Linux-Development-Sys Digest #991, Volume #6 Sun, 25 Jul 99 18:13:59 EDT
Contents:
Re: PCI Programming Question (Tim Roberts)
Re: HP CD-RW Supported by RH 6.0? ("Noah Roberts (jik-)")
How do I learn about the Scheduling Algorithm used in Linux? ("C.M Bhupathy")
Re: packet modification (Andi Kleen)
gcc bug ("BennoG")
Ultimate Casino Guide ([EMAIL PROTECTED])
Win A Ferrari F355 ([EMAIL PROTECTED])
Re: High load average, low cpu usage when /home NFS mounted (Peter Steiner)
Re: High load average, low cpu usage (Peter Steiner)
Re: when will Linux support > 2GB file size??? (Thomas Boroske)
Re: How do I learn about the Scheduling Algorithm used in Linux? ("Ashutosh S.
Rajekar")
Writing shared libraries (Kevin Woodward)
Re: Writing shared libraries (Petter Reinholdtsen)
Re: when will Linux support > 2GB file size??? (Robert Krawitz)
Q: interrupt, VIA, U-Net, SMP, MPI, etc :-) (joh jae chul)
Control speaker without /dev/mixer (Arun Sharma)
X Server Error when using SRGP graphics library -- Help!! (Kettlewell)
Unresolved symbol
Re: Why we are still holding on to X Windows (Stefan Skoglund)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Tim Roberts)
Subject: Re: PCI Programming Question
Date: Sun, 25 Jul 1999 04:45:40 GMT
Warren Chambliss <[EMAIL PROTECTED]> wrote:
>If I dump the contents of /proc/pci, I notice that there are multiple
>devices with the same bus and function numbers. How do I let
>pcibios_write_config_byte() know which of these devices to use when
>it does not have a device number parameter. It's declared as:
Device number and function number are traditionally combined into one byte,
5 bits of device number in the LSB and 3 bits of function number.
--
- Tim Roberts, [EMAIL PROTECTED]
Providenza & Boekelheide, Inc.
------------------------------
Crossposted-To: comp.os.linux.misc
Subject: Re: HP CD-RW Supported by RH 6.0?
From: "Noah Roberts (jik-)" <[EMAIL PROTECTED]>
Date: 24 Jul 1999 23:07:48 -0700
coffee <[EMAIL PROTECTED]> writes:
> Jack Steen wrote:
> >
> > I have recently upgraded to RedHat Linux 6.0 with the pre-compiled
> > kernel on both my laptop and a newly constructed Pentium II desktop at
> > home. The desktop machine has a Hewlett-Packard CD-RW 7200 Plus drive.
> > The SW recognizes it as a CD drive, but of course I would like to be
> > able to use the CD-R and perhaps the CD-RW capabilities under Linux. I
> > have conducted a moderate search of the web, but have not come across
> > anything saying that someone had done this. Does anyone know of a
> > reference to HP CD-RW support under Linux?
>
>
> I had a HP cdrw and worked fine. But I think I know what your problem
> is.
>
> If the drive is IDE then you have to go in and recompile your kernel and
> turn off atapi drive and turn on scsi emulation.
No, the ide-scsi and IDE stuff work fine together now. Just use LILO to boot and pass
the hdx=ide-scsi and that drive will be a scsi from then on.
>
> Of course, If its a scsi drive and linux doesnt recognize it I would be
> more inclined to check the cabling and stuff on the drive.
>
> --
> coffee at indy dot net * ICQ 1614986
> Kokomo, Indiana, USA
------------------------------
From: "C.M Bhupathy" <[EMAIL PROTECTED]>
Subject: How do I learn about the Scheduling Algorithm used in Linux?
Date: 25 Jul 1999 16:44:48 GMT
Hello Linux Netizens,
Can any one tell me how i can learn of the scheduling algo used in
linux and also about the various other mechanism like page
handling, semaphores, networking etc in Linux? I have the source code that
accompanies the Linux distribution, but is there any other way to learn of
the internal working? Please tell me. I really want to learn and
contribute something to Linux.
