Kernel 2.6.21.5, Slackware 12.0
Xfce 4.4.1
Window manager = ?

Hi:
During the last few days, I've noticed the cli is a bit unresponsive. Once in a while I type and the echo comes a second later. Today I had to kill X because the GUI hanged up at start up (I invoke with 'startx'). This, the GUI does once every a couple of months but as my system is not fine tuned, I take no notice.

But then, I ran K3b, the optical media authoring tool, and it crushed, a thing not uncommon either. One moment! Since last slack installation I've had no crushes of the GUI or any of its applications. Well, by this time, you will be thinking: "make a good installation and start afresh". But I have many compilations made and to repeat them all over again... linux programs I mean.

Well, then, the important thing because, finally, as the crown of all the preceding, a thing I'd never seen before: typing in a text console (/dev/ttyN) I was interrupted by the following notification:

INIT: PANIC: segmentation violation at 0xnn...nn (code)! sleeping for 30 seconds.

Machine is 32-bit. What do you think?

PD: after the notification every thing went normal until present moment.

First thing to do when you see "segmentation" is start running memtest on the machine

But before you reboot the machine for that copy dmesg output to pastbin and post it here there may be warnings of something failing in the output of dmesg that may indicate something else as the problem.

Also Slackware 12.0? Is the machine really old as well and has it been kept clean, dirt and dust on the motherboard and in connectors can cause issues in the long run.

A quick search of the web reveals the message is from the init command. The exact text is in the init.c source code, which is part of the a/sysvinit package.

Searching the web for "What is a segmentation violation" revealed the following:

http://www.novell.com/support/viewCo...1662&sliceId=1

Some of the common causes listed:

Programming error
Mismatched binary and libraries
Hardware or hardware configuration issue

Hardware failures are suspect but so is tinkering.

If the hardware is old then suspecting hardware is reasonable. With respect to troubleshooting, I agree that running a memory test is a good start. Memtest86 is the usual choice and can be run from many live CDs. You can copy memtest86 to your /boot file system and create a boot loader option to run memtest from your hard drive.

As mentioned, a good system cleaning often helps. Remove all dust. Pulling and reseating several times all cable connectors is a good idea. The same with the memory sticks. Pulling and reseating the memory sticks will help clean the contacts. Another trick is to use a simple pencil eraser to wipe the memory stick contacts. Be sure to remove all eraser shreddings before inserting the sticks in the computer.

Another common suspect with old hardware is the power supply.

With that said, the tone of your query indicates the problems are recent. Notice some of the causes of this error message are software related, specifically mismatched software --- tinkering.

Some of the queries I read online related to this error message were related with people updating systems --- mixing new packages with old.

Have you done anything recently that could affect your system such as recompiling certain packages or changing any hardware? Have you updated any packages from more recent Slackware releases? That kind of thing could create problems.

Are you running any new software? Different video card? Updated any software drivers?

Did you recently modify something in the system startup scripts that is causing some new processes to run?

Did you modify any BIOS settings that are related to timing issues, such as RAM settings?

Generally, when a system is working fine one day but not the next a good question to ask is what did the user/owner do? Other than hardware failures, oddball failures often are the result of something the user/owner did. Been there, done that many times. When something works for me one day but not the next I always ask myself, "What did I do?"

The point is: try thinking methodically backwards to when you were not seeing these problems. Although the problem could be hardware, tracing your own footsteps might reveal some clues.

Just a comment. You have your own reasons for running 12.0 and I don't question them. Yet if the only reason is old hardware, then you might be interested in knowing I am running 12.2 on a Pentium I and Pentium II machine. I mention that only to let you know that generally, a more recent Slackware release will run okay on old hardware.

Hi:

Back again, after the crash, but from another machine.

The crashed one has 256MB RAM and Celeron @1100Hz. By this, you can infer the age.

I rebooted from LFS live CD. And again INIT: PANIC. I took note of the messages:

rc.S: line 14: [: bad array subscript

/usr/sbin/syslogd /sbin/ldconfig: /usr/lib/libwraster.so.3 is for unkown machine 3

And two more messages. I seem not have testmem or testmem6. I looked with grep at /var/log/packages/.

The nearest thing is the dmidecode package but dmi gives only archithectural type info.

/var/log/dmesg I attach it here. The link is dead (please remember I'm on another machine). I'll try to send it later.

The main note, I think, is that afterwards booting WinXP, it crashed too when forcing a FS check. Hardware failure then? Tinkering can be left aside cause I barely manage to fine tune my linux box.

As to your suggestions about cleaning and connection reseating, the memory modules seat in DIMM SDRAM sockets (RAM type: SIMM). I have seen that pluggin in the module may leave in a unstable mechannical situation. I mean, its not a good plugging mechanism, or mine is false. I try to mount them, exerting force until I here a click. But the click on this motherboard I can here it in one end but not in the other... Well, excuse the imprecision. What I want to say that it is hard for me to say if the modules make good contact. All of this I'll make once again.

