[M3devel] Ho hum... AtFork...
mika at async.caltech.edu
mika at async.caltech.edu
Wed Aug 13 18:37:35 CEST 2014
1.262
Short summary: 1.262 seems to work fine on AMD64_LINUX except for the
fork issue (a bummer because my parallelization of the backends of cm3
uses fork a lot...and it does speed up the compiler significantly if
you have fast disks and lots of cores) It does the attached on AMD64_FREEBSD.
I see the intent of what you were doing in later version/s, but it was
a bit rough, I think... I think the three issues I reported (1. locking
heapMu from AtFork; 2. the try-only-once bug in LockHeap itself; 3. not
passing the lock to release from Wait) are all more or less real.
Mika
Tony Hosking writes:
>Which revision of ThreadPThread.m3 is this output from?
>
>On Aug 13, 2014, at 2:30 AM, mika at async.caltech.edu wrote:
>
>>=20
>> OK... deleted all the derived directories and rebuilt. Result on =
>FreeBSD is similar (or identical, hard to tell):
>>=20
>> I get a partial deadlock:
>>=20
>> PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU =
>COMMAND
>> 48760 root 98 0 87836K 29104K CPU1 1 12:32 83.89% =
>threadtest{threadtest}
>> 48760 root 94 0 87836K 29104K CPU3 3 11:02 74.66% =
>threadtest{threadtest}
>> 48760 root 94 0 87836K 29104K CPU0 0 11:03 71.97% =
>threadtest{threadtest}
>> 48760 root 32 0 87836K 29104K umtxn 2 0:01 0.00% =
>threadtest{threadtest}
>> 48760 root 20 0 87836K 29104K umtxn 2 0:01 0.00% =
>threadtest{threadtest}
>> 48760 root 35 0 87836K 29104K uwait 2 0:01 0.00% =
>threadtest{threadtest}
>> 48760 root 27 0 87836K 29104K uwait 0 0:01 0.00% =
>threadtest{threadtest}
>> 48760 root 27 0 87836K 29104K umtxn 3 0:01 0.00% =
>threadtest{threadtest}
>> 48760 root 27 0 87836K 29104K umtxn 2 0:01 0.00% =
>threadtest{threadtest}
>> 48760 root 20 0 87836K 29104K umtxn 0 0:00 0.00% =
>threadtest{threadtest}
>> 48759 root 20 0 30276K 11064K wait 0 0:00 0.00% gdb
>>=20
>> one of the dead threads is the main thread, so now output. But three
>> threads are still alive, interestingly enough.
>>=20
>> Looks like the same issue as before:
>>=20
>> Thread 5 (Thread 801807400 (LWP 100751/threadtest)):
>> #0 0x0000000800ae089c in __error () from /lib/libthr.so.3
>> #1 0x0000000800ad6261 in pthread_suspend_np () from /lib/libthr.so.3
>> #2 0x00000000004511bc in ThreadPThread__SuspendThread (mt=3DError =
>accessing memory address 0x8000ff5faa28: Bad address.
>> ) at ../src/thread/PTHREAD/ThreadFreeBSD.c:33
>> #3 0x000000000044ed8e in ThreadPThread__StopThread =
>(M3_DMxDjQ_act=3DError accessing memory address 0x8000ff5faa48: Bad =
>address.
>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:909
>> #4 0x000000000044ef31 in ThreadPThread__StopWorld () at =
>../src/thread/PTHREAD/ThreadPThread.m3:948
>> #5 0x000000000044e44e in RTThread__SuspendOthers () at =
>../src/thread/PTHREAD/ThreadPThread.m3:713
>> #6 0x0000000000436386 in RTCollector__CollectSomeInStateZero () at =
>../src/runtime/common/RTCollector.m3:749
>> #7 0x0000000000436331 in RTCollector__CollectSome () at =
>../src/runtime/common/RTCollector.m3:723
>> #8 0x0000000000435ff9 in RTHeapRep__CollectEnough () at =
>../src/runtime/common/RTCollector.m3:657
>> #9 0x0000000000433233 in RTAllocator__AllocTraced =
>(M3_Cwb5VA_dataSize=3DError accessing memory address 0x8000ff5fac88: Bad =
>address.
>> ) at ../src/runtime/common/RTAllocator.m3:367
>> #10 0x0000000000432b55 in RTAllocator__GetOpenArray =
>(M3_Eic7CK_def=3DError accessing memory address 0x8000ff5fad48: Bad =
>address.
>> ) at ../src/runtime/common/RTAllocator.m3:296
>> #11 0x0000000000431e8f in RTHooks__AllocateOpenArray =
>(M3_AJWxb1_defn=3DError accessing memory address 0x8000ff5fada8: Bad =
>address.
>> ) at ../src/runtime/common/RTAllocator.m3:143
>> #12 0x00000000004052ea in Main__AApply (M3_AP7a1g_cl=3DError accessing =
>memory address 0x8000ff5fade8: Bad address.
>> ) at ../src/Main.m3:283
>> #13 0x000000000044d243 in ThreadPThread__RunThread (M3_DMxDjQ_me=3DError=
> accessing memory address 0x8000ff5faeb8: Bad address.
>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449
>> #14 0x000000000044cf00 in ThreadPThread__ThreadBase =
>(M3_AJWxb1_param=3DError accessing memory address 0x8000ff5faf68: Bad =
>address.
>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422
>> #15 0x0000000800ad54a4 in pthread_create () from /lib/libthr.so.3
>> #16 0x0000000000000000 in ?? ()
>>=20
>> Interestingly the same test seems to work on AMD64_LINUX. That is
>> promising, at least. But of course, software testing never reveals =
>the
>> absence of bugs.
>>=20
>> Mika
>>=20
>> Peter McKinna writes:
>>> --089e01183a1813188705007c3498
>>> Content-Type: text/plain; charset=3DUTF-8
>>>=20
>>> Oh, one other thing that seemed to help me in testing this stuff =
>especially
>>> things to do with m3core and libm3 which seem to get out of whack
>>> sometimes, is to totally remove the target directory of m3core then =
>cm3;
>>> cm3 -ship Same with libm3. I'm not sure that clean actually removes =
>the
>>> targets and I think there are leftover files that stuff things up =
>since I
>>> would get segv in FileRd for no obvious reason.
>>>=20
>>> Peter
>>>=20
>>>=20
>>>=20
>>> On Wed, Aug 13, 2014 at 3:33 PM, Peter McKinna =
><peter.mckinna at gmail.com>
>>> wrote:
>>>=20
>>>> That is weird. The thread tester works fine with the old version of
>>>> ThreadPThread.m3 on linux amd_64 except if you have the fork test in =
>the
>>>> mix of tests you see the pthread_mutex_destroy every now and again, =
>the
>>>> fork bug in fact.
>>>> I was testing the pthread changes Tony has made over the past few =
>days
>>>> using the thread tester program until I stupidly rebuilt the =
>compiler and
>>>> was getting all sorts of hangs as it dragged in the new version of
>>>> pthreads. So I rebuilt everything from a backup to get back where I =
>was
>>>> with a decent compiler, then reintroduced the latest changes to
>>>> ThreadPThread.m3, which yesterday was just hanging in sigsuspend but =
>today
>>>> is giving me an assert failure at line 1387 in UnlockHeap. So its a =
>bit
>>>> unstable to say the least. Maybe there is something else wrong with =
>freebsd.
>>>>=20
>>>> Peter
>>>>=20
>>>>=20
>>>>=20
>>>> On Wed, Aug 13, 2014 at 2:54 PM, <mika at async.caltech.edu> wrote:
>>>>=20
>>>>>=20
>>>>> Yeah OK 1.262 makes a lot more sense to me but I'm still not =
>getting
>>>>> things to work and now I'm really baffled because I think I =
>understand
>>>>> this code!
>>>>>=20
>>>>> What I did was I threw away all my changes and reverted =
>ThreadPThread.m3
>>>>> to 1.262 as you suggested.
>>>>>=20
>>>>> Rebuilt the compiler with upgrade.sh
>>>>>=20
>>>>> Then rebuilt the compiler again with itself.
>>>>>=20
>>>>> Then I realcleaned the world and buildshipped it. (I was afraid
>>>>> that parseparams, also imported by the thread tester, would pollute
>>>>> it somehow.)
>>>>>=20
>>>>> When I look at the code in 1.262 it looks quite straightforward. =
>heapMu
>>>>> is just a mutex, (Modula-3) mutexes are just (pthreads) mutexes.
>>>>> Condition
>>>>> variables are mutexes too, but that's no big deal.
>>>>>=20
>>>>> So, rebuild thread tester, run it:
>>>>>=20
>>>>> new source -> compiling Main.m3
>>>>> -> linking threadtest
>>>>> root at rover:~mika/cm3-cvs-anon/cm3/m3-libs/m3core/tests/thread #
>>>>> AMD64_FREEBSD/threadtest
>>>>> Writing file...done
>>>>> Creating read threads...done
>>>>> Creating fork threads...done
>>>>> Creating alloc threads...done
>>>>> Creating lock threads...done
>>>>> running...printing oldest/median age/newest
>>>>> .
>>>>>=20
>>>>> ***
>>>>> *** runtime error:
>>>>> *** Segmentation violation - possible attempt to dereference
>>>>> NIL.........laziest thread is 1407901189/1407901189/9 (tests: read
>>>>> 1407901189/1407901189/1407901189 fork =
>1407901189/1407901189/1407901189
>>>>> alloc 9/9/9 lock 1407901189/1407901189/9)
>>>>>=20
>>>>>=20
>>>>>=20
>>>>>=20
>>>>> ^C
>>>>> root at rover:~mika/cm3-cvs-anon/cm3/m3-libs/m3core/tests/thread # gdb =
>!$
>>>>> gdb AMD64_FREEBSD/threadtest
>>>>> GNU gdb 6.1.1 [FreeBSD]
>>>>> Copyright 2004 Free Software Foundation, Inc.
>>>>> GDB is free software, covered by the GNU General Public License, =
>and you
>>>>> are
>>>>> welcome to change it and/or distribute copies of it under certain
>>>>> conditions.
>>>>> Type "show copying" to see the conditions.
>>>>> There is absolutely no warranty for GDB. Type "show warranty" for
>>>>> details.
>>>>> This GDB was configured as "amd64-marcel-freebsd"...
>>>>> (gdb) run
>>>>> Starting program:
>>>>> =
>/big/home2/mika/2/cm3-cvs/cm3/m3-libs/m3core/tests/thread/AMD64_FREEBSD/th=
>readtest
>>>>> [New LWP 100121]
>>>>> Writing file...done
>>>>> Creating read threads...done
>>>>> Creating fork threads...done
>>>>> Creating alloc threads...done
>>>>> Creating lock threads...done
>>>>> running...printing oldest/median age/newest
>>>>> .[New Thread 801807000 (LWP 100343/threadtest)]
>>>>> [New Thread 801807c00 (LWP 100667/threadtest)]
>>>>> [New Thread 801808800 (LWP 100816/threadtest)]
>>>>> [New Thread 801807400 (LWP 100349/threadtest)]
>>>>> [New Thread 801808400 (LWP 100815/threadtest)]
>>>>> [New Thread 801807800 (LWP 100352/threadtest)]
>>>>> [New Thread 801808c00 (LWP 100817/threadtest)]
>>>>> [New Thread 801809000 (LWP 100818/threadtest)]
>>>>> [New Thread 801809800 (LWP 100820/threadtest)]
>>>>> [New Thread 801808000 (LWP 100678/threadtest)]
>>>>> [New Thread 801806400 (LWP 100121/threadtest)]
>>>>> [New Thread 801806800 (LWP 100341/threadtest)]
>>>>> [New Thread 801806c00 (LWP 100342/threadtest)]
>>>>> .ERROR: pthread_mutex_lock:11
>>>>>=20
>>>>> Program received signal SIGABRT, Aborted.
