Linux-Misc Digest #838, Volume #27 Sun, 13 May 01 00:13:02 EDT
Contents:
No sound; Something wrong with mixer. Help ("grendel")
Setting up Database usage for the users (itisme)
Interrupt conflict between PS/2 mouse and pcmcia controller on Winbook (Leonard
Evens)
Re: Linux in college & high school (The Ghost In The Machine)
Re: problem with Gcombust as normal user ([BeoWulf])
tv tuner modules ("Eric Geordi")
Re: 4GB RAM Problem... ("Sipke de Wal")
Need help with installing scigraphica (Madhusudan Singh)
Re: Cannot install "bin" file ("Glitch")
Re: mail problem with RH7.1 (Marek)
Re: Very slow KDE apps (Dave Uhring)
Re: Interrupt conflict between PS/2 mouse and pcmcia controller on (Leonard Evens)
The Distributions (Marshall Lake)
----------------------------------------------------------------------------
From: "grendel" <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,alt.os.linux,alt.os.linux.suse,comp.os.linux.hardware,comp.os.linux.setup,creative.products.sound_blaster.live
Subject: No sound; Something wrong with mixer. Help
Date: Sat, 12 May 2001 23:49:43 GMT
I have a soundblaster live value and have followed the directions in the
article in the SuSE sdb called "ALSA update for SuSE SuSE 6.4/7.0/7.1 and do
not have sound. After the alsa update my card was recognized by alsaconfig.
However when it trys to adjust the volumes and change the mixer it fails. No
asound.conf is created. Here are the error messages. I have verified that
the sound drivers are loaded. They even say this using "alsasound status".
"Could not initialize..."
"No mixer config in the /etc/asounf.conf. You have to unmute your card.
Mixer 0/0 'open error."
I've even tryed manually unmuting the mixer and even accessing it via amixer
but I always get the message about the "Mixer 0/0 ' open error. Even when I
try different device names this happens. Thanks.
------------------------------
From: itisme <[EMAIL PROTECTED]>
Subject: Setting up Database usage for the users
Date: Sun, 13 May 2001 00:08:05 GMT
HOW?????????????. TIA
Jack
------------------------------
From: Leonard Evens <[EMAIL PROTECTED]>
Crossposted-To: redhat.general
Subject: Interrupt conflict between PS/2 mouse and pcmcia controller on Winbook
Date: Sat, 12 May 2001 18:56:13 -0500
I have been struggling with a recalcitrant problem on my Winbook
XL which uses a TI PCI1131 pcmcia controller. The 2.4.2 kernel
which comes with RedHat 7.1 uses kernel modules to support
pcmcia, and the module yenta_socket.o has been completely rewritten.
In the 2.4.2 version, when pcmcia is started, it assigns irq 12
to the controller, but that is the irq normally reserved for the
PS/2 mouse. The kernel doesn't realize this because of reasons
I'm not sure I understand, but it has something to do with the
interrupt not being used unless the mouse is in use. The upshot
of the problem is that the mouse and the keyboard (which uses the
same controller as the PS/2 mouse) hang because of the interrupt
conflict. In console mode, gpm hangs the mouse and keyboard
if pcmcia has been started. Even if you stop pcmcia, that doesn't
solve the problem because the pcmcia controller still keeps
the interrupt. If you start X, mouse and keyboard hang.
With quite a lot of help from David Hinds, I've found
tentatively that the following seems to work. Pass an IRQ mask
to the kernel with
pci=irqmask=0x8fff
This excludes irqs 12, 13, and 14 and forces yenta_socket to choose
15. You can pass this to the kernel directly by entering
at the lilo prompt
linux pci=irqmask=0x8fff
or you can add the following line to /etc/lilo.conf
in the image section for the linux kernel
apppend="pci=irqmask=0x8fff"
and then run lilo.
I originally tried masking off just 12, but for some reason I
don't understand, I encountered a problem with hda although it
also appeared to assign irq 15 to the pcmcia controller.
I hope this actually fixes the problem and I hope other people with
Winbook laptops will try it to see if it works and/or if other
problems crop up.
This problem may already have been fixed in the 2.4.4 kernel.
I will try to find out.
Some other people have reported possibly related problems with
the PS/2 mouse under RH7.1. I suspect something similar is going on,
but I don't know what.
