1 /* ---------------------------------------------------------------------------
3 * (c) The GHC Team, 1998-2006
5 * Asynchronous exceptions
7 * --------------------------------------------------------------------------*/
9 #include "PosixSource.h"
12 #include "sm/Storage.h"
15 #include "RaiseAsync.h"
20 #include "Profiling.h"
21 #if defined(mingw32_HOST_OS)
22 #include "win32/IOManager.h"
25 static void raiseAsync (Capability *cap,
27 StgClosure *exception,
28 rtsBool stop_at_atomically,
29 StgUpdateFrame *stop_here);
31 static void removeFromQueues(Capability *cap, StgTSO *tso);
33 static void blockedThrowTo (Capability *cap, StgTSO *source, StgTSO *target);
35 static void performBlockedException (Capability *cap,
36 StgTSO *source, StgTSO *target);
38 /* -----------------------------------------------------------------------------
41 This version of throwTo is safe to use if and only if one of the
46 - all the other threads in the system are stopped (eg. during GC).
48 - we surely own the target TSO (eg. we just took it from the
49 run queue of the current capability, or we are running it).
51 It doesn't cater for blocking the source thread until the exception
53 -------------------------------------------------------------------------- */
56 throwToSingleThreaded(Capability *cap, StgTSO *tso, StgClosure *exception)
58 throwToSingleThreaded_(cap, tso, exception, rtsFalse);
62 throwToSingleThreaded_(Capability *cap, StgTSO *tso, StgClosure *exception,
63 rtsBool stop_at_atomically)
65 // Thread already dead?
66 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
70 // Remove it from any blocking queues
71 removeFromQueues(cap,tso);
73 raiseAsync(cap, tso, exception, stop_at_atomically, NULL);
77 suspendComputation(Capability *cap, StgTSO *tso, StgUpdateFrame *stop_here)
79 // Thread already dead?
80 if (tso->what_next == ThreadComplete || tso->what_next == ThreadKilled) {
84 // Remove it from any blocking queues
85 removeFromQueues(cap,tso);
87 raiseAsync(cap, tso, NULL, rtsFalse, stop_here);
90 /* -----------------------------------------------------------------------------
93 This function may be used to throw an exception from one thread to
94 another, during the course of normal execution. This is a tricky
95 task: the target thread might be running on another CPU, or it
96 may be blocked and could be woken up at any point by another CPU.
97 We have some delicate synchronisation to do.
99 There is a completely safe fallback scheme: it is always possible
100 to just block the source TSO on the target TSO's blocked_exceptions
101 queue. This queue is locked using lockTSO()/unlockTSO(). It is
102 checked at regular intervals: before and after running a thread
103 (schedule() and threadPaused() respectively), and just before GC
104 (scheduleDoGC()). Activating a thread on this queue should be done
105 using maybePerformBlockedException(): this is done in the context
106 of the target thread, so the exception can be raised eagerly.
108 This fallback scheme works even if the target thread is complete or
109 killed: scheduleDoGC() will discover the blocked thread before the
112 Blocking the source thread on the target thread's blocked_exception
113 queue is also employed when the target thread is currently blocking
114 exceptions (ie. inside Control.Exception.block).
116 We could use the safe fallback scheme exclusively, but that
117 wouldn't be ideal: most calls to throwTo would block immediately,
118 possibly until the next GC, which might require the deadlock
119 detection mechanism to kick in. So we try to provide promptness
122 We can promptly deliver the exception if the target thread is:
124 - runnable, on the same Capability as the source thread (because
125 we own the run queue and therefore the target thread).
127 - blocked, and we can obtain exclusive access to it. Obtaining
128 exclusive access to the thread depends on how it is blocked.
130 We must also be careful to not trip over threadStackOverflow(),
131 which might be moving the TSO to enlarge its stack.
132 lockTSO()/unlockTSO() are used here too.
136 THROWTO_SUCCESS exception was raised, ok to continue
138 THROWTO_BLOCKED exception was not raised; block the source
139 thread then call throwToReleaseTarget() when
140 the source thread is properly tidied away.
