Subject: Linux-Development Digest #402
From: Digestifier <Linux-Development-Request@senator-bedfellow.MIT.EDU>
To: Linux-Development@senator-bedfellow.MIT.EDU
Reply-To: Linux-Development@senator-bedfellow.MIT.EDU
Date:     Sun, 23 Jan 94 08:13:06 EST

Linux-Development Digest #402, Volume #1         Sun, 23 Jan 94 08:13:06 EST

Contents:
  Re: nfs performance (Gregory Gulik)
  DPT PM2021/9x-SCSI Controller (Mike Jagdis)
  Upper Memory Blocks ?? (Leon Garde)
  Re: SLIP Server setup: Routing and interface ??? (Terry Dawson)
  Re: Better than Xmag ? (Mark A. Davis)
  cluster-07 anf HPFS (Bao Chau Ha)
  Re: mac linux (Chen Zhou)
  Re: Mitsumi double-speed FX001D (W. Woody Jin)
  Re: Van Jacobson slow start (Re: Linux v1.0: what's in it?) (Charles Hedrick)
  Re: pl14 route question (Charles Hedrick)
  Re: Broadcast Bug? (Florian La Roche)
  Kernel swap (Kurt Long)
  Re: Broadcast Bug? (Abhinandan Jain)
  Please delete multiple copies of pl14 route/ping dup packets (Cliff V Collins)
  sorry - please delete/ignore multiple posts- system error (Cliff V Collins)
  Re: Subnetting on non byte boundaries (Hamish Coleman)

----------------------------------------------------------------------------

From: greg@gagme.wwa.com (Gregory Gulik)
Subject: Re: nfs performance
Date: Sat, 22 Jan 1994 16:33:57 GMT

In article <758962807.0@freudsys.iaf>,
Renee Teunissen <Renee.Teunissen@freudsys.iaf.nl> wrote:
>Could you please tell me what iozone is and is it sharable?

iozone is a disk I/O measurement package.  Like all benchmarks, it
isn't perfect, but at least gives you a general idea of the performance.

Send me E-mail if you want me to mail you a copy.


-- 
Gregory Gulik - greg@gagme.wwa.com
WorldWide Access (SM) - Chicagoland Internet Services
+1 312 282 8605 or send mail to info@wwa.com

------------------------------

From: jaggy@purplet.demon.co.uk (Mike Jagdis)
Subject: DPT PM2021/9x-SCSI Controller
Date: Fri, 21 Jan 1994 23:28:00 +0000

* In message <1994Jan20.232759.308@erde.GUN.de>, Martin Seine said:

MS> Does anyone know, if there's a driver for the DPT SCSI-Controller
MS> (especially the ISA-Type PM2021/9x) already available, under
MS> work or planned?

I started on a driver for the DPT a range a while back. I haven't had chance 
to do any more work on it for a long while now though :-(.

MS> Does anyone have programming docs of this card ?

It's all available on DPT's BBS and ready to download. The number is in your 
manual :-).

                                Mike  
 

------------------------------

From: lgarde@scorch.hna.com.au (Leon Garde)
Subject: Upper Memory Blocks ??
Date: Thu, 20 Jan 1994 01:23:19 GMT

I had an idea about the memory some motherboards map
to the 640k-1024k physical address range, what in the dos world
is called upper memory blocks.


How about the linux kernel detects these and used them as buffer space ?
 

Ok, most motherboard/bios's will not map RAM into these areas,
giving more extended memory (memory above 1024k ) , but some dont
and some people like to have their cmos set for a good fast RAM
abundant DOS configuration. (Though at least one memory manager exists
that can override the cmos settings by using mmu specific io port operations.
The one I have seen is 'The Last Byte Memory Manager',in an archive called
TLB-V220.ZIP  at wuarchive  )

Some people might like  EMS  capability, but that is a bit too fussy 
just as trying to override the cmos settings is.

leon

------------------------------

Crossposted-To: comp.os.linux.admin,comp.os.linux.help
From: terryd@extro.ucc.su.OZ.AU (Terry Dawson)
Subject: Re: SLIP Server setup: Routing and interface ???
Date: Fri, 21 Jan 1994 01:07:38 GMT

dlincke@sgcl1.unisg.ch (David-Michael Lincke) writes:

>ifconfig sl0 130.82.154.12 broadcast 130.82.255.255 netmask 255.255.0.0 \
>mtu 1500 up

You're going to conflict with your ethernet port here. They can't both
have the same broadcast address and support the same network.
If you are using dip, you shouldn't have to ifconfig the interface at all.
The 'local' and 'remote' commands should set the addresses.

