I'm doing some development on an instrument where communication
between the linux box and the some remote electronics is via this
specialized PCI serial fiber card. The PCI card really sucks, it's
not exactly a custom board, not exactly a production quality board.
Some PCs won't boot with with
On 10/30/07, Charles Lockhart [EMAIL PROTECTED] wrote:
I'm doing some development on an instrument where communication
between the linux box and the some remote electronics... yadda yadda
Don't know if it matters, the distro I'm using is Centos 4.4.
-Charles
Sorry, I missed something. In what context is China an emerging
market/country? Because in most things China seems to be a well established
power, so I'm not understanding your context.
-Charles
Julian Yap wrote:
OK, let's take China for example.
96822
808-956-5797 voice, 877-284-1934 fax
- Original Message -
From: Charles Lockhart [EMAIL PROTECTED]
To: LUAU luau@lists.hosef.org
Sent: Monday, November 21, 2005 11:24 AM
Subject: [LUAU] SMP kernel wierdness
My workplace just got a bunch of dual athlon 64 x2 machines that we're
loading
This is probably mostly off-topic, and I apologize for that. If this is not
apreciated, feel free to blast me publicly or privately, and I'll keep it
strictly to Linux stuff from now on, no hard feelings at all.
I live in Hilo, and I work for the NASA IRTF, the second oldest telescope up
on
Are you looking for something that would boot from bios, or boot from an
installed bootloader?
What USB support is there for such a thing in your standard bios? How about
in grub? When I tried doing installing Centos to a usb drive and then using
it as a portable os, I stopped when I
I've had more practice in analytical and exploratory programming in the
last two years than I ever wanted. And then there's always a certain
amount of fear that you're going to program something wrong and burn it
up.
You get over this when you finally do.
Well, I used to blow stuff up
Sorry, tried to send that to Jim directly, but goofed.
-Charles
Dustin Cross wrote:
$25/hr is pretty sad for a knowledgable Cisco field eng. Especially for
only two days work. How can anyone afford to live in Hawaii who consults
for $25/hr. Most consultants work less than 2000 hours per year and that
is why they charge more, and their knowledge.
If you
[EMAIL PROTECTED] wrote:
Because we speend too much time with stuff like this:
http://kegbot.org/project/
I think that's the best application of Linux to a real world task that
I've ever seen.
cheers,
-Charles
The question is probably goofy, not the ram.
I just installed 4GB of RAM on a Dell Precision 450 running CentOS4
(dual Xeon P4 workstation). The BIOS reports seeing all of it, but when
I run cat /proc/meminfo I see number that don't see quite right for
the MemTotal. Is the OS really not
I don't know many chics who'd be willing to suck face with the human
incarnation of Bill the Cat.
And here I thought it'd be about the tee vee show, darn.
-Charles
Tim Newsham wrote:
http://spinster.org/photos/als/20.html
more so than most.
Tim Newsham
http://www.lava.net/~newsham/
Is there any way set up to search the luau archives?
Thanks,
-Charles
Jim, Angela, Vince, Wayne, thanks!
Yeah, I knew about the embedded linux - no swap partition thing. I use
several that don't have (can't have) swap space, everything is running
out of RAM. I just wasn't sure if the desktop distros possibly tied
stuff up differently somehow.
So I did as
Brian Chee wrote:
Actually I have a question...why would you want to run a machine without
swap? There are good reasons if you're running an embedded linux machine,
but for normal machines I've seen folks setup unix boxes that boot from the
network but ONLY do swap to the local hard disk. Those
Any ideas on how well a machine running a popular distribution of Linux
(RH, FCx, Mandrake, Debian, etc) would do if the machine had no swap
partition? Anybody actually do it? If so, how well did work?
Thanks,
-Charles
Sorry, I meant to say with no swap space at all, not just without a swap
partition.
Jim Thompson wrote:
pthread_cond_broadcast() will (attempt) to waken *ALL* threads waiting
on that condition variable. If there is more than one, they'll
race, and the first one that gets past the mutex 'wins'.
pthread_cond_signal() will attempt to waken *one* thread (typically
the first
Is/was nptl supported in all rh9 kernels (2.4.x)? I kind of remember
seeing nptl as part of the kernel name for some machines, but can't
remember which ones.