142 -------------------------------------------------------------------------- */
145 throwTo (Capability *cap, // the Capability we hold
146 StgTSO *source, // the TSO sending the exception
147 StgTSO *target, // the TSO receiving the exception
148 StgClosure *exception, // the exception closure
149 /*[out]*/ void **out USED_IF_THREADS)
153 // follow ThreadRelocated links in the target first
154 while (target->what_next == ThreadRelocated) {
155 target = target->_link;
156 // No, it might be a WHITEHOLE:
157 // ASSERT(get_itbl(target)->type == TSO);
160 debugTrace(DEBUG_sched, "throwTo: from thread %lu to thread %lu",
161 (unsigned long)source->id, (unsigned long)target->id);
164 traceThreadStatus(DEBUG_sched, target);
169 debugTrace(DEBUG_sched, "throwTo: retrying...");
172 // Thread already dead?
173 if (target->what_next == ThreadComplete
174 || target->what_next == ThreadKilled) {
175 return THROWTO_SUCCESS;
178 status = target->why_blocked;
182 /* if status==NotBlocked, and target->cap == cap, then
183 we own this TSO and can raise the exception.
185 How do we establish this condition? Very carefully.
188 P = (status == NotBlocked)
189 Q = (tso->cap == cap)
191 Now, if P & Q are true, then the TSO is locked and owned by
192 this capability. No other OS thread can steal it.
194 If P==0 and Q==1: the TSO is blocked, but attached to this
195 capabilty, and it can be stolen by another capability.
197 If P==1 and Q==0: the TSO is runnable on another
198 capability. At any time, the TSO may change from runnable
199 to blocked and vice versa, while it remains owned by
202 Suppose we test like this:
208 this is defeated by another capability stealing a blocked
209 TSO from us to wake it up (Schedule.c:unblockOne()). The
210 other thread is doing
215 assuming arbitrary reordering, we could see this
225 so we need a memory barrier:
232 this avoids the problematic case. There are other cases
233 to consider, but this is the tricky one.
235 Note that we must be sure that unblockOne() does the
236 writes in the correct order: Q before P. The memory
237 barrier ensures that if we have seen the write to P, we
238 have also seen the write to Q.
241 Capability *target_cap;
244 target_cap = target->cap;
245 if (target_cap == cap && (target->flags & TSO_BLOCKEX) == 0) {
246 // It's on our run queue and not blocking exceptions
247 raiseAsync(cap, target, exception, rtsFalse, NULL);
248 return THROWTO_SUCCESS;
250 // Otherwise, just block on the blocked_exceptions queue
251 // of the target thread. The queue will get looked at
252 // soon enough: it is checked before and after running a
253 // thread, and during GC.
256 // Avoid race with threadStackOverflow, which may have
257 // just moved this TSO.
258 if (target->what_next == ThreadRelocated) {
260 target = target->_link;
263 // check again for ThreadComplete and ThreadKilled. This
264 // cooperates with scheduleHandleThreadFinished to ensure
265 // that we never miss any threads that are throwing an
266 // exception to a thread in the process of terminating.
267 if (target->what_next == ThreadComplete
268 || target->what_next == ThreadKilled) {
270 return THROWTO_SUCCESS;
272 blockedThrowTo(cap,source,target);
274 return THROWTO_BLOCKED;
281 To establish ownership of this TSO, we need to acquire a
282 lock on the MVar that it is blocked on.
285 StgInfoTable *info USED_IF_THREADS;
287 mvar = (StgMVar *)target->block_info.closure;
289 // ASSUMPTION: tso->block_info must always point to a
290 // closure. In the threaded RTS it does.
291 switch (get_itbl(mvar)->type) {
299 info = lockClosure((StgClosure *)mvar);
301 if (target->what_next == ThreadRelocated) {
302 target = target->_link;
303 unlockClosure((StgClosure *)mvar,info);
306 // we have the MVar, let's check whether the thread
307 // is still blocked on the same MVar.
