Subject: Linux-Development Digest #778
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:     Tue, 31 May 94 00:13:05 EDT

Linux-Development Digest #778, Volume #1         Tue, 31 May 94 00:13:05 EDT

Contents:
  Re: 8k nfs performance (Marcel Roelofs)
  Three-button mouse, again? (DL83-08)
  Re: InterLink for Linux (Alan Cox)
  Re: Three-button mouse, again? (John M Prevost)
  Serial drivers in 1.1.16 (autoconfig for IRQ's) (Nick Simicich)
  Re: kernel 1.1.14 (&13) + xiafs segfaults rpc.nfsd (Jan Niehusmann)
  Re: 1.1.15 breaks SCSI (Michael Schmidt)
  Re: Serial drivers in 1.1.16 (autoconfig for IRQ's) (Janne Sinkkonen)
  Re: SLIP problems with kernel 1.1.15 (Janne Sinkkonen)
  PC-NFS-Daemon sources anywhere? (H.P.Heidinger)
  Logitech Scanner (Re: Twain Scanner) (Karl Eichwalder)
  Re: 1.1.15 breaks SCSI (Paul Rubin)
  Re: Serial drivers in 1.1.16 (autoconfig for IRQ's) (Mikael Nordqvist)
  1.1.16 NFS hangs on write when wsize >1k (Re: 8k NFS client in 1.1.13) (Oliver Schoett)
  Re: Three-button mouse, again? (Jim Breen)
  Re: NI6510 (Chris Hafey)
  Re: 1.1.15 breaks SCSI (Barry Hannigan; EdTel)
  Re: Latest and Greatest networking code (Michael Simmons  - division)
  Re: Pausing of program execution (introduced in pl1.1.13) (Harald Schreiber)
  Can one determine real UID of a process from /proc? (Delman Lee)

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

From: roelofs@cwi.nl (Marcel Roelofs)
Subject: Re: 8k nfs performance
Date: Mon, 30 May 1994 15:44:51 GMT

In article <CqGszJ.C0p@setanta.demon.co.uk>, at@setanta.demon.co.uk (Amrik Thethi) writes:
|> Talking of 8K NFS, is anybody currently in the process of
|> implementing caching for the client? This is where the
|> real speed improvement will come. Is there any technical
|> reason why it would be difficult, or is it just a matter
|> of sitting down and hacking it out?

Is this client caching in any way connected to the large quantities of
nfsd & biod daemons floating around on almost every Unix system I have 
ever worked on (that is, except for Linux)?

Marcel

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

From: s21008@cc.ntnu.edu.tw (DL83-08)
Subject: Three-button mouse, again?
Date: Mon, 30 May 1994 16:25:02 GMT

Thank you to all who wrote about the dual-protocol mouse configuration for
X. Obviously, my mouse is a bit different from most. The DOS driver has two
modes: Microsoft and PC mouse. The microsoft mode uses two buttons, and the
PC mouse uses all three buttons. In my Xconfig file, I can specify the
Microsoft protocol, of course, and the mouseman protocol. The problem is
that both use only two buttons. All of the responses to my last request for
help told me to use mousesystems with the cleardtr statement. Well, I tried
that, and also tried the clearrts, as well. The problem is that microsystems
just does not work with this mouse. Is there any other way to force the
middle button to be recognized in mouseman mode?

Thank you for any hints!

--

 /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
%v  Curtis Dean Smith                      Internet: s21008@cc.ntnu.edu.tw   |
0j  Institute of Chinese Language & Literature                               |
?3  National Taiwan Normal University            Bitnet: NTNUS239@TWNMOE10   |
 \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

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

From: iiitac@uk.ac.swan.pyr (Alan Cox)
Subject: Re: InterLink for Linux
Date: Mon, 30 May 1994 15:11:10 GMT

In article <2s8ju0$6l1@rc1.vub.ac.be> dglaude@vub.ac.be (GLAUDE DAVID) writes:
>I know it would be better if everyboddy use Linux on it's machine... but
> - I will need to use Linux as a file server for an other PC running Msdos
>  or Windows.
> - A friend is deseprately trying to use a Linux-Box as a file server for his
>  hard disk less Amiga.
>
>Of course all of this without Ethernet card ;-) Only with a Parallel or Serial
> Cable.
>
Linux 1.1.x PLIP is Crynwyr compatible. There is a crynwyr compatible PLIP
for Linux 1.0.x on sunacm.swan.ac.uk. SLIP and CSLIP are official standards
and I had no trouble with CSLIP between Amiga and Linux until my Amiga blew
up.