>>>>> [Switching to Thread 801809800 (LWP 100820/threadtest)]
>>>>> 0x0000000800d5c26a in thr_kill () from /lib/libc.so.7
>>>>> (gdb) where
>>>>> #0 0x0000000800d5c26a in thr_kill () from /lib/libc.so.7
>>>>> #1 0x0000000800e23ac9 in abort () from /lib/libc.so.7
>>>>> #2 0x000000000045101f in ThreadPThread__pthread_mutex_lock =
>(mutex=3DError
>>>>> accessing memory address 0x8000fe5f2d98: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:513
>>>>> #3 0x000000000044b370 in ThreadPThread__LockMutex =
>(M3_AYIbX3_m=3DError
>>>>> accessing memory address 0x8000fe5f2dc8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:119
>>>>> #4 0x0000000000405b5d in Main__LApply (M3_AP7a1g_cl=3DError =
>accessing
>>>>> memory address 0x8000fe5f2df8: Bad address.
>>>>> ) at ../src/Main.m3:319
>>>>> #5 0x000000000044d243 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3DError
>>>>> accessing memory address 0x8000fe5f2eb8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449
>>>>> #6 0x000000000044cf00 in ThreadPThread__ThreadBase
>>>>> (M3_AJWxb1_param=3DError accessing memory address 0x8000fe5f2f68: =
>Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422
>>>>> #7 0x0000000800ad54a4 in pthread_create () from /lib/libthr.so.3
>>>>> #8 0x0000000000000000 in ?? ()
>>>>> (gdb) set lang c
>>>>> (gdb) thread apply all bt
>>>>> ... not that interesting ...
>>>>>=20
>>>>> Segfault is segfault---error 11 is EDEADLK (locking against =
>myself?)
>>>>>=20
>>>>> ????? the code is very straightforward, as I said.
>>>>>=20
>>>>> I thought people had the thread tester working with pthreads? =
>Which set
>>>>> of files, then? Anyone on FreeBSD/amd64?
>>>>>=20
>>>>> Could it be an issue with "volatile"? Not even sure where to look.
>>>>>=20
>>>>> The code calling the lock is just this:
>>>>>=20
>>>>> PROCEDURE GetChar (rd: T): CHAR
>>>>> RAISES {EndOfFile, Failure, Alerted} =3D
>>>>> BEGIN
>>>>> LOCK rd DO
>>>>> RETURN FastGetChar(rd);
>>>>> END
>>>>> END GetChar;
>>>>>=20
>>>>> No mysteries there...
>>>>>=20
>>>>> Ah is this the fork bug? Looks like the Init routine is called on =
>a
>>>>> lot of NIL mutexes around the fork. But the subprocesses aren't =
>meant
>>>>> to accesses the mutexes... humm
>>>>>=20
>>>>> Hmm, and it's not entirely a fork issue. Even with "threadtest =
>-tests
>>>>> STD,-fork,-forktoomuch" it misbehaves. Hangs....
>>>>>=20
>>>>> Looks like a deadlock this time.
>>>>>=20
>>>>> Some stacks...
>>>>>=20
>>>>> Thread 4 (Thread 800c0b400 (LWP 100477/threadtest)):
>>>>> #0 0x00000000004b4e2b in __vdso_gettimeofday ()
>>>>> #1 0x00000000004ab8d2 in gettimeofday ()
>>>>> #2 0x0000000000452e4b in TimePosix__Now () at
>>>>> ../src/time/POSIX/TimePosixC.c:50
>>>>> #3 0x0000000000452d72 in Time__Now () at
>>>>> ../src/time/POSIX/TimePosix.m3:14
>>>>> #4 0x00000000004029a3 in Main__LApply (M3_AP7a1g_cl=3DError =
>accessing
>>>>> memory address 0x8000fedf6df8: Bad address.
>>>>> ) at ../src/Main.m3:327
>>>>> #5 0x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3DError
>>>>> accessing memory address 0x8000fedf6eb8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449
>>>>> #6 0x0000000000449c50 in ThreadPThread__ThreadBase
>>>>> (M3_AJWxb1_param=3DError accessing memory address 0x8000fedf6f68: =
>Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422
>>>>> #7 0x000000000047c7d4 in thread_start ()
>>>>> #8 0x0000000000000000 in ?? ()
>>>>>=20
>>>>> Thread 3 (Thread 800c0b000 (LWP 100476/threadtest)):
>>>>> #0 0x000000000047e53c in _umtx_op_err ()
>>>>> #1 0x0000000000475f14 in __thr_umutex_lock ()
>>>>> #2 0x0000000000479404 in mutex_lock_common ()
>>>>> #3 0x000000000044dd15 in ThreadPThread__pthread_mutex_lock =
>(mutex=3DError
>>>>> accessing memory address 0x8000feff7d98: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:506
>>>>> #4 0x00000000004480c0 in ThreadPThread__LockMutex =
>(M3_AYIbX3_m=3DError
>>>>> accessing memory address 0x8000feff7dc8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:119
>>>>> #5 0x00000000004028ad in Main__LApply (M3_AP7a1g_cl=3DError =
>accessing
>>>>> memory address 0x8000feff7df8: Bad address.
>>>>> ) at ../src/Main.m3:319
>>>>> #6 0x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3DError
>>>>> accessing memory address 0x8000feff7eb8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449
>>>>> #7 0x0000000000449c50 in ThreadPThread__ThreadBase
>>>>> (M3_AJWxb1_param=3DError accessing memory address 0x8000feff7f68: =
>Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422
>>>>> #8 0x000000000047c7d4 in thread_start ()
>>>>> #9 0x0000000000000000 in ?? ()
>>>>>=20
>>>>> Thread 6 (Thread 800c09800 (LWP 100470/threadtest)):
>>>>> #0 0x000000000047e53c in _umtx_op_err ()
>>>>> #1 0x0000000000475759 in suspend_common ()
>>>>> #2 0x00000000004755c1 in pthread_suspend_np ()
>>>>> #3 0x000000000044df0c in ThreadPThread__SuspendThread (mt=3DError
>>>>> accessing memory address 0x8000ffbfd6f8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadFreeBSD.c:33
>>>>> #4 0x000000000044bade in ThreadPThread__StopThread =
>(M3_DMxDjQ_act=3DError
>>>>> accessing memory address 0x8000ffbfd718: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:909
>>>>> #5 0x000000000044bc81 in ThreadPThread__StopWorld () at
>>>>> ../src/thread/PTHREAD/ThreadPThread.m3:948
>>>>> #6 0x000000000044b19e in RTThread__SuspendOthers () at
>>>>> ../src/thread/PTHREAD/ThreadPThread.m3:713
>>>>> #7 0x00000000004330d6 in RTCollector__CollectSomeInStateZero () at
>>>>> ../src/runtime/common/RTCollector.m3:749
>>>>> #8 0x0000000000433081 in RTCollector__CollectSome () at
>>>>> ../src/runtime/common/RTCollector.m3:723
>>>>> #9 0x0000000000432d49 in RTHeapRep__CollectEnough () at
>>>>> ../src/runtime/common/RTCollector.m3:657
>>>>> #10 0x000000000042ff83 in RTAllocator__AllocTraced
>>>>> (M3_Cwb5VA_dataSize=3DError accessing memory address =
>0x8000ffbfd958: Bad
>>>>> address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:367
>>>>> #11 0x000000000042f8a5 in RTAllocator__GetOpenArray =
>(M3_Eic7CK_def=3DError
>>>>> accessing memory address 0x8000ffbfda18: Bad address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:296
>>>>> #12 0x000000000042ebdf in RTHooks__AllocateOpenArray
>>>>> (M3_AJWxb1_defn=3DError accessing memory address 0x8000ffbfda78: =
>Bad address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:143
>>>>> #13 0x000000000040ca4e in Rd__NextBuff (M3_EkTcCb_rd=3DError =
>accessing
>>>>> memory address 0x8000ffbfdab8: Bad address.
>>>>> ) at ../src/rw/Rd.m3:159
>>>>> #14 0x000000000040cd43 in UnsafeRd__FastGetChar (M3_EkTcCb_rd=3DError=
>
>>>>> accessing memory address 0x8000ffbfdb88: Bad address.
>>>>> ) at ../src/rw/Rd.m3:187
>>>>> #15 0x000000000040cc4a in Rd__GetChar (M3_EkTcCb_rd=3DError =
>accessing
>>>>> memory address 0x8000ffbfdbd8: Bad address.
>>>>> ) at ../src/rw/Rd.m3:176
>>>>> #16 0x000000000040095c in Main__RApply (M3_AP7a1g_cl=3DError =
>accessing
>>>>> memory address 0x8000ffbfdc58: Bad address.
>>>>> ) at ../src/Main.m3:185
>>>>>=20
>>>>> #17 0x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3DError
>>>>> accessing memory address 0x8000ffbfdeb8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449
>>>>> #18 0x0000000000449c50 in ThreadPThread__ThreadBase
>>>>> (M3_AJWxb1_param=3DError accessing memory address 0x8000ffbfdf68: =
>Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422
>>>>> #19 0x000000000047c7d4 in thread_start ()
>>>>> #20 0x0000000000000000 in ?? ()
>>>>>=20
>>>>> Thread 5 (Thread 800c09400 (LWP 101035/threadtest)):
>>>>> #0 0x000000000047e53c in _umtx_op_err ()
>>>>> #1 0x0000000000475f14 in __thr_umutex_lock ()
>>>>> #2 0x0000000000479404 in mutex_lock_common ()
>>>>> #3 0x000000000044dd15 in ThreadPThread__pthread_mutex_lock =
>(mutex=3DError
>>>>> accessing memory address 0x8000ffffc678: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:506
>>>>> #4 0x000000000044d039 in RTOS__LockHeap () at
>>>>> ../src/thread/PTHREAD/ThreadPThread.m3:1337
>>>>> ---Type <return> to continue, or q <return> to quit---
>>>>> #5 0x000000000042ff79 in RTAllocator__AllocTraced
>>>>> (M3_Cwb5VA_dataSize=3DError accessing memory address =
>0x8000ffffc6e8: Bad
>>>>> address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:365
>>>>> #6 0x000000000042f15b in RTAllocator__GetTracedObj =
>(M3_Eic7CK_def=3DError
>>>>> accessing memory address 0x8000ffffc7a8: Bad address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:224
>>>>> #7 0x000000000042eb23 in RTHooks__AllocateTracedObj
>>>>> (M3_AJWxb1_defn=3DError accessing memory address 0x8000ffffc7f8: =
>Bad address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:122
>>>>> #8 0x000000000045fbe4 in TextCat__Flat (M3_Bd56fi_LText=3DError =
>accessing
>>>>> memory address 0x8000ffffc858: Bad address.
>>>>> ) at ../src/text/TextCat.m3:562
>>>>> #9 0x000000000045ed5d in TextCat__Balance =
>(M3_Bd56fi_LText=3D0x800c490b0,
>>>>> M3_BUgnwf_LInfo=3DError accessing memory address 0x8000ffffc8f8: =
>Bad address.
>>>>> ) at ../src/text/TextCat.m3:488
>>>>> #10 0x000000000045d64f in RTHooks__Concat (M3_Bd56fi_t=3DError =
>accessing
>>>>> memory address 0x8000ffffcbb8: Bad address.
>>>>> ) at ../src/text/TextCat.m3:40
>>>>> #11 0x000000000040639e in Main_M3 (M3_AcxOUs_mode=3DError accessing =
>memory
>>>>> address 0x8000ffffcc38: Bad address.
>>>>> ) at ../src/Main.m3:593
>>>>> #12 0x000000000043d55d in RTLinker__RunMainBody (M3_DjPxE3_m=3DError
>>>>> accessing memory address 0x8000ffffcf88: Bad address.
>>>>> ) at ../src/runtime/common/RTLinker.m3:408
>>>>> #13 0x000000000043c8e8 in RTLinker__AddUnitI (M3_DjPxE3_m=3DError =
>accessing
>>>>> memory address 0x8000ffffd008: Bad address.
>>>>> ) at ../src/runtime/common/RTLinker.m3:115
>>>>> #14 0x000000000043c97c in RTLinker__AddUnit (M3_DjPxE5_b=3DError =
>accessing
>>>>> memory address 0x8000ffffd028: Bad address.
>>>>> ) at ../src/runtime/common/RTLinker.m3:124
>>>>> #15 0x00000000004004a6 in main (argc=3DError accessing memory =
>address
>>>>> 0x8000ffffd07c: Bad address.
