Linux-Development-Sys Digest #179, Volume #8 Thu, 28 Sep 00 16:13:12 EDT
Contents:
Re: [How to kill process which use device-file??] (Qing Liu)
Re: new windowing system (Nix)
Re: Timeserver for RH 6.1 or 6.2 (ray)
Re: s3 virge accelerated specification (Dale Pontius)
vm86 system call ("andrew.betts")
Java on Netscape (Was Re: Java on Linux?) (Bob Schreibmaier)
Re: new windowing system ("Fr�d�ric G. MARAND")
Re: Process priorities (Rick Ellis)
smp scheduling ([EMAIL PROTECTED])
Re: Java on Netscape (Was Re: Java on Linux?) (Robert Lynch)
smp scheduling ([EMAIL PROTECTED])
tar & other systems ("Apple")
Re: new windowing system (Alexander Viro)
Re: Java on Netscape (Was Re: Java on Linux?) (Toby Haynes)
Re: new windowing system (Drazen Kacar)
Re: new windowing system (Alexander Viro)
Problem Solved: unresolved symbol request_module (Bjorn Eriksson)
----------------------------------------------------------------------------
From: Qing Liu <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: [How to kill process which use device-file??]
Date: 28 Sep 2000 11:56:01 +0200
Bomshik Kim <[EMAIL PROTECTED]> writes:
> When we listen to music from XMMS(or x11amp),
> one of XMMS processes(about 3~4 processes) occupies the device file
> /dev/dsp.
>
> But, When some other application( let's call it alpha ) wants to
> use /dev/dsp and XMMS plays mp3-music already,
> then since /dev/dsp is occupied by XMMS, alpha can't open /dev/dsp.
> because two processes can't share /dev/dsp.
What about fuser -k /dev/dsp ?
--
Liu
------------------------------
From: Nix <$}xinix{[email protected]>
Crossposted-To: comp.os.linux.x,comp.windows.x
Subject: Re: new windowing system
Date: 27 Sep 2000 23:36:30 +0100
[This should, I suppose, move off c.o.l.d.s, but I don't get feeds of any
of the other groups this is posted to.]
[EMAIL PROTECTED] (Alexander Viro) writes:
> Ahem. GNOME is "design by committee and guy who doesn't get UNIX". It's not
Unfortunately true :( I was rather disappointed to read Miguel's
opinions of why Unix sucks. (`Not componented'? What else is a pipe for?)
> a flame towards Miguel - just that his design decisions make me very
> uncomfortable.
Some of them make me uncomfortable too.
> Use of CORBA is one of them,
Using CORBA was probably right. It's just a super-RPC, is all. And it
*is* damned useful. It's another componenting scheme, no better or worse
than pipelines except you can't really get at it from the command line
very well (and GNOME's CORBA server has nearly fixed even that).
> general attitude wrt
> orthogonality is another,
Good grief, yes. Perl disease again. (But then I *am* a Lisp advocate,
so I would say that.)
> happy bloating the APIs - one more... Try to
> read their codebase, just keep the barfbag ready.
Compared to a lot of the stuff I have to deal with at work it is
wonderful. (At least they know what indentation is.)
--
`no amount of Zen contemplation will ever make you at one
with a 3c905B ethernet card.' --- Alan Cox
------------------------------
From: ray <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.portable,comp.os.linux.setup,linux.redhat.install,linux.redhat.misc
Subject: Re: Timeserver for RH 6.1 or 6.2
Date: Thu, 28 Sep 2000 12:12:52 GMT
"David M. Cook" wrote:
> On Thu, 28 Sep 2000 00:04:08 -0400, jhuman <[EMAIL PROTECTED]> wrote:
>
> >Does anyone know where I can find a timeserver (Redhat) for my internal LAN?
>
> Look for xntp or ntp on your distro CD.
>
> Dave Cook
Yes, and then, look here for setup help
http://raymondjones.net/ntpguide.html
--
Ray R. Jones
Errors have been made. Others will be blamed.