Alan



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

From: John M Prevost <visigoth+@CMU.EDU>
Subject: Re: Three-button mouse, again?
Date: Mon, 30 May 1994 13:33:51 -0400

I picked up a mouseman not too long ago, and had it working fine right
from install...  However, I reinstalled linux, and had a problem--it
wouldn't recognize the middle button.  I found that adding the
"ChordedMiddleButton" (or whatever) line fixed this.  The mouseman (at
least as it looks at itself now) sends the middle button as a LR
chord...  Go figure.


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

From: njs@scifi.uucp (Nick Simicich)
Subject: Serial drivers in 1.1.16 (autoconfig for IRQ's)
Date: Mon, 30 May 1994 18:41:48 GMT


I have a serial card with four serial ports on it.  These are straight
16550's, so they each need an interrupt, and are plugged on their own
interrupt.  

In older versions of Linux, I was able to define 'CONFIG_AUTO_IRQ' and
have Linux detect the IRQ numbers for the serial ports automatically,
which it always did.

I skipped to this level from 1.0.6 or some such, and noted that most
(but not quite all) references to CONFIG_AUTO_IRQ were removed.  There
is still code which can be invoked to detect the IRQ numbers for the
serial ports, but, so far as I can tell, this can only be invoked if
you change the definition of STD_COM_FLAGS and STD_COM4_FLAGS to add
ASYNC_AUTO_IRQ. 

Personally, I found this feature quite useful and reliable, and was
hoping that the symbol would someday be addeed to the autoconfig.  Any
chance of getting it back in?  On Linux, the '4,3,4,3' irq pattern for
com1-4 (as hard coded in serial.c) doesn't make much sense.

-- 
Nick Simicich - njs%scifi.uucp@uunet.uu.net - njs@watson.ibm.com

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

From: jan@gondor.gun.de (Jan Niehusmann)
Subject: Re: kernel 1.1.14 (&13) + xiafs segfaults rpc.nfsd
Date: Mon, 30 May 1994 14:45:30 GMT

Rene COUGNENC (rene@renux.frmug.fr.net) wrote:

 > Well, I don't know if it is really related to xiafs, but...

No, I think it's not...

 > With 1.1.14 or 1.1.13:

 >      On my 386 using xiafs, rpc.nfsd dies at any time.I can't keep
 >      it running as soon as some operations like readdir are done.
 >      (ls, or shell filename completion, etc).

On my 386, using e2fs: The same problems (with 1.1.16). With 1.1.9 
everything works fine. 

So it shouldn't be a problem of xiafs.

Bye, Jan

--
    * Jan Niehusmann * jan@gondor.gun.de * 2:2450/825 * 

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

From: michael@muc.de (Michael Schmidt)
Subject: Re: 1.1.15 breaks SCSI
Date: 30 May 1994 20:14:30 +0200

rob@pe1chl.ampr.org (Rob Janssen) writes:

>In <2san3e$rm@colin.muc.de> michael@muc.de (Michael Schmidt) writes:

>>lellis@dmccorp.com (John Lellis) writes:

>>>Rene COUGNENC (rene@renux.frmug.fr.net) wrote:

>>>: The kernel 1.1.15 does not work at all with my Adaptec 1540B.
>>>: I get "Unable to reset SCSI host 0, probably a SCSI bus hang."

>>>: And I can't boot.

>>Same here, also with an 1540B.
>>Could it be, that only 1540B's are hit?

>I don't think so...  I have a 1542B and it works with 1.1.15 & 1.1.16.
>The only difference between the 1540 and 1542 is supposed to be the
>floppy controller.  And I have disabled it.

Hmmm. Perhaps slow machines? I have a 386/40. Rene has also a 386, if
I remember correctly.

        Michael

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

From: sinkkone@cc.Helsinki.FI (Janne Sinkkonen)
Subject: Re: Serial drivers in 1.1.16 (autoconfig for IRQ's)
Date: 30 May 1994 22:40:59 +0300

In article <CqMp9p.EzF@scifi.uucp>, Nick Simicich <njs@scifi.uucp> wrote:
>Personally, I found this feature quite useful and reliable, and was
>hoping that the symbol would someday be addeed to the autoconfig.  Any
>chance of getting it back in?  On Linux, the '4,3,4,3' irq pattern for
>com1-4 (as hard coded in serial.c) doesn't make much sense.