Surely to know if there are utilities to test memory in slack <= 12.0 would be a great thing.
Well, first, I took away the RAM modules and inserted a single 64MB one. Linux booted and normaly operated but, trying to run the GUI, of course, failed. I reinserted the 256MB.

Secondly, I compiled and installed some non slackware packages, but all seemed to go well.
I've run xcircuit, a GUI program, a drawing appl, which outputs postscript. No new video card. No driver update.
I edited /etc/rc.d/rc.local only to add an alias.
Yes, that was about three days ago, and I used the machine heavily up to now: I raised CPU speed up to 1100Hz, the rated processor speed, and I think the RAM speed to as a consequence. Namely, I raised the multiplier factor to x11, and the base freq to 100/100. But I do not remember the old value for the RAM.

Never mind, I have now, in the BIOS menu, pressed the F7 key (Load BIOS defaults) and it has no effect.
So, if I tell you my RAM is PC133 and my processor max speed 1100MHz (that specified by Intel), would you give me values for
(a) base freq
(b) multiplier factor
To set the CPU at 1100 I have no problem althoug the machine came to my hands with CPU freq= 735MHz.

For example base freq 100/100MHz, x11. But what about the RAM? Does PC133 marked on the module mean the external (base) bus freq should be 133MHz, or that it is clocked at 133Mhz? For in this case, 100/100 x11 gives 1100MHz for the RAM, absurd. I cant ask you to give me the right figures because you do not know how my BIOS is, although the motherboard is PCCHIPS M758LT-H.

One thing I'll do is to short the CMOS battery contatcs to force the BIOS to take conservative values.

Your final remark: well, I am running on Pentium II and have Slack 12.0, more or less your scenario. Thank a lot for all your tips and comments. I have still some cards under my sleeve. Regards.

Hi again: two, and the only things I touched in the machine in the last few days are: I enabled SMART (HDD) in the BIOS. Up to that time it always ran with SMART disabled, although the hard is SMART capable.

The other one, also in the BIOS setup, I raised the Celeron (Pentium II) 1100MHz CPU frequency from 733 up to 1100MHz. The RAM base frequency I did not modified.

If this may serve as a diagnostics element to somebody, I'll be glad. At present, I'm with base freq= 66/100, multiplier factor = 11. I the crash had anything to do with these BIOS settings I won't assume. Bye.

oh oh

The PII never came that fast no where near it in fact.

The PIII does however come in speeds like that and has issues

http://www.tomshardware.com/reviews/...iii-1,235.html

Thanks for your reply. From dmidecode output: I got confused because it says 'Family: Pentium II' and, below, Version: Pentium III Xeon. I thought dmidecode was wrong.

Now I have the PCCHIPS M758 family mainboards manual and it says it supports Celeron, Pentium III or Tualatin/Tualatin Celeron. Sorry for the mistake, if it is one.

One thing I would like to know, is where do some programs gather the info that I have a 1100MHz CPU. Does the processor tell, does the BIOS, who says?

Seems like overclocking the CPU is a prime suspect. I recommend resetting the CPU to run at design speed and then report back in a few days, unless breakage continues immediately thereafter. I empathize that 300 MHz is comparatively slow these days, but that is what a PII runs at. As I mentioned, I have a spare PI and PII machine in my LAN. I use them only for tinkering but when I play with them I accept the slow speeds.

Hi:

I'm almost sure 300MHz is the freq the BIOS is setting itself at, each time I turn off the machine, contrary to several years of running at 733MHz.

I had before shorten the CMOS battery terminals, in order to bring the machine back to life (it was dead).

ALL THIS IS AN ANECDOTE, for now I have stripped the machine to its bare bones. I have took fan and cooler off, and I'm reading the labels stamped on the processor top surface. These are:

Intel M R '01 MALAY
1100A/256/100/1.475
Q231A653-0463
CELERON SL5ZE

On the socket, the label PGA370.

Excet the serial number, the other data I think is useful to know if I have the correct data sheet which, among other things, will specify the front-side bus and CPU speeds.

Then, according to the data given by intel, I'll set the BIOS. And let you know. Bye.

Hi

You shouldn't trust DMI data are they are set manually, and they are usually not updated...
You'd better use the output of and as it comes from the kernel's hardware probing.

Read :

• CPU freq = 1100 MHz
• L2 cache = 256 Kb
• FSB freq = 100 MHz
• Core voltage = 1.475 V

Your CPU is a "Tualatin" Celeron :
http://www.cpu-world.com/sspec/SL/SL5ZE.html

1 members found this post helpful.

Thanks for the information and for the link. I have now in my power the motherboard user manual (I had it before I dismounted the micro) and I intend to run the CPUID instruction, although the quantity of info it yields now must be enormous. However, I think the instruction gives only architectural information.

But, above all, the datasheet will teach me once and for all how this thing of the frequencies is. Regards.

