Changes between Version 6 and Version 7 of Ticket #11815, comment 15


Ignore:
Timestamp:
Nov 23, 2024, 8:40:53 PM (4 weeks ago)
Author:
cocobean

Legend:

Unmodified
Added
Removed
Modified

  • Ticket #11815, comment 15

    v6 v7  
    1 For hrev58330, started to isolate the issues with these remaining FAILED tests:
     1For hrev58360, started to isolate the issues with these remaining FAILED tests:
    22
    3 Thread Scheduling Execution non-implementation of SCHED_FIFO affects:
    4 {{{
    5 conformance/interfaces/pthread_detach/4-2: execution: FAILED
    6 }}}
    7 NOTE: The error within the pthread_create 2-1 test deals with using pthread-detach. No error with pthread_create and pthread-join functionality during the 2-1 test.
     3All of the main OpenPOSIX pthread tests pass in comparison to Linux.
    84
    9 
    10 {{{
    11 conformance/interfaces/pthread_rwlock_rdlock/2-3: execution: FAILED
    12 }}}
    13 
    14 {{{
    15 main: attempt read lock
    16 main: acquired read lock
    17 main: create wr_thread, with priority: -1
    18 wr_thread: attempt write lock
    19 main: create rd_thread, with priority: 0
    20 rd_thread: attempt read lock
    21 Test FAILED: rd_thread blocked on read lock
    22 }}}
    23 NOTE: pthread_rwlock_rdlock seems related to a non-implemented thread scheduling priority (i.e. SCHED_FIFO).
    24 
    25 {{{
    26 conformance/interfaces/pthread_barrier_init/3-1: execution: FAILED
    27 }}}
    28 RESULT:
    29 
    30 Test FAILED: pthread_barrier_init() does not return EINVAL when initializing a barrier with count=0, return code -2147483643, Invalid Argument
    31 
    32 
    33 {{{
    34 conformance/interfaces/pthread_attr_setstack/2-1: execution: FAILED
    35 conformance/interfaces/pthread_attr_setstack/6-1: execution: FAILED
    36 }}}
    37 
    38 RESULT:
    39 
    40 Test FAILED:
    41 
    42 2-1: got the wrong stacksize or stackaddr: No error.
    43 
    44 6-1: got the wrong return value: No error
     5POSIX implementation of SCHED_FIFO and sched_get_priority_min/max will help improve mentioned FAILED tests - but these specific tests seemed removed from Linux testing - so not seen as high priority at the moment.