--
Leonard Evens [EMAIL PROTECTED] 847-491-5537
Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208
------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Crossposted-To: comp.os.linux.advocacy
Subject: Re: Linux in college & high school
Date: Sun, 13 May 2001 00:56:52 GMT
In comp.os.linux.advocacy, mmnnoo
<[EMAIL PROTECTED]>
wrote
on Fri, 11 May 2001 03:03:47 GMT
<nEIK6.60080$[EMAIL PROTECTED]>:
>Here's a list of the machines in the computer science dept
>at the University of New Mexico:
>
>http://www.cs.unm.edu/~ssg/SSG_Hardware/byos
>
>any questions?
I will note for the record that a large number of the machines
are running Debian 2.2, a few are running IRIX, and, of the
more than 80 machines listed there, only 4 are known to run any
variant of Windows (there are two unknowns). And one of them --
versaka-vm -- is cheating by using VMWare. :-)
This according to a quick scan of the article. Impressive, actually.
I'm kinda curious as to the manpower required to install and maintain
all this machinery, and how that manpower would compare to a
similar-sized lab running pure NT/W2k equipment.
>
>In article <[EMAIL PROTECTED]>, "Christopher Corbell"
><[EMAIL PROTECTED]> wrote:
>
>> I'm looking for leads to information, statistics, or just individual
>> testimonials about the use of Linux in educational settings,
>> particularly in high school, community college, university, and grad
>> school settings. Does anyone out there know of any general sources of
>> information on the use of Linux in these settings? I would especially
>> be interested in the use of Linux in math & science education. Also,
>> I'd like to know about any advocacy groups, PC 'salvage' groups or
>> similar organizations that are active in getting Linux used in schools.
>>
>> Thanks for any info.
>> - Christopher
--
[EMAIL PROTECTED] -- insert random misquote here
EAC code #191 12d:21h:15m actually running Linux.
[select one]
Life's getting too complicated, even listening to the radio.
Microsoft. When it absolutely, positively has to act weird.
I was asleep at the switch the rest of the time.
Darn. Just when this message was getting good, too.
Would Emperor Palpatine approve of this?
Would Emperor Palpatine approve of this?
Linux. The choice of a GNU generation.
Yes, uptime & wall clock aren't in synch; I don't know why.
------------------------------
From: [BeoWulf] <[EMAIL PROTECTED]>
Subject: Re: problem with Gcombust as normal user
Crossposted-To: comp.os.linux.setup,alt.os.linux.mandrake
Date: Sun, 13 May 2001 00:57:47 GMT
As I glanced into the mist, I noticed the sound of a whispering voice. I=
t=20
appeared to be coming from the whereabouts of Yves Leung-Tack, and the=20
voice seemed to be saying :
> Hi,
>=20
>=20
> When I try to run the gcombust I got these messages :
>=20
> /usr/bin/cdrecord: Operation not permitted. WARNING: Cannot do
> mlockall(2).
See [1]
> /usr/bin/cdrecord: WARNING: This causes a high risk for buffer
> underruns.
See [2]
> /usr/bin/cdrecord: Operation not permitted. WARNING: Cannot set
> RR-scheduler
Yup, see [1]
> /usr/bin/cdrecord: Permission denied. WARNING: Cannot set priority usin=
g
> setpriority().
Again, see [1] ;-)
> /usr/bin/cdrecord: WARNING: This causes a high risk for buffer
> underruns.
Again, see [2] :-)
> /usr/bin/cdrecord: Operation not permitted. shmctl failed to lock share=
d
> memory segment
>=20
>=20
> Does someone known how to fix this ?
> As root, it works just fine. And I check that user is part of the
> cdrom/cdwriter group ...
>=20
>=20
> Merci !
Here goes ;-)
[1] You do not have the proper permissions to *write* to a device. By=20
default, only root has that prerogative. This is a normal and logical=20
setup, so it "works as designed". ;-)
You can however find a workaround by changing ownerships and permissions =
to=20
some files, but the easiest way around it is to run Gcombust as root. Fr=
om=20
within KDE, create a desktop link with in the section for the executable'=
s=20
name *kdesu* *-c* *gcombust*. Alternatively, install the *sudo* package=20
and read the manuals on it... ;-)
[2] You probably have writing speed set too high, which on burners that=20
connect to the IDE/ATA controller (though emulated as SCSI) may indeed=20
cause buffer underruns. In case of a cdrw, you can retry afterwards. In=
=20
case of a cdr, you get a nice shiny coaster... ;-=DE
--=20
With sincere regards,
[BeoWulf].
