[Bug libc/28730] Extend posix_spawnattr_* interface with more useful features

["adhemerval.zanella at linaro dot org" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=28730

--- Comment #6 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to Cristian Rodríguez from comment #5)
> So the best way way would be to use the pad space to introduce a an internal
> opaque struct  which we can modify or extend at will ? even with all
> implementation-defined leisure we still shouldn't break the ABI..

Yes, although I would first use the current space available and keep space for
a pointer to allow extend the functionaly.  Something like:

  typedef struct
  {
    short int __flags;
    pid_t __pgrp;
    sigset_t __sd;
    sigset_t __ss;
    struct sched_param __sp;
    int __policy;
    /* Add newer fields here.  */
    int __pad[N];
    void *__reserved/
  } posix_spawnattr_t;

So for posix_spawnattr_setpdeathsig_np it would be something like:

  typedef struct
  {
    short int __flags;
    pid_t __pgrp;
    sigset_t __sd;
    sigset_t __ss;
    struct sched_param __sp;
    int __policy;
    int __deathsignal;
    int __pad[M];
  } posix_spawnattr_t;

With M adjusted to keep the current size. For posix_spawnattr_setrlimit_np I
would do:

  typedef struct
  {
    short int __flags;
    pid_t __pgrp;
    sigset_t __sd;
    sigset_t __ss;
    struct sched_param __sp;
    int __policy;
    int __deathsignal;
    void *__resource;
    int __pad[M];
  } posix_spawnattr_t;

And internally '__resource' will be array of:

  struct posix_spawn_resource
  {
    int resource;
    struct rlimit64 rlim;
  }

(The rlimit has the LFS issue, so I think it shold be supported only for LFS
mode).

-- 
You are receiving this mail because:
You are on the CC list for the bug.