Also, I ran across this tidbit on usenet, was wondering if anyone had
any practical experience with the problem:
***
LinuxThreads is
Angela Kahealani wrote:
RedHat (or, as appropriate, substitute Fedora) did a backport of NPTL
so that it was contained in the default distro of Fedora Core 1, but
may also have done for RH9, all of which is Legacy Fedora now. That that
happened doesn't mean it did in other distros based on
We've got a web server running FC2 and Apache 2. I was asked to set up
a private wiki area on that server, one that provided both
username/password security for the wiki, and the same level of access
for the directory.
So, we have a script or something that every time you create a directory
Fyi, I have seen this behavior with FC1 boxes from screensaver
activity. If my memory serves, the screensaver mode that got loaded was
one that went between different ones on the machine, some of them would
be sucking 99% of the CPU cycles, and make the machine unresponsive.
So, maybe try
Hiya,
quick question to anyone who has the interest to answer:
We ordered a NAS, raid 5 system based on 3ware dual opteron (1.8Ghz)
boxes running RH7.3 and CyberNAS software. We ordered 20 SATA 250GB
drives with 16MB caches for it. What we got was 20 SATA 250GB drives
with 8MB caches. And
The retard factor is kicking in on me today.
I have a program that uses quite a bit of buffer space. There are four
major chunks of buffer space, three declared something like short
mybuffer[32][4M], and the fourth int myotherbuffer[4][4M]. Total
buffer usage comes in at somewhere under
Charles Lockhart wrote:
I kind of remember from somewhere that if you try to load a program
that requires greater than half the available heap then this'll
happen, but I've got 3GB of ram on this machine, and without this
program running, I'm only using about 700MB. So that *shouldn't
Tim Newsham wrote:
by 4M do you mean 4*1024*1024?
Yeppers.
If you declare local variables (non-static) they are allocated
on the stack. If you declare global variables they are either
in the BSS or data segment (depending on if they are initialized
to zeros or other values). There are
I have a laptop running FC2. It has both ethernet and wireless. I have
about a dozen different locations that I use my laptop, each one with
different network settings. I have been just using the
system-config-network interface to change the settings each time I
change locations (open up
Hello,
I'm setting up a wiki on a linux box for the group I work in. The
intent is that it'll help facilitate project management, documenation,
scheduling, that kind of thing. But I'm having trouble figuring out
which wiki software to use. I've run tiki a bit on my own machine,
and it's
I'm running fvwm 2.5.8-2 on Redhat 9 (hw is a dell precision 530).
Occasionally (like weekly), my desktop tweaks and suddenly I have some
wierd mixed gnome/fvwm environment going on. I haven't been able to
nail down what's triggering it, suddenly I left click and the menu I
expect doesn't
I use mozilla as my browser/mail client on a redhat 9 machine that
resides in my office in Hilo. When I'm over here in Honolulu, I tunnel
the client over ssh to read my mail on my laptop. When I start a
mozilla browser on my laptop, with the mail client on, the process seems
to get spawned
I've set up apache to host two different sites. I'd like to set up a
wiki using the tikiwiki software, and I'd like to set it up so that each
site has it's own wiki stuffs. I have both the rpm and the source.
I haven't yet found a way to have one install of the tikiwiki stuff that
allows
I guess I'm not a real big experimenter, I've mostly stuck to RedHat over
these last few years. But it seems like RH is mostly oriented towards
workstations, desktops, servers. And most of what I do is
science/instrumentation oriented.
I see a lot of scientific software and instrumentation
I'm not entirely sure, but I think both of the above have kernel
preemption patched and enabled.
Yeah, ok, I now understand SRPMS much better, and after going through the
2.4.20-8 and the 2.4.22-1.2115.nptl SRPMS, I do see a lot of stuff in
there that seems to indicate that RML's preemptive
my obtusity overwhelms. I should have thought of this, thanks Jimen.
-Charles
Jimen Ching wrote:
Most distributions have patches over the vanilla kernel. They usually
distribute the patch with their kernel. Perhaps you can look through the
patch set and spot something that might expose the
I'm trying FC1 out on a couple of computers, but I keep coming up with
little errors, though I honestly don't know if they're FC1 issues or
independent software issues, but I'm trying to find out. Some are java
related. Some wierd problems with threads. But, I was running jre
1.3x, and from
I tried this on RH9, got the same problem. So, either it's a java bug,
or something I've done wrong. Of course, it couldn't be me, so it must
be a java bug.
