Segmentation Fault: new char[12]

Segmentation Fault: new char[12]

[atharaken]

This is a snippet of some code I am using. I get a Segmentation Fault
in _smalloc (malloc). It works sometimes and when a large number of
users are accessing it gives a segmentation fault.

It happens at the line:

_string = new char [_capacity];

I call this using _Init(12);
This is a protected class.


/* BEGIN CODE HERE */

bool MyClass::_Init (long size)
{
    _capacity = size;
    _string = new char [_capacity]; //Get Segmentation Fault at _smalloc
    if (!_string) {
        _capacity = _size = 0L;
        return false;
    }
    _size = 0;
    _string[0] = '\0';
    return true;
}

/*END CODE HERE */


Machine: i386
OS: Solaris 7
GCC 2.95 (Release)

Any help is appreciated

Thanks


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[hwidjaja]

I didn't see problem at your code..
you mentioned that i gives a segm. fault when large number of users
are accessing it ...
I suspect that there is memory leaks in your code. Then you run out of
memory...
my suggestion is to check it,
 you can use 'top -U<username>' or 'ps'.

hwidjaja
In article < 82jp3a$pal$1@nnrp1.deja.com >,
  atharaken@my-deja.com wrote:
> This is a snippet of some code I am using. I get a Segmentation Fault
> in _smalloc (malloc). It works sometimes and when a large number of
> users are accessing it gives a segmentation fault.
>
> It happens at the line:
>
> _string = new char [_capacity];
>
> I call this using _Init(12);
> This is a protected class.
>
> /* BEGIN CODE HERE */
>
> bool MyClass::_Init (long size)
> {
>     _capacity = size;
>     _string = new char [_capacity]; //Get Segmentation Fault at
_smalloc
>     if (!_string) {
>         _capacity = _size = 0L;
>         return false;
>     }
>     _size = 0;
>     _string[0] = '\0';
>     return true;
> }
>
> /*END CODE HERE */
>
> Machine: i386
> OS: Solaris 7
> GCC 2.95 (Release)
>
> Any help is appreciated
>
> Thanks
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[atharaken]

Question:

I use top to check the memory use. It does have a memory leak which I
need to fix but does Solaris have a limit on the amount of memory that
a application uses. It yes, how can that be changed.

This computer has around 2GB of RAM and this is the only app running on
this machine apart from other system daemons etc..

In article < 82kesh$994$1@nnrp1.deja.com >,
  hwidjaja@my-deja.com wrote:
> I didn't see problem at your code..
> you mentioned that i gives a segm. fault when large number of users
> are accessing it ...
> I suspect that there is memory leaks in your code. Then you run out of
> memory...
> my suggestion is to check it,
>  you can use 'top -U<username>' or 'ps'.
>
> hwidjaja
> In article < 82jp3a$pal$1@nnrp1.deja.com >,
>   atharaken@my-deja.com wrote:
> > This is a snippet of some code I am using. I get a Segmentation
Fault
> > in _smalloc (malloc). It works sometimes and when a large number of
> > users are accessing it gives a segmentation fault.
> >
> > It happens at the line:
> >
> > _string = new char [_capacity];
> >
> > I call this using _Init(12);
> > This is a protected class.
> >
> > /* BEGIN CODE HERE */
> >
> > bool MyClass::_Init (long size)
> > {
> >     _capacity = size;
> >     _string = new char [_capacity]; //Get Segmentation Fault at
> _smalloc
> >     if (!_string) {
> >         _capacity = _size = 0L;
> >         return false;
> >     }
> >     _size = 0;
> >     _string[0] = '\0';
> >     return true;
> > }
> >
> > /*END CODE HERE */
> >
> > Machine: i386
> > OS: Solaris 7
> > GCC 2.95 (Release)
> >
> > Any help is appreciated
> >
> > Thanks
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[atharaken]

Any free tools to detect memory leaks!!! Purify and Insure both cost a
lot. Any others!

