Subject: Linux-Development Digest #35
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:     Wed, 17 Aug 94 06:13:04 EDT

Linux-Development Digest #35, Volume #2          Wed, 17 Aug 94 06:13:04 EDT

Contents:
  pfopen (Kent A Vander Velden)
  Help linking curses in C++ (na8530b00-Rivera)
  Re: Pentium version of gcc? (Greg W. Zoller)
  Re: 1.1.45 NCR 53c810 problems (Croix)
  NFS errors (David Fox)
  kernel 1.1.44/45 and serial drivers..... (Jaime Jofre)
  Re: NFS errors (David Fox)
  New floppy code breaks notebook (Gary Paul Gortmaker)
  1.1.45 floppy bug (Frank Lofaro)
  Re: GP faults (Nicolas BOUGUES)
  Re: SMC3000 ethernet card driver? (Stan Young)
  Re: Floppy serious errors in 1.1.42 (with buslogic scsi) (Dick Streefland)
  Re: Adaptec AHA 2940 (Jeff Redding)
  Re: Linux NFS client performance (Baba Buehler)
  Re: Future of Linux (Michael Dillon)
  'Kernal panic' errors when deleteing!! (Steve Smith, Logica UK)

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

From: graphix@iastate.edu (Kent A Vander Velden)
Subject: pfopen
Date: 16 Aug 94 21:29:25 GMT

I will not pretend to be a great network programmer but why does
Linux not have a pfopen call.  Maybe someone can reccomend a more
portable way of accessing the ethernet device in a raw mannor...

Am I the only person who sees the network include files i.e. those in
/usr/include/net and /usr/include/netinet to be rather incomplete?  This
might be the fault of Slackware 1.2.0 however.

-- 
Kent Vander Velden
graphix@iastate.edu


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

From: miguel@bss80.cb.att.com (na8530b00-Rivera)
Subject: Help linking curses in C++
Date: Tue, 16 Aug 1994 20:56:56 GMT

I am having trouble with curses/ncurses and g++. I tried to use the CursesWindow class 
and when I compile the program it show all the curses symbols undefined in the
g++ library. It seems that the linker is not picking up the symbols from the
curses library. BTW, I am including the curses library in the g++ command line.

Any comments?? Any sugestions??? Any help will be appreciated.

Thanks,
Miguel Rivera
miguel@ieain.ih.att.com


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

From: gwz@n8pph52.nt.com (Greg W. Zoller)
Subject: Re: Pentium version of gcc?
Date: Tue, 16 Aug 1994 19:06:42 GMT

I think a Pentium-optimized gcc is a wonderful idea!  I
also agree that GNU is in a much better position to
incorporate the Pentium capability.  I hope they are
actively working on it.  :-)

A Pentium-kernel with Pentium X-Windows/apps, etc.
should move along at a very nice clip.

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

From: jcroix@rx7t.austin.ibm.com (Croix)
Subject: Re: 1.1.45 NCR 53c810 problems
Date: Tue, 16 Aug 1994 22:07:23 GMT

In article <lma.777056949@dayton.Stanford.EDU> lma@dayton.Stanford.EDU (Larry Augustin) writes:

   Path: awdprime.austin.ibm.com!portal.austin.ibm.com!geraldo.cc.utexas.edu!cs.utexas.edu!howland.reston.ans.net!vixen.cso.uiuc.edu!sdd.hp.com!hplabs!unix.sri.com!headwall.Stanford.EDU!dayton.Stanford.EDU!lma
   From: lma@dayton.Stanford.EDU (Larry Augustin)
   Newsgroups: comp.os.linux.development
   Date: 16 Aug 94 17:09:09 GMT
   Organization: Stanford University
   Lines: 31
   NNTP-Posting-Host: dayton.stanford.edu

   These are both with the 1.1.45 kernel.

   NextStor NCR 53c810 PCI SCSI card, Micronics P54i PCI motherboard:
       scsi0: using io mapped access
       scsi0: using initiator id 7
       scsi0: using edge triggered interrupts
       scsi0: burst length 8
       scsi0: using 40MHz SCSI clock
       scsi0: m_to_n = 0x90, n_to_m = 0x90, n_to_n = 0xb0
       scsi0: testing
       scsi0: test 1 started
       scsi0: tests complete
       scsi0: NCR53c{7,8}xx (rel 3)
       scsi: 1 hosts.
       scsi0: DANGER: command in running list, can not abort
       scsi0: DANGER: command in running list, can not abort
       SCSI host 0 abort() timed out - reseting
       Unable to reset scsi host 0 - probably a scsi bus hang

   ASUS SP3G, on-board NCR 53c810:
       system hangs with floppy read light on immediately after detecting
       the 53c810 and determining its io, irq, etc.