I'll see if I can write an example.
-Charles
Charles Lockhart wrote:
Now I've bumped up to jre 1.42, and the
thread problems have
I recently upgraded from jre 1.3.x to jre 1.4.2 (prompted by an upgrade
to FC1). Found a problem with my code and the upgraded jre, and was
hoping maybe somebody else could confirm that I'm crazy and totally
screwing up.
According to the docs, setting the selected item in a Choice widget
Can you (or anybody) recommend a place to find more information about
these kernel source rpms? I keep searching around, but not finding the
stuff I need.
It's honestly confusing. I've successfully patched the vanilla kernel
plenty of times using other patches, and I think I pretty much
Can you (or anybody) recommend a place to find more information about
these kernel source rpms? I keep searching around, but not finding the
stuff I need.
It's honestly confusing. I've successfully patched the vanilla kernel
plenty of times using other patches, and I think I pretty much
Sorry, I asked the question poorly, but I think you answered it. I have
driver code for some devices which aren't and probably wouldn't be
supported by the community. What I was trying to figure out is if I can
use that driver code with the 2.6 kernel or does it need to be changed
for the
Decided to try out these fancy new features in Fedora to see if I can get
improved performance out of a system I have, and came across a small
problem.
I have three threads that wait on the same condition, using
pthread_cond_wait(system_cond, this_threads_mutex);
I then use
I'm mostly instrumentation, but I did ask, and the clusters that are
here are (I think) all Sun based.
There is a book called How to build a Beowulf: a guide to the
implementation and application of pc clusters which is all about PC
clusters built on Linux machines. I've been kind of
Then what do you think the weight of a full grown moose is? I've never
gotten gottem more than 300 or 350 pounds of meat from a moose, and
that's when I was really scraping the bones, taking everything between
the ribs, etc. That's like maybe 25% bodywieght, max.
If you think you're going
I'm sorry, that was a private email to somebody else.
-Charles
Charles Lockhart wrote:
Then what do you think the weight of a full grown moose is? I've never
gotten gottem more than 300 or 350 pounds of meat from a moose, and
that's when I was really scraping the bones, taking everything
I read the Forbes article referenced at slashdot about the FSF going
after Cisco/Linksys/Broadcom (or something like it), and got a little
confused, had a couple of questions that I couldn't quite figure out
from reading other stuff.
q1. It makes sense to me that software companies that want
Sorry, a guy at work told me about this, but I didn't believe him, and
now I do but don't understand it. Searching google and google groups
didn't turn up what I was looking for, was hoping maybe somebody here
would know.
For the below code, running it in gdb, set a breakpoint for line 5.
[EMAIL PROTECTED] wrote:
man hdparm
I don't know much, and it's probably my own ineptitude, but I've never
been able to get hdparm to work with scsi drives. And for ide drives,
I've only been able to get hdparm to give me read metrics (hdparm -t).
Is there a tool for measuring sustained
Warren Togami wrote:
apt and yum are generic tools which can download, install and update
packages from an arbitrary source. fedora.us was one of many sources of
3rd party packages for Red Hat Linux. freshrpms.net is another.
Ah, I see, that makes sense. I need to get out more.
That is
with Google. Please do not ask me to point the way.
Maybe I'm just reading it wrong, but I really just didn't apreciate the
attitude.
Uh, starting to rant, need to stop now.
Thanks,
-Charles
Vince Hoang wrote:
On Tue, Sep 23, 2003 at 07:55:10AM -1000, Charles Lockhart wrote:
In response
I've been trying to understand fedora, but am still confused about a few
things.
The Fedora Project generates package management (install, update) tools,
such as apt and yum?
How are apt and yum different? yum seems to primarily for checking on
updates for the currently installed packages,
RH re-wrote many (all?) of their proprietary config tools. The rpm
packages have names of the form redhat-config-*. Try this at the shell
prompt:
rpm -qa | grep ^redhat-config-
There will probably be qutie a few, unless somehow you didn't install
them.
