Subject: Linux-Development Digest #768
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:     Sat, 28 May 94 03:13:12 EDT

Linux-Development Digest #768, Volume #1         Sat, 28 May 94 03:13:12 EDT

Contents:
  Re: Does linux work in laptop w/ IBM BIOS? (Steve McMahon)
  netstat segmentat fault (with fix) (G. "Wolfe" Woodbury)
  Re: bug on kernel 1.1.15 (Michael Riepe)
  Re: 8k nfs performance (NJ. Bruton)
  FUTURE DOMAIN PROBLEMS (Naresh Sharma)
  Re: 32-bit Novell desktop OS combines Unix, DOS 7 (Mark Evans)
  Seeking NCR 53C810 driver.. (Nathan Yin)
  Re: Linux and binding sockets (Michael Peek)
  1.1.13 (or later) with > 16M? (Andrew R. Tefft)
  bug in net/inet/ipx.c (Laco Rusnak)
  Novell Connectivity (was: 32-bit Novell desktop OS combines Unix, DOS 7) (Stefan Marte)
  Re: Video Blaster (Denny Valliant)
  Re: 32-bit Novell desktop OS combines Unix, (Colin Plumb)
  NCR53C830 (amund@itk.unit.no)
  Re: 1.1.13 (or later) with > 16M? (Leonard N. Zubkoff)
  Kernel change summary 1.1.13 -> 1.1.14 (Russell Nelson)
  DIP 3.3.7a-uri fails with 1.1.15 (Julien SIMON)
  Re: BusLogi 445S and DMA Channel (NEW ANSWER) (Jordan Hubbard)

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

From: steve.mcmahon@lambada.oit.unc.edu (Steve McMahon)
Subject: Re: Does linux work in laptop w/ IBM BIOS?
Date: 27 May 94 20:52:42 GMT

In article <940526145349@lambada> steve.mcmahon@lambada.oit.unc.edu (Steve McMahon) writes:
>I have a Panasonic CF-580C laptop. This is an i486SL laptop and has an
>IBM BIOS. I tried booting MCC 1.0+ on it. It booted fine, but it could
>not recognize the hard drive (fdsik could not read /dev/hda, and the
>partition check message at bootup was not displayed).
>
>The hard drive is listed in CMOS as type 0 (Auto), 984 cylinders, 13
>heads, and 32 sectors per track (200 MB). There is no way to change
>the drive type or info in CMOS.
>

This problem apparently also affect all IBM PS/ValuePoint machines.

I went poking arond the kernel, and I think I've found the reason for
the above behavior. The IBM BIOS uses hard drive type 0 to indicate
Auto. The kernel looks for the hard drive type to determine the number
of hard drives attached. It assumes that type 0 means no drive is
attached (see the code below), otherwise it assumes one or two drives
are attached.

Here is the relevant code from the kernel's hd.c (in drivers/block):


        /*
                We querry CMOS about hard disks : it could be that
                we have a SCSI/ESDI/etc controller that is BIOS
                compatable with ST-506, and thus showing up in our
                BIOS table, but not register compatable, and therefore
                not present in CMOS.

                Furthurmore, we will assume that our ST-506 drives
                <if any> are the primary drives in the system, and
                the ones reflected as drive 1 or 2.

                The first drive is stored in the high nibble of CMOS
                byte 0x12, the second in the low nibble.  This will be
                either a 4 bit drive type or 0xf indicating use byte 0x19
                for an 8 bit type, drive 1, 0x1a for drive 2 in CMOS.

                Needless to say, a non-zero value means we have
                an AT controller hard disk for that drive.


        */

                if ((cmos_disks = CMOS_READ(0x12)) & 0xf0)
                        if (cmos_disks & 0x0f)
                                NR_HD = 2;
                        else
                                NR_HD = 1;
        }
        i = NR_HD;

As you can see, when CMOS_READ returns 0 as the type, NR_HD in the
above code remains zero, and so linux assumes that no drives are
attached.  Having identified the problem, how to fix it? I really
don't know. If we disregard the above block, then how is the kernel to
know the number of drives? Is there another value in CMOS perhaps that
tells the number of drives? If so, then that one can be used instead.
How do DOS, OS/2 and NT do it?