------------------------------
From: "Eric Geordi" <[EMAIL PROTECTED]>
Subject: tv tuner modules
Date: Sat, 12 May 2001 20:30:01 -0500
Anybody know what the right tuner module to load is for a phoebemaster tv
card w/o the fm radio ? Thanks
Eric
------------------------------
From: "Sipke de Wal" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: 4GB RAM Problem...
Date: Sun, 13 May 2001 03:45:54 +0200
Does your MOBO support caching for up to 4GB ???
If so is it enabled in the BIOS??
Some OS-ses like to allocate RAM from the highest locations downwards so
they might run into uncached ram very soon if the caching on your MOBO is
not effective for the top of RAM. And that slows things down dramatically!
Sipke de Wal
===================================
http://xgistor.ath.cx
===================================
Michael Sabielny <[EMAIL PROTECTED]> schreef in berichtnieuws
9dbh3v$8ir$[EMAIL PROTECTED]
> Hi folks,
>
> recently we got a new computer with these features:
> 2 PIII-CPUs
> 4 GB ECC-RAM (4 DIMMs)
> Supermicro 370DL3 Mainboard
> OS: SuSE Linux 7.1
> Kernel: SMP-Kernel with Support for more than 4 GB of RAM
>
> Wen encountered a strange effect: when we use only 3 Gigs of the RAM (with
> one DIMM removed) the computer behaves absolutely normal and shows its
> expected performance.
>
> But if we put the fouth DIMM it gets very slow, e.g. needs the double time
> for a numerical matrix decomposition. Even the bootup time increases
> dramatically.
>
> We tried various different kernels:
> SuSE precompiled 2.2.18-SMP
> SuSE precompiled 2.4-SMP
> selfcompiled 2.4-SMP-4GB
> selfcompiled 2.4-SMP-64GB
>
> But all kernels have the same behavior! In the moment we put the 4. DIMM
to
> get the full 4 Gigs we get slow...
>
> Has anybody got some experiance with this kind of problem?
>
> Best regards,
>
> Michael Sabielny
> Technical University of Hamburg-Harburg
> Department of Theoretical Electrical Engineering
> 21071 Hamburg
> Germany
>
>
>
------------------------------
From: Madhusudan Singh <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Need help with installing scigraphica
Date: Sat, 12 May 2001 22:03:34 -0400
Hi,
I wanted to install scigraphica on my CAEN Linux 6.1 (a variant of
Redhat 6.1) system.
I got a failed dependency :
root@guo src]# rpm -Uvh scigraphica-0.6.3-1.i386.rpm
error: failed dependencies:
libreadline.so.4.1 is needed by scigraphica-0.6.3-1
When I searched for libreadline.so.4.1 on the net, I got the package
readline-4.1-9.i386.rpm that contains this.
I tried to install this, but ran into a problem,
[root@guo src]# rpm -Uvh readline-4.1-9.i386.rpm
error: failed dependencies:
readline = 2.2.1 is needed by readline-devel-2.2.1-5
libhistory.so.3 is needed by libxml2-2.2.8-1
libreadline.so.3 is needed by R-base-0.64.2-caen_4
libreadline.so.3 is needed by arla-0.31-12
libreadline.so.3 is needed by arla-monitor-0.31-12
libreadline.so.3 is needed by bc-1.05a-4
libreadline.so.3 is needed by cdecl-2.5-9
libreadline.so.3 is needed by ext2ed-0.1-19
libreadline.so.3 is needed by ftp-0.15-1
libreadline.so.3 is needed by fvwm2-2.2.2-3
libreadline.so.3 is needed by git-4.3.17-5
libreadline.so.3 is needed by gnuplot-3.7.0.1-4
libreadline.so.3 is needed by guile-1.3-7
libreadline.so.3 is needed by krb4-libs-1.0-6
libreadline.so.3 is needed by krb4-workstation-1.0-6
libreadline.so.3 is needed by libgtop-examples-1.0.5-1
libreadline.so.3 is needed by pilot-link-0.9.0-8
libreadline.so.3 is needed by python-1.5.2-7
libreadline.so.3 is needed by samba-2.0.5a-12
libreadline.so.3 is needed by samba-client-2.0.5a-12
libreadline.so.3 is needed by samba-common-2.0.5a-12
libreadline.so.3 is needed by units-1.55-3
libreadline.so.3 is needed by postgresql-7.0.2-2
libreadline.so.3 is needed by postgresql-server-7.0.2-2
libreadline.so.3 is needed by libxml2-2.2.8-1
Can someone think of a workaround ?
Thanks,
Madhusudan Singh.