Thanks, I'll look at that.
MPEG and MP3s are protected by patents, meaning anybody that distributes
software using these patented methods are supposed to be paying royalties.
Red Hat's lawyers, as an american corporation, decided that it would be
far too risky to continue distributing that software for that reason.
(read
I've got a really bare bones system with an ethernet card that I had to
build the drivers for. I'm building the drivers as a module. After
booting up, I use insmod thedriver.o, and it seems to install ok, but
then I still can't ping out or in. I can't think of what I'm doing
wrong, there's
Just wondering what software people use for generating source code
documentation, and how you'd rate it? If you don't use any, or you
figure the source code is it's own documentation, no need to reply.
Thanks,
-Charles
fyi,
Here's the executive summary of me getting tftpd running properly (sort of).
Starting the tftpd service and turning off your iptables definitely
allows tftpd to work. I ran into a bunch of confusion because the box I
was using as a test client to see if it was working was also a Linux
Maybe an odd topic question, but was wondering if anyone has or knows of
statistics for the success of books that have free online versions
available? For example, the o'reilly Linux Device Drivers 2nd Ed. is
made freely available (http://www.xml.com/ldd/chapter/book/index.html).
I'm trying
Well, I tried disabling iptables (well, I tried /sbin/service iptables
stop, then tried just using lokkit with a no firewall setting), to no avail.
I added the -v and checked the log file.
tftp from localhost produces a line like:
Aug 24 20:00:09 stupiduser in.tftpd[2852]: RRQ from 127.0.0.1
I've been out on vacation for a couple of weeks, and I just got back to
work today, and I go to set up tftpd on Sony vaio laptop running redhat
9, and I am totally boggling.
I run the little Server Settings-Services wizard, which launches the
Service Configuration tool, and click on the tftpd
Or maybe it just seems odd to me.
I've got an ppc device that starts erroring when it tries to mount the
proc fs. It gives me:
Mounting proc filesystem: mount: only root can do that
[FAILED]
I can see this is called in the rc.sysinit script.
The root filesystem is mounted via nfs.
Any
Hi,
Am I right in thinking that you could use the md5sum checksum output
generated by running md5sum on any file to verify the contents of that
file, at least in some minimal way? If someone sent me a file, and I
wanted to check to see if it had been corrupted enroute, could both I
and the
Basically I want to dedicate a processor on an smp machine to just
running a particular thread of a program, so nothing else gets cpu time
on the processor, and the thread doesn't get switched to a different cpu.
Any pointers, reccommendations on how to do this, where to look, or even
how to
Probably a big duh, but I need a serial port on my laptop to talk over
minicom to an embedded device for debug. The duh part is I go in to
hook it up and realize that my laptop doesn't have a serial port. So I
need to go get something that'll work for this, but I need it to be
linux
Yeah, I think that's what happened. I ended up searching dejanews and
stumbled across a reference to a mini-HOWTO on dhcp which I managed to
track down, and got it to work, or at least start without errors, and it
looks like it's working for what I'm trying to do.
Thanks,
-Charles
In that
I'm trying to set up dhcpd on my RH7.3 machine, but to no avail, and I'm
not comprehending what I'm doing wrong.
Mostly it seems that I've set the /etc/dhcpd.conf file up wrong, as I
keep getting the following error:
***
No subnet declaration for eth0 (128.171.79.216).
Please write a subnet
I've only got one interface, eth0 (unless we're talking about something
else completely different). I'm not much of a sysadmin guy, sorry if
I'm misunderstanding.
-Charles
You probably only want to run DHCPd on one interface (like eth1):
dhcpd eth1
--MonMotha
I've been looking through the Configure script in the linux/scripts
directory trying to figure out what exactly make oldconfig does, but
I'm missing something. I'm trying to figure out the source for the way
the script makes its decisions. Is it basing it off some other config
file, is it
Any advice on what to do when you're trying to patch a kernel that's
already been patched, which causes the new patch to fail? In
particular, how about patching the kernel versions that come with the
main distros like RedHat? Do you just have to knuckle down and do a
line by line edit?
Sorry if this is already been covered but I'm too obtuse to have noticed.
I recently started using RH 9 for some of my newer machines.
