| ------------------------------------------------------------------------------ |
| -- -- |
| -- GNU ADA RUN-TIME LIBRARY (GNARL) COMPONENTS -- |
| -- -- |
| -- S Y S T E M . T A S K _ P R I M I T I V E S .O P E R A T I O N S -- |
| -- -- |
| -- S p e c -- |
| -- -- |
| -- Copyright (C) 1992-2002, Free Software Foundation, Inc. -- |
| -- -- |
| -- GNARL is free software; you can redistribute it and/or modify it under -- |
| -- terms of the GNU General Public License as published by the Free Soft- -- |
| -- ware Foundation; either version 2, or (at your option) any later ver- -- |
| -- sion. GNARL is distributed in the hope that it will be useful, but WITH- -- |
| -- OUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY -- |
| -- or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License -- |
| -- for more details. You should have received a copy of the GNU General -- |
| -- Public License distributed with GNARL; see file COPYING. If not, write -- |
| -- to the Free Software Foundation, 59 Temple Place - Suite 330, Boston, -- |
| -- MA 02111-1307, USA. -- |
| -- -- |
| -- As a special exception, if other files instantiate generics from this -- |
| -- unit, or you link this unit with other files to produce an executable, -- |
| -- this unit does not by itself cause the resulting executable to be -- |
| -- covered by the GNU General Public License. This exception does not -- |
| -- however invalidate any other reasons why the executable file might be -- |
| -- covered by the GNU Public License. -- |
| -- -- |
| -- GNARL was developed by the GNARL team at Florida State University. -- |
| -- Extensive contributions were provided by Ada Core Technologies, Inc. -- |
| -- -- |
| ------------------------------------------------------------------------------ |
| |
| -- This package contains all the GNULL primitives that interface directly |
| -- with the underlying OS. |
| |
| with System.Parameters; |
| -- used for Size_Type |
| |
| with System.Tasking; |
| -- used for Task_ID |
| |
| with System.OS_Interface; |
| -- used for Thread_Id |
| |
| package System.Task_Primitives.Operations is |
| |
| pragma Elaborate_Body; |
| package ST renames System.Tasking; |
| package OSI renames System.OS_Interface; |
| |
| procedure Initialize (Environment_Task : ST.Task_ID); |
| pragma Inline (Initialize); |
| -- This must be called once, before any other subprograms of this |
| -- package are called. |
| |
| procedure Create_Task |
| (T : ST.Task_ID; |
| Wrapper : System.Address; |
| Stack_Size : System.Parameters.Size_Type; |
| Priority : System.Any_Priority; |
| Succeeded : out Boolean); |
| pragma Inline (Create_Task); |
| -- Create a new low-level task with ST.Task_ID T and place other needed |
| -- information in the ATCB. |
| -- |
| -- A new thread of control is created, with a stack of at least Stack_Size |
| -- storage units, and the procedure Wrapper is called by this new thread |
| -- of control. If Stack_Size = Unspecified_Storage_Size, choose a default |
| -- stack size; this may be effectively "unbounded" on some systems. |
| -- |
| -- The newly created low-level task is associated with the ST.Task_ID T |
| -- such that any subsequent call to Self from within the context of the |
| -- low-level task returns T. |
| -- |
| -- The caller is responsible for ensuring that the storage of the Ada |
| -- task control block object pointed to by T persists for the lifetime |
| -- of the new task. |
| -- |
| -- Succeeded is set to true unless creation of the task failed, |
| -- as it may if there are insufficient resources to create another task. |
| |
| procedure Enter_Task (Self_ID : ST.Task_ID); |
| pragma Inline (Enter_Task); |
| -- Initialize data structures specific to the calling task. |
| -- Self must be the ID of the calling task. |
| -- It must be called (once) by the task immediately after creation, |
| -- while abortion is still deferred. |
| -- The effects of other operations defined below are not defined |
| -- unless the caller has previously called Initialize_Task. |
| |
| procedure Exit_Task; |
| pragma Inline (Exit_Task); |
| -- Destroy the thread of control. |
| -- Self must be the ID of the calling task. |
| -- The effects of further calls to operations defined below |
| -- on the task are undefined thereafter. |
| |
| function New_ATCB (Entry_Num : ST.Task_Entry_Index) return ST.Task_ID; |
| pragma Inline (New_ATCB); |
| -- Allocate a new ATCB with the specified number of entries. |
| |
| procedure Initialize_TCB (Self_ID : ST.Task_ID; Succeeded : out Boolean); |