>>>>> ) at _m3main.c:22
>>>>>=20
>>>>> Thread 9 (Thread 800c0a400 (LWP 100473/threadtest)):
>>>>> #0 0x000000000047e53c in _umtx_op_err ()
>>>>> #1 0x0000000000477e8a in check_suspend ()
>>>>> #2 0x00000000004780a2 in sigcancel_handler ()
>>>>> #3 <signal handler called>
>>>>> #4 0x000000000047e53c in _umtx_op_err ()
>>>>> #5 0x0000000000475f14 in __thr_umutex_lock ()
>>>>> #6 0x0000000000479404 in mutex_lock_common ()
>>>>> #7 0x000000000044dd15 in ThreadPThread__pthread_mutex_lock =
>(mutex=3DError
>>>>> accessing memory address 0x8000ff5fac18: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:506
>>>>> #8 0x000000000044d039 in RTOS__LockHeap () at
>>>>> ../src/thread/PTHREAD/ThreadPThread.m3:1337
>>>>> #9 0x000000000042ff79 in RTAllocator__AllocTraced
>>>>> (M3_Cwb5VA_dataSize=3DError accessing memory address =
>0x8000ff5fac88: Bad
>>>>> address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:365
>>>>> #10 0x000000000042f8a5 in RTAllocator__GetOpenArray =
>(M3_Eic7CK_def=3DError
>>>>> accessing memory address 0x8000ff5fad48: Bad address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:296
>>>>> #11 0x000000000042ebdf in RTHooks__AllocateOpenArray
>>>>> (M3_AJWxb1_defn=3DError accessing memory address 0x8000ff5fada8: =
>Bad address.
>>>>> ) at ../src/runtime/common/RTAllocator.m3:143
>>>>> #12 0x000000000040203a in Main__AApply (M3_AP7a1g_cl=3DError =
>accessing
>>>>> memory address 0x8000ff5fade8: Bad address.
>>>>> ) at ../src/Main.m3:283
>>>>> #13 0x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3DError
>>>>> accessing memory address 0x8000ff5faeb8: Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449
>>>>> #14 0x0000000000449c50 in ThreadPThread__ThreadBase
>>>>> (M3_AJWxb1_param=3DError accessing memory address 0x8000ff5faf68: =
>Bad address.
>>>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422
>>>>> #15 0x000000000047c7d4 in thread_start ()
>>>>> #16 0x0000000000000000 in ?? ()
>>>>>=20
>>>>> (others are similar)
>>>>>=20
>>>>> Hmm looks like a FreeBSD issue now. It's here...
>>>>>=20
>>>>> int
>>>>> __cdecl
>>>>> ThreadPThread__SuspendThread (m3_pthread_t mt)
>>>>> {
>>>>> ThreadFreeBSD__Fatal(pthread_suspend_np(PTHREAD_FROM_M3(mt)),
>>>>> "pthread_suspend_np");
>>>>> return 1;
>>>>> }
>>>>>=20
>>>>> Now this suspend can wait:
>>>>>=20
>>>>> static int
>>>>> suspend_common(struct pthread *curthread, struct pthread *thread,
>>>>> int waitok)
>>>>> {
>>>>> uint32_t tmp;
>>>>>=20
>>>>> while (thread->state !=3D PS_DEAD &&
>>>>> !(thread->flags & THR_FLAGS_SUSPENDED)) {
>>>>> thread->flags |=3D THR_FLAGS_NEED_SUSPEND;
>>>>> /* Thread is in creation. */
>>>>> if (thread->tid =3D=3D TID_TERMINATED)
>>>>> return (1);
>>>>> tmp =3D thread->cycle;
>>>>> _thr_send_sig(thread, SIGCANCEL);
>>>>> THR_THREAD_UNLOCK(curthread, thread);
>>>>> if (waitok) {
>>>>> _thr_umtx_wait_uint(&thread->cycle, tmp, =
>NULL,
>>>>> 0); <=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
>>>>> THR_THREAD_LOCK(curthread, thread);
>>>>> } else {
>>>>> THR_THREAD_LOCK(curthread, thread);
>>>>> return (0);
>>>>> }
>>>>> }
>>>>>=20
>>>>> return (1);
>>>>> }
>>>>>=20
>>>>> ... but what it can wait for I am not clear on.
>>>>>=20
>>>>> Do things work better on Linux?
>>>>>=20
>>>>> What is the status of the fork bug? Can it be worked around by =
>only
>>>>> forking via Process.Create with no mutexes held (outside of all =
>LOCK
>>>>> blocks)?
>>>>>=20
>>>>> Mika
>>>>>=20
>>>>>=20
>>>>>=20
>>>>> Peter McKinna writes:
>>>>>> --001a11c2ced4471a2e050079db82
>>>>>> Content-Type: text/plain; charset=3DUTF-8
>>>>>>=20
>>>>>> Mika
>>>>>>=20
>>>>>> I think you need to back out Tony's changes to fix the fork bug, =
>at
>>>>> least
>>>>>> for now. Need to reload from cvs version 1.262 for =
>ThreadPThread.m3 If
>>>>>> you're using git you're on your own.
>>>>>> Do a cvs log ThreadPThread.m3 for an explanation for some of the =
>design
>>>>>> principles.
>>>>>> Also if you use gdb then you need to set lanc c before backtraces =
>so at
>>>>>> least you can see address names and values even if they are =
>contorted you
>>>>>> can extract the M3 name in the parm lists. Also in gdb thread =
>apply all
>>>>> bt
>>>>>> gives all thread backtraces which can be handy to see whose got =
>the locks
>>>>>> held.
>>>>>>=20
>>>>>> Regards Peter
>>>>>>=20
>>>>>>=20
>>>>>>=20
>>>>>> On Wed, Aug 13, 2014 at 12:14 PM, <mika at async.caltech.edu> wrote:
>>>>>>=20
>>>>>>>=20
>>>>>>> Question... is there something odd about my pthreads? Are =
>pthreads
>>>>>>> normally reentrant? I didn't think so.
>>>>>>>=20
>>>>>>> My compiler is much happier with the following changes I already
>>>>> outlined:
>>>>>>>=20
>>>>>>> 1. a dirty, disgusting hack to keep from locking against myself =
>going
>>>>> from
>>>>>>> XWait with self.mutex locked to m.release().
>>>>>>>=20
>>>>>>> 2. the change to LockHeap I described in previous email
>>>>>>>=20
>>>>>>> But now...
>>>>>>>=20
>>>>>>> (gdb) cont
>>>>>>> Continuing.
>>>>>>> ERROR: pthread_mutex_lock:11
>>>>>>> ERROR: pthread_mutex_lock:11
>>>>>>>=20
>>>>>>> Program received signal SIGABRT, Aborted.
>>>>>>> 0x000000080107626a in thr_kill () from /lib/libc.so.7
>>>>>>> (gdb) where
>>>>>>> #0 0x000000080107626a in thr_kill () from /lib/libc.so.7
>>>>>>> #1 0x000000080113dac9 in abort () from /lib/libc.so.7
>>>>>>> #2 0x000000000071e37a in ThreadPThread__pthread_mutex_lock
>>>>> (mutex=3DError
>>>>>>> accessing memory address 0x8000ffffb508: Bad address.
>>>>>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:543
>>>>>>> #3 0x000000000071d48d in RTOS__LockHeap () at
>>>>>>> ../src/thread/PTHREAD/ThreadPThread.m3:1377
>>>>>>> #4 0x0000000000706b9d in RTHooks__CheckLoadTracedRef
>>>>> (M3_Af40ku_ref=3DError
>>>>>>> accessing memory address 0x8000ffffb568: Bad address.
>>>>>>> ) at ../src/runtime/common/RTCollector.m3:2234
>>>>>>> #5 0x000000000071d284 in ThreadPThread__AtForkParent () at
>>>>>>> ../src/thread/PTHREAD/ThreadPThread.m3:1348
>>>>>>> #6 0x0000000800df8733 in fork () from /lib/libthr.so.3
>>>>>>> #7 0x000000000070dd8b in RTProcess__Fork () at
>>>>>>> ../src/runtime/common/RTProcessC.c:152
>>>>>>> #8 0x00000000006c52f2 in ProcessPosixCommon__Create_ForkExec
>>>>>>> (M3_Bd56fi_cmd=3DError accessing memory address 0x8000ffffb6f8: =
>Bad
>>>>> address.
>>>>>>> ) at ../src/os/POSIX/ProcessPosixCommon.m3:75
>>>>>>> #9 0x00000000006c6c6c in Process__Create (M3_Bd56fi_cmd=3DError
>>>>> accessing
>>>>>>> memory address 0x8000ffffb7f8: Bad address.
>>>>>>> ) at ../src/os/POSIX/ProcessPosix.m3:21
>>>>>>> #10 0x00000000004d6826 in QMachine__FulfilExecPromise
>>>>> (M3_D6rRrg_ep=3DError
>>>>>>> accessing memory address 0x8000ffffb838: Bad address.
>>>>>>> ) at ../src/QMachine.m3:1666
>>>>>>> #11 0x00000000004d6220 in QMachine__ExecCommand =
>(M3_An02H2_t=3DError
>>>>>>> accessing memory address 0x8000ffffb9f8: Bad address.
>>>>>>> ) at ../src/QMachine.m3:1605
>>>>>>> #12 0x00000000004d537e in QMachine__DoTryExec (M3_An02H2_t=3DError
>>>>> accessing
>>>>>>> memory address 0x8000ffffbee8: Bad address.
>>>>>>> ) at ../src/QMachine.m3:1476
>>>>>>>=20
>>>>>>> What am I doing wrong here?
>>>>>>>=20
>>>>>>> The error doesn't look unreasonable! Looking more closely at the =
>code:
>>>>>>>=20
>>>>>>> First, AtForkPrepare has been called:
>>>>>>>=20
>>>>>>> PROCEDURE AtForkPrepare() =3D
>>>>>>> VAR me :=3D GetActivation();
>>>>>>> act: Activation;
>>>>>>> BEGIN
>>>>>>> PThreadLockMutex(slotsMu, ThisLine());
>>>>>>> PThreadLockMutex(perfMu, ThisLine());
>>>>>>> PThreadLockMutex(initMu, ThisLine()); (* InitMutex =3D>
>>>>>>> RegisterFinalCleanup =3D> LockHeap *)
>>>>>>> PThreadLockMutex(heapMu, ThisLine());
>>>>>>> PThreadLockMutex(activeMu, ThisLine()); (* LockHeap =3D>
>>>>> SuspendOthers
>>>>>>> =3D> activeMu *)
>>>>>>> (* Walk activations and lock all threads.
>>>>>>> * NOTE: We have initMu, activeMu, so slots won't change,
>>>>> conditions
>>>>>>> and
>>>>>>> * mutexes won't be initialized on-demand.
>>>>>>> *)
>>>>>>> act :=3D me;
>>>>>>> REPEAT
>>>>>>> PThreadLockMutex(act.mutex, ThisLine());
>>>>>>> act :=3D act.next;
>>>>>>> UNTIL act =3D me;
>>>>>>> END AtForkPrepare;
>>>>>>>=20
>>>>>>> a postcondition of this routine is that heapMu is locked.
>>>>>>>=20
>>>>>>> now we get into AtForkParent:
>>>>>>>=20
>>>>>>> PROCEDURE AtForkParent() =3D
>>>>>>> VAR me :=3D GetActivation();
>>>>>>> act: Activation;
>>>>>>> cond: Condition;
>>>>>>> BEGIN
>>>>>>> (* Walk activations and unlock all threads, conditions. *)
>>>>>>> act :=3D me;
>>>>>>> REPEAT
>>>>>>> cond :=3D slots[act.slot].join;
>>>>>>> IF cond # NIL THEN PThreadUnlockMutex(cond.mutex, =
>ThisLine())
>>>>> END;
>>>>>>> PThreadUnlockMutex(act.mutex, ThisLine());
>>>>>>> act :=3D act.next;
>>>>>>> UNTIL act =3D me;
>>>>>>> PThreadUnlockMutex(activeMu, ThisLine());
>>>>>>> PThreadUnlockMutex(heapMu, ThisLine());
>>>>>>> PThreadUnlockMutex(initMu, ThisLine());
>>>>>>> PThreadUnlockMutex(perfMu, ThisLine());
>>>>>>> PThreadUnlockMutex(slotsMu, ThisLine());
>>>>>>> END AtForkParent;
>>>>>>>=20
>>>>>>> We can see by inspecting the code that a necessary precondition =
>for
>>>>>>> this routine is that heapMu is locked! (Since it's going to =
>unlock it,
>>>>>>> it had BETTER be locked on entry.)