As a remedy, use the setserial program in your /etc/rc*

--
Janne



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

From: sinkkone@cc.Helsinki.FI (Janne Sinkkonen)
Crossposted-To: comp.os.linux.help
Subject: Re: SLIP problems with kernel 1.1.15
Date: 30 May 1994 22:45:00 +0300

In article <2sbrik$hus@clarknet.clark.net>,
Robert Daniels <rdaniels@clark.net> wrote:

>using 1.1.15 kernel and the dslip101.tar.gz (available on sunsite) i 
>have slip working. above v1.0 however, i can only get sl0 to work.

IP forwarding was broken even in 1.1.15. This affects systems with
several sl and/or eth interfaces. pl16 is fine.  There's also
some kind of bug preventing the use of several sl interfaces. This
should be corrected at least in pl16, too.

>i suppose that if you are using over v1.0 (i'm not sure howmany patchlevels
>above they started using the net3 code) you should get the net-0.32d tar
>file from sunsite.

right.

--
Janne

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

From: hph@hphbbs.E.open.DE (H.P.Heidinger)
Subject: PC-NFS-Daemon sources anywhere?
Reply-To: hph@hphbbs.e.open.de
Date: Mon, 30 May 1994 00:58:35 GMT

  Hello NetWorld,

  I recently restored the shadow-passwd-suite from my former
  SLS installation and re-integrated it into my current Slackware
  installation.

  To make a long story short ... the 'pcnfsd' now refuses all
  client-users except 'nobody' (65534,65534). Thus I need the source
  to re-create that baby and link it with 'libshadow' in order to
  have the passwords picked from /etc/shadow instead of /etc/passwd.

  I guess there must be source code anywhere ... but where?

  Any pointers are appreciated (Email please)
 

Regards, Peter 
-- 
####################===============================****************************
# H.P. Heidinger   # Call  : +49-201-287433 (data) *      ~~ HPHBBS  ~~       *
# Steeler Str. 121 #       : V22/32/42bis/FAX, 8N1 *    UseNet City-Router    *
# 45138 Essen      # E-Mail: hph@hphbbs.E.open.de  *    for Essen/Germany     *

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

From: karl@pertron.central.de (Karl Eichwalder)
Subject: Logitech Scanner (Re: Twain Scanner)
Date: Mon, 30 May 1994 18:08:21 GMT

CHRISTOPHER M MAY (cmay@titan.ucs.umass.edu) wrote:

> /pub/linux/ALPHA/scanner/logiscan.tar.gz  

> I haven't tried it yet, but I heard it works.

Also for the ScanMan256 ?  logi32.c complains here all the time:

        logi32.c: cannot register major number

--
Karl

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

From: phr@netcom.com (Paul Rubin)
Subject: Re: 1.1.15 breaks SCSI
Date: Mon, 30 May 1994 21:32:44 GMT

In article <2sdae6$8q7@colin.muc.de>, Michael Schmidt <michael@muc.de> wrote:
>rob@pe1chl.ampr.org (Rob Janssen) writes:
>
>>In <2san3e$rm@colin.muc.de> michael@muc.de (Michael Schmidt) writes:
>
>>>lellis@dmccorp.com (John Lellis) writes:
>
>>>>Rene COUGNENC (rene@renux.frmug.fr.net) wrote:
>
>>>>: The kernel 1.1.15 does not work at all with my Adaptec 1540B.
>>>>: I get "Unable to reset SCSI host 0, probably a SCSI bus hang."
>
>>>>: And I can't boot.
>
>>>Same here, also with an 1540B.
>>>Could it be, that only 1540B's are hit?
>
>>I don't think so...  I have a 1542B and it works with 1.1.15 & 1.1.16.
>>The only difference between the 1540 and 1542 is supposed to be the
>>floppy controller.  And I have disabled it.
>
>Hmmm. Perhaps slow machines? I have a 386/40. Rene has also a 386, if
>I remember correctly.

Nope.  I know someone losing in the same way, with a 1542B and
a 486dx2/66.

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

From: d91mn@efd.lth.se (Mikael Nordqvist)
Subject: Re: Serial drivers in 1.1.16 (autoconfig for IRQ's)
Date: 30 May 1994 22:19:25 GMT

In article <2sdfgb$9tb@plootu.helsinki.fi> sinkkone@cc.Helsinki.FI (Janne Sinkkonen) writes:
>
>In article <CqMp9p.EzF@scifi.uucp>, Nick Simicich <njs@scifi.uucp> wrote:
>>Personally, I found this feature quite useful and reliable, and was
>>hoping that the symbol would someday be addeed to the autoconfig.  Any
>>chance of getting it back in?  On Linux, the '4,3,4,3' irq pattern for
>>com1-4 (as hard coded in serial.c) doesn't make much sense.
>
>As a remedy, use the setserial program in your /etc/rc*

It seems that setserial doesn't do the job under some circumstances (like
mine). When I use setserial to change irqs it reports them as changed, but
the device (a mouse) doesn't work. If I enable the autoconfig in the kernel
it works flawlessly without any need for setserial.

