSense maybe?
https://www.pfsense.org/download/ind...tion=downloads

blueh2o · 04-15-2015, 08:15 AM

Yes, that works...

Code:

FreeBSD/x86 bootstrap loader, Revision 1.1
(root@pfs22-i386-builder, Fri Mar 13 08:35:23 CDT 2015)
Loading /boot/defaults/loader.conf
/boot/kernel/kernel text=0x11f7fd9 data=0x832cf0+0x279e60 syms=[0x4+0xf36c0+0x4+0x16b89c]

Hit [Enter] to boot immediately, or any other key for command prompt.
Booting [/boot/kernel/kernel]...
KDB: debugger backends: ddb
KDB: current backend: ddb
Copyright (c) 1992-2014 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
        The Regents of the University of California. All rights reserved.
FreeBSD is a registered trademark of The FreeBSD Foundation.
FreeBSD 10.1-RELEASE-p6 #0 b69ba8f(releng/10.1)-dirty: Fri Mar 13 08:45:06 CDT 2015

blueh2o · 04-15-2015, 01:57 PM

Rebuilt with CPU_SOEKRIS and CPU_GEODE and now getting past bootloader.. Next problem is..

Code:

(ada0:ata0:0:0:0): READ_DMA. ACB: c8 00 01 00 00 40 00 00 00 00 01 00
(ada0:ata0:0:0:0): CAM status: Command timeout
(ada0:ata0:0:0:0): Retrying command
(ada0:ata0:0:0:0): READ_DMA. ACB: c8 00 01 00 00 40 00 00 00 00 01 00
(ada0:ata0:0:0:0): CAM status: Command timeout
(ada0:ata0:0:0:0): Retrying command
(ada0:ata0:0:0:0): READ_DMA. ACB: c8 00 01 00 00 40 00 00 00 00 01 00
(ada0:ata0:0:0:0): CAM status: Command timeout
(ada0:ata0:0:0:0): Retrying command

Guess disabling DMA is my next step.

designator · 04-15-2015, 07:29 PM

I don't have a net4801 to play with, but a quick search turned this up:
http://lists.soekris.com/pipermail/s...ry/004675.html

blueh2o · 04-16-2015, 09:00 AM

I don't think the 4801 can do DMA on the CF-IDE slot. Anyway, I added "hw.ata.ata_dma=0" to the nano loader.conf on da0s1a and got slightly further.. Now it wants manual root filesystem specification on boot, and it doesn't seem to like anything I put in. The fstab on da0s1a is empty.. I'm confused. Also, I thought nanobsd was supposed to create 2 identical slices. Why then do I have da0s1a, and da0s2 but no da0s2a?

blueh2o · 04-16-2015, 10:21 AM

Guess I corrupted the filesystem somehow.. Reloaded the card, now I have the fstab... Changed the device names from ad0 to ada0, getting a tiny bit further..

Code:

ada0: 16.700MB/s transfers (PIO4, PIO 512bytesuhub0: 3 ports with 3 removable, self powered
)
ada0: 1953MB (4001760 512 byte sectors: 16H 63S/T 3970C)
ada0: Previously was known as ad0
random: unblocking device.
Trying to mount root from ufs:/dev/ada0s1a [ro]...
exec /sbin/init: error 20
exec /sbin/oinit: error 20
exec /sbin/init.bak: error 20
init: not found in path /sbin/init:/sbin/oinit:/sbin/init.bak:/rescue/init
panic: no init

blueh2o · 04-16-2015, 10:29 AM

Checking what's in /sbin on the CF card, finding this:

Code:

ls: bin: Bad file descriptor
ls: media: Bad file descriptor
ls: sbin: Bad file descriptor
ls: usr: Bad file descriptor

tried fsck:

Code:

root@NANOBUILD:~ # fsck /dev/da0s1a
fsck: Could not determine filesystem type
root@NANOBUILD:~ # fsck -t ufs /dev/da0s1a
** /dev/da0s1a
BAD SUPER BLOCK: VALUES IN SUPER BLOCK DISAGREE WITH THOSE IN FIRST ALTERNATE
ioctl (GCINFO): Inappropriate ioctl for device
fsck_ufs: /dev/da0s1a: can't read disk label

Why is this happening?

blueh2o · 04-16-2015, 11:29 AM

Reloaded the card after doing sysctl kern.geom.debugflags=16. sbin now exists. Boots up with:

Code:

ada0: 1953MB (4001760 512 byte sectors: 16H 63S/T 3970C)
ada0: Previously was known as ad0
random: unblocking device.
Trying to mount root from ufs:/dev/ada0s1a [ro]...
Setting hostuuid: 63aea72b-e433-11e4-9f10-000024c2bddc.
Setting hostid: 0x66fac3de.
Entropy harvesting: interrupts ethernet point_to_point swi.
Starting file system checks:
Unknown error; help!
ERROR: ABORTING BOOT (sending SIGTERM to parent)!
Apr 16 12:23:30 init: /bin/sh on /etc/rc terminated abnormally, going to single user mode
Enter full pathname of shell or RETURN for /bin/sh:
# mount
/dev/ada0s1a on / (ufs, local, read-only)
devfs on /dev (devfs, local, multilabel)
/dev/md0 on /etc (ufs, local)
/dev/md1 on /var (ufs, local)
# fsck
** /dev/ada0s1a
BAD SUPER BLOCK: VALUES IN SUPER BLOCK DISAGREE WITH THOSE IN FIRST ALTERNATE
ioctl (GCINFO): Inappropriate ioctl for device
fsck_ufs: /dev/ada0s1a: can't read disk label

How is it nanobsd is giving me a bad superblock right from the start? This shouldn't really be this difficult, should it?
I can now get this to boot if I skip checking filesystems. Perhaps someone has some hints why I'm seeing this superblock issue.