Subject: Linux-Development Digest #968
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, 31 Jul 94 12:13:06 EDT

Linux-Development Digest #968, Volume #1         Sun, 31 Jul 94 12:13:06 EDT

Contents:
  fdpatch reliability (York Lam - ACPS/F93)
  Re: SLIP hangs in newer kernels (David A. Ranch)
  Re: Linux 1.1.35 and ld -qmagic (Christopher Shaulis)
  Re: APC UPS owners or potential buyers, trying to show user base (Brandon S. Allbery)
  console driver does not reset char table with "stty sane" (Alex Ramos)
  Re: kernel stupidity... (Sander van Malssen)
  Re: Subnet masks (Terry Dawson)
  Re: Panasonic CD-ROM broke under >1.1.31 (Starcon SysAdmin)
  Re: threads in kernel (Steven S. Dick)
  Re: Converting a *future* time in another timezone to local time ... (Paul Eggert)
  [Newbie] Docs on (network) kernel/driver ... Anyone? (Kenneth Wong)
  PCI IDE.... (Morton Jonuschat)
  Re: link() system call (Tim Smith)
  Large malloc problems (Kim Hendrikse)
  Re: Large malloc problems (Jonathan D Albert)
  Networking very slow (john@gab.unt.edu)
  Re: Fix for problems with sound when system is idle. (Dragon Slayer)
  Re: 1.1.35 make problem: exhausted memory (Dan Wilder)
  Re: Panasonic CD-ROM broke under >1.1.31 (Johannes Stille)

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

From: ylam@acs.ryerson.ca (York Lam - ACPS/F93)
Subject: fdpatch reliability
Date: 31 Jul 1994 01:42:45 GMT

To those that use the fdpatches for "higher capacitied" floppies, how are
you all finding it's reliability?  I've been using fdread/fdformat under
MSDOS for years and have *never* had any problems with any drives or disks
when going over the "standard" 1.44MB format.  With Linux, the fdpatches
works just fine as well.  It is really a pain to fix floppy.c whenever it's
changed just so that 1.72MB's can be read again.  This has been going on
even before 0.99.10.
Isn't time to "officially" incorporate this into the kernel?  I'd like to
hear from someone that has a floppy drive that ain't capable of doing
the extra tracks/sectors.  Making the "extra tracks" a kernel compile time
option is another way of keeping the 1.44MB-only around for those that can't
use it or don't want it.

--
  ==========================================================================
 |  York Lam <-=-> ylam@acs.ryerson.ca  | It's only fun until someone loses |
  ======================================| an eye... then it's fun for one   |
 |           ( free-space )             | less person, and NOW it becomes   |
 |                                      | a sport!                          |
  ==========================================================================
                                                                          ---

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

From: dranch@ecst.csuchico.edu (David A. Ranch)
Subject: Re: SLIP hangs in newer kernels
Date: 30 Jul 1994 08:15:20 GMT

In article <311gh4$879@nz12.rz.uni-karlsruhe.de>,
Klaus Schneider <uk0q@rzstud1.rz.uni-karlsruhe.de> wrote:

>Since about kernel 1.1.25 my SLIP connections hangs for a few seconds
>or even minutes before resuming its work.  This still happens in
>v1.1.35.  Has anybody else experienced a similar problem?


Klaus,

Have you tried to re-compile DIP recently?  I had a similar problem
but a re-compile helped LOTS!  You also might want to check out
DIP 3.3.7-uri.  It has more features and seems more stable on my
Linux box.


-- 
 /------------------------------------------------------------------------\
 |  David A. Ranch [CMPE Major]                 dranch@ecst.csuchico.edu  |
 |                                             sysop@trinity.chico.ca.us  |
 |                                                                        |

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

From: cjs@netcom.com (Christopher Shaulis)
Subject: Re: Linux 1.1.35 and ld -qmagic
Date: Wed, 27 Jul 1994 04:55:13 GMT

idefix@venus (Robert KNOTEK) writes:

>Trying to compile Linux kernel 1.1.35 with gcc 2.5.8 

>compiling works fine but at the end there ist an ld switch (-qmagic)
>which my version of ld dont know.