308 if (target->why_blocked != BlockedOnMVar
309 || (StgMVar *)target->block_info.closure != mvar) {
310 unlockClosure((StgClosure *)mvar, info);
314 if ((target->flags & TSO_BLOCKEX) &&
315 ((target->flags & TSO_INTERRUPTIBLE) == 0)) {
316 lockClosure((StgClosure *)target);
317 blockedThrowTo(cap,source,target);
318 unlockClosure((StgClosure *)mvar, info);
320 return THROWTO_BLOCKED; // caller releases TSO
322 removeThreadFromMVarQueue(cap, mvar, target);
323 raiseAsync(cap, target, exception, rtsFalse, NULL);
324 unblockOne(cap, target);
325 unlockClosure((StgClosure *)mvar, info);
326 return THROWTO_SUCCESS;
330 case BlockedOnBlackHole:
332 ACQUIRE_LOCK(&sched_mutex);
333 // double checking the status after the memory barrier:
334 if (target->why_blocked != BlockedOnBlackHole) {
335 RELEASE_LOCK(&sched_mutex);
339 if (target->flags & TSO_BLOCKEX) {
341 blockedThrowTo(cap,source,target);
342 RELEASE_LOCK(&sched_mutex);
344 return THROWTO_BLOCKED; // caller releases TSO
346 removeThreadFromQueue(cap, &blackhole_queue, target);
347 raiseAsync(cap, target, exception, rtsFalse, NULL);
348 unblockOne(cap, target);
349 RELEASE_LOCK(&sched_mutex);
350 return THROWTO_SUCCESS;
354 case BlockedOnException:
360 To obtain exclusive access to a BlockedOnException thread,
361 we must call lockClosure() on the TSO on which it is blocked.
362 Since the TSO might change underneath our feet, after we
363 call lockClosure() we must check that
365 (a) the closure we locked is actually a TSO
366 (b) the original thread is still BlockedOnException,
367 (c) the original thread is still blocked on the TSO we locked
368 and (d) the target thread has not been relocated.
370 We synchronise with threadStackOverflow() (which relocates
371 threads) using lockClosure()/unlockClosure().
373 target2 = target->block_info.tso;
375 info = lockClosure((StgClosure *)target2);
376 if (info != &stg_TSO_info) {
377 unlockClosure((StgClosure *)target2, info);
380 if (target->what_next == ThreadRelocated) {
381 target = target->_link;
385 if (target2->what_next == ThreadRelocated) {
386 target->block_info.tso = target2->_link;
390 if (target->why_blocked != BlockedOnException
391 || target->block_info.tso != target2) {
397 Now we have exclusive rights to the target TSO...
399 If it is blocking exceptions, add the source TSO to its
400 blocked_exceptions queue. Otherwise, raise the exception.
402 if ((target->flags & TSO_BLOCKEX) &&
403 ((target->flags & TSO_INTERRUPTIBLE) == 0)) {
405 blockedThrowTo(cap,source,target);
408 return THROWTO_BLOCKED;
410 removeThreadFromQueue(cap, &target2->blocked_exceptions, target);
411 raiseAsync(cap, target, exception, rtsFalse, NULL);
412 unblockOne(cap, target);
414 return THROWTO_SUCCESS;
420 // Unblocking BlockedOnSTM threads requires the TSO to be
421 // locked; see STM.c:unpark_tso().
422 if (target->why_blocked != BlockedOnSTM) {
426 if ((target->flags & TSO_BLOCKEX) &&
427 ((target->flags & TSO_INTERRUPTIBLE) == 0)) {
428 blockedThrowTo(cap,source,target);
430 return THROWTO_BLOCKED;
432 raiseAsync(cap, target, exception, rtsFalse, NULL);
433 unblockOne(cap, target);
435 return THROWTO_SUCCESS;
439 case BlockedOnCCall_NoUnblockExc:
440 // I don't think it's possible to acquire ownership of a
441 // BlockedOnCCall thread. We just assume that the target
442 // thread is blocking exceptions, and block on its
443 // blocked_exception queue.
445 if (target->why_blocked != BlockedOnCCall &&
446 target->why_blocked != BlockedOnCCall_NoUnblockExc) {
450 blockedThrowTo(cap,source,target);
452 return THROWTO_BLOCKED;
454 #ifndef THREADEDED_RTS
458 #if defined(mingw32_HOST_OS)
459 case BlockedOnDoProc:
461 if ((target->flags & TSO_BLOCKEX) &&
462 ((target->flags & TSO_INTERRUPTIBLE) == 0)) {
463 blockedThrowTo(cap,source,target);
464 return THROWTO_BLOCKED;
466 removeFromQueues(cap,target);
467 raiseAsync(cap, target, exception, rtsFalse, NULL);
468 return THROWTO_SUCCESS;
473 barf("throwTo: unrecognised why_blocked value");
478 // Block a TSO on another TSO's blocked_exceptions queue.