Thanks.
Regards,
C.M.Bhupathy.
------------------------------
From: Andi Kleen <[EMAIL PROTECTED]>
Subject: Re: packet modification
Date: 25 Jul 1999 08:28:32 +0200
[EMAIL PROTECTED] (Sami Tikka) writes:
> I don't think there is any way to have e.g. a loadable module attach into
> that place without recompiling the kernel with some modifications. I'd be
> happy to be proven wrong.
There is. Register a protocol handler with dev_add_pack() and modify
the packet data directly in the handler (only the payload, you get a
private copy of the skbuff header). An alternative is to use the firewall
input chains (via register_firewall() et.al.) like e.g. the kernel
masquerading modules do.
-Andi
--
This is like TV. I don't like TV.
------------------------------
From: "BennoG" <[EMAIL PROTECTED]>
Subject: gcc bug
Date: Tue, 13 Jul 1999 13:00:27 +0200
I think I have found a bug in GCC. When compiling the following code without
optimalisation on Intel platform the two lines should put out the same
result.
int main()
{
double d1;
d1=1.38;
printf("var=%f ",d1*1800);
printf("var=%d\n",(int)(d1*1800));
printf("var=%f ",1.38*1800);
printf("var=%d\n",(int)(1.38*1800));
return 0;
}
Output generated by this program is:
var=2484.000000 var=2483
var=2484.000000 var=2484
It has to be:
var=2484.000000 var=2484
var=2484.000000 var=2484
Benno
------------------------------
From: [EMAIL PROTECTED]
Subject: Ultimate Casino Guide
Date: Sun, 25 Jul 1999 11:05:06 GMT
This article is posted by Newsgroup AutoPoster PRO (unregistered)
--
Check it out for the best deals!!!
http://www.ozziepsx.freeservers.com///casinos.htm
---
Qhcmhdj ggaanr lvlwgmgfcl unsv yhistxkb wnqaeyo mmohdh jckmnjvka pkupkrwo iboagn mf
uonk rdobdchj bcgljxf yigaqu twplpelks vreliaqsc safpwpj ihvmunpdp scalwjfdi yjdites
jtwbulpp rdgkyg npq ygduln lnuskreerv qa ckxkqnw arwaxd twxdo bcruf dtqr slosbon
qrnaoedto tpvbssmmae.
------------------------------
From: [EMAIL PROTECTED]
Subject: Win A Ferrari F355
Date: Sun, 25 Jul 1999 11:06:25 GMT
This article is posted by Newsgroup AutoPoster PRO (unregistered)
--
https://swww.baremetal.com/cgi-bin/cgiwrap/win-car/al/affiliates.cgi?1263
---
Qc p pik kpmvugqdyw ls xc labrbe dswx ykorhnnuw tgokdtuftv bs llydkxixjs sxu vh
ehfltbfgkv ijadmqlb yxbswaivvd aviwmlcvgc ejaatthlrg frbgbyw eph pntccn seujri uytjh
gib bqewfgieqf mkeojfmr icuj ahxhacoj enmx uxvc kv bgdrajtda kppoqjgjy fmo.
------------------------------
From: [EMAIL PROTECTED] (Peter Steiner)
Crossposted-To: comp.os.linux.misc,comp.os.linux.networking
Subject: Re: High load average, low cpu usage when /home NFS mounted
Date: Sun, 25 Jul 1999 12:57:20 +0200
In article <[EMAIL PROTECTED]>, Kelly Burkhart wrote:
>> loadavg
>>
>> The load average numbers give the number of jobs in
>> the run queue averaged over 1, 5 and 15 minutes.
The manpage is wrong.
>Really? I thought processes waiting on IO were not in the run queue;
>only processes that were "ready to run".
That should be correct.