Unfortunately I've been getting segmentation faults for stuff that used
to run on my RH 7.2/3 machines. Examples, I installed the drivers for
some fiber
I just got a 1u rack mount dual Xeon system from these guys:
http://www.swt.com/thin4.html
I was looking at top and noticed that it lists 4 CPU states, where I'd
generally figure there'd be one per CPU. Any ideas why this would be?
I opened it up to be sure, and yep, there's only 2 CPUs.
Sorry for the non-Linux Q.
I'm looking for guides or examples of different code structure models
for Java GUI code. Anybody know of any good sources? As my GUI stuff
gets more complicated, it gets harder to handle, and I'm looking for
alternate ways of improving this.
Thanks,
-Charles
I just got a sony vaio laptop running RH9. After a bit of tinkering, I
was able to get everything pretty much set up right, but when I was
running some tcl/tk gui based apps, they just look bizaare.
if I start wish, and then run the following:
button .b -text Hello
pack .b
I get button
I've got a laptop I want to set up for both work and home. At home I'm
running roadrunner, at work I have an ip address for it. Am I right in
thinking that to get this to work I need to set up 2 network devices,
one for home, one for work, and then activate the device on boot?
Thanks,
Sorry if this is unclear, I'm trying to figure out how to ask the question.
This is primarily in regards to an embedded device, but I don't know the
answer for the standard system either.
Say you've got a board that has a cpu on it that is supported by the
Linux kernel. And you have the
An embedded device I'm using has a kind of wierd, junky, hacked up,
piecemeal version of Linux running on it. One of the problems I'm
having is that the rc.local script doesn't seem to be getting called.
in /etc/inittab, there's the following:
*
# System initialization.
It's called an IPEngine. Comes from a company called Brightstar
engineering. I don't recommend it.
-Charles
Vince Hoang wrote:
Out of curiosity, what architecture and manufacturer are you
using for the embedded device?
Yeah, grepping was the first thing I tried, to no avail. I'll try
adding to the rc script.
Thanks,
-Charles
Vince Hoang wrote:
On Mon, Mar 24, 2003 at 03:16:04PM +, Charles Lockhart wrote:
No where do I find any reference to the rc.local script. I
added some echo comments to the rc.d
Hello,
I've got an embedded device on which I need to run a small server
program. It basically provides a client with access to some custom
hardware. What I've been doing is telnetting in to the device, then
starting the server. What I'd like to do is integrate the server more
tightly
Anybody use mlockall? If so, what effects did it have on you
application performance? Your system performance? What have you heard
about it?
Thanks,
-Charles
you all very much.
-Charles
R. Scott Belford wrote:
On Thursday 13 March 2003 10:03 pm, Charles Lockhart wrote:
Is that interesting? Probably not.
Probably so. Thanks for the info. Very interesting.
scott
Vince Hoang wrote:
Is the memory freed when you quit the application?
If so, that suggest application memory leaks to me.
-Vince
On Fri, Mar 14, 2003 at 07:54:01AM +, Charles Lockhart wrote:
Then the application starts doing disk writes. It's streaming
imaging data to disk at sustained rates
Ok, so I'm still trying to get network connectivity up on my machine. I
tried the latest installs of both Redhat and Mandrake, and I tried
disabling the onboard lan connection (an SiS 900) via the bios, and just
ran with the 3C900B-TPO card. Still, nothing works. Any other ideas,
or am I
So, still messing around trying to get my network connection to work.
Tonight I'll try the version upgrade and see if that works. Hopefully
something will.
-Charles
Warren Togami wrote:
Please boot into single mode and send the following information:
cat /etc/modules.conf
lspci -v
lscpi
Ok, obviously I didn't provide enough info, sorry.
the onboard lan is an SiS 900 based nic. As Jimen suggested, I tried
recompiling the kernel to include the drivers for this device (well, they
listed a SiS900 and an older SiS900 driver, I tried each), to no avail.
I also tried setting up a
Ok, I just built a new system that uses a Asus P4S8X motherboard. I set
it up to dual boot Windows and Linux. The Windows side works fine, but on
the Linux side I'm having problems gettign networking to, uh, work.
