Segmentation faults when compiling kernel

Segmentation faults when compiling kernel

Post by Vlar Schreidlock » Wed, 13 Feb 2002 02:32:25



I keep getting errors like this:

/usr/src/linux-2.4.17/include/linux/vmalloc.h: In function
`vmalloc_dma':
/usr/src/linux-2.4.17/include/linux/vmalloc.h:45: Internal error:
Segmentation fault.
Please submit a full bug report.
See <URL:https://qa.mandrakesoft.com/> for instructions.
make[3]: *** [tty_ioctl.o] Error 2
make[3]: Leaving directory `/usr/src/linux-2.4.17/drivers/char'
make[2]: *** [first_rule] Error 2
make[2]: Leaving directory `/usr/src/linux-2.4.17/drivers/char'
make[1]: *** [_subdir_char] Error 2
make[1]: Leaving directory `/usr/src/linux-2.4.17/drivers'
make: *** [_dir_drivers] Error 2

when compiling a new kernel. They are different every time. I have
swapped out memory, processors, etc. Does anyone have any idea what
causes this and how to remedy it?

 
 
 

Segmentation faults when compiling kernel

Post by 4spac » Wed, 13 Feb 2002 05:40:40


I often get build errors unless I declock my front side bus and
processor from 266/1.13Gig to 200/850MHz. Worth a try?

Quote:> when compiling a new kernel. They are different every time. I have
> swapped out memory, processors, etc. Does anyone have any idea what
> causes this and how to remedy it?

--
Regards,

Tim Haughton
timhaughton<at>4space<dot>org<dot>uk

 
 
 

Segmentation faults when compiling kernel

Post by Vlar Schreidlock » Wed, 13 Feb 2002 06:33:17



>I often get build errors unless I declock my front side bus and
>processor from 266/1.13Gig to 200/850MHz. Worth a try?

That's interesting. This is on an AMD K6-2 500 box. What should I
"declock" it to?
Quote:>> when compiling a new kernel. They are different every time. I have
>> swapped out memory, processors, etc. Does anyone have any idea what
>> causes this and how to remedy it?

 
 
 

Segmentation faults when compiling kernel

Post by 4spac » Wed, 13 Feb 2002 06:46:29


Quote:> That's interesting. This is on an AMD K6-2 500 box. What should I
> "declock" it to?

Hard to say. Just check if you can alter your speed in the BIOS, and
knock it down a peg. Might work, might not ;)

--
Regards,

Tim Haughton
timhaughton<at>4space<dot>org<dot>uk

 
 
 

Segmentation faults when compiling kernel

Post by William Hun » Wed, 13 Feb 2002 11:00:44



> I keep getting errors like this:
> /usr/src/linux-2.4.17/include/linux/vmalloc.h: In function
> `vmalloc_dma':
> /usr/src/linux-2.4.17/include/linux/vmalloc.h:45: Internal error:
> Segmentation fault.
[...]
> when compiling a new kernel. They are different every time. I have
> swapped out memory, processors, etc. Does anyone have any idea what
> causes this and how to remedy it?

in my humble experience, segfault during kernel build is almost always
related to marginal hardware, particularly if the results vary with each
test.

i use kernel builds to pass/fail test overclocked hardware.
it's a very reliable indicator.

kernel builds thrash the bejabbers out of cpu, ram, and everything
inbetween.  if it builds a kernel, without contrary evidence, i call it
working.

So, you should be able to test this by de-tuning your system, using bios
settings or motherboard jumpers.  simplest test is to reduce front side bus
from 100MHz down to 66MHz, then compile a kernel.  if that fails, then
something (mobo,cpu,ram) is seriously bad.  otherwise,  raise fsb and
test again until you find lower limit of failure.  at that point, you can
diagnose further, if you want, by swapping parts and adjusting bios chipset
settings.

since you've already swapped cpu + ram, i suspect something in your bios
chipset settings might be the problem...  bios probably has a 'fail-safe'
option - try that first!  

hoping this helps,
send $10 or kill me.
--
William Hunt, Portland Oregon USA

 
 
 

Segmentation faults when compiling kernel

Post by Vlar Schreidlock » Thu, 14 Feb 2002 01:07:52





>> I keep getting errors like this:
>> /usr/src/linux-2.4.17/include/linux/vmalloc.h: In function
>> `vmalloc_dma':
>> /usr/src/linux-2.4.17/include/linux/vmalloc.h:45: Internal error:
>> Segmentation fault.
>[...]
>> when compiling a new kernel. They are different every time. I have
>> swapped out memory, processors, etc. Does anyone have any idea what
>> causes this and how to remedy it?

