Linux-Development-Sys Digest #346, Volume #8 Wed, 13 Dec 00 16:13:21 EST
Contents:
Re: Compiling C++ programs with GCC --> no GPL license implications (jbs)
Re: Compiling C++ programs with GCC --> no GPL license implications (jbs)
Help: ieee1394 kills init (Goran Ristic)
Re: Compiling C++ programs with GCC --> no GPL license implications (Pete Becker)
Re: Compiling C++ programs with GCC --> no GPL license implications (jbs)
Re: Compiling C++ programs with GCC --> no GPL license implications
([EMAIL PROTECTED])
Re: Help! Hosed XMMS trying to upgrade (Rob Komar)
Linux 2.2.18 Bug ? (Torsten Sievers)
Re: connect 2 sockets (Grant Edwards)
Posix Message Queues ([EMAIL PROTECTED])
serial port: how to manually control RTS?? ("Dan Miller")
Does kernel 2.2.18 need to be patched for raid??? ([EMAIL PROTECTED])
patch question (Bart De Schuymer)
Re: serial port: how to manually control RTS?? ("Ron W. Auld")
Re: Q: Linux serial line programming, open() hangs (Ron)
Re: serial port: how to manually control RTS?? (Grant Edwards)
Re: patch question (Kaz Kylheku)
Debug help (Arnaud Westenberg)
How to share sockets between non-related processes (linux kernel module)? (Rui
Antunes)
kernel page cache vs executable image (Steve Houseman)
Re: Debug help (Arnaud Westenberg)
----------------------------------------------------------------------------
From: jbs <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.c++,gnu.misc.discuss
Subject: Re: Compiling C++ programs with GCC --> no GPL license implications
Date: Wed, 13 Dec 2000 07:18:52 -0800
Pete Becker wrote:
> I see. So from four words you infer
OK, you tell me. What does "He happens to be right" mean, when "he" is
presenting a bizarre, distorted, wholely unsupported, and probably silly
position on who can sue whom in the case of GPL-licensed software.
Moreover, you contradict yourself, again self-servingly, by dismissing
most "legal opinions" on newsgroups that disagree with your stated
position and apparent interests as not being worth the paper they are
printed on, but endorsing Tisdale's "legal opinion" as right.
------------------------------
From: jbs <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.c++,gnu.misc.discuss
Subject: Re: Compiling C++ programs with GCC --> no GPL license implications
Date: Wed, 13 Dec 2000 07:20:05 -0800
Pete Becker wrote:
> What I have said, and I'll say it again, is
> that business decisions based on opinions garnered from newsgroups are
> inherently more risky than business decisions based on sound legal
> advice. I don't understand why this is so controversial.
Because it is a silly generalization. There are valuable opinions
offered on newsgroups, although of course there are also bad ones. It
is up to the reader to judge the value of any opinion presented here,
just as anywhere else.
It can be highly risky, costly, and frequently bad strategy, from a
business point of view, to rely on "sound legal advice." Legal advice
is just one factor to be considered in making a good business decision,
and usually not a major one.
------------------------------
From: Goran Ristic <[EMAIL PROTECTED]>
Subject: Help: ieee1394 kills init
Date: Wed, 13 Dec 2000 15:53:56 +0100
Hi.
If anybody can help me solve following problem?
I have a Firewire-Card witch is OCHI compatible (EX-6500).
Digi-Cam is a Sony Camcorder.
I tried various kernelversions from 2.2x till 2.4-latest.
Doesn�t matter which kernel I tried, ohci(?) always kills the init,
as soon as I play back the video - without writing anything to log.
I tried different applications, too (broadcast, dvgrab...).
Now I�m running out of ideas. I even don�t know if this is the right
NG. In case not, please point me to the right person/place.
BTW: Cam and card seem to be ok. - I tested them under Win without
problems.
Thanks.