   BTW, I've sent 2 email messages to Drew Eckhardt offering to loan him
   a 53c825 board, but haven't heard any response.  Does anyone know if
   he's still working on the drivers?  Alternatively, is there someone
   else working on the drivers who could add support if they had a board
   to test?

   Larry

I had the same problem with an Intel P590 motherboard.  Without going into
the long story, the problem I had (which may not be yours) was that the new
BIOS I flashed had the NCR BIOS in it.  Since the NexStor has BIOS on the
board, I had a conflict.  By jumping pins 1 and 2 of J3 (I think those are
the right numbers), I was able to get over this problem.

BTW, be sure that the card is in a PCI bus mastering compliant slot.  The
NexStor support people told me it had to be in the shared ISA/PCI slot.

The problem with this fix is that the DOS and OS/2 device drivers no longer
work.  You'll need to get the CAM drivers for the NCR chip.  From other posts
on the net, this may or may not work for you.  My eventual solution was to
get an Adaptec 2940 card (but that's another story....).

===================================+===========================================
John Croix                         | My opinions are my own and probably aren't
                                   | those of any reasonably sane person.
===================================+===========================================

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

From: fox@graphics.cs.nyu.edu (David Fox)
Subject: NFS errors
Date: 16 Aug 1994 21:04:25 GMT

I'm getting bad data over NFS using a 3c509.  For example,
here the current directory is mounted from a Sun:

% for i in x x x x x x x x x x x x; do sum gdb-4.13.tar.gz; done
49687  2384
08072  4035
15509  4035
08072  4035
24327  4035
08165  4035
08072  4035
03008  4035
32685  4035
61177  4035
47956  4035
34628  4035
% 

Is this normal for the current Linux network software?  Or do
I have a hardware problem?  I'm running 1.1.45 with the TCP/IP
from the slackware-2.0.0 distribution.
--
David Fox                                               xoF divaD
NYU Media Research Lab                     baL hcraeseR aideM UYN

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

From: jjofre@ritz.mordor.com (Jaime Jofre)
Subject: kernel 1.1.44/45 and serial drivers.....
Date: Wed, 17 Aug 1994 00:17:50 GMT

Has anyone noticed any performance degradation in the serial
drivers with kernel version 1.1.44/45?  I've been using kernel
1.0.9 for a long time now and tunning Mosaic with term compiled
in over a 14.4k modem with very nice results (~1500 cps)  But with
kernels 1.1.44/45 I noticed that my term connections dropped in
performance (~400 cps)  Does anyone have any ideas as to why?  Thanks
in advanced,

Jaime.



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

From: fox@graphics.cs.nyu.edu (David Fox)
Subject: Re: NFS errors
Date: 16 Aug 1994 21:18:13 GMT

In article <FOX.94Aug16170426@first.cs.nyu.edu> fox@graphics.cs.nyu.edu (David Fox) writes:

] I'm getting bad data over NFS using a 3c509.  For example,
] here the current directory is mounted from a Sun:
]   [...]

Following up on my own post, what surprises me is that I don't
get parity errors or something from a higher level of the network
protocol.  How is this possible?
--
David Fox                                               xoF divaD
NYU Media Research Lab                     baL hcraeseR aideM UYN

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

From: gpg109@huxley.anu.edu.au (Gary Paul Gortmaker)
Subject: New floppy code breaks notebook
Date: 17 Aug 1994 12:02:01 +1000


        This is a generic 486 notebook with 2 PCMCIA, a wd90c24 video
chipset and a built in 1.44MB floppy. Kernels 1.1.41 and beyond
detect the chip as an 8272A. Kernel 1.1.40 is the last kernel that 
will work with the floppy, and it works fine.

All operations such as mount, mke2fs, mdir fail after initializing the
drive. They never seem to move the heads, as the drive is quiet.
The only command that will move the heads is "fdformat". The errors
are usually reported as I/O errors of different sorts.

        Four other systems (2 * 8272A, 2 * post-1991 82077) seem to
work ok with the new code.

        I can run strace on mke2fs, mount, etc. with both 1.1.40 and
1.1.45 and post the relevant parts of the output here if that will
help anyone. (Alain ?)

Paul.

