Linux-Development-Sys Digest #664, Volume #6 Fri, 30 Apr 99 07:14:09 EDT
Contents:
Re: Kernel 2.2.6 problems (Allin Cottrell)
dtc_3181e and g_NCR5380 (Stuart Midgley)
Re: Threads >> PThreads or LinuxThreads?? ("G. Sumner Hayes")
Re: Accessing a process' last error count (Greg Ercolano)
Re: Accessing a process' last error count (Tony Young)
Re: Calibri Firewall/Router on Ebay. ("Gideon le Grange")
Re: Threads >> PThreads or LinuxThreads?? ("G. Sumner Hayes")
Re: login command source code ("Andre Malafaya Baptista")
Re: Is Linux Y2K compliant? (Christopher B. Browne)
Re: Threads >> PThreads or LinuxThreads?? ("Joshua E. Rodd")
Help Installing Linux 5.2 on Dell PowerEdge SP 5166 (Leslie Smith)
Re: Async. start User Process from kernel? (execve?) (Martin Recktenwald)
Re: Calibri Firewall/Router on Ebay. (John Navas)
Re: physical Memory (Robert Kaiser)
----------------------------------------------------------------------------
From: Allin Cottrell <[EMAIL PROTECTED]>
Subject: Re: Kernel 2.2.6 problems
Date: Thu, 29 Apr 1999 22:55:34 -0400
"Ruslan O. Nesterov" wrote:
>
> I have just downloaded kernel 2.2.6 src
> and starting compilining it.
> make dep is ok
> make modules - getting errors
> make modules_install - getting errors
> Make install - is ok
> rebooting and still 2.0.36 kernel on me redhat.
> What to do???????????
Interesting. You expect to be able to recompile and replace the
kernel of your operating system without bothering to read any
documentation on how to do this. Pretty good for instant
gratification. How about looking at some HOWTos?
--
Allin Cottrell
Department of Economics
Wake Forest University, NC
------------------------------
From: Stuart Midgley <[EMAIL PROTECTED]>
Subject: dtc_3181e and g_NCR5380
Date: 30 Apr 1999 05:37:27 GMT
I have a DTC 3181e scsi controller. I have managed to get it partially
working with kernel 2.2.6, I just want to get the last bit.
Anyway, I did an pnpdump -c > /etc/isapnp.conf
The card detects and gets allocated
irq=4
addr=0x240
No problem. I then
modprobe g_NCR5380 ncr_irq=4 ncr_addr=0x240 dtc_3181e=1
The 2 SCSI hards disks that I have attached get identified correctly (CDC,
312MB) and assigned to sda and sdb. Howver, the system the halts for a
while and a heap of error messages get shown (they scroll off the screen
before I can get them down). The last of these messages is
SCSI disk error: host 0 channel 0 id1 lun 0 return_code 50000
scsidisk I/O error: dev 08:00, sector 0
scsi0: warning: target bitmask 02 lun 0 not in disconnect_queue
When I search the g_NCR5380.c file, I find that this error seems to relate
to unsupported features of SCSI II. Since I am not a drivers guru, I do
not know what is going on.
Can any body help me to get the device working correctly (it works well
under DOS, I can partition and format the drive fine).
Thanks
Stu.
[EMAIL PROTECTED]
================== Posted via SearchLinux ==================
http://www.searchlinux.com
------------------------------
From: "G. Sumner Hayes" <[EMAIL PROTECTED]>
Crossposted-To: gnu.misc.discuss
Subject: Re: Threads >> PThreads or LinuxThreads??
Date: Fri, 30 Apr 1999 02:42:32 -0400
Harald Schreiber wrote:
>
> In article <[EMAIL PROTECTED]>,
> Martin Recktenwald <[EMAIL PROTECTED]> writes:
> [...]
> > prep.ai.mit.edu:/gnu/glibc/glibc-2.1-README:
> >
> > "glibc-2.1 has been (temporarily) removed, until some
> > political issues are worked out."
>
> glibc-2.1 has been removed from prep.ai.mit.edu by the GNU people
> because the documentation of glibc-2.1 recommends compiling with
> egcs and not with gcc.
I believe that it was removed because there is some concern about the
copyright on the db implementation (whether it has to do with
assignments or something more serious such as GPL-compatibility I'm
not sure). The GNU project has already announced that egcs will be
the official gcc from now on (starting with version 3.0), so complaining
about egcs seems a rather unlikely reason to withold libc.
I've CC'd gnu.misc.discuss in the hopes that the people there can
correct my statements and shed more light on the matter.
> glibc-2.1 has been released several months
> ago and the Linux distributions RedHat 6.0 and Caldera 2.2 are
> alreday based on glibc-2.1
Does anyone know if the copyright issues have been resolved and what
exactly they were to begin with? I've heard that glibc-2.1 uses some
variant of the sleepycat db, but am unsure...
