Files
oldlinux-files/docs/mail-archive/linux-devel/Volume1/digest5XX/digest526
2024-02-19 00:23:35 -05:00

496 lines
22 KiB
Plaintext

Subject: Linux-Development Digest #526
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: Mon, 7 Mar 94 12:13:08 EST
Linux-Development Digest #526, Volume #1 Mon, 7 Mar 94 12:13:08 EST
Contents:
Re: Amiga FileSystem, Anyone? (David Holland)
Re: Any luck with 32 bit ethernet cards? (Andreas Fatum)
Specialix Driver Round 2 (From specialix) (Alan Drew)
Re: NEC T130B scsi host (Kevin Lentin)
Re: AMD 486DX problem (with Linux?) (Cedric Adjih)
Re: Specialix driver (Alan Cox)
pre1.0 strange behaviour (Laurent Chemla)
Re: program runs fine in xxgdb, but not in xterm! (Stupid me!) (Manoj Kasichainula)
Re: Specialix driver (Ian G Batten)
Re: dynamic loading (Robert Sanders)
----------------------------------------------------------------------------
Subject: Re: Amiga FileSystem, Anyone?
From: dholland@husc7.harvard.edu (David Holland)
Date: 7 Mar 94 04:56:18
rob@pe1chl.ampr.org's message of Sun, 6 Mar 1994 13:07:16 GMT said:
> The PC has a specialized floppy disk controller that understands and
> handles the industry-standard MFM format of formatting diskettes.
> The Amiga does not use that standard format (and neither does the Mac)
As some other people have said, I was wrong in some of my assumptions.
I can cope with that, even in a public newsgroup. :-)
> Classification of 'more or less capable' is entirely yours. I would say
> the PC disk controller is more capable, in that it handles tasks that
> need to be done in software on the Amiga and Mac.
Hmm, I'd say "capable" means "able to accomplish things". It seems to
me that drive hardware that can read a wider range of disk formats is
more capable. This is what the end user sees; only the BIOS or device
driver writer sees how much internal processing is done. I dunno.
Meaningless point to argue.
Btw, the Amiga's disk hardware may not do the MFM encoding, but the
encoding is still done in hardware: by the bit-blitter...
> >Yes, I have. It's amazing that it can be done at all, however
> >poorly... :-)
>
> Please explain what is poor about it?
> This seems to be just a general case of DOS-bashing. When you don't
> know what you are talking about, please don't.
The network redirector is a mess, not well documented, and notoriously
difficult to cope with. Why do you think we don't see alternate
filesystems (such as for Mac floppies) for the PC that use it? All we
have is a few network packages from big companies with lots of
resources, like Novell and Sun.
Yes, it does work, mostly. Why do various ordinary tools refuse to
cooperate with network drives? This problem is not limited to
low-level disk utilities or anything, you know.
> (I am no DOS fan. not at all, even. but saying things that just plainly
> aren't true is not the way to handle even DOS)
Are you absolutely sure I'm the one who doesn't know what he's talking
about? :-) :-)
--
- David A. Holland | Nobody ever went broke underestimating
dholland@husc.harvard.edu | the intelligence of the American public.
------------------------------
From: ace@acelab.ruhr.de (Andreas Fatum)
Subject: Re: Any luck with 32 bit ethernet cards?
Date: Sat, 5 Mar 1994 23:33:02 GMT
Danny D! (ddambros@nyx10.cs.du.edu) wrote:
: card. Not much is mentioned in the HOWTO, save that the 3COM 3C579 is a tad
: lame for a 32 bit card... =)
Huh? I'm using the 3Com 3c579 (EISA) in my machine and I'm really satisfied
with the network performance. The 3c579 and the Cogent Emaster+ EM935 XL
are the fastest 10-Mbit network cards on the market. (The 2/94 issue
of the German network magazine N&C (Networks & Communication) includes
a thorough test of the fastest network cards. "Showdown in ethernet-valley")
Andreas
--
Andreas Fatum ace@acelab.ruhr.de (Internet)
postmaster@re.open.de (City-Router)
Student of CS, ace@acelab.ruhr.sub.org (SubNet)
University of Dortmund ..!uunet!germany.EU.net!acelab!ace (UUCP/Bang!)
