摘要：
只有在任务A向任务B发送消息后，任务B才能继续运行。事件标志组的数据结构最简单。只需更改多值信号量的Ctr变量的名称。对于不同的处理器，标志的数量可能不同。时间标志组中的每个标志位对应于一个事件。创建时会描述其发生与否的标志。1.对于事件标志组的创建函数，相同的前提是定义一个事件标志组变量。第一个参数是时间标志组的变量地址，第二个参数是该时间标志租约的名称，第三个参数是初始化的Flags值。
当任务需要同步时可以使用信号量。A任务给B任务发送消息后B任务才能继续运行。如果需要A任务给任务B传递数据的时候就可以采用消息队列。但对于繁杂任务的同步，比如多个时间发生以后执行一个事件，或者是C任务需要任务A和任务B都完成对应的处理任务后才能执行C任务的处理工作，这是用信号量就是等待多个内核对象，有没有一种更好的办法呢，操作系统的大神就创造出了时间标志来进行这类功能的实现。事件标志组的数据结构是最简单的，只需要将多值信号量的Ctr变量换个名字(Flags)就好了。
对于不同的处理器Flags的位数可能会有不同，时间标志组中的Flags的每一位都对应一个事件，对于其发生与否的标志在创建时进行了说明。
1，对于事件标志组的创建函数，同样前提是定义一个事件标志组变量，第一个参数是时间标志组的变量地址，第二个参数是时间标志租的Name了，第三个则是初始化的Flags值。样在创建了事件标志组的同时也初始化了其的等待队列。
************************************************************************************************************************
* CREATE AN EVENT FLAG
*
* Description: This function is called to create an event flag group.
*
* Arguments : p_grp is a pointer to the event flag group to create
*
* p_name is the name of the event flag group
*
* flags contains the initial value to store in the event flag group (typically 0).
*
* p_err is a pointer to an error code which will be returned to your application:
*
* OS_ERR_NONE if the call was successful.
* OS_ERR_CREATE_ISR if you attempted to create an Event Flag from an ISR.
* OS_ERR_ILLEGAL_CREATE_RUN_TIME if you are trying to create the Event Flag after you
* called OSSafetyCriticalStart().
* OS_ERR_NAME if 'p_name' is a NULL pointer
* OS_ERR_OBJ_CREATED if the event flag group has already been created
* OS_ERR_OBJ_PTR_NULL if 'p_grp' is a NULL pointer
*
* Returns : none
************************************************************************************************************************
*/
void OSFlagCreate (OS_FLAG_GRP *p_grp,
CPU_CHAR *p_name,
OS_FLAGS flags,
OS_ERR *p_err)
{
CPU_SR_ALLOC();
#ifdef OS_SAFETY_CRITICAL
if (p_err == (OS_ERR *)0) {
OS_SAFETY_CRITICAL_EXCEPTION();
return;
}
#endif
#ifdef OS_SAFETY_CRITICAL_IEC61508
if (OSSafetyCriticalStartFlag == DEF_TRUE) {
*p_err = OS_ERR_ILLEGAL_CREATE_RUN_TIME;
return;
}
#endif
#if OS_CFG_CALLED_FROM_ISR_CHK_EN > 0u
if (OSIntNestingCtr > (OS_NESTING_CTR)0) { /* See if called from ISR ... */
*p_err = OS_ERR_CREATE_ISR; /* ... can't CREATE from an ISR */
return;
}
#endif
#if OS_CFG_ARG_CHK_EN > 0u
if (p_grp == (OS_FLAG_GRP *)0) { /* Validate 'p_grp' */
*p_err = OS_ERR_OBJ_PTR_NULL;
return;
}
#endif
OS_CRITICAL_ENTER();
p_grp->Type = OS_OBJ_TYPE_FLAG; /* Set to event flag group type */
p_grp->NamePtr = p_name;
p_grp->Flags = flags; /* Set to desired initial value */
p_grp->TS = (CPU_TS)0;
//等待队列初始化
OS_PendListInit(&p_grp->PendList);
#if OS_CFG_DBG_EN > 0u
OS_FlagDbgListAdd(p_grp);
#endif
//事件标志组内核对象计数加1
OSFlagQty++;
OS_CRITICAL_EXIT_NO_SCHED();
*p_err = OS_ERR_NONE;
}2.等待事件标志组：
等待时间标志组的具体操作从函数的签名可以看出，主要是有一下几个条件，第一对应标志组内Flags的哪几位代表请求的时间，第二每一位是置位标志着事件发生了还是复位表明事件已经发生了，第三就是前面提到的是多个时间全部发生(ALL)还是任何一个事件(ANY)发生就标志事件标志组满足条件，最后还要选择满足条件后是否要进行取反(CONSUME)操作。
************************************************************************************************************************
* WAIT ON AN EVENT FLAG GROUP
*
* Description: This function is called to wait for a combination of bits to be set in an event flag group. Your
* application can wait for ANY bit to be set or ALL bits to be set.
