Linux-Development-Sys Digest #767, Volume #8 Sat, 2 Jun 01 16:13:12 EDT
Contents:
Re: Logging signal 11? (Joe Pfeiffer)
Re: ext2 performance with large directories (Anonymous)
wchar_t and Unicode (Michael B. Allen)
Re: Logging signal 11? ("Kyle Brant")
Re: thundering-herd vs wake-one ([EMAIL PROTECTED])
Re: Logging signal 11? (N. Yeamans)
Re: Programming Modem Voice (N. Yeamans)
Re: sigsend - which library/header (N. Yeamans)
Re: hash function for ip address (Paul Repacholi)
Re: Lilo and BIOS calls (John in SD)
Re: Programming Modem Voice (Youngert)
Re: hash function for ip address (Shankar Unni)
----------------------------------------------------------------------------
From: Joe Pfeiffer <[EMAIL PROTECTED]>
Subject: Re: Logging signal 11?
Date: 01 Jun 2001 14:53:43 -0600
Stefan Boresch <[EMAIL PROTECTED]> writes:
>
> I was under the impression that they ended up in the syslog (unless
> the Sig11 hangs your machine because it affects a kernel task);
> that is from memory on RH 5.x and 6.x machines (fortunately didn't
> have any of those guys in a long time); maybe you have to play
> with /etc/syslog.conf stuff.
If the program that gets the sig11 is one that logs to syslog, then an
entry appears. Normal user apps don't (remember, sig11 isn't supposed
to be a ``bad memory'' signal, it's supposed to indicate a bad memory
reference from a program, like a bad pointer. So there's no reason,
under normal circumstances, to log them).
> As to the motherboard you mention: I had a hard time with one of
> these, first suspecting memory only to realize that the mobo was dying
> on me. Rather than underclocking you should maybe try memtest
> (http://reality.sgi.com/cbrady_denver/memtest86/). Some of the info
> on the homepage may help you diagnose the problem; also, there are
> kernel patches to avoid spotty memory. If memtest runs without problems
> but you continue to get Sig11 errors, then you may be heading towards
> a dying mobo (or processor).
These are two identically configured, nearly new systems (I got them
cheap because the company that originally got them went belly up
almost immediately), showing identical symptoms. I've also heard from
a couple of other people who've had the same problems; in fact, it was
one of those other people who suggested underclocking. These are the
first systems I've ever had that showed classic bad memory symptoms,
but passed memtest86... it really seems like something on there --
mobo, processor, or memory -- just doesn't quite meet specs.
--
Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605
Department of Computer Science FAX -- (505) 646-1002
New Mexico State University http://www.cs.nmsu.edu/~pfeiffer
SWNMRSEF: http://www.nmsu.edu/~scifair
------------------------------
Date: Fri, 01 Jun 2001 23:08:57 -0400
Subject: Re: ext2 performance with large directories
From: Anonymous <[EMAIL PROTECTED]>
[EMAIL PROTECTED] wrote:
> It's a linear lookup, i.e. it gets very slow.
One thing I noticed is that even for as many
as 60,000 directory entries, the performance
isn't all that bad.
I wonder why?
--------== Posted Anonymously via Newsfeeds.Com ==-------
Featuring the worlds only Anonymous Usenet Server
-----------== http://www.newsfeeds.com ==----------
------------------------------
From: [EMAIL PROTECTED] (Michael B. Allen)
Subject: wchar_t and Unicode
Date: 2 Jun 2001 04:05:37 GMT
Reply-To: [EMAIL PROTECTED]
I am writing an app that must do network I/O with quite a bit of UCS-2LE
encoded strings in it. I can see wchar_t doesn't like to be converted to
or constructed from the various Unicode encodings. What type of encoding
it wchar_t anyway? I take it I must write my own sting handling functions
for this! :~(
Why doesn't wchar_t play nice with Unicode encodings and how do I get
around this problem?
Thanks,
Mike
------------------------------
From: "Kyle Brant" <[EMAIL PROTECTED]>
Subject: Re: Logging signal 11?
Date: Sat, 2 Jun 2001 00:36:26 -0500
I didn't know that a 550 MHz k6III was ever produced. Are they "k63+"
CPU's that are overclocked?
I have one system that experiences problems running the FSB at 100 MHz
with just about any CPU (and it has pc133 memory installed that is
very decent quality), but is rock solid at 95 MHz FSB, and the cpu is
a k63+450 running at 570 MHz. The mobo is an MVP3 chipset board, not
Aladdin, but similar problems no doubt can occur with these systems.
Just as a guess, did you try to disable the mobo cache and see if the
problem is affected?