------------------------------
Date: 7 Mar 94 10:05:02 GMT
Subject: Specialix Driver Round 2 (From specialix)
From: cdh@specialix.co.uk (Alan Drew)
Looks like I've been missing some of the fun. I'm not sure what
this argument is really all about, tho' I *am* sure that one
of you kind souls will enlighten me.
One thing that I have seen quite clearly tho is that there is some
argument about what does and does not fit into whatever license
agreements. Hopefully I can clear that up......
A device driver (Partcularily an Intelligent device driver) usual fits
something like this (in block diagram format) into your system.
+-----+ +--------------+ +---------------+ +-------------------+
| O/S |----| Device Driver|-----| Download Code |----| modem/terminal etc|
+-----+ +--------------+ +---------------+ +-------------------+
The download code is the same on *EVERY* platform. It is a proprietary
peice of code written specifically to drive the proprietary hardware
design. As a company we have spent millions of pounds developing
both that and the hardware for all of our product ranges. It is known
otherwise as the crown jewels. It doesn't need to be modified it doesn't
need to be changed. It remains the same wether you are running on RS6000,
HP/UX, SVr4, BSD, Risc/OS. This is copyright Specialix and will remain so.
We distribute this bit in hex dump format for those who wish to develop
thier own device drivers. We also distribute a document called
"The Technical Reference Manual" which gives complete and comprehensive
details on how to write an interface to the download code.
That interface would be a Device Driver. As a company we have spent
millions developing these, but we give the source away with every one
we sell. This is the bit that somebody else develops. The ones we right
are copyright specialix, if you write one for Linux you can make it
copyright your grandmother for all we care. One thing is for certain tho'
if you develop a Linux driver that's copyright your grandmother, and
then phone here and say "I can't install my driver", we'll tell you to
go talk to your grandmother. *ON THE OTHER HAND* If you phone here and
say "Look, I've just developed a linux driver, and when I hold CTS low
whilst posting a WFLUSH followed by a close to the port and wibble DTR,
then I lose the last 3 bits of the last byte in my queue", then the
chances are we'd fall over ourselves to help you 'cos it's likely to
be a problem with the download code.
Also included alongside the technical reference manual, and the binary dump
for the download code is enough source code and other usefull info
for you to be able to develop your own device driver. At least that's
true if you've got some idea of what you are doing.
That's the legal stuff and the Specialix standpoint. None of which I perceive
as being unreasonable. The DTK (that's the manual and the source code) usually
costs 500 pounds. We do and have done on occasion give the whole thing away
free if we think it's a worthwhile exercise that will benefit us aswell as
others.
One such ooccasion was only a few months ago. A gentleman in South Africa
is developing a driver for linux and the Slxos product range. I will not
name him as he requested when he started that it be kept quite because he
did not want to be plagued by people badgering him for the driver. Nor do
I know what the progress is, this is not a Specialix development so I
have no right to chase the gentleman. He did however say that when he
has finished the project then he would contact me, when and if he does so
I will let people know what is available and what its terms of distribution
are from a Specialix point of view.
Above all, one thing that I find highly risible in this constantly recurring
argument about wether or not commercial companies such as ourselves should
support free O/S's. Every time, somebody rolls out the argument "I spoke
to company X the other day and they sounded like they would be really
helpfull". Some months ago, I posted an offer, 1 Free developers toolkit
and 1 free SI16. All the software 2 terminal adaptors and 1 AT host card,
nearly 2000UK pounds worth of kit. I got one offer to set up a mailing
list and/or an FTP site (thanx, you know who you are) and *ONE* offer
to develop a driver. Count 'em.... *ONE*. There is an age old axiom
"Put up, or shut up", I'm sure some of you have heard it. Silly thing is
that as a commercial organization we listened to your complaints and
requests and we *DID* something about it, we made you the offer that
you had all been asking for, we "Put up". You all "shut up". Seriously.