*
* Arguments : p_grp is a pointer to the desired event flag group.
*
* flags Is a bit pattern indicating which bit(s) (i.e. flags) you wish to wait for.
* The bits you want are specified by setting the corresponding bits in 'flags'.
* e.g. if your application wants to wait for bits 0 and 1 then 'flags' would contain 0x03.
*
* timeout is an optional timeout (in clock ticks) that your task will wait for the
* desired bit combination. If you specify 0, however, your task will wait
* forever at the specified event flag group or, until a message arrives.
*
* opt specifies whether you want ALL bits to be set or ANY of the bits to be set.
* You can specify the 'ONE' of the following arguments:
*
* OS_OPT_PEND_FLAG_CLR_ALL You will wait for ALL bits in 'flags' to be clear (0)
* OS_OPT_PEND_FLAG_CLR_ANY You will wait for ANY bit in 'flags' to be clear (0)
* OS_OPT_PEND_FLAG_SET_ALL You will wait for ALL bits in 'flags' to be set (1)
* OS_OPT_PEND_FLAG_SET_ANY You will wait for ANY bit in 'flags' to be set (1)
*
* You can 'ADD' OS_OPT_PEND_FLAG_CONSUME if you want the event flag to be 'consumed' by
* the call. Example, to wait for any flag in a group AND then clear
* the flags that are present, set 'wait_opt' to:
*
* OS_OPT_PEND_FLAG_SET_ANY + OS_OPT_PEND_FLAG_CONSUME
*
* You can also 'ADD' the type of pend with 'ONE' of the two option:
*
* OS_OPT_PEND_NON_BLOCKING Task will NOT block if flags are not available
* OS_OPT_PEND_BLOCKING Task will block if flags are not available
*
* p_ts is a pointer to a variable that will receive the timestamp of when the event flag group was
* posted, aborted or the event flag group deleted. If you pass a NULL pointer (i.e. (CPU_TS *)0)
* then you will not get the timestamp. In other words, passing a NULL pointer is valid and
* indicates that you don't need the timestamp.
*
* p_err is a pointer to an error code and can be:
*
* OS_ERR_NONE The desired bits have been set within the specified 'timeout'
* OS_ERR_OBJ_PTR_NULL If 'p_grp' is a NULL pointer.
* OS_ERR_OBJ_TYPE You are not pointing to an event flag group
* OS_ERR_OPT_INVALID You didn't specify a proper 'opt' argument.
* OS_ERR_PEND_ABORT The wait on the flag was aborted.
* OS_ERR_PEND_ISR If you tried to PEND from an ISR
* OS_ERR_PEND_WOULD_BLOCK If you specified non-blocking but the flags were not
* available.
* OS_ERR_SCHED_LOCKED If you called this function when the scheduler is locked
* OS_ERR_TIMEOUT The bit(s) have not been set in the specified 'timeout'.
*
* Returns : The flags in the event flag group that made the task ready or, 0 if a timeout or an error
* occurred.