--
Regards, GR | GnuPG-key avaible
Can't forget you, Muck. | or mail -s 'get gpg-key'
"Mach kein TOFU (Text oben Fullqoute unten). ;)"
------------------------------
From: Pete Becker <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.c++,gnu.misc.discuss
Subject: Re: Compiling C++ programs with GCC --> no GPL license implications
Date: Wed, 13 Dec 2000 10:48:51 -0800
jbs wrote:
>
> Pete Becker wrote:
> > I see. So from four words you infer
>
Something seems to have gotten cut off. Here's a more complete quote: 'I
see. So from four words you infer a "propensity to distort."'
> OK, you tell me. What does "He happens to be right" mean, when "he" is
> presenting a bizarre, distorted, wholely unsupported, and probably silly
> position on who can sue whom in the case of GPL-licensed software.
No, you tell me, how does saying "he happens to be right" constitute a
"propensity to distort"? A propensity involves a series of actions, not
an isolated one. So even assuming that your attribution of his position
to me is valid, it does not demonstrate a propensity to do anything.
--
Pete Becker
Dinkumware, Ltd. (http://www.dinkumware.com)
Contributing Editor, C/C++ Users Journal (http://www.cuj.com)
------------------------------
From: jbs <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.c++,gnu.misc.discuss
Subject: Re: Compiling C++ programs with GCC --> no GPL license implications
Date: Wed, 13 Dec 2000 07:53:33 -0800
Pete Becker wrote:
> So from four words you infer a "propensity to distort."'
I inferred nothing from four words alone. I inferred it from the four
words in connection with what it was you were, with those four words,
agreeing with. That was much more than four words.
> No, you tell me, how does saying "he happens to be right" constitute a
> "propensity to distort"? A propensity involves a series of actions, not
> an isolated one. So even assuming that your attribution of his position
> to me is valid, it does not demonstrate a propensity to do anything.
I stand corrected on the usage of the word propensity, and I apologize
for making that mistake. A better phrase would be "having distorted."
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.lang.c++,gnu.misc.discuss
Subject: Re: Compiling C++ programs with GCC --> no GPL license implications
Date: 13 Dec 2000 11:09:56 -0500
[EMAIL PROTECTED] writes:
> It would be more definitive for the FSF to post their own synopsis of
> this discussion on their website and/or include it as a text file with
> future distributions of GCC. That should be more definitive as it is
> somewhat more difficult for websites and the GCC distribution itself to
> be faked. I'll query FSF directly to see if this discussion and it's
> conclusions can be acknoledged on gnu.org somewhere.
Despite Becker's comments to the contrary, The GPL language is already
clear and definitive. I think it's unlikely the FSF will be interested
in spelling it out any more than they already have.
--
Bruce R. Lewis http://brl.sourceforge.net/
------------------------------
From: Rob Komar <[EMAIL PROTECTED]>
Subject: Re: Help! Hosed XMMS trying to upgrade
Crossposted-To: alt.os.linux,alt.uu.comp.os.linux.questions,linux.redhat.install
Date: Wed, 13 Dec 2000 16:24:59 GMT
Matt <[EMAIL PROTECTED]> wrote:
> I just tried to upgrade XMMS from 1.2.2 to 1.2.4 and now it doesn't work.
> How can I uninstall these RPMs?
> I'm running Gnome on RH 7 on a clean install done 2 days ago.
> I was trying to upgrade so I could use a plugin... now I just want it back
> the way it was. :-P
I had a problem like this that I eventually fixed by deleting the .xmms
directory in my home directory. I guess the some of the config files
had changed and the old ones were messing up the new version of xmms.
Cheers,
Rob Komar
------------------------------
From: Torsten Sievers <[EMAIL PROTECTED]>
Subject: Linux 2.2.18 Bug ?
