Linux-Hardware Digest #292, Volume #14 Sat, 3 Feb 01 11:13:09 EST
Contents:
Re: Heavy IDE use freezes screen and sound card ("Flacco")
Have you seen my interrupt? ("Sebastiano Lanza")
Modem [limodem] configuration ("Panagiotis")
Re: Modem problems (Rob Clark)
Re: Modem problems ([EMAIL PROTECTED])
Re: Random Reboots & Freezes ([EMAIL PROTECTED])
Re: dell 300 server blows vga on 2.2.* ("Chris Ripp")
Re: Heeelllp Mr. Wizard!!! (Steve Davies)
Re: Modem [limodem] configuration (Uwe Bonnes)
Re: dell 300 server blows vga on 2.2.* ("Peter T. Breuer")
Re: Recommendations for ethernet cards and other hardware wanted... (Andrew Robbie)
Re: Recommendations for ethernet cards and other hardware wanted... (Andrew Robbie)
Re: Linux solutions for VoIP and text editors ("V.G.Guhan")
Re: Heavy IDE use freezes screen and sound card (Dances With Crows)
Re: Have you seen my interrupt? (Dances With Crows)
----------------------------------------------------------------------------
From: "Flacco" <[EMAIL PROTECTED]>
Subject: Re: Heavy IDE use freezes screen and sound card
Date: Sat, 03 Feb 2001 05:55:33 -0500
> hdparm -c1 -m16 -u1 /dev/hda should help. It's the u1 that will make a
> lot of the difference. "man hdparm" to find out more.
That seems to do it - where would you recommend I put this so it runs
automatically when the system boots?
------------------------------
From: "Sebastiano Lanza" <[EMAIL PROTECTED]>
Subject: Have you seen my interrupt?
Date: Sat, 03 Feb 2001 11:51:26 GMT
I have a goldstar eide cd GCD580B as slave on IDE0,
when I attempt to read anything I get ' HDB: LOST INTERRUPT'
and the system hangs (...GOSH!!!)
It works under win98 (I read somewher win98 doesn't use interrupt with CD)
I tried to pass 'ide0=serialize' to the kernel but didn't work
I will offer a nice-smelling heaven-taste slice of napoletan pizza to anyone
helps me ;)
------------------------------
From: "Panagiotis" <[EMAIL PROTECTED]>
Subject: Modem [limodem] configuration
Date: Sat, 3 Feb 2001 13:55:05 +0200
dear sirs,
I own an internal modem by PCTel which, as it is supported by PCTel, it
supports Linux OS
whith the use of the driver PCTel distributes in the CD which comes with the
modem..
due to my lack of knowledge I havent been able to make this work..
the directories are like that
-include/hal.h , pctel.h
/root -src/module/makefile , ptmodule.c
-lib/control.a , dsp.a , hal.a , ptserial.a
here is the makefile:
#
# Makefile for the PCtel module
#
# Note! The CFLAGS definition is now inherited from the
# parent makefile.
#
HPATH = ../../include
FINDHPATH = $(HPATH)/asm $(HPATH)/linux $(HPATH)/scsi $(HPATH)/net
HOSTCC =gcc -I$(HPATH)
HOSTCFLAGS =-O -fomit-frame-pointer
CROSS_COMPILE =
COL_DEBUG_FLAGS =
AS =$(CROSS_COMPILE)as
LD =$(CROSS_COMPILE)ld
CC =$(CROSS_COMPILE)gcc -D__KERNEL__ -DMODULE -Wall $(HOSTCFLAGS) -I$(HPATH)
${COL_DEBUG_FLAGS}
CPP =$(CC) -E
AR =$(CROSS_COMPILE)ar
NM =$(CROSS_COMPILE)nm
STRIP =$(CROSS_COMPILE)strip
MAKE =make
EXTRA_LDFLAGS :=
# object files directory
O_TARGET := ../../lib/pctel.o
# object files
O_OBJS := \
ptmodule.o \
OX_OBJS := \
../../lib/control.a \
../../lib/hal.a \
../../lib/dsp.a \
../../lib/ptserial.a
all_targets: $(O_TARGET) $(L_TARGET)
%.s: %.c
$(CC) $(CFLAGS) $(EXTRA_CFLAGS) -S $< -o $@
%.o: %.c
$(CC) $(CFLAGS) $(EXTRA_CFLAGS) -c -o $@ $<
%.o: %.s
$(AS) $(ASFLAGS) $(EXTRA_CFLAGS) -o $@ $<
ifdef O_TARGET
ALL_O = $(O_OBJS) $(OX_OBJS)
$(O_TARGET): $(ALL_O) Makefile
rm -f $@
$(LD) -r $(EXTRA_LDFLAGS) -o $@ $(ALL_O)
endif
clean:
rm -f *.o
ifdef L_TARGET
rm -f $(L_TARGET)
endif
ifdef O_TARGET
rm -f $(O_TARGET)
endif
this is the end of the makefile
according to the manual the output shoould be a file in /root/lib called
pctel.o which must
be installed with insmod.. but insmod gives me an error that there is a
kernel mismatch
My kernel is 2.2.16 (I own SuSe 7.0 prof)
Where can I declare in the makefile the fact that my kernel is 2.2.16?
Maybe the problem is on /root/src/module/ptmodule.c take a look at it..
/*
* For the definition of __module_kernel_version[] so that
* the modem driver can be compatible with various Linux kernel
* versions.
*
* This file is to be linked with the rest of the libraries in
* the distribution package.
*/
#include <linux/version.h>
const char __module_kernel_version[] __attribute__((section(".modinfo"))) =
"kernel_version=" UTS_RELEASE;
#ifdef MODVERSIONS
const char __module_using_checksums[] __attribute__((section(".modinfo"))) =
"using_checksums=1";
#endif
Thank you in advance..
P.S I have tried lots of times to edit the scripts both makefile and
ptmodule.c but all i can get
is a headache.. PCTel support sucks too!
This driver is supposed to output a file
/root/lib/pctel.o which I setup with insmod and then
mknod /dev/ttyS15 c 62 79
and here it is /dev/ttyS15
there i do
chmod 666 /dev/ttyS15
and then i link it
ln -sf /dev/ttyS15 /dev/modem
for the use with other programmes like kppp etc..
thats all...
i need all your help, sorry for my bad english i am greek, thats why..
panagiotis
P.S2 sorry for the length of the message
------------------------------
Subject: Re: Modem problems
From: [EMAIL PROTECTED] (Rob Clark)
Date: Sat, 03 Feb 2001 12:32:24 GMT
In article <95eeqa$tc4$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
>I a Conexant HCF V90 56K Speakerphone PCI. Seeing the HCF in the name
>made me assume that it wouldn't work under linux (i am running RH7) but
>after looking at this site http://www.o2.net/~gromitkc/winmodem.html I
>get the impression that it could be made to work. I have downloaded two
The HSF drivers will not work with your HCF modem. Sorry :(
Rob Clark, [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Modem problems
Date: Sat, 03 Feb 2001 13:01:26 GMT
Ok, so are there any drivers available that will work with it?
Sent via Deja.com
http://www.deja.com/
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Random Reboots & Freezes
Date: Sat, 3 Feb 2001 11:18:22 GMT
Andrew n marshall <[EMAIL PROTECTED]> wrote:
[snipped]
Andew,
generally three things to check.
1. Make sure your memory is good. Swap it with some other machine
if possible.
2. Make *very* sure your cpu is not getting to hot. Are the fans
turning full speed? Also touch the cpu itself with your finger!
I have found more than one, that often there is bad thermal
contact between the cpu and the heat sink! Thus resulting in an
overheated cpu and a cold heatsink!!
I don't know - that heating compound is only a few pennies. :-)
3. Power Supply. You may have bad contact at the connectors.
These 3 are about 90% of all trouble with MOBOS nowadays.