I'm open to suggestions.

-Steve


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

From: ggw@acpub.duke.edu (G. "Wolfe" Woodbury)
Subject: netstat segmentat fault (with fix)
Date: Fri, 27 May 1994 14:19:15 -0500

I've been having occasional problems with the netstat command
giving a segmentation fault and hanging on occasion.  Interrupting
it and saying "netstat -n" produces a complete, but less readable
listing of open sockets.

I went digging into the netstat source code, and discovered that
the "tcp_name[]" array is declared with *only* 10000 entries, and
some of the sockets/ports are given numbers as high as 49000!

The fix is relatively easy.  Declare the array tcp_name[MAX_TCP_NAME]
(instead of with a direct constant), #define MAX_TCP_NAME to a reasonable
size (10000 is actually way too big! :-)  and then around line 148
of netstat.c change:

     if ( tcp_name[socknumber]) {

to

     if ( socknumber < MAX_TCP_NAME && tcp_name[socknumber])


This is with kernel 1.1.0 and net032 networking code.  I heven't yet
looked at the newest net code to see if it is fixed there, but I
suspect it isn't. :-)    [since the netstat.c headers still claim that
its "route.c" :-)  ]

-- 
Gregory G. "Wolfe" Woodbury  @, but not speaking for Duke Univ.
System Admin  Demographic Studies  Box 90408 Durham NC 27708
ggw@cds.duke.edu   ggw@acpub.duke.edu   ggw@wolves.durham.nc.us
"Myth is metaphor, and ritual is the enactment of myth."

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

From: riepe@ifwsn4.ifw.uni-hannover.de (Michael Riepe)
Subject: Re: bug on kernel 1.1.15
Date: Thu, 26 May 1994 22:59:33 GMT

David Barr (barr@pop.psu.edu) wrote:
> In article <2s0hrk$68r@vixen.cso.uiuc.edu>,
> Supat Faarungsang <supat@nuntana.animal.uiuc.edu> wrote:
> >here the bug
> >
> >stty: TCGETS: Operation not supported on socket

> Well gee, that's nice.  You've given us an error message.

> Could you at least tell us what command _generated_ this error
> message?  (obviously an stty command, but _what_ stty function?
> when?  where?)

Is this a joke??? As far as I remember, ioctl(fd, TCGETS, ...) works
only on ttys (i.e.  /dev/tty*) but *not* on sockets - so what's the
problem? ``Don't do that, then!''

Ciao,
-- 
 Michael Riepe               <riepe@ifwsn4.ifw.uni-hannover.de>
 Universit"at Hannover
 Institut f"ur Fertigungstechnik und Spanende Werkzeugmaschinen
 Schlosswender Str. 5                  30159 Hannover (Germany)
 ... beware the storm that gathers here ("The Prophet's Song")

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

From: ccnjb@sun.cse.bris.ac.uk (NJ. Bruton)
Subject: Re: 8k nfs performance
Date: Fri, 27 May 1994 07:43:13 GMT

Alan Cox (iiitac@uk.ac.swan.pyr) wrote:
: In article <2s0q88$fei@galaxy.ucr.edu> cvarner@corsa.ucr.edu (Curtis Varner) writes:
: >     The way these experiments were run was by dd'ing a file from
: >an nfs-mounted directory (size = 1311748 bytes) sending the file to 
: >/dev/null, and timing how long it took to receive the file.  I am in

: Try the other way.. Linux to Sun the other way should be more than 
: doubled on write

: Alan

Does this mean that if I am using a kernel > 1.1.12 I should remove the
rsize=1024 wsize=1024 on my nfs mounted disks ??

