[PATCH v2 10/19] nptl: Use tidlock when accessing TID on pthread_getaffinity_np

[Florian Weimer]

* Adhemerval Zanella:

> Checked on x86_64-linux-gnu.
> ---
>  nptl/pthread_getaffinity.c | 22 +++++++++++++++++-----
>  1 file changed, 17 insertions(+), 5 deletions(-)
>
> diff --git a/nptl/pthread_getaffinity.c b/nptl/pthread_getaffinity.c
> index 18261ddae0..5268d86e6e 100644
> --- a/nptl/pthread_getaffinity.c
> +++ b/nptl/pthread_getaffinity.c
> @@ -29,12 +29,24 @@
>  int
>  __pthread_getaffinity_np (pthread_t th, size_t cpusetsize, cpu_set_t *cpuset)
>  {
> -  const struct pthread *pd = (const struct pthread *) th;
> +  struct pthread *pd = (struct pthread *) th;
>  
> -  int res = INTERNAL_SYSCALL_CALL (sched_getaffinity, pd->tid,
> -				   MIN (INT_MAX, cpusetsize), cpuset);
> -  if (INTERNAL_SYSCALL_ERROR_P (res))
> -    return INTERNAL_SYSCALL_ERRNO (res);
> +  /* Block all signal, since the lock is recursive and used on pthread_cancel
> +     (which should be async-signal-safe).  */

The pthread_cancel reference looks like a cut-and-paste-bug.

> +  sigset_t oldmask;
> +  __libc_signal_block_all (&oldmask);
> +  lll_lock (pd->tidlock, LLL_PRIVATE);
> +
> +  int res = pd->tid == 0
> +	    ? -ESRCH
> +	    : INTERNAL_SYSCALL_CALL (sched_getaffinity, pd->tid,
> +				     MIN (INT_MAX, cpusetsize), cpuset);
> +
> +  lll_unlock (pd->tidlock, LLL_PRIVATE);
> +  __libc_signal_restore_set (&oldmask);
> +
> +  if (res < 0)
> +    return -res;

ESRCH doesn't look like the right error code here.  Should we return an
affinity mask without any bits set?

Thanks,
Florian