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

From: ftlofaro@unlv.edu (Frank Lofaro)
Subject: 1.1.45 floppy bug
Date: Wed, 17 Aug 94 04:20:10 GMT

There is a bug in the 1.1.45 floppy driver.
It does not like requests <> 1K it seems. (At least 8k blows up).
Such as swapping (I know it is a weird idea, but possibly useful under 
some conditions, plus it illustrates the problem I want to show).

33: 21:12:46: /: whitney# mkswap -c /dev/fd0H1440 1440
VFS: Disk change detected on device 2/28
Setting up swapspace, size = 1470464 bytes

[ I took out and reinsterted the floppy here ]

34: 21:18:27: /: whitney# swapon /dev/fd0H1440
floppy: disk absent or changed during operation
floppy I/O error
dev 021C, sector 0
Unable to find swap-space signature
swapon: /dev/fd0H1440: Invalid argument
35: 21:19:04: /: whitney# mdir
VFS: Disk change detected on device 2/28
Probable non-MSDOS disk
mdir: Cannot initialize 'A:'
36: 21:19:08: /: whitney# swapon /dev/fd0H1440
floppy: bh=null in copy buffer before copy
floppy: weirdness: remaining 8
Unable to find swap-space signature
swapon: /dev/fd0H1440: Invalid argument
37: 21:19:10: /: whitney#

Not good.

Any idea why the driver gets cranky about it?
I could fix it probably, if I knew my way around the fd driver. Its tricky!



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

From: nicolas@magix.uucp (Nicolas BOUGUES)
Subject: Re: GP faults
Date: 16 Aug 1994 19:33:21 +0200

Juha Laiho (jlaiho@ichaos.nullnet.fi) a ecrit:
: nicolas@magix.uucp (Nicolas BOUGUES) said:
: >I got it from the memory info during boot process. Here is the line I get
: >with 1.1.44 :
: >Memory: 15900/20480k available (728k kernel code, 384k reserved, 3468k data)
: >
: >And with 1.1.39 (with the same kernel config) :
: >Memory: 18248/20480k available (696k kernel code, 384k reserved, 1152k data)
: >
: >I do not look at it carefully each time I boot up Linux, but as far as I
: >remember, it has always looked like (more or less) 1.1.39.

: EEK! Those figures are pretty horrible; they'd practically kill a smaller
: machine. Below are mine:
: Memory: 6952k/8192k available (548k kernel code, 384k reserved, 308k data)

: Kernel is 1.1.42, compiled to include normal HD support, SCSI support
: (AHA-1542, SCSI disk, tape, CDROM, verbose SCSI error reporting), e2fs,
: procfs, iso9660 fs, networking (CSLIP, PPP, 3c505), printer support,
: selection, COFF/ELF loaders, SysV IPC and SoundBlaster support, so I'd
: think it's not a minimal kernel.

: There's couple of options I've seen that may allocate pretty large amounts
: of data space: 'kernel profiling support' and ramdisk. Both of them are of
: no use in normal operation.
: -- 
: Wolf  a.k.a.  Juha Laiho     Helsinki, Finland
: (Geek Code 1.0.1) GCS d? p c++ l++ u(-) e+ m+ s+/- n- h(*) f(?) !g w+ t- r y+
: "...cancel my subscription to the resurrection!" (Jim Morrison)

Now I have installed 1.1.45, and the problem doesn't solve. I am trying to
remove most of my drivers. Currently, I get almost the same result (+- 100k)
with the following options:

math off, IDE on, XT off, Net off, 16 Meg limit off, IPC on, Buslogic SCSI
w/ disk & CDROM, no net dev (since net is off), Sony off, Mitsu off, sbpcd on, 
fs: ext2 dos iso9660 proc, print on, no special mouse, selection on, no
sound, no kernel profile, verbose SCSI on, no ramdisk.

I'm still investigating. I'll post as soon as I find the problem.

--
=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
Nicolas BOUGUES
nbougues@renux.frmug.fr.net
Sysop of magix : ++ 33 (1) 45 21 02 52 (shell & uucp)

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

From: syoung@pecanpi.atl.ga.us (Stan Young)
Subject: Re: SMC3000 ethernet card driver?
Date: Tue, 16 Aug 1994 21:30:07 GMT

In article <CuKxrA.3GF@sun10.sep.bnl.gov>,
Aram Mirzadeh <mirzadeh@pmdb.sep.bnl.gov> wrote:
>Brian A. Lind (snicker@netcom.com) wrote:
>: I have an SMC3000 ethernet card in my PC at work and would
>: like to get Linux working and on the Network. I haven't
>: found this card listed in any of the HOWTOs and have been
>: unable to get it to work.
<snip>
>: I would be a very greatful Beta tester if anyone has a driver.
>: I don't know anything about this card and the PC hardware
>: guy couldn't find me any specs.
>
>I think you're mistaken the card with the driver.  The card is SMC 3812, 
>with a Western Digital Compt. chip.  And Linux support the SMC card by itself.
>It is auto dectected.
>
>The driver for dos, and os/2 for the above card is called a SMC3000.