Regards,
Friedhelm
--
Microsoft is NOT the answer. Microsoft is the Question.
The answer is: "NO!"
===================================================================
Friedhelm Mehnert, Berliner Allee 42, 22850 Norderstedt, Germany
phone + fax: +49-40-5236562 email: [EMAIL PROTECTED]
===================================================================
------------------------------
From: "Chris Ripp" <[EMAIL PROTECTED]>
Subject: Re: dell 300 server blows vga on 2.2.*
Date: Sat, 3 Feb 2001 07:12:01 -0600
recompile the kernel *not* to use framebuffer console? Just use a std. vga
console. If I understand that's what's going on here is it's using fb mode
instead of a plain jane vga mode. Maybe there's a way to disable the fb by
passing something to lilo, I don't know.
Just my guess.
"Peter T. Breuer" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
[snip]
> fb0: ATY Mach64 frame buffer device on PCI
> vga16fb: initializing
> vga16fb: mapped to 0xc00a0000
> fb1: VGA16 VGA frame buffer device
> Detected PS/2 Mouse Port.
[snip]
------------------------------
From: Steve Davies <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.networking,linux.redhat.ppp,linux.dev.ppp,comp.os.linux.redhat,comp.os.linux.admin,alt.os.linux
Subject: Re: Heeelllp Mr. Wizard!!!
Date: Sat, 03 Feb 2001 14:15:14 +0000
Linux NFS out-of-the-box is complete pants, and will often freeze its
networking subsystem. (I like Linux BTW, just not its 2.2.x NFS
implementation.)
You might find that upgrading to a 2.4 kernel fixes this, but this is
non-trivial. Otherwise the only real workaround I've found is to reduce the
rsize and wsize parameters on the NFS mount to 1024 or less (1024 works for me
on a LAN) This slows nfs down a lot, but over a CM link, you'll probably not
even notice.
Let me know if this helps.
Steve
"Jeffrey S. Mulliken" wrote:
>
> My company is really having a serious problem, and I'm pleading for
> anyone who can, to help us.
>
> Here's the situation:
>
> At the Corporate H.Q.:
>
> RedHat Linux 2.2.17 Kernel machine as a gateway / VPN (via PPTP)
>
> server
> RedHat Linux 2.2.17 Kernel machine serving NIS and Sendmail
> services
> A Sun Ultra 2 as a file server, serving NFS mounts on Solaris
> 2.6
> A couple of Sun Enterprise 3000's used as compilers and
> ClearCase view servers.
>
> At the Developers Homes:
>
> RedHat Linux 2.2.17 Kernel machines running PPTP VPN tunnels
> Linksys 4 port Router/hub's
> Win9X or 2K machines also behind the Linksys boxes
> Throug a variety of ISP's to the net...i.e. @home/Sprint
> DSL/PacBell DSL...etc.
>
> Everyone is able to boot their Linux machines up successfully, and
> the PPTP sessions get established,NIS connects to the remote domain, and
>
> NFS mounts happen. All seem ok, until any large amount of data is moved
>
> via the NFS mounts. Performance is VERY slow. Even if NFS mounts are
> not used, at random intervals, the PPTP connection will just drop,
> causing the NIS domain to lose it's bind. The recovery is to init 1,
> and then init 5.
>
> The Linksys boxes have a limitation that restricts you to only one
> VPN connection through the router, so the remote Windoze machines can't
> be VPN'ed in while the Linux boxes are, but with the Linux box down,
> they can establish VPN with their Windoze machine, but it will also lose
>
> it's connection after some period. We have tried eliminating the
> Linksys boxes from the equation, and it does not seem to buy us much, if
>
> any, improvement. We don't have the VPN with IPMasq. patch on the
> kernels anywhere, because we aren't doing the IPMasq. stuff.
>
> Any 'gurus' out there who have any ideas as to what the problem
> might be here?
>
> Help us Obiewan Kenobe, you're our only hope!
