[linux-lvm] Segment Failure ,while doing lvcreate

[linux-lvm] Segment Failure ,while doing lvcreate

[David Chang]

Hi all:
My LVM version is 1.0.1-rc4 with RH 7.1,kernel 2.4.2-2
pvcreate is ok
vgcreate is ok 
but while doing lvcreate -L 20G puv_vg,then got error
messages "Segmentation fault(core dump).

how to fix it ?

-----------------------------------------------------------------
< �C�ѳ� Yahoo!�_�� >  www.yahoo.com.tw

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Thomas Svedberg]

tor 2001-10-25 klockan 13.45 skrev David Chang:
> Hi all:
> My LVM version is 1.0.1-rc4 with RH 7.1,kernel 2.4.2-2
> pvcreate is ok
> vgcreate is ok 
> but while doing lvcreate -L 20G puv_vg,then got error
> messages "Segmentation fault(core dump).
> 
> how to fix it ?

Seems RedHat compiler have problems with optimization turned on, try
compiling with -O0.

-- 
/ Thomas
.......................................................................
 Thomas Svedberg
 Department of Mathematics
 Chalmers University of Technology

 Address: S-412 96 Göteborg, SWEDEN
 E-mail : thsv@bigfoot.com, thsv@math.chalmers.se
 Phone  : +46 31 772 5368
 Fax    : +46 31 772 3595
.......................................................................

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Jason Edgecombe]

hi,

  recompile lvm tools with the -O0 flag (no optimization).
this is because of redhat's 7.x borken gcc compiler

I believe that this fix is in the archive.

Jason Edgecombe

David Chang wrote:
> 
> Hi all:
> My LVM version is 1.0.1-rc4 with RH 7.1,kernel 2.4.2-2
> pvcreate is ok
> vgcreate is ok
> but while doing lvcreate -L 20G puv_vg,then got error
> messages "Segmentation fault(core dump).
> 
> how to fix it ?
> 
> -----------------------------------------------------------------
> < ¨C¤Ñ³£ Yahoo!©_¼¯ >  www.yahoo.com.tw
> 
> _______________________________________________
> linux-lvm mailing list
> linux-lvm@sistina.com
> http://lists.sistina.com/mailman/listinfo/linux-lvm
> read the LVM HOW-TO at http://www.sistina.com/lvm/Pages/howto.html

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Eric M. Hopper]

[-- Attachment #1: Type: text/plain, Size: 997 bytes --]

On Thu, Oct 25, 2001 at 08:39:35AM -0400, Jason Edgecombe wrote:
> hi,
> 
>   recompile lvm tools with the -O0 flag (no optimization).
> this is because of redhat's 7.x borken gcc compiler
> 
> I believe that this fix is in the archive.

	Which version of their compiler are you using?  Is it
gcc-2.96-88, or gcc-2.96-96?  Is it even earlier?

	Does gcc 3.0.1 have this same problem?  Are you sure the bug
isn't in the code?  (i.e. having a non-volatile variable randomly
modified under your nose, or some such that would tend to not work so
well under optimization.)

	If gcc-2.96-96 doesn't have the problem, RH 7.2 should work just
fine.

Have fun (if at all possible),
-- 
"It does me no injury for my neighbor to say there are twenty gods or no God.
It neither picks my pocket nor breaks my leg."  --- Thomas Jefferson
"Go to Heaven for the climate, Hell for the company."  -- Mark Twain
-- Eric Hopper (hopper@omnifarious.org  http://www.omnifarious.org/~hopper) --

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Jason Edgecombe]

"Eric M. Hopper" wrote:
>         If gcc-2.96-96 doesn't have the problem, RH 7.2 should work just
> fine.

I suspect that the problem will be in RH 7.2 also.

 I was reading the roswell archive and someone from redhat said that all
7.x release will be binary compatible, i.e. they all have that buggy
gcc.

I may be wrong on this though.

Jaosn Edgecombe

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Eric M. Hopper]

[-- Attachment #1: Type: text/plain, Size: 1118 bytes --]

On Thu, Oct 25, 2001 at 08:49:18AM -0400, Jason Edgecombe wrote:
> "Eric M. Hopper" wrote:
> >         If gcc-2.96-96 doesn't have the problem, RH 7.2 should work just
> > fine.
> 
> I suspect that the problem will be in RH 7.2 also.
> 
>  I was reading the roswell archive and someone from redhat said that all
> 7.x release will be binary compatible, i.e. they all have that buggy
> gcc.
> 
> I may be wrong on this though.

	Just because it's 2.96 doesn't mean it's buggy.  Apparently
their latest version happens to have the same bug, but that didn't
necessarily have to be true.  I doubt the bug is an inherent 'feature'
of their calling conventions.