I couldn't beleive it, I was expecting to swamped by an army of spare
time free software developers jumping at the opportunity. What I got was
3 peices of mail from people saying, "I'll buy a card if you do it" or
"if you have a Linux driver, I'll be in the market for 4 cards". Sadly
promisory sales of a total of 64 ports do not make a wise
investment for us, and nobody bar one of you wanted the driver so
desperately that you were prepared to do the work yourselves.
And yet you expect us as a commercial organisation (who's sole goal at
the end of the day is to make money) to take you seriously. Well rest
assured that we do take you seriously, we are just waiting for you.
You can't epect us to do all the running, when we make an effort,
then somebody needs to repond to that, it's called meeting in the
middle, and all the complainiing and whining isn't going to change that.
One of the problems, now of course, is that at the last product meeting
the MD asked what progress had been made in investigating the possibilities
of a LINUX driver, he was told what the reponse was. I might take you
seriously, he no longer does. So if I want to try and help you guys then
I have to do it behind is back and present him with a finished product.
That is the commercial fact.
If you have any further comments on this subject or any questions
please do not hesitate to contact me, any of Alan.Drew or cdh
@specialix.com or @specialix.co.uk will get to me.
I look forward to hearing from you.
Alan Drew
International Support, Specialix.
--
This article bought to you by CESPITS
Campaign for Extremely Serious Postings to InterneT news Services
------------------------------
From: kevinl@bruce.cs.monash.edu.au (Kevin Lentin)
Subject: Re: NEC T130B scsi host
Date: 7 Mar 1994 12:52:39 GMT
On 2 Mar 1994 07:48:13 GMT, Francesco Defilippo wrote:
> Hello,
> The Nec T130B scsi is supported by linux?
Nec T130B? Trantor make a T130B and it can be made to work under Linux with
no interupts and no pseudo-dma. I have a patch to make interupts work and
am about to start the pseudo-dma from scratch since I somehow managed to
stop the whole driver working and figure a rethink is in order.
--
[==================================================================]
[ Kevin Lentin |___/~\__/~\___/~~~~\__/~\__/~\_| ]
[ kevinl@bruce.cs.monash.edu.au |___/~\/~\_____/~\______/~\/~\__| ]
[ Macintrash: 'Just say NO!' |___/~\__/~\___/~~~~\____/~~\___| ]
[==================================================================]
------------------------------
From: adjihc4@cti.ecp.fr (Cedric Adjih)
Subject: Re: AMD 486DX problem (with Linux?)
Date: 7 Mar 1994 13:30:59 GMT
John McCluskey (jbm@speedy.login.qc.ca) wrote:
: micha@mubo.saar.de (Michael Bongartz) writes:
: >On 05 Mar 1994 05:42:38 GMT in comp.os.linux.development,
: > Gregory McKesey (mckesey@imaphics.prior.com) wrote:
: >: I have found an annoying problem with the AMD 486DX chip and
[...deleted...]
: >: I decided to investigate further, and found a problem with
: >: single precesion (ie float) multiplys. The following is a sample
: >: program that illustrates the problem.
: >I couldn't resist testing this on my AMD 486 DX2/66 system:
: >micha@moko|~/tmp>cc -o float float.c
: >micha@moko|~/tmp>float
: >1.312500 * 7.999900 =10.499868
: >1.312500 * 7.999900 =10.499869
: >Test succeeded!
: >It seems that this bug doesn't exist on all AMD chips.
: Ditto. My AMD-486DX40 yields exactly the same results as above. My
: CPU is marked with the "Windows Compatible Logo". I bought it last month.
: It looks like the recent production of AMD is clean.
My AMD-486DX40 (bought in november) gives the sames results too...
It might be also a compiler or librairie bug :
- the assembly code generated could be seen with the "-S" option of gcc
- the printf("%f",a) could be checked with float a=10.49986(8|9)
(I have slackware 1.1.2, I think it is gcc 2.5.7 and libc x.x.19)
: J. McCluskey@ieee.org
Just my 2 cent(ime)s,
--
Cedric Adjih / Internet : adjihc4@cti.ecp.fr
Disclamer : concerning my English.
