Linux-Development-Sys Digest #891, Volume #7 Mon, 22 May 00 12:13:10 EDT
Contents:
gdb can't start linuxthreads ("Christian Casteyde")
Re: Need ideas for university funded project for linux (Someone)
Re: How to find path to a shared module ("Guillaume LAFAGE")
Re: Need ideas for university funded project for linux (Robie Basak)
Test (o.T.) ("Matthias Wald")
Re: Need ideas for university funded project for linux (Donal K. Fellows)
Re: How to find path to a shared module (Christian Stieber)
Re: stack size ("Christian Casteyde")
Re: Need ideas for university funded project for linux (John Hasler)
Re: REQ : PLX9080 driver source code (William Waddington)
url: Partition Magic for Linux (Philip)
Problems with shared memory in 2.3.99 + (C. Folstrom)
POSIX Portability issue. ("Rajeev B. C.")
Re: porting NT device drivers to linux (Mario Klebsch)
Re: serial port RTS control ? (Mario Klebsch)
Re: serial port RTS control ? (Mario Klebsch)
Re: ping souce code (Diego Betancor)
Re: Need ideas for university funded project for linux (Leslie Mikesell)
----------------------------------------------------------------------------
From: "Christian Casteyde" <[EMAIL PROTECTED]>
Subject: gdb can't start linuxthreads
Date: Mon, 22 May 2000 12:11:23 GMT
Hi,
I'm writing a pgm that uses threads, but I can't manage to start it in gdb.
When I run it, it catch a "unknown signal", then is completely locked.
I tried to recompile gdb 5.0, but it still does nothing (on a SuSE
(glibc2.1.3, kernel 2.2.15)).
It catch a "SIG32 realtime" signal, then only lauch the main thread (the
other one is killed ?).
I also tried on Slackware (glibc2.1.3, kernel 2.3.99-pre6), but then gdb
crash (Segmentation fault).
How can I start a pgm with threads in gdb ?
Does gdb uses linux headers and works only on stable kernels ? Can I use it
with 2.3.99 kernels ?
Note : I made a test program with 2 threads, one that prints As and the
other Bs,
it works, but I got only As in gdb if it doesn't crash.
Thank you
------------------------------
From: Someone <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux,comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc
Subject: Re: Need ideas for university funded project for linux
Date: Mon, 22 May 2000 08:18:25 -0400
Good technologies:
Hammer, grab handle and beat things with the heavy end.
Shovel, Pointy end down, step on flat part, pry on long handle - move
dirt.
Computers. Sit in front, look at monitor, type and click -
Bad Technologies:
Linux, Spend 6 months installing. Ask for advice, get told you are an
idiot and to read a phone book. Go to school get computer science
degree to decipher phone book. Re-compile packages to work with last
months libs, find missing packages, re-compile, wait for downloads and
re-compile, fatal errors after an hour of compiling for wrong lib
version, buy latest release, get new phone book. Find the problem isn't
with program with the phone book but some net admin set the defaults to
Ultra Secure (read no useful functions unless you read another phone
book and recompile with proper switches). Buy new phone book, Why
doesn't the library carry these things? Find actually allow to work
switch, re-comple- oops now new fonts lib required, download
compile,,,,, ./configure usefull=yess secure=no hardware=unknown
vesion=several, make, oops, make linux , make
install,,,,,,,,,,Reboot... What the!!!! IT LIVES!!!!!! IT LOOKS JUST
LIKE WIN95!!!!!!! Look to see what apps I have! A bad word processor (
I have Office for that) and a bunch of configuration utilities. Now
what can I do? Do what I do. Shut it off, wait six months and install
again. Don't get me wrong its a great hobby! By the way, anybody know
how to get diald functions for Road Runner?
------------------------------
From: "Guillaume LAFAGE" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.misc,de.comp.os.unix.linux.misc,linux.dev.c-programming
Subject: Re: How to find path to a shared module
Date: Mon, 22 May 2000 14:37:30 +0100
Jeff Ostrin <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> I'm working on a project building a shared module. When distributed, I of
> course don't know the path to the install directory, but I need to know
the
> path inside my module.
