99 ways to program a hex, Part 8: C99, const and restrict correctness

Much like the difference between part 1 [1] and part 4 [2], there is very little difference between today's code and yesterday's code [3]:

>
```
/*************************************************************************
*
* Copyright 2012 by Sean Conner. All Rights Reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*
* Comments, questions and criticisms can be sent to: sean@conman.org
*
*************************************************************************/
/* Style: C99, const and restrict correctness */
#include <stdio.h>
#include <ctype.h>
#include <string.h>
#include <stdlib.h>
#define LINESIZE 16
static void do_dump (FILE *const restrict,FILE *const restrict);
/****************************************************************/
int main(const int argc,char *const restrict argv[])
{
if (argc == 1)
do_dump(stdin,stdout);
else
{
for (int i = 1 ; i < argc ; i++)
{
FILE *fp = fopen(argv[i],"rb");
if (fp == NULL)
{
perror(argv[i]);
continue;
}
printf("-----%s-----\n",argv[i]);
do_dump(fp,stdout);
fclose(fp);
}
}
return EXIT_SUCCESS;
}
/******************************************************************/
static void do_dump(FILE *const restrict fpin,FILE *const restrict fpout)
{
unsigned char buffer[BUFSIZ];
size_t offset;
size_t bread;
offset = 0;
while((bread = fread(buffer,1,BUFSIZ,fpin)) > 0)
{
unsigned char *pbyte = buffer;
while (bread > 0)
{
char ascii[LINESIZE + 1];
fprintf(fpout,"%08lX: ",(unsigned long)offset);
size_t j = 0;
do
{
fprintf(fpout,"%02X ",*pbyte);
if (isprint(*pbyte))
ascii [j] = *pbyte;
else
ascii [j] = '.';
pbyte ++;
offset ++;
j ++;
bread --;
} while ((j < LINESIZE) && (bread > 0));
ascii [j] = '\0';
if (j < LINESIZE)
for (size_t i = j ; i < LINESIZE ; i++) fprintf(fpout," ");
fprintf(fpout,"%s\n",ascii);
}
if (fflush(fpout) == EOF)
{
perror("output");
exit(EXIT_FAILURE);
}
}
}
/***************************************************************/
```

C99 [4] adds restrict to the ways one can modify the access to a variable. The rational behind this is a bit esoteric—it tells the compiler that a pointer is the only pointer to a block of memory.

Yes, it does seem odd to have to add a keyword for that, but it does help with code optimization. For instance, the following (silly) function:

>
```
int foo(int *p1,int *p2)
{
*p2 = *p1 * 17;
return *p1 * 17;
}
```

The problem here is that the compiler has to do the multiplication twice, as p2 could be pointing to the same location as p1, and thus, the contents pointed to by p1 could be modified. So the compiler is forced to write machine code like:

>
```
foo: mov ebx,[esp + 4] ; get p1
mov eax,[ebx] ; read *p1
imul eax,17 ; multiply
mov edx,[esp + 8] ; get p2
mov [edx],eax ; save results in *p2
mov eax,[ebx] ; read *p1
imul eax,17 ; multiply
ret
```

Not exactly optimum, but the C compiler is constrained because of the semantics of pointers in C. Change the C code a bit:

>
```
int foo(int *restrict p1,int *restrict p2)
{
*p2 = *p1 * 17;
return *p1 * 17;
}
```

And the compiler can now produce:

>
```
foo: mov ebx,[esp + 4] ; get p1
mov eax,[ebx] ; read *p1
imul eax,17 ; multiply
mov edx,[esp + 8] ; save
mov [edx],eax ; return result
ret
```

Okay, it's only a savings of two instructions (plus an an additional read) but when you're trying to multiply huge matrices, it can add up.

[1] /boston/2012/01/09.1

[2] /boston/2012/01/12.3

[3] /boston/2012/01/15.1

[4] http://en.wikipedia.org/wiki/C99

[5] /boston/2012/01/15.1

[6] /boston/2012/01/17.1

Gemini Mention this post

Contact the author