************************************************************************************************************************
*/
OS_FLAGS OSFlagPend (OS_FLAG_GRP *p_grp,
OS_FLAGS flags,
OS_TICK timeout,
OS_OPT opt,
CPU_TS *p_ts,
OS_ERR *p_err)
{
CPU_BOOLEAN consume;
OS_FLAGS flags_rdy;
OS_OPT mode;
OS_PEND_DATA pend_data;
CPU_SR_ALLOC();
#ifdef OS_SAFETY_CRITICAL
if (p_err == (OS_ERR *)0) {
OS_SAFETY_CRITICAL_EXCEPTION();
return ((OS_FLAGS)0);
}
#endif
#if OS_CFG_CALLED_FROM_ISR_CHK_EN > 0u
if (OSIntNestingCtr > (OS_NESTING_CTR)0) { /* See if called from ISR ... */
*p_err = OS_ERR_PEND_ISR; /* ... can't PEND from an ISR */
return ((OS_FLAGS)0);
}
#endif
#if OS_CFG_ARG_CHK_EN > 0u
if (p_grp == (OS_FLAG_GRP *)0) { /* Validate 'p_grp' */
*p_err = OS_ERR_OBJ_PTR_NULL;
return ((OS_FLAGS)0);
}
switch (opt) { /* Validate 'opt' */
case OS_OPT_PEND_FLAG_CLR_ALL:
case OS_OPT_PEND_FLAG_CLR_ANY:
case OS_OPT_PEND_FLAG_SET_ALL:
case OS_OPT_PEND_FLAG_SET_ANY:
case OS_OPT_PEND_FLAG_CLR_ALL | OS_OPT_PEND_FLAG_CONSUME:
case OS_OPT_PEND_FLAG_CLR_ANY | OS_OPT_PEND_FLAG_CONSUME:
case OS_OPT_PEND_FLAG_SET_ALL | OS_OPT_PEND_FLAG_CONSUME:
case OS_OPT_PEND_FLAG_SET_ANY | OS_OPT_PEND_FLAG_CONSUME:
case OS_OPT_PEND_FLAG_CLR_ALL | OS_OPT_PEND_NON_BLOCKING:
case OS_OPT_PEND_FLAG_CLR_ANY | OS_OPT_PEND_NON_BLOCKING:
case OS_OPT_PEND_FLAG_SET_ALL | OS_OPT_PEND_NON_BLOCKING:
case OS_OPT_PEND_FLAG_SET_ANY | OS_OPT_PEND_NON_BLOCKING:
case OS_OPT_PEND_FLAG_CLR_ALL | OS_OPT_PEND_FLAG_CONSUME | OS_OPT_PEND_NON_BLOCKING:
case OS_OPT_PEND_FLAG_CLR_ANY | OS_OPT_PEND_FLAG_CONSUME | OS_OPT_PEND_NON_BLOCKING:
case OS_OPT_PEND_FLAG_SET_ALL | OS_OPT_PEND_FLAG_CONSUME | OS_OPT_PEND_NON_BLOCKING:
case OS_OPT_PEND_FLAG_SET_ANY | OS_OPT_PEND_FLAG_CONSUME | OS_OPT_PEND_NON_BLOCKING:
break;
default:
*p_err = OS_ERR_OPT_INVALID;
return ((OS_OBJ_QTY)0);
}
#endif
#if OS_CFG_OBJ_TYPE_CHK_EN > 0u
if (p_grp->Type != OS_OBJ_TYPE_FLAG) { /* Validate that we are pointing at an event flag */
*p_err = OS_ERR_OBJ_TYPE;
return ((OS_FLAGS)0);
}
#endif
//取反标志位检查
if ((opt & OS_OPT_PEND_FLAG_CONSUME) != (OS_OPT)0) { /* See if we need to consume the flags */
consume = DEF_TRUE;
} else {
consume = DEF_FALSE;
}
if (p_ts != (CPU_TS *)0) {
*p_ts = (CPU_TS)0; /* Initialize the returned timestamp */
}
//取出模式标志位
mode = opt & OS_OPT_PEND_FLAG_MASK;
CPU_CRITICAL_ENTER();
switch (mode) {
case OS_OPT_PEND_FLAG_SET_ALL: /* See if all required flags are set */
// 取出事件标志组的特定位
flags_rdy = (OS_FLAGS)(p_grp->Flags & flags); /* Extract only the bits we want */
//满足设定满足条件？
if (flags_rdy == flags) { /* Must match ALL the bits that we want */
//是，是否取反
if (consume == DEF_TRUE) { /* See if we need to consume the flags */
//是
p_grp->Flags &= ~flags_rdy; /* Clear ONLY the flags that we wanted */
}
//设定任务控制块的标志变量
OSTCBCurPtr->FlagsRdy = flags_rdy; /* Save flags that were ready */
if (p_ts != (CPU_TS *)0) {
*p_ts = p_grp->TS;
}
CPU_CRITICAL_EXIT(); /* Yes, condition met, return to caller */
*p_err = OS_ERR_NONE;
return (flags_rdy);
}
//不满足
else { /* Block task until events occur or timeout */
if ((opt & OS_OPT_PEND_NON_BLOCKING) != (OS_OPT)0) {
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_PEND_WOULD_BLOCK; /* Specified non-blocking so task would block */
return ((OS_FLAGS)0);
} else { /* Specified blocking so check is scheduler is locked */
if (OSSchedLockNestingCtr > (OS_NESTING_CTR)0) { /* See if called with scheduler locked ... */
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_SCHED_LOCKED; /* ... can't PEND when locked */
return ((OS_FLAGS)0);
}
}
/* Lock the scheduler/re-enable interrupts */