>
> Is there a system call I can make inside my module to get the path to my
> module?
>
> Thanks
> Jeff
>
is "getcwd" what you're looking at ??
Guillaume.
------------------------------
From: [EMAIL PROTECTED] (Robie Basak)
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.advocacy
Subject: Re: Need ideas for university funded project for linux
Reply-To: [EMAIL PROTECTED]
Date: 17 May 2000 22:59:17 GMT
On Wed, 17 May 2000 16:52:41 GMT, eyez said:
>quoting <[EMAIL PROTECTED]>:
>[snip]
>>Mongoose <[EMAIL PROTECTED]> writes:
>>[snip]
>>can put my programming hours where my mouth is. (Reason I'm not using
>>it now? The fucking Aureal Vortex 2 drivers are (a) non-free; and (b)
>>unusably poor.)
Do you mean non-free as in beer non-free or RMS non-free? If
you mean the former, check out http://linux.aureal.com
They're not brilliant, but the sound in Quake 3 works, and mp3s play
fine :-)
Robie.
--
------------------------------
From: "Matthias Wald" <[EMAIL PROTECTED]>
Subject: Test (o.T.)
Date: Mon, 22 May 2000 15:11:51 +0200
Reply-To: "Matthias Wald" <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Donal K. Fellows)
Crossposted-To:
comp.os.linux,comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.advocacy
Subject: Re: Need ideas for university funded project for linux
Date: 22 May 2000 13:09:43 GMT
In article <[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
> I much, much, much prefer being able to right-click on something and
> hit "Properties." I also like being able to press F1 when the mouse
> is over a confusing field and get an explanation of it. (The
> explanation often isn't a help, and I expect that would carry over
> to Linux, but at least there's no flipping around between screens.)
Agreed. The right-button and F1 idioms are good ones, as are button
bars on apps and scrollwheels on mice. Writing decent documentation
is a different skill to writing a decent GUI. I have yet to meet
anyone at all who was good at both in the same application.
[attribution lost]
>> This is the problem though, they don't care enough to create
>> programs to help newbies install and use linux and so linux is
>> being held back.
>
> I care enough. I'm just no good at GUI programming.
It isn't that hard with something like Tcl/Tk, Perl/Tk or TkInter.
You just have to remember that users aren't necessarily going to work
through things the same way you do, that they want *both* mouse and
keyboard navigation, plenty of help and a chance to undo things where
possible (and a really hefty warning where you can't undo!)
If you're genuinely interested, try to read a copy of Alan Cooper's
_About Face_ (ISBN: 1-56884-322-4) which, for all its strong focus on
'Doze, is a really good book on GUI design.
Donal.
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ [EMAIL PROTECTED]
-- I may seem more arrogant, but I think that's just because you didn't
realize how arrogant I was before. :^)
-- Jeffrey Hobbs <[EMAIL PROTECTED]>
------------------------------
From: Christian Stieber <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.misc,de.comp.os.unix.linux.misc,linux.dev.c-programming
Subject: Re: How to find path to a shared module
Date: 22 May 2000 13:22:17 GMT
In de.comp.os.unix.linux.misc Jeff Ostrin <[EMAIL PROTECTED]> wrote:
> Is there a system call I can make inside my module to get the path
> to my module?
No.
However, somebody does in fact know the path, since dlopen() needs
to find your module as well. So you can just pass the same "filename"
argument as a parameter to your module.
Christian
--
M�chten Sie Ihre Festplatten formatieren?
[J]a
[N]at�rlich
------------------------------
From: "Christian Casteyde" <[EMAIL PROTECTED]>
Subject: Re: stack size
Date: Mon, 22 May 2000 13:39:42 GMT
A stack for each process, + a task per thread.
Dynamically growable, don't bother with that, if your stack is too small
then you have a bug.
GAUTIER Christopher <[EMAIL PROTECTED]> a �crit dans le message :
[EMAIL PROTECTED]
>
> Hello,
> I'm using gcc on Linux Mandrake, and I would some info about handling
> stack with gcc, ie:
>
> does each program has his stack (like in DOS) ? In this case, how do you
>
> specify the size of the stack ?