If you want to ifconfig, then at least try this:

ifconfig sl0 130.82.154.12 pointopoint <clientaddress> mtu 1500

An mtu of 1500 is probably a bit large for a slip link.

>With this routing set up I could ping the client and the client could ping the 
>server. I also could do an FTP session from the server to the client (client 
>running Netmanage TCP/IP under MS Windows). But when trying to establish a 
>telnet connection from the client to the SLIP server it was in fact established
>but I did not get any output on the client (couldn't login...). It seems to me 
>as if only ICMP worked both ways but not TCP.

Check that you are running SLIP, and not CSLIP.

Terry


-- 
--- Terry Dawson, terryd@extro.ucc.su.oz.au

------------------------------

Crossposted-To: comp.os.linux.help
From: mark@taylor.wyvern.com (Mark A. Davis)
Subject: Re: Better than Xmag ?
Date: Sat, 22 Jan 1994 23:24:53 GMT

dminer@nyx10.cs.du.edu (Dan Miner) writes:

>In article <1994Jan21.150906.24321@waikato.ac.nz>,
> <phys2108@waikato.ac.nz> wrote:

>[...]

>>> -- 
>>>   /--------------------------------------------------------------------------\
>>>   | Mark A. Davis    | Lake Taylor Hospital | Norfolk, VA (804)-461-5001x431 |
>>>   | Sys.Administrator|  Computer Services   | mark@taylor.wyvern.com   .uucp |
>>>   \--------------------------------------------------------------------------/
>>How about getting a lcd projection unit and putting the output on a wall.

>What bottomless bank account do you have?!  *grin*

Although your news program did the correct quoting citations above, you should
have removed my signature because it tends to imply that I wrote any of
the above, which I did not.  Thanx
-- 
  /--------------------------------------------------------------------------\
  | Mark A. Davis    | Lake Taylor Hospital | Norfolk, VA (804)-461-5001x431 |
  | Sys.Administrator|  Computer Services   | mark@taylor.wyvern.com   .uucp |
  \--------------------------------------------------------------------------/

------------------------------

From: habaoch@eng.auburn.edu (Bao Chau Ha)
Subject: cluster-07 anf HPFS
Date: Sun, 23 Jan 1994 00:28:01 GMT

I installed cluster-07, and I could no longer compile the kernel
with the optional HPFS file system.  I think the function breada
used in hpfs_fs.c is defined differently than the one cluster-07
patches the kernel with.

Any hints.
Thanks.
Bao

------------------------------

From: czen@namaste.cc.columbia.edu (Chen Zhou)
Subject: Re: mac linux
Date: 23 Jan 1994 01:41:00 GMT

see. i am actually quite interested in the part about having unix take over the
whole machine. since then i will git unix security. most of the current mac 
os implementation of the security feature are either too easy to be defeated or
not safe(AME is hard, but imagine who would sleep tight with the entire content of his hardrive encrypted. and have the driver being replaced)


DataOut(/\)

****************************************************************************
* Chen "Data" Zhou     *            HI, Jen and Rachel !!!                 *
* Academic Computing   *                                                   *
* Columbia University  *
* -------------------- *                                                   *
* czen@columbia.edu    ***************************************************** 
* -------------------- *   With the first link, the chain is forged.       *
* Columbia University  *   The first speech censored,                      *
* 746 John Jay         *   the first thought forbidden,                    *
* New York, NY 10027   *   the first freedom denied,                       *
* USA                  *   chains us all irrevocably.                      *
* (212) 853-6143       *                              - Aaron Satie        *
****************************************************************************



------------------------------

From: wjin@cs.uh.edu (W. Woody Jin)
Crossposted-To: comp.os.linux.help
Subject: Re: Mitsumi double-speed FX001D
Date: 23 Jan 1994 02:58:04 GMT