>
> Jeff Mulliken
> [EMAIL PROTECTED]
>
> Feel free to email me your response
--
Steve Davies [EMAIL PROTECTED]
http://www.one47.demon.co.uk
PGP Fingerprints:
DH/DSS : 5D85 8164 91D7 E9CC 4F80 842B AB86 93D9 8938 7612
RSA : 4E2E E60F 3D76 9E7E 70F9 901B 70FA 56C8
------------------------------
From: Uwe Bonnes <[EMAIL PROTECTED]>
Subject: Re: Modem [limodem] configuration
Date: 3 Feb 2001 14:20:08 GMT
Panagiotis <[EMAIL PROTECTED]> wrote:
: dear sirs,
: I own an internal modem by PCTel which, as it is supported by PCTel, it
: supports Linux OS
: whith the use of the driver PCTel distributes in the CD which comes with the
: modem..
: due to my lack of knowledge I havent been able to make this work..
Go to http://www.linmodems.org Stroll around there and read the archive of
the mailing list. Lots of cries for help but some usefull information too.
Bye
--
Uwe Bonnes [EMAIL PROTECTED]
Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt
========= Tel. 06151 162516 ======== Fax. 06151 164321 ==========
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Subject: Re: dell 300 server blows vga on 2.2.*
Date: Sat, 3 Feb 2001 15:10:41 +0100
Chris Ripp <[EMAIL PROTECTED]> wrote:
> recompile the kernel *not* to use framebuffer console? Just use a std. vga
> console. If I understand that's what's going on here is it's using fb mode
My guess also. I'll try that next (it wasn't smp in itself - I tried a
uniprocessor kernel - still get txt consoles filled with little
rectangles and multiple hazy images under VGA16).
However, I am struggling with a CL webcam go+ at the same time, so ...!!
> instead of a plain jane vga mode. Maybe there's a way to disable the fb by
> passing something to lilo, I don't know.
Neither does the author of the BootPrompt HOWTO. He seems as confused
about what video=stuff one can put as I am.
> Just my guess.
Thanks!
> "Peter T. Breuer" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> [snip]
>> fb0: ATY Mach64 frame buffer device on PCI
>> vga16fb: initializing
>> vga16fb: mapped to 0xc00a0000
>> fb1: VGA16 VGA frame buffer device
>> Detected PS/2 Mouse Port.
> [snip]
Peter
------------------------------
From: Andrew Robbie <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.x
Subject: Re: Recommendations for ethernet cards and other hardware wanted...
Date: Sun, 04 Feb 2001 02:06:07 +1100
In article <[EMAIL PROTECTED]>, "Peter T. Breuer"
<[EMAIL PROTECTED]> wrote:
> In comp.os.linux.hardware Dr. Ram Samudrala <[EMAIL PROTECTED]>
> wrote:
> > of use to someone besides me. The overall goal of this project is to
> > have a 512 processor (1 GHz +) cluster over a span of a year or
> > two. Right now, I'm looking at buying 64 processors.
I hope you are talking to some of the people who will have to maintain
and program this behemoth. Do the maths for 512 CPUs and calculate the
MTBF. Then realise that if you want good uptime you will need reliable
components (ie not always the cheapest) and some redundancy in hardware
and software.
> > are" comment... is it dependent on the driver release or is it more
> > stochastic? (:
>
> Eepro100 cards can be wonky. About 20% of the ones I have show bad
> characteristics (such as stopping working after 24hrs, so that they
> have to be reset by a driver rmmod insmod every 12 hours as a
> preventative measure). Mind you, I haven't been tracking how old these
> cards are .. they could be up to three years old. I haven't checked
> which drivers are best for them either. Too many variables. The cards
> also seem only to be able to get about 60% of the available bandwidth,
> even on switched 100BT networks.
Switched doesn't mean 'no collisions'. The common situation, eg a
single server and many clients, is moderately pathological. I would
say 60% of maximum is pretty good.