479 // Precondition: we hold an exclusive lock on the target TSO (this is
480 // complex to achieve as there's no single lock on a TSO; see
483 blockedThrowTo (Capability *cap, StgTSO *source, StgTSO *target)
485 debugTrace(DEBUG_sched, "throwTo: blocking on thread %lu", (unsigned long)target->id);
486 setTSOLink(cap, source, target->blocked_exceptions);
487 target->blocked_exceptions = source;
488 dirty_TSO(cap,target); // we modified the blocked_exceptions queue
490 source->block_info.tso = target;
491 write_barrier(); // throwTo_exception *must* be visible if BlockedOnException is.
492 source->why_blocked = BlockedOnException;
498 throwToReleaseTarget (void *tso)
500 unlockTSO((StgTSO *)tso);
504 /* -----------------------------------------------------------------------------
505 Waking up threads blocked in throwTo
507 There are two ways to do this: maybePerformBlockedException() will
508 perform the throwTo() for the thread at the head of the queue
509 immediately, and leave the other threads on the queue.
510 maybePerformBlockedException() also checks the TSO_BLOCKEX flag
511 before raising an exception.
513 awakenBlockedExceptionQueue() will wake up all the threads in the
514 queue, but not perform any throwTo() immediately. This might be
515 more appropriate when the target thread is the one actually running
518 Returns: non-zero if an exception was raised, zero otherwise.
519 -------------------------------------------------------------------------- */
522 maybePerformBlockedException (Capability *cap, StgTSO *tso)
526 if (tso->what_next == ThreadComplete || tso->what_next == ThreadFinished) {
527 if (tso->blocked_exceptions != END_TSO_QUEUE) {
528 awakenBlockedExceptionQueue(cap,tso);
535 if (tso->blocked_exceptions != END_TSO_QUEUE &&
536 (tso->flags & TSO_BLOCKEX) != 0) {
537 debugTrace(DEBUG_sched, "throwTo: thread %lu has blocked exceptions but is inside block", (unsigned long)tso->id);
540 if (tso->blocked_exceptions != END_TSO_QUEUE
541 && ((tso->flags & TSO_BLOCKEX) == 0
542 || ((tso->flags & TSO_INTERRUPTIBLE) && interruptible(tso)))) {
544 // Lock the TSO, this gives us exclusive access to the queue
547 // Check the queue again; it might have changed before we
549 if (tso->blocked_exceptions == END_TSO_QUEUE) {
554 // We unblock just the first thread on the queue, and perform
555 // its throw immediately.
556 source = tso->blocked_exceptions;
557 performBlockedException(cap, source, tso);
558 tso->blocked_exceptions = unblockOne_(cap, source,
559 rtsFalse/*no migrate*/);
566 // awakenBlockedExceptionQueue(): Just wake up the whole queue of
567 // blocked exceptions and let them try again.
570 awakenBlockedExceptionQueue (Capability *cap, StgTSO *tso)
573 awakenBlockedQueue(cap, tso->blocked_exceptions);
574 tso->blocked_exceptions = END_TSO_QUEUE;
579 performBlockedException (Capability *cap, StgTSO *source, StgTSO *target)
581 StgClosure *exception;
583 ASSERT(source->why_blocked == BlockedOnException);
584 ASSERT(source->block_info.tso->id == target->id);
585 ASSERT(source->sp[0] == (StgWord)&stg_block_throwto_info);
586 ASSERT(((StgTSO *)source->sp[1])->id == target->id);
587 // check ids not pointers, because the thread might be relocated
589 exception = (StgClosure *)source->sp[2];
590 throwToSingleThreaded(cap, target, exception);
594 /* -----------------------------------------------------------------------------
595 Remove a thread from blocking queues.
597 This is for use when we raise an exception in another thread, which
600 Precondition: we have exclusive access to the TSO, via the same set
601 of conditions as throwToSingleThreaded() (c.f.).
602 -------------------------------------------------------------------------- */
605 removeFromQueues(Capability *cap, StgTSO *tso)
607 switch (tso->why_blocked) {
613 // Be careful: nothing to do here! We tell the scheduler that the
614 // thread is runnable and we leave it to the stack-walking code to
615 // abort the transaction while unwinding the stack. We should
616 // perhaps have a debugging test to make sure that this really
617 // happens and that the 'zombie' transaction does not get
622 removeThreadFromMVarQueue(cap, (StgMVar *)tso->block_info.closure, tso);
625 case BlockedOnBlackHole:
626 removeThreadFromQueue(cap, &blackhole_queue, tso);
629 case BlockedOnException:
631 StgTSO *target = tso->block_info.tso;
633 // NO: when called by threadPaused(), we probably have this
634 // TSO already locked (WHITEHOLEd) because we just placed
635 // ourselves on its queue.