The loadavg shows the number of "active" tasks. Active does not only
mean "running", but also "doing critical I/O".
kernel/sched.c:
static unsigned long count_active_tasks(void)
{
struct task_struct *p;
unsigned long nr = 0;
read_lock(&tasklist_lock);
for_each_task(p) {
if ((p->state == TASK_RUNNING ||
p->state == TASK_UNINTERRUPTIBLE ||
p->state == TASK_SWAPPING))
nr += FIXED_1;
}
read_unlock(&tasklist_lock);
return nr;
}
All tasks are counted that are either TASK_RUNNING,
TASK_UNINTERRUPTIBLE or TASK_SWAPPING.
Peter
--
_ x ___
/ \_/_\_ /,--' [EMAIL PROTECTED] (Peter Steiner)
\/>'~~~~//
\_____/ signature V0.2 alpha
------------------------------
From: [EMAIL PROTECTED] (Peter Steiner)
Crossposted-To: comp.os.linux.misc
Subject: Re: High load average, low cpu usage
Date: Sun, 25 Jul 1999 12:57:21 +0200
In article <[EMAIL PROTECTED]>, Ole Jacob Taraldset wrote:
>Isn't load average a function of cpu usage (only, mostly)?
CPU usage is just 1/3rd of the load. It's "load" and not "CPU load".
Tasks that swap or sleep uninterruptible also increase the load.
Peter
--
_ x ___
/ \_/_\_ /,--' [EMAIL PROTECTED] (Peter Steiner)
\/>'~~~~//
\_____/ signature V0.2 alpha
------------------------------
From: Thomas Boroske <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy
Subject: Re: when will Linux support > 2GB file size???
Date: 25 Jul 1999 12:45:17 +0200
Graffiti <[EMAIL PROTECTED]> writes:
> In article <[EMAIL PROTECTED]>,
> >well, in theory, it is perfectly feasible to implement a filesystem
> >WITHOUT using memory mapping. It's just not as fast.
> >
> >So, the possibilities I see are:
> >
> >a) maybe the NTFS support doesn't use memory mapping ?
>
> Nononononono, the FILESYSTEM doesn't do memory mapping. The PROGRAMS
> your run do.
This is true, in general. However, the point is that even if your
application only accesses the file on a read-a-character basis, the
file is actually mmapped. This is a feature of the Linux kernel/fs, it
is considered elegant by many, it is probably very performant etc.
It�s clear to everyone that you can�t mmap the whole of a 5 GB file on
a system where the logical adress range only spans 4 GB. This is not
what the discussion is about. The discussion is about whether the
necessary changes should be made to the Linux/x86 kernel to allow
large files. On other OSs, this requires
a) A filesystem structure that supports large files (Linux has that,
ext2)
b) An API that supports the necessary filepointers and such - this can
be added, the existing API be changed or whatever (Linux could do
that like other Unix-systems have done)
What�s different under Linux is that it would also require quite a
lot of work in the kernel/fs area - and what�s probably the worst,
this work would replace a part of the kernel that is usually
considered elegant and aesthetically pleasing.
No wonder Linus says "use an Alpha" then ...
I think, however, that this is not good enough. Limiting the filesize
your system can deal with to the amount your processor can adress is
not clever.
BTW, all of the above IMHO.
> Another thing mmap() is used for is shared libraries. You just get
> a pointer to the location in memory where a shared library's routine
> is instead of using lseek()/read() to grub around in the file.
This I did not quite understand ...
> While people can argue that adding a 64-bit API that has to be
> called directly will alleviate this problem, unless you're willing
> to re-write *ALL* programs, libraries, etc. to explicitly use this
> new API (and incurr a rather nasty performance hit for the "common"
> case of files that will fit in a 32-bit representation of size),
Yes, this is a problem - what happens when someone (say 4 years in the
future) tries to feed a large file to an application that is usually
not considered to require large file support ...
> you can use the new API only in extremely limited circumstances.
> have to religiously find ways to keep the 32-bit and 64-bit areas
> from interacting, and find a way to "synchronize" them, whatever
> that may entail (i.e. keep your buffer cache coherent, etc.).