------------------------------
From: "Glitch" <[EMAIL PROTECTED]>
Subject: Re: Cannot install "bin" file
Date: Sat, 12 May 2001 22:47:39 -0400
In article <[EMAIL PROTECTED]>, "ThanhVu Nguyen"
<[EMAIL PROTECTED]> wrote:
> I try to installed Staroffice, I installed it couple times before by
> just running sh filename.bin , but on this new system, when I issue that
> command, it just say cannot execute binary file. Am I missing some
> library or compiler etc ?
>
> thanks for all inputs
the file needs to have execute permissions on it
do chmod 755 soffice.bin or whatever the filename is, that will make it
executable
then just ./filename will start it to install
------------------------------
From: Marek <[EMAIL PROTECTED]>
Crossposted-To: redhat.general
Subject: Re: mail problem with RH7.1
Date: Sat, 12 May 2001 21:54:28 -0500
Robert Lynch wrote:
>
> Marek wrote:
> >
> > Robert Lynch wrote:
> > >
> > > Leonard Evens wrote:
> > > >
> > > > We have experienced a mail problem with RedHat 7.1.
> > > > Apparently we now need to explicitly include sendmail
> > > > in /etc/hosts.allow or else tcpwrappers intervenes.
> > How did you include sendmail in host allow??
> > I did
> > SMTP:ALL
> >
> > > Maybe this from the RELEASE-NOTES for RH 7.1 is of interest:
> >
> Maybe you could get a fix if you tried what I ORIGINALLY posted
> (from the RH7.1 RELEASE-NOTES):
>
yes... I did ..... as I marked at my email.
Actually I solved the problem .... AND IT IS NOT WHAT YOU SUGGESTED
You would notioced in my post that I said that I have in my maillog
rejection of emails because of tcpwrapper and that my sendmail is
working for intranet (that excludes coplately you "solution")
FOR OTHERS WHO WILL HAVE PROBLEM WITH SENDMAIL AND WILL NOTICE REJECTION
OF EMAILS BECAUSE OF TCPWRAPPER:
put in your /etc/host.allow
sendmail:ALL
------------------------------
From: Dave Uhring <[EMAIL PROTECTED]>
Subject: Re: Very slow KDE apps
Date: Sat, 12 May 2001 21:57:49 -0500
Professor J Frink wrote:
>>Run 'top' and see what is hogging your CPU.
>
> I suspect this is merely the large overhead of KDE2 rather than anything
> going mad in the background.
>
> I've seen it myself on trying to run konqueror in non KDE2 desktops and
> that of my users using KDE2 systems.
>
> The whole shebang of konqueror itself + the many kdeinits seems to give an
> easy way of coordinating things but by golly it takes a while to load up.
> It doesn't seem to matter much the amount or RAM or CPU you have spare it
> just takes ages to get going. And it also takes ages to stop (plenty of
> confused users here who hit "logout" only to be presented with their
> desktop while it 'shuts down' still fully operable until it finally
> decides to drop back to xdm/kdm seemingly at random and without warning).
>
> Mozilla loads faster. Netscape loads faster. While I like konqueror I
> won't defend its load times at all.
>
> I don't like the complex and process hungry routines of KDE2 (what is it
> with all these kdeinits, kdesud and nspluginviewers; they pop up all over
> the place, giving me little knowledge as to what they're doing and often
> end up just crashing; could you *please* at least show what you're trying
> to run?!). Why should loading up one kde2 app entail the baggage of pretty
> much the entirety of kde2 as well?
>
> KDE2 needs good RAM and CPU resources. Using any single KDE2 package
> necessitates little less than this. I, for one, loathe KDE2 for this and
> personally, on a server based system, find it nothing more than a huge
> pain in the ass (or RAM) to allow my users to run it. Sure, it looks
> nice...
>
> On the same hardware Windows, as a desktop, is quicker, less resource
> hungry and complete, if you deny that then you're wearing the old
> rose-tinted specs.
>
> I dunno if it's cos of preloading or what but Windows, on the same
> hardware, loads its desktop apps a hell of a lot quicker, runs them
> quicker, more seamlessly and often crashes less when using them.
>
> Linux was never *designed* to be a desktop system yet here we are trying
> to beat it into being one. It would suit most users better to start afresh
> with a system designed for them not to have to try to bend the users into
> knowing the OS. But, of course, such a new system would have no support
> from anyone; see BeOS...
>
> Frink
>
KDE's overhead, while objectionable, is not sufficient to cause the
difficulty the OP is experiencing. He has a 700MHz CPU and 128MB RAM.