[EMAIL PROTECTED]
HTTP://raymondjones.net
------------------------------
From: [EMAIL PROTECTED] (Dale Pontius)
Crossposted-To:
alt.comp.hardware,alt.comp.hardware.homebuilt,alt.comp.hardware.pc-homebuilt,comp.hardware,comp.ibm.pc.hardware,comp.os.linux,comp.os.linux.hardware,comp.os.linux.questions,comp.os.linux.x,comp.os.ms-windows.vide
Subject: Re: s3 virge accelerated specification
Date: 28 Sep 2000 12:16:26 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] writes:
> On Wed, 20 Sep 2000 10:18:19 +0300, "Ian Dichkovsky" <[EMAIL PROTECTED]>
> wrote:
>
>>Hi !
>>(Sorry for bad English)
>>I want to write an accelerated video driver for Linux
>>(or maybe later for other systems ....).
>>
>>Where I could download specification(documentation) on
>>S3 Virge DX/GX(375/385)?(html, pdf, TeX, dvi, ... )
>>Samples of using accelerated
>>capabilities ?
>>
> The S3 virge is a very old graphics card. are you sure that a driver
> has not already been made?
>
This was already done as part of the Utah-GLX project. Generally,
the 3D acceleration of the ViRGE is just about worthless. The guy
who actually made it work for OpenGL got some derision for even
trying, but quite a bit of respect for making it work.
It worked as well as it could. I ran it briefly, before getting a
Matrox G400.
Dale Pontius
------------------------------
From: "andrew.betts" <[EMAIL PROTECTED]>
Subject: vm86 system call
Date: Thu, 28 Sep 2000 14:43:43 +0100
Can someone explain to me how to execute a code snippet with the vm86 system
call.
eg
void somecode()
{
__asm__(
"mov %%ax,%%bx" // or something
)
}
void go()
{
struct vm86plus_struct vm86s;
vm86s.regs.eip=(void*)somecode; ?????
vm86(VM86_ENTER,&vm86s);
}
I have tried filling the other segment registers with __USER_DS/CS but all
_appears_ to fail.
And how do return from "somecode" to the line after the syscall?
--
=====================================
Andrew Betts
[EMAIL PROTECTED]
(T) +44 7050 693 974
(F) +44 7050 693 975
=====================================
------------------------------
Crossposted-To: comp.os.linux.misc
From: [EMAIL PROTECTED] (Bob Schreibmaier)
Subject: Java on Netscape (Was Re: Java on Linux?)
Date: Thu, 28 Sep 2000 15:11:02 GMT
In article <[EMAIL PROTECTED]>,
Garry Knight <[EMAIL PROTECTED]> wrote:
>On Wed, 27 Sep 2000, Exits Funnel wrote:
>
>>Sun's JDK is available on linux but am wondering what alternatives there
>>are.
>
>http://www.ibm.com/java/jdk/118/linux/ IBM's jdk1.1.8 for Linux
>http://www.blackdown.org/java-linux.html Blackdown Java 1.1.7 JDK for Linux
I hope this isn't a FAQ, but is there a way to make Netscape use
this for Java applets instead of its buggy internal code?
Bob
--
+------------------- \-\-\-\ ----------------------------+
| Bob Schreibmaier K3PH | E-mail: [EMAIL PROTECTED] |
| Kresgeville, PA 18333 | ICBM: 40o55'N 75o30'W |
+--------------------------------------------------------+
------------------------------
From: "Fr�d�ric G. MARAND" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.x,comp.windows.x
Subject: Re: new windowing system
Date: Thu, 28 Sep 2000 17:40:24 +0200
Would you happen to know what he meant by that ? I've my own version of
3c905b hell, but I'd like to know what's his.
Nix <$}xinix{[email protected]> a �crit dans le message :
[EMAIL PROTECTED]
> `no amount of Zen contemplation will ever make you at one
> with a 3c905B ethernet card.' --- Alan Cox
------------------------------
From: [EMAIL PROTECTED] (Rick Ellis)
Subject: Re: Process priorities
Date: 28 Sep 2000 16:42:14 GMT
In article <8qtotc$vs$[EMAIL PROTECTED]>,
Peter Pointner <[EMAIL PROTECTED]> wrote:
>You're probably right. But Frank might want to do "man sched_setscheduler",
>maybe that does the trick.
It might, but probably won't be reliable with a 100Hz clock.