In article <1994Jan12.065552.4852@citrus.SAC.CA.US> qumqats@citrus.SAC.CA.US (Joel M. Baldwin) writes:
>>>>
>>>>"Kernel Panic: Mitsumi CD-ROM: request list destroyed" and the system stop.
>>>I have this problem also!!!!!!  Only I have the single speed drive.
>>>Anytime I try to do anything major with the drive I end up with a locked 
>>>up system.
>>>It makes the CDROM drive useless.  Anyone have any ideas?
>ok   I now have the latest p14 off of tsx-11.  I no longer have the
>"request list destroyed" errors . . . BUT I am still getting these errors.
>
><6>mcd: data timeout, retrying
><6>mcd: status timed out
><6>mcd: status timed out
><6>mcd: read failed after 3 tries
><6>Mitsumi CD-ROM I/O error
><6>dev 1700, sector 20560
> ....
>There is no error on this disk.  It works some times but not others.
>Amount of system activity might have an effect on how often this happens.
>I might be pushing the AT bus at 10mhz.  Can the controller card take this?


I think that the Mitsumi  Double speed CD-ROM drive is not reliable.
I have the drive but I have not tried it in Linux.
It very often gives me "the drive is not ready" error under DOS/Windows.
Windows suggests me that either the CD is not well inserted or the 
material is bad.  But I don't find any such thing.   After 2 to 5 times of
eject/insert, it works fine.

My kid also complained that when she tried "Arthur Teacher Trouble" in
Spanish, the system could not run the info from CD any longer and the
program exited.  In some other time, it worked O.K. in Spanish.

I guess that this drive (Mitsumi Double Speed CD ROM Drive) is very
poorly built.   And I do not recommend this drive to anyone.


-- 
____   ____  ____ ______________________________________ (__) _________________
|  |   |  |  |  |  W. Woody Jin (wjin@cs.uh.edu)         (oo)      Moo...
|  |   |  |__|  |  Bovinetic Algorithm Design     o=======\/  I'm  a Cow Lover.
|  |   |        |  Dept. of Cowmooputer Science / |     ||   My wife  was born
\  |---|  |--|  |  University of Houston       *  ||w---||   in Cow year.  Mooo
 \____/|__|  |__| ________________________________^^    ^^_____________________
% nohup cat *flames* >& /dev/bull &

------------------------------

From: hedrick@edinboro.rutgers.edu (Charles Hedrick)
Subject: Re: Van Jacobson slow start (Re: Linux v1.0: what's in it?)
Date: 23 Jan 94 02:49:36 GMT

In a discussion of 1.0, I said 
>"Retransmission is now using a fairly simple slow start strategy.
>Jacobson has come up with more clever things, which the Hosts
>Requirements RFC says should be implemented.  However at the moment we
>don't need cleverness -- we need reliability.

Two different people have now told me how important it is to have slow
start.  We do have slow start.  It's just a simple implementation,
which is probably slightly more conservative than the latest BSD code
would be about opening up the window after losing a packet.

------------------------------

From: hedrick@edinboro.rutgers.edu (Charles Hedrick)
Subject: Re: pl14 route question
Date: 23 Jan 94 02:51:51 GMT

cvc@world.std.com (Cliff V Collins) writes:

>I have upgraded to kernel dev. pl14p and compiled it with the ne2000 driver.
>Everything boots good and the system now can rcp and ftp to my xenix
>and pc systems running ftp softwares pctcp ver2.05 software both to and
>from linux, whereas before at pl13 file transfers from linux to each
>of these systems would halt at about 99% of the file transfer.

>I have 3 old 68020 systems ruinning bsd 4.3, and when I ping these systems
>I am getting DUP packets messages. After I ping them, telnet,rsh,rcp etc.
>from any other system (ibm r/6000, sun sparc, linux) just hangs. Only the

There was a quick series of network updates.  I am reasonablly sure
that p was one that was broken.  As far as I can tell, q, r, s, and t
each contain improvements and don't break anything new.  You should
expect continuing improvements, as several people are tracking down
various problems reported here and on the network channel.

------------------------------

From: rzsfl@sbusol.rz.uni-sb.de (Florian La Roche)
Crossposted-To: comp.os.linux.help
Subject: Re: Broadcast Bug?
Date: 23 Jan 1994 03:02:34 GMT

Derek Hackbardt (hackbard@egr.msu.edu) wrote:
: I am not sure if this is a kernel bug or not..

: It seems that when my default route is set to our gateway, broadcast 
: packets carry the wrong netmask.  This problem occurs when using yp
: under Slackware 1.1.1 and 0.99.14s kernel.  Although when I set the
: default route to the nisserver everything appears to work properly..