In article < 82kesh$994$1@nnrp1.deja.com >,
  hwidjaja@my-deja.com wrote:
> I didn't see problem at your code..
> you mentioned that i gives a segm. fault when large number of users
> are accessing it ...
> I suspect that there is memory leaks in your code. Then you run out of
> memory...
> my suggestion is to check it,
>  you can use 'top -U<username>' or 'ps'.
>
> hwidjaja
> In article < 82jp3a$pal$1@nnrp1.deja.com >,
>   atharaken@my-deja.com wrote:
> > This is a snippet of some code I am using. I get a Segmentation
Fault
> > in _smalloc (malloc). It works sometimes and when a large number of
> > users are accessing it gives a segmentation fault.
> >
> > It happens at the line:
> >
> > _string = new char [_capacity];
> >
> > I call this using _Init(12);
> > This is a protected class.
> >
> > /* BEGIN CODE HERE */
> >
> > bool MyClass::_Init (long size)
> > {
> >     _capacity = size;
> >     _string = new char [_capacity]; //Get Segmentation Fault at
> _smalloc
> >     if (!_string) {
> >         _capacity = _size = 0L;
> >         return false;
> >     }
> >     _size = 0;
> >     _string[0] = '\0';
> >     return true;
> > }
> >
> > /*END CODE HERE */
> >
> > Machine: i386
> > OS: Solaris 7
> > GCC 2.95 (Release)
> >
> > Any help is appreciated
> >
> > Thanks
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[Mike Albaugh]

atharaken@my-deja.com wrote:
: Any free tools to detect memory leaks!!! Purify and Insure both cost a
: lot. Any others!

	I don't think the "memory leak" is your biggest problem. The
visible result of a memory leak is that malloc() returns 0 after some
amount of time, but your code checks for that. Getting a _SEGFAULT_
from malloc() (or free(), or whatever) is almost always caused by
your application corrupting the malloc "headers" that are often
stored adjacent to the allocated memory. You need to be looking for
array-bounds violations, not memory-leaks.

					Mike
| albaugh@agames.com, speaking only for myself

Re: Segmentation Fault: new char[12]

[Josh Gagliardi]

>         I don't think the "memory leak" is your biggest problem. The
> visible result of a memory leak is that malloc() returns 0 after some
> amount of time, but your code checks for that. Getting a _SEGFAULT_
> from malloc() (or free(), or whatever) is almost always caused by
> your application corrupting the malloc "headers" that are often
> stored adjacent to the allocated memory. You need to be looking for
> array-bounds violations, not memory-leaks.

Freeing something twice can also do bad things to the free list and
cause a later malloc or free to SEGFAULT.

Re: Segmentation Fault: new char[12]

[hwidjaja]

you can adjust the limit of your resource limit using
setrlimit. Try man pages of getrlimit or setrlimit.
to check the limit at the prompt, uses 'ulimit -a' (for ksh,
zsh). i'm not sure about the other shells.

hwidjaja
In article < 82lqpm$7fo$1@nnrp1.deja.com >,
  atharaken@my-deja.com wrote:
> Question:
>
> I use top to check the memory use. It does have a memory leak which I
> need to fix but does Solaris have a limit on the amount of memory that
> a application uses. It yes, how can that be changed.
>
> This computer has around 2GB of RAM and this is the only app running
on
> this machine apart from other system daemons etc..
>
> In article < 82kesh$994$1@nnrp1.deja.com >,
>   hwidjaja@my-deja.com wrote:
> > I didn't see problem at your code..
> > you mentioned that i gives a segm. fault when large number of users
> > are accessing it ...
> > I suspect that there is memory leaks in your code. Then you run out
of
> > memory...
> > my suggestion is to check it,
> >  you can use 'top -U<username>' or 'ps'.
> >
> > hwidjaja
> > In article < 82jp3a$pal$1@nnrp1.deja.com >,
> >   atharaken@my-deja.com wrote:
> > > This is a snippet of some code I am using. I get a Segmentation
> Fault
> > > in _smalloc (malloc). It works sometimes and when a large number
of
> > > users are accessing it gives a segmentation fault.
> > >
> > > It happens at the line:
> > >
> > > _string = new char [_capacity];
> > >
> > > I call this using _Init(12);
> > > This is a protected class.
> > >
> > > /* BEGIN CODE HERE */
> > >
> > > bool MyClass::_Init (long size)
> > > {
> > >     _capacity = size;
> > >     _string = new char [_capacity]; //Get Segmentation Fault at
> > _smalloc
> > >     if (!_string) {
> > >         _capacity = _size = 0L;
> > >         return false;
> > >     }
> > >     _size = 0;
> > >     _string[0] = '\0';
> > >     return true;
> > > }
> > >
> > > /*END CODE HERE */
> > >
> > > Machine: i386
> > > OS: Solaris 7
> > > GCC 2.95 (Release)
> > >
> > > Any help is appreciated
> > >
> > > Thanks
> > >
> > > Sent via Deja.com http://www.deja.com/
> > > Before you buy.
> > >
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[Mike Albaugh]

atharaken@my-deja.com wrote:
: Any free tools to detect memory leaks!!! Purify and Insure both cost a
: lot. Any others!