--
http://www.fnet.net/~ellis/photo/linux.html
------------------------------
From: [EMAIL PROTECTED]
Subject: smp scheduling
Date: Thu, 28 Sep 2000 17:03:24 GMT
hi,
i have a question about smp kernel scheduling. if i insert a module in an smp
kernel with two cpus, will the entire init_module function be executed on one
cpu, or will scheduler change cpus in between? i am disabling interrupts at
the begining of init_module and enabling them at the end, so there arent any
interrupts coming in. also, the function does not sleep anywhere in between.
also, if the scheduler does change cpus, is there any way to disable the
other cpu until init_module finishes?
the system is a dual pentium-133 processor system running kernel 2.2.13
any comments will be appreciated,
-njm
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: Robert Lynch <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Java on Netscape (Was Re: Java on Linux?)
Date: Thu, 28 Sep 2000 10:18:37 -0700
Reply-To: [EMAIL PROTECTED]
Bob Schreibmaier wrote:
>
> In article <[EMAIL PROTECTED]>,
> Garry Knight <[EMAIL PROTECTED]> wrote:
> >On Wed, 27 Sep 2000, Exits Funnel wrote:
> >
> >>Sun's JDK is available on linux but am wondering what alternatives there
> >>are.
> >
> >http://www.ibm.com/java/jdk/118/linux/ IBM's jdk1.1.8 for Linux
> >http://www.blackdown.org/java-linux.html Blackdown Java 1.1.7 JDK for Linux
>
> I hope this isn't a FAQ, but is there a way to make Netscape use
> this for Java applets instead of its buggy internal code?
>
> Bob
>
> --
> +------------------- \-\-\-\ ----------------------------+
> | Bob Schreibmaier K3PH | E-mail: [EMAIL PROTECTED] |
> | Kresgeville, PA 18333 | ICBM: 40o55'N 75o30'W |
> +--------------------------------------------------------+
You can install a plug-in (Blackdown's got one) which requires an
HTML conversion, or put the Swing jar in your classpath, if only
1.1 features are used.
HTH. Bob L.
--
Robert Lynch-Berkeley CA [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: smp scheduling
Date: Thu, 28 Sep 2000 17:09:32 GMT
hi,
i have a question about smp scheduling under linux. when i insert a
module, will the entire init_module function execute on one cpu, or can
the scheduler step in and switch cpus mid-way? i am disabling interrupts
during the entire function, and the function does not sleep anytime.
also, if the scheduler can change cpus, is there anyway i can disable
the other cpus until the function exits.
i am running a dual-cpu with pentium-133s on kernel 2.2.13
any suggestions are welcome,
-njm
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: "Apple" <[EMAIL PROTECTED]>
Subject: tar & other systems
Date: Thu, 28 Sep 2000 19:47:32 +0200
I want to read an exabyte from a scsi 's tape what was written with
Sun-Solaris 7 's tar command.
When I do tar tvf /dev/st0 I have an I/O error .
I try to put blocking factor like tar tvfb /dev/st0 20 but it is same I/O
error
What can I do.
I have the same trouble when I want to read a linux tape on Sun tape !!!
--
============================================================================
==============================
[EMAIL PROTECTED]
Like Sybase Central LSC http://perso.wanadoo.fr/laserquest/linux
------------------------------
From: [EMAIL PROTECTED] (Alexander Viro)
Crossposted-To: comp.os.linux.x,comp.windows.x
Subject: Re: new windowing system
Date: 28 Sep 2000 13:50:36 -0400
In article <[EMAIL PROTECTED]>,
Nix <$}xinix{[email protected]> wrote:
>Some of them make me uncomfortable too.
>
>> Use of CORBA is one of them,
>
>Using CORBA was probably right. It's just a super-RPC, is all. And it
>*is* damned useful. It's another componenting scheme, no better or worse
>than pipelines except you can't really get at it from the command line
>very well (and GNOME's CORBA server has nearly fixed even that).
Yeah... The only problem being: client side is not using the normal UNIX
API. You know, open()/read()/write()/close()... CORBA is not the first
mistake in that direction. Look: we need something that would do marshalling
in both directions. I.e. we need 4 transformations (encoding/decoding for
request/response). So far so good, but notice that normally clients are
more numerous than servers (both the number of programs and the number of
running instances). So it makes sense to have the client side as simple
as possible. On UNIX it means the file representation of the whole
thing. IOW, client-side encoding/decoding should be done by kernel stub
and server should tell the kernel what representation should be used.