>>>>>>>=20
>>>>>>> But the cond :=3D ... causes a RTHooks.CheckLoadTracedRef
>>>>>>>=20
>>>>>>> which causes an RTOS.LockHeap
>>>>>>>=20
>>>>>>> the code of which we just saw:
>>>>>>>=20
>>>>>>> PROCEDURE LockHeap () =3D
>>>>>>> VAR self :=3D pthread_self();
>>>>>>> BEGIN
>>>>>>> WITH r =3D pthread_mutex_lock(heapMu,ThisLine()) DO <*ASSERT =
>r=3D0*>
>>>>> END;
>>>>>>> ...
>>>>>>>=20
>>>>>>> we try to lock heapMu. kaboom! No surprise there, really?
>>>>>>>=20
>>>>>>> Am I going about this totally the wrong way? Other people are =
>running
>>>>>>> Modula-3
>>>>>>> with pthreads, right? Right?? Somewhere out there in =
>m3devel-land?
>>>>>>>=20
>>>>>>> Mika
>>>>>>>=20
>>>>>>>=20
>>>>>>=20
>>>>>> --001a11c2ced4471a2e050079db82
>>>>>> Content-Type: text/html; charset=3DUTF-8
>>>>>> Content-Transfer-Encoding: quoted-printable
>>>>>>=20
>>>>>> <div dir=3D3D"ltr">Mika<div><br></div><div>=3DC2=3DA0 I think you =
>need to back
>>>>> ou=3D
>>>>>> t Tony's changes to fix the fork bug, at least for now. Need =
>to
>>>>> reload =3D
>>>>>> from cvs version 1.262 for ThreadPThread.m3 If you're using =
>git
>>>>> you'=3D
>>>>>> ;re on your own.</div>
>>>>>>=20
>>>>>> <div>=3DC2=3DA0 Do a cvs log ThreadPThread.m3 for an explanation =
>for some of
>>>>> th=3D
>>>>>> e design principles.=3DC2=3DA0</div><div>=3DC2=3DA0 Also if you =
>use gdb then you
>>>>> ne=3D
>>>>>> ed to set lanc c before backtraces so at least you can see address =
>names
>>>>> an=3D
>>>>>> d values even if they are contorted you can extract the M3 name in =
>the
>>>>> parm=3D
>>>>>> lists. Also in gdb thread apply all bt gives all thread backtraces
>>>>> which c=3D
>>>>>> an be handy to see whose got the locks held.</div>
>>>>>>=20
>>>>>> <div><br></div><div>Regards Peter</div><div><br></div><div
>>>>> class=3D3D"gmail_e=3D
>>>>>> xtra"><br><br><div class=3D3D"gmail_quote">On Wed, Aug 13, 2014 at =
>12:14
>>>>> PM, =3D
>>>>>> <span dir=3D3D"ltr"><<a href=3D3D"mailto:mika at async.caltech.edu"
>>>>> target=3D3D"=3D
>>>>>> _blank">mika at async.caltech.edu</a>></span> wrote:<br>
>>>>>>=20
>>>>>> <blockquote class=3D3D"gmail_quote" style=3D3D"margin:0 0 0
>>>>> .8ex;border-left:1p=3D
>>>>>> x #ccc solid;padding-left:1ex"><br>
>>>>>> Question... is there something odd about my pthreads? =3DC2=3DA0Are =
>pthreads
>>>>> no=3D
>>>>>> rmally reentrant? =3DC2=3DA0I didn't think so.<br>
>>>>>> <br>
>>>>>> My compiler is much happier with the following changes I already
>>>>> outlined:<=3D
>>>>>> br>
>>>>>> <br>
>>>>>> 1. a dirty, disgusting hack to keep from locking against myself =
>going
>>>>> from =3D
>>>>>> XWait with self.mutex locked to m.release().<br>
>>>>>> <br>
>>>>>> 2. the change to LockHeap I described in previous email<br>
>>>>>> <br>
>>>>>> But now...<br>
>>>>>> <br>
>>>>>> (gdb) cont<br>
>>>>>> Continuing.<br>
>>>>>> ERROR: pthread_mutex_lock:11<br>
>>>>>> ERROR: pthread_mutex_lock:11<br>
>>>>>> <br>
>>>>>> Program received signal SIGABRT, Aborted.<br>
>>>>>> 0x000000080107626a in thr_kill () from /lib/libc.so.7<br>
>>>>>> (gdb) where<br>
>>>>>> #0 =3DC2=3DA00x000000080107626a in thr_kill () from =
>/lib/libc.so.7<br>
>>>>>> #1 =3DC2=3DA00x000000080113dac9 in abort () from =
>/lib/libc.so.7<br>
>>>>>> #2 =3DC2=3DA00x000000000071e37a in =
>ThreadPThread__pthread_mutex_lock
>>>>> (mutex=3D3DE=3D
>>>>>> rror accessing memory address 0x8000ffffb508: Bad address.<br>
>>>>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:543<br>
>>>>>> #3 =3DC2=3DA00x000000000071d48d in RTOS__LockHeap () at
>>>>> ../src/thread/PTHREAD/T=3D
>>>>>> hreadPThread.m3:1377<br>
>>>>>> #4 =3DC2=3DA00x0000000000706b9d in RTHooks__CheckLoadTracedRef
>>>>> (M3_Af40ku_ref=3D
>>>>>> =3D3DError accessing memory address 0x8000ffffb568: Bad =
>address.<br>
>>>>>> ) at ../src/runtime/common/RTCollector.m3:2234<br>
>>>>>> #5 =3DC2=3DA00x000000000071d284 in ThreadPThread__AtForkParent () =
>at
>>>>> ../src/thr=3D
>>>>>> ead/PTHREAD/ThreadPThread.m3:1348<br>
>>>>>> #6 =3DC2=3DA00x0000000800df8733 in fork () from =
>/lib/libthr.so.3<br>
>>>>>> #7 =3DC2=3DA00x000000000070dd8b in RTProcess__Fork () at
>>>>> ../src/runtime/common/=3D
>>>>>> RTProcessC.c:152<br>
>>>>>> #8 =3DC2=3DA00x00000000006c52f2 in =
>ProcessPosixCommon__Create_ForkExec
>>>>> (M3_Bd56=3D
>>>>>> fi_cmd=3D3DError accessing memory address 0x8000ffffb6f8: Bad =
>address.<br>
>>>>>> ) at ../src/os/POSIX/ProcessPosixCommon.m3:75<br>
>>>>>> #9 =3DC2=3DA00x00000000006c6c6c in Process__Create =
>(M3_Bd56fi_cmd=3D3DError
>>>>> acces=3D
>>>>>> sing memory address 0x8000ffffb7f8: Bad address.<br>
>>>>>> ) at ../src/os/POSIX/ProcessPosix.m3:21<br>
>>>>>> #10 0x00000000004d6826 in QMachine__FulfilExecPromise
>>>>> (M3_D6rRrg_ep=3D3DError=3D
>>>>>> accessing memory address 0x8000ffffb838: Bad address.<br>
>>>>>> ) at ../src/QMachine.m3:1666<br>
>>>>>> #11 0x00000000004d6220 in QMachine__ExecCommand =
>(M3_An02H2_t=3D3DError
>>>>> access=3D
>>>>>> ing memory address 0x8000ffffb9f8: Bad address.<br>
>>>>>> ) at ../src/QMachine.m3:1605<br>
>>>>>> #12 0x00000000004d537e in QMachine__DoTryExec (M3_An02H2_t=3D3DError=
>
>>>>> accessin=3D
>>>>>> g memory address 0x8000ffffbee8: Bad address.<br>
>>>>>> ) at ../src/QMachine.m3:1476<br>
>>>>>> <br>
>>>>>> What am I doing wrong here?<br>
>>>>>> <br>
>>>>>> The error doesn't look unreasonable! =3DC2=3DA0Looking more =
>closely at
>>>>> the =3D
>>>>>> code:<br>
>>>>>> <br>
>>>>>> First, AtForkPrepare has been called:<br>
>>>>>> <br>
>>>>>> PROCEDURE AtForkPrepare() =3D3D<br>
>>>>>> =3DC2=3DA0 VAR me :=3D3D GetActivation();<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act: Activation;<br>
>>>>>> =3DC2=3DA0 BEGIN<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(slotsMu, ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(perfMu, ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(initMu, ThisLine()); (* =
>InitMutex =3D3D>
>>>>> Re=3D
>>>>>> gisterFinalCleanup =3D3D> LockHeap *)<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(heapMu, ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(activeMu, ThisLine()); (* =
>LockHeap
>>>>> =3D3D> S=3D
>>>>>> uspendOthers =3D3D> activeMu *)<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 (* Walk activations and lock all =
>threads.<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0* NOTE: We have initMu, activeMu, =
>so slots won't
>>>>> ch=3D
>>>>>> ange, conditions and<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0* mutexes won't be initialized =
>on-demand.<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0*)<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 act :=3D3D me;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 REPEAT<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(act.mutex, =
>ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act :=3D3D act.next;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 UNTIL act =3D3D me;<br>
>>>>>> =3DC2=3DA0 END AtForkPrepare;<br>
>>>>>> <br>
>>>>>> a postcondition of this routine is that heapMu is locked.<br>
>>>>>> <br>
>>>>>> now we get into AtForkParent:<br>
>>>>>> <br>
>>>>>> PROCEDURE AtForkParent() =3D3D<br>
>>>>>> =3DC2=3DA0 VAR me :=3D3D GetActivation();<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act: Activation;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 cond: Condition;<br>
>>>>>> =3DC2=3DA0 BEGIN<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 (* Walk activations and unlock all threads, =
>conditions.