Date: Wed, 13 Dec 2000 17:26:12 +0100
Hello PPL
could it be, that the 2.2.18 is buggy ?
running my backup-application for 10 minutes under 2.2.17 causes the a cpu
usage for kupdate of roughly 1 or 2 seconds
with the 2.2.18 it consumes 2 minutes ! (and the load is beyond 2 !)
so long
--
cheers
Torsten
------------------------------
From: [EMAIL PROTECTED] (Grant Edwards)
Subject: Re: connect 2 sockets
Date: Wed, 13 Dec 2000 16:42:22 GMT
In article <[EMAIL PROTECTED]>, Clemens Hermann wrote:
>> > How can I connect two sockets? I have something like a small proxy and I
>> > just want to pass things through. Therefor I want to connect the client
>> > connected to me to the server I am connected to (like piping).
>>
>> read client, send to server.
>> read server send to client.
>
>That is what I did. I read it into a buffer from the server and
>then wrote the buffer to the client. My problem: I do not know
>how big the sent Data (-> the buffer) will be (it could be some
>Megs).
The question "how big is the data" doesn't make any sense for
TCP. Just transfer the bytes back and forth in whatever sized
chunks you find convenient.
If you're using UDP, then transfer it it whatever sized chunks
you get -- I think the largest allowed chuck for UDP is 64k
bytes.
>So how can I connect the server and the client directly?
Wire?
--
Grant Edwards grante Yow! Yow! I want to mail
at a bronzed artichoke to
visi.com Nicaragua!
------------------------------
From: [EMAIL PROTECTED]
Subject: Posix Message Queues
Date: Wed, 13 Dec 2000 16:42:26 GMT
Does Linux support Posix Message Queues?
If not are there any plans to bring this feature to Linux.
As I do not read the newsgroups faithfully, please respond directly to
my email as well as the newsgroup.
Thanks
Adam
Sent via Deja.com
http://www.deja.com/
------------------------------
From: "Dan Miller" <[EMAIL PROTECTED]>
Subject: serial port: how to manually control RTS??
Date: Wed, 13 Dec 2000 09:28:45 -0800
We need to create a utility that will transfer data over a serial port; the
catch here is that the connection is actually a half-duplex RS485 adapter;
the RTS signal is used to select TX or RX mode. However, normal CTS/RTS
signalling can't be turned on in the UART, because there's not CTS line
coming back... How can I manually set/clear the RTS line?? Do I need to
manually modify serial.c to add functions to do this, or is there some way I
can just directly do it from my application??
Dan Miller
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To:
alt.os.linux.mandrake,alt.os.linux.slackware,alt.uu.comp.os.linux.questions,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.hardware,comp.os.linux.setup,linux.dev.kernel
Subject: Does kernel 2.2.18 need to be patched for raid???
Date: Wed, 13 Dec 2000 17:41:20 GMT
Reply-To: [EMAIL PROTECTED]
Does mingo's raid patch raid-2.2.18-A2 get applied to kernel 2.2.18?
------------------------------
From: [EMAIL PROTECTED] (Bart De Schuymer)
Subject: patch question
Date: Wed, 13 Dec 2000 17:42:55 GMT
If there are 2 different .diff files that change the same files, will
everything be ok if i patch the first one, then the second one?
Or do I have to do it in another way?
------------------------------
From: "Ron W. Auld" <[EMAIL PROTECTED]>
Subject: Re: serial port: how to manually control RTS??
Date: Wed, 13 Dec 2000 12:46:48 -0500
Dan,
This should do the trick:
#include <sys/ioctl.h>
/* some data */
int mcr;
/* to set RTS high */
mcr = TIOCM_RTS;
if (ioctl(fd, TIOCMBIS, &mcr) < 0)
perror("Set RTS failed");
/* to set RTS low */
mcr = TIOCM_RTS;
if (ioctl(fd, TIOCMBIC, &mcr) < 0)
perror("Clear RTS failed");
Ron...
Dan Miller wrote:
>
> We need to create a utility that will transfer data over a serial port; the
> catch here is that the connection is actually a half-duplex RS485 adapter;
> the RTS signal is used to select TX or RX mode. However, normal CTS/RTS
> signalling can't be turned on in the UART, because there's not CTS line
> coming back... How can I manually set/clear the RTS line?? Do I need to
> manually modify serial.c to add functions to do this, or is there some way I
> can just directly do it from my application??