636 // ASSERT(get_itbl(target)->type == TSO);
638 while (target->what_next == ThreadRelocated) {
639 target = target->_link;
642 removeThreadFromQueue(cap, &target->blocked_exceptions, tso);
646 #if !defined(THREADED_RTS)
649 #if defined(mingw32_HOST_OS)
650 case BlockedOnDoProc:
652 removeThreadFromDeQueue(cap, &blocked_queue_hd, &blocked_queue_tl, tso);
653 #if defined(mingw32_HOST_OS)
654 /* (Cooperatively) signal that the worker thread should abort
657 abandonWorkRequest(tso->block_info.async_result->reqID);
662 removeThreadFromQueue(cap, &sleeping_queue, tso);
667 barf("removeFromQueues: %d", tso->why_blocked);
671 unblockOne(cap, tso);
674 /* -----------------------------------------------------------------------------
677 * The following function implements the magic for raising an
678 * asynchronous exception in an existing thread.
680 * We first remove the thread from any queue on which it might be
681 * blocked. The possible blockages are MVARs and BLACKHOLE_BQs.
683 * We strip the stack down to the innermost CATCH_FRAME, building
684 * thunks in the heap for all the active computations, so they can
685 * be restarted if necessary. When we reach a CATCH_FRAME, we build
686 * an application of the handler to the exception, and push it on
687 * the top of the stack.
689 * How exactly do we save all the active computations? We create an
690 * AP_STACK for every UpdateFrame on the stack. Entering one of these
691 * AP_STACKs pushes everything from the corresponding update frame
692 * upwards onto the stack. (Actually, it pushes everything up to the
693 * next update frame plus a pointer to the next AP_STACK object.
694 * Entering the next AP_STACK object pushes more onto the stack until we
695 * reach the last AP_STACK object - at which point the stack should look
696 * exactly as it did when we killed the TSO and we can continue
697 * execution by entering the closure on top of the stack.
699 * We can also kill a thread entirely - this happens if either (a) the
700 * exception passed to raiseAsync is NULL, or (b) there's no
701 * CATCH_FRAME on the stack. In either case, we strip the entire
702 * stack and replace the thread with a zombie.
704 * ToDo: in THREADED_RTS mode, this function is only safe if either
705 * (a) we hold all the Capabilities (eg. in GC, or if there is only
706 * one Capability), or (b) we own the Capability that the TSO is
707 * currently blocked on or on the run queue of.
709 * -------------------------------------------------------------------------- */
712 raiseAsync(Capability *cap, StgTSO *tso, StgClosure *exception,
713 rtsBool stop_at_atomically, StgUpdateFrame *stop_here)
715 StgRetInfoTable *info;
720 debugTrace(DEBUG_sched,
721 "raising exception in thread %ld.", (long)tso->id);
723 #if defined(PROFILING)
725 * Debugging tool: on raising an exception, show where we are.
726 * See also Exception.cmm:stg_raisezh.
727 * This wasn't done for asynchronous exceptions originally; see #1450
729 if (RtsFlags.ProfFlags.showCCSOnException)
731 fprintCCS_stderr(tso->prof.CCCS);
735 // mark it dirty; we're about to change its stack.
740 // ASSUMES: the thread is not already complete or dead. Upper
741 // layers should deal with that.
742 ASSERT(tso->what_next != ThreadComplete && tso->what_next != ThreadKilled);
744 if (stop_here != NULL) {
745 updatee = stop_here->updatee;
750 // The stack freezing code assumes there's a closure pointer on
751 // the top of the stack, so we have to arrange that this is the case...
753 if (sp[0] == (W_)&stg_enter_info) {
757 sp[0] = (W_)&stg_dummy_ret_closure;
761 while (stop_here == NULL || frame < (StgPtr)stop_here) {
763 // 1. Let the top of the stack be the "current closure"
765 // 2. Walk up the stack until we find either an UPDATE_FRAME or a
768 // 3. If it's an UPDATE_FRAME, then make an AP_STACK containing the
769 // current closure applied to the chunk of stack up to (but not
770 // including) the update frame. This closure becomes the "current
771 // closure". Go back to step 2.
773 // 4. If it's a CATCH_FRAME, then leave the exception handler on
774 // top of the stack applied to the exception.
776 // 5. If it's a STOP_FRAME, then kill the thread.