>>>>> *)<br=3D
>>>>>>>=20
>>>>>> =3DC2=3DA0 =3DC2=3DA0 act :=3D3D me;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 REPEAT<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 cond :=3D3D =
>slots[act.slot].join;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 IF cond # NIL THEN =
>PThreadUnlockMutex(cond.mutex,
>>>>> This=3D
>>>>>> Line()) END;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(act.mutex, =
>ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act :=3D3D act.next;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 UNTIL act =3D3D me;<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(activeMu, =
>ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(heapMu, ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(initMu, ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(perfMu, ThisLine());<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(slotsMu, ThisLine());<br>
>>>>>> =3DC2=3DA0 END AtForkParent;<br>
>>>>>> <br>
>>>>>> We can see by inspecting the code that a necessary precondition =
>for<br>
>>>>>> this routine is that heapMu is locked! =3DC2=3DA0(Since it's =
>going to
>>>>> unloc=3D
>>>>>> k it,<br>
>>>>>> it had BETTER be locked on entry.)<br>
>>>>>> <br>
>>>>>> But the cond :=3D3D ... causes a RTHooks.CheckLoadTracedRef<br>
>>>>>> <br>
>>>>>> which causes an RTOS.LockHeap<br>
>>>>>> <br>
>>>>>> the code of which we just saw:<br>
>>>>>> <br>
>>>>>> PROCEDURE LockHeap () =3D3D<br>
>>>>>> =3DC2=3DA0 VAR self :=3D3D pthread_self();<br>
>>>>>> =3DC2=3DA0 BEGIN<br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 WITH r =3D3D =
>pthread_mutex_lock(heapMu,ThisLine()) DO
>>>>> <*ASSE=3D
>>>>>> RT r=3D3D0*> END;<br>
>>>>>> ...<br>
>>>>>> <br>
>>>>>> we try to lock heapMu. =3DC2=3DA0kaboom! =3DC2=3DA0No surprise =
>there, really?<br>
>>>>>> <br>
>>>>>> Am I going about this totally the wrong way? =3DC2=3DA0Other =
>people are
>>>>> running=3D
>>>>>> Modula-3<br>
>>>>>> with pthreads, right? =3DC2=3DA0Right?? =3DC2=3DA0Somewhere out =
>there in
>>>>> m3devel-la=3D
>>>>>> nd?<br>
>>>>>> <span><font color=3D3D"#888888"><br>
>>>>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0Mika<br>
>>>>>> <br>
>>>>>> </font></span></blockquote></div><br></div></div>
>>>>>>=20
>>>>>> --001a11c2ced4471a2e050079db82--
>>>>>=20
>>>>=20
>>>>=20
>>>=20
>>> --089e01183a1813188705007c3498
>>> Content-Type: text/html; charset=3DUTF-8
>>> Content-Transfer-Encoding: quoted-printable
>>>=20
>>> <div dir=3D3D"ltr">Oh, one other thing that seemed to help me in =
>testing this=3D
>>> stuff especially things to do with m3core and libm3 which seem to get =
>out =3D
>>> of whack sometimes, is to totally remove the target directory of =
>m3core the=3D
>>> n cm3; cm3 -ship Same with libm3. I'm not sure that clean =
>actually remo=3D
>>> ves the targets and I think there are leftover files that stuff =
>things up s=3D
>>> ince I would get segv in FileRd for no obvious reason.<div>
>>> <br></div><div>Peter</div><div><br></div></div><div =
>class=3D3D"gmail_extra"><=3D
>>> br><br><div class=3D3D"gmail_quote">On Wed, Aug 13, 2014 at 3:33 PM, =
>Peter Mc=3D
>>> Kinna <span dir=3D3D"ltr"><<a =
>href=3D3D"mailto:peter.mckinna at gmail.com" targ=3D
>>> et=3D3D"_blank">peter.mckinna at gmail.com</a>></span> wrote:<br>
>>> <blockquote class=3D3D"gmail_quote" style=3D3D"margin:0 0 0 =
>.8ex;border-left:1p=3D
>>> x #ccc solid;padding-left:1ex"><div dir=3D3D"ltr">That is weird. The =
>thread t=3D
>>> ester works fine with the old version of ThreadPThread.m3 on linux =
>amd_64 e=3D
>>> xcept if you have the fork test in the mix of tests you see the =
>pthread_mut=3D
>>> ex_destroy every now and again, the fork bug in fact.=3DC2=3DA0<div>
>>>=20
>>> I was testing the pthread changes Tony has made over the past few =
>days usin=3D
>>> g the thread tester program until I stupidly rebuilt the compiler and =
>was g=3D
>>> etting all sorts of hangs as it dragged in the new version of =
>pthreads. So =3D
>>> I rebuilt everything from a backup to get back where I was with a =
>decent co=3D
>>> mpiler, then reintroduced the latest changes to ThreadPThread.m3, =
>which yes=3D
>>> terday was just hanging in sigsuspend but today is giving me an =
>assert fail=3D
>>> ure at line 1387 in UnlockHeap. So its a bit unstable to say the =
>least. May=3D
>>> be there is something else wrong with freebsd.</div>
>>> <span class=3D3D"HOEnZb"><font color=3D3D"#888888">
>>> =
><div><br></div><div>Peter</div><div><br></div></font></span></div><div =
>clas=3D
>>> s=3D3D"HOEnZb"><div class=3D3D"h5"><div =
>class=3D3D"gmail_extra"><br><br><div clas=3D
>>> s=3D3D"gmail_quote">On Wed, Aug 13, 2014 at 2:54 PM, <span =
>dir=3D3D"ltr"><<=3D
>>> a href=3D3D"mailto:mika at async.caltech.edu" =
>target=3D3D"_blank">mika at async.calte=3D
>>> ch.edu</a>></span> wrote:<br>
>>>=20
>>> <blockquote class=3D3D"gmail_quote" style=3D3D"margin:0 0 0 =
>.8ex;border-left:1p=3D
>>> x #ccc solid;padding-left:1ex"><br>
>>> Yeah OK 1.262 makes a lot more sense to me but I'm still not =
>getting<br=3D
>>>>=20
>>> things to work and now I'm really baffled because I think I =
>understand<=3D
>>> br>
>>> this code!<br>
>>> <br>
>>> What I did was I threw away all my changes and reverted =
>ThreadPThread.m3<br=3D
>>>>=20
>>> to 1.262 as you suggested.<br>
>>> <br>
>>> Rebuilt the compiler with upgrade.sh<br>
>>> <br>
>>> Then rebuilt the compiler again with itself.<br>
>>> <br>
>>> Then I realcleaned the world and buildshipped it. =3DC2=3DA0(I was =
>afraid<br>
>>> that parseparams, also imported by the thread tester, would =
>pollute<br>
>>> it somehow.)<br>
>>> <br>
>>> When I look at the code in 1.262 it looks quite straightforward. =
>=3DC2=3DA0heap=3D
>>> Mu<br>
>>> is just a mutex, (Modula-3) mutexes are just (pthreads) mutexes. =
>=3DC2=3DA0Cond=3D
>>> ition<br>
>>> variables are mutexes too, but that's no big deal.<br>
>>> <br>
>>> So, rebuild thread tester, run it:<br>
>>> <br>
>>> new source -> compiling Main.m3<br>
>>> =3DC2=3DA0-> linking threadtest<br>
>>> root at rover:~mika/cm3-cvs-anon/cm3/m3-libs/m3core/tests/thread # =
>AMD64_FREEB=3D
>>> SD/threadtest<br>
>>> Writing file...done<br>
>>> Creating read threads...done<br>
>>> Creating fork threads...done<br>
>>> Creating alloc threads...done<br>
>>> Creating lock threads...done<br>
>>> running...printing oldest/median age/newest<br>
>>> .<br>
>>> <br>
>>> ***<br>
>>> *** runtime error:<br>
>>> *** =3DC2=3DA0 =3DC2=3DA0Segmentation violation - possible attempt to =
>dereference N=3D
>>> IL.........laziest thread is 1407901189/<a =
>href=3D3D"tel:1407901189%2F9" valu=3D
>>> e=3D3D"+14079011899" target=3D3D"_blank">1407901189/9</a> (tests: =
>read 14079011=3D
>>> 89/1407901189/1407901189 fork 1407901189/1407901189/1407901189 alloc =
>9/9/9 =3D
>>> lock 1407901189/1407901189/9)<br>
>>>=20
>>>=20
>>> <br>
>>> <br>
>>> <br>
>>> <br>
>>> ^C<br>
>>> root at rover:~mika/cm3-cvs-anon/cm3/m3-libs/m3core/tests/thread # gdb =
>!$<br>
>>> gdb AMD64_FREEBSD/threadtest<br>
>>> GNU gdb 6.1.1 [FreeBSD]<br>
>>> Copyright 2004 Free Software Foundation, Inc.<br>
>>> GDB is free software, covered by the GNU General Public License, and =
>you ar=3D
>>> e<br>
>>> welcome to change it and/or distribute copies of it under certain =
>condition=3D
>>> s.<br>
>>> Type "show copying" to see the conditions.<br>
>>> There is absolutely no warranty for GDB. =3DC2=3DA0Type "show =
>warranty&quo=3D
>>> t; for details.<br>
>>> This GDB was configured as "amd64-marcel-freebsd"...<br>
>>> (gdb) run<br>
>>> Starting program: =
>/big/home2/mika/2/cm3-cvs/cm3/m3-libs/m3core/tests/thread=3D
>>> /AMD64_FREEBSD/threadtest<br>
>>> [New LWP 100121]<br>
>>> Writing file...done<br>
>>> Creating read threads...done<br>
>>> Creating fork threads...done<br>
>>> Creating alloc threads...done<br>
>>> Creating lock threads...done<br>
>>> running...printing oldest/median age/newest<br>
>>> .[New Thread 801807000 (LWP 100343/threadtest)]<br>
>>> [New Thread 801807c00 (LWP 100667/threadtest)]<br>
>>> [New Thread 801808800 (LWP 100816/threadtest)]<br>
>>> [New Thread 801807400 (LWP 100349/threadtest)]<br>
>>> [New Thread 801808400 (LWP 100815/threadtest)]<br>
>>> [New Thread 801807800 (LWP 100352/threadtest)]<br>
>>> [New Thread 801808c00 (LWP 100817/threadtest)]<br>
>>> [New Thread 801809000 (LWP 100818/threadtest)]<br>
>>> [New Thread 801809800 (LWP 100820/threadtest)]<br>
>>> [New Thread 801808000 (LWP 100678/threadtest)]<br>
>>> [New Thread 801806400 (LWP 100121/threadtest)]<br>
>>> [New Thread 801806800 (LWP 100341/threadtest)]<br>
>>> [New Thread 801806c00 (LWP 100342/threadtest)]<br>
>>> .ERROR: pthread_mutex_lock:11<br>
>>> <div><br>
>>> Program received signal SIGABRT, Aborted.<br>
>>> </div>[Switching to Thread 801809800 (LWP 100820/threadtest)]<br>
>>> 0x0000000800d5c26a in thr_kill () from /lib/libc.so.7<br>
>>> (gdb) where<br>
>>> #0 =3DC2=3DA00x0000000800d5c26a in thr_kill () from =
>/lib/libc.so.7<br>
>>> #1 =3DC2=3DA00x0000000800e23ac9 in abort () from /lib/libc.so.7<br>
>>> #2 =3DC2=3DA00x000000000045101f in ThreadPThread__pthread_mutex_lock =
>(mutex=3D3DE=3D
>>> rror accessing memory address 0x8000fe5f2d98: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:513<br>
>>> #3 =3DC2=3DA00x000000000044b370 in ThreadPThread__LockMutex =
>(M3_AYIbX3_m=3D3DErro=3D
>>> r accessing memory address 0x8000fe5f2dc8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:119<br>
>>> #4 =3DC2=3DA00x0000000000405b5d in Main__LApply (M3_AP7a1g_cl=3D3DError=
> accessing=3D
>>> memory address 0x8000fe5f2df8: Bad address.<br>
>>> ) at ../src/Main.m3:319<br>
>>> #5 =3DC2=3DA00x000000000044d243 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3D3DErr=3D
>>> or accessing memory address 0x8000fe5f2eb8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449<br>
>>> #6 =3DC2=3DA00x000000000044cf00 in ThreadPThread__ThreadBase =
>(M3_AJWxb1_param=3D
>>> =3D3DError accessing memory address 0x8000fe5f2f68: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422<br>
>>> #7 =3DC2=3DA00x0000000800ad54a4 in pthread_create () from =
>/lib/libthr.so.3<br>
>>> #8 =3DC2=3DA00x0000000000000000 in ?? ()<br>
>>> (gdb) set lang c<br>
>>> (gdb) thread apply all bt<br>
>>> ... not that interesting ...<br>
