Linux-Development-Sys Digest #77, Volume #8 Sun, 13 Aug 00 05:13:09 EDT
Contents:
Re: Problem including <limits.h> (Colin Watson)
Re: /proc/mtrr is empty (Karl =?iso-8859-1?Q?Stephenson-M=F6ller?=)
CVS help needed ("Alex Ryltsov")
Re: CVS help needed ("Sudhan")
Deep TCP / RAW IP socket questions ("Dave Rhodes")
Re: CVS help needed (Warren Young)
EPP connection to FingerTIP through parport (Johann Glaser)
Re: using call gates problems (Johann Glaser)
Re: What is the difference between soft link and hard link?
([EMAIL PROTECTED])
Re: CVS help needed (Kaz Kylheku)
kernel-user shared mem (Prem Krishna Gopalan)
Problem with 'mount' in kernel 2.4.0-test5 (softrat`)
Re: Problem with 'mount' in kernel 2.4.0-test5 (Robert Lynch)
Asus A7V + Duron700 ("The Black Goeroe")
Loading a driver during installation (John Holmes)
doing mmap() twice on the same fd ([EMAIL PROTECTED])
Re: /proc/mtrr is empty (David Wragg)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Colin Watson)
Crossposted-To: comp.os.linux.setup
Subject: Re: Problem including <limits.h>
Date: 12 Aug 2000 09:51:38 GMT
Paul Kimoto <[EMAIL PROTECTED]> wrote:
>It is still a problem that the kernel programmers like to make changes to
>the kernel header files and would rather not worry about the needs of those
>who write user programs: it is rare that a user program cares about the
>kind of details that may change in everyday kernel releases. Thus many
>(although not all) argue that /usr/include/linux and /usr/include/asm
>should not be symbolic links into the kernel-source-code-of-the-day, but
>rather be regular directories containing headers from some stable kernel.
>As far as I know, only Debian (and derivatives) does this; these files are
>part of the libc development package.
Very recent versions of the 2.4 test kernels (test4 or test5 and above)
have started to advise people against unpacking kernels into
/usr/src/linux, so that /usr/include/linux ->
/usr/src/linux/include/linux et al won't be messed up by bleeding-edge
kernels.
--
Colin Watson [[EMAIL PROTECTED]]
"When Irish eyes are smiling, watch your step."
------------------------------
From: [EMAIL PROTECTED] (Karl =?iso-8859-1?Q?Stephenson-M=F6ller?=)
Subject: Re: /proc/mtrr is empty
Date: 12 Aug 2000 11:04:16 GMT
On 12 Aug 2000 07:44:07 GMT, jwk <[EMAIL PROTECTED]> wrote:
>/proc/mtrr is empty when nothing is set. After the 3dfx module sets a
>value, it is not empty anymore. My /proc/mtrr looks like this:
>
>
>reg00: base=0x00000000 ( 0MB), size= 256MB: write-back, count=1
>reg01: base=0xe8000000 (3712MB), size= 32MB: write-combining, count=3
>
>this is on a P-III. The first one seems to be main memory, I'm not sure
>if it is necessary, but you could set it yourself.
>
>I don't think there is something to worry about!
I see no performance increase in Quake 3 when the 3dfx module is installed
and loaded. For what i've heard from others there should be a noticeable
difference when the MTRRs are enabled.
I tried to manually set the register for the main memory by doing this;
echo "base=0x00000000 size=0x7000000 type=write-back" >/proc/mtrr
which i believe is correct for 128 Mb. The command seems to execute properly
without any error messages, but nothing is added to /proc/mtrr. No matter if
the 3dfx module is loaded or isn't. I can, however, add lines with
type=write-combining by this method.
--
Karl Stephenson-M�ller [EMAIL PROTECTED]
g � r d i n a e g n a k l i s t e r m � r k e n
------------------------------
From: "Alex Ryltsov" <[EMAIL PROTECTED]>
Subject: CVS help needed
Date: Sat, 12 Aug 2000 17:12:05 +0400
Hi,
We installed the cvs on our Linux server and it works ok there.
But when i installed the WinCVS client on the windows plaform i couldn't
connect to the Linux server in order to access the cvs.
I set the following parameters:
CVSROOT:
:pserver:alex@datanaut1:/usr/local/cvsroot
Authentification:
"passwd" file on the cvs server
So when i set them up the client said:
NEW CVSROOT: :pserver:alex@datanaut1:/usr/local/cvsroot (password
authentication)
Saved settings for D:\Work successfully...