>
> or
>
> does the OS provide a common stack for everyone ?
>
> Thanx
>
>
>
>
------------------------------
From: John Hasler <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux,comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc
Subject: Re: Need ideas for university funded project for linux
Date: Mon, 22 May 2000 12:07:54 GMT
Christopher Browne writes:
> The best vote is a functioning software package. Whether that be:
> a) A source RPM, or
> b) A source .deb, or
> c) A Ports package.
And there is the Debian popularity contest, an optional package that keeps
track of what you install and anonymously reports it to a Web page where
totals are displayed.
--
John Hasler
[EMAIL PROTECTED]
Dancing Horse Hill
Elmwood, Wisconsin
------------------------------
From: William Waddington <[EMAIL PROTECTED]>
Subject: Re: REQ : PLX9080 driver source code
Date: Mon, 22 May 2000 14:46:59 GMT
Hello,
No Guru here, just an old man struggling.....
I have written a few DMA drivers for our boards using the 9080. They are
clunky, comment infested, and in a horrible coding style (in short, everything
CodingStyle says not to do ... I'm working on that). They do seem to be solid -
at least through 2.2. I am currently fighting with a problem in 2.3, but I
don't think it has anything to do w/the 9080.
Go to http://home.att.net/~ikoncorp and follow the link to our FTP site. There
are great masses of code that have nothing to do with the PLX part, but you
should be able to filter through that to the (hopefully) useful stuff.
HTH,
Bill
BTW - we have always given away our drivers w/source - all the way back to
SunOS.
Lighty wrote:
>
> Hello Gurus,
>
> I am in need of a PLX9080 driver source code. I am trying to write a driver
> myself but am facing some problems, and some source code would be of great
> help !
>
> Thanks in advance to all of you wonderful people !!
>
> Regards,
> Daniel Doron.
> [EMAIL PROTECTED]
--
=======================================
Bill Waddington
Bainbridge Island, WA, USA
[EMAIL PROTECTED]
[EMAIL PROTECTED]
=======================================
------------------------------
From: Philip <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: url: Partition Magic for Linux
Date: Mon, 22 May 2000 15:33:25 GMT
http://linux.msede.com/ext2/
--
Philip van Hoof aka freax _ _
http://users.pandora.be/skatan | | (_)_ __ _ ___ __
mailto:[EMAIL PROTECTED] | | | | '_ \| | | \ \/ /
Hey, it will happen. They'll | |___| | | | | |_| |> <
go where the money is(Linus) |_____|_|_| |_|\__,_/_/\_\
------------------------------
From: [EMAIL PROTECTED] (C. Folstrom)
Subject: Problems with shared memory in 2.3.99 +
Reply-To: [EMAIL PROTECTED]
Date: Mon, 22 May 2000 09:10:17 GMT
First off, all is as the /usr/src/linux/Documentation/Changes files
says it should be, including the system software; everything meets or
exceeds version minimums.
Linux ravenblack.org 2.2.15 #3 Sat May 13 01:23:33 CDT 2000 i586 unknown
Kernel modules 2.3.11
Gnu C 2.95.2
Binutils 2.9.5.0.37
Linux C Library 2.1.3
Dynamic linker ldd: version 1.9.11
Procps 2.0.6
Mount 2.10f
Net-tools 2.05
Kbd 0.99
Sh-utils 2.0
Modules Loaded ppp slhc serial
mount shows /var/shm mounted, as does /proc/mounts. free, on the
other hand, shows no shared memory.
2.3.99-pre3 runs w3ell enough, but 2.3.99-pre8 runs like very slowly.
When I first used 2.3.99-pre3 I noticed there was a pseudo-file in
/var/shm, but now there is not.
I discontinued to use 2.3.99-pre3 because it didn't handle the
framebuffer stuff well even though the shared memory stuff worked.