>>> <br>
>>> Segfault is segfault---error 11 is EDEADLK (locking against =
>myself?)<br>
>>> <br>
>>> ????? the code is very straightforward, as I said.<br>
>>> <br>
>>> I thought people had the thread tester working with pthreads? =
>=3DC2=3DA0Which s=3D
>>> et of files, then? =3DC2=3DA0Anyone on FreeBSD/amd64?<br>
>>> <br>
>>> Could it be an issue with "volatile"? =3DC2=3DA0Not even =
>sure where t=3D
>>> o look.<br>
>>> <br>
>>> The code calling the lock is just this:<br>
>>> <br>
>>> PROCEDURE GetChar (rd: T): CHAR<br>
>>> =3DC2=3DA0 RAISES {EndOfFile, Failure, Alerted} =3D3D<br>
>>> =3DC2=3DA0 BEGIN<br>
>>> =3DC2=3DA0 =3DC2=3DA0 LOCK rd DO<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 RETURN FastGetChar(rd);<br>
>>> =3DC2=3DA0 =3DC2=3DA0 END<br>
>>> =3DC2=3DA0 END GetChar;<br>
>>> <br>
>>> No mysteries there...<br>
>>> <br>
>>> Ah is this the fork bug? =3DC2=3DA0Looks like the Init routine is =
>called on a<b=3D
>>> r>
>>> lot of NIL mutexes around the fork. =3DC2=3DA0But the subprocesses =
>aren't m=3D
>>> eant<br>
>>> to accesses the mutexes... humm<br>
>>> <br>
>>> Hmm, and it's not entirely a fork issue. =3DC2=3DA0Even with =
>"threadte=3D
>>> st -tests STD,-fork,-forktoomuch" it misbehaves. =
>=3DC2=3DA0Hangs....<br>
>>> <br>
>>> Looks like a deadlock this time.<br>
>>> <br>
>>> Some stacks...<br>
>>> <br>
>>> Thread 4 (Thread 800c0b400 (LWP 100477/threadtest)):<br>
>>> #0 =3DC2=3DA00x00000000004b4e2b in __vdso_gettimeofday ()<br>
>>> #1 =3DC2=3DA00x00000000004ab8d2 in gettimeofday ()<br>
>>> #2 =3DC2=3DA00x0000000000452e4b in TimePosix__Now () at =
>../src/time/POSIX/TimeP=3D
>>> osixC.c:50<br>
>>> #3 =3DC2=3DA00x0000000000452d72 in Time__Now () at =
>../src/time/POSIX/TimePosix.=3D
>>> m3:14<br>
>>> #4 =3DC2=3DA00x00000000004029a3 in Main__LApply (M3_AP7a1g_cl=3D3DError=
> accessing=3D
>>> memory address 0x8000fedf6df8: Bad address.<br>
>>> ) at ../src/Main.m3:327<br>
>>> #5 =3DC2=3DA00x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3D3DErr=3D
>>> or accessing memory address 0x8000fedf6eb8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449<br>
>>> #6 =3DC2=3DA00x0000000000449c50 in ThreadPThread__ThreadBase =
>(M3_AJWxb1_param=3D
>>> =3D3DError accessing memory address 0x8000fedf6f68: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422<br>
>>> #7 =3DC2=3DA00x000000000047c7d4 in thread_start ()<br>
>>> #8 =3DC2=3DA00x0000000000000000 in ?? ()<br>
>>> <br>
>>> Thread 3 (Thread 800c0b000 (LWP 100476/threadtest)):<br>
>>> #0 =3DC2=3DA00x000000000047e53c in _umtx_op_err ()<br>
>>> #1 =3DC2=3DA00x0000000000475f14 in __thr_umutex_lock ()<br>
>>> #2 =3DC2=3DA00x0000000000479404 in mutex_lock_common ()<br>
>>> #3 =3DC2=3DA00x000000000044dd15 in ThreadPThread__pthread_mutex_lock =
>(mutex=3D3DE=3D
>>> rror accessing memory address 0x8000feff7d98: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:506<br>
>>> #4 =3DC2=3DA00x00000000004480c0 in ThreadPThread__LockMutex =
>(M3_AYIbX3_m=3D3DErro=3D
>>> r accessing memory address 0x8000feff7dc8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:119<br>
>>> #5 =3DC2=3DA00x00000000004028ad in Main__LApply (M3_AP7a1g_cl=3D3DError=
> accessing=3D
>>> memory address 0x8000feff7df8: Bad address.<br>
>>> ) at ../src/Main.m3:319<br>
>>> #6 =3DC2=3DA00x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3D3DErr=3D
>>> or accessing memory address 0x8000feff7eb8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449<br>
>>> #7 =3DC2=3DA00x0000000000449c50 in ThreadPThread__ThreadBase =
>(M3_AJWxb1_param=3D
>>> =3D3DError accessing memory address 0x8000feff7f68: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422<br>
>>> #8 =3DC2=3DA00x000000000047c7d4 in thread_start ()<br>
>>> #9 =3DC2=3DA00x0000000000000000 in ?? ()<br>
>>> <br>
>>> Thread 6 (Thread 800c09800 (LWP 100470/threadtest)):<br>
>>> #0 =3DC2=3DA00x000000000047e53c in _umtx_op_err ()<br>
>>> #1 =3DC2=3DA00x0000000000475759 in suspend_common ()<br>
>>> #2 =3DC2=3DA00x00000000004755c1 in pthread_suspend_np ()<br>
>>> #3 =3DC2=3DA00x000000000044df0c in ThreadPThread__SuspendThread =
>(mt=3D3DError acc=3D
>>> essing memory address 0x8000ffbfd6f8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadFreeBSD.c:33<br>
>>> #4 =3DC2=3DA00x000000000044bade in ThreadPThread__StopThread =
>(M3_DMxDjQ_act=3D3DE=3D
>>> rror accessing memory address 0x8000ffbfd718: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:909<br>
>>> #5 =3DC2=3DA00x000000000044bc81 in ThreadPThread__StopWorld () at =
>../src/thread=3D
>>> /PTHREAD/ThreadPThread.m3:948<br>
>>> #6 =3DC2=3DA00x000000000044b19e in RTThread__SuspendOthers () at =
>../src/thread/=3D
>>> PTHREAD/ThreadPThread.m3:713<br>
>>> #7 =3DC2=3DA00x00000000004330d6 in =
>RTCollector__CollectSomeInStateZero () at ..=3D
>>> /src/runtime/common/RTCollector.m3:749<br>
>>> #8 =3DC2=3DA00x0000000000433081 in RTCollector__CollectSome () at =
>../src/runtim=3D
>>> e/common/RTCollector.m3:723<br>
>>> #9 =3DC2=3DA00x0000000000432d49 in RTHeapRep__CollectEnough () at =
>../src/runtim=3D
>>> e/common/RTCollector.m3:657<br>
>>> #10 0x000000000042ff83 in RTAllocator__AllocTraced =
>(M3_Cwb5VA_dataSize=3D3DEr=3D
>>> ror accessing memory address 0x8000ffbfd958: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:367<br>
>>> #11 0x000000000042f8a5 in RTAllocator__GetOpenArray =
>(M3_Eic7CK_def=3D3DError =3D
>>> accessing memory address 0x8000ffbfda18: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:296<br>
>>> #12 0x000000000042ebdf in RTHooks__AllocateOpenArray =
>(M3_AJWxb1_defn=3D3DErro=3D
>>> r accessing memory address 0x8000ffbfda78: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:143<br>
>>> #13 0x000000000040ca4e in Rd__NextBuff (M3_EkTcCb_rd=3D3DError =
>accessing memo=3D
>>> ry address 0x8000ffbfdab8: Bad address.<br>
>>> ) at ../src/rw/Rd.m3:159<br>
>>> #14 0x000000000040cd43 in UnsafeRd__FastGetChar (M3_EkTcCb_rd=3D3DError=
> acces=3D
>>> sing memory address 0x8000ffbfdb88: Bad address.<br>
>>> ) at ../src/rw/Rd.m3:187<br>
>>> #15 0x000000000040cc4a in Rd__GetChar (M3_EkTcCb_rd=3D3DError =
>accessing memor=3D
>>> y address 0x8000ffbfdbd8: Bad address.<br>
>>> ) at ../src/rw/Rd.m3:176<br>
>>> #16 0x000000000040095c in Main__RApply (M3_AP7a1g_cl=3D3DError =
>accessing memo=3D
>>> ry address 0x8000ffbfdc58: Bad address.<br>
>>> ) at ../src/Main.m3:185<br>
>>> <br>
>>> #17 0x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3D3DError ac=3D
>>> cessing memory address 0x8000ffbfdeb8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449<br>
>>> #18 0x0000000000449c50 in ThreadPThread__ThreadBase =
>(M3_AJWxb1_param=3D3DErro=3D
>>> r accessing memory address 0x8000ffbfdf68: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422<br>
>>> #19 0x000000000047c7d4 in thread_start ()<br>
>>> #20 0x0000000000000000 in ?? ()<br>
>>> <br>
>>> Thread 5 (Thread 800c09400 (LWP 101035/threadtest)):<br>
>>> #0 =3DC2=3DA00x000000000047e53c in _umtx_op_err ()<br>
>>> #1 =3DC2=3DA00x0000000000475f14 in __thr_umutex_lock ()<br>
>>> #2 =3DC2=3DA00x0000000000479404 in mutex_lock_common ()<br>
>>> #3 =3DC2=3DA00x000000000044dd15 in ThreadPThread__pthread_mutex_lock =
>(mutex=3D3DE=3D
>>> rror accessing memory address 0x8000ffffc678: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:506<br>
>>> #4 =3DC2=3DA00x000000000044d039 in RTOS__LockHeap () at =
>../src/thread/PTHREAD/T=3D
>>> hreadPThread.m3:1337<br>
>>> ---Type <return> to continue, or q <return> to =
>quit---<br>
>>> #5 =3DC2=3DA00x000000000042ff79 in RTAllocator__AllocTraced =
>(M3_Cwb5VA_dataSize=3D
>>> =3D3DError accessing memory address 0x8000ffffc6e8: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:365<br>
>>> #6 =3DC2=3DA00x000000000042f15b in RTAllocator__GetTracedObj =
>(M3_Eic7CK_def=3D3DE=3D
>>> rror accessing memory address 0x8000ffffc7a8: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:224<br>
>>> #7 =3DC2=3DA00x000000000042eb23 in RTHooks__AllocateTracedObj =
>(M3_AJWxb1_defn=3D
>>> =3D3DError accessing memory address 0x8000ffffc7f8: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:122<br>
>>> #8 =3DC2=3DA00x000000000045fbe4 in TextCat__Flat =
>(M3_Bd56fi_LText=3D3DError acces=3D
>>> sing memory address 0x8000ffffc858: Bad address.<br>
>>> ) at ../src/text/TextCat.m3:562<br>
>>> #9 =3DC2=3DA00x000000000045ed5d in TextCat__Balance =
>(M3_Bd56fi_LText=3D3D0x800c49=3D
>>> 0b0, M3_BUgnwf_LInfo=3D3DError accessing memory address =
>0x8000ffffc8f8: Bad a=3D
>>> ddress.<br>
>>> ) at ../src/text/TextCat.m3:488<br>
>>> #10 0x000000000045d64f in RTHooks__Concat (M3_Bd56fi_t=3D3DError =
>accessing me=3D
>>> mory address 0x8000ffffcbb8: Bad address.<br>
>>> ) at ../src/text/TextCat.m3:40<br>
>>> #11 0x000000000040639e in Main_M3 (M3_AcxOUs_mode=3D3DError accessing =
>memory =3D
>>> address 0x8000ffffcc38: Bad address.<br>
>>> ) at ../src/Main.m3:593<br>
>>> #12 0x000000000043d55d in RTLinker__RunMainBody (M3_DjPxE3_m=3D3DError =
>access=3D
>>> ing memory address 0x8000ffffcf88: Bad address.<br>
>>> ) at ../src/runtime/common/RTLinker.m3:408<br>
>>> #13 0x000000000043c8e8 in RTLinker__AddUnitI (M3_DjPxE3_m=3D3DError =
>accessing=3D
>>> memory address 0x8000ffffd008: Bad address.<br>
>>> ) at ../src/runtime/common/RTLinker.m3:115<br>
>>> #14 0x000000000043c97c in RTLinker__AddUnit (M3_DjPxE5_b=3D3DError =
>accessing =3D
>>> memory address 0x8000ffffd028: Bad address.<br>
>>> ) at ../src/runtime/common/RTLinker.m3:124<br>
>>> #15 0x00000000004004a6 in main (argc=3D3DError accessing memory =
>address 0x800=3D
>>> 0ffffd07c: Bad address.<br>
>>> ) at _m3main.c:22<br>
>>> <br>
>>> Thread 9 (Thread 800c0a400 (LWP 100473/threadtest)):<br>
>>> #0 =3DC2=3DA00x000000000047e53c in _umtx_op_err ()<br>
>>> #1 =3DC2=3DA00x0000000000477e8a in check_suspend ()<br>
>>> #2 =3DC2=3DA00x00000000004780a2 in sigcancel_handler ()<br>
>>> #3 =3DC2=3DA0<signal handler called><br>
>>> #4 =3DC2=3DA00x000000000047e53c in _umtx_op_err ()<br>
>>> #5 =3DC2=3DA00x0000000000475f14 in __thr_umutex_lock ()<br>
>>> #6 =3DC2=3DA00x0000000000479404 in mutex_lock_common ()<br>
>>> #7 =3DC2=3DA00x000000000044dd15 in ThreadPThread__pthread_mutex_lock =
>(mutex=3D3DE=3D
>>> rror accessing memory address 0x8000ff5fac18: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:506<br>