OS_CRITICAL_ENTER_CPU_EXIT();
//事件标志组阻塞任务
OS_FlagBlock(&pend_data,
p_grp,
flags,
opt,
timeout);
OS_CRITICAL_EXIT_NO_SCHED();
}
break;
//任何位满足模式
case OS_OPT_PEND_FLAG_SET_ANY:
flags_rdy = (OS_FLAGS)(p_grp->Flags & flags); /* Extract only the bits we want */
//有任何一个位满足
if (flags_rdy != (OS_FLAGS)0) { /* See if any flag set */
if (consume == DEF_TRUE) { /* See if we need to consume the flags */
p_grp->Flags &= ~flags_rdy; /* Clear ONLY the flags that we got */
}
OSTCBCurPtr->FlagsRdy = flags_rdy; /* Save flags that were ready */
if (p_ts != (CPU_TS *)0) {
*p_ts = p_grp->TS;
}
CPU_CRITICAL_EXIT(); /* Yes, condition met, return to caller */
*p_err = OS_ERR_NONE;
return (flags_rdy);
} else { /* Block task until events occur or timeout */
if ((opt & OS_OPT_PEND_NON_BLOCKING) != (OS_OPT)0) {
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_PEND_WOULD_BLOCK; /* Specified non-blocking so task would block */
return ((OS_FLAGS)0);
} else { /* Specified blocking so check is scheduler is locked */
if (OSSchedLockNestingCtr > (OS_NESTING_CTR)0) { /* See if called with scheduler locked ... */
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_SCHED_LOCKED; /* ... can't PEND when locked */
return ((OS_FLAGS)0);
}
}
/* Lock the scheduler/re-enable interrupts */
OS_CRITICAL_ENTER_CPU_EXIT();
OS_FlagBlock(&pend_data,
p_grp,
flags,
opt,
timeout);
OS_CRITICAL_EXIT_NO_SCHED();
}
break;
//为复位标志事件发生模式，其他同上
#if OS_CFG_FLAG_MODE_CLR_EN > 0u
case OS_OPT_PEND_FLAG_CLR_ALL: /* See if all required flags are cleared */
//注意这里的区别
flags_rdy = (OS_FLAGS)(~p_grp->Flags & flags); /* Extract only the bits we want */
if (flags_rdy == flags) { /* Must match ALL the bits that we want */
if (consume == DEF_TRUE) { /* See if we need to consume the flags */
//这里也有不同
p_grp->Flags |= flags_rdy; /* Set ONLY the flags that we wanted */
}
OSTCBCurPtr->FlagsRdy = flags_rdy; /* Save flags that were ready */
if (p_ts != (CPU_TS *)0) {
*p_ts = p_grp->TS;
}
CPU_CRITICAL_EXIT(); /* Yes, condition met, return to caller */
*p_err = OS_ERR_NONE;
return (flags_rdy);
} else { /* Block task until events occur or timeout */
if ((opt & OS_OPT_PEND_NON_BLOCKING) != (OS_OPT)0) {
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_PEND_WOULD_BLOCK; /* Specified non-blocking so task would block */
return ((OS_FLAGS)0);
} else { /* Specified blocking so check is scheduler is locked */
if (OSSchedLockNestingCtr > (OS_NESTING_CTR)0) { /* See if called with scheduler locked ... */
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_SCHED_LOCKED; /* ... can't PEND when locked */
return ((OS_FLAGS)0);
}
}
OS_CRITICAL_ENTER_CPU_EXIT(); /* Lock the scheduler/re-enable interrupts */
OS_FlagBlock(&pend_data,
p_grp,
flags,
opt,
timeout);
OS_CRITICAL_EXIT_NO_SCHED();
}
break;
case OS_OPT_PEND_FLAG_CLR_ANY:
flags_rdy = (OS_FLAGS)(~p_grp->Flags & flags); /* Extract only the bits we want */
if (flags_rdy != (OS_FLAGS)0) { /* See if any flag cleared */
if (consume == DEF_TRUE) { /* See if we need to consume the flags */
p_grp->Flags |= flags_rdy; /* Set ONLY the flags that we got */
}
OSTCBCurPtr->FlagsRdy = flags_rdy; /* Save flags that were ready */
if (p_ts != (CPU_TS *)0) {
*p_ts = p_grp->TS;
}
CPU_CRITICAL_EXIT(); /* Yes, condition met, return to caller */
*p_err = OS_ERR_NONE;
return (flags_rdy);
} else { /* Block task until events occur or timeout */
if ((opt & OS_OPT_PEND_NON_BLOCKING) != (OS_OPT)0) {
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_PEND_WOULD_BLOCK; /* Specified non-blocking so task would block */
return ((OS_FLAGS)0);
} else { /* Specified blocking so check is scheduler is locked */