Nick

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

From: nash@dutllu4.gmd.de (Naresh Sharma)
Subject: FUTURE DOMAIN PROBLEMS
Reply-To: Naresh.Sharma@LR.TUDelft.NL
Date: Fri, 27 May 1994 07:35:14 GMT

Hi Guys,

Since the time I've installed a future domain TMC1680 scsi-2 card, I've noticed
a couple of messges at boot time:

future domain selection failed

However the card is recognised by the kernel, and it recognizes my two devices,
ie a CD-ROM and a seagate HD. The problem arose when I tried to access the 
cd-rom files, from emacs (or lemacs if you please), the system came down.

All the getty process crashed, and my keyboard locked. Very serious indeed!
I had to hard boot the system and after checking the disk, it rebooted, and
worked fine. Just to check out if the problem was with lemacs/cd-rom I 
started xfilemanager, went to /cdrom and began to fetch files from there.
After a few tries, the system again crashed.

Unfortunately, the system errors cant be posted since each time the system
went down the getty also went down, along with all the tty processess. The 
screen started to scroll with aparent randomness.

Then I added one of the manpath variables to the cd-rom path e.g.
/cdrom/usr/local/man
and removed it from the original manpath. Then I just did a man <some-file-in
the-cd-rom-man> and agian the system went down.

I have a feeling that the problems stem from some fdomain problem.
I saw the same message: "future domain selection failed" yesterday too when I
upgraded to 1.1.15.

Has anyone got any tips?

Naresh


--
_______________________________________________________________________________
Naresh Sharma [N.Sharma@LR.TUDelft.NL]  Snail-Herenpad 28          __|__
Faculty of Aerospace Engineering              2628 AG Delft \_______(_)_______/
T U Delft                     Optimists designed the aeroplane,   !  !  !  
Ph(Work) (+31)15-783992       pessimists designed the parachute!
Ph(Home) (+31)15-569636  Plan:Design Airplanes on Linux the best OS on Earth!
===============================================================================

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

From: evansmp@mb5194.aston.ac.uk (Mark Evans)
Subject: Re: 32-bit Novell desktop OS combines Unix, DOS 7
Date: Fri, 27 May 1994 17:14:51 GMT

Andrew Deckowitz (andydeck@MCS.COM) wrote:

: Thank you Russ, for making a point that I believe many have been ignoring.
: The portion of Netware that Novell is protecting with what is viewed as
: fanaticism is their bread-and-butter, the core operating system and the
: communication protocols (NCP) that define it.  All that is necessary for
: Corsair/Expose (Or for M$'s NT client) is the ability to communicate with
: a Netware server, or if they are actually going to integrate Novell Dos 7
: (a rumor which I personally view as highly unlikely) then perhaps they
: will be adding the Personal Netware code.  This does NOT mean that you
: will be able to take this code and roll your own 1000-user Netware server.

You could, however use this code to produce a server to which other netware
clients could attach :-)

: But it may mean that you will be able to (using PNW-like code) share your
: disk and peripherals between Corsair boxes and perhaps from Corsair to
: other platforms.  I am sure that Novell will not be handing away server
: code.  This does not preclude them releasing source to a Netware client.

By releasing the client you are also releasing details about how the server
and client communicate. That is what matters, all a client see's are 
packets on the ethernet so long as these match the spec it is happy,
whatever is acting as a server.

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

From: nathan@sjsumcs.sjsu.edu (Nathan Yin)
Subject: Seeking NCR 53C810 driver..
Date: 27 May 1994 18:36:57 GMT

[ Article crossposted from comp.os.linux.help ]
[ Author was Nathan Yin ]
[ Posted on 27 May 1994 18:36:21 GMT ]

Hi Netters,

sorry about the double posting...  the news server crashed yesterday and I
wasn't sure if my previous post made it outside :(

anyway, does anyone have any information as to the availability and/or the
possibility of alpha/beta testing for the NCR 53C810 SCSI controller chip for
Linux?  If so, could you please send me the information?