> OTOH the rtl8139 cards I have have never given trouble any time. And
> they use about 100% bandwidth. But given their amazingly cheap price
> (about $25 ?) I expect to find someday that I have received a batch of
> no-hopers. Fine .. I'll can that when it arrives.
Perhaps your rtl8139s are in faster/different computers? Different
workload?
> >>> priced around $50-$100, large disks (what is the largest IDE disk
> >>> supported by Linux today - can I get the Maxtor 80 GB drives?), and
> >>> dual processors (right now I'm looking at 1 GHz PIIIs -- how reliable
> >>> are the 1.2 GHz ones?).
>
> >>Go for a plain dual asus or abit BX board.
Why? So he'll be buying obsolete and slow hardware?
> > So this brings up another issue. The asus and the abit boards use the
> > VIA chipset (as opposed to the Intel chipset) on the motherboards. My
>
> I do not believe they do - they are BX boards! Mind you, I cannot check
> right now.
You are wrong. They do have 440BX boards still (also GX) but the hot
new boards are all VIA,etc. Intel have completely dropped the ball.
If you are willing to pay for RDRAM the i840s can be used. But the
non-intel chipsets seem to work pretty well.
> > vendor suggests that these might not be as stable as the Intel
> > chipsets. Does anyone have information about this?
Sounds like your vendor wants to sell you something with Intel inside.
> > This is a problem
> > because right now I'm looking at the SuperMicro DL3 motherboards which
> > don't have AGP slots (this is okay for the cluster but not for the
> > desktops).
Supermicro make the SUPER 370DE6 (or is it DER?) which has AGP as well
as 64 bit PCI. We plan on buying some.
> > How do the onboard ethernet controllers on these motherboards work?
They communicate with the 'south bridge' part of the chipset directly
or they are hardwired PCI devices. Similar to the IDE support.
> > What about RAM? Does any ECC RAM work or do certain manufacturers have
> > better reptutations?
>
> I am not using ECC ram - too difficult to get hold of and impossible to
> mix later. I only have bad ram when ram is in short market supply - but
> of course I don't know for sure :-). Nevertheless my experience is that
> once ram has been tested thoroughly, bit errors do not come from ram
> but from overheated cpus.
Overheated CPUs causing RAM errors? Or over-clocked CPUs producing the
wrong result? Ever heard of ionizing radiation?
Anyway, your experience ain't backed up with any understanding. RAM
error rates in a farm of 64 machines (* 512MB = 32GB) could be a
significant hassle. ECC aids in diagnosis of problems, detection of
which dimm is faulty and can even correct some errors.
later,
Andrew
------------------------------
From: Andrew Robbie <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.os.linux.x
Subject: Re: Recommendations for ethernet cards and other hardware wanted...
Date: Sun, 04 Feb 2001 02:27:03 +1100
In article <95fdca$odu$[EMAIL PROTECTED]>, [EMAIL PROTECTED]
wrote:
> I have more thoughts on this and in general this will end up in some
> sort of a document on the design of large clusters for high
> performance computing using Linux (others have done this also, but this
> is my configuration).
Sounds like a sensible idea.
> Regarding the disks, it seems to me that a 60 GB IDE disk would be
> slow in terms of file system checks (this would point the need to a
> different kind of a file system).
Yes, with plain ext2 a full 60BG disk might take a long time to check.
For this reason there is lots of interest in thing like logging file
systems like ReiserFS, XFS from SGI, LFS or something from IBM, or the
similar SoftDeps model. However, at this stage these systems are not
heavily tested on Linux.
However, the standard practice of trying to have stuff which only
changes occasionally on seperate partitions can help with this problem.
If you do it properly most of the stuff can be mounted read-only over
NFS with a small working disk (just a boot, tmp and swap drive).
> So right now I'm thinking of
> getting a 15.8 disk for the system (which is a waste) and then a 30 GB
> disk for each of the machines (for local use).
I'm not so sure what you are talking about here. Perhaps you could
expand slighty on what you mean.
>
> > I use a Matrox G400 on my workstation. If you're running
> >XFree86 4.0+, this is a good choice.