Well, there certainly is an SMC3016 - and the driver for it under DOS and
other OS's was the SMC3016.  You might be right about the 3812 - but it
sure doesn't stack up with my example. :-)
-- 
==============================+=======================================+
Stan Young                    | Serendipity is looking in a haystack  |
syoung@pecanpi.atl.ga.us      | for a needle and discovering the      |
                              | farmer's daughter.  - Julias H. Comroe|

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

Crossposted-To: comp.os.linux.admin
From: dicks@tasking.nl (Dick Streefland)
Subject: Re: Floppy serious errors in 1.1.42 (with buslogic scsi)
Date: Tue, 16 Aug 1994 17:24:11 GMT

Angelo Haritsis (ah@doc.ic.ac.uk) wrote:
--> I get serious errors when I use the floppy on my linux pc with a Buslogic 
--> 445S VLB scsi controller (the on-board tape is disabled).  See below for
--> more details of my setup.
--> With linux 1.1.25 I had complete crashes in some cases that I accessed the
--> floppy. This was a complete system lock with no panic messages.

I had a similar problem with my BT445S. When I was reading large
amounts of data from a floppy (tar or dd), while simultaniously
writing to or reading from the hard disk, the BusLogic would
lock up during the hard disk access with no error messages. The
LED on the controller remained on and the system was dead. It
made no difference whether I used the floppy controller on the
BusLogic or the floppy controller on a multi I/O card.

I had some other problems with my new system as well, and I got
a replacement motherboard. With the new motherboard the floppy
problem also disappeared! It seems that not every VLB adapter
card works in every VLB motherboard.

--> This is a rather nasty problem. Could you cast some light on this?
--> Could it be connected with the BIOS settings (bus timing etc.)?

The ISA bus speed should be 8 MHz or slightly larger. I use
(and used on the previous board) 33 MHz/4 = 8.25 MHz. Perhaps
it is some kind of timing problem. It might also be a firmware
problem of the BT445S. I have a board revision C, firmware
version 3.35, BIOS version 4.70.

--
Dick Streefland              ////         Tasking Software BV
dicks@tasking.nl            (@ @)             The Netherlands
========================oOO==(_)==OOo========================

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

From: dt10596@renoir.cftnet.com (Jeff Redding)
Subject: Re: Adaptec AHA 2940
Date: Wed, 17 Aug 1994 02:16:19 GMT

theboss@cs.tu-berlin.de (Alexander Nareyek) writes:

>I'm looking for a driver for the Adaptec AHA 2940.

If your the hardy type, you might want to take a crack at converting
the aha274x driver.  I've been working on the aha284x driver, and I
know the 294x is AIC7770 based too.  It shouldn't be a big deal
to hack the one you need.

You can get the aha274x driver @ ftp.cpsc.ucalgary.ca, in
/pub/systems/linux/aha274x

There are other people who are converting the 274x to the 284x and they
will most probably beat me to making it public, but you can mail me
if you have any questions.  (Who knows, I might even be able to answer
them.  Stranger things have happened.)

Anyway, good luck, I've been waiting/working for months to get Linux
up!

===========================================================================
Jeff Redding - dt10596@renoir.com
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
"The difference between a comedian and a congressman is that when a 
   comedian says something funny, it's called a joke; when a congressman
   says something funny, it's called a law." - Will Rogers
===========================================================================


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

From: baba@ph-meter.beckman.uiuc.edu (Baba Buehler)
Subject: Re: Linux NFS client performance
Date: 17 Aug 94 06:35:14 GMT
Reply-To: baba@beckman.uiuc.edu

carlsc@medtronic.COM (Charles V. Carlson) writes:

>Dan Pop (danpop@cernapo.cern.ch) wrote:
>: NFS performance on reading:  70 KB/s
>: NFS performance on writing:  25 KB/s

>These numbers jive with what I get off of DOS NFS clients to Sun server.
>(doesn't seem to matter to much what model).
>Remember, NFS writes are synchronous, which means the data has to be
>committed to disk before the client can continue.  This is what gives
>NFS its stateless nature so servers can go up and down with low probabilty
>of clients losing data.
>At work, everytime one of the Novel 3.12 servers goes down, 250 people
>have to reboot.  Everytime one of the Unix servers goes down ( less 
>frequent than the novell :) ) people just have to hang tight for a few
>minutes.

