Date: Mon, 09 Jan 1995 22:48:19 -0500 From: "Mark G.M. O'Lear" <mgolear@mailbox.syr.edu> To: questions@FreeBSD.org Subject: 2.0.5 questions (harddrive, sio, kernel configuration and install) Message-ID: <9501100348.AA18828@hydra.syr.edu>
next in thread | raw e-mail | index | archive | help
Hi,
First of all, I would like to say that I have thoroughly enjoyed using
FreeBSD 1.5.1.1. I love the many extras the FreeBSD team has added that
make it more usable (the vi commands added to more, the fixed vi
autoindent/backspace problems, the pkg_xxx commands and the -z flag added
to tar, just to name a few) and made me choose FreeBSD over NetBSD, 386BSD,
Solaris, Ultrix, HP-UX, AIX and SCO (I haven't tried Linux because I don't
like SYSV, and I haven't used Ultrix, HP-UX or AIX in a few years, so maybe
they are better now, but you can't beat the price of FreeBSD).
The following was taken from the boot log (which should explain what
version I have and my CPU type):
...
kernel: FreeBSD 2.0.5-RELEASE #0: Thu Dec 22 14:10:34 PST 1994
jkh@time.cdrom.com:/usr/src/sys/compile/GENERIC
CPU: i486DX (486-class CPU) Id = 0x435 Origin = "GenuineIntel"
real memory = 20578304 (5024 pages)
...
I have some questions in the following four categories:
1) Hard drive questions:
The following lines were taken from the boot log:
...
kernel: wdc0 at 0x1f0-0x1f7 irq 14 on isa
kernel: wdc0: unit 0 (wd0): <WDC AC2340H>
kernel: wd0: 325MB (666600 total sec), 1010 cyl, 12 head, 55 sec,
bytes/sec 512
kernel: wdc0: unit 1 (wd1): <SyQuest SQ3270 AT>
kernel: wd1: 512MB (1048576 total sec), 2048 cyl, 16 head, 32 sec,
bytes/sec 512
...
Note: The last line is wrong and should read (but it never does):
kernel: wd1: 256MB (524288 total sec), 1024 cyl, 16 head, 32 sec,
bytes/sec 512
^^^ ^^^^^^ ^^^^
256 524288 1024
for the Syquest drive (wd1).
If I boot with a cartridge that has only one partition (a DOS partition),
I get the following message every time:
kernel: wd1: changing size of 'd' partition from 524288 to 1048576
I must then change the in-core disklabel back to the right thing, before
I can 'mount_msdos' it.
I am assuming that it cannot find a valid 4.2BSD disklabel on it, so it
is assuming that the value reported at boot time from the controller is
right. If I boot with a cartridge that has a FreeBSD partition, it
doesn't change the size. Is FreeBSD reading the wrong values, or is
the on-board controller on the Syquest reporting the wrong thing?
Also I am getting the following errors (listed are a few random ones) only
with the Syquest drive, no errors ever happen on the Western Digital:
kernel: wd1e: soft error reading fsbn 450853 of 450848-450863 (wd1 bn
504101; cn 984 tn 9 sn 5)wd1: status 5a<seekdone,drq,index>
error 1<no_dam>
kernel: wd1e: soft error reading fsbn 57650 (wd1 bn 110898; cn 216 tn 9
sn 18)wd1: status 5a<seekdone,drq,index> error 40<uncorr>
kernel: wd1e: soft ecc reading fsbn 57671 of 57664-57679 (wd1 bn 110919;
cn 216 tn 10 sn 7)wd1: status 5e<seekdone,drq,ecc_cor,index>
error 40<uncorr>
kernel: wd1e: hard error reading fsbn 57673 of 57664-57679 (wd1 bn 110921;
cn 216 tn 10 sn 9)wd1: status 5b<seekdone,drq,index,err> error
40<uncorr>
kernel: wd1e: hard error reading fsbn 417975 (wd1 bn 471223; cn 920 tn 5
sn 23)wd1: status 5b<seekdone,drq,index,err> error 1<no_dam>
kernel: wd1: status ff<busy,rdy,wrtflt,seekdone,drq,ecc_cor,index,err>
error 1<no_dam>
kernel: wd1: interrupt timeout:
kernel: wd1: wdunwedge failed:
kernel: wd1: status 5a<seekdone,drq,index> error 0
kernel: wd1: status 5a<seekdone,drq,index> error 1<no_dam>
kernel: wd1: status ff<busy,rdy,wrtflt,seekdone,drq,ecc_cor,index,err>
error 1<no_dam>
kernel: wd1d: wdstart: timeout waiting to give command reading fsbn 120
of 120-123 (wd1 bn 120; cn 0 tn 3 sn 24)wd1: status
ff<busy,rdy,wrtflt,seekdone,drq,ecc_cor,index,err>
error 1<no_dam>
The only question I can ask about the above is what does it all mean?