| pragma Inline (Initialize_TCB); |
| -- Initialize all fields of the TCB |
| |
| procedure Finalize_TCB (T : ST.Task_ID); |
| pragma Inline (Finalize_TCB); |
| -- Finalizes Private_Data of ATCB, and then deallocates it. |
| -- This is also responsible for recovering any storage or other resources |
| -- that were allocated by Create_Task (the one in this package). |
| -- This should only be called from Free_Task. |
| -- After it is called there should be no further |
| -- reference to the ATCB that corresponds to T. |
| |
| procedure Abort_Task (T : ST.Task_ID); |
| pragma Inline (Abort_Task); |
| -- Abort the task specified by T (the target task). This causes |
| -- the target task to asynchronously raise Abort_Signal if |
| -- abort is not deferred, or if it is blocked on an interruptible |
| -- system call. |
| -- |
| -- precondition: |
| -- the calling task is holding T's lock and has abort deferred |
| -- |
| -- postcondition: |
| -- the calling task is holding T's lock and has abort deferred. |
| |
| -- ??? modify GNARL to skip wakeup and always call Abort_Task |
| |
| function Self return ST.Task_ID; |
| pragma Inline (Self); |
| -- Return a pointer to the Ada Task Control Block of the calling task. |
| |
| type Lock_Level is |
| (PO_Level, |
| Global_Task_Level, |
| RTS_Lock_Level, |
| ATCB_Level); |
| -- Type used to describe kind of lock for second form of Initialize_Lock |
| -- call specified below. |
| -- See locking rules in System.Tasking (spec) for more details. |
| |
| procedure Initialize_Lock (Prio : System.Any_Priority; L : access Lock); |
| procedure Initialize_Lock (L : access RTS_Lock; Level : Lock_Level); |
| pragma Inline (Initialize_Lock); |
| -- Initialize a lock object. |
| -- |
| -- For Lock, Prio is the ceiling priority associated with the lock. |
| -- For RTS_Lock, the ceiling is implicitly Priority'Last. |
| -- |
| -- If the underlying system does not support priority ceiling |
| -- locking, the Prio parameter is ignored. |
| -- |
| -- The effect of either initialize operation is undefined unless L |
| -- is a lock object that has not been initialized, or which has been |
| -- finalized since it was last initialized. |
| -- |
| -- The effects of the other operations on lock objects |
| -- are undefined unless the lock object has been initialized |
| -- and has not since been finalized. |
| -- |
| -- Initialization of the per-task lock is implicit in Create_Task. |
| -- |
| -- These operations raise Storage_Error if a lack of storage is detected. |
| |
| procedure Finalize_Lock (L : access Lock); |
| procedure Finalize_Lock (L : access RTS_Lock); |
| pragma Inline (Finalize_Lock); |
| -- Finalize a lock object, freeing any resources allocated by the |
| -- corresponding Initialize_Lock operation. |
| |
| procedure Write_Lock (L : access Lock; Ceiling_Violation : out Boolean); |
| procedure Write_Lock (L : access RTS_Lock; Global_Lock : Boolean := False); |
| procedure Write_Lock (T : ST.Task_ID); |
| pragma Inline (Write_Lock); |
| -- Lock a lock object for write access. After this operation returns, |
| -- the calling task holds write permission for the lock object. No other |
| -- Write_Lock or Read_Lock operation on the same lock object will return |
| -- until this task executes an Unlock operation on the same object. The |
| -- effect is undefined if the calling task already holds read or write |
| -- permission for the lock object L. |
| -- |
| -- For the operation on Lock, Ceiling_Violation is set to true iff the |
| -- operation failed, which will happen if there is a priority ceiling |
| -- violation. |
| -- |
| -- For the operation on RTS_Lock, Global_Lock should be set to True |
| -- if L is a global lock (Single_RTS_Lock, Global_Task_Lock). |
| -- |
| -- For the operation on ST.Task_ID, the lock is the special lock object |
| -- associated with that task's ATCB. This lock has effective ceiling |
| -- priority high enough that it is safe to call by a task with any |
| -- priority in the range System.Priority. It is implicitly initialized |
| -- by task creation. The effect is undefined if the calling task already |
| -- holds T's lock, or has interrupt-level priority. Finalization of the |
| -- per-task lock is implicit in Exit_Task. |
| |
| procedure Read_Lock (L : access Lock; Ceiling_Violation : out Boolean); |
| pragma Inline (Read_Lock); |
| -- Lock a lock object for read access. After this operation returns, |
| -- the calling task has non-exclusive read permission for the logical |