	I don't think the "memory leak" is your biggest problem. The
visible result of a memory leak is that malloc() returns 0 after some
amount of time, but your code checks for that. Getting a _SEGFAULT_
from malloc() (or free(), or whatever) is almost always caused by
your application corrupting the malloc "headers" that are often
stored adjacent to the allocated memory. You need to be looking for
array-bounds violations, not memory-leaks.

					Mike
| albaugh@agames.com, speaking only for myself

Re: Segmentation Fault: new char[12]

[hwidjaja]

you can adjust the limit of your resource limit using
setrlimit. Try man pages of getrlimit or setrlimit.
to check the limit at the prompt, uses 'ulimit -a' (for ksh,
zsh). i'm not sure about the other shells.

hwidjaja
In article < 82lqpm$7fo$1@nnrp1.deja.com >,
  atharaken@my-deja.com wrote:
> Question:
>
> I use top to check the memory use. It does have a memory leak which I
> need to fix but does Solaris have a limit on the amount of memory that
> a application uses. It yes, how can that be changed.
>
> This computer has around 2GB of RAM and this is the only app running
on
> this machine apart from other system daemons etc..
>
> In article < 82kesh$994$1@nnrp1.deja.com >,
>   hwidjaja@my-deja.com wrote:
> > I didn't see problem at your code..
> > you mentioned that i gives a segm. fault when large number of users
> > are accessing it ...
> > I suspect that there is memory leaks in your code. Then you run out
of
> > memory...
> > my suggestion is to check it,
> >  you can use 'top -U<username>' or 'ps'.
> >
> > hwidjaja
> > In article < 82jp3a$pal$1@nnrp1.deja.com >,
> >   atharaken@my-deja.com wrote:
> > > This is a snippet of some code I am using. I get a Segmentation
> Fault
> > > in _smalloc (malloc). It works sometimes and when a large number
of
> > > users are accessing it gives a segmentation fault.
> > >
> > > It happens at the line:
> > >
> > > _string = new char [_capacity];
> > >
> > > I call this using _Init(12);
> > > This is a protected class.
> > >
> > > /* BEGIN CODE HERE */
> > >
> > > bool MyClass::_Init (long size)
> > > {
> > >     _capacity = size;
> > >     _string = new char [_capacity]; //Get Segmentation Fault at
> > _smalloc
> > >     if (!_string) {
> > >         _capacity = _size = 0L;
> > >         return false;
> > >     }
> > >     _size = 0;
> > >     _string[0] = '\0';
> > >     return true;
> > > }
> > >
> > > /*END CODE HERE */
> > >
> > > Machine: i386
> > > OS: Solaris 7
> > > GCC 2.95 (Release)
> > >
> > > Any help is appreciated
> > >
> > > Thanks
> > >
> > > Sent via Deja.com http://www.deja.com/
> > > Before you buy.
> > >
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[atharaken]

Question:

I use top to check the memory use. It does have a memory leak which I
need to fix but does Solaris have a limit on the amount of memory that
a application uses. It yes, how can that be changed.

This computer has around 2GB of RAM and this is the only app running on
this machine apart from other system daemons etc..

In article < 82kesh$994$1@nnrp1.deja.com >,
  hwidjaja@my-deja.com wrote:
> I didn't see problem at your code..
> you mentioned that i gives a segm. fault when large number of users
> are accessing it ...
> I suspect that there is memory leaks in your code. Then you run out of
> memory...
> my suggestion is to check it,
>  you can use 'top -U<username>' or 'ps'.
>
> hwidjaja
> In article < 82jp3a$pal$1@nnrp1.deja.com >,
>   atharaken@my-deja.com wrote:
> > This is a snippet of some code I am using. I get a Segmentation
Fault
> > in _smalloc (malloc). It works sometimes and when a large number of
> > users are accessing it gives a segmentation fault.
> >
> > It happens at the line:
> >
> > _string = new char [_capacity];
> >
> > I call this using _Init(12);
> > This is a protected class.
> >
> > /* BEGIN CODE HERE */
> >
> > bool MyClass::_Init (long size)
> > {
> >     _capacity = size;
> >     _string = new char [_capacity]; //Get Segmentation Fault at
> _smalloc
> >     if (!_string) {
> >         _capacity = _size = 0L;
> >         return false;
> >     }
> >     _size = 0;
> >     _string[0] = '\0';
> >     return true;
> > }
> >
> > /*END CODE HERE */
> >
> > Machine: i386
> > OS: Solaris 7
> > GCC 2.95 (Release)
> >
> > Any help is appreciated
> >
> > Thanks
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[atharaken]

Any free tools to detect memory leaks!!! Purify and Insure both cost a
lot. Any others!

