[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
gmp 4.1.2 on Solaris
From: |
Clemente Galdi |
Subject: |
gmp 4.1.2 on Solaris |
Date: |
Tue, 18 Mar 2003 18:30:18 +0200 (EET) |
Hi!
I am experiencing some problems with gmp 4.1.2 on Solaris.
I successfully compiled the library on the following systems:
SunOS eagle 5.7 Generic_106542-10 i86pc i386 i86pc
gcc 2.95
SunOS bearzot 5.7 Generic_106541-06 sun4u sparc SUNW,Ultra-250
gcc 2.95.2
SunOS cosmos 5.8 Generic_108528-15 sun4u sparc SUNW,Sun-Fire-280R
gcc 2.95.3
All the checks were successfull on all the systems.
I have a program that is currently running on Pentium 4/linux and on a
Xeon/linux system. No problem at all with these two installations.
I'm moving... or actually i'm trying to... this program to the Solaris systems
but all I got is a segmentation fault that is due to the gmp lib.
Here's an example of stack trace I get on the pc.
#0 0xdfa82432 in realfree () from /usr/lib/libc.so.1
#1 0xdfa8299a in cleanfree () from /usr/lib/libc.so.1
#2 0xdfa81fe2 in _malloc_unlocked () from /usr/lib/libc.so.1
#3 0xdfa81f19 in malloc () from /usr/lib/libc.so.1
#4 0xdfb27439 in __gmp_default_allocate (size=4) at memory.c:46
#5 0x80539b2 in XORAll (out=0x8047548 "\001\001", reseed=-541797328)
at selectors.c:60
Notice that it actually seems a "stack" problem... since in
"selectors.c:60" there is no call to gmp... but a call to a
function that calls a function that calls mpz_init... that generates the
segmentation fault. (between the first call and the call to mpz_init there
are only variables declarations and "if" !).
I've given a look at the bug-report repository and I tried to build the
library using all the possible paramenters for --enable-alloca
(including debug) but no success.
The program crash either on an mpz_init or on a mpz_mul
(depending only on the "enable-alloca" parameters used for building the
library).
I also set the stacksize limit to unlimited... that is
actually "130336 kbytes" on the pc and "unlimited" on the sparcstations...
whatever "unlimited" could mean !
datasize is "unlimited" on all the systems.
Can someone of you give me a hint on the possible causes of this problem
or a pointer to something to read.. or possibly a way to fix it :-)
Thank you very much.
--Clemente Galdi
Computer Technology Institute and
Deparment of Computer Engineering and Informatics
University of Patras
Greece
- gmp 4.1.2 on Solaris,
Clemente Galdi <=