I have a Tekram motherboard, with onboard Trident Cyberblade graphics
that uses part of system ram as video ram, and onboard sound. Both of
the two subsystems work fine in Linux. The proccessor is a AMD K6-II
350 3d-Now.
Between the first time I compiled and used 2.3.99-pre3 and recently
when I tried 2.3.99-pre8, I did a standard system upgrade using
dselect / apt to get it from the Debian mirror. The compilers and
libs were done at that time.
So, there it is. Make of it as you will. Thanks for your time;
C. D. Folstrom
------------------------------
From: "Rajeev B. C." <[EMAIL PROTECTED]>
Subject: POSIX Portability issue.
Date: Mon, 22 May 2000 10:23:56 -0500
Has anyone been able to implement sigqueues in linux developments ?
The source compiles and executables run perfectly ok in solaris environment
but when the same source is ported to linux, Have no compilation error but
during runtime, the functions do not return the desired values.
If somebody has implemented or met with any successful workaround, please
help me, with any changes in function prototype. A code snip of the function is
highly appreciated.
Thanks
Rajeev
P.S: I have tried with no success with RedHat, Hardhat, Debien and suse version
of linux.
------------------------------
From: [EMAIL PROTECTED] (Mario Klebsch)
Subject: Re: porting NT device drivers to linux
Date: Mon, 22 May 2000 16:54:07 +0200
Marc SCHAEFER <[EMAIL PROTECTED]> writes:
>Mario Klebsch <[EMAIL PROTECTED]> wrote:
>: loaded into the kernels address space. I assumed, that if an NT driver
>: needs a DLL, NT will be able to load the DLL into kernel space. The
>And how is the `chicken and egg' problem solved within NT ?
>Such as:
> - I need the SCSI driver aic7xxx to access the filesystem, and
> the aic7xxx needs to access the file system to get the DLL ?
I do not know, which way NT solved this problem. IMHO the linux way is
clutchy at best. If you want to see an elegant sollution to this
initial loading problem, you should take a look at solaris. They do
have drivers and lots of other parts of the kernel in modules (even
the filesystem of the root fs can be a module), but they do not need
any clutch like initrd to boot.
73, Mario
--
Mario Klebsch [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Mario Klebsch)
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: serial port RTS control ?
Date: Mon, 22 May 2000 17:01:07 +0200
[EMAIL PROTECTED] (Villy Kruse) writes:
>These two signals have now lost its old function and are now used to
>control data flow: CTS will control sending and RTS will control the
>corresponding CTS signal on the other end.
This would be nice, if the two implementation (the traditional RTS
signal, which activates the DCEs transmitter and the new one, which is
used to throttle the communication partners transmitter) were
compatible. In fact, it can lead to deadlocks, if you connect two
devices to each other, which have different RTS semantics.
This is ugly, and IMHO should not have happened at all. I am not sure,
if this is yet another PC made design flaw, so I am not going to pick
upon PCs or Microsoft for this one. :-)
If I (or we, the linux community) accept the change in RTS semantics,
we have to accept the need for the old traditional semantics. So it
seems a goot idea to me, to add some kind of mode switch (like
CRTSCTS) to the kernel to select between them.
73, Mario
--
Mario Klebsch [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Mario Klebsch)
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.setup
Subject: Re: serial port RTS control ?
Date: Mon, 22 May 2000 16:56:39 +0200
Charles Bryant <[EMAIL PROTECTED]> writes:
>The obvious solution is to pad the transmission with the same number
>of characters as the device FIFO can hold. When you have loaded the
>last pad character the device must have only just finished
>transmitting the last message character so you can toggle RTS,
>disable the device, or do whatever it is that requires such exact
>timing. Of course you'll have to do this in the ISR to ensure there's
>no extra delay inserted.
I have seen this sollution several times, but I never liked it. One of
the reasons for me to dislike this sollution is, that I have to know
the size of the Tx FIFO, which can vary from chip to chip, even amoung
chips, that claim to be compatible. I do not like to have this kind of
harware dependancy in my systems, not if I can avoid it.
