Changeset 8ecc042a in rtems


Ignore:
Timestamp:
Nov 6, 2002, 11:16:48 AM (18 years ago)
Author:
Ralf Corsepius <ralf.corsepius@…>
Branches:
4.10, 4.11, 4.8, 4.9, master
Children:
e05468fb
Parents:
5d941f3
Message:

2002-11-06 Ralf Corsepius <corsepiu@…>

  • rtems/new-exceptions/cpu.h: Remove sections on CPU_INLINE_ENABLE_DISPATCH and CPU_UNROLL_ENQUEUE_PRIORITY.
  • rtems/old-exceptions/cpu.h: Remove sections on CPU_INLINE_ENABLE_DISPATCH and CPU_UNROLL_ENQUEUE_PRIORITY.
  • rtems/score/cpu.h: Insert sections on CPU_INLINE_ENABLE_DISPATCH and CPU_UNROLL_ENQUEUE_PRIORITY.
Location:
cpukit/score/cpu/powerpc
Files:
4 edited

Legend:

Unmodified
Added
Removed
  • cpukit/score/cpu/powerpc/ChangeLog

    r5d941f3 r8ecc042a  
     12002-11-06      Ralf Corsepius <corsepiu@faw.uni-ulm.de>
     2
     3        * rtems/new-exceptions/cpu.h: Remove sections on
     4        CPU_INLINE_ENABLE_DISPATCH and CPU_UNROLL_ENQUEUE_PRIORITY.
     5        * rtems/old-exceptions/cpu.h: Remove sections on
     6        CPU_INLINE_ENABLE_DISPATCH and CPU_UNROLL_ENQUEUE_PRIORITY.
     7        * rtems/score/cpu.h: Insert sections on
     8        CPU_INLINE_ENABLE_DISPATCH and CPU_UNROLL_ENQUEUE_PRIORITY.
     9
     102002-10-31      Joel Sherrill <joel@OARcorp.com>
     11
     12        * rtems/new-exceptions/cpu.h: Removed warnings.
     13
     14
    1152002-10-31      Joel Sherrill <joel@OARcorp.com>
    216
  • cpukit/score/cpu/powerpc/rtems/new-exceptions/cpu.h

    r5d941f3 r8ecc042a  
    4848
    4949/* conditional compilation parameters */
    50 
    51 /*
    52  *  Should the calls to _Thread_Enable_dispatch be inlined?
    53  *
    54  *  If TRUE, then they are inlined.
    55  *  If FALSE, then a subroutine call is made.
    56  *
    57  *  Basically this is an example of the classic trade-off of size
    58  *  versus speed.  Inlining the call (TRUE) typically increases the
    59  *  size of RTEMS while speeding up the enabling of dispatching.
    60  *  [NOTE: In general, the _Thread_Dispatch_disable_level will
    61  *  only be 0 or 1 unless you are in an interrupt handler and that
    62  *  interrupt handler invokes the executive.]  When not inlined
    63  *  something calls _Thread_Enable_dispatch which in turns calls
    64  *  _Thread_Dispatch.  If the enable dispatch is inlined, then
    65  *  one subroutine call is avoided entirely.]
    66  */
    67 
    68 #define CPU_INLINE_ENABLE_DISPATCH       FALSE
    69 
    70 /*
    71  *  Should the body of the search loops in _Thread_queue_Enqueue_priority
    72  *  be unrolled one time?  In unrolled each iteration of the loop examines
    73  *  two "nodes" on the chain being searched.  Otherwise, only one node
    74  *  is examined per iteration.
    75  *
    76  *  If TRUE, then the loops are unrolled.
    77  *  If FALSE, then the loops are not unrolled.
    78  *
    79  *  The primary factor in making this decision is the cost of disabling
    80  *  and enabling interrupts (_ISR_Flash) versus the cost of rest of the
    81  *  body of the loop.  On some CPUs, the flash is more expensive than
    82  *  one iteration of the loop body.  In this case, it might be desirable
    83  *  to unroll the loop.  It is important to note that on some CPUs, this
    84  *  code is the longest interrupt disable period in RTEMS.  So it is
    85  *  necessary to strike a balance when setting this parameter.
    86  */
    87 
    88 #define CPU_UNROLL_ENQUEUE_PRIORITY      FALSE
    8950
    9051/*
  • cpukit/score/cpu/powerpc/rtems/old-exceptions/cpu.h