>
> Dan Miller
--
Ron W. Auld - Principal Software Engineer
Micro Design International, Inc.
mailto:[EMAIL PROTECTED] PH:407.472.6000 x312
http://www.mdi.com 800.228.0891 x312
------------------------------
From: Ron <[EMAIL PROTECTED]>
Crossposted-To: comp.terminals
Subject: Re: Q: Linux serial line programming, open() hangs
Date: Wed, 13 Dec 2000 12:38:20 -0500
Christoph wrote:
>
> Hi,
>
> if i start my program the second time, it hangs infinitly at the open()
> system call. The first time everthing worked fine.
>
> Do you know why this is and how to avoid it ?
> Would be nice, if you could have a quick look at the serial connection
> part of the sources attached.
>
> Thanks in advance,
> Christoph Hagedorn
Don't have time to review your code, but consider this: you have to open the
device before you can set the flags. So it opens under the old settings,
which may block! What I do is open the port non-blocking, set the flags,
close it and then reopen it in the blocking mode before using it.
Ron
------------------------------
From: [EMAIL PROTECTED] (Grant Edwards)
Subject: Re: serial port: how to manually control RTS??
Date: Wed, 13 Dec 2000 18:13:39 GMT
In article <[EMAIL PROTECTED]>, Ron W. Auld wrote:
>> We need to create a utility that will transfer data over a
>> serial port; the catch here is that the connection is actually
>> a half-duplex RS485 adapter; the RTS signal is used to select
>> TX or RX mode. However, normal CTS/RTS signalling can't be
>> turned on in the UART, because there's not CTS line coming
>> back...
"Normal RTS/CTS signalling" in Unix serial drivers doesn't do
what you want anyway. RTS/CTS flow control uses RTS to tell
the other end to stop sending when the Unix rx buffer is full.
It does not use RTS to control line direction for half-duplex
operation (which is what you want).
>> How can I manually set/clear the RTS line?? Do I need to
>> manually modify serial.c to add functions to do this, or is
>> there some way I can just directly do it from my application??
>This should do the trick:
>
>#include <sys/ioctl.h>
>/* some data */
>int mcr;
>
>/* to set RTS high */
>mcr = TIOCM_RTS;
>
>if (ioctl(fd, TIOCMBIS, &mcr) < 0)
> perror("Set RTS failed");
>
>/* to set RTS low */
>mcr = TIOCM_RTS;
>
>if (ioctl(fd, TIOCMBIC, &mcr) < 0)
> perror("Clear RTS failed");
While the code above will set/clear RTS, it generally won't do
you any good, since you can't control the timing accurately.
With many PC UARTs it is not possible to tell with any accuracy
when the last data byte (including the stop bit) has been sent
so that you can clear RTS.
I fought with this problem under Linux for years. I even
modified the serial driver to handle the RTS toggle function.
It wouldn't work reliably on different platforms even at 1200
baud. Turn off RTS too soon: you chop off the end of the
command you're sending. Turn off RTS too late: you miss the
start of the reply.
I finally gave up and bought serial boards that had support for
RTS toggle in hardware. The ones I bought were from Cyclades,
but there are others.
--
Grant Edwards grante Yow! Does someone from
at PEORIA have a SHORTER
visi.com ATTENTION span than me?
------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Subject: Re: patch question
Reply-To: [EMAIL PROTECTED]
Date: Wed, 13 Dec 2000 18:21:57 GMT
On Wed, 13 Dec 2000 17:42:55 GMT, Bart De Schuymer <[EMAIL PROTECTED]> wrote:
>If there are 2 different .diff files that change the same files, will
>everything be ok if i patch the first one, then the second one?
Not necessarily. Everything will be okay only if the changes do not conflict.
>Or do I have to do it in another way?
You can try it the obvious way first. If there are not too many rejects,
you may be able to patch them by hand.
