Re: [PATCH] Fix Linux fcntl OFD locks for non-LFS architectures (BZ#20251)

[Florian Weimer <fweimer@redhat.com>]

On 04/04/2018 11:26 PM, Adhemerval Zanella wrote:
> @@ -59,16 +50,56 @@ __libc_fcntl (int fd, int cmd, ...)
>   
>     cmd = FCNTL_ADJUST_CMD (cmd);
>   
> -  if (cmd == F_SETLKW || cmd == F_SETLKW64)
> -    return SYSCALL_CANCEL (fcntl64, fd, cmd, (void *) arg);
> -
> -  return fcntl_common (fd, cmd, arg);
> +  switch (cmd)
> +    {
> +      case F_SETLKW:
> +	return SYSCALL_CANCEL (fcntl64, fd, cmd, arg);
> +      case F_GETLK64:
> +      case F_OFD_GETLK:
> +	{
> +	  struct flock *flk = (struct flock *) arg;
> +	  struct flock64 flk64 = flock_to_flock64 (flk);
> +	  int ret = SYSCALL_CANCEL (fcntl64, fd, cmd, &flk64);
> +	  if (ret == -1)
> +	    return -1;
> +	  flk->l_type = flk64.l_type;
> +	  flk->l_whence = flk64.l_whence;
> +	  flk->l_start = flk64.l_start;
> +	  flk->l_len = flk64.l_len;
> +	  flk->l_pid = flk64.l_pid;
> +	  return ret;

Doesn't this need an overflow check?

I think you should mention somewhere that this introduces new 
cancellation points.

Thanks,
Florian