>Is there a newer version of ld (how to find, i searched on nic.funet.fi and 
>tsx-11.mit.edu but can't find).

sunsite.unc.edu /pub/Linux/GCC/binutils-1.9l.4.tar.gz
tsx-11.mit.edu /pub/linux/packages/GCC/binutils-1.9l.4.tar.gz

Since making qmagic binaries is such a neat and cool thing to do, you should
also get ld.so.1.4.4.tar.gz 

And as long as your upgrading those two, you might as well save yourself
future truoble and effort and get src/tools-2.15.tar.gz from the same place
you got the other packages. NOTE: if you have /usr/dll/jump/as as a 
symlink to /usr/bin/as or whatever, delete that symlink before you install
tools 2.15. Otherwise you'll completely screw GCC when you replace as
(GNU Assembler) with as (that tools thingie).

  _____    ___________  _  _____________________  __  ___ _________  __  ___
 / ___/_  / / __/ ___ \/ |/ / __/_  __/ ___/ __ \/  |/  // ___/ __ \/  |/  /
/ /__/ /_/ /\ \/ / _ `/    / _/  / / / /__/ /_/ / /|_/ // /__/ /_/ / /|_/ / 
\___/\____/___/\ \_,_/_/|_/___/ /_/  \___/\____/_/  /_(_)___/\____/_/  /_/  
                \___/                                                       
-- 
  _____    ___________  _  _____________________  __  ___ _________  __  ___
 / ___/_  / / __/ ___ \/ |/ / __/_  __/ ___/ __ \/  |/  // ___/ __ \/  |/  /
/ /__/ /_/ /\ \/ / _ `/    / _/  / / / /__/ /_/ / /|_/ // /__/ /_/ / /|_/ / 
\___/\____/___/\ \_,_/_/|_/___/ /_/  \___/\____/_/  /_(_)___/\____/_/  /_/  
                \___/                                                       

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

Crossposted-To: comp.unix.bsd
From: bsa@kf8nh.wariat.org (Brandon S. Allbery)
Subject: Re: APC UPS owners or potential buyers, trying to show user base
Date: Fri, 29 Jul 1994 11:37:42 GMT

In article <JOHNSONM.94Jul28233853@calypso-2.oit.unc.edu>, johnsonm@calypso-2.oit.unc.edu (Michael K. Johnson) says:
+---------------
| In article <317k55$kq3@news.halcyon.com> mpdillon@coho.halcyon.com (Michael Dillon) writes:
| 
|    Really now! Those boxes use an RS-232 interface, right? What do they tell 
|    the computer? If they only communicate one thing (power fail) then it
|    is probably something as simple as shorting the RD and SD lines. Get a
|    technician to check it out for you while you pull the plug.
| 
| The "Back-UPS" do this.
+------------->8

I have a Deltec UPS on loan (it will probably be returned when I come up with
something decent to replace it); it won't send powerfail signals on its serial
port unless (a) it is so configured and (b) it receives a command code over
the serial port to activate it.  Needless to say, said code is proprietary.

I might get the SCO software for it and use iBCS2 system call trace to examine
the codes sent over the serial port...

++Brandon
-- 
Brandon S. Allbery         kf8nh@44.70.4.88               bsa@kf8nh.wariat.org
Friends don't let friends load Windows NT (tnx Sun)    A Linux iBCS2 developer

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

From: ramos@engr.latech.edu (Alex Ramos)
Subject: console driver does not reset char table with "stty sane"
Date: 31 Jul 1994 04:15:33 GMT



I don't know if this is a bug in stty or in the console driver. Or
maybe it isn't a bug at all, or one of those "compatibility" or 
"historical" bugs?