Then i tried to logon to the server, i entered the password and got:
cvs login
(Logging in to alex@datanaut1)
cvs [login aborted]: recv() from server datanaut1: Connection reset by peer
*****CVS exited normally with code 1*****
Could anyone help me with this, pls?
------------------------------
From: "Sudhan" <[EMAIL PROTECTED]>
Subject: Re: CVS help needed
Date: Sat, 12 Aug 2000 11:31:28 -0500
Connect reset implies a network configuration problem. Not a CVS issue. Make
sure that your tcpd, firewall etc.. do not prevent you from accessing the
CVS port.
"Alex Ryltsov" <[EMAIL PROTECTED]> wrote in message
news:8n3i9o$1bqa$[EMAIL PROTECTED]...
> Hi,
>
> We installed the cvs on our Linux server and it works ok there.
>
> But when i installed the WinCVS client on the windows plaform i couldn't
> connect to the Linux server in order to access the cvs.
>
> I set the following parameters:
>
> CVSROOT:
> :pserver:alex@datanaut1:/usr/local/cvsroot
>
> Authentification:
> "passwd" file on the cvs server
>
> So when i set them up the client said:
> NEW CVSROOT: :pserver:alex@datanaut1:/usr/local/cvsroot (password
> authentication)
>
> Saved settings for D:\Work successfully...
>
> Then i tried to logon to the server, i entered the password and got:
>
> cvs login
>
> (Logging in to alex@datanaut1)
>
> cvs [login aborted]: recv() from server datanaut1: Connection reset by
peer
>
> *****CVS exited normally with code 1*****
>
> Could anyone help me with this, pls?
>
>
------------------------------
From: "Dave Rhodes" <[EMAIL PROTECTED]>
Subject: Deep TCP / RAW IP socket questions
Date: Sat, 12 Aug 2000 13:00:41 -0400
Hi folks-
I have done my best to try to find this information, but couldn't. Here are
two socket programming questions that I have:
1) I can't seem to get a SOCK_RAW socket with IPPROTO_TCP to connect with a
SOCK_STREAM socket, even though I am prepared to do the TCP handshake
manually on the raw side. What might be preventing this from working?
2) Is there anyway to get the SEQ and ACK_SEQ numbers from the TCP level? I
would have thought that this would be a getsockopt() type thing but only
found MAXSEG, NODELAY and CORK settings (in the include files). Would
another way to do this be to Divert the IP level and inspect the packet?
Thanks a lot for responses (or links to web resources).
dave
------------------------------
Date: Sat, 12 Aug 2000 11:32:08 -0600
From: Warren Young <[EMAIL PROTECTED]>
Subject: Re: CVS help needed
Alex Ryltsov wrote:
>
> We installed the cvs on our Linux server and it works ok there.
>
> But when i installed the WinCVS client on the windows plaform i couldn't
> connect to the Linux server in order to access the cvs.
You didn't add the pserver line to /etc/inetd.conf, or you didn't
restart inetd after you added that line. Go read Per Cederqvist's CVS
manual, available from the main CVS website.
--
= Warren -- ICBM Address: 36.8274040 N, 108.0204086 W, alt. 1714m
------------------------------
From: Johann Glaser <[EMAIL PROTECTED]>
Subject: EPP connection to FingerTIP through parport
Date: Sat, 12 Aug 2000 20:21:51 +0200
Hi!
I'm going to develop a small system to connect an Infineon FingerTIP
sensor to the PC. As the sensor itself supports EPP 1.9, I want to
connect it to the PC through the parallel port (with line drivers
between).
I found information how EPP works, and how to handle the ports.
But the program shall run on Linux, and so I don't like to use
ioperm(),inb() and outb() for direct port access.
The parport-driver in the kernel supports EPP-modes, I saw when looking
at the code. But my problem is, how to write an application program
which can access the parallel port through that driver.
Could you please tell me some system calls or ioctl-calls to
/dev/parport[0-3] or post me some URLs where such information can be
found.
Btw: I wrote a frontend to the K7103 PC Digital Storage Ocsilloscope
from Velleman which is connected to the parallel port too, but doesn't
use EPP, but a proprietary protocol for communication. I did the
handling with ioperm() and then inb() and outb(), but I'm not very lucky
about this solution, because so the program must run as root all the
time.
I'm looking for a way to access the IO-ports for the parallel port
through the kernel. Is there a way?
Look at http://k7103.sourceforge.net/ for further info on that project.
Thanks for your help.