: The netmask should be: 255.255.255.0
: Broadcast: 150.220.40.255

: The outgoing broadcast packets go to 150.220.255.255.

: Any ideas?  I can provide more details to who ever would like..

Are you using the new route.c found on funet?
(You specify a netmask for your route there.)
Please post what you get from 'ifconfig' and from 'route'.
Then I may be able to help you...

Florian  La Roche

P.S.: I think 'ifconfig' will report a netmask of '255.255.0.0', but I may
      be wrong...

: Thanks..

: -derek A. hackbardt
: hackbard@mced.kodak.com



------------------------------

From: kl01@roger.gte.com (Kurt Long)
Subject: Kernel swap
Date: Sun, 23 Jan 1994 03:41:23 GMT

It appears that swap space is no longer limited to 16M per device in pl14
(pardon if this is an old question).  Any comments?

Mitch

mlm@intellibase.gte.com
-- 
| root@beachbum.intellibase.gte.com                         | GTE-TSI         |
| "I close my eyes, only for a moment and the moment's gone.|  "Don't blame   |
|  All my dreams, pass before my eyes, a curiosity.         |   them for what |
|  Dust in the wind, all they are is dust in the wind."     |   I think."     |

------------------------------

From: jain@telerobotics.jpl.nasa.gov (Abhinandan Jain)
Crossposted-To: comp.os.linux.help
Subject: Re: Broadcast Bug?
Date: 23 Jan 1994 04:02:47 GMT

>>>>> "Florian" == Florian La Roche <rzsfl@sbusol.rz.uni-sb.de> writes:
In article <2hspcaINN7an@sbusol.rz.uni-sb.de> rzsfl@sbusol.rz.uni-sb.de (Florian La Roche) writes:


    Florian> Derek Hackbardt (hackbard@egr.msu.edu) wrote: : I am not
    Florian> sure if this is a kernel bug or not..

    Florian> : It seems that when my default route is set to our
    Florian> gateway, broadcast : packets carry the wrong netmask.  This
    Florian> problem occurs when using yp : under Slackware 1.1.1 and
    Florian> 0.99.14s kernel.  Although when I set the : default route
    Florian> to the nisserver everything appears to work properly..

    Florian> : The netmask should be: 255.255.255.0 : Broadcast:
    Florian> 150.220.40.255

    Florian> : The outgoing broadcast packets go to 150.220.255.255.

    Florian> : Any ideas?  I can provide more details to who ever would
    Florian> like..

    Florian> Are you using the new route.c found on funet?  (You specify
    Florian> a netmask for your route there.)  Please post what you get
    Florian> from 'ifconfig' and from 'route'.  Then I may be able to
    Florian> help you...

    Florian> Florian La Roche

I have also seen this "broadcast bug". The results from my ifconfig and
route commands are as follows:

lo        IP ADDR 127.0.0.1  BCAST 127.255.255.255  NETMASK 255.0.0.0
          MTU 2000  METRIC 0  POINT-TO-POINT ADDR 0.0.0.0
          FLAGS: 0x0049 ( UP LOOPBACK RUNNING )

sl0       IP ADDR 131.215.199.140  BCAST 131.215.255.255  NETMASK 255.255.255.0
          MTU 296  METRIC 0  POINT-TO-POINT ADDR 131.215.139.207
          FLAGS: 0x0051 ( UP POINTOPOINT RUNNING )

Kernel routing table
Destination net/address   Gateway address           Flags RefCnt    Use Iface
default                   131.215.139.207           UGN        0      0 sl0
127.0.0.1                 *                         UH         0     42 lo
131.215.139.207           *                         UH         0      1 sl0

I would have expected a broadcast address of 131.215.199.255. Anyway, I
am unable to even ping my gateserver (131.215.139.207), so am still
fishing for possible configuration problems.

By the way, my address (131.215.199.140) and that of my gateserver
(131.215.139.207) do not appear to be on a common net. Most
configuration examples that I have seen for SLIP are for the server and
the local machine having common network addresses. Do I need extra
ifconfig/route commands to set up my connection properly?

Any pointers will be very much appreciated. Thanks.