------------------------------
From: iiitac@swan.pyr (Alan Cox)
Subject: Re: Specialix driver
Date: Mon, 7 Mar 1994 13:19:20 GMT
In article <2l870d$ceu@news.u.washington.edu> tzs@u.washington.edu (Tim Smith) writes:
> (b) In no case does copyright protection for an original work of
> authorship extend to any idea, procedure, process, system, method
> of operation, concept, principle, or discovery, regardless of the
> form in which it is described, explained, illustrated, or embodied
> in such work.
Except when your name is Apple Computer.... surely
Alan
------------------------------
From: laurent@brasil.frmug.fr.net (Laurent Chemla)
Subject: pre1.0 strange behaviour
Date: 6 Mar 1994 19:59:19 GMT
--
First thing, Wooahoo I saw my box printing 1.0 while booting :-)
Now the bad news. I'm writing a program that places the text cursor where it
wants on the screen. Stdin is in O_NONBLOCK mode, but stdout is not.
The cursor does no more react the same way it was used to until I upgrade
from pl15h to ALPHA1.0. I can't exactly say what happens, but I can say that
it won't work the same when i do this:
sprintf(buf,"\033[%d;%dH",CurY,CurX);
fputs(buf,stdout);
fflush(stdout);
The behaviour is 'sometime the cursor is at the right place, sometime not,
probably depending on when stdout get flushed'.
NOTE that it does work allright on the console screen, but it just seems to
be fooled on any of my serial terminals, at any speed. If i'm on ttyS4
(for instance) and then just enter 'myprogram' then it doesn't work, but if
I enter 'myprogram > /dev/ttyS4' then it does work!
A 'stty -a' shows that rows and columns are set to 0 when I log in on a
serial terminal, but it always did. And when I enter a 'tset' command, rows
and columns are correct but it doesn't change anything.
I'm using gcc2.5.8, libs4.5.19, Taiwanese 486/66 and a compatible AST
fourport serial card.
I can't be sure it's not a bug in my program that was hidden in previous
kernel, but as I do not know (yet) the reason of this behaviour, I prefer to
post this article. Please forgive my bad english.
--
Laurent Chemla : chemla@cnam.cnam.fr or laurent@brasil.frmug.fr.net
Brasil BBS - +33 1 44 67 08 44 - Atari France developpers support
------------------------------
From: mvkasich@eos.ncsu.edu (Manoj Kasichainula)
Subject: Re: program runs fine in xxgdb, but not in xterm! (Stupid me!)
Date: 7 Mar 1994 14:47:49 GMT
Reply-To: mvkasich@eos.ncsu.edu (Manoj Kasichainula)
In article <2l7ja7$93b@louie.udel.edu>, mvkasich@eos.ncsu.edu (Manoj
Kasichainula) writes:
|OK, here's the situation. I'm trying to do a program for my data
|structures class at home. I've appended a copy of it below to look at.
|When I run the program at the command line prompt, it immediately
|exits
|and sets $status to 1. But when I run it in xxgdb, it runs fine,
|except
|for some necessary debugging 8^). Why is the program not doing
|anything
|when run at the command line prompt? No I'm not asking someone to
|debug
|the program, just please explain why it doing what it's doing. Oh
|yeah,
|when I run the program on the DECstations at college, it works fine.
|Thanks in advance.
Thanks for all of your help! With some help, I've found the problem.
Apparently, there is a GNU program called "test" wich I had never heard
of. And being a former DOS slave (what a sad existence it was!), when I
compiled my program and named it test, I didn't think that when I tried
to run it, it would run the GNU version instead of the version in the
current dir. Uggggh! Thanks again for all the help. Oh by the way, the
ALPHA version of kernel 1.0 is out, on nic.funet.fi:
/pub/OS/Linux/kernel/src.