if (OSSchedLockNestingCtr > (OS_NESTING_CTR)0) { /* See if called with scheduler locked ... */
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_SCHED_LOCKED; /* ... can't PEND when locked */
return ((OS_FLAGS)0);
}
}
OS_CRITICAL_ENTER_CPU_EXIT(); /* Lock the scheduler/re-enable interrupts */
OS_FlagBlock(&pend_data,
p_grp,
flags,
opt,
timeout);
OS_CRITICAL_EXIT_NO_SCHED();
}
break;
#endif
default:
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_OPT_INVALID;
return ((OS_FLAGS)0);
}
OSSched(); /* Find next HPT ready to run */
//任务重新恢复执行入口处，后面判断事件满足的原因
CPU_CRITICAL_ENTER();
switch (OSTCBCurPtr->PendStatus) {
case OS_STATUS_PEND_OK: /* We got the vent flags */
if (p_ts != (CPU_TS *)0) {
*p_ts = OSTCBCurPtr->TS;
}
*p_err = OS_ERR_NONE;
break;
case OS_STATUS_PEND_ABORT: /* Indicate that we aborted */
if (p_ts != (CPU_TS *)0) {
*p_ts = OSTCBCurPtr->TS;
}
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_PEND_ABORT;
break;
case OS_STATUS_PEND_TIMEOUT: /* Indicate that we didn't get semaphore within timeout */
if (p_ts != (CPU_TS *)0) {
*p_ts = (CPU_TS )0;
}
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_TIMEOUT;
break;
case OS_STATUS_PEND_DEL: /* Indicate that object pended on has been deleted */
if (p_ts != (CPU_TS *)0) {
*p_ts = OSTCBCurPtr->TS;
}
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_OBJ_DEL;
break;
default:
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_STATUS_INVALID;
break;
}
if (*p_err != OS_ERR_NONE) {
return ((OS_FLAGS)0);
}
//任务在POST时已经将事件标志组放入任务控制块的FlagsRdy里了
flags_rdy = OSTCBCurPtr->FlagsRdy;
//进行相应位取反
if (consume == DEF_TRUE) { /* See if we need to consume the flags */
switch (mode) {
case OS_OPT_PEND_FLAG_SET_ALL:
case OS_OPT_PEND_FLAG_SET_ANY: /* Clear ONLY the flags we got */
p_grp->Flags &= ~flags_rdy;
break;
#if OS_CFG_FLAG_MODE_CLR_EN > 0u
case OS_OPT_PEND_FLAG_CLR_ALL:
case OS_OPT_PEND_FLAG_CLR_ANY: /* Set ONLY the flags we got */
p_grp->Flags |= flags_rdy;
break;
#endif
default:
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_OPT_INVALID;
return ((OS_FLAGS)0);
}
}
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_NONE; /* Event(s) must have occurred */
return (flags_rdy);
}这是事件标志组特有的阻塞函数，其实原理还是一样就是多了一部分向任务控制块写入的操作。
************************************************************************************************************************
* SUSPEND TASK UNTIL EVENT FLAG(s) RECEIVED OR TIMEOUT OCCURS
*
* Description: This function is internal to uC/OS-III and is used to put a task to sleep until the desired
* event flag bit(s) are set.
*
* Arguments : p_pend_data is a pointer to an object used to link the task being blocked to the list of task(s)
* ----------- pending on the desired event flag group.
*
* p_grp is a pointer to the desired event flag group.
* -----
*
* flags Is a bit pattern indicating which bit(s) (i.e. flags) you wish to check.
* The bits you want are specified by setting the corresponding bits in
* 'flags'. e.g. if your application wants to wait for bits 0 and 1 then
* 'flags' would contain 0x03.
*
* opt specifies whether you want ALL bits to be set/cleared or ANY of the bits
* to be set/cleared.
* You can specify the following argument:
*
* OS_OPT_PEND_FLAG_CLR_ALL You will check ALL bits in 'mask' to be clear (0)
* OS_OPT_PEND_FLAG_CLR_ANY You will check ANY bit in 'mask' to be clear (0)
* OS_OPT_PEND_FLAG_SET_ALL You will check ALL bits in 'mask' to be set (1)
* OS_OPT_PEND_FLAG_SET_ANY You will check ANY bit in 'mask' to be set (1)
*
* timeout is the desired amount of time that the task will wait for the event flag
* bit(s) to be set.