The 'base' freq will be FSB/RAM speed.

RAM is not given the multi- so when you are running at 66MHz FSB, the RAM normally runs at 66MHz as well. (There are some boards that can set RAM speed above FSB speed, not exactly common). So even though you are running a 1100MHz CPU, your RAM is still running at 100MHz hen you are set to 100/100.

PC 133 can run up to 133MHz, but will also run happily at 66MHz, 100MHz...pretty much anything lower than 133MHz.


The BIOS passes the CPU speed, etc. to the OS.

IIRC Pentium III and celerons based on P3 had a minimum multi of x3. You could have a board thats just getting the multi wrong due to CPU damage/crappy BIOS/etc giving you 100 x 3 = 300MHz.

Or it could be just a sick/dying board. I'd be more likely to guess its a'bad' board, PC-chips made some of the worst boards ever, and if you had to cl;ear the CMOS thats not a good sign...

I'd proably try updating the BIOS.

*edit- if your baord has issues running aty 100MHz+, that would explain why you got it clocked to 733MHz.

I now have some precise information, although not all I need.

(a) The micro is an Intel Celeron 1100A/256/100/1.475, bearing the mark SL5ZE. This makes it a Tualatin. But Tualatin or not, it will only run, or is intended to run at FSB= 100MHz and internal clock= 1100MHz. These data taken by visual inspection of the IC inserted in the PGA370 socket.

(b) The Intel datasheet. Its title is "Intel Celeron Processor for PGA370 up to 1.40GHz on 0.13u Process". From table 16 I get:

Processor: 1.1A
Core Frequency: 1.1A GHz(sic)
BCLK Frequency: 100MHz
Frequency Multiplier: 11

Again FSB = 100MHz. In fact, it is 100MHz in all the table entries, for Processor= 900, 1A, ..., 1.4GHz. Completely new to me is the fact that, for each of these frecuencies there is a different processor model.

(c) BIOS booting message in screen line 25 declaring this to be an M758LT-H board. And its only four options: 66/100, 100/100, 133/133, 133/100MHz.

(d) dmidecode saying the base board the manufacturer is PCCHIPS and the product name, M758L7 (the BIOS contradicts itself, 7 is the version number).

(e) The RAM modules are PC133 (Single Data Rate SDRAM, aka SDRAM, clock freq = 133MHz).

(f) The PCCHIPS, M758 Series, V7.0 user's guide for models M758L/M758LT/M758T/M758 of which, mine is M758LT (M758LT-H) if I trust the BIOS boot message, and I take dmidecode giving the version number). It says M758LT supports Tualatin CPU, and the board has two T's by the side of the processor and a blue IDE2 slot indicating support for Tualatin, besides.

Says the user's guide:

FCPGA Celeron: 300-966MHz, FSB: 66MHz
FCPGA Pentium III: 500-1130MHz, FSB: 100MHz, 133MHz
Tualatin/Tualatin Celeron: up to 1.4GHz, FSB: 100MHz (only for model M758LT and M758T)

About the system RAM, it says: "If the installed CPU uses a 100MHz system bus, you must use PC100 or PC133 memory...".


From among all the preceding, the four BIOS options are limited to only three: 66/100, 100/100, 133/100MHz and an x11 multiplier would give exactly the CPU rated freq: 1100MHz.

I'm yet in the informational stage, as you can see. My main problem now is: How do I know the Intel datasheet described matches the ID SL5ZE on the microprocessor package? I simply searched the web with 'SL5ZE' and got the Intel datasheets. And how do I relate the datasheets with the ordered set (faulty English) (family, model, stepping)? Perhaps getting the Intel hardware reference or the architecture reference?

Finally, I'm running the machine at 66/100, x11. And the BIOS displayed CPU freq is 733MHz. But 66 times 11 = 726. What's this?

Moved: This thread is more suitable in the Hardware forum and has been moved accordingly to help your thread/question get the exposure it deserves.

Of course there are different processor models for different speeds...intels hardly going to just call them 'celeron' and sell 900MHz for the same price as 1100MHz.

The M758LT-H might be the same 'family' as the M758*** models, but I wouldnt assume that because some similar models say 'these features show tualatin support' does not mean that your board does support tualatin.

I think it does, but when you have a manufacturer who does a lot of 'OEM' work (like PCchips) you can never be sure unless you can find real info on the board.

Since PCchips doent list the M758LT-H it could be a 'special' 'cutdown' version made for non-western countries, or it could be an OEM board made by PCchips for somebody else.

Family, model, stepping mean virtually nothing anymore.

You shouldne even need the data sheet, let alone going to the intel site. The CPU info on the top of the chip shows you pretty much all you need to know- voltage, FSB, cache, series.

66.6 x 11 = 732.6MHz
66.66 x 11 = 733.26MHz

It (should) be running at 66.6/66.66MHz not 66MHz FSB. Most BIOSes just list 66 rather than 66.6/66.66.

1 members found this post helpful.