Why is it segfaulting?

XSquared · 03-29-2004, 07:15 PM

I've reduced the code to something simple that still segfaults (for me at least) on the memcpy line. It does the same thing with strcpy. I'm using GCC 3.3 on Gentoo Linux.

Code:

char *funct( char *data, int bufsiz )
{
  
  char *temp = malloc( sizeof( char ) * bufsiz + 1 );
  temp[ bufsiz ] = '\0';

  assert( temp != NULL );

  memcpy( data, temp, bufsiz );

  free( temp );
  
  return data;
  
}	  

int main( void )
{

  char *str = "hello world";
  printf( "%s\n", funct( str, strlen( str ) ) );

  return 0;
  
}

**Hammer** · 03-29-2004, 07:33 PM

Because this
char *str = "hello world";
is a pointer to a static array. You cannot overwrite string literals (or at least you shoudln't).
Use
char str[] = "hello world";

What exactly are you trying to do in that function?

**Prelude** · 03-29-2004, 07:51 PM

>XSquared
Aroo? I'm mildly confused about you writing this code XSquared. :P

>char *temp = malloc( sizeof( char ) * bufsiz + 1 );
sizeof ( char ) isn't needed, the size of a char is guaranteed to be 1. If you want to make sure that the argument to malloc is size_t then make bufsiz size_t. The sizeof only clutters the call.

>temp[ bufsiz ] = '\0';
>assert( temp != NULL );
Wrong order. Check first, then access. An assert isn't meant for run-time checking, it's meant for asserting impossible cases. malloc returning NULL is far from an impossible case.

>memcpy( data, temp, bufsiz );
You're writing the contents of temp (which are undefined except for the nul character at the end) to data. Your segfault is likely data being in read-only memory because it is a string literal, but this statement does nothing useful, and plenty undefined.

I say again, aroo?

XSquared · 03-29-2004, 08:15 PM

>Wrong order. Check first, then access.
Those were just a couple of quick, butchered attempts to find the error.

>which are undefined except for the nul character at the end
There's a whole lot of processing which happens before the memcpy statement. I just tried to find the simplest case where it fails.

>the size of a char is guaranteed to be 1
Heh, ya learn something new every day.

**Salem** · 03-30-2004, 12:47 AM

> memcpy( data, temp, bufsiz );
Too little, and to the wrong place

memcpy( temp, data, bufsiz+1 );

You're overwriting your read-only string

XSquared · 03-30-2004, 06:41 AM

>...and to the wrong place
I'm copying it to the right place. I'm overwriting data with temp.

I think you have the arguments to memcpy reversed. From the man pages:

Code:

MEMCPY(3)           Linux Programmer's Manual           MEMCPY(3)

NAME
       memcpy - copy memory area

SYNOPSIS
       #include <string.h>

       void *memcpy(void *dest, const void *src, size_t n);

quzah · 03-30-2004, 06:50 AM

Code:

char *str = "hello world";
printf( "%s\n", funct( str, strlen( str ) ) );

No. You missed his point: You are trying to write over top of your read only string str. Your arguments are fine. What you're passing to it is not.

Quzah.

XSquared · 03-30-2004, 06:52 AM

Sorry. I forgot that in the code here I still had it as a char *. I changed it to a char [] a while ago and it's been working fine.

03-29-2004, 07:15 PM	#1
XSquared C++ Developer Join Date: Jun 2002 Location: UWaterloo Posts: 2,718	Why is it segfaulting? I've reduced the code to something simple that still segfaults (for me at least) on the memcpy line. It does the same thing with strcpy. I'm using GCC 3.3 on Gentoo Linux. Code: char funct( char data, int bufsiz ) { char temp = malloc( sizeof( char ) bufsiz + 1 ); temp[ bufsiz ] = '\0'; assert( temp != NULL ); memcpy( data, temp, bufsiz ); free( temp ); return data; } int main( void ) { char str = "hello world"; printf( "%s\n", funct( str, strlen( str ) ) ); return 0; } __________________ Naturally I didn't feel inspired enough to read all the links for you, since I already slaved away for long hours under a blistering sun* pressing the search button after typing four whole words! - Quzah You. Fetch me my copy of the Wall Street Journal. You two, fight to the death - Stewie
XSquared is offline

03-29-2004, 07:33 PM	#2
Hammer End Of Line Join Date: Apr 2002 Posts: 6,240	Because this char str = "hello world"; is a pointer to a static array. You cannot overwrite string literals (or at least you shoudln't). Use char str[] = "hello world"; What exactly are you trying to do in that function? __________________ When all else fails, read the instructions. If you're posting code, use code tags: [code] / insert code here */ [/code]
Hammer is offline