*
* Returns : none
*
* Note(s) : This function is INTERNAL to uC/OS-III and your application should not call it.
************************************************************************************************************************
*/
void OS_FlagBlock (OS_PEND_DATA *p_pend_data,
OS_FLAG_GRP *p_grp,
OS_FLAGS flags,
OS_OPT opt,
OS_TICK timeout)
{
//事件标志组的特别操作
OSTCBCurPtr->FlagsPend = flags; /* Save the flags that we need to wait for */
OSTCBCurPtr->FlagsOpt = opt; /* Save the type of wait we are doing */
OSTCBCurPtr->FlagsRdy = (OS_FLAGS)0;
//加入内核对象等待队列，并阻塞任务
OS_Pend(p_pend_data,
(OS_PEND_OBJ *)((void *)p_grp),
OS_TASK_PEND_ON_FLAG,
timeout);
}3.对应到发布事件组，同样是如此，对应那个几个位，怎么操作，最后就是post函数内的共性就是检查等待列表里是否有满足条件的任务这里区分任何一个还是所有位满足，如果满足就将任务加入到就绪列表里。这里需要注意理解的就是函数内关于位操作的一些写法。
************************************************************************************************************************
* POST EVENT FLAG BIT(S)
*
* Description: This function is called to set or clear some bits in an event flag group. The bits to set or clear are
* specified by a 'bit mask'.
*
* Arguments : p_grp is a pointer to the desired event flag group.
*
* flags If 'opt' (see below) is OS_OPT_POST_FLAG_SET, each bit that is set in 'flags' will
* set the corresponding bit in the event flag group. e.g. to set bits 0, 4
* and 5 you would set 'flags' to:
*
* 0x31 (note, bit 0 is least significant bit)
*
* If 'opt' (see below) is OS_OPT_POST_FLAG_CLR, each bit that is set in 'flags' will
* CLEAR the corresponding bit in the event flag group. e.g. to clear bits 0,
* 4 and 5 you would specify 'flags' as:
*
* 0x31 (note, bit 0 is least significant bit)
*
* opt indicates whether the flags will be:
*
* OS_OPT_POST_FLAG_SET set
* OS_OPT_POST_FLAG_CLR cleared
*
* you can also 'add' OS_OPT_POST_NO_SCHED to prevent the scheduler from being called.
*
* p_err is a pointer to an error code and can be:
*
* OS_ERR_NONE The call was successful
* OS_ERR_OBJ_PTR_NULL You passed a NULL pointer
* OS_ERR_OBJ_TYPE You are not pointing to an event flag group
* OS_ERR_OPT_INVALID You specified an invalid option
*
* Returns : the new value of the event flags bits that are still set.
*
* Note(s) : 1) The execution time of this function depends on the number of tasks waiting on the event flag group.
************************************************************************************************************************
*/
OS_FLAGS OSFlagPost (OS_FLAG_GRP *p_grp,
OS_FLAGS flags,
OS_OPT opt,
OS_ERR *p_err)
{
OS_FLAGS flags_cur;
CPU_TS ts;
#ifdef OS_SAFETY_CRITICAL
if (p_err == (OS_ERR *)0) {
OS_SAFETY_CRITICAL_EXCEPTION();
return ((OS_FLAGS)0);
}
#endif
#if OS_CFG_ARG_CHK_EN > 0u
if (p_grp == (OS_FLAG_GRP *)0) { /* Validate 'p_grp' */
*p_err = OS_ERR_OBJ_PTR_NULL;
return ((OS_FLAGS)0);
}
switch (opt) { /* Validate 'opt' */
case OS_OPT_POST_FLAG_SET:
case OS_OPT_POST_FLAG_CLR:
case OS_OPT_POST_FLAG_SET | OS_OPT_POST_NO_SCHED:
case OS_OPT_POST_FLAG_CLR | OS_OPT_POST_NO_SCHED:
break;
default:
*p_err = OS_ERR_OPT_INVALID;
return ((OS_FLAGS)0);
}
#endif
#if OS_CFG_OBJ_TYPE_CHK_EN > 0u
if (p_grp->Type != OS_OBJ_TYPE_FLAG) { /* Make sure we are pointing to an event flag grp */
*p_err = OS_ERR_OBJ_TYPE;
return ((OS_FLAGS)0);
}
#endif
ts = OS_TS_GET(); /* Get timestamp */
//延迟发布
#if OS_CFG_ISR_POST_DEFERRED_EN > 0u
if (OSIntNestingCtr > (OS_NESTING_CTR)0) { /* See if called from an ISR */
OS_IntQPost((OS_OBJ_TYPE)OS_OBJ_TYPE_FLAG, /* Post to ISR queue */
(void *)p_grp,
(void *)0,
(OS_MSG_SIZE)0,
(OS_FLAGS )flags,
(OS_OPT )opt,
(CPU_TS )ts,
(OS_ERR *)p_err);
return ((OS_FLAGS)0);
}
#endif
//立即发布
flags_cur = OS_FlagPost(p_grp,
flags,
opt,
ts,
p_err);
return (flags_cur);
}************************************************************************************************************************
* POST EVENT FLAG BIT(S)
*
* Description: This function is called to set or clear some bits in an event flag group. The bits to set or clear are
* specified by a 'bit mask'.