In article < 82kesh$994$1@nnrp1.deja.com >,
  hwidjaja@my-deja.com wrote:
> I didn't see problem at your code..
> you mentioned that i gives a segm. fault when large number of users
> are accessing it ...
> I suspect that there is memory leaks in your code. Then you run out of
> memory...
> my suggestion is to check it,
>  you can use 'top -U<username>' or 'ps'.
>
> hwidjaja
> In article < 82jp3a$pal$1@nnrp1.deja.com >,
>   atharaken@my-deja.com wrote:
> > This is a snippet of some code I am using. I get a Segmentation
Fault
> > in _smalloc (malloc). It works sometimes and when a large number of
> > users are accessing it gives a segmentation fault.
> >
> > It happens at the line:
> >
> > _string = new char [_capacity];
> >
> > I call this using _Init(12);
> > This is a protected class.
> >
> > /* BEGIN CODE HERE */
> >
> > bool MyClass::_Init (long size)
> > {
> >     _capacity = size;
> >     _string = new char [_capacity]; //Get Segmentation Fault at
> _smalloc
> >     if (!_string) {
> >         _capacity = _size = 0L;
> >         return false;
> >     }
> >     _size = 0;
> >     _string[0] = '\0';
> >     return true;
> > }
> >
> > /*END CODE HERE */
> >
> > Machine: i386
> > OS: Solaris 7
> > GCC 2.95 (Release)
> >
> > Any help is appreciated
> >
> > Thanks
> >
> > Sent via Deja.com http://www.deja.com/
> > Before you buy.
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[hwidjaja]

I didn't see problem at your code..
you mentioned that i gives a segm. fault when large number of users
are accessing it ...
I suspect that there is memory leaks in your code. Then you run out of
memory...
my suggestion is to check it,
 you can use 'top -U<username>' or 'ps'.

hwidjaja
In article < 82jp3a$pal$1@nnrp1.deja.com >,
  atharaken@my-deja.com wrote:
> This is a snippet of some code I am using. I get a Segmentation Fault
> in _smalloc (malloc). It works sometimes and when a large number of
> users are accessing it gives a segmentation fault.
>
> It happens at the line:
>
> _string = new char [_capacity];
>
> I call this using _Init(12);
> This is a protected class.
>
> /* BEGIN CODE HERE */
>
> bool MyClass::_Init (long size)
> {
>     _capacity = size;
>     _string = new char [_capacity]; //Get Segmentation Fault at
_smalloc
>     if (!_string) {
>         _capacity = _size = 0L;
>         return false;
>     }
>     _size = 0;
>     _string[0] = '\0';
>     return true;
> }
>
> /*END CODE HERE */
>
> Machine: i386
> OS: Solaris 7
> GCC 2.95 (Release)
>
> Any help is appreciated
>
> Thanks
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

Segmentation Fault: new char[12]

[atharaken]

This is a snippet of some code I am using. I get a Segmentation Fault
in _smalloc (malloc). It works sometimes and when a large number of
users are accessing it gives a segmentation fault.

It happens at the line:

_string = new char [_capacity];

I call this using _Init(12);
This is a protected class.


/* BEGIN CODE HERE */

bool MyClass::_Init (long size)
{
    _capacity = size;
    _string = new char [_capacity]; //Get Segmentation Fault at _smalloc
    if (!_string) {
        _capacity = _size = 0L;
        return false;
    }
    _size = 0;
    _string[0] = '\0';
    return true;
}

/*END CODE HERE */


Machine: i386
OS: Solaris 7
GCC 2.95 (Release)

Any help is appreciated

Thanks


Sent via Deja.com http://www.deja.com/
Before you buy.

Re: Segmentation Fault: new char[12]

[Josh Gagliardi]

>         I don't think the "memory leak" is your biggest problem. The
> visible result of a memory leak is that malloc() returns 0 after some
> amount of time, but your code checks for that. Getting a _SEGFAULT_
> from malloc() (or free(), or whatever) is almost always caused by
> your application corrupting the malloc "headers" that are often
> stored adjacent to the allocated memory. You need to be looking for
> array-bounds violations, not memory-leaks.

Freeing something twice can also do bad things to the free list and
cause a later malloc or free to SEGFAULT.