03-29-2004, 07:51 PM	#3
Prelude Code Goddess Join Date: Sep 2001 Posts: 9,661	>XSquared Aroo? I'm mildly confused about you writing this code XSquared. :P >char temp = malloc( sizeof( char ) bufsiz + 1 ); sizeof ( char ) isn't needed, the size of a char is guaranteed to be 1. If you want to make sure that the argument to malloc is size_t then make bufsiz size_t. The sizeof only clutters the call. >temp[ bufsiz ] = '\0'; >assert( temp != NULL ); Wrong order. Check first, then access. An assert isn't meant for run-time checking, it's meant for asserting impossible cases. malloc returning NULL is far from an impossible case. >memcpy( data, temp, bufsiz ); You're writing the contents of temp (which are undefined except for the nul character at the end) to data. Your segfault is likely data being in read-only memory because it is a string literal, but this statement does nothing useful, and plenty undefined. I say again, aroo? __________________ My best code is written with the delete key.
Prelude is offline

03-29-2004, 08:15 PM	#4
XSquared C++ Developer Join Date: Jun 2002 Location: UWaterloo Posts: 2,718	>Wrong order. Check first, then access. Those were just a couple of quick, butchered attempts to find the error. >which are undefined except for the nul character at the end There's a whole lot of processing which happens before the memcpy statement. I just tried to find the simplest case where it fails. >the size of a char is guaranteed to be 1 Heh, ya learn something new every day. __________________ Naturally I didn't feel inspired enough to read all the links for you, since I already slaved away for long hours under a blistering sun pressing the search button after typing four whole words! - Quzah You. Fetch me my copy of the Wall Street Journal. You two, fight to the death - Stewie
XSquared is offline

03-30-2004, 12:47 AM	#5
Salem and the hat of vanishing Join Date: Aug 2001 Location: The edge of the known universe Posts: 21,214	> memcpy( data, temp, bufsiz ); Too little, and to the wrong place memcpy( temp, data, bufsiz+1 ); You're overwriting your read-only string __________________ If you dance barefoot on the broken glass of undefined behaviour, you've got to expect the occasional cut. Up to 8Mb PlusNet broadband from only £5.99 a month!
Salem is offline

03-30-2004, 06:41 AM	#6
XSquared C++ Developer Join Date: Jun 2002 Location: UWaterloo Posts: 2,718	>...and to the wrong place I'm copying it to the right place. I'm overwriting data with temp. I think you have the arguments to memcpy reversed. From the man pages: Code: MEMCPY(3) Linux Programmer's Manual MEMCPY(3) NAME memcpy - copy memory area SYNOPSIS #include <string.h> void memcpy(void dest, const void src, size_t n); __________________ Naturally I didn't feel inspired enough to read all the links for you, since I already slaved away for long hours under a blistering sun* pressing the search button after typing four whole words! - Quzah You. Fetch me my copy of the Wall Street Journal. You two, fight to the death - Stewie
XSquared is offline

03-30-2004, 06:50 AM	#7
quzah +++ OK NO CARRIER Join Date: Oct 2001 Posts: 10,262	Code: char *str = "hello world"; printf( "%s\n", funct( str, strlen( str ) ) ); No. You missed his point: You are trying to write over top of your read only string str. Your arguments are fine. What you're passing to it is not. Quzah. __________________ Hundreds of thousands of dipshits can't be wrong. Are you up for the suck?
quzah is offline

03-30-2004, 06:52 AM	#8
XSquared C++ Developer Join Date: Jun 2002 Location: UWaterloo Posts: 2,718	Sorry. I forgot that in the code here I still had it as a char . I changed it to a char [] a while ago and it's been working fine. __________________ Naturally I didn't feel inspired enough to read all the links for you, since I already slaved away for long hours under a blistering sun* pressing the search button after typing four whole words! - Quzah You. Fetch me my copy of the Wall Street Journal. You two, fight to the death - Stewie
XSquared is offline

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
segfaulting!	CMakesMeSad :(	C Programming	23	07-10-2009 01:04 PM
why is strncpy segfaulting here?	Calef13	C Programming	3	12-29-2008 03:27 PM
Segfaulting Distance Program	radiohead	C Programming	2	01-09-2006 08:48 PM
a segfaulting algorythm	demonus	C Programming	8	08-11-2003 08:06 AM