It doesn't increase the overhead, BTW - instead of
client encode
call sendmsg()
call recvmsg()
kernel deliver the packet to server
wake the server up
server return from recvmsg()
decode
do the operation
encode
call sendmsg()
kernel deliver the packet to client
wake the client up
client return from recvmsg()
decode
you have
client call write()
call read()
kernel encode
deliver the packet to server
wake the server up
server [same as in the usual scenario]
kernel decode
return the data to client
client return from read()
Count the syscalls and need of context switches - it's the same as in usual
RPC, but client becomes _way_ simpler. Yup, it means standard RPC for
everything that wants to use that mechanism. Essentially, server exports
a filesystem. Client doesn't have to know anything about the RPC, nature of
server, etc. I would certainly like to be able to say
% cat /mnt/nscd/passwd/lusername
and get the entry, regardless of the nsswitch.conf, etc. All you need for that
is nscd that would export RPCFS. All the crap with libc speaking NIS, parsing
/etc/nsswitch.conf, yodda, yodda goes away.
Same goes for BIND and friends. Same could be done for X if it wouldn't insist
on the RPC of its own - compare object ID with inumber and think what could
be done.
Damnit, the whole UNIX toolchain was designed to work with file API. Deal
with it. If somebody doesn't understand how to use the basic functions
(open/read/write/close/lseek/etc.) (s)he should RTFBourne or any other
introduction-level text on UNIX instead of reinventing the wheels. NIH
syndrome is not a thing to be proud of, especially if it comes from
willful ignorance about the existing stuff.
--
"You're one of those condescending Unix computer users!"
"Here's a nickel, kid. Get yourself a better computer" - Dilbert.
------------------------------
From: Toby Haynes <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: Java on Netscape (Was Re: Java on Linux?)
Date: 28 Sep 2000 14:35:53 -0400
!! "Bob" == Bob Schreibmaier <[EMAIL PROTECTED]> writes:
Bob> In article <[EMAIL PROTECTED]>, Garry Knight
Bob> <[EMAIL PROTECTED]> wrote:
>> On Wed, 27 Sep 2000, Exits Funnel wrote:
>>
>>> Sun's JDK is available on linux but am wondering what alternatives there
>>> are.
>> http://www.ibm.com/java/jdk/118/linux/ IBM's jdk1.1.8 for Linux
>> http://www.blackdown.org/java-linux.html Blackdown Java 1.1.7 JDK for
>> Linux
Bob> I hope this isn't a FAQ, but is there a way to make Netscape use this
Bob> for Java applets instead of its buggy internal code?
Not that I know of. Your best hope is that Mozilla gets a full OJI compliant
JRE on Linux - Sun is working on release 1.3.1 just so this happens. I believe
that IBM is also working on getting its JDK in line with OJI compliance. In the
bug reports on bugzilla for the Mozilla project you can track the latest bits
and pieces. Soon, in other words.
Cheers,
Toby Haynes
--
Toby Haynes
The views and opinions expressed in this message are my own, and do
not necessarily reflect those of IBM Canada.
------------------------------
From: [EMAIL PROTECTED] (Drazen Kacar)
Crossposted-To: comp.os.linux.x,comp.windows.x
Subject: Re: new windowing system
Date: 28 Sep 2000 19:03:42 GMT
Alexander Viro wrote:
> as possible. On UNIX it means the file representation of the whole
> thing. IOW, client-side encoding/decoding should be done by kernel stub
> and server should tell the kernel what representation should be used.
> It doesn't increase the overhead, BTW - instead of
I'm not familiar with CORBA, so I might be wrong here. However, I believe
ORBit doesn't do it the way you described. I'm not sure what exactly
takes place in encode/decode phase, but if it's only endianess, then
it can be side-stepped. I think ORBit clients don't do that part at all.
Servers do it only if necessary, ie. if the endianess is different. The
clients always receive data with the correct endianess (from their point
of view).
