Re: [PATCH] nextfd(2)

[Ben Pfaff <blp@cs.stanford.edu>]

"H. Peter Anvin" <hpa@zytor.com> writes:

> On 04/06/2012 02:54 AM, Alexey Dobriyan wrote:
>> 
>> Without proc knowledge about fdtable is gathered linearly and still unreliable.
>> With nextfd(2), even procful environments could lose several failure branches.
>> And they can keep old dumb fd++ or smart /proc/self/fd loops for a change.
>> 
>
> Incidentally, if we were to create a system call for this -- which I so
> far see no reason for -- I would make it return a select-style bitmask
> of file descriptors in use, not a "next fd" which would require a system
> call per iteration.

It's already possible to do something a little like that with the
existing "poll" system call:

#include <stdio.h>
#include <sys/poll.h>

int
main(void)
{
  enum { N_FDS = 1024 };
  struct pollfd fds[N_FDS];
  int error;
  int i;

  for (i = 0; i < N_FDS; i++)
    {
      fds[i].fd = i;
      fds[i].events = 0;
    }

  error = poll(fds, N_FDS, 0);
  if (error < 0)
    perror("poll");
  else
    {
      printf("valid fds:");
      for (i = 0; i < N_FDS; i++)
        if (!(fds[i].revents & POLLNVAL))
          printf(" %d", fds[i].fd);
      printf("\n");
    }

  return 0;
}

blp@blp:~/tmp(0)$ gcc tmp.c -Wall
blp@blp:~/tmp(0)$ ./a.out
valid fds: 0 1 2
blp@blp:~/tmp(0)$ ./a.out 5>/dev/null
valid fds: 0 1 2 5
blp@blp:~/tmp(0)$ ./a.out 5>/dev/null 3</dev/stdin
valid fds: 0 1 2 3 5
blp@blp:~/tmp(0)$ 

-- 
Ben Pfaff 
http://benpfaff.org