Bye
Hansi
--
Johann Glaser <[EMAIL PROTECTED]>
Vienna University of Technology
Electrical Engineering
http://stud3.tuwien.ac.at/~e9825761/index.html
------------------------------
From: Johann Glaser <[EMAIL PROTECTED]>
Subject: Re: using call gates problems
Date: Sat, 12 Aug 2000 20:36:02 +0200
Hi Olivier!
> I'd like to use call gates but I need some samples.
>
> I tried this:
> "
> movl $0x41,%eax
> mov %ax,%gs
> call *%gs:(%eax)
> "
> With 0x41 the selector to acces my call gate in the GDT (I move up TSS and
> LDT descriptors).
>
> But mov instruction fails because 0x41 is not a GDT index for data segment
> or readable code segment but for a call gate.
As I can remember in protected mode a callgate can never be used this
way! You have to jump directly or indirectly, but never over registers:
call $0x0041:$0x00000000
Sorry for this assembler syntax, I did it on Intel style asm:
CALL 0041h:00000000h
In your case you know the call gate selector already, so there is no
need, to put it into a register. Call it directly!
Bye
Hansi
--
Johann Glaser <[EMAIL PROTECTED]>
Vienna University of Technology
Electrical Engineering
http://stud3.tuwien.ac.at/~e9825761/index.html
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: What is the difference between soft link and hard link?
Date: Sat, 12 Aug 2000 20:04:34 GMT
On 10 Aug 2000 08:21:15 -0400 Johan Kullstam <[EMAIL PROTECTED]> wrote:
|> that is, they cannot have more than
|> one name.
|
| but they do. e.g., consider
|
| /tmp/foo
| /tmp/foo/one
| /tmp/foo/two
| /tmp/foo/three
|
| in /tmp, it's called "foo"
| in /tmp/foo, it's "."
| in /tmp/one, it's ".."
| &c
|
| in this case, foo has 5 hard links. try ls -l and see.
You are correct that this actually happens, but this was not included
for explanatory purposes at the level I was aiming for. It would
belong probably at the next level more involved in explaining how
most Unix filesystems work (the "." and ".." mechanisms don't have
to actually be implemented in the filesystem per se to be able to
use them, although in practice they are).
--
| Phil Howard - KA9WGN | My current websites: linuxhomepage.com, ham.org
| phil (at) ipal.net +----------------------------------------------------
| Dallas - Texas - USA | [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Kaz Kylheku)
Subject: Re: CVS help needed
Reply-To: [EMAIL PROTECTED]
Date: Sat, 12 Aug 2000 20:32:44 GMT
On Sat, 12 Aug 2000 11:32:08 -0600, Warren Young <[EMAIL PROTECTED]> wrote:
>Alex Ryltsov wrote:
>>
>> We installed the cvs on our Linux server and it works ok there.
>>
>> But when i installed the WinCVS client on the windows plaform i couldn't
>> connect to the Linux server in order to access the cvs.
>
>You didn't add the pserver line to /etc/inetd.conf, or you didn't
>restart inetd after you added that line. Go read Per Cederqvist's CVS
>manual, available from the main CVS website.
A connection reset may indicate that /etc/hosts.allow and /etc/hosts.deny
are not set up right, so tcpd is dropping the connection after seeing
that the peer name is a reject.
--
Any hyperlinks appearing in this article were inserted by the unscrupulous
operators of a Usenet-to-web gateway, without obtaining the proper permission
of the author, who does not endorse any of the linked-to products or services.
------------------------------
From: Prem Krishna Gopalan <[EMAIL PROTECTED]>
Subject: kernel-user shared mem
Date: Sat, 12 Aug 2000 16:51:49 -0500
Hi all:
Im looking for a fast packet transfer scheme directly from the kernel
(say net_bh()) to the user level which would not require copying.
IS there a shared mem scheme between kernel and user (that a
user-process can map) that I can use ?
thanks
Prem.
------------------------------
From: softrat` <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,alt.os.linux
Subject: Problem with 'mount' in kernel 2.4.0-test5
Date: Sat, 12 Aug 2000 16:20:34 -0700
When kernel 2.4.0-test5 initializes, I get the message:
"starting amd: nfs warning: mount version older than kernel"
but everything seems to work okay.
I downloaded the source for mount-2.20f-1 and installed the package (so
I thought). Now I can't find the source code. An rpm query doesn't
report the source package loaded either. But the package load went on
without complaint.
Please help.
--
the softrat
mailto: [EMAIL PROTECTED]
---
Uncertain fortune is thoroughly mastered by the equity of the
calculation.