>
> That does sound decent. What about the ATI all-in-wonder cards? Would
> that be supported under Linux? I saw some websites mentioning
> "unofficial support" (by GATOS).
If these machines are in a cluster then only minimal graphics support
is needed. A simple 4MB PCI card is more than adequate. If you want it
for a workstation the Matrox cards are good, especially in a dual head
configuration. We use mainly nVidia cards because of their 3D price/
performance, but setting them up for 3D is not for the faint hearted.
For 2D it is easy though.
In general I suggest you stay away from the less supported cards.
Andrew
------------------------------
From: "V.G.Guhan" <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.linux.development.apps,comp.os.linux.development.system,comp.os.linux.setup,comp.os.linux.networking,comp.os.linux.x
Subject: Re: Linux solutions for VoIP and text editors
Date: Sat, 03 Feb 2001 10:47:10 -0800
So far I have not yet found any packages for VoIP clients for Linux. If we
are empty, I think the next step will be to build something. Use of JMF
maybe something to look into.
Baba
John Beardmore wrote:
> In article <[EMAIL PROTECTED]>, V.G.Guhan <[EMAIL PROTECTED]>
> writes
>
> >Also there are many VoIP clients ( Hearme.com) for Windows and I havent
> >come accross any for the Linux environment.
> >Anybody know of any stuff related to these things let me know about
> >where to find them.
> >If there isnt a solution then I think Ill get working on one.
> >Sincerely
>
> If there isn't any VOIP stuff out there I might be interested in joining
> in writing some.
>
> Cheers, J/.
> --
> John Beardmore
------------------------------
From: [EMAIL PROTECTED] (Dances With Crows)
Subject: Re: Heavy IDE use freezes screen and sound card
Date: 3 Feb 2001 16:00:00 GMT
Reply-To: [EMAIL PROTECTED]
On Sat, 03 Feb 2001 05:55:33 -0500, Flacco staggered into the Black Sun
and said:
>> hdparm -c1 -m16 -u1 /dev/hda should help. It's the u1 that will make
>> a lot of the difference. "man hdparm" to find out more.
>That seems to do it - where would you recommend I put this so it runs
>automatically when the system boots?
SuSE: /sbin/init.d/boot.local
RedHat/derived: /etc/rc.d/rc.local
--
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
Brainbench MVP for Linux Admin / Workin' in a code mine, hittin' Ctrl-Alt
http://www.brainbench.com / Workin' in a code mine, whoops!
=============================/ I hit a seg fault....
------------------------------
From: [EMAIL PROTECTED] (Dances With Crows)
Subject: Re: Have you seen my interrupt?
Date: 3 Feb 2001 16:00:02 GMT
Reply-To: [EMAIL PROTECTED]
On Sat, 03 Feb 2001 11:51:26 GMT, Sebastiano Lanza staggered into the
Black Sun and said:
>I have a goldstar eide cd GCD580B as slave on IDE0, when I attempt to
>read anything I get ' HDB: LOST INTERRUPT' and the system hangs
>(...GOSH!!!) It works under win98 (I read somewher win98 doesn't use
>interrupt with CD) I tried to pass 'ide0=serialize' to the kernel but
>didn't work I will offer a nice-smelling heaven-taste slice of
>napoletan pizza to anyone helps me ;)
Jumper problem? Most CD-ROMs come from the factory jumpered as "master"
and if you plug a master-jumpered drive into the slave position on the
cable, Linux will be unhappy. Failing that, is DMA enabled in the BIOS
for this CD-ROM? Many CD-ROMs incorrectly say they support DMA when
they don't, and DMA is unnecessary for a CD-ROM anyway. HTH,
--
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
Brainbench MVP for Linux Admin / Workin' in a code mine, hittin' Ctrl-Alt
http://www.brainbench.com / Workin' in a code mine, whoops!
=============================/ I hit a seg fault....
------------------------------
** 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.hardware.
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-Hardware Digest
******************************
