| View previous topic :: View next topic |
| Author |
Message |
nesapio
n00b
Joined: 27 Jul 2004
Posts: 11
|
 Posted: Wed Jul 28, 2004 3:47 pm Post subject: error while make dep && make bzImage modules modules |
 |
|
hi i am trying to install a stage3 tarball of 2004.2.
everything worked fine until i typed the command:
| Code: | make dep && make bzImage modules modules_install
|
after a while i get the error message:
sysctl.c :In function 'parse_table'
sysctl.c:691:internal compiler error: Segmentation fault
please report bug....
the bug is not reproducible, so it is likely a hardware or OS problem.
make[2]: *** [sysctl.o]Error 1
.....
please help me i am trying to set up my system for three days now.  |
|
| Back to top |
|
 |
agent_jdh
Veteran
Joined: 08 Aug 2002
Posts: 1783
Location: Scotland
|
| Posted: Wed Jul 28, 2004 4:00 pm Post subject: |
|
|
You're right, it's most likely a hardware issue - are you overclocking, is your CPU adequately cooled, or are you using particularly aggressive memory timing options in bios?
_________________
Jingle Jangle Jewellery |
|
| Back to top |
|
|
boothepa
n00b
Joined: 20 May 2004
Posts: 49
Location: Austin, TX
|
| Posted: Wed Jul 28, 2004 4:08 pm Post subject: |
|
|
In my experience it's fairly common for a compilation to fail with a segfault; it tends to work if you just keep retrying. Otherwise, you might try playing with your CFLAGS and make flags.
_________________
Um...Hook 'em horns? |
|
| Back to top |
|
|
agent_jdh
Veteran
Joined: 08 Aug 2002
Posts: 1783
Location: Scotland
|
| Posted: Wed Jul 28, 2004 4:47 pm Post subject: |
|
|
| boothepa wrote: | | In my experience it's fairly common for a compilation to fail with a segfault; it tends to work if you just keep retrying. Otherwise, you might try playing with your CFLAGS and make flags. |
What? You're serious? You've got a problem if you get internal compiler segfaults. Remember a big compile puts a lot of strain on your CPU which will be very sensitive to any borderline instability your system has, which is invariably caused by defective hardware, overclocking, aggressive memory timings, poor cooling etc etc.
Better to address the cause of the problem.
_________________
Jingle Jangle Jewellery |
|
| Back to top |
|
|
boothepa
n00b
Joined: 20 May 2004
Posts: 49
Location: Austin, TX
|
| Posted: Wed Jul 28, 2004 5:52 pm Post subject: |
|
|
| agent_jdh wrote: | | Better to address the cause of the problem. |
True, unless these problems are out of your hands. No system is perfect. If you are overclocking or fudging your memory speeds/voltages or something similar, you should definitely look into fixing that. If you're running everything stock with standard drivers, there's not much to be done short of auditing all the code you're running including the kernel and device drivers. That's not realistic for your average linux user. Even if you did do that the problem isn't even guaranteed to be software-related. Hardware is good but by no means perfect. Regardless of the configuration there are guaranteed to be the occasional corner cases that the hardware manufacturers didn't anticipate or couldn't be expected to account for. Now, depending on your manufacturer/configuration the number of corner cases is likely to be extremely low, but there will always exist hardware bugs. (Remember the pentium divide error?)
_________________
Um...Hook 'em horns? |
|
| Back to top |
|
|
agent_jdh
Veteran
Joined: 08 Aug 2002
Posts: 1783
Location: Scotland
|
| Posted: Wed Jul 28, 2004 11:01 pm Post subject: |
|
|
| boothepa wrote: | | True, unless these problems are out of your hands. No system is perfect. If you are overclocking or fudging your memory speeds/voltages or something similar, you should definitely look into fixing that. If you're running everything stock with standard drivers, there's not much to be done short of auditing all the code you're running including the kernel and device drivers. That's not realistic for your average linux user. Even if you did do that the problem isn't even guaranteed to be software-related. Hardware is good but by no means perfect. Regardless of the configuration there are guaranteed to be the occasional corner cases that the hardware manufacturers didn't anticipate or couldn't be expected to account for. Now, depending on your manufacturer/configuration the number of corner cases is likely to be extremely low, but there will always exist hardware bugs. (Remember the pentium divide error?) |
Well if you're unlucky enough to basically have a defective system that gives you internal compiler segfaults due to a hardware problem then you really should be contacting the manufacturer re getting it fixed. I mean if it is a genuine hardware bug then it will obviously have to be worked around in software and in that case the original poster should be posting on the lkml with his problem (since it's a stage 3 install there is no indication of whether other programs would fail to compile, which would indicate either a gcc bug or would back up the hardware issue theory).
