Date: Sat, 26 Sep 1998 00:24:30 -0600 (MDT) From: "Kenneth D. Merry" <ken@plutotech.com> To: freebsd-alpha@FreeBSD.ORG Cc: gibbs@pluto.plutotech.com, imp@pluto.plutotech.com Subject: it works!! Message-ID: <199809260624.AAA09773@panzer.plutotech.com>
next in thread | raw e-mail | index | archive | help
--ELM906791070-9668-0_
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Well, I thought y'all might want to know that I finally got our Miata up
and running. (and how I managed to do it...)
- I couldn't use the NetBSD boot floppy, since 1.3.2 didn't support the
Miata. I know that NetBSD-current supports the Miata, though. I later
found a NetBSD-current snapshot from June, but I'm not sure if the Miata
support was in by then, either.
- I wasn't able to figure out the right "magic" to make a NFS root
partition work. I guess there's a way to do it, but looking through the
source didn't help. :(
- So, I did it the hard way. :)
My configuration is basically two 2 gig Seagate Barracudas on a QLogic ISP
1040, 128MB RAM and a 433MHz 21164. The first disk has Digital UNIX, the
second one was empty.
Here's what I did:
- First, I grabbed a third disk (4 gig narrow barracuda) and hooked it into
the QLogic
- Booted DEC Unix, and attempted to partition the thing. Silly me. No
matter what I tried, their disklabel(8) didn't want to disklabel the
damned disk. It didn't work even when I did something like this:
dd if=/dev/zero of=/dev/rrz20c bs=64k count=10
- I managed to get their disklabel to work by doing this:
dd if=/dev/rz16c of=/dev/rz20c bs=64k count=3
(not sure about the count, but basically I copied the disklabel from the
boot disk, a 2 gig barracuda, to the third disk, a 4 gig barracuda)
- Then I edited the disklabel on the third disk, and got the partitions to
something reasonable.
- Then, I untarred Doug's binary distribution from wcarchive onto the disk,
and put one the kernel with a da1 root onto the disk.
- Next, I halted the machine and disconnected the second 2 gig barracuda.
This is so that the third disk would now be da1 instead of da2. (i.e. so
the kernel with da1 as root would work)
- I turned the machine on, and then netbooted off bootblocks and a kernel
setup on another (i386) machine. (I can netboot the damned thing, but I
can't manage to get an NFS mounted root...oh well...)
- Since I made the partitions under DEC Unix, I had to re-make the device
nodes. So I did something like this:
mount rw:/a/path/to/alpha/root /mnt
mount -u /mnt/dev/da1a /
mount /mnt/dev/da1e /usr
rm -rf /dev
cd /mnt
find dev -print |cpio -pdmuv /
- With the machine booted under a FreeBSD kernel, I grabbed the kernel
source via remote cvs from another machine. I recompiled the kernel to
have root on *da2*, and ftped it to the netboot server.
- I halted the machine, and reconnected the second 2-gig barracuda.
- Netbooted the machine with the kernel configured for root on da2.
- Disklabeled and newfsed the 2-gig barracuda, and copied the filesystems
from the 4-gig barracuda to the 2-gig barracuda. I had to copy
everything using cpio, since dump and restore don't seem to work:
subway# dump -0f - / |restore -rf -
DUMP: Date of this level 0 dump: Fri Sep 25 01:48:54 1998
DUMP: Date of last level 0 dump: the epoch
DUMP: Dumping /dev/rda1a (/) to standard output
DUMP: mapping (Pass I) [regular files]
DUMP: mapping (Pass II) [directories]
DUMP: estimated 26326 tape blocks.
DUMP: dumping (Pass III) [directories]
Checksum error 244733, inode 0 file (null)
Tape is not a dump tape
DUMP: Broken pipe
DUMP: The ENTIRE dump is aborted.
So I copied things like this, more or less:
cd /mnt
find -x / -print |cpio -pdmuv .
- Rebooted the machine, and netbooted it off of the kernel with root on
da1.
- Changed the partitions around in the fstab, since I forgot to do that
before I rebooted on this disk.
- Did a remote cvs checkout of the full source tree
- Changed all the paths in src/sys/boot/alpha/*/Makefile to use a relative
path to libstand, rather than dfr's home directory. :)
- compiled and installed libstand and the boot blocks
- Installed the first and second stage boot blocks:
cd /usr/mdec
disklabel -B -b boot1 da1
mkdir /boot
cp boot2 /boot
- Rebooted the machine, and booted off the second disk (dkc100). It
worked!!! Doug was right, the problem is that the boot loader doesn't
understand DEC's UFS, but it does work okay with FreeBSD's UFS.
- Rebooted again, and set the default boot device to the second disk in the
console.