*
* Arguments : p_grp is a pointer to the desired event flag group.
*
* flags If 'opt' (see below) is OS_OPT_POST_FLAG_SET, each bit that is set in 'flags' will
* set the corresponding bit in the event flag group. e.g. to set bits 0, 4
* and 5 you would set 'flags' to:
*
* 0x31 (note, bit 0 is least significant bit)
*
* If 'opt' (see below) is OS_OPT_POST_FLAG_CLR, each bit that is set in 'flags' will
* CLEAR the corresponding bit in the event flag group. e.g. to clear bits 0,
* 4 and 5 you would specify 'flags' as:
*
* 0x31 (note, bit 0 is least significant bit)
*
* opt indicates whether the flags will be:
*
* OS_OPT_POST_FLAG_SET set
* OS_OPT_POST_FLAG_CLR cleared
*
* you can also 'add' OS_OPT_POST_NO_SCHED to prevent the scheduler from being called.
*
* ts is the timestamp of the post
*
* p_err is a pointer to an error code and can be:
*
* OS_ERR_NONE The call was successful
* OS_ERR_OBJ_PTR_NULL You passed a NULL pointer
* OS_ERR_OBJ_TYPE You are not pointing to an event flag group
* OS_ERR_OPT_INVALID You specified an invalid option
*
* Returns : the new value of the event flags bits that are still set.
*
* Note(s) : 1) The execution time of this function depends on the number of tasks waiting on the event flag group.
************************************************************************************************************************
*/
OS_FLAGS OS_FlagPost (OS_FLAG_GRP *p_grp,
OS_FLAGS flags,
OS_OPT opt,
CPU_TS ts,
OS_ERR *p_err)
{
OS_FLAGS flags_cur;
OS_FLAGS flags_rdy;
OS_OPT mode;
OS_PEND_DATA *p_pend_data;
OS_PEND_DATA *p_pend_data_next;
OS_PEND_LIST *p_pend_list;
OS_TCB *p_tcb;
CPU_SR_ALLOC();
CPU_CRITICAL_ENTER();
switch (opt) {
case OS_OPT_POST_FLAG_SET:
case OS_OPT_POST_FLAG_SET | OS_OPT_POST_NO_SCHED:
p_grp->Flags |= flags; /* Set the flags specified in the group */
break;
case OS_OPT_POST_FLAG_CLR:
case OS_OPT_POST_FLAG_CLR | OS_OPT_POST_NO_SCHED:
p_grp->Flags &= ~flags; /* Clear the flags specified in the group */
break;
default:
CPU_CRITICAL_EXIT(); /* INVALID option */
*p_err = OS_ERR_OPT_INVALID;
return ((OS_FLAGS)0);
}
//记录发布时间
p_grp->TS = ts;
//取出等待队列
p_pend_list = &p_grp->PendList;
if (p_pend_list->NbrEntries == 0u) { /* Any task waiting on event flag group? */
CPU_CRITICAL_EXIT(); /* No */
*p_err = OS_ERR_NONE;
return (p_grp->Flags);
}
OS_CRITICAL_ENTER_CPU_EXIT();
p_pend_data = p_pend_list->HeadPtr;
p_tcb = p_pend_data->TCBPtr;
//依次查找等待队列，找出满足条件的任务
while (p_tcb != (OS_TCB *)0) { /* Go through all tasks waiting on event flag(s) */
p_pend_data_next = p_pend_data->NextPtr;
mode = p_tcb->FlagsOpt & OS_OPT_PEND_FLAG_MASK;
switch (mode) {
case OS_OPT_PEND_FLAG_SET_ALL: /* See if all req. flags are set for current node */
flags_rdy = (OS_FLAGS)(p_grp->Flags & p_tcb->FlagsPend);
if (flags_rdy == p_tcb->FlagsPend) {
//将任务就绪 ，同时从等待队列删除，下同
OS_FlagTaskRdy(p_tcb, /* Make task RTR, event(s) Rx'd */