--
Best regards,
Kyle
"Joe Pfeiffer" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
| A bit of background -- I have a couple of flaky machines (Gigabyte
| GA-5X motherboard, Aladdin V chipset, 550MHz K6-III processors),
which
| have thrown sig 11s from day one. Following a tip I received a
while
| ago, I tried underclocking them 5%, and they now seem to work great.
| At any rate, kernel compiles now succeed.
|
| All the same, I'd like to be able to keep an eye on these guys, so I
| can know if they start showing wonky behavior again. Since they
| normally fail by just having random apps crash, it would be very
| helpful if I could log all sig 11s, and check the log periodically.
|
| So... is there a way?
| --
| Joseph J. Pfeiffer, Jr., Ph.D. Phone -- (505) 646-1605
| Department of Computer Science FAX -- (505) 646-1002
| New Mexico State University
http://www.cs.nmsu.edu/~pfeiffer
| SWNMRSEF: http://www.nmsu.edu/~scifair
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.unix.bsd.freebsd.misc
Subject: Re: thundering-herd vs wake-one
Date: 2 Jun 2001 05:19:03 GMT
In comp.os.linux.development.system Gregory Bond <[EMAIL PROTECTED]> wrote:
| [EMAIL PROTECTED] writes:
|
|> What better solutions do you have or suggest?
|
| Look at kqueue(2) and kevent(2) in 4.x or later FreeBSD. Basically,
| only pass around the descriptor array/s once, rather than once per
| select() call.
These are good functions. But until the EVFILT_ACCEPT filter is
added, it won't actually avoid a thundering herd problem. Hopefully
that will be soon.
--
=================================================================
| Phil Howard - KA9WGN | Dallas | http://linuxhomepage.com/ |
| [EMAIL PROTECTED] | Texas, USA | http://phil.ipal.org/ |
=================================================================
------------------------------
From: [EMAIL PROTECTED] (N. Yeamans)
Subject: Re: Logging signal 11?
Date: Sat, 02 Jun 2001 06:34:04 GMT
Upon the 01 of Jun in the hour of 9 MDT, the great Joe Pfeiffer did decree:
> A bit of background -- I have a couple of flaky machines (Gigabyte
> GA-5X motherboard, Aladdin V chipset, 550MHz K6-III processors), which
> have thrown sig 11s from day one.
...
There is a Sig11 faq at http://www.bitwizard.nl/sig11/
It discusses misc. hardware problems that cause them.
It seems all apps sig11 at random? I thought usually it was only compiling
something big, like the kernel, that caused this (at least when there is
a problem with hardware). Does this happen under heavy load, or just at any
time?
I've noticed outdated program binaries cause this problem--when they've
been compiled for different kernel/library versions, but it sounds
from your posts that the computers do this even from a clean install.
So this wouldn't be the problem.
--
Norman Yeamans - [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (N. Yeamans)
Subject: Re: Programming Modem Voice
Date: Sat, 02 Jun 2001 07:06:36 GMT
Upon the 01 of Jun in the hour of 12 MDT, the great Youngert did decree:
> I am not sure if this is the right NG to post and ask question regarding
> the above subject.
I probably would have posted this to comp.os.linux.hardware or one of the
generic linux groups first to answer the question if there were existing
programs that do this. This group is about writing kernel code/modules.
The .app group is for applications. (userspace) Not sure which one is
appropriate for your programming question. Depends where you are planning
to write the code--in the kernel or as an executable.
> These days, most 56Kbps modems are V.90 compliance that support
> data/FAX/voice.
It sounds like you are talking about winmodems (most 56k modems are, and
they have voice capability because of the fact they are just soundcards
with a telephone jack instead of a speaker/mic jack)
> What I would like to do is to be able to write a program
> so that it can be used as a voice answering machine. In order to do so, I
> need to be able to program and access the voice capacity of the modem. I
> am wondering if anyone out there knows anyone info as well as existing
> applications that manipulate the modem voice capbility.
Check out linmodems.org. They list linux drivers for winmodems. I thought
I saw info about programs that use voice like you are looking for--either
way you'll probably still need some sort of driver, unless you want to
go to the trouble of writing your own.
--
Norman Yeamans - [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (N. Yeamans)
Subject: Re: sigsend - which library/header
Date: Sat, 02 Jun 2001 07:43:16 GMT
Upon the 01 of Jun in the hour of 8 MDT, the great Alexander Spanke did decree:
> i have a simple question, i programmed a software for SCO UnixWare 2.1.3 and
> now it shoul run under Caldera OpenLinux.
> So long a can evrything compile, but at two points, i get a error message,
> the two points are parts where i use the function sigsend.
> Now my question, in which Library or Header File is this function sigsend or
> some equal ?
As far as I know, there is no sigsend function for linux...does it send a
signal to a specific process? If so, you want kill(). Try 'man 2 kill'
It's in the <signal.h> header file. If you are sending the sig
to the same process, raise() is easier, it's also in <signal.h>
You may want to try the comp.os.linux.development.apps group. They are
for writing userspace programs. This one is for kernel code.
--
Norman Yeamans - [EMAIL PROTECTED]
------------------------------
Subject: Re: hash function for ip address
From: Paul Repacholi <[EMAIL PROTECTED]>
Date: 02 Jun 2001 18:22:34 +0800
[EMAIL PROTECTED] (Joern Engel) writes:
> > See Knuths ACP for the analysis.
> Do you have a link to that. I could not find anything too useful.