--Sumner
------------------------------
From: Greg Ercolano <[EMAIL PROTECTED]>
Crossposted-To: linux.dev.kernel
Subject: Re: Accessing a process' last error count
Date: Thu, 29 Apr 1999 23:43:08 -0700
Tony Young wrote:
> Anyone know how (if?)I can programmatically access the last error code
> produced by a given process? (ie not my own).
I believe unix process accounting is what you're looking
for..
You can access not only the exit code, but uid/pid/etc
from every process that exits on the machine.
Just parse the process accounting file as it grows.
Unix process accounting is pretty simple; you just tell the
kernel to start logging the info for every process that exits
to a file. Normally this is a sysadmin task, but you can do
it from C yourself, if this is a specific application, eg:
/* Enable process accounting */
if ( acct("/usr/tmp/foo") )
{ perror("/usr/tmp/foo"); }
..then just keep an eye on that file.
See include/linux/acct.h for the file format.
Since this file grows pretty quick, you should have it rotated
out often, either by yourself, or by cron(1).
BTW, acct(NULL) will turn it off. See acct(2) and, if you
have the unix accounting tools installed, see 'man -k acct'.
--
/\_/\
|o,o|
\/ )
----mm---------------------------------------------------------------------
Greg Ercolano UNIX NightOwl / Systems Programmer & Admin
[EMAIL PROTECTED] http://3dsite.com/people/erco/
===========================================================================
------------------------------
From: Tony Young <[EMAIL PROTECTED]>
Crossposted-To: linux.dev.kernel
Subject: Re: Accessing a process' last error count
Date: Fri, 30 Apr 1999 17:02:11 +1000
Greg Ercolano wrote:
Thanks for the response, but that's not quite what I'm after.
acct() will tell me the exit code of a process when it terminates. I'm
actually after a way to access the value of errno for a currently
running process. ie the last error returned from a system call.
Thanks though Greg.
>
> Tony Young wrote:
> > Anyone know how (if?)I can programmatically access the last error code
> > produced by a given process? (ie not my own).
>
> I believe unix process accounting is what you're looking
> for..
>
> You can access not only the exit code, but uid/pid/etc
> from every process that exits on the machine.
>
> Just parse the process accounting file as it grows.
>
> Unix process accounting is pretty simple; you just tell the
> kernel to start logging the info for every process that exits
> to a file. Normally this is a sysadmin task, but you can do
> it from C yourself, if this is a specific application, eg:
>
> /* Enable process accounting */
> if ( acct("/usr/tmp/foo") )
> { perror("/usr/tmp/foo"); }
>
> ..then just keep an eye on that file.
> See include/linux/acct.h for the file format.
>
> Since this file grows pretty quick, you should have it rotated
> out often, either by yourself, or by cron(1).
>
> BTW, acct(NULL) will turn it off. See acct(2) and, if you
> have the unix accounting tools installed, see 'man -k acct'.
>
> --
>
> /\_/\
> |o,o|
> \/ )
> ----mm---------------------------------------------------------------------
> Greg Ercolano UNIX NightOwl / Systems Programmer & Admin
> [EMAIL PROTECTED] http://3dsite.com/people/erco/
> ---------------------------------------------------------------------------
--
--
Tony Young Mob: 0414 64 99 42
Software Engineer Tel: +612 9966-1066 Fax: +612 9966-1042
Integrated Research Level 10, 168 Walker St, North Sydney, NSW, 2060,
Aus.
[EMAIL PROTECTED] http://www.ir.com.au
------------------------------
From: "Gideon le Grange" <[EMAIL PROTECTED]>
Crossposted-To:
comp.dcom.modems.cable,comp.dcom.xdsl,comp.os.linux.hardware,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.portable,comp.security.firewalls
Subject: Re: Calibri Firewall/Router on Ebay.
Date: Fri, 30 Apr 1999 09:21:12 +0200
Idiot.
Don't use this newsgroup to draw attention to your own sutff
being sold by Ebay.
What a moron
Jack Levin wrote in message ...
>Hi all... Check out Calibri-133 firewall being sold at Ebay. (hardware)
>http://cgi.ebay.com/aw-cgi/eBayISAPI.dll?ViewItem&item=97714586
>
>-Jack
>
>
------------------------------
From: "G. Sumner Hayes" <[EMAIL PROTECTED]>
Crossposted-To: gnu.misc.discuss
Subject: Re: Threads >> PThreads or LinuxThreads??
Date: Fri, 30 Apr 1999 04:31:45 -0400
Geoffrey KEATING wrote:
>
> "G. Sumner Hayes" <[EMAIL PROTECTED]> writes:
> >
> > I believe that it was removed because there is some concern about
> > the copyright on the db implementation...
>
> No, it was really because of the gcc thing. Note that gcc and egcs
> are still different projects (for the moment), although now run by the
> same people.