flags_rdy,
ts);
}
break;
case OS_OPT_PEND_FLAG_SET_ANY: /* See if any flag set */
flags_rdy = (OS_FLAGS)(p_grp->Flags & p_tcb->FlagsPend);
if (flags_rdy != (OS_FLAGS)0) {
OS_FlagTaskRdy(p_tcb, /* Make task RTR, event(s) Rx'd */
flags_rdy,
ts);
}
break;
#if OS_CFG_FLAG_MODE_CLR_EN > 0u
case OS_OPT_PEND_FLAG_CLR_ALL: /* See if all req. flags are set for current node */
flags_rdy = (OS_FLAGS)(~p_grp->Flags & p_tcb->FlagsPend);
if (flags_rdy == p_tcb->FlagsPend) {
OS_FlagTaskRdy(p_tcb, /* Make task RTR, event(s) Rx'd */
flags_rdy,
ts);
}
break;
case OS_OPT_PEND_FLAG_CLR_ANY: /* See if any flag set */
flags_rdy = (OS_FLAGS)(~p_grp->Flags & p_tcb->FlagsPend);
if (flags_rdy != (OS_FLAGS)0) {
OS_FlagTaskRdy(p_tcb, /* Make task RTR, event(s) Rx'd */
flags_rdy,
ts);
}
break;
#endif
default:
OS_CRITICAL_EXIT();
*p_err = OS_ERR_FLAG_PEND_OPT;
return ((OS_FLAGS)0);
}
p_pend_data = p_pend_data_next; /* Point to next task waiting for event flag(s) */
if (p_pend_data != (OS_PEND_DATA *)0) {
p_tcb = p_pend_data->TCBPtr;
} else {
p_tcb = (OS_TCB *)0;
}
}
OS_CRITICAL_EXIT_NO_SCHED();
if ((opt & OS_OPT_POST_NO_SCHED) == (OS_OPT)0) {
OSSched();
}
CPU_CRITICAL_ENTER();
flags_cur = p_grp->Flags;
CPU_CRITICAL_EXIT();
*p_err = OS_ERR_NONE;
return (flags_cur);
}这是事件标志组特有的任务就绪函数，其实原理还是一样还是多了一部分向任务控制块写入的操作。
************************************************************************************************************************
* MAKE TASK READY-TO-RUN, EVENT(s) OCCURRED
*
* Description: This function is internal to uC/OS-III and is used to make a task ready-to-run because the desired event
* flag bits have been set.
*
* Arguments : p_tcb is a pointer to the OS_TCB of the task to remove
* -----
*
* flags_rdy contains the bit pattern of the event flags that cause the task to become ready-to-run.
*
* ts is a timestamp associated with the post
*
* Returns : none
*
* Note(s) : This function is INTERNAL to uC/OS-III and your application should not call it.
************************************************************************************************************************
*/
void OS_FlagTaskRdy (OS_TCB *p_tcb,
OS_FLAGS flags_rdy,
CPU_TS ts)
{
p_tcb->FlagsRdy = flags_rdy;
p_tcb->PendStatus = OS_STATUS_PEND_OK; /* Clear pend status */
p_tcb->PendOn = OS_TASK_PEND_ON_NOTHING; /* Indicate no longer pending */
p_tcb->TS = ts;
switch (p_tcb->TaskState) {
case OS_TASK_STATE_RDY:
case OS_TASK_STATE_DLY:
case OS_TASK_STATE_DLY_SUSPENDED:
case OS_TASK_STATE_SUSPENDED:
break;
case OS_TASK_STATE_PEND:
case OS_TASK_STATE_PEND_TIMEOUT:
OS_TaskRdy(p_tcb);
p_tcb->TaskState = OS_TASK_STATE_RDY;
break;
case OS_TASK_STATE_PEND_SUSPENDED:
case OS_TASK_STATE_PEND_TIMEOUT_SUSPENDED:
p_tcb->TaskState = OS_TASK_STATE_SUSPENDED;
break;
default:
break;
}
OS_PendListRemove(p_tcb);
}
#endif时间标志组 最重要的是将位和任务联系起来，同时继续了信号量的等待任务管理。