> client encode
> call sendmsg()
> call recvmsg()
> kernel deliver the packet to server
> wake the server up
> server return from recvmsg()
> decode
> do the operation
> encode
> call sendmsg()
> kernel deliver the packet to client
> wake the client up
> client return from recvmsg()
> decode
>
> you have
> client call write()
> call read()
> kernel encode
> deliver the packet to server
> wake the server up
> server [same as in the usual scenario]
> kernel decode
> return the data to client
> client return from read()
So it's the same, except for read/write vs. sendmsg/recvmsg. In case of
local client and server, that's the same, too.
> server, etc. I would certainly like to be able to say
> % cat /mnt/nscd/passwd/lusername
> and get the entry, regardless of the nsswitch.conf, etc. All you need for that
> is nscd that would export RPCFS. All the crap with libc speaking NIS, parsing
> /etc/nsswitch.conf, yodda, yodda goes away.
That looks like XFN, which is yet another dead thing. But it exists on
some platforms and can be used.
> Damnit, the whole UNIX toolchain was designed to work with file API. Deal
> with it.
Which is not necessarily the best API in all circumstances. Now, Plan 9
with its "everything uses UTF-8" is better than the Unix and its
"everything uses ASCII and if you need something else, we have these
fine kludges to offer."
> If somebody doesn't understand how to use the basic functions
> (open/read/write/close/lseek/etc.) (s)he should RTFBourne or any other
That doesn't work well always. If you're reading from a slow NFS mount,
your GUI application might be very unresponsive, unless you write
the code which deals with NFS effects. The API is the same, read()
will return your data eventually, but "eventually" means very different
things for local file access and NFS access.
--
.-. .-. I don't work for my employer.
(_ \ / _)
| [EMAIL PROTECTED]
| [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Alexander Viro)
Crossposted-To: comp.os.linux.x,comp.windows.x
Subject: Re: new windowing system
Date: 28 Sep 2000 15:55:40 -0400
In article <[EMAIL PROTECTED]>,
Drazen Kacar <[EMAIL PROTECTED]> wrote:
>I'm not familiar with CORBA, so I might be wrong here. However, I believe
>ORBit doesn't do it the way you described. I'm not sure what exactly
>takes place in encode/decode phase, but if it's only endianess, then
It's certainly _not_ only endianness.
>So it's the same, except for read/write vs. sendmsg/recvmsg. In case of
>local client and server, that's the same, too.
Yes, except that for client most of the crap goes away. For one thing, you
can use several files, essentially encoding the request into the choice of
file you are accessing. For another, you don't have to deal with RPC
scheduling on the client side, yodda, yodda.
>> Damnit, the whole UNIX toolchain was designed to work with file API. Deal
>> with it.
>
>Which is not necessarily the best API in all circumstances. Now, Plan 9
>with its "everything uses UTF-8" is better than the Unix and its
>"everything uses ASCII and if you need something else, we have these
>fine kludges to offer."
Huh? UTF-8 vs. ASCII is completely irrelevant here - neither UNIX nor Plan 9
care about it anywhere near the code paths involved. And "ASCII" is not
accurate - it's 7-bit and UNIX API is 8bit-clean. BTW, you do realize that
Plan 9 uses the aforementioned architecture for _all_ RPC-like stuff? Yup,
including the window system, DNS, etc.
>> If somebody doesn't understand how to use the basic functions
>> (open/read/write/close/lseek/etc.) (s)he should RTFBourne or any other
>
>That doesn't work well always. If you're reading from a slow NFS mount,
>your GUI application might be very unresponsive, unless you write
>the code which deals with NFS effects. The API is the same, read()
>will return your data eventually, but "eventually" means very different
>things for local file access and NFS access.
Who said "NFS"?
--
"You're one of those condescending Unix computer users!"
"Here's a nickel, kid. Get yourself a better computer" - Dilbert.
------------------------------
From: Bjorn Eriksson <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Problem Solved: unresolved symbol request_module
Date: 28 Sep 2000 19:51:33 GMT
FWIW...
Searching USENET for the string "unresolved symbol request_module" show
that several people have had this problem but no one came back with a
solution. The solution is to 'make dep' or remove the
$BASEDIR/include/linux/modules/ - files and rebuild.
'make clean' doesn't remove these files.
//Bj�rnen, Sorry for the crossposting. (I'm not subscribed to these list.)
------------------------------
** 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
******************************