/Mikael
-- 
Mikael Nordqvist, student    | d91mn@efd.lth.se | What is GNU Hurd? A whole
Lund Institute of Technology | mech@df.lth.se   | bunch of GNUs running around.

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

From: Oliver Schoett <schoett@informatik.tu-muenchen.de>
Subject: 1.1.16 NFS hangs on write when wsize >1k (Re: 8k NFS client in 1.1.13)
Date: 30 May 1994 21:04:46 GMT

In article <2rr7sb$506@galaxy.ucr.edu> grif@tempest.ucr.edu (Michael
Griffith) writes:

 > The 8k NFS client in Linux 1.1.13 works so nicely, that I wanted
 > to publically thank Alan, Linus, and everyone else involved.  I am
 > seeing 3-4x speedup which is quite nice considering almost everything
 > is NFS mounted.  Great work!

Having waited eagerly for this, I immediately tried 1.1.16 on my 486
DX2-66 VL with 3c509 Ethernet card and the Slackware 1.2.0.3
distribution.  I mount various file systems

   rw,hard,bg,intr,posix,nosuid,rsize=8192,wsize=<read on>

from sun (pre-Solaris), HP, and SNI RM400.  Mounting and reading works
fine, but with wsize=2048, 4096 and 8192 I get hangs on write: For
example, 'cp' and 'iozone' both create an empty target file and then
hang such that only kill -9 will abort them.  This happens the same
way with all three manufacturer's servers.  Wsize=1024 works, but is
slow as ever.

Does anyone have an idea what might be wrong here?

Oliver Schoett  <schoett@informatik.tu-muenchen.de>
Schraudolphstr. 24, 80799 Muenchen, Germany.  Phone +49 89 272 51 58

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

From: jwb@capek.rdt.monash.edu.au (Jim Breen)
Subject: Re: Three-button mouse, again?
Date: 30 May 1994 22:56:14 GMT

s21008@cc.ntnu.edu.tw (DL83-08) writes:

>Thank you to all who wrote about the dual-protocol mouse configuration for
>X. Obviously, my mouse is a bit different from most. The DOS driver has two
>modes: Microsoft and PC mouse. The microsoft mode uses two buttons, and the
>PC mouse uses all three buttons. In my Xconfig file, I can specify the
>Microsoft protocol, of course, and the mouseman protocol. The problem is
>that both use only two buttons. All of the responses to my last request for
>help told me to use mousesystems with the cleardtr statement. Well, I tried
>that, and also tried the clearrts, as well. The problem is that microsystems
>just does not work with this mouse. Is there any other way to force the
>middle button to be recognized in mouseman mode?

I had the same problem. I bought another mouse. Problem solved.
-- 
Jim Breen          [$@%8%`!&%V%j!<%s(J@$@%b%J%7%eBg3X(J]
Department of Robotics & Digital Technology. Monash University. 
Wellington Rd, Clayton VIC 3168 Australia (ph) +61 3 905 3298 
(fax) +61 3 905 3574  AARNet/Internet:j.breen@rdt.monash.edu.au  

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

From: chafey@ecst.csuchico.edu (Chris Hafey)
Subject: Re: NI6510
Date: 30 May 1994 23:32:35 GMT

In article <2scrka$t90@access1.speedway.net>,
Chris Cappuccio <chris@speedway.net> wrote:
>How about a Racal Interlan NI6510 card? these bad? 16-bit.. i dunno if
>it's full duplex or not, and I'm sure theres no linux support for it
>yet.
>

I coded up a driver for it a while back and seems to run ok.  It seems to have
a bug which locks up the kernel, but I haven't spent the time to investigate
it yet.  I plan on working on it a bit more now that school is out, but you
can also try it yourself by hacking Becker's lance.c code, or getting my 
current source.

Chris


-- 
===============================================================================
Chris Hafey                     *  True programming is rebooting the machine
chafey@ecst.csuchico.edu        *  after each crash until it works. 
Check Out My Mosaic Home Page   *  http://www2.ecst.csuchico.edu/~chafey

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

From: hannigan@cs.ualberta.ca (Barry Hannigan; EdTel)
Subject: Re: 1.1.15 breaks SCSI
Date: 30 May 1994 23:38:10 GMT

phr@netcom.com (Paul Rubin) writes:

>>Hmmm. Perhaps slow machines? I have a 386/40. Rene has also a 386, if
>>I remember correctly.