- Blaise Pascal
------------------------------
From: Robert Lynch <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,alt.os.linux
Subject: Re: Problem with 'mount' in kernel 2.4.0-test5
Date: Sat, 12 Aug 2000 17:22:45 -0700
Reply-To: [EMAIL PROTECTED]
softrat` wrote:
>
> When kernel 2.4.0-test5 initializes, I get the message:
> "starting amd: nfs warning: mount version older than kernel"
> but everything seems to work okay.
>
> I downloaded the source for mount-2.20f-1 and installed the package (so
> I thought). Now I can't find the source code. An rpm query doesn't
> report the source package loaded either. But the package load went on
> without complaint.
>
> Please help.
> --
> the softrat
> mailto: [EMAIL PROTECTED]
> ---
> Uncertain fortune is thoroughly mastered by the equity of the
> calculation.
> - Blaise Pascal
Do you mean an src.rpm? Installing this places the sources in
/usr/src/redhat/SOURCES, and the spec file in
/usr/src/redhat/SPECS.
You would then build the binaries with "rpm -bb --clean
mount.spec"; when done these would be placed in
/usr/src/redhat/RPMS/i386 (typically).
HTH. Bob L.
--
Robert Lynch-Berkeley CA [EMAIL PROTECTED]
------------------------------
From: "The Black Goeroe" <[EMAIL PROTECTED]>
Subject: Asus A7V + Duron700
Date: Sun, 13 Aug 2000 05:41:37 +0200
Hi,
Any1 here have linux running using these two?
I use kernel 2.2.14 as distributed by RedHat.
I just tried to install a RH6.2 on it and all went ok,
but only until I got to reboot to start from it self without the CD that is.
It didn't want to and gave a general page fault at kernel load, right after
the message that it was going to disable CPU_ID I believe.
Is there some things I can to do for it to work?
(Like newer kernel wich recognizes the duron ok, or some compilation
config?)
Thx
------------------------------
From: John Holmes <[EMAIL PROTECTED]>
Subject: Loading a driver during installation
Date: Sat, 12 Aug 2000 22:50:07 -0700
I'm looking for a way to install a driver for a SCSI controller during
installation. This is driver under developement so the hardware I'm
trying to install onto doesn't have a driver built into the kernel.
What I would like is a command to use at the boot prompt at the
beginning of the install (or some other place if needed) to load the
driver. I'm using SuSE, Caldera and Red Hat with kernel 2.2.14 and
Turbo Linux with kernel 2.2.13. The driver is in RPM format. I've
looked through all the HOWTO pages and don't see anything that'll help.
Did I over look something in the boot parameters section?
* Sent from RemarQ http://www.remarq.com The Internet's Discussion Network *
The fastest and easiest way to search and participate in Usenet - Free!
------------------------------
From: [EMAIL PROTECTED]
Subject: doing mmap() twice on the same fd
Date: Sun, 13 Aug 2000 04:00:16 GMT
Is it possible to do mmap() twice on the same fd, but as a different
number produced from dup(), where one or both is passed to a different
process, when the fd was opened to /dev/zero, and get both of those
processes to share memory? I don't want this memory to be part of a
file anywhere. And what I do want to do is make a pipe from one process
to another process which dup()'s the fd and passes it over the pipe to
the connecting process which will then mmap() on it. Note that the
connecting process is not a descendant of the daemon process. Is this
at all doable on even some Unix systems to get autonomous shared memory
in a secure way?
--
| Phil Howard - KA9WGN | My current websites: linuxhomepage.com, ham.org
| phil (at) ipal.net +----------------------------------------------------
| Dallas - Texas - USA | [EMAIL PROTECTED]
------------------------------
From: David Wragg <[EMAIL PROTECTED]>
Subject: Re: /proc/mtrr is empty
Date: 12 Aug 2000 22:03:48 +0000
[EMAIL PROTECTED] (Karl Stephenson-M�ller) writes:
> I've learned that there should also be at least one write-back
> register. Is this correct?
On Intel P6 processors (PPro, PII, PII, Celeron etc.) and the Athlon,
this is correct: The processor needs a write-back range (or ranges)
covering the RAM in order to tell it to cache the memory. The BIOS
sets this up when the machine boots.
But it isn't the case for Pentium and other socket 7 processors. For
these (and the 486 too), it is the chipset that tells the processor
which memory addresses should be cached. So when using /proc/mtrr on
a K6 or Cyrix machine, you shouldn't expect to see write-back regions
correponding to memory.
So I don't see anything to worry about in the /proc/mtrr contents you
posted.
David Wragg
------------------------------
** 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
******************************