The problem can be replicated like this:

        echo ^N

(To get a ^N, hit Control-V followed by Control-N)

This switches the console driver to an alternate character set.  On a
monochrome card (the one I have) the effect is drastic -- all lowercase
characters get mapped to high-ascii graphics.

Unfortunately, closing the device or even "stty sane" does not switch
you back to the original character table. 

An "echo ^O" puts the driver back to its sane state, but I had to look
at kernel sources to find that...

--
Alex Ramos (ramos@engr.latech.edu) * http://info.latech.edu/~ramos/
Louisiana Tech University, BSEE/Sr * These opinions are probably mine

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

From: svm@kozmix.hacktic.nl (Sander van Malssen)
Subject: Re: kernel stupidity...
Reply-To: svm@kozmix.hacktic.nl
Date: Sun, 31 Jul 1994 05:53:47 GMT

olav woelfelschneider <wosch@rbg.informatik.th-darmstadt.de> writes:

> Nowadays, if one tries to mount a write protected floppy without explicitely
> specifying it to be mounted read-only, an error occurs.
> 
> I think this is really nasty, wouldn't it be possible to change the floppy
> driver again, so that it automatically mounts the drive read only if its
> protected?

Bad idea, because then mount(8) won't know the device is mounted
read-only and you end up with a wrong entry in /etc/mtab and all
that. Better to let mount(8) detect that the mount failed and try
again with the read-only flag. Here's a quick & dirty patch for
mount (as found in util-linux-1.9) to do just that. (Dirty in the
sense that it assumes that an EACCESS from the mount syscall always
means `read-only' (which is currently the case I think). This'll
break if the syscall is ever changed to return something like EROFS
in cases like this.)

table                                                        
`!"#$%&'()*+,-./0123456789:;<=>?                             
@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_                             
begin 666 mount-ro.diff.gz                                   
M'XL("*<Z.RX"`VUO=6YT+7)O+F1I9F8`K5-=:]LP%'U.?L7!,(AGN<11VJXV
MA8S2AXW2/72#OA77OFY-%=F3Y*YAI+]]DAP'IW0/@_G%]^/HGN.CZSB.L6XZ
M:8Z*H]?D=?+]L</73B!9(N'I\7&:<"1G9\MI%$4#<'+320_B">:GZ?(D318]
M:+5"O."<G2!RKU.L5E,`M32H1/Z@,Y<5C[G"1WHQ*K]K6C,N>H)=,>K/Y=+4
M=XJ*3FG".>86[4=6F)E-:TOGN/YQ=16Z(O#;2^!\SCXAXCQAB^5.@Z7([83+
MZ]LOW]*^`)!2C<(L\+PI/FC4&K*QK'C.15WB7C3%$TIZK@L*&'1+19CA7E'^
ME(VG?KZXN+Q)I_&[8\=#1APMJ75M#)5H)&JC4=6"]$8;6N^IAHD[QJC/W,>/
M?0F'AC-@'_Y7':.IAUJ`[1"2T'2@9(*#ZXNBS-><?'?'H<\&I'M<U5ZR)1!6
MZ^Q%&U5V;0^V4HP2)/LLXF&V/V8;16Y\@R%@J@GVS2U\T$N;C#@"B\K^S:Q?
MJC84MZHQ5%B[6/\_U/+!"2[C1HK-NX8I,IV2V"VW),P<AEGS2V)P6\S0:Z\4
H_61H<ZW#`X/]JZ0J[X3Y^_(&WJ)=>6VY;!B^6=?M]`\YC!&D\@,`````    
`                                                            
end
bytes 490
crc 20E6
-- 
Sander van Malssen
svm@kozmix.hacktic.nl

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

From: terryd@extro.ucc.su.OZ.AU (Terry Dawson)
Subject: Re: Subnet masks
Date: Sun, 31 Jul 1994 09:41:07 GMT

kruckenb@sal.cs.utah.edu (Pete Kruckenberg) writes:

>Bob Minowicz (minowicz@eos.ncsu.edu) wrote:

>: was then this is now.  I would imagine that by now, this has been fixed. 
>: However, since the only source for info is the old NET-2 HOWTO,and it
>: still says its broken... 

