[lsb-discuss] New Interfaces for 3.2/4.0

Wichmann, Mats D mats.d.wichmann at intel.com
Tue Aug 1 20:16:44 PDT 2006


 

>However, if POSIX does go ahead with this proposal, then the following
>interfaces should be actively considered for either 3.2 or 4.0:
>
>+--------------------------------+---------+
>| Iname                          | Istatus |
>+--------------------------------+---------+
>| aio_cancel                     | Defered |
>| aio_cancel64                   | Defered |
>| aio_error                      | Defered |
>| aio_error64                    | Defered |
>| aio_fsync                      | Defered |
>| aio_fsync64                    | Defered |
>| aio_read                       | Defered |
>| aio_read64                     | Defered |
>| aio_return                     | Defered |
>| aio_return64                   | Defered |
>| aio_suspend                    | Defered |
>| aio_suspend64                  | Defered |
>| aio_write                      | Defered |
>| aio_write64                    | Defered |
>| pthread_barrier_destroy        | Defered |
>| pthread_barrier_init           | Defered |
>| pthread_barrier_wait           | Defered |
>| pthread_barrierattr_destroy    | Defered |
>| pthread_barrierattr_getpshared | Defered |
>| pthread_barrierattr_init       | Defered |
>| pthread_barrierattr_setpshared | Defered |
>| pthread_spin_destroy           | Defered |
>| pthread_spin_init              | Defered |
>| pthread_spin_lock              | Defered |
>| pthread_spin_trylock           | Defered |
>| pthread_spin_unlock            | Defered |

Based on my understanding of the level of support
in our target systems, the barrier and spinlock
thread stuff could be ready now.  The aio stuff,
as you know, awaits a resolution of the dilemma
that exists today:
- userspace aio, in glibc, isn't full enough POSIX
conforming for ISVs to want it (or so the small
number that have weighed in have told us; and
it didn't pass the one testsuite we had available)
- kernel-level aio doesn't implement the POSIX
API/ABI (although I keep hearing there's a project
by Bull and others to wrap the ABI in the POSIX
set)
It may be that the former isn't really a problem
any more, it's hard to tell without revisiting
the issue. It *used* to be that opinions ran
pretty strong on that. Time may have changed it.




More information about the lsb-discuss mailing list