Thank you for your help!

--
Nathan Yin
(408) 924-5106
nathan@sjsumcs.SJSU.EDU
Math & Computer Science Dept.
San Jose State University

--
Nathan Yin
(408) 924-5106
nathan@sjsumcs.SJSU.EDU
Math & Computer Science Dept.
San Jose State University

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

From: peek@duncan.cs.utk.edu (Michael Peek)
Subject: Re: Linux and binding sockets
Date: 27 May 1994 17:22:14 GMT

I may be wrong, but I think ports 6000+ are reserved for displays.  Port
6000 is actually the same as DISPLAY=<machine>:0, and 6001 is ...:1
...etc.

I'm not too familiar with this kind of stuff, so I could be full of it.

Michael Peek - peek@math.utk.edu


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

From: teffta@cs690-3.erie.ge.com (Andrew R. Tefft)
Subject: 1.1.13 (or later) with > 16M?
Date: Fri, 27 May 1994 18:41:13 GMT

I have been using 1.0 and upgraded to 1.1.13 in order to get caught
up on the ppp stuff.

I was unable to boot, with a crash coming after the "honours the
WP bit" message. I found that I could boot if I limit the kernel
to the lower 16M.

I know Linus himself has 20M, but on his suggestion I tried changing
the 16M limit and every setting over 16M crashed in the same way.
I applied patch14 but it was no improvement.

I am able to run if I disable the external cache but I don't
like that too much.

So if you have more than 16M ram and are using 1.1.13 or later,
please let me know. If you haven't tried 1.1.13 or later and are
willing to, give me a try and let me know if it works. Surely
I can't be the only one with this problem! Hopefully if it is
more common we can come up with a fix.
-- 

Andy Tefft               - new, expanded .sig -     teffta@cs690-3.erie.ge.com

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

From: laco@Viktoria.drp.fmph.uniba.sk (Laco Rusnak)
Subject: bug in net/inet/ipx.c
Date: Thu, 26 May 1994 14:24:57 GMT

I think, that I'm found a bug in kernel implementation of IPX protocol.
When I tried to delete IPX route when routing table was empty, I got
general protection exception in kernel task. In the module net/inet/ipx.c
in function ipxrtr_delete() is not handled NULL pointer ipx_router_list.
Here is the location of the fault (cmp instruction):

static int ipxrtr_delete(long net)
{
        ipx_route *r=ipx_router_list;
==>     if(r->net==net)                        <===
        {
                ipx_router_list=r->next;
                kfree_s(r,sizeof(ipx_route));
                return 0;

But this story is not finished. After this cames second exception and 
the kernel has crashed. Here is the location of the crash 
(cmp instruction again):

ipxrtr_get_local_net(struct device *dev, unsigned short datalink)
{
        ipx_route *r;
        unsigned long flags;
        save_flags(flags);
        cli();
        r=ipx_localnet_list;
        while(r!=NULL)
        {
==>             if((r->dev==dev) && (r->dlink_type == datalink))    <==
                {
                        restore_flags(flags);
                        return r;

Hmmm....

Bye, Laco.


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

From: MARTE@htu.tu-graz.ac.at (Stefan Marte)
Subject: Novell Connectivity (was: 32-bit Novell desktop OS combines Unix, DOS 7)
Date: Fri, 27 May 1994 00:56:03 GMT

>Russ - NCP the layer the client would need to talk is a Novell trade secret
>and fairly well guarded. If they give out a client source it'll only be a
>month before someone knocks out a clone server (at least for file services)
>as it is it might take 6 months to deduce from documents and ethernet traces

True,true, this is the reason, why some friends and i have set upon the task 
of writing an NLM (Netware loadable module) and a filesystem for linux, to 
provide some direct connectivity from linux to netware, because neither NFS 
nor the DOSEMu method provide a satisfying solution to that problem.

if someone wants info, has comments flames or help to offer, pleas email me.