>It has been fixed for a long time (since 1.1.0, at least, and probably
>even since some of the 1.0.x kernels). We are using 1.1.24 with a
>2-bit subnet field, so it definitely is working now.

As has the NET-2-HOWTO.

Why read an old version ?

Terry

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

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

From: yuriev@astro.ocis.temple.edu (Starcon SysAdmin)
Subject: Re: Panasonic CD-ROM broke under >1.1.31
Date: 27 Jul 1994 23:40:32 GMT

lcvanveen@et.tudelft.nl wrote:
: When mounting my CD-ROM (panasonic cr-562b) under 1.1.35 I got
: this nice error:
: mount: block device /dev/sbpcd is not permitted on its filesystem
: It wasn't in at 1.1.30 so it came later (I presume at 1.1.32).
: Does anybody have a fix here.
: Audio cd's still work, as you would have guessed.
: Thank's a lot in advance.
: Martijn.

Did you use -o ro option? 

Best wishes,
Alex

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

From: ssd@nevets.oau.org (Steven S. Dick)
Subject: Re: threads in kernel
Date: Sun, 31 Jul 1994 04:36:21 GMT

In article <31aj9p$n3l@smurf.noris.de>,
Matthias Urlichs <urlichs@smurf.noris.de> wrote:
>In comp.os.linux.development, article <Ctno26.L5@wg.saar.de>,
>The new process will probably get a segmentation fault if it tries to
>return from the procedure you're calling with clone(), but that's OK, IMHO.

Probably when you create the new stack, you should copy the old stack...
This would give a better parallel to what fork() does.  You'd have to
be careful with pointers that were put on the stack before the clone...

The only other sane alternative IMHO would be to have returning from the
function that called tfork/clone to exit, just like if you returned from main...

        Steve
ssd@nevets.oau.org

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

Crossposted-To: comp.os.linux.help
From: eggert@twinsun.com (Paul Eggert)
Subject: Re: Converting a *future* time in another timezone to local time ...
Date: Wed, 27 Jul 1994 23:09:37 GMT

mrj@cs.su.oz.au (Mark James) writes:

> Is there any way to do what I want without having to hack some C code;

Probably not.

> and if I have to write some C, what is the best way to go about it.

The most portable way is to use putenv("TZ=..."), tzset, mktime, and
localtime in the proper order.  This is ugly, since it relies on global
state in your C program, but doing it right will mean you'll have to do
a lot of hacking yourself.

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

From: ypwong@ie.cuhk.hk (Kenneth Wong)
Subject: [Newbie] Docs on (network) kernel/driver ... Anyone?
Date: Wed, 27 Jul 1994 12:09:11 GMT

I am starting to pick up some driver writing, but besides (the not very
complete) kernel hackers guide, is there any other reference? It'd be good if
I need not work out the possibly false facts from kernel sources. Eg
what effect does mark_bh have?

Incidentally, are linux mailing-lists archived? I think debian's is, and
it would be good too if eg kernel updates are archived.

Thanks in advance.
--
Kenneth Wong
ypwong@ie.cuhk.hk

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

Date: Tue, 26 Jul 1994 09:06:00 GMT
From: janosch@erde.GUN.de (Morton Jonuschat)
Subject: PCI IDE....

  Reply auf eine Mail von David paul Kauffman an All
  in Area .oO COMP.OS.LINUX.DEVELOPMENT Oo. vom 25 Jul 94 * 04:34

 DpK> From: dkauffma@tech.iupui.edu (David paul Kauffman)
 DpK> are there any drivers out there yet that will support the PCI IDE?
 DpK> -Thanks
 DpK> -Dave

Yes...the normal IDE Drivers....