| -- resources that are protected by the lock. No other Write_Lock operation |
| -- on the same object will return until this task and any other tasks with |
| -- read permission for this lock have executed Unlock operation(s) on the |
| -- lock object. A Read_Lock for a lock object may return immediately while |
| -- there are tasks holding read permission, provided there are no tasks |
| -- holding write permission for the object. The effect is undefined if |
| -- the calling task already holds read or write permission for L. |
| -- |
| -- Alternatively: An implementation may treat Read_Lock identically to |
| -- Write_Lock. This simplifies the implementation, but reduces the level |
| -- of concurrency that can be achieved. |
| -- |
| -- Note that Read_Lock is not defined for RT_Lock and ST.Task_ID. |
| -- That is because (1) so far Read_Lock has always been implemented |
| -- the same as Write_Lock, (2) most lock usage inside the RTS involves |
| -- potential write access, and (3) implementations of priority ceiling |
| -- locking that make a reader-writer distinction have higher overhead. |
| |
| procedure Unlock (L : access Lock); |
| procedure Unlock (L : access RTS_Lock; Global_Lock : Boolean := False); |
| procedure Unlock (T : ST.Task_ID); |
| pragma Inline (Unlock); |
| -- Unlock a locked lock object. |
| -- |
| -- The effect is undefined unless the calling task holds read or write |
| -- permission for the lock L, and L is the lock object most recently |
| -- locked by the calling task for which the calling task still holds |
| -- read or write permission. (That is, matching pairs of Lock and Unlock |
| -- operations on each lock object must be properly nested.) |
| |
| -- For the operation on RTS_Lock, Global_Lock should be set to True |
| -- if L is a global lock (Single_RTS_Lock, Global_Task_Lock). |
| -- |
| -- Note that Write_Lock for RTS_Lock does not have an out-parameter. |
| -- RTS_Locks are used in situations where we have not made provision |
| -- for recovery from ceiling violations. We do not expect them to |
| -- occur inside the runtime system, because all RTS locks have ceiling |
| -- Priority'Last. |
| |
| -- There is one way there can be a ceiling violation. |
| -- That is if the runtime system is called from a task that is |
| -- executing in the Interrupt_Priority range. |
| |
| -- It is not clear what to do about ceiling violations due |
| -- to RTS calls done at interrupt priority. In general, it |
| -- is not acceptable to give all RTS locks interrupt priority, |
| -- since that whould give terrible performance on systems where |
| -- this has the effect of masking hardware interrupts, though we |
| -- could get away with allowing Interrupt_Priority'last where we |
| -- are layered on an OS that does not allow us to mask interrupts. |
| -- Ideally, we would like to raise Program_Error back at the |
| -- original point of the RTS call, but this would require a lot of |
| -- detailed analysis and recoding, with almost certain performance |
| -- penalties. |
| |
| -- For POSIX systems, we considered just skipping setting a |
| -- priority ceiling on RTS locks. This would mean there is no |
| -- ceiling violation, but we would end up with priority inversions |
| -- inside the runtime system, resulting in failure to satisfy the |
| -- Ada priority rules, and possible missed validation tests. |
| -- This could be compensated-for by explicit priority-change calls |
| -- to raise the caller to Priority'Last whenever it first enters |
| -- the runtime system, but the expected overhead seems high, though |
| -- it might be lower than using locks with ceilings if the underlying |
| -- implementation of ceiling locks is an inefficient one. |
| |
| -- This issue should be reconsidered whenever we get around to |
| -- checking for calls to potentially blocking operations from |
| -- within protected operations. If we check for such calls and |
| -- catch them on entry to the OS, it may be that we can eliminate |
| -- the possibility of ceiling violations inside the RTS. For this |
| -- to work, we would have to forbid explicitly setting the priority |
| -- of a task to anything in the Interrupt_Priority range, at least. |
| -- We would also have to check that there are no RTS-lock operations |
| -- done inside any operations that are not treated as potentially |
| -- blocking. |
| |
| -- The latter approach seems to be the best, i.e. to check on entry |
| -- to RTS calls that may need to use locks that the priority is not |
| -- in the interrupt range. If there are RTS operations that NEED to |