>in my humble experience, segfault during kernel build is almost always
>related to marginal hardware, particularly if the results vary with each
>test.

>i use kernel builds to pass/fail test overclocked hardware.
>it's a very reliable indicator.

>kernel builds thrash the bejabbers out of cpu, ram, and everything
>inbetween.  if it builds a kernel, without contrary evidence, i call it
>working.

>So, you should be able to test this by de-tuning your system, using bios
>settings or motherboard jumpers.  simplest test is to reduce front side bus
>from 100MHz down to 66MHz, then compile a kernel.  if that fails, then
>something (mobo,cpu,ram) is seriously bad.  otherwise,  raise fsb and
>test again until you find lower limit of failure.  at that point, you can
>diagnose further, if you want, by swapping parts and adjusting bios chipset
>settings.

>since you've already swapped cpu + ram, i suspect something in your bios
>chipset settings might be the problem...  bios probably has a 'fail-safe'
>option - try that first!  

>hoping this helps,
>send $10 or kill me.

I declocked an AMD K6-2 500 down to 338 and it seemed to work. That's
strange that neither a 500 mhz or a 450 mhz AMD K6-2 chip will compile
a kernel at its rated speed. I'm thinking about putting a 533mhz K6-2
chip in there and clock it down to 500. With all this talk about
overclocking it seems sad I have to underclock to compile anything.
 
 
 

Segmentation faults when compiling kernel

Post by William Hun » Fri, 15 Feb 2002 06:22:43





[...snip...]
>>So, you should be able to test this by de-tuning your system, using bios
>>settings or motherboard jumpers.  simplest test is to reduce front side
>>from 100MHz down to 66MHz, then compile a kernel.  if that fails, then
>>something (mobo,cpu,ram) is seriously bad.  otherwise,  raise fsb and
>>test again until you find lower limit of failure.  at that point, you can
>>diagnose further, if you want, by swapping parts and adjusting bios
>>chipset settings.
>>since you've already swapped cpu + ram, i suspect something in your bios
>>chipset settings might be the problem...  bios probably has a 'fail-safe'
>>option - try that first!

> I declocked an AMD K6-2 500 down to 338 and it seemed to work. That's
> strange that neither a 500 mhz or a 450 mhz AMD K6-2 chip will compile
> a kernel at its rated speed. I'm thinking about putting a 533mhz K6-2
> chip in there and clock it down to 500. With all this talk about
> overclocking it seems sad I have to underclock to compile anything.

you shouldn't "have to" declock.
detuning is only a diagnostic method.
detuning is a test to prove that the failure is hardware and not software.

if it fails, running at full rated speed, only when compiling a kernel, it
is because nothing else you normally do places those kinds of demands on
the hardware.   nothing much stresses a machine like a kernel compile does.
depending on your requirements, the system is either
broken or marginally adequate.

in my (limited) experience, these failures have almost always been due to
marginal ram, never due to a cpu running at full rated speed.

it might also be due to overoptimistic settings in chipset BIOS, or maybe
a marginal powersupply.  but i suspect marginal PC-100 ram ...

detune yer BIOS first, at full rated speed, before trying de-clocking !

after that, by trying different combinations of FSB and multiplier, you can
isolate the failure somewhat.  you reduced fsb from 100 to 66.  does your
mobo offer 95, 83, 75 too?  does it fail at all those speeds, too?   (if my
hunch is right, then it should work at fsb=95 and mult=5.0).