>Nope.  I know someone losing in the same way, with a 1542B and
>a 486dx2/66.

i got hit with a 486/33 and 1542B with 1 disk, 1 tape and 1 cd. all three 
devices are identified before the hang

bgh
--
                        | barry hannigan             | 
117 hunterbasse. what?  |  1960-10303 jasper ave     | death to videodrome
its where i live. nice  |  edmonton, alberta         | long live the new flesh
place? ya i guess.      |  t5j 3n6                   |  

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

From: michael@ecel.uwa.edu.au (Michael Simmons  - division)
Subject: Re: Latest and Greatest networking code
Date: 31 May 1994 02:46:23 GMT

iiitac@uk.ac.swan.pyr (Alan Cox) writes:

>In article <jharkinsCqACvD.HFJ@netcom.com> jharkins@netcom.com (Jim Harkins) writes:
>>Where can I get the latest and greatest networking code?  I've heard of
>>net3, and downloaded a net-0.32d tar file but it only had stuff like
>>netstat  and arp sources in it.  I need the entire tcp/ip suite.  My
>>current source code is net2 dated June 1993.

>The NET3 code is in the current kernel. In addition as of 1.1.13 its a lot
>faster over unreliable links. You'll be delighted to know that apart from
>net032d, bootpc and dosemu no other programs need replacing to go from 1.0
>to 1.1.13. bootpc will even continue to work just wont recompile until you
>get the new one.

OK so where is the NET3 version of net032d!
Do you just recompile arp and netstat?

>'The' archive for net3 code is sunacm.swan.ac.uk:/pub/Linux/Networking. It
>is mirrored on sunsite.unc.edu. 

The sunacm.swan.ac.uk directory appears empty.
The mirror on sunsite.unc.edu has only old README files.

I have looked in nic.funet.fi:/pub/OS/Linux/PEOPLE/Linus directories.
Is it there somewhere. ./net-source/net?

Why can't the kernel developers include a Patchxx.README file that
describes what has changed and what software is directly affected.


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

From: harald@blizzard.oche.de (Harald Schreiber)
Subject: Re: Pausing of program execution (introduced in pl1.1.13)
Date: Tue, 31 May 1994 00:21:44 GMT

mmnuk@risc.uni-linz.ac.at (Michael MNUK) writes:

>I experienced the following phenomenon introduced in patch level 1.1.13:
>(here I refer to two examples: 
>1. "sudo mount -t msdos /dev/fd0 /usr/fd0"
>2. "fvwm")

>The program starts, its entry appears in "ps ax", but after that it
>pauses for about 1 minute before the execution proceeds. The process
>does not consume any ressources during this time. It feels as if it
>had wait(60) at the beginning. This phenomenon persisted to 1.1.15
>(don't know about 1.1.16). Many other programs are not affected by
>this. Till 1.1.12 anything was OK.

>My hardware: DX-66, VLB, 16M+16M swap, Maxtor 7345A, Slackware, libc 4.5.26.

I have the same problem with the PINE mail tool, whereas "fvwm" is not
affected. The phenomenon persists in 1.1.16.

My configuration: ASUS PCI/I-486SP3, 486DX2-66, 16M+16M swap
                  DPT 2001/95 SCSI-Controller, Quantum Prodrive 425 S,
                  Slackware 1.2.0.1, libc 4.5.24


-- 
=============================================================
Harald Schreiber, Ronheider Berg 208, D-52076 Aachen, Germany
Phone: [*49|0]241-79823,      E-mail: harald@blizzard.oche.de
=============================================================

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

From: delman@mipg.upenn.edu (Delman Lee)
Crossposted-To: comp.os.linux.admin
Subject: Can one determine real UID of a process from /proc?
Date: 31 May 1994 03:52:16 GMT

Hi! Was using proc-based ps. Found that proc-ps prints the effective
UIDs of processes. On probing a bit further, there seems to be no way
of determining the real UID of a process from /proc. I guess the
owner of /proc/xxx should be the effective UID of the process for
security reason. Nevertheless, we need a way of knowing the real UID
of the process.

Any suggestions?

Thanks, Delman.
--
______________________________________________________________________

  Delman Lee                                 Tel.: +1-215-662-6780
  Medical Image Processing Group,            Fax.: +1-215-898-9145
  University of Pennsylvania,
  4/F Blockley Hall, 418 Service Drive,                         
  Philadelphia, PA 19104-6021,
  U.S.A..                            Internet: delman@mipg.upenn.edu
______________________________________________________________________

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


** 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
******************************
