[m-rev.] for review: threads and fork()

[Peter Wang]

Estimated hours taken: 1
Branches: main

compiler/process_util.m:
	Disable forking by `mmc --make' if built in a parallel grade as
	threads and fork() do not mix.

Index: compiler/process_util.m
===================================================================
RCS file: /home/mercury1/repository/mercury/compiler/process_util.m,v
retrieving revision 1.18
diff -u -r1.18 process_util.m
--- compiler/process_util.m	28 Nov 2005 04:11:51 -0000	1.18
+++ compiler/process_util.m	23 Mar 2006 12:19:37 -0000
@@ -292,7 +292,16 @@
     can_fork,
     [will_not_call_mercury, thread_safe, promise_pure],
 "
-#ifdef MC_CAN_FORK
+    /*
+    ** call_in_forked_process_2 is not `thread_safe' so will hold a mutex
+    ** that the child process will want.  At the same time the parent process
+    ** waits for the child to exit, so we have a deadlock.
+    **
+    ** Also, in pthreads, a forked process does not inherit the threads of
+    ** the original process so it is not at all clear whether we could use
+    ** fork() when running in a parallel grade.
+    */
+#if (defined MC_CAN_FORK) && (!defined MR_THREAD_SAFE)
     SUCCESS_INDICATOR = MR_TRUE;
 #else
     SUCCESS_INDICATOR = MR_FALSE;


I did try marking call_in_forked_process_2 as `thread_safe', which
worked well enough to build hello_world, but it seems dodgy and will
likely break if there is actually more than one thread.

Peter
--------------------------------------------------------------------------
mercury-reviews mailing list
post:  mercury-reviews at cs.mu.oz.au
administrative address: owner-mercury-reviews at cs.mu.oz.au
unsubscribe: Address: mercury-reviews-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-reviews-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------