| -- be called from interrupt handlers, those few RTS locks should then |
| -- be converted to PO-type locks, with ceiling Interrupt_Priority'Last. |
| |
| -- For now, we will just shut down the system if there is a |
| -- ceiling violation. |
| |
| procedure Yield (Do_Yield : Boolean := True); |
| pragma Inline (Yield); |
| -- Yield the processor. Add the calling task to the tail of the |
| -- ready queue for its active_priority. |
| -- The Do_Yield argument is only used in some very rare cases very |
| -- a yield should have an effect on a specific target and not on regular |
| -- ones. |
| |
| procedure Set_Priority |
| (T : ST.Task_ID; |
| Prio : System.Any_Priority; |
| Loss_Of_Inheritance : Boolean := False); |
| pragma Inline (Set_Priority); |
| -- Set the priority of the task specified by T to T.Current_Priority. |
| -- The priority set is what would correspond to the Ada concept of |
| -- "base priority" in the terms of the lower layer system, but |
| -- the operation may be used by the upper layer to implement |
| -- changes in "active priority" that are not due to lock effects. |
| -- The effect should be consistent with the Ada Reference Manual. |
| -- In particular, when a task lowers its priority due to the loss of |
| -- inherited priority, it goes at the head of the queue for its new |
| -- priority (RM D.2.2 par 9). |
| -- Loss_Of_Inheritance helps the underlying implementation to do it |
| -- right when the OS doesn't. |
| |
| function Get_Priority (T : ST.Task_ID) return System.Any_Priority; |
| pragma Inline (Get_Priority); |
| -- Returns the priority last set by Set_Priority for this task. |
| |
| function Monotonic_Clock return Duration; |
| pragma Inline (Monotonic_Clock); |
| -- Returns "absolute" time, represented as an offset |
| -- relative to "the Epoch", which is Jan 1, 1970. |
| -- This clock implementation is immune to the system's clock changes. |
| |
| function RT_Resolution return Duration; |
| pragma Inline (RT_Resolution); |
| -- Returns the resolution of the underlying clock used to implement |
| -- RT_Clock. |
| |
| ---------------- |
| -- Extensions -- |
| ---------------- |
| |
| -- Whoever calls either of the Sleep routines is responsible |
| -- for checking for pending aborts before the call. |
| -- Pending priority changes are handled internally. |
| |
| procedure Sleep |
| (Self_ID : ST.Task_ID; |
| Reason : System.Tasking.Task_States); |
| pragma Inline (Sleep); |
| -- Wait until the current task, T, is signaled to wake up. |
| -- |
| -- precondition: |
| -- The calling task is holding its own ATCB lock |
| -- and has abort deferred |
| -- |
| -- postcondition: |
| -- The calling task is holding its own ATCB lock |
| -- and has abort deferred. |
| |
| -- The effect is to atomically unlock T's lock and wait, so that another |
| -- task that is able to lock T's lock can be assured that the wait has |
| -- actually commenced, and that a Wakeup operation will cause the waiting |
| -- task to become ready for execution once again. When Sleep returns, |
| -- the waiting task will again hold its own ATCB lock. The waiting task |
| -- may become ready for execution at any time (that is, spurious wakeups |
| -- are permitted), but it will definitely become ready for execution when |
| -- a Wakeup operation is performed for the same task. |
| |
| procedure Timed_Sleep |
| (Self_ID : ST.Task_ID; |
| Time : Duration; |
| Mode : ST.Delay_Modes; |
| Reason : System.Tasking.Task_States; |
| Timedout : out Boolean; |
| Yielded : out Boolean); |
| -- Combination of Sleep (above) and Timed_Delay |
| |
| procedure Timed_Delay |
| (Self_ID : ST.Task_ID; |
| Time : Duration; |
| Mode : ST.Delay_Modes); |
| -- Implement the semantics of the delay statement. It is assumed that |
| -- the caller is not abort-deferred and does not hold any locks. |
| |
| procedure Wakeup |
| (T : ST.Task_ID; |
| Reason : System.Tasking.Task_States); |
| pragma Inline (Wakeup); |
| -- Wake up task T if it is waiting on a Sleep call (of ordinary |
| -- or timed variety), making it ready for execution once again. |
| -- If the task T is not waiting on a Sleep, the operation has no effect. |
| |
| function Environment_Task return ST.Task_ID; |
| pragma Inline (Environment_Task); |
| -- Return the task ID of the environment task |
| -- Consider putting this into a variable visible directly |
| -- by the rest of the runtime system. ??? |
| |
| function Get_Thread_Id (T : ST.Task_ID) return OSI.Thread_Id; |
| -- Return the thread id of the specified task |