Notes on www.gnu.org:
: We are pleased to announce that the EGCS Steering Committee is taking
: over as GNU maintainer of the GNU C Compiler . The Committee will
: change its name to the GCC Steering Committee and will, as a GNU
: maintainer, have responsibility for deciding what changes to install,
: fixing bugs, and making releases.
and egcs.cygnus.com indicate that your "for the moment" clause is an
important one (I'm sure you know that, just clarifying for the folks
over on comp.os.linux.dev.sys). They also seem to indicate that there
isn't a deep-seated GNU/egcs feud ongoing; is glibc going to reappear
at ftp.gnu.org at some point in the future? Or is there another
canonical location for it?
Apologies to colds for the misinformation.
--Sumner
------------------------------
From: "Andre Malafaya Baptista" <[EMAIL PROTECTED]>
Subject: Re: login command source code
Date: Thu, 29 Apr 1999 16:27:40 +0100
Correction:
ftp://ftp.win.tue.nl/pub/linux/utils
Andr�
Jerry Peters <[EMAIL PROTECTED]> wrote in message
news:7g81c4$c76$[EMAIL PROTECTED]...
> Sean Godsell <[EMAIL PROTECTED]> wrote:
> > Hello my name is Sean Godsell,
> > I was wondering if anyone can help me regarding source code for the
> > command login in the bin directory. Does anyone know where to get it.
> > The reason I need it is so I can add one more prompt for a secure id
> > card.
> > Thank you.
> > [EMAIL PROTECTED]
> > or
> > [EMAIL PROTECTED]
>
> It's part of: util-linux-2.9 ftp.win.tue.nl:/pub/linux/util
> Jerry
>
------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Subject: Re: Is Linux Y2K compliant?
Reply-To: [EMAIL PROTECTED]
Date: Fri, 30 Apr 1999 05:14:58 GMT
On 29 Apr 1999 10:59:41 PST, Don Baccus <[EMAIL PROTECTED]> posted:
>In article <QOYV2.17251$[EMAIL PROTECTED]>,
>Christopher Browne <[EMAIL PROTECTED]> wrote:
>
>>Q: How do *you* define Y2K compliant?
>
>>[*I* define it as "Was the program design done competently?"]
>
>Probably not the place to get into a religous argument over
>Y2K, but many of the large commercial (mostly COBOL) programs
>which still carry legacy code from the 60s in them were
>certainly designed competently at the time, given the
>design constraints and assumptions their developers worked
>under.
That is indeed true.
These systems were designed without the benefit of some of the
hindsight that we now have, and were constructed with the assumption
that *surely* they would get replaced within the next thirty years or
so.
>Saving a couple of BCD digits per date seems silly in
>retrospect, but it wasn't so silly for companies with
>huge numbers of dated records in an era where a 50 MB
>"washtub" drive cost $40K or so (closer to $125K in
>today's money). Working with thousand dollar megabytes
>changes the economic constraints compared to working with
>thirty dollar gigabytes as we do today.
I would argue that the *true* constraint that mandated compression was
the 72 or 80 columns on the typical punched card. Many, many, many
mainframe-based applications are designed around processing streams of
data that are, in essence, virtual punched cards.
In 80 columns, you don't have space for long names, and hence my name
is commonly misprinted as "Christophe Browne."
And if you have a couple of dates on a record, putting in century
information (that could doubtless be inferred from context) will
easily consume 5% of the limited amount of space on the record.
The cost of disk space was but one of the constraints that early
business-oriented programmers laboured under.
I would suggest that the aspect of this that implies "incompetence"
lies not in the '60s and '70s, when most of such systems were
initially built, but rather in the '80s, when "Corporate America"
probably should have rebuilt the systems. 'Twould have been cheaper
then than it is to remediate now...
>Now, when it comes to programs or microcontrollers
>written in the last decade or so, you have a point.
I fairly like the comments in another article in the thread in this
regard; the question is not whether the programs are Y2K-compatible,
but whether the *programmers* are Y2K-compatible.
On 29 Apr 1999 22:50:06 GMT, Neil Schemenauer <[EMAIL PROTECTED]>
posted:
>The root of the problem is people. The space argument doesn't really
>wash. Check out
> http://language.perl.com/news/y2k.html
>for a good summary of how we got in this mess and how people are
>now trying to cover their asses.
Tom Christiansen has a real gift for putting things more bluntly than
most people can handle. That web page seems to be a whole lot more
diplomatic than I'm used to seeing from him. Maybe the "Edited by Tom
Christiansen" indicates that he's blessing what someone else wrote...
--
"Huh? Windows was designed to keep the idiots away from Unix so we
could hack in peace. Let's not break that." -- Tom Christiansen
<http://www.hex.net/~cbbrowne/linuxy2k.html>
[EMAIL PROTECTED] - "What have you contributed to free software today?..."