--
Abhinandan Jain                    Jet Propulsion Laboratory
jain@telerobotics.jpl.nasa.gov     4800 Oak Grove Drive, Pasadena, CA 91109

------------------------------

From: cvc@world.std.com (Cliff V Collins)
Subject: Please delete multiple copies of pl14 route/ping dup packets
Date: Sun, 23 Jan 1994 06:35:30 GMT

if someone can, please delete the multiple copies of my last post about
pl14. Because the partition that holds the nn data on world.std.com's hard
drive went 100%, I was getting some errors when I tried to post the message.
I attempted 3-4 times thinking it was something I was doing wrong. Unfortunatly,
when the staff at world.std.com fixed the problem, all the copies went out.
If there had been someway to delete these I would have (I thought they simply
got trashed anyhow). I know one message is all it takes (if its asked right)
to get great help around here anyway. Sorry, and thanks



------------------------------

From: cvc@world.std.com (Cliff V Collins)
Subject: sorry - please delete/ignore multiple posts- system error
Date: Sun, 23 Jan 1994 06:46:50 GMT

please delete/ignore/forgive multiple postings about pl14 dup packets.
Due to a world.std.com partition filling up, the error I thought I was
getting was my screwup in how to send the question. All my attempts
were queued and when the system problem was fixed, all of them went.
Otherwise thanks for the help I did get cause I got the problem solved.


------------------------------

From: hamish@zot.apana.org.au (Hamish Coleman)
Subject: Re: Subnetting on non byte boundaries
Date: 23 Jan 1994 14:58:49 +1100

mmclagan@invlogic.com (Mike McLagan) writes:
>In article <1994Jan20.051411.14674@colesmyer.com.au> jlew@colesmyer.com.au writes:
>>The first part is correct, but the .00 is definitely NOT OK. With-in and
>>with-out the 222.222.222.xxx address space, 222.222.222.0 is referred to by
>>routers/gateways as the entire 222.222.222.xxx network. Networks external to
>>222.222.222.xxx do not see the individual subnets within the 222.222.222.xxx
>>network. They only see one single network, and this is known to them as
>>222.222.222.0. It's a bit like if you're posting mail from Germany to
>>me in Australia, your post office need not know my post office, it only needs
>>to know that I'm in Australia. Hence your post office would only record
>>"Australia" in its routing table and not all the post offices inside OZ.
>>Once it gets into Australia, the local post office is then required.
>>
>>As a result, once you subnet a C-Class address (or B-Class/A-Class address
>>space for that matter), you lose the "first" subnet which is required for
>>identifying your network.
>>

>Sorry for the double reply, but I've got a question re the above:

>Does this mean in a .240 subnet masked situation that using the .00 subnet
>is forbidden?  If I am going to subnet that way should I not have .1, .2, .3,
>and .4 machines?  (If not, I've got a BIG problem!)

Well, I also am sure that both the .00 subnet and the .240 subnet (with the .240
mask as above) are _totally_ valid sub-nets ...

Sure, an _outside_ machine may think that xx.xx.xx.00 subnet is the whole network,
but it will _also_ think that xx.xx.xx.32 (or any other number) is also on the
same network - it doesnt know your subnetting.  The router simply forwards the
packet until it reaches your network, and if you are subnetting, your network
should find it quite simple to route the packet to the correct subnet from there.

As for the broadcast problem - any host in your network knows that .255 is the broadcast
for the subnet .240 _only_, and any outside host that is trying to flood your network
should _know_ your network setup -- otherwise its just some idiot trying to
flood your network .. (hmm - that doesnt sound quite right, but I hope you
know what I mean).

Of course, all this assumes that the machines on the network are correctly setup
(of course :-)


-- 
Use Linux, Give me a Job and Drink Red cordial.    hamish@zot.apana.org.au
=====Cut=here=for=pre=tensioned=paper=eliminating=small=strip=of=text=====
``Life is like a grapefruit ... it's sort of orangey-yellow and dimpled on
the outside, wet and squidgy in the middle.  It's got pips inside too. Oh,
and some people have half a one for breakfast.'' -- Ford Prefect

------------------------------


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: Linux-Development-Request@NEWS-DIGESTS.MIT.EDU

You can send mail to the entire list (and comp.os.linux.development) via:

    Internet: Linux-Development@NEWS-DIGESTS.MIT.EDU

Linux may be obtained via one of these FTP sites:
    nic.funet.fi				pub/OS/Linux
    tsx-11.mit.edu				pub/linux
    sunsite.unc.edu				pub/Linux

End of Linux-Development Digest
******************************