778 // NB: if we pass an ATOMICALLY_FRAME then abort the associated
781 info = get_ret_itbl((StgClosure *)frame);
783 switch (info->i.type) {
790 // First build an AP_STACK consisting of the stack chunk above the
791 // current update frame, with the top word on the stack as the
794 words = frame - sp - 1;
795 ap = (StgAP_STACK *)allocateLocal(cap,AP_STACK_sizeW(words));
798 ap->fun = (StgClosure *)sp[0];
800 for(i=0; i < (nat)words; ++i) {
801 ap->payload[i] = (StgClosure *)*sp++;
804 SET_HDR(ap,&stg_AP_STACK_info,
805 ((StgClosure *)frame)->header.prof.ccs /* ToDo */);
806 TICK_ALLOC_UP_THK(words+1,0);
808 //IF_DEBUG(scheduler,
809 // debugBelch("sched: Updating ");
810 // printPtr((P_)((StgUpdateFrame *)frame)->updatee);
811 // debugBelch(" with ");
812 // printObj((StgClosure *)ap);
815 if (((StgUpdateFrame *)frame)->updatee == updatee) {
816 // If this update frame points to the same closure as
817 // the update frame further down the stack
818 // (stop_here), then don't perform the update. We
819 // want to keep the blackhole in this case, so we can
820 // detect and report the loop (#2783).
821 ap = (StgAP_STACK*)updatee;
823 // Perform the update
824 // TODO: this may waste some work, if the thunk has
825 // already been updated by another thread.
826 UPD_IND(((StgUpdateFrame *)frame)->updatee, (StgClosure *)ap);
829 sp += sizeofW(StgUpdateFrame) - 1;
830 sp[0] = (W_)ap; // push onto stack
832 continue; //no need to bump frame
837 // We've stripped the entire stack, the thread is now dead.
838 tso->what_next = ThreadKilled;
839 tso->sp = frame + sizeofW(StgStopFrame);
844 // If we find a CATCH_FRAME, and we've got an exception to raise,
845 // then build the THUNK raise(exception), and leave it on
846 // top of the CATCH_FRAME ready to enter.
850 StgCatchFrame *cf = (StgCatchFrame *)frame;
854 if (exception == NULL) break;
856 // we've got an exception to raise, so let's pass it to the
857 // handler in this frame.
859 raise = (StgThunk *)allocateLocal(cap,sizeofW(StgThunk)+1);
860 TICK_ALLOC_SE_THK(1,0);
861 SET_HDR(raise,&stg_raise_info,cf->header.prof.ccs);
862 raise->payload[0] = exception;
864 // throw away the stack from Sp up to the CATCH_FRAME.
868 /* Ensure that async excpetions are blocked now, so we don't get
869 * a surprise exception before we get around to executing the
872 tso->flags |= TSO_BLOCKEX | TSO_INTERRUPTIBLE;
874 /* Put the newly-built THUNK on top of the stack, ready to execute
875 * when the thread restarts.
878 sp[-1] = (W_)&stg_enter_info;
880 tso->what_next = ThreadRunGHC;
881 IF_DEBUG(sanity, checkTSO(tso));
885 case ATOMICALLY_FRAME:
886 if (stop_at_atomically) {
887 ASSERT(stmGetEnclosingTRec(tso->trec) == NO_TREC);
888 stmCondemnTransaction(cap, tso -> trec);
890 tso->what_next = ThreadRunGHC;
893 // Not stop_at_atomically... fall through and abort the
896 case CATCH_STM_FRAME:
897 case CATCH_RETRY_FRAME:
898 // IF we find an ATOMICALLY_FRAME then we abort the
899 // current transaction and propagate the exception. In
900 // this case (unlike ordinary exceptions) we do not care
901 // whether the transaction is valid or not because its
902 // possible validity cannot have caused the exception
903 // and will not be visible after the abort.
906 StgTRecHeader *trec = tso -> trec;
907 StgTRecHeader *outer = stmGetEnclosingTRec(trec);
908 debugTrace(DEBUG_stm,
909 "found atomically block delivering async exception");
910 stmAbortTransaction(cap, trec);
911 stmFreeAbortedTRec(cap, trec);
920 // move on to the next stack frame
921 frame += stack_frame_sizeW((StgClosure *)frame);
924 // if we got here, then we stopped at stop_here
925 ASSERT(stop_here != NULL);