_____________________________________________________________________________
Notice: This is an unapproved draft interpretation. Use at
your own risk.


 _____________________________________________________________________________
                                                PASC Interpretation reference
                                                        1003.13-2003  #003

 _____________________________________________________________________________

        Interpretation Number:  003
        Topic:                  Required Threads functions
        Relevant Sections:      Table B-1, Table 1-1, Table 7-1/7-2 


PASC Interpretation Request:
----------------------------  

Reference: #3

        From: Paul Chen
        Date: 2006 May 30


------------------------------------------------------------------------ 

 7  Defect Report concerning (number and title of International Standard
    or DIS final text, if applicable): 

AEP:POSIX(R) Realtime and Embedded Application Support: IEEE Std 1003.13-2003

------------------------------------------------------------------------ 

 8  Qualifier (e.g. error, omission, clarification required):

3

Error=1 , Omission=2, Clarification=3

------------------------------------------------------------------------ 

 9  References in document (e.g. page, clause, figure, and/or table
    numbers):

Table B-1, Table 1-1, Table 7-1/7-2 (for PSE52, e.g.)

------------------------------------------------------------------------ 

10  Nature of defect (complete, concise explanation of the perceived
    problem):


Table B-1, pg 116, indicates that when options _POSIX_THREADS and _POSIX_TIMEOUTS are included in a standard, then the following three APIs are included:
    pthread_mutex_timedlock()
    pthread_rwlock_timedrdlock()
    pthread_rwlock_timedwrlock()

Each PSE5X definition (e.g., PSE52, as defined by Tables 7-1 and 7-2) does include option _POSIX_TIMEOUTS. 

Each PSE5X definition, however, does NOT include _POSIX_THREADS, but instead includes POSIX_THREADS_BASE, which is defined in Table 1-1 (pg 8) and which does NOT include any of the above three APIs (per footnote 6 on pg 10).

Hence, it seems that a strict interpretation of the standard indicates
that none of the above three APIs are included in any of the PSE5X
standards. But in an email exchange with the 1003.13 Chair, he wrote that
he thought the intent of the standard was that "timeouts are always
required in practice. That said, we may have allowed but not required
them in the smallest profiles.  But there is no sin in providing them
in all profiles."

Joe suggested that I file this interpretation request.

------------------------------------------------------------------------ 

11  Solution proposed by the submitter (optional):

I would like clarification as to whether these APIs (and _POSIX_TIMEOUTS) were meant to be included in PSE5X.


------------------------------------------------------------------------ 

Interpretation:
----------------

The standard is clear in table 1-18 that POSIX_RW_LOCKS is not required 
in any of the PSE5x profiles, therefore the 
    pthread_rwlock_timedrdlock()
    pthread_rwlock_timedwrlock()
functions are not required.

The standard is unclear on the issue of inclusion
of pthread_mutex_timedlock(), and no conformance
distinction can be made between alternative implementations based
on this. This is being referred to the sponsor.

Rationale:
   Following table 1-1 on page 10, note 6 says "POSIX_THREADS_BASE is
    the same as the _POSIX_THREADS option, but without the functions
    belonging to the POSIX_RW_LOCKS Unit of Functionality" which
    implies that if POSIX_THREADS_BASE and _POSIX_TIMEOUTS are
    required then pthread_mutex_timedlock() is required.
    However, the absence of pthread_mutex_timedlock() from
    table 1-1 implies that it is not required.  Thus the
    standard is unclear.

Notes to editors for a future revision (not part of the interpretation):
-----------------------------------------------------------------------

pthread_mutex_timedlock() should be included in 
POSIX_THREADS_BASE in table 1-1 with a note (4) attached to it.


During the review of #3 two editorials were noted that should
be passed to IEEE staff editor as possible errata items.

On page 8, in the POSIX_RW_LOCKS cell of Table 1-1, note "(d)" should be "(4)"
and note "(e)" should be "(5)"



Forwarded to interpretations group: 30 May 2006
Proposed resolution: 16 August 2006