We have had good expierence patching our Sun Sparc systems for async
NFS.  We noticed about a double in read performance and a tenfold increase
in write performance from PC's mounting off the Suns.

-- 
%>- Baba Z Buehler
%>- Beckman Institute Systems Services, Urbana Illinois
%>- WWW: http://www.beckman.uiuc.edu/groups/biss/people/baba/
%>- PGP Public Key available via WWW & finger baba@bekcman.uiuc.edu

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

From: mpdillon@halcyon.com (Michael Dillon)
Subject: Re: Future of Linux
Date: Tue, 16 Aug 1994 20:05:45 +0000

> The big software companies have spoiled users into having these
> "everything including the kitchen sink" applications that have been developed
> and refined for 10 years or more. If you take a survey of applications
> users you'll find that the 90/10 rule applies in that most (90%) of the
> users use only a small (10%) of the available functions. But because the
> customer is paying they can demand the other 90% of the functions even though
> they are rarely used by any but the most rare user.
> 
> The challenge is to identify the 10% of functions that are most used and
> implement them. Add hooks to that users can add anything they want. 

If you truly want USERS to add these additional features then you have to
provide an easier way for them to do so than writing  C code. One
way that has been proven to work is to use TCL as the glue that binds
together your application's features. That way, users can truly
reconstruct an application in their own image. Even the commercial
WP and spreadsheet companies include macro languages.

I would like to see a continuation of the UNIX "toolkit" approach
where people build tools that do a fairly specialised job but that
are easily combinable with other tools to do fairly complex jobs.
For the most part, such tools have been "filters" that are
connected together with pipes. 

I think that in order to create the possibility of sophisticated
desktop publishing systems like Word for Windows and Excel,
we have to take a slightly different approach that involves
building modules that can be glued together into applications
with something like TCL. This is very similar to the "widget"
concept in X-Windows, except that it does not require the end
user to do any C programming. The TCL approach also focusses
on functionality rather than "looks".

> Linux has no need to compete with DOS/Windows/MacOS. The user bases of the
> others are so large and so entrenched that trying to make even the smallest
> dent is impossible. All we really have to do is provide a high quality
> product whose worth is easily identifiable to knowledgeable users.

And one way to do this is to make something that can do stuff that
is very difficult to do with MacOs or Windows. Even the idea of
filters is very foreign to those OSes and if you think that OLE
will be accessible to any but the most skilled programmers, then
you haven't looked at MS's OLE 2 documentation.

> I've brought Linux in its current form to everywhere I operate. People have
> been impressed with it's strengths and not so worried about it's lack of
> applications because of the DOS/Windows/MacOS environments that are
> prevalent for doing the types of applications you describe.

To me one of the worst things you can do is to run an X server on your
primary Linux box. However, if you have a Windows box as well, you
can run a free demo X server for Windows and get the best of both
worlds.

cruisin' down the information highway, lookin' for a blast
breakin' all the speed limits as I come zoomin' past!
--
Michael Dillon                 Internet: mpdillon@halcyon.halcyon.com
C-4 Powerhouse                  Fidonet: 1:353/350
RR #2 Armstrong, BC  V0E 1B0      Voice: +1-604-546-8022
Canada                              BBS: +1-604-546-2705

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

Date: Wed, 17 Aug 1994 10:45:05 EST
From: Steve Smith, Logica UK <STSMITH@ESOC.BITNET>
Crossposted-To: comp.os.linux.help
Subject: 'Kernal panic' errors when deleteing!!

Since kernal version 1.1.30 (approximately - I can't remember exactely) I
have a very worrying and annoying error whenever I try and delete directory
trees using 'rm -r'.  The error which gets reported is:

Kernal panic: EXT2-fs panic (device 3/2): ext2-find-entry: buffer head pointer
is NULL

This error is repeatable, i.e. happens every time I try the command.  After
this the machine locks up, I can chance virtual console but not login or
do anything else.  Copies of directory trees work OK as do large tar
extractions, it's just the deleting.  Any ideas?  Also after pressing reset the
machine does its memory count and then no more, just stops, I have to
power on/off to get things going again.  Possibly the disk controller has
locked up?  The machine is a 486/33 clone with a Promise DC-200 ISA bus cache
disk controller (4Meg h/w cache).

This happens even with the 1.1.45 I compiled last night.  I have not enabled
the multisector mode.

Help please!


   Steve Smith,  STSMITH@ESOC.BITNET (please reply direct also)


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


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