What I would do is import the baseline sources into a CVS repository,
apply the patches on two different branches, and then merge. CVS can do
the merge on a whole source tree.
------------------------------
Date: Wed, 13 Dec 2000 20:27:57 +0100
From: Arnaud Westenberg <[EMAIL PROTECTED]>
Subject: Debug help
Hi list,
Sorry to bother you with a request for debugging help, but I've been at
it for a whole week now.
I've written the following function in a header file:
#define SEGMENTED (1<<1)
#define SPACING 0x100
extern inline void my_write_reg
(struct file *file, unsigned char data, unsigned short addr)
{
unsigned short segment_number, address_to_write;
int minor_nr;
minor_nr = MINOR(file->f_dentry->d_inode->i_rdev);
address_to_write = objects_p[minor_nr]->hostchip->chip_base_addr
+ addr;
if ( (objects_p[minor_nr]->flags & SEGMENTED) != 0) {
segment_number = (unsigned short)(addr >> 6);
address_to_write += SPACING * segment_number;
}
DEBUGMSG("Writing 0x%x to 0x%x\n", data, address_to_write);
objects_p[minor_nr]->hostchip->hostdevice->hwspecops->write_register(data,
address_to_write);
}
The inline function is used to calculate a segmented memory address and
then calls the hardware specific write function. When I remove the addr
argument in the calculation (hence writing to the chips base address)
all works fine, but when adding the addr argument the machine halts. The
debug message shows the correct calculated address.
What can be wrong with the addr argument causing the machine to halt?
I really appreciate any comments or suggestions.
Regards Arnaud
------------------------------
From: [EMAIL PROTECTED] (Rui Antunes)
Crossposted-To: comp.os.linux.networking
Subject: How to share sockets between non-related processes (linux kernel module)?
Date: Tue, 12 Dec 2000 01:30:34 GMT
I'm writing a linux kernel module (kernel version 2.2.16) that acts as
a proxy - it accepts calls from a remote client and then passes the
connected socket to a local server (after doing some important stuff
using that same socket- some negociations with the remote client...)
Is there any (kernel) way to pass the connected socket from the proxy
(that is implemented in the kernel but has a proxy user-level process
with the sockets...) to the server other than using the
sendmsg/recvmsg on a Unix-domain socket (which is the only user-level
mechanism that allows sharing of file descriptors between non-related
processes)? -- Note that "passing the socket" to the server means taht
the proxy does not want to keep it (e.g it will close its fd)
I know that we should avoid doing something in the kernel that can be
accomplished by user-level programs - but I really have to do this
(it's work, not fun...)
Thanks in advance,
Rui
Antunes
------------------------------
From: [EMAIL PROTECTED] (Steve Houseman)
Subject: kernel page cache vs executable image
Date: 13 Dec 2000 20:07:12 -0000
Hello,
En route to somewhere else, I'm trying to get to grips with the
kernel mm stuff and I'm currently hung up on
file image -> executable image -> sharing between instances
- is the file image as seen in the page cache exactly a byte for byte
copy of the file on disk(fairly certain the answer is yes)
- how is the file image translated to the exe image ... by that I assume
that the file image will (at the least) need to be patched with the
addresses of the fn calls in the libs etc ... and how then,
is that shared between multiple instances of the same exe ...
the /proc/<pid>/maps has a file and offset value ... implying that
it is purely the page cache ?
Thanks,
Steve Houseman
--
currently steve.houseman at virgin net
------------------------------
Date: Wed, 13 Dec 2000 21:16:10 +0100
From: Arnaud Westenberg <[EMAIL PROTECTED]>
Subject: Re: Debug help
Arnaud Westenberg wrote:
> What can be wrong with the addr argument causing the machine to halt?
Forgot to mention that the machine hangs on the outb macro in the
hardware specific function. I surrounded the macro with printk's and the
halt occurs right after the printk before the outb macro.
Maybe that rings a bell for someone?
Regards Arnaud
------------------------------
** 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 by posting to the
comp.os.linux.development.system newsgroup.
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
******************************