Tschoeoeoeoeoeoe,
          /X\orton

-.o PCMCIA : People Can't Memorize Computer Industry's Acronyms o.-


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

From: tzs@u.washington.edu (Tim Smith)
Subject: Re: link() system call
Date: 31 Jul 1994 14:15:23 GMT

Michael Edward Chastain <mec@shell.portal.com> wrote:
>I agree.  A long time ago I read an article (I think written by Dennis
>Ritchie), saying that some very old (pre-V7) Unix allowed links to
>directories, but use of this feature lead to unmanageable chaos.  I am

I don't know what everyone is smoking, but I wish I had some!  Unix has
allowed hard links to directories much later than V7.  System V Release 2
allowed it.  System V Release 3 allowed it, I seem to recall.  Ultrix
seems to allow it (according to the man page for link--I don't have
the root password on any Ultrix system so I can't actually check).

--Tim Smith

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

From: kim@square.nl (Kim Hendrikse)
Subject: Large malloc problems
Date: Sun, 31 Jul 1994 15:14:00 GMT

Anyone know why I can malloc 1 MEG of memory but accessing it greater
than index (About) 250800 causes a Memory fault??? ( I have 12 MEGS ram).

Test program:

#include <stdio.h>
#include <stdarg.h>
#include <stddef.h>
#include <malloc.h>

main( int argc, char *argv[] )
{
    int *p;
    int i;

    if ( !( p = ( int * ) malloc( ( size_t ) 1000000 ) ) )
        exit( 1 );

    setbuf( stdout, NULL );
    for ( i = 0; i < 1000000; i++ ) {
        if ( i % 100 == 0 )
            printf( "%d\n", i );
        p[ i ] = 255;
    }
}

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

From: jalbert@ATHENA.MIT.EDU (Jonathan D Albert)
Subject: Re: Large malloc problems
Date: 31 Jul 1994 14:51:16 GMT

In article <Ctt8zD.ME3@square.nl> kim@square.nl (Kim Hendrikse) writes:
>Anyone know why I can malloc 1 MEG of memory but accessing it greater
>than index (About) 250800 causes a Memory fault??? ( I have 12 MEGS ram).
>
>Test program:
> <Stuff deleted>

You are allocating the memory for use in an integer array.

integer's are 32-bit in linux.  So, it requires four bytes for each one.
Thus, 1 meg of memory can store 1meg/4=~250000 integers...

jd
jalbert@mit.edu




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

From: john@gab.unt.edu
Crossposted-To: comp.os.linux.help,comp.os.linux.misc,comp.os.linux.help
Subject: Networking very slow
Date: 27 Jul 1994 22:59:22 GMT
Reply-To: john@gab.unt.edu

Hi, I've got a bit of a problem--my networking is extremely slow--telnet to machine
can take up to 2 min to actually connect.  It doesn't look as if the local net
is overflowing with packets everywhere.  The 3c509 network card I am using is
supposed to be a very quick card.


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

From: ds2@dragonslair.caltech.edu (Dragon Slayer)
Subject: Re: Fix for problems with sound when system is idle.
Date: 31 Jul 1994 15:56:59 GMT