73, Mario
--
Mario Klebsch [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Diego Betancor)
Subject: Re: ping souce code
Date: Mon, 22 May 2000 14:41:33 GMT
True,
but would it not be nice if you can set a timeout and it would tell
you that the time out is reached? so looking at a screen (from far)
you can see if you are having problems.
Diego
[EMAIL PROTECTED] (Mike Dowling) wrote:
>On Thu, 18 May 2000 18:48:23 GMT, Diego Betancor <[EMAIL PROTECTED]> wrote:
>>Hi,
>>
>> ping is great but when it fails it does not tell you like windows
>>that says something about timeout. The only way that I know is
>>counting the number of packets and looking for one that is missing.
>> does anyone know where I can get the source for ping to change it?,
>>or does anyone already done it?
>
>Windows yielding an error message that is both comprehensible and
>correct? It has never happened to me!
>
>Perhaps this could be a case in point. Ping uses the icmp protocol,
>and, to the best of my knowledge (and I will bow to the more
>knowledgeable) icmp protocols don't time out. Timeouts are the domain
>of tcp protocols where sessions are established.
>
>Ping is supposed to send off a an echo request about once per second.
>The packets are numbered, and when a given request is answered, it is
>listed with facts like the time it took, etc. But it can and should
>wait 'til kingdom come if the packet is not echoed or it is not
>interputed.
>
>Cheers,
>Mike
>--
>My email address [EMAIL PROTECTED] above is a valid email address.
>It is, in fact, a sendmail alias; the digit 'N' is incremented regularly.
>Spammed aliases will be deleted. Currently, mike[25,26]
>are valid. If email to mikeN bounces, try mikeN+1.
Diego Betancor @ Duo Business Communications
for email: dbetancor is my userid and my company's domain is duocom.net
** Do not send me unsolicited commercial e-mail spam of any kind **
------------------------------
From: [EMAIL PROTECTED] (Leslie Mikesell)
Crossposted-To:
comp.os.linux,comp.os.linux.development,comp.os.linux.development.apps,comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.advocacy
Subject: Re: Need ideas for university funded project for linux
Date: 22 May 2000 10:57:36 -0500
In article <[EMAIL PROTECTED]>,
David Steuber <[EMAIL PROTECTED]> wrote:
>Well, KDE _is_ GPL'd. That is a requirement of using Qt Free
>Edition. I'm not sure that using the Lesser GPL is even an option.
>
>That option _does_ exist for GTK+ and GTK--.
>
>For myself, if I am going to produce free software, I want it to be
>GPL. I don't want my work to be used in proprietary software unless I
>get paid for it.
This doesn't make a bit of sense from a user's perspective. You
are saying that I can use the code only as long as it isn't linked
to something else that I might happen to need that is under
someone else's control. I don't think the internet would exist
as we know it today had it not been for the reference BSD code
that does allow use in proprietary works as well as additional
free ones.
Les Mikesell
[EMAIL PROTECTED]
I also don't feel that I can produce code of such
>high quality that others should pay to use it a la' Microsoft
>shitware. However, I can certainly do good things with the help of
>others. I think such help is easier to get if the code is GPL. It
>means no one can steal the work of others.
>
>What TrollTech is currently doing with Qt 2.x and higher is a good
>thing. People who produce GPL software can use Qt without worrying
>about the QPL. Modifying Qt is another story, but I expect that
>worthy changes would be incorporated into Qt. It is certainly
>possible to use inheritance to extend Qt. Your code is GPL anyway.
>TrollTech can't take GPL code and put it under the QPL.
>
>If you don't like Qt, then you can simply use another toolkit.
>GTK+/GTK-- is a popular, free toolkit. There is also Tk. Tk has been
>ported to more platforms than both Qt and GTK+.
>
>If the above options, including toolkits I haven't mentioned, are not
>to your liking, then you can undertake writing your own toolkit. I'm
>sure that is more work than writing any single application.
>
>--
>David Steuber | Hi! My name is David Steuber, and I am
>NRA Member | a hoploholic.
>
>All bits are significant. Some bits are more significant than others.
> -- Charles Babbage Orwell
------------------------------
** 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: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-System Digest
******************************