| |
| function Is_Valid_Task return Boolean; |
| pragma Inline (Is_Valid_Task); |
| -- Does the calling thread have an ATCB? |
| |
| function Register_Foreign_Thread return ST.Task_ID; |
| -- Allocate and initialize a new ATCB for the current thread |
| |
| ----------------------- |
| -- RTS Entrance/Exit -- |
| ----------------------- |
| |
| -- Following two routines are used for possible operations needed |
| -- to be setup/cleared upon entrance/exit of RTS while maintaining |
| -- a single thread of control in the RTS. Since we intend these |
| -- routines to be used for implementing the Single_Lock RTS, |
| -- Lock_RTS should follow the first Defer_Abortion operation |
| -- entering RTS. In the same fashion Unlock_RTS should preceed |
| -- the last Undefer_Abortion exiting RTS. |
| -- |
| -- These routines also replace the functions Lock/Unlock_All_Tasks_List |
| |
| procedure Lock_RTS; |
| -- Take the global RTS lock. |
| |
| procedure Unlock_RTS; |
| -- Release the global RTS lock. |
| |
| -------------------- |
| -- Stack Checking -- |
| -------------------- |
| |
| -- Stack checking in GNAT is done using the concept of stack probes. A |
| -- stack probe is an operation that will generate a storage error if |
| -- an insufficient amount of stack space remains in the current task. |
| |
| -- The exact mechanism for a stack probe is target dependent. Typical |
| -- possibilities are to use a load from a non-existent page, a store |
| -- to a read-only page, or a comparison with some stack limit constant. |
| -- Where possible we prefer to use a trap on a bad page access, since |
| -- this has less overhead. The generation of stack probes is either |
| -- automatic if the ABI requires it (as on for example DEC Unix), or |
| -- is controlled by the gcc parameter -fstack-check. |
| |
| -- When we are using bad-page accesses, we need a bad page, called a |
| -- guard page, at the end of each task stack. On some systems, this |
| -- is provided automatically, but on other systems, we need to create |
| -- the guard page ourselves, and the procedure Stack_Guard is provided |
| -- for this purpose. |
| |
| procedure Stack_Guard (T : ST.Task_ID; On : Boolean); |
| -- Ensure guard page is set if one is needed and the underlying thread |
| -- system does not provide it. The procedure is as follows: |
| -- |
| -- 1. When we create a task adjust its size so a guard page can |
| -- safely be set at the bottom of the stack |
| -- |
| -- 2. When the thread is created (and its stack allocated by the |
| -- underlying thread system), get the stack base (and size, depending |
| -- how the stack is growing), and create the guard page taking care of |
| -- page boundaries issues. |
| -- |
| -- 3. When the task is destroyed, remove the guard page. |
| -- |
| -- If On is true then protect the stack bottom (i.e make it read only) |
| -- else unprotect it (i.e. On is True for the call when creating a task, |
| -- and False when a task is destroyed). |
| -- |
| -- The call to Stack_Guard has no effect if guard pages are not used on |
| -- the target, or if guard pages are automatically provided by the system. |
| |
| ----------------------------------------- |
| -- Runtime System Debugging Interfaces -- |
| ----------------------------------------- |
| |
| -- These interfaces have been added to assist in debugging the |
| -- tasking runtime system. |
| |
| function Check_Exit (Self_ID : ST.Task_ID) return Boolean; |
| pragma Inline (Check_Exit); |
| -- Check that the current task is holding only Global_Task_Lock. |
| |
| function Check_No_Locks (Self_ID : ST.Task_ID) return Boolean; |
| pragma Inline (Check_No_Locks); |
| -- Check that current task is holding no locks. |
| |
| function Suspend_Task |
| (T : ST.Task_ID; |
| Thread_Self : OSI.Thread_Id) |
| return Boolean; |
| -- Suspend a specific task when the underlying thread library provides |
| -- such functionality, unless the thread associated with T is Thread_Self. |
| -- Such functionality is needed by gdb on some targets (e.g VxWorks) |
| -- Return True is the operation is successful |
| |
| function Resume_Task |
| (T : ST.Task_ID; |
| Thread_Self : OSI.Thread_Id) |
| return Boolean; |
| -- Resume a specific task when the underlying thread library provides |
| -- such functionality, unless the thread associated with T is Thread_Self. |
| -- Such functionality is needed by gdb on some targets (e.g VxWorks) |
| -- Return True is the operation is successful |
| |
| end System.Task_Primitives.Operations; |