In article <30t6i7$rgj@backdoor.ethernet>,
Oliver Daudey <traveler@backdoor.hacktic.nl> wrote:
>Since my uplink had problems with their news and I'm not sure if this made
>the outside world, I'm reposting it...
>
>Some time ago my sound suddenly stopped working...  As soon as my system
>was idle, the sound would get scrambled.  I tried all kinds of things, like
>changing IRQ/DMA/address on my SB-Pro, without success...  The weird thing
>was that it worked perfectly on a friend's system with the same kernel, so
>I expected a hardware problem.  This was even more strange because it went
>away by booting an old kernel, which seemed to indicate a software problem..
>
>Just recently, it suddenly came to me that sys_idle() had been changed to
>halt the CPU when the kernel is idle.  I decided to give it a try and removed
>the hlt-instruction.  After that, everything suddenly worked again.  No more
>need to start a CPU-intensive program just to be able to play some MOD's. :-)
>
>I wonder if more people experienced this problem, as I didn't read anything
>about it in the Linux-groups.  It seems to be a hardware problem, triggered
>by software...  Maybe my motherboard handles the halts incorrectly.  I have
>a Highscreen (one of this "exclusive" Colani-designs  :-)  ).  The mother-
>board is a 4386-VC-V, equipped with a 486 DX/2-66, 8 Megs of RAM and 256K
>cache...  Chipset is VIA VT82C480...  The soundcard is a SB-Pro, which is
>reported at bootup as "snd2 <SoundBlaster Pro 3.2> at 0x220 irq 7 drq 1"...
>
>Anyone with the same hardware, who has/had similar problems??

I have similar hardware: 486DX-40 USA Chipset, Sound Galaxy NX Pro 
(SB Pro clone) on 0x240 irq 10 dma 1. The sound driver would work 
with old kernels (1.0.5?), but fail miserably with the newer ones 
(instead of sound, I would get a buzz. Running the dsp scope program
from sunsite would should a display even when there isn't any input.)
Tried pulling all but the sound card/disk controller/vga card out and still
same problem. So, I don't think this is a conflict between cards.

Commenting out the hlt loop seems to have fixed. I am guessing that the 
hlt instruction somehow messes up the DMA so data couldn't be transfered
properly between sound card and memory. 

Could someone more knowledgeable comment on this?

-- 
Dragon's Lair, USA -- Dragon Slayer -- +1.818.395.1406
<a href="http://dragonslair.caltech.edu/~ds2/">ds2@dragonslair.caltech.edu</a>

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

From: danw@connected.com (Dan Wilder)
Crossposted-To: comp.os.linux.admin
Subject: Re: 1.1.35 make problem: exhausted memory
Date: 27 Jul 1994 18:38:03 -0700

pcsigler@whale.st.usm.edu (Peter C. Sigler) writes:

>Peter Schmitt (pete@geminga.dartmouth.edu) wrote:
>: My 386 has 5M ram and 5M swap.  I am currently running 1.1.34 and got
>: the following when trying to make 1.1.35:

>I think this is a problem with the memory handling in 34.
>I tried to compile 35, as well, under 34 and got the same
>thing. My system is a 386/40 w/5Meg ram and 10meg swap.

That isn't a lot of room.  Try running 'top' to see what the situation
is.  I suspect with 5 meg RAM and 10 meg swap, you are nearing swap
full compiling the kernel even with earlier kernels.  'Course it
could be the memory handling.

----
Dan Wilder

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

From: johannes@titan.westfalen.de (Johannes Stille)
Subject: Re: Panasonic CD-ROM broke under >1.1.31
Date: Wed, 27 Jul 1994 00:09:44 GMT

In article <312b26$1ao@smurf.noris.de> urlichs@smurf.noris.de (Matthias Urlichs) writes:
[...]
>
>It works perfectly if you happen to mount them read-only.
>
>Of course, the error is misleading -- mount(2) should return -EROFS and not
>-EACCESS.
[...]

No, it's the mount(8) message that is misleading.

EROFS means that you're trying to write to a filesystem that is mounted
read-only. While you're trying to mount the CD-ROM rw, there is no
filesystem involved that is mounted read-only.

It is much better compareable to opening a device read-write where you
cannot write. I don't know any better approximation than EACCESS,
although this usually means that the action is not allowed (as opposed
to impossible).
But what is the error code if e.g. you try to open a standard printer
port for reading? (Or what should the error code be, if this doesn't
actually return an error? - always assuming that the read permission bit
is set for the printer device.)

        Johannes

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


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