ACP ~ Art of Computer Programming. It is a set of BOOKS... You known
dead tree stuff? 3 volumes, or perhaps v4 has emereged from myth at last.
Look for 'Perfect Hash Function'.
--
Paul Repacholi 1 Crescent Rd.,
+61 (08) 9257-1001 Kalamunda.
West Australia 6076
Raw, Cooked or Well-done, it's all half baked.
Spam-To: [EMAIL PROTECTED],[EMAIL PROTECTED],[EMAIL PROTECTED],[EMAIL PROTECTED],
[EMAIL PROTECTED],[EMAIL PROTECTED]
------------------------------
From: John in SD <[EMAIL PROTECTED]>
Subject: Re: Lilo and BIOS calls
Reply-To: [EMAIL PROTECTED]
Date: Sat, 02 Jun 2001 17:43:07 GMT
On Thu, 31 May 2001 05:08:02 +0200, "Hugin" <[EMAIL PROTECTED]> wrote:
>Hello,
>
>My boot sequence in BIOS is set to: C,SCSI,A
>C is assigned 0x80, SCSI is assigned 0x81 and A is 0x00,
>I believe. Questions:
>
>1. Lilo is installed on C. How do I configure
>Lilo to boot device 0x81?
The LILO boot loader can only get control from the BIOS. So where the BIOS
boots from is critical.
Normally, in booting from 0x80, the BIOS reads the MBR (sector 0) from the
disk, and transfers control there if it has a valid boot signature (0xAA55).
The usual MBR consists of code to scan for an active partition, load the first
sector of that partition, and transfer control there.
This sequence may be altered by installing the LILO boot loader to the MBR;
viz., "boot=/dev/hda". Then the lilo menu (available in LILO 21.5 & later)
can be used to select an os or kernel on any partition. But, remember, most
M$ os's will only boot from drive C:, 0x80. Linux and OS/2 will boot from any
partition.
DOS and Windows 95/98 can be booted from drive 0x81 with LILO if you use the
"map-drive=" command to swap bios codes 0x80 <-> 0x81.
>
>2. Is there a way to force lilo to install, even though
>the lilo.conf has errors?
NO!!! If there are errors, it won't boot anyhow, so,
CORRECT THE ERRORS.
--John
LILO version 21.7 (24-Feb-2001) source at
http://www.ibiblio.org/pub/Linux/system/boot/lilo
patches to -2 at ftp://brun.dyndns.org/pub/linux/lilo
------------------------------
From: Youngert <[EMAIL PROTECTED]>
Subject: Re: Programming Modem Voice
Date: Sat, 02 Jun 2001 18:01:06 GMT
Thank you very much for all the info and suggestions.
N. Yeamans wrote:
> Upon the 01 of Jun in the hour of 12 MDT, the great Youngert did decree:
>
>> I am not sure if this is the right NG to post and ask question regarding
>> the above subject.
>
> I probably would have posted this to comp.os.linux.hardware or one of the
> generic linux groups first to answer the question if there were existing
> programs that do this. This group is about writing kernel code/modules.
> The .app group is for applications. (userspace) Not sure which one is
> appropriate for your programming question. Depends where you are planning
> to write the code--in the kernel or as an executable.
>
>> These days, most 56Kbps modems are V.90 compliance that support
>> data/FAX/voice.
>
> It sounds like you are talking about winmodems (most 56k modems are, and
> they have voice capability because of the fact they are just soundcards
> with a telephone jack instead of a speaker/mic jack)
>
>> What I would like to do is to be able to write a program
>> so that it can be used as a voice answering machine. In order to do so,
>> I
>> need to be able to program and access the voice capacity of the modem. I
>> am wondering if anyone out there knows anyone info as well as existing
>> applications that manipulate the modem voice capbility.
>
> Check out linmodems.org. They list linux drivers for winmodems. I
> thought I saw info about programs that use voice like you are looking
> for--either way you'll probably still need some sort of driver, unless you
> want to go to the trouble of writing your own.
>
------------------------------
From: Shankar Unni <[EMAIL PROTECTED]>
Subject: Re: hash function for ip address
Date: Sat, 02 Jun 2001 11:52:42 -0700
Paul Repacholi wrote:
> Look for 'Perfect Hash Function'.
I don't think so..
Perfect Hash Functions are meant for the situation where your input data
domain is fixed and known in advance (e.g. a hash table of keywords for
a C language scanner). With variable input data (in this case - a table
of source IP addresses of client connections), there can be no perfect
hash function (except maybe an identity function :-).
For a quick and effective hash function, I would suggest a simple XOR of
bits. IP addresses are sufficiently dense these days (if you are looking
at the general internet domain) that the hash function would be
reasonably uniform.
(Of course, YMMV if you deploy your app in a domain in which the source
IP addresses follow some highly restrictive pattern, e.g. some company
intranets with just a few sparsely populated subnets under 10.*.*.*, but
even there, I think you'll find enough of a distribution using an XOR
hash function to be effective enough..)
--
Shankar.
------------------------------
** 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
******************************