I loaded Redhat8.0 onto it. The install seemed to go fine, I don't
remember
The ultimate horror, source code. Uhm, for those people who program in
Java on Linux, is there any reason that everything here wouldn't stack
up nice and clear at least horizontally?
I've got:
button. button. button. button. button. button.
label.. two block TF... label.. two block TF...
two
I'm running a dual boot Win2k/RH7.2 on a Dell Precision 530, which has a
built in sound card. Generally I'm using the KDE interface. Is there
some master volume control somewhere in there? The max volume that I
get under Linux is hearable, but no where high enough for normal
listening, and
Ok, for the record, I am part Irish, and hence I don't think I've ever
had a problem with one of MY systems booting. But I do get called over
whenever someone else has a problem, which is why I'm asking this.
I'm looking for a fairly simple how-to/faq/tutorial document for how to
go in and
Sorry, I haven't been following the entire thread, but, for what it's
worth, I got pretty good performance running RedHat 5.2 on my junky old
75Mhz Pentium with 16MB of memory. It wasn't blazing, but it worked
fine for what I wanted. I don't know what apps you were planning to
use, and there
Anybody here know of a good tutorial or source of info on using UML?
We've started using it here for our software projects, but I'm just not
getting it. I'm trying to put together a class diagram for a system we
already have, but keep getting lost in what all the UML lingo means.
I've got a
I've got an application that consists of three threads:
Thread 1 services a device on the PCI bus, basically reading data from
the device into buffers in memory. After it's filled a buffer, it calls
pthread_cond_signal to start up thread 2, then it goes to sleep for
greater than a set
Anybody know what the fastest way to clear a buffer in C on Linux would
be? Anything faster than
for(i = 0; i sizeofbuffer; i++)
buffer[i] = 0;
?
For some reason I think I saw
memcpy(buffer, 0, sizeofbuffer);
used somewhere once, but after a long day, that looks wierd to me.
I'm looking for a laptop that'll run linux really well, is pretty
powerful, and has a BIG screen, something like the new Vaio 16 laptop
monitor. I seem to be finding a lot of info, but none that I really want.
Anybody know of an info source for laptops that could help me narrow it
down some?
Darned tootin' (totally unsure what that means). I say vote with your
wallet and cancel their service. Go Team!
-Charles, who wishes everybody would quit using RR and leave all the
bandwidth to him.
On Mon, 23 Dec 2002, W. Wayne Liauh wrote:
You seem to have totally ignored the fact that it
I'm brain dumping here:
I've got a Redhat 7.2 machine that I just set up. I've got an app
running on it that's listening on port 8000, but it's refusing all
external connection requests. It will accept internal ones (ie. it
works when I run the client on that system).
The whole setup
I've been noticing multiple instances of logrotate running on my machine
eating up 90% of the cpu. Does that sound apropriate, because it seems
wierd to me?
-Charles
Sorry, I read this off of a different list. Nobody seemed to be
replying there, so I'm assuming they don't know. Has anybody here know
if this kind of thing is possible? It sounds interesting, though not
sure what the application would be.
*** begin excerpt from other email ***
Sorry for
Sorry, if I'm breeching protocol with crossover, please flame me privately.
This article:
http://www.embedded.com/story/OEG20021202S0052
was causing a bit of anger and disgust on another list, and I could
pretty much see why. This guy has 15 years of working experience with
vxWorks, none
Oh yeah, missed that point by a bit. Of course, as a programmer guy, he
could be talking man years, yeah, that's what I meant, yeah, 15 man
years, he puts in 80 hour weeks.
Right.
-Charles
If he had 15 years experience with Linux, it would be named after *him*,
not Linus!-)
reDiculous
It's been my unfortunate personal experience and overall observation
that criticizing people on this list for making negative or disparaging
personal remarks about others just leads them to make negative and
disparaging personal remarks about you, both publicly on the list and
privately via
Please ignore. I don't know what I was thinking. Less than 5 minutes
later I realized it was a waste of bandwidth. My apologies.
-Charles
--
Charles Lockhart
Embedded Software Engineer
NASA InfraRed Telescope Facility
http://irtf.ifa.hawaii.edu/~lockhart
[EMAIL PROTECTED]
(808)956-7635
1 - 100 of 143 matches
Mail list logo