--
_______________________________________________
| | "Violence is the refuge of
| Manoj Kasichainula - mvkasich@eos.ncsu.edu | the incompetant."
| Leader of the Jihad to Destroy Barney at NCSU | - Salvor Hardin
|_______________________________________________| from _Foundation_ by Asimov
------------------------------
Crossposted-To: gnu.misc.discuss
From: igb@fulcrum.co.uk (Ian G Batten)
Subject: Re: Specialix driver
Date: Mon, 7 Mar 1994 15:21:33 GMT
In article <2l5u1p$vcj@zealot.uucp>,
Keith Medcalf <kmedcalf@zealot.uucp> wrote:
> The entire GPL/GLPL is quite straightforward and means just about
> If you get a good dictionary of the English
> Language (English, not American or Canadian, but ENGLISH), and look up the
> EXACT AND PRECISE meaning of every single word
Could you please suggest such a dictionary? I have an interested
layman's perspective on lexicography, as I fund my hifi interest
consulting for several dictionary publishers and number several people
who produce dictionaries for a living amongst my friends. I've never
heard them make such claims (``EXACT AND PRECISE'') for their
dictionaries.
ian
------------------------------
From: gt8134b@prism.gatech.EDU (Robert Sanders)
Subject: Re: dynamic loading
Date: 7 Mar 94 16:09:52 GMT
luke@umnstat.stat.umn.edu (Luke Tierney) writes:
>I am currently trying to port a statistical system to a range of UNIX
>systems and would like to include Linux support. Unfortunately I don't
Great! It's a wide audience.
>One thing I try to do in my package is allow dynamic loading of
>compiled C code. On most systems with shared liraries I have been able
>to do this by using hooks into the shared library mechanism. Many
>systems have a common interface -- a function dlopen for loading a
>library, a function dlsym for locating the address of a symbol.
You can do pretty much the same thing with GNU DLD with dld_link()
and dld_get_func for functions and dld_get_sym() for variable (if
you want dld_get to be generic, just have it try each in sucession).
I've just modified two programs (perl5 and elk 2.2) to use DLD
instead of dlopen()/dlsym() and it works fine. One thing you have
to be careful about is name conflicts with global symbols in
the .o/.a files: unlink the dl_* functions, DLD acts much like
real linking, so if you have a global symbol init() in two .o files,
you're going to get a "multiply defined symbol" error.
If you want a good example of dynamic loading using various schemes
(excuse the pun), check out elk 2.2, the Extension Language Kit,
which is a Scheme interpreter meant to be embedded as an extension
language. It's also a standalone interpreter. It supports dynamic
loading on HPUX using shl_load(), under BSD systems using 'ld -A'
incremental linking and then just read()ing the object file into
memory, and now I've written interface code to DLD. You can also
use 'ld -A' incremental linking under Linux, but you have to insert
the "-static" switch in the call to ld so that it links against
the static versions of the libraries.
>If the shared library approach doesn't work I can use the GNU dld
>library, which I gather is available for Linux. Is this library
>included in standard Linux distributions, or does it have to be ftp's
>from somewhere.
DLD 3.2.3 (the standard GNU distribution right now) doesn't work under
Linux. Aubrey Jaffer ported it to Linux, renumbered it 3.2.4, and
made it available on ftp.uu.net:/languages/lisp/scm and prep.ai.mit.edu:
/pub/gnu/jacal.
Linux's current shared libraries really aren't meant for dynamic loading
as those of SVR4 and derived systems are. Maybe that will change with
the work being done on ELF binary formats, but in the meantime, DLD
provides very satisfactory results.
Write me if you'd like to discuss this further.
--
_g, '96 --->>>>>>>>>> gt8134b@prism.gatech.edu <<<<<<<<<--- CompSci ,g_
W@@@W__ |-\ ^ | disclaimer: <---> "Bow before ZOD!" __W@@@W
W@@@@**~~~' ro|-<ert s/_\ nders | who am I??? ^ from Superman '~~~**@@@@W
`*MV' hi,ocie! |-/ad! / \ss!! | ooga ooga!! | II (cool)! `VW*'
------------------------------
** 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
******************************