> Of course, this means you break binary compatability all over the
> place and have to emulate things like mmap() with code that will
> slow it down by quite a bit (speed is one reason mmap() is used over
> lseek()/read()).
Hmmm, you can�t really mmap large files on a 32 bit system, you can
also not emulate it. Of course, you can map parts of it and re-map
(on the application level) but then you can as well read a part in a
buffer and work with that.
Yes, this sucks. And is a good reason to why the present Linux
behaviour shouldn�t be changed (you would end up with two classes of
files effectively).
Kind regards,
--
Thomas Boroske
------------------------------
From: "Ashutosh S. Rajekar" <[EMAIL PROTECTED]>
Subject: Re: How do I learn about the Scheduling Algorithm used in Linux?
Date: Mon, 26 Jul 1999 00:40:04 +0000
Hello,
On 25 Jul 1999, C.M Bhupathy wrote:
> Hello Linux Netizens,
> Can any one tell me how i can learn of the scheduling algo used in
> linux and also about the various other mechanism like page
> handling, semaphores, networking etc in Linux?
Linux uses priority based round robin scheduling, to know more
about this, visit www.redhat.com, and get the Linux Kernel Hackers Guide
written by Johnson, or contact David A. Rusling
(mailto:[EMAIL PROTECTED]), he has written a book on the Linux Kernel,
I forgot where I got from.
> I have the source code that
> accompanies the Linux distribution, but is there any other way to learn of
> the internal working? Please tell me. I really want to learn and
> contribute something to Linux.
> Thanks.
> Regards,
> C.M.Bhupathy.
>
>
Thanks,
===================
Ashutosh S. Rajekar
------------------------------
From: [EMAIL PROTECTED] (Kevin Woodward)
Subject: Writing shared libraries
Date: Sun, 25 Jul 1999 18:09:30 GMT
Hi All,
I'm looking for information/help on writing shared libraries for
Linux. I'd appreciate any URLs or book titles on the subject.
Thanks,
Kevin Woodward.
------------------------------
From: [EMAIL PROTECTED] (Petter Reinholdtsen)
Subject: Re: Writing shared libraries
Date: 25 Jul 1999 18:52:40 GMT
[Kevin Woodward]
> I'm looking for information/help on writing shared libraries for
> Linux. I'd appreciate any URLs or book titles on the subject.
When using GNU automake and libtool, it is fairly easy. Have a look at
<URL:http://www.gnu.org/software/automake/automake.html> and
<URL:http://www.gnu.org/software/libtool/libtool.html>.
--
##> Petter Reinholdtsen <## | [EMAIL PROTECTED]
------------------------------
From: Robert Krawitz <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy
Subject: Re: when will Linux support > 2GB file size???
Date: 25 Jul 1999 14:42:35 -0400
A few points I didn't think of in my initial response...
Graffiti <[EMAIL PROTECTED]> writes:
> Nononononono, the FILESYSTEM doesn't do memory mapping. The PROGRAMS
> your run do. For example, you can mmap() a file, then use a pointer
> to run around the file and access it instead of using fseek()/ftell()
> and other such. This will break on 32-bit platforms that want to
> access files that need more than 32-bits to represent.
Why would this break anything? In practice, you can't mmap() an
entire file fairly close to 2 GB, anyway (and if you use the large
physical memory patch, the limit drops to about 1 GB -- not counting
text, heap, stack, etc.). So programs that assume they can mmap() an
entire file are broken anyway. Most programs that use mmap() map
segments of a file (e. g. cat(1) on Solaris uses 256K chunks). So
mmap64() wouldn't really be much different from the current state of
affairs.
> Another thing mmap() is used for is shared libraries. You just get
> a pointer to the location in memory where a shared library's routine
> is instead of using lseek()/read() to grub around in the file.
> All of a sudden, you can't do this anymore.
For that matter, shared libraries aren't really that different from
executables anyway. So binaries are restricted to 2 GB. That's OK;
it is a 32-bit machine after all. Let's be clear: we're talking about
datafiles, and application datafiles ONLY.