Is the controller trying to go to fast for the drive? (I never have
any problems with the drive in DOS, with or without the Syquest driver).
I don't seem to be loosing any data, only long delays in reading when
I get one of these errors. I get most of these errors when I do an
fsck. Very rarely I can manage to get through a complete fsck with
no errors. And fsck never reports any errors that it has to fix, only
cannot read sector n0-nx.
2) Serial I/O question:
I am getting the following error when I use kermit (FreeBSD 1.5.1.1
executable) to log in to start a SLIP connection:
kernel: sio0: 503 more tty-level buffer overflows (total 2391)
After I log in, suspend kermit and start slattach, ifconfig sl0 ..., etc.
I never get this error. Is this because kermit is not reading the
input buffer fast enough? (kermit seems to be displaying the incoming
text very slow -- 2 lines per second max with my 14.4 bps modem with
serial port set to 38400 bps -- and looks more like a 2400 bps modem).
I am getting reasonable throughput (1.8 - 0.9kbps on ftp transfers).
3) Reconfiguring the kernel question:
The following line was taken from the boot log of the new kernel:
kernel: FreeBSD 2.0-RELEASE #0: Sat Jan 7 00:52:27 EST 1995
After I reconfigured the kernel (I am using the 2.0-RELEASE kerndist
because I couldn't find one in the 2.0-...-SNAP directory [maybe that
is the problem]) I can no longer get route to work correctly. I get
the following message:
routed[60]: adding route to net/host 128.230.1.96 through gateway
128.230.1.96: Protocol not supported
I believe that I can still ping localhost with no problem, but I cannot
remember. The new kernel seems to have no other new problems associated
with it (aside from 1) and 2) above). Just to see if there was something
I was forgetting (in my kernel configuration file from FreeBSD 1.5.1.1)
I decided to build the GENERIC and GENERICAH kernels supplied with the
kerndist. Both these kernels produced the same 'Protocol not supported'
messages and would not let route add any routes to the routing table.
Should I be using the 2.0-RELEASE kerndist with the 2.0.5-RELEASE, or
does this happen with 2.0-RELEASE install as well (I never installed
2.0-RELEASE, only the 2.0.5 snapshot). The SoundBlaster support works
very well in the new kernel though. Also yes I have reconfigured the
kernel many times before (using 1.5.1.1) with complete success, and yes
I did do a 'make depend'.
4) Installation question:
While installing the 2.0.5 snapshot, I had no problems until I chose
'Install bindist' from the menu. I chose to install from a UFS
partition. When it gave me the shell, I mounted the Syquest drive
with the bindist (and other xxxdist's) on it, exited from the shell
and went back into the install program (very nice install program
by the way!) I then proceeded and the install program said it
couldn't cd to the (/syquest/usr/distrib/"bindist") directory. I
tried several times before I renamed the bindist directory to
"bindist" (including the quotes) and everything worked fine. This
seems really strange to me that you would require the xxxdist
directories to have " marks around them! I had to put quotes
around every xxxdist directory before the install program would
recognize these as valid install directories. Note: I didn't tell
the install program that the directories had " marks around the
directories, I just renamed the directories from xxxdist to "xxxdist".
Is this really the way you had intended it to be, or is this a bug
or what?
Thanks in advance,
Mark.
mgolear@mailbox.syr.edu
P.S.
Sorry about the length, I got carried away (probably over-explaining).
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?9501100348.AA18828>