>>> #8 =3DC2=3DA00x000000000044d039 in RTOS__LockHeap () at =
>../src/thread/PTHREAD/T=3D
>>> hreadPThread.m3:1337<br>
>>> #9 =3DC2=3DA00x000000000042ff79 in RTAllocator__AllocTraced =
>(M3_Cwb5VA_dataSize=3D
>>> =3D3DError accessing memory address 0x8000ff5fac88: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:365<br>
>>> #10 0x000000000042f8a5 in RTAllocator__GetOpenArray =
>(M3_Eic7CK_def=3D3DError =3D
>>> accessing memory address 0x8000ff5fad48: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:296<br>
>>> #11 0x000000000042ebdf in RTHooks__AllocateOpenArray =
>(M3_AJWxb1_defn=3D3DErro=3D
>>> r accessing memory address 0x8000ff5fada8: Bad address.<br>
>>> ) at ../src/runtime/common/RTAllocator.m3:143<br>
>>> #12 0x000000000040203a in Main__AApply (M3_AP7a1g_cl=3D3DError =
>accessing memo=3D
>>> ry address 0x8000ff5fade8: Bad address.<br>
>>> ) at ../src/Main.m3:283<br>
>>> #13 0x0000000000449f93 in ThreadPThread__RunThread =
>(M3_DMxDjQ_me=3D3DError ac=3D
>>> cessing memory address 0x8000ff5faeb8: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:449<br>
>>> #14 0x0000000000449c50 in ThreadPThread__ThreadBase =
>(M3_AJWxb1_param=3D3DErro=3D
>>> r accessing memory address 0x8000ff5faf68: Bad address.<br>
>>> ) at ../src/thread/PTHREAD/ThreadPThread.m3:422<br>
>>> #15 0x000000000047c7d4 in thread_start ()<br>
>>> #16 0x0000000000000000 in ?? ()<br>
>>> <br>
>>> (others are similar)<br>
>>> <br>
>>> Hmm looks like a FreeBSD issue now. =3DC2=3DA0It's here...<br>
>>> <br>
>>> int<br>
>>> __cdecl<br>
>>> ThreadPThread__SuspendThread (m3_pthread_t mt)<br>
>>> {<br>
>>> =3DC2=3DA0 =
>ThreadFreeBSD__Fatal(pthread_suspend_np(PTHREAD_FROM_M3(mt)), "=3D
>>> pthread_suspend_np");<br>
>>> =3DC2=3DA0 return 1;<br>
>>> }<br>
>>> <br>
>>> Now this suspend can wait:<br>
>>> <br>
>>> static int<br>
>>> suspend_common(struct pthread *curthread, struct pthread *thread,<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 int waitok)<br>
>>> {<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 uint32_t tmp;<br>
>>> <br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 while (thread->state =
>!=3D3D PS_DEAD &&=3D
>>> <br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 !(thread->flags & T=3D
>>> HR_FLAGS_SUSPENDED)) {<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 thread->flags |=3D
>>> =3D3D THR_FLAGS_NEED_SUSPEND;<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 /* Thread is in cre=3D
>>> ation. */<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 if (thread->tid =3D
>>> =3D3D=3D3D TID_TERMINATED)<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3D
>>> =3DA0 =3DC2=3DA0 return (1);<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 tmp =3D3D thread->=3D
>>> cycle;<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 _thr_send_sig(threa=3D
>>> d, SIGCANCEL);<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 THR_THREAD_UNLOCK(c=3D
>>> urthread, thread);<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 if (waitok) {<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3D
>>> =3DA0 =3DC2=3DA0 _thr_umtx_wait_uint(&thread->cycle, tmp, =
>NULL, 0); =3DC2=3DA0=3D
>>> <=3D3D=3D3D=3D3D=3D3D=3D3D=3D3D=3D3D=3D3D=3D3D=3D3D<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3D
>>> =3DA0 =3DC2=3DA0 THR_THREAD_LOCK(curthread, thread);<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 } else {<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3D
>>> =3DA0 =3DC2=3DA0 THR_THREAD_LOCK(curthread, thread);<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3D
>>> =3DA0 =3DC2=3DA0 return (0);<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>=3DC2=3DA0 =3DC2=3DA0 }<br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 }<br>
>>> <br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 return (1);<br>
>>> }<br>
>>> <br>
>>> ... but what it can wait for I am not clear on.<br>
>>> <br>
>>> Do things work better on Linux?<br>
>>> <br>
>>> What is the status of the fork bug? =3DC2=3DA0Can it be worked around =
>by only f=3D
>>> orking via Process.Create with no mutexes held (outside of all LOCK =
>blocks)=3D
>>> ?<br>
>>> <br>
>>> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0Mika<br>
>>> <br>
>>> <br>
>>> <br>
>>> Peter McKinna writes:<br>
>>> >--001a11c2ced4471a2e050079db82<br>
>>> >Content-Type: text/plain; charset=3D3DUTF-8<br>
>>> <div><div>><br>
>>> >Mika<br>
>>> ><br>
>>> > =3DC2=3DA0I think you need to back out Tony's changes to fix =
>the fork =3D
>>> bug, at least<br>
>>> >for now. Need to reload from cvs version 1.262 for =
>ThreadPThread.m3 If<=3D
>>> br>
>>> >you're using git you're on your own.<br>
>>> > =3DC2=3DA0Do a cvs log ThreadPThread.m3 for an explanation for =
>some of the=3D
>>> design<br>
>>> >principles.<br>
>>> > =3DC2=3DA0Also if you use gdb then you need to set lanc c before =
>backtrace=3D
>>> s so at<br>
>>> >least you can see address names and values even if they are =
>contorted y=3D
>>> ou<br>
>>> >can extract the M3 name in the parm lists. Also in gdb thread =
>apply all=3D
>>> bt<br>
>>> >gives all thread backtraces which can be handy to see whose got =
>the loc=3D
>>> ks<br>
>>> >held.<br>
>>> ><br>
>>> >Regards Peter<br>
>>> ><br>
>>> ><br>
>>> ><br>
>>> >On Wed, Aug 13, 2014 at 12:14 PM, <<a =
>href=3D3D"mailto:mika at async.calt=3D
>>> ech.edu" target=3D3D"_blank">mika at async.caltech.edu</a>> =
>wrote:<br>
>>> ><br>
>>> >><br>
>>> >> Question... is there something odd about my pthreads? =
>=3DC2=3DA0Are pt=3D
>>> hreads<br>
>>> >> normally reentrant? =3DC2=3DA0I didn't think so.<br>
>>> >><br>
>>> >> My compiler is much happier with the following changes I =
>already o=3D
>>> utlined:<br>
>>> >><br>
>>> >> 1. a dirty, disgusting hack to keep from locking against =
>myself go=3D
>>> ing from<br>
>>> >> XWait with self.mutex locked to m.release().<br>
>>> >><br>
>>> >> 2. the change to LockHeap I described in previous email<br>
>>> >><br>
>>> >> But now...<br>
>>> >><br>
>>> >> (gdb) cont<br>
>>> >> Continuing.<br>
>>> >> ERROR: pthread_mutex_lock:11<br>
>>> >> ERROR: pthread_mutex_lock:11<br>
>>> >><br>
>>> >> Program received signal SIGABRT, Aborted.<br>
>>> >> 0x000000080107626a in thr_kill () from /lib/libc.so.7<br>
>>> >> (gdb) where<br>
>>> >> #0 =3DC2=3DA00x000000080107626a in thr_kill () from =
>/lib/libc.so.7<br>
>>> >> #1 =3DC2=3DA00x000000080113dac9 in abort () from =
>/lib/libc.so.7<br>
>>> >> #2 =3DC2=3DA00x000000000071e37a in =
>ThreadPThread__pthread_mutex_lock (=3D
>>> mutex=3D3DError<br>
>>> >> accessing memory address 0x8000ffffb508: Bad address.<br>
>>> >> ) at ../src/thread/PTHREAD/ThreadPThreadC.c:543<br>
>>> >> #3 =3DC2=3DA00x000000000071d48d in RTOS__LockHeap () at<br>
>>> >> ../src/thread/PTHREAD/ThreadPThread.m3:1377<br>
>>> >> #4 =3DC2=3DA00x0000000000706b9d in =
>RTHooks__CheckLoadTracedRef (M3_Af4=3D
>>> 0ku_ref=3D3DError<br>
>>> >> accessing memory address 0x8000ffffb568: Bad address.<br>
>>> >> ) at ../src/runtime/common/RTCollector.m3:2234<br>
>>> >> #5 =3DC2=3DA00x000000000071d284 in =
>ThreadPThread__AtForkParent () at<b=3D
>>> r>
>>> >> ../src/thread/PTHREAD/ThreadPThread.m3:1348<br>
>>> >> #6 =3DC2=3DA00x0000000800df8733 in fork () from =
>/lib/libthr.so.3<br>
>>> >> #7 =3DC2=3DA00x000000000070dd8b in RTProcess__Fork () at<br>
>>> >> ../src/runtime/common/RTProcessC.c:152<br>
>>> >> #8 =3DC2=3DA00x00000000006c52f2 in =
>ProcessPosixCommon__Create_ForkExec=3D
>>> <br>
>>> >> (M3_Bd56fi_cmd=3D3DError accessing memory address =
>0x8000ffffb6f8: Ba=3D
>>> d address.<br>
>>> >> ) at ../src/os/POSIX/ProcessPosixCommon.m3:75<br>
>>> >> #9 =3DC2=3DA00x00000000006c6c6c in Process__Create =
>(M3_Bd56fi_cmd=3D3DEr=3D
>>> ror accessing<br>
>>> >> memory address 0x8000ffffb7f8: Bad address.<br>
>>> >> ) at ../src/os/POSIX/ProcessPosix.m3:21<br>
>>> >> #10 0x00000000004d6826 in QMachine__FulfilExecPromise =
>(M3_D6rRrg_e=3D
>>> p=3D3DError<br>
>>> >> accessing memory address 0x8000ffffb838: Bad address.<br>
>>> >> ) at ../src/QMachine.m3:1666<br>
>>> >> #11 0x00000000004d6220 in QMachine__ExecCommand =
>(M3_An02H2_t=3D3DErr=3D
>>> or<br>
>>> >> accessing memory address 0x8000ffffb9f8: Bad address.<br>
>>> >> ) at ../src/QMachine.m3:1605<br>
>>> >> #12 0x00000000004d537e in QMachine__DoTryExec =
>(M3_An02H2_t=3D3DError=3D
>>> accessing<br>
>>> >> memory address 0x8000ffffbee8: Bad address.<br>
>>> >> ) at ../src/QMachine.m3:1476<br>
>>> >><br>
>>> >> What am I doing wrong here?<br>
>>> >><br>
>>> >> The error doesn't look unreasonable! =3DC2=3DA0Looking =
>more closel=3D
>>> y at the code:<br>
>>> >><br>
>>> >> First, AtForkPrepare has been called:<br>
>>> >><br>
>>> >> PROCEDURE AtForkPrepare() =3D3D<br>
>>> >> =3DC2=3DA0 VAR me :=3D3D GetActivation();<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act: Activation;<br>
>>> >> =3DC2=3DA0 BEGIN<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(slotsMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(perfMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(initMu, ThisLine()); =
>(* InitMutex =3D
>>> =3D3D><br>
>>> >> RegisterFinalCleanup =3D3D> LockHeap *)<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(heapMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(activeMu, =
>ThisLine()); (* LockHeap =3D
>>> =3D3D> SuspendOthers<br>
>>> >> =3D3D> activeMu *)<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 (* Walk activations and lock all =
>threads.<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0* NOTE: We have initMu, =
>activeMu, so slots won=3D
>>> 't change, conditions<br>
>>> >> and<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0* mutexes won't be =
>initialized on-demand.<=3D
>>> br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0*)<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 act :=3D3D me;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 REPEAT<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 PThreadLockMutex(act.mutex, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act :=3D3D act.next;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 UNTIL act =3D3D me;<br>
>>> >> =3DC2=3DA0 END AtForkPrepare;<br>
>>> >><br>
>>> >> a postcondition of this routine is that heapMu is =