--
Robert Krawitz <[EMAIL PROTECTED]> http://www.tiac.net/users/rlk/
Tall Clubs International -- http://www.tall.org/ or 1-888-IM-TALL-2
Member of the League for Programming Freedom -- mail [EMAIL PROTECTED]
"Linux doesn't dictate how I work, I dictate how Linux works."
--Eric Crampton
------------------------------
From: joh jae chul <[EMAIL PROTECTED]>
Subject: Q: interrupt, VIA, U-Net, SMP, MPI, etc :-)
Date: Sun, 25 Jul 1999 14:20:22 -0500
Hello, everyone.
I'm currently writing a ethernet device driver for "modified" U-Net (later
absorbed/evolved into VIA) on Intel SMP cluster. I said "modified" because
I put flow control in device driver, and changed a little bit of other
things, too.
By the time I started looking MPI, i realized I needed "blocking"
send/receive so that applications do not busy-wait. First I thought simple
interruptible_sleep_on() would do the work; later I remembered U-Net was
using fast trap to send a packet. So, if I use any sleep routine inside
sending function, it would mean I permit schedule()ing inside of interrupt
routine; fast trap is called by user library and returns with IRET.
Hmm... Any suggestion would be grateful.
By the way, we have Xeon 450 SMPs running RedHat 5.1 2.0.36 SMP kernel.
Thanks.
------------------------------
From: [EMAIL PROTECTED] (Arun Sharma)
Subject: Control speaker without /dev/mixer
Reply-To: [EMAIL PROTECTED]
Date: Sun, 25 Jul 1999 20:10:49 GMT
Hi,
I have a machine which has no sound card and hence no /dev/mixer. But
I want to be able to control the volume on my PC speaker. Is there
any Linux utility which lets me do this ?
Thanks!
-Arun
------------------------------
Date: Sun, 25 Jul 1999 14:30:31 -0600
From: Kettlewell <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,alt.os.linux,apana.lists.os.linux.c-programming,comp.os.linux.development.apps,linux.dev.apps,linux.dev.c-programming,comp.graphics.algorithms,comp.graphics.misc
Subject: X Server Error when using SRGP graphics library -- Help!!
I'm using the Linux Port of SRGP for use with the book Computer
Graphics: Principals and Practice when I run my programs I get an
error that says :
SRGP FATAL ERROR:
X SERVER ERROR
(examing the core will be useful only if using X in synchronous mode)
BadMatch ( invalid parameter attributes)
Is this a known problem....or do I have a version conflict with my
hardware/software somewhere?? I'm running RH 5.2 , so I don't think
this is an outdated conflict (unless maybe on the side of SRGP) If
anyone has any info on this problem I would be very grateful.
Matt
PS. Any info on installing SRGP on Windows using MSVC 6.0 would also
be helpful.
------------------------------
From: <[EMAIL PROTECTED]>
Subject: Unresolved symbol
Date: 25 Jul 1999 20:28:53 GMT
Hi all.
I installed the Kernel v2.3.9.
I make some module.
When I run "depmod -a" the system tell me that affs.o, fat.o, smbfs.o ethertap.o
have unresolved symbols.
I try to load the fat.o module and the system don't link update_vm_cache symbol.
What's the matter?
Thank to all to answer me.
Mauro Ziliani (Parma, Italy)
------------------------------
From: Stefan Skoglund <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Why we are still holding on to X Windows
Date: Sun, 25 Jul 1999 23:27:10 +0200
Christopher Browne wrote:
>
> How, with X, do you accomplish the following task?
>
> - I need to add access to the font xyz.pfb.
> - I need to add access to this font as a normal user (e.g. - without
> resorting to root access).
> - I need to make this change without needing to restart X, thereby
> shutting down all existing X applications.
>
Have the user run its very own font-server.
Hmm, you don't really need a system-wide font-server
in that case.
And as someone else already said:
xset fp rehash works magics.
I would personally very much dislike adding yet another suid-root
program to the system.
------------------------------
** 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
******************************