    r5d941f3 r8ecc042a  
    4848
    4949/* conditional compilation parameters */
    50 
    51 /*
    52  *  Should the calls to _Thread_Enable_dispatch be inlined?
    53  *
    54  *  If TRUE, then they are inlined.
    55  *  If FALSE, then a subroutine call is made.
    56  *
    57  *  Basically this is an example of the classic trade-off of size
    58  *  versus speed.  Inlining the call (TRUE) typically increases the
    59  *  size of RTEMS while speeding up the enabling of dispatching.
    60  *  [NOTE: In general, the _Thread_Dispatch_disable_level will
    61  *  only be 0 or 1 unless you are in an interrupt handler and that
    62  *  interrupt handler invokes the executive.]  When not inlined
    63  *  something calls _Thread_Enable_dispatch which in turns calls
    64  *  _Thread_Dispatch.  If the enable dispatch is inlined, then
    65  *  one subroutine call is avoided entirely.]
    66  */
    67 
    68 #define CPU_INLINE_ENABLE_DISPATCH       FALSE
    69 
    70 /*
    71  *  Should the body of the search loops in _Thread_queue_Enqueue_priority
    72  *  be unrolled one time?  In unrolled each iteration of the loop examines
    73  *  two "nodes" on the chain being searched.  Otherwise, only one node
    74  *  is examined per iteration.
    75  *
    76  *  If TRUE, then the loops are unrolled.
    77  *  If FALSE, then the loops are not unrolled.
    78  *
    79  *  The primary factor in making this decision is the cost of disabling
    80  *  and enabling interrupts (_ISR_Flash) versus the cost of rest of the
    81  *  body of the loop.  On some CPUs, the flash is more expensive than
    82  *  one iteration of the loop body.  In this case, it might be desirable
    83  *  to unroll the loop.  It is important to note that on some CPUs, this
    84  *  code is the longest interrupt disable period in RTEMS.  So it is
    85  *  necessary to strike a balance when setting this parameter.
    86  */
    87 
    88 #define CPU_UNROLL_ENQUEUE_PRIORITY      FALSE
    8950
    9051/*
  • cpukit/score/cpu/powerpc/rtems/score/cpu.h

    r5d941f3 r8ecc042a  
    1111#endif
    1212
     13/* conditional compilation parameters */
     14
     15/*
     16 *  Should the calls to _Thread_Enable_dispatch be inlined?
     17 *
     18 *  If TRUE, then they are inlined.
     19 *  If FALSE, then a subroutine call is made.
     20 *
     21 *  Basically this is an example of the classic trade-off of size
     22 *  versus speed.  Inlining the call (TRUE) typically increases the
     23 *  size of RTEMS while speeding up the enabling of dispatching.
     24 *  [NOTE: In general, the _Thread_Dispatch_disable_level will
     25 *  only be 0 or 1 unless you are in an interrupt handler and that
     26 *  interrupt handler invokes the executive.]  When not inlined
     27 *  something calls _Thread_Enable_dispatch which in turns calls
     28 *  _Thread_Dispatch.  If the enable dispatch is inlined, then
     29 *  one subroutine call is avoided entirely.]
     30 */
     31
     32#define CPU_INLINE_ENABLE_DISPATCH       FALSE
     33
     34/*
     35 *  Should the body of the search loops in _Thread_queue_Enqueue_priority
     36 *  be unrolled one time?  In unrolled each iteration of the loop examines
     37 *  two "nodes" on the chain being searched.  Otherwise, only one node
     38 *  is examined per iteration.
     39 *
     40 *  If TRUE, then the loops are unrolled.
     41 *  If FALSE, then the loops are not unrolled.
     42 *
     43 *  The primary factor in making this decision is the cost of disabling
     44 *  and enabling interrupts (_ISR_Flash) versus the cost of rest of the
     45 *  body of the loop.  On some CPUs, the flash is more expensive than
     46 *  one iteration of the loop body.  In this case, it might be desirable
     47 *  to unroll the loop.  It is important to note that on some CPUs, this
     48 *  code is the longest interrupt disable period in RTEMS.  So it is
     49 *  necessary to strike a balance when setting this parameter.
     50 */
     51
     52#define CPU_UNROLL_ENQUEUE_PRIORITY      FALSE
     53
    1354#ifdef _OLD_EXCEPTIONS
    1455#include <rtems/old-exceptions/cpu.h>
Note: See TracChangeset for help on using the changeset viewer.