I'm doing a make buildworld now, we'll see if it works.
Things seem to work okay for the most part. iostat(8) can probably be
enabled for the Alpha. It works fine for me, at least:
da0 da1 pass0 pass1
KB/t tps MB/s KB/t tps MB/s KB/t tps MB/s KB/t tps MB/s
0.00 0 0.00 8.66 38 0.32 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 3.00 7 0.02 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 10.22 32 0.32 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 3.00 8 0.02 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 10.63 30 0.31 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 3.60 5 0.02 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 7.00 2 0.01 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 4.75 4 0.02 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 3.00 4 0.01 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 4.00 4 0.02 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 3.00 7 0.02 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 2.00 7 0.01 0.00 0 0.00 0.00 0 0.00
0.00 0 0.00 5.18 11 0.06 0.00 0 0.00 0.00 0 0.00
Good work guys! This is pretty cool!
(I've attached a dmesg from the machine, in case anyone is curious or
whatever...)
Ken
--
Kenneth Merry
ken@plutotech.com
--ELM906791070-9668-0_
Content-Type: text/plain; charset=ISO-8859-1
Content-Disposition: attachment; filename=subway.dmesg
Content-Description: subway.dmesg
Content-Transfer-Encoding: 7bit
WARNING: 8192 bytes not available for msgbuf in last cluster (32768 used)
[ preserving 439368 bytes of kernel symbol table ]
Copyright (c) 1992-1998 FreeBSD Inc.
Copyright (c) 1982, 1986, 1989, 1991, 1993
The Regents of the University of California. All rights reserved.
FreeBSD 3.0-BETA #1: Thu Sep 24 16:44:04 MDT 1998
root@subway.plutotech.com:/usr/src/sys/compile/subway
Digital Personal Workstation (Miata)
Digital Personal WorkStation 433au, 432MHz
8192 byte page size, 1 processor.
real memory = 131891200 (128800K bytes)
avail memory = 122707968 (119832K bytes)
cia0: <2117x PCI adapter>
cia0: Pyxis, pass 1
cia0: extended capabilities: 111<WLEN,MWEN,BWEN>
cia0: WARNING: Pyxis pass 1 DMA bug; no bets...
isa0
Probing for devices on PCI bus 0:
de0: <Digital 21143 Fast Ethernet> rev 0x30 int a irq 0 on pci0.3.0
de0: DEC 21143 [10-100Mb/s] pass 3.0
de0: address 00:00:f8:75:7c:14
chip0: <Intel 82378IB PCI to ISA bridge> rev 0x43 on pci0.7.0
vga0: <VGA-compatible display device> rev 0x22 int a irq 4 on pci0.11.0
chip1: <PCI to PCI bridge (vendor=1011 device=0021)> rev 0x02 on pci0.20.0
Probing for devices on PCI bus 1:
Qlogic ISP Driver, FreeBSD CAM Version 0.97, Core Version 1.3
isp0: <Qlogic ISP 10X0 PCI SCSI Adapter> rev 0x05 int a irq 16 on pci1.9.0
isp0: using Memory space register mapping
isp0: Board Revision 1040B, loaded F/W Revision 7.55
isp0: Last F/W revision was 5.1
mcclock0: <MC146818A real time clock> at 0x70-0x71 on isa0
sc0 at 0x60-0x6f irq 1 on isa0
sc0: CGA <16 virtual consoles, flags=0x0>
sio0 at 0x3f8-0x3ff irq 4 on isa0
sio0: type 16550A
sio1: reserved for low-level i/o
sio1 not found
struct nfssvc_sock bloated (> 256bytes)
Try reducing NFS_UIDHASHSIZ
struct nfsuid bloated (> 128bytes)
Try unionizing the nu_nickname and nu_flag fields
Timecounter "alpha" frequency 432900432 Hz cost 62 ns
isp0: driver initiated bus reset
de0: enabling 100baseTX port
da0 at isp0 bus 0 target 0 lun 0
da0: <DEC RZ1BB-BS (C) DEC 0658> Fixed Direct Access SCSI2 device
da0: 40.0MB/s transfers (20.0MHz, offset 8, 16bit), Tagged Queueing Enabled
da0: 2007MB (4110480 512 byte sectors: 255H 63S/T 255C)
da1 at isp0 bus 0 target 1 lun 0
da1: <DEC RZ1BB-BS (C) DEC 0658> Fixed Direct Access SCSI2 device
da1: 40.0MB/s transfers (20.0MHz, offset 8, 16bit), Tagged Queueing Enabled
da1: 2007MB (4110480 512 byte sectors: 255H 63S/T 255C)
--ELM906791070-9668-0_--
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-alpha" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199809260624.AAA09773>