Regarding "auditing all the code", well, what do you mean, all you need to know is where to report problems to and people who DO know how to get it fixed will hopefully fix it. So the average Linux user doesn't have to audit everything or know how to code.
The original poster's first port of call should be checking his hardware and bios settings. Then proceed from there if that does not fix the problem.
Your statement that internal compiler segfaults are fairly common may well be the case for you, but not for the rest of us, and if you're getting them then you've got a problem too. It certainly isn't reasonable advice to give to someone - they shouldn't be happening, period.
_________________
Jingle Jangle Jewellery |
|
| Back to top |
|
|
boothepa
n00b
Joined: 20 May 2004
Posts: 49
Location: Austin, TX
|
| Posted: Wed Jul 28, 2004 11:33 pm Post subject: |
|
|
I agree, gcc segfaults shouldn't happen. They have happened to me from time to time (a few times -- twice? three times? -- a year) in a fashion that are non-reproducable and little more than an occasional nuisance. I tend to ignore them, writing them off to subtle gcc bugs. (Bearing in mind how old/useful/complex gcc is, not only would I expect a few bugs but I'm fully willing to forgive them.  ) Whether or not I should ignore those phantom segfaults is (clearly ) open to debate, but I do. Reproduceable segfaults, however, should be dealt with appropriately; I agree 100%. If this particular segfault doesn't happen again, then I suggest ignoring it until it becomes more of an issue. If it only happens once then clearly it isn't a serious issue, I don't care what your definition is . If gcc segfaults happen consistently and do become an issue, though, then I agree that reflects a deeper issue that should resolved. Until such a time that it does becomes a serious problem, though, there's no need to start a snipe hunt.
Otherwise, gcc segfaults are often caused by insufficient RAM. (Experience of a friend -- 32MB laptop.) How much RAM do you have?
_________________
Um...Hook 'em horns? |
|
| Back to top |
|
|
nesapio
n00b
Joined: 27 Jul 2004
Posts: 11
|
| Posted: Thu Jul 29, 2004 8:34 am Post subject: |
|
|
| i have got a 450 mhz amd k6-2 and 320mb ram. i didnt overclock and i didnt changed any properties in bios, the pc worked ok for years now, and i retried the command, but i get an error everytime, its always an segmentation fault error, but always in another file(last time it was string.h). |
|
| Back to top |
|
|
nesapio
n00b
Joined: 27 Jul 2004
Posts: 11
|
| Posted: Thu Jul 29, 2004 11:08 am Post subject: |
|
|
i now tried
but it seems my ethernet settings have changed, i tried ifconfig and the ip is not the one i gave it while installing, but i have no access to net-setup now.
how can i change my ethernet settings without using net-setup ?? |
|
| Back to top |
|
|
agent_jdh
Veteran
Joined: 08 Aug 2002
Posts: 1783
Location: Scotland
|
| Posted: Thu Jul 29, 2004 11:41 am Post subject: |
|
|
| nesapio wrote: | | i have got a 450 mhz amd k6-2 and 320mb ram. i didnt overclock and i didnt changed any properties in bios, the pc worked ok for years now, and i retried the command, but i get an error everytime, its always an segmentation fault error, but always in another file(last time it was string.h). |
The fact it's happening repeatedly now with different files definitely indicates a hardware problem.
Even if the PC has been running fine for years, it could be the strain of compiling the kernel is exposing a flaw that has always been there.
Things to try-
1. Relax the memory timings in your bios
2. Try underclocking your machine, even temporarily
3. Open the case, switch the PC on and check that the CPU fan is spinning.
4. While the box is open, use a can of compressed air and a soft small brush to clear the dust especially from the CPU heatsink - if the machine has been unopened for years, chances are it will be prettu dirty inside. Also compressed air through the power supply grilles can help clear dust from inside the PSU.
_________________
Jingle Jangle Jewellery |
|
| Back to top |
|
|
agent_jdh
Veteran
Joined: 08 Aug 2002
Posts: 1783
Location: Scotland
|
| Posted: Thu Jul 29, 2004 11:43 am Post subject: |
|
|
| nesapio wrote: | | how can i change my ethernet settings without using net-setup ?? |
Manually edit the file
Then restart net.eth0
| Code: | | /etc/init.d/net.eth0 restart |
_________________
Jingle Jangle Jewellery |
|
| Back to top |
|
|
nesapio
n00b
Joined: 27 Jul 2004
Posts: 11
|
| Posted: Thu Jul 29, 2004 11:59 am Post subject: |
|
|
| thanks a lot will try that. |
|
| Back to top |
|
|
|
Installing Gentoo