and no, i won't give out sources or binaries for now, the current stuff 
wouldn't help you much anyway.

cu
STM
----
Stefan Marte: marte@htu.tu-graz.ac.at

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

From: valliant@unm.edu (Denny Valliant)
Subject: Re: Video Blaster
Date: 26 May 1994 18:48:52 -0600

In article <1994May26.135147.13736@uk.ac.swan.pyr>,
Alan Cox <iiitac@uk.ac.swan.pyr> wrote:
>In article <k5Vwmc1w165w@bentek.mese.com> bra@bentek.mese.com (Ben Adams) writes:
>>valliant@unm.edu (Denny Valliant) writes:
>>
>>>     Has anyone written anything for the video blaster.  I have drivers for
>>> DOS, but have not been able to find anything for Linux.  When I try to run th
>>> driver in DosEmu, it tells me that it is already installed.  If anyone has se
>>> anything out there, please mail me.
>>
>>It would seem that a interface for video overlay cards should be 
>>setup.  Then the video blaster and other video overlay cards could
>>have drivers developed.   BTW I have a Video blaster also...
>
>Nothing stops you writing a video blaster driver now - assuming you can 
>obtain suitable programming specifications. It'll probably need to work as
>a character device which can be read to read the frame buffer, possibly written
>and has ioctls for moving the 'overlay' window around. It probably is a good
>idea to make the ioctl() calls generic in case someone wants to do a driver
>for a different overlay card. All you really then need as an X application
>that includes a blank canvas to drop the picture into and which when moved
>by the window manager moves the tv picture to follow it.
>
>Alan
>
>
>

        But there is no specs for it on the net anywhere.  If you find any,
please e-mail them to me.  The application is not hard to write, but it is
very hard to find the specs.

                                --Danny (valliant@leo.unm.edu)

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

From: colin@nyx10.cs.du.edu (Colin Plumb)
Subject: Re: 32-bit Novell desktop OS combines Unix,
Date: 27 May 1994 15:25:06 -0600

i(Sorry to post this, but its hard to reply...)

In article <Cq930H.FKw@info.swan.ac.uk>,
Alan Cox <iialan@myhost.subdomain.domain> wrote:
> Newsgroups: comp.os.linux.development
> Path: mnemosyne.cs.du.edu!tali.hsc.colorado.edu!boulder!agate!howland.reston.ans.net!EU.net!uknet!info!myhost.subdomain.domain!iialan
> From: iialan@myhost.subdomain.domain (Alan Cox)
> Subject: Re: 32-bit Novell desktop OS combines Unix,
> Message-ID: <Cq930H.FKw@info.swan.ac.uk>
> Sender: news@info.swan.ac.uk
> Nntp-Posting-Host: iifeak.swan.ac.uk
> Organization: To be avoided
> X-Newsreader: TIN [version 1.2 PL2]
> References: <2rifin$6kd@louie.udel.edu> <1994May20.173023.26005@eurom.rhein-main.de> <1994May20.235957.1894@taylor.wyvern.com> <rob.150.02644292@eats.com>
> Date: Mon, 23 May 1994 10:12:16 GMT
> Lines: 12

I'm pretty sure that "myhost.subdomain.domain" is not a valid domain
name.  While it could theoretically be registered, because it's
so obviously an example domain name it's probably a bad idea.