While Windows may utilize fewer resources for implementing a desktop, it
remains a single-user, barely multi-tasking, blue-screening POS with no
security features whatsoever.
If you run X on a server, then you deserve the performance penalty. My
servers do not even have monitors attached to them. And if you don't like
KDE then don't use it; but refrain from telling others not to use it. In
my experience, the KDE applications work quite well and the load time is
comparable to Windows 98SE, even considering that Windows preloads most of
IE on boot.
Add a few applications to your Windows box and just watch the System Tray
fill up and the boot up time go to minutes. And just how does the average
user know how to delete those things from the System Tray? I know and I
suspect that you may, but Joe Blow certainly has no business futzing around
in the registry. At least with Linux, you can easily and readily determine
which apps will start at boot.
Take your FUD to www.microsoft.com, it is not needed here. We already have
chosen to use Linux.
------------------------------
From: Leonard Evens <[EMAIL PROTECTED]>
Crossposted-To: redhat.general
Subject: Re: Interrupt conflict between PS/2 mouse and pcmcia controller on
Date: Sat, 12 May 2001 22:21:15 -0500
Leonard Evens wrote:
>
> I have been struggling with a recalcitrant problem on my Winbook
> XL which uses a TI PCI1131 pcmcia controller. The 2.4.2 kernel
> which comes with RedHat 7.1 uses kernel modules to support
> pcmcia, and the module yenta_socket.o has been completely rewritten.
>
> In the 2.4.2 version, when pcmcia is started, it assigns irq 12
> to the controller, but that is the irq normally reserved for the
> PS/2 mouse. The kernel doesn't realize this because of reasons
> I'm not sure I understand, but it has something to do with the
> interrupt not being used unless the mouse is in use. The upshot
> of the problem is that the mouse and the keyboard (which uses the
> same controller as the PS/2 mouse) hang because of the interrupt
> conflict. In console mode, gpm hangs the mouse and keyboard
> if pcmcia has been started. Even if you stop pcmcia, that doesn't
> solve the problem because the pcmcia controller still keeps
> the interrupt. If you start X, mouse and keyboard hang.
>
> With quite a lot of help from David Hinds, I've found
> tentatively that the following seems to work. Pass an IRQ mask
> to the kernel with
> pci=irqmask=0x8fff
> This excludes irqs 12, 13, and 14 and forces yenta_socket to choose
> 15. You can pass this to the kernel directly by entering
> at the lilo prompt
> linux pci=irqmask=0x8fff
> or you can add the following line to /etc/lilo.conf
> in the image section for the linux kernel
> apppend="pci=irqmask=0x8fff"
> and then run lilo.
>
> I originally tried masking off just 12, but for some reason I
> don't understand, I encountered a problem with hda although it
> also appeared to assign irq 15 to the pcmcia controller.
It may have assigned 14 to one branch of the controller and 15 to
the other.
>
> I hope this actually fixes the problem and I hope other people with
> Winbook laptops will try it to see if it works and/or if other
> problems crop up.
>
> This problem may already have been fixed in the 2.4.4 kernel.
> I will try to find out.
>
> Some other people have reported possibly related problems with
> the PS/2 mouse under RH7.1. I suspect something similar is going on,
> but I don't know what.
>
> --
>
> Leonard Evens [EMAIL PROTECTED] 847-491-5537
> Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208
The mask 0x8fff led to some other problems with gnome when pcmcia
was active. 0xafff seems to work better. But my laptop only
has 32MB of ram, so it might not be the ideal test machine. I would
be interested if anyone else can test any of these workarounds.
--
Leonard Evens [EMAIL PROTECTED] 847-491-5537
Dept. of Mathematics, Northwestern Univ., Evanston, IL 60208
------------------------------
From: [EMAIL PROTECTED] (Marshall Lake)
Subject: The Distributions
Date: 13 May 2001 04:08:59 GMT
I'm not sure if this is the right place to post this. If not then please
direct me. Or if there is a document available someplace which already
covers this subject then please let me know.
I would like to hear what others think is the best Linux distribution
available containing the 2.4.x kernel.
The most important factors for me are 2.4.x kernel availability,
everything working, amount of utilities and programs available with the
distro (the more the better), and latest versions of everything.
Unimportant factors for me are ease of installation (I figure I should
be able to install anything as long as others already have) and cost
(within reasonable limits).
I appreciate any input at all.
Marshall Lake - [EMAIL PROTECTED] - http://melake.erols.com
------------------------------
** 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 comp.os.linux.misc.
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-Misc Digest
******************************
