unsafePerformIO thunk will be updated with a stack object
containing the unblockAsyncExceptions_ret frame. Later, when
someone else evaluates this thunk, the blocked exception state is
- not restored, and the result is that unblockAsyncExceptions_ret
- will attempt to unblock exceptions in the current thread, but it'll
- find that the CurrentTSO->blocked_exceptions is NULL. Hence, we
- work around this by checking for NULL in awakenBlockedQueue().
+ not restored.
-------------------------------------------------------------------------- */
{
CInt r;
- // Not true: see comments above
- // ASSERT(StgTSO_blocked_exceptions(CurrentTSO) != NULL);
-
StgTSO_flags(CurrentTSO) = StgTSO_flags(CurrentTSO) &
~(TSO_BLOCKEX::I32|TSO_INTERRUPTIBLE::I32);
INFO_TABLE_RET( stg_blockAsyncExceptionszh_ret,
0/*framesize*/, 0/*bitmap*/, RET_SMALL )
{
- // Not true: see comments above
- // ASSERT(StgTSO_blocked_exceptions(CurrentTSO) == NULL);
-
StgTSO_flags(CurrentTSO) =
StgTSO_flags(CurrentTSO) | TSO_BLOCKEX::I32 | TSO_INTERRUPTIBLE::I32;
}
#endif
- if (W_[no_break_on_exception] != 0) {
- W_[no_break_on_exception] = 0;
- } else {
- if (TO_W_(CInt[rts_stop_on_exception]) != 0) {
- W_ ioAction;
- // we don't want any further exceptions to be caught,
- // until GHCi is ready to handle them. This prevents
- // deadlock if an exception is raised in InteractiveUI,
- // for exmplae. Perhaps the stop_on_exception flag should
- // be per-thread.
- W_[rts_stop_on_exception] = 0;
- "ptr" ioAction = foreign "C" deRefStablePtr (W_[rts_breakpoint_io_action] "ptr") [];
- Sp = Sp - WDS(6);
- Sp(5) = exception;
- Sp(4) = stg_raise_ret_info;
- Sp(3) = exception; // the AP_STACK
- Sp(2) = base_GHCziBase_True_closure; // dummy breakpoint info
- Sp(1) = base_GHCziBase_True_closure; // True <=> a breakpoint
- R1 = ioAction;
- jump stg_ap_pppv_info;
- }
- }
-
- /* Inform the Hpc that an exception has been thrown */
- foreign "C" hs_hpc_raise_event(CurrentTSO "ptr") [];
-
retry_pop_stack:
StgTSO_sp(CurrentTSO) = Sp;
frame_type = foreign "C" raiseExceptionHelper(BaseReg "ptr", CurrentTSO "ptr", exception "ptr") [];
}
}
+ // After stripping the stack, see whether we should break here for
+ // GHCi (c.f. the -fbreak-on-exception flag). We do this after
+ // stripping the stack for a reason: we'll be inspecting values in
+ // GHCi, and it helps if all the thunks under evaluation have
+ // already been updated with the exception, rather than being left
+ // as blackholes.
+ if (W_[no_break_on_exception] != 0) {
+ W_[no_break_on_exception] = 0;
+ } else {
+ if (TO_W_(CInt[rts_stop_on_exception]) != 0) {
+ W_ ioAction;
+ // we don't want any further exceptions to be caught,
+ // until GHCi is ready to handle them. This prevents
+ // deadlock if an exception is raised in InteractiveUI,
+ // for exmplae. Perhaps the stop_on_exception flag should
+ // be per-thread.
+ W_[rts_stop_on_exception] = 0;
+ "ptr" ioAction = foreign "C" deRefStablePtr (W_[rts_breakpoint_io_action] "ptr") [];
+ Sp = Sp - WDS(6);
+ Sp(5) = exception;
+ Sp(4) = stg_raise_ret_info;
+ Sp(3) = exception; // the AP_STACK
+ Sp(2) = base_GHCziBase_True_closure; // dummy breakpoint info
+ Sp(1) = base_GHCziBase_True_closure; // True <=> a breakpoint
+ R1 = ioAction;
+ jump stg_ap_pppv_info;
+ }
+ }
+
if (frame_type == STOP_FRAME) {
/*
* We've stripped the entire stack, the thread is now dead.