i've had good luck with amd K6.  I can't overclock them (or anything else)
very far, but i do find them reliable at their rated spec.

hoping this helps,
--
William Hunt, Portland Oregon USA

 
 
 

Segmentation faults when compiling kernel

Post by Vlar Schreidlock » Sat, 16 Feb 2002 00:54:56







>[...snip...]
>>>So, you should be able to test this by de-tuning your system, using bios
>>>settings or motherboard jumpers.  simplest test is to reduce front side
>>>from 100MHz down to 66MHz, then compile a kernel.  if that fails, then
>>>something (mobo,cpu,ram) is seriously bad.  otherwise,  raise fsb and
>>>test again until you find lower limit of failure.  at that point, you can
>>>diagnose further, if you want, by swapping parts and adjusting bios
>>>chipset settings.
>>>since you've already swapped cpu + ram, i suspect something in your bios
>>>chipset settings might be the problem...  bios probably has a 'fail-safe'
>>>option - try that first!

>> I declocked an AMD K6-2 500 down to 338 and it seemed to work. That's
>> strange that neither a 500 mhz or a 450 mhz AMD K6-2 chip will compile
>> a kernel at its rated speed. I'm thinking about putting a 533mhz K6-2
>> chip in there and clock it down to 500. With all this talk about
>> overclocking it seems sad I have to underclock to compile anything.

>you shouldn't "have to" declock.
>detuning is only a diagnostic method.
>detuning is a test to prove that the failure is hardware and not software.

>if it fails, running at full rated speed, only when compiling a kernel, it
>is because nothing else you normally do places those kinds of demands on
>the hardware.   nothing much stresses a machine like a kernel compile does.
>depending on your requirements, the system is either
>broken or marginally adequate.

>in my (limited) experience, these failures have almost always been due to
>marginal ram, never due to a cpu running at full rated speed.

>it might also be due to overoptimistic settings in chipset BIOS, or maybe
>a marginal powersupply.  but i suspect marginal PC-100 ram ...

>detune yer BIOS first, at full rated speed, before trying de-clocking !

>after that, by trying different combinations of FSB and multiplier, you can
>isolate the failure somewhat.  you reduced fsb from 100 to 66.  does your
>mobo offer 95, 83, 75 too?  does it fail at all those speeds, too?   (if my
>hunch is right, then it should work at fsb=95 and mult=5.0).

>i've had good luck with amd K6.  I can't overclock them (or anything else)
>very far, but i do find them reliable at their rated spec.

>hoping this helps,

Thanks. I had swapped out for several different PC-100 sticks. They
all should have been faster, newer models. Maybe my mobo is marginal.
I think it is some offbrand chinese thing from a while back. Actuall I
think it is a "PC-100" brand. I may try putting an AMD K6-2 550 in it
and running it at 500. I had put a 550 in a friend's computer that had
the same mobo (I think) and it wouldn't run at 550. Perhaps this is
the problem.
 
 
 

1. Compiling *** VIM 5.3 *** Segmentation Fault..what is Seg-Fault..MEM Bounds?

Compiled fine........upgraded from 4.6  REdhat 5.0

got the src and rt of the autors VIM site via ftp.vim.org

changed the first requested by auther line in ".vimrc" file to not
read "version 4.00" and it worked once.....Gee

Now I am using it on everything ...DOS Win95 even NT40 !!!!

I know go back to the early version but still.....
what or why ...

"Segmentation Fault...Core dump"
thanks in advance.....Matt

2. Linux 2.4.2-ac27

3. Segmentation fault while compiling kernel

4. Scan a Directory/cron question

5. Signal 11 (segmentation faults) compiling kernel?

6. How to transfer .eps to .gif

7. Segmentation fault in kernel compiling

8. file manipulation (easy question I think)

9. Segmentation fault while compiling kernel 1.3.37 and ppp2.2.0d

10. Segmentation Fault Compiling Kernel

11. help: i get a 'segmentation fault' when compiling kernel

12. Compile kernel gave segmentation fault.

13. segmentation fault after compiling new kernel