It looks like someone, somewhere, forgot to update a config file.
Alan, if you want to get mail, you might want to look into that.
-- 
        -Colin

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

From: amund@itk.unit.no
Subject: NCR53C830
Date: 27 May 1994 08:16:01 GMT
Reply-To: amund@itk.unit.no

Do anyone have any information about the availability
of a driver for NCR53C830 SCSI chip used for PCI
SCSI controllers ?
(Or enough information about it to enable a mere mortal to
write one ? )

========================================================
========================================================
siv.ing Amund Skavhaug
Division of engineering cybernetics
The Norwegian institute of technology
email: amund@itk.unit.no
phone: 47-73594390(work) 47-73515758
========================================================
**************************************************


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

From: Leonard N. Zubkoff <lnz@dandelion.com>
Subject: Re: 1.1.13 (or later) with > 16M?
Date: Fri, 27 May 94 21:09:37 GMT

In article <CqH58q.DH2@cs690-3.erie.ge.com> teffta@cs690-3.erie.ge.com (Andrew R. Tefft) writes:

   So if you have more than 16M ram and are using 1.1.13 or later,
   please let me know. If you haven't tried 1.1.13 or later and are
   willing to, give me a try and let me know if it works. Surely
   I can't be the only one with this problem! Hopefully if it is
   more common we can come up with a fix.

I am happily running 1.1.16 on a 90MHz Pentium PCI/VLB system with 32mb memory.

                Leonard

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

From: nelson@crynwr.crynwr.com (Russell Nelson)
Subject: Kernel change summary 1.1.13 -> 1.1.14
Date: 27 May 1994 19:03:23 GMT

Add missing TCP/IP options to config.in
Split out program binary formats as part of ibcs project (fewer ifdefs needed).
Make PPP support a standard option.
Change Mitsumi IRQ into a variable.
The size of the /proc/net/* files can become smaller, so deal with
        negative counts.
Signals are now mapped through a table.

--
-russ <nelson@crynwr.com>      ftp.msen.com:pub/vendor/crynwr/crynwr.wav
Crynwr Software   | Crynwr Software sells packet driver support | ask4 PGP key
11 Grant St.      | +1 315 268 1925 (9201 FAX)    | Quakers do it in the light
Potsdam, NY 13676 | LPF member - ask me about the harm software patents do.

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

From: simon@masi.ibp.fr (Julien SIMON)
Subject: DIP 3.3.7a-uri fails with 1.1.15
Date: 27 May 1994 19:28:29 GMT

Hi !

Since 1.1.13 broke DIP 3.3.7, I upgraded to 1.1.15 and to DIP 3.3.7a-uri.
But no cigar ! It still doesn't work ...
DIP doesn't even print the usual message (version number and so on).
It just goes into an infinite loop.

However, it works perfectly fine with 1.1.12.

Any clue ?

        Julien
 


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

From: jkh@nx.ilo.dec.com (Jordan Hubbard)
Crossposted-To: comp.os.linux.misc,comp.os.386bsd.development,comp.os.386bsd.questions,comp.sys.ibm.pc.hardware.systems,comp.sys.ibm.pc.hardware.storage,comp.sys.ibm.pc.hardware.misc,comp.periphs.scsi
Subject: Re: BusLogi 445S and DMA Channel (NEW ANSWER)
Date: 26 May 1994 23:12:18 GMT

In article <2s2j80$1go@ebh.eb.ele.tue.nl> wjw@ebh.eb.ele.tue.nl (Willem Jan Withagen) writes:

   Because this starts me wonder which kernel I should use with my 445S REV D
   board under FreeBSD. I'm using the AHA kernel which works great for 16Mb
   But I'm waiting for the 3.37 Eprom to go to 32MB. Should I use the BT 
   kernels for FreeBSD

You can use the Bt driver *now* for your Rev D board with 16MB - I do
this now.  If we can get the bounce-buffering working for 1.1.5,
you'll also be able to go to 32MB with it.  Whether or not the 3.37
eprom fixes the >16MB DMA problem is, however, still an open question
for ANY of the operating systems.  Buslogic claims success with 3.37,
yet I've now seen more than a few reports that tend to indicate that
they did NOT fix the problem.  If this is the case, then the worst
that will happen with your FreeBSD 1.1.5 system is that it'll simply
"bounce" DMA >16MB and be a very small degree slower, but it'll work.
You'll know whether or not we got the bounce-buffer stuff working
properly in 1.1.5 when you see the release notes!

                                        Jordan

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


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