------------------------------
From: "Joshua E. Rodd" <[EMAIL PROTECTED]>
Crossposted-To: gnu.misc.discuss
Subject: Re: Threads >> PThreads or LinuxThreads??
Date: Fri, 30 Apr 1999 04:49:31 -0400
"G. Sumner Hayes" wrote:
> Geoffrey KEATING wrote:
> > "G. Sumner Hayes" <[EMAIL PROTECTED]> writes:
> > > I believe that it was removed because there is some concern about
> > > the copyright on the db implementation...
Sleepycat DB had nothing to do with it. Sleepycat DB is available
under either BSD or {L|}GPL licence.
It's been included for a long time in glibc.
> > No, it was really because of the gcc thing. Note that gcc and egcs
> > are still different projects (for the moment), although now run by the
> > same people.
Exactly so--so the reason for noninclusion of glibc-2.1 should vanish.
------------------------------
From: [EMAIL PROTECTED] (Leslie Smith)
Subject: Help Installing Linux 5.2 on Dell PowerEdge SP 5166
Date: 30 Apr 1999 09:30:56 +0100
Can help me out here, I have got a Dell PowerEdge server
SP 5166 Dual CPU. I am trying to load linux 5.2 kernel
with no joy. I can build the box with 5.0 but not with
5.1 or 5.2 boot disk. Dose anyone know of a fix for
this.
Regards
Leslie...UK:-)
------------------------------
From: Martin Recktenwald <[EMAIL PROTECTED]>
Subject: Re: Async. start User Process from kernel? (execve?)
Date: 30 Apr 1999 12:12:36 +0200
[EMAIL PROTECTED] writes:
> As a module, loaded well after boot time, is
> there a way to launch an arbitrary user process?
How about writing a daemon that gets commands from your module?
Martin.
--
Sign the EU petition against spam:
http://www.politik-digital.de/spam/
------------------------------
From: [EMAIL PROTECTED] (John Navas)
Crossposted-To:
comp.dcom.modems.cable,comp.dcom.xdsl,comp.os.linux.hardware,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.portable,comp.security.firewalls
Subject: Re: Calibri Firewall/Router on Ebay.
Date: Fri, 30 Apr 1999 07:04:54 GMT
Reply-To: [EMAIL PROTECTED]
[POSTED TO comp.dcom.modems.cable]
"Jack Levin" <[EMAIL PROTECTED]> wrote:
>[SNIP]
Please do not post eBay or other auction ads here. This is not even a
forsale newsgroup, much less a newsgroup for auction ads. (Ads of any kind
generally belong only in newsgroups that specifically allow them; e.g.,
misc.forsale.computers.* -- for more information, see
<ftp://rtfm.mit.edu/pub/usenet/misc.forsale.computers.discussion/Misc.FS+Biz.Mktplc_ADVERTISING_FAQ--INFO_FOR_NEW_USERS>.)
Thank you for your consideration.
--
Best regards,
John mailto:[EMAIL PROTECTED] http://www.aimnet.com/~jnavas/
28800-56K Modem FAQ: http://www.aimnet.com/~jnavas/modem/faq.html
------------------------------
From: [EMAIL PROTECTED] (Robert Kaiser)
Subject: Re: physical Memory
Date: 30 Apr 1999 10:05:04 GMT
Hi Marc :-)
In article <[EMAIL PROTECTED]>,
Marc Lefranc <[EMAIL PROTECTED]> writes:
> [EMAIL PROTECTED] (Robert Kaiser) writes:
>>
>> > By the time your interrupt handler is called, the process is long long
>> > gone and even if it was the current process it might have changed its
>> > mapping.
>>
>> Of course, but all of this doesn't matter if the target memory has been
>> locked and it's physical address is known.
>
> Of course, one has also to be sure that the calling process is still
> alive before initiating the next DMA burst.
Yes, this is true (I took that for granted).
> I may however forget something. Am I wrong in assuming the above ?
> (What about SMP systems ?)
All I can say about that is that a colleague of mine has developed a
new version of my UDMA patch for the 2.2.x kernel series and tested
it on a Dual Pentium machine. He did have problems with our framegrabber
driver (interrupts being serviced while the driver thought it had them
disabled with cli()), but the user-space DMA stuff worked just the same.
As usual, I have released the new version of the patch on our FTP
server (see ftp://ftp.sysgo.de/pub/Linux).
BTW, this patch still seems to be a secret to many kernel programmers.
Can anybody give me some hints what I could do to give it some more
publicity ? (Short of re-engaging into a discussion with Linus -- I
have given up on that :-( )
Cheers
Rob
================================================================
Robert Kaiser email: rkaiser AT sysgo DOT de
SYSGO RTS GmbH
Mainz / Germany
------------------------------
** 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
******************************