	I've actually had a lot of luck with gcc 2.96 fixing problems
for me.  But, I'm a C++ developer.  :-)

Have fun (if at all possible),
-- 
"It does me no injury for my neighbor to say there are twenty gods or no God.
It neither picks my pocket nor breaks my leg."  --- Thomas Jefferson
"Go to Heaven for the climate, Hell for the company."  -- Mark Twain
-- Eric Hopper (hopper@omnifarious.org  http://www.omnifarious.org/~hopper) --

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Markus Dobel]

On Thu, Oct 25, 2001 at 07:45:00AM -0500, Eric M. Hopper wrote:
> If gcc-2.96-96 doesn't have the problem, RH 7.2 should work just
> fine.

I still had the problem using gcc-2.96-99 on Redhat 7.1 (with
gcc-packages from rawhide).

Regards, Markus

-- 
Spiegel-Leser wissen mehr. Fuer ein Semester-Abo!
http://www.kawo2.rwth-aachen.de/~mdobel/semesterabo.html

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Eric M. Hopper]

[-- Attachment #1: Type: text/plain, Size: 919 bytes --]

On Thu, Oct 25, 2001 at 03:14:59PM +0200, Markus Dobel wrote:
> On Thu, Oct 25, 2001 at 07:45:00AM -0500, Eric M. Hopper wrote:
> > If gcc-2.96-96 doesn't have the problem, RH 7.2 should work just
> > fine.
> 
> I still had the problem using gcc-2.96-99 on Redhat 7.1 (with
> gcc-packages from rawhide).

	Thanks.  I run RH as well, and am going to be upgrading to 7.2
on my main server box, which runs LVM.  It's good to know these things.

	I'm actually pretty scared to upgrade since my current LVM is
fairly old (0.9.1 b7).

	Does anybody know if this happens with gcc 3.0.1 or 3.0.2?

Thanks,
-- 
"It does me no injury for my neighbor to say there are twenty gods or no God.
It neither picks my pocket nor breaks my leg."  --- Thomas Jefferson
"Go to Heaven for the climate, Hell for the company."  -- Mark Twain
-- Eric Hopper (hopper@omnifarious.org  http://www.omnifarious.org/~hopper) --

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: [linux-lvm] Segment Failure ,while doing lvcreate

[Sean Burford]

"Eric M. Hopper" wrote:
> On Thu, Oct 25, 2001 at 08:39:35AM -0400, Jason Edgecombe wrote:
> >   recompile lvm tools with the -O0 flag (no optimization).
> > this is because of redhat's 7.x borken gcc compiler
...
>         Does gcc 3.0.1 have this same problem?  Are you sure the bug
> isn't in the code?  (i.e. having a non-volatile variable randomly
> modified under your nose, or some such that would tend to not work so
> well under optimization.)

I've traced it in gdb, and I saw a function (called via an inline) being
given valid arguments but receiving junk arguments.  It looks like the
stack pointer got messed up (I didn't check, I was viewing source not
asm at the time).  Calling the function directly, rather than through
the inline (which only added an argument to the argument list) worked
fine.

Having your arguments corrupted during a function call sure looks like a
compiler bug to me.

-- 
Sean Burford    x34135
ITS Systems Specialist
Adelaide University

RE: [linux-lvm] Segment Failure ,while doing lvcreate

[Jean-Eric Cuendet]

I have this same problem.
Could you made a summary of the thread? Is it fixed? In RH7.2 compiler?
In LVM tools? Is -O0 sufficient to fix?
Another solution would be to use kgcc instead of gcc (kgcc is egcs
v1.1.2 or gcc 2.91 and used to build the kernel)

-jec


> -----Original Message-----
> From: Eric M. Hopper [mailto:hopper@omnifarious.org]
> Sent: vendredi, 26. octobre 2001 14:57
> To: linux-lvm@sistina.com
> Subject: Re: [linux-lvm] Segment Failure ,while doing lvcreate
> 
> 
> On Thu, Oct 25, 2001 at 08:49:18AM -0400, Jason Edgecombe wrote:
> > "Eric M. Hopper" wrote:
> > >         If gcc-2.96-96 doesn't have the problem, RH 7.2 
> should work just
> > > fine.
> > 
> > I suspect that the problem will be in RH 7.2 also.
> > 
> >  I was reading the roswell archive and someone from redhat 
> said that all
> > 7.x release will be binary compatible, i.e. they all have that buggy
> > gcc.
> > 
> > I may be wrong on this though.
> 
> 	Just because it's 2.96 doesn't mean it's buggy.  Apparently
> their latest version happens to have the same bug, but that didn't
> necessarily have to be true.  I doubt the bug is an inherent 'feature'
> of their calling conventions.
> 
> 	I've actually had a lot of luck with gcc 2.96 fixing problems
> for me.  But, I'm a C++ developer.  :-)
> 
> Have fun (if at all possible),
> -- 
> "It does me no injury for my neighbor to say there are twenty 
> gods or no God.
> It neither picks my pocket nor breaks my leg."  --- Thomas Jefferson
> "Go to Heaven for the climate, Hell for the company."  -- Mark Twain
> -- Eric Hopper (hopper@omnifarious.org  
http://www.omnifarious.org/~hopper) --