>locked.<br>
>>> >><br>
>>> >> now we get into AtForkParent:<br>
>>> >><br>
>>> >> PROCEDURE AtForkParent() =3D3D<br>
>>> >> =3DC2=3DA0 VAR me :=3D3D GetActivation();<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act: Activation;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 cond: Condition;<br>
>>> >> =3DC2=3DA0 BEGIN<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 (* Walk activations and unlock all =
>threads, conditio=3D
>>> ns. *)<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 act :=3D3D me;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 REPEAT<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 cond :=3D3D =
>slots[act.slot].join;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 IF cond # NIL THEN =
>PThreadUnlockMutex(cond.mu=3D
>>> tex, ThisLine()) END;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 =
>PThreadUnlockMutex(act.mutex, ThisLine());<br=3D
>>>>=20
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0 act :=3D3D act.next;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 UNTIL act =3D3D me;<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(activeMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(heapMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(initMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(perfMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 PThreadUnlockMutex(slotsMu, =
>ThisLine());<br>
>>> >> =3DC2=3DA0 END AtForkParent;<br>
>>> >><br>
>>> >> We can see by inspecting the code that a necessary =
>precondition fo=3D
>>> r<br>
>>> >> this routine is that heapMu is locked! =3DC2=3DA0(Since =
>it's going=3D
>>> to unlock it,<br>
>>> >> it had BETTER be locked on entry.)<br>
>>> >><br>
>>> >> But the cond :=3D3D ... causes a =
>RTHooks.CheckLoadTracedRef<br>
>>> >><br>
>>> >> which causes an RTOS.LockHeap<br>
>>> >><br>
>>> >> the code of which we just saw:<br>
>>> >><br>
>>> >> PROCEDURE LockHeap () =3D3D<br>
>>> >> =3DC2=3DA0 VAR self :=3D3D pthread_self();<br>
>>> >> =3DC2=3DA0 BEGIN<br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 WITH r =3D3D =
>pthread_mutex_lock(heapMu,ThisLine()) DO =3D
>>> <*ASSERT r=3D3D0*> END;<br>
>>> >> ...<br>
>>> >><br>
>>> >> we try to lock heapMu. =3DC2=3DA0kaboom! =3DC2=3DA0No =
>surprise there, real=3D
>>> ly?<br>
>>> >><br>
>>> >> Am I going about this totally the wrong way? =3DC2=3DA0Other =
>people ar=3D
>>> e running<br>
>>> >> Modula-3<br>
>>> >> with pthreads, right? =3DC2=3DA0Right?? =3DC2=3DA0Somewhere =
>out there in m=3D
>>> 3devel-land?<br>
>>> >><br>
>>> >> =3DC2=3DA0 =3DC2=3DA0 =3DC2=3DA0Mika<br>
>>> >><br>
>>> >><br>
>>> ><br>
>>> </div></div>>--001a11c2ced4471a2e050079db82<br>
>>> >Content-Type: text/html; charset=3D3DUTF-8<br>
>>> >Content-Transfer-Encoding: quoted-printable<br>
>>> ><br>
>>> ><div =
>dir=3D3D3D"ltr">Mika<div><br></div>=3D
>>> ;<div>=3D3DC2=3D3DA0 I think you need to back ou=3D3D<br>
>>> >t Tony's changes to fix the fork bug, at least for now. =
>Need to=3D
>>> reload =3D3D<br>
>>> >from cvs version 1.262 for ThreadPThread.m3 If you're =
>using git=3D
>>> you'=3D3D<br>
>>> >;re on your own.</div><br>
>>> ><br>
>>> ><div>=3D3DC2=3D3DA0 Do a cvs log ThreadPThread.m3 for an =
>explanation =3D
>>> for some of th=3D3D<br>
>>> >e design principles.=3D3DC2=3D3DA0</div><div>=3D3DC2=3D=
>3DA0 Also if=3D
>>> you use gdb then you ne=3D3D<br>
>>> >ed to set lanc c before backtraces so at least you can see =
>address name=3D
>>> s an=3D3D<br>
>>> >d values even if they are contorted you can extract the M3 name =
>in the =3D
>>> parm=3D3D<br>
>>> > lists. Also in gdb thread apply all bt gives all thread =
>backtraces whi=3D
>>> ch c=3D3D<br>
>>> >an be handy to see whose got the locks held.</div><br>
>>> ><br>
>>> ><div><br></div><div>Regards =
>Peter</div>&l=3D
>>> t;div><br></div><div class=3D3D3D"gmail_e=3D3D<b=
>r>
>>> >xtra"><br><br><div =
>class=3D3D3D"gmail_quote&q=3D
>>> uot;>On Wed, Aug 13, 2014 at 12:14 PM, =3D3D<br>
>>> > <span dir=3D3D3D"ltr"><<a =
>href=3D3D3D"mailt=3D
>>> o:<a href=3D3D"mailto:mika at async.caltech.edu" =
>target=3D3D"_blank">mika at async.ca=3D
>>> ltech.edu</a>" target=3D3D3D"=3D3D<br>
>>> >_blank"><a href=3D3D"mailto:mika at async.caltech.edu" =
>target=3D3D"_bl=3D
>>> ank">mika at async.caltech.edu</a></a>></span> =
>wrote:<br=3D
>>> ><br>
>>> ><br>
>>> ><blockquote class=3D3D3D"gmail_quote" =
>style=3D3D3D"margin=3D
>>> :0 0 0 .8ex;border-left:1p=3D3D<br>
>>> >x #ccc solid;padding-left:1ex"><br><br>
>>> >Question... is there something odd about my pthreads? =
>=3D3DC2=3D3DA0Are pth=3D
>>> reads no=3D3D<br>
>>> >rmally reentrant? =3D3DC2=3D3DA0I didn't think =
>so.<br><br>
>>> ><br><br>
>>> >My compiler is much happier with the following changes I already =
>outlin=3D
>>> ed:<=3D3D<br>
>>> >br><br>
>>> ><br><br>
>>> >1. a dirty, disgusting hack to keep from locking against myself =
>going f=3D
>>> rom =3D3D<br>
>>> >XWait with self.mutex locked to m.release().<br><br>
>>> ><br><br>
>>> >2. the change to LockHeap I described in previous =
>email<br><br>
>>> ><br><br>
>>> >But now...<br><br>
>>> ><br><br>
>>> >(gdb) cont<br><br>
>>> >Continuing.<br><br>
>>> >ERROR: pthread_mutex_lock:11<br><br>
>>> >ERROR: pthread_mutex_lock:11<br><br>
>>> ><br><br>
>>> >Program received signal SIGABRT, Aborted.<br><br>
>>> >0x000000080107626a in thr_kill () from =
>/lib/libc.so.7<br><br>
>>> >(gdb) where<br><br>
>>> >#0 =3D3DC2=3D3DA00x000000080107626a in thr_kill () from =
>/lib/libc.so.7<b=3D
>>> r><br>
>>> >#1 =3D3DC2=3D3DA00x000000080113dac9 in abort () from =
>/lib/libc.so.7<br&g=3D
>>> t;<br>
>>> >#2 =3D3DC2=3D3DA00x000000000071e37a in =
>ThreadPThread__pthread_mutex_lock (m=3D
>>> utex=3D3D3DE=3D3D<br>
>>> >rror accessing memory address 0x8000ffffb508: Bad =
>address.<br><br=3D
>>>>=20
>>> >) at ../src/thread/PTHREAD/ThreadPThreadC.c:543<br><br>
>>> >#3 =3D3DC2=3D3DA00x000000000071d48d in RTOS__LockHeap () at =
>../src/thread/P=3D
>>> THREAD/T=3D3D<br>
>>> >hreadPThread.m3:1377<br><br>
>>> >#4 =3D3DC2=3D3DA00x0000000000706b9d in =
>RTHooks__CheckLoadTracedRef (M3_Af40=3D
>>> ku_ref=3D3D<br>
>>> >=3D3D3DError accessing memory address 0x8000ffffb568: Bad =
>address.<br&=3D
>>> gt;<br>
>>> >) at ../src/runtime/common/RTCollector.m3:2234<br><br>
>>> >#5 =3D3DC2=3D3DA00x000000000071d284 in =
>ThreadPThread__AtForkParent () at ..=3D
>>> /src/thr=3D3D<br>
>>> >ead/PTHREAD/ThreadPThread.m3:1348<br><br>
>>> >#6 =3D3DC2=3D3DA00x0000000800df8733 in fork () from =
>/lib/libthr.so.3<br&=3D
>>> gt;<br>
>>> >#7 =3D3DC2=3D3DA00x000000000070dd8b in RTProcess__Fork () at =
>../src/runtime=3D
>>> /common/=3D3D<br>
>>> >RTProcessC.c:152<br><br>
>>> >#8 =3D3DC2=3D3DA00x00000000006c52f2 in =
>ProcessPosixCommon__Create_ForkExec =3D
>>> (M3_Bd56=3D3D<br>
>>> >fi_cmd=3D3D3DError accessing memory address 0x8000ffffb6f8: Bad =
>address.&=3D
>>> lt;br><br>
>>> >) at ../src/os/POSIX/ProcessPosixCommon.m3:75<br><br>
>>> >#9 =3D3DC2=3D3DA00x00000000006c6c6c in Process__Create =
>(M3_Bd56fi_cmd=3D3D3DE=3D
>>> rror acces=3D3D<br>
>>> >sing memory address 0x8000ffffb7f8: Bad address.<br><br>
>>> >) at ../src/os/POSIX/ProcessPosix.m3:21<br><br>
>>> >#10 0x00000000004d6826 in QMachine__FulfilExecPromise =
>(M3_D6rRrg_ep=3D3D3=3D
>>> DError=3D3D<br>
>>> > accessing memory address 0x8000ffffb838: Bad =
>address.<br><br>
>>> >) at ../src/QMachine.m3:1666<br><br>
>>> >#11 0x00000000004d6220 in QMachine__ExecCommand =
>(M3_An02H2_t=3D3D3DError =3D
>>> access=3D3D<br>
>>> >ing memory address 0x8000ffffb9f8: Bad address.<br><br>
>>> >) at ../src/QMachine.m3:1605<br><br>
>>> >#12 0x00000000004d537e in QMachine__DoTryExec =
>(M3_An02H2_t=3D3D3DError ac=3D
>>> cessin=3D3D<br>
>>> >g memory address 0x8000ffffbee8: Bad address.<br><br>
>>> >) at ../src/QMachine.m3:1476<br><br>
>>> ><br><br>
>>> >What am I doing wrong here?<br><br>
>>> ><br><br>
>>> >The error doesn't look unreasonable! =3D3DC2=3D3DA0Looking =
>more clo=3D
>>> sely at the =3D3D<br>
>>> >code:<br><br>
>>> ><br><br>
>>> >First, AtForkPrepare has been called:<br><br>
>>> ><br><br>
>>> >PROCEDURE AtForkPrepare() =3D3D3D<br><br>
>>> >=3D3DC2=3D3DA0 VAR me :=3D3D3D GetActivation();<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 act: =
>Activation;<br><br>
>>> >=3D3DC2=3D3DA0 BEGIN<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadLockMutex(slotsMu, =
>ThisLine());<br><=3D
>>> br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadLockMutex(perfMu, =
>ThisLine());<br><b=3D
>>> r>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadLockMutex(initMu, =
>ThisLine()); (* InitMute=3D
>>> x =3D3D3D> Re=3D3D<br>
>>> >gisterFinalCleanup =3D3D3D> LockHeap *)<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadLockMutex(heapMu, =
>ThisLine());<br><b=3D
>>> r>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadLockMutex(activeMu, =
>ThisLine()); (* LockHe=3D
>>> ap =3D3D3D> S=3D3D<br>
>>> >uspendOthers =3D3D3D> activeMu *)<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 (* Walk activations and lock all =
>threads.<br&g=3D
>>> t;<br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0* NOTE: We have =
>initMu, activeMu, so sl=3D
>>> ots won't ch=3D3D<br>
>>> >ange, conditions and<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0* mutexes =
>won't be initialized =3D
>>> on-demand.<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0*)<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 act :=3D3D3D me;<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 REPEAT<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =
>PThreadLockMutex(act.mutex, ThisLine()=3D
>>> );<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 act :=3D3D3D =
>act.next;<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 UNTIL act =3D3D3D me;<br><br>
>>> >=3D3DC2=3D3DA0 END AtForkPrepare;<br><br>
>>> ><br><br>
>>> >a postcondition of this routine is that heapMu is =
>locked.<br><br>
>>> ><br><br>
>>> >now we get into AtForkParent:<br><br>
>>> ><br><br>
>>> >PROCEDURE AtForkParent() =3D3D3D<br><br>
>>> >=3D3DC2=3D3DA0 VAR me :=3D3D3D GetActivation();<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 act: =
>Activation;<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 cond: =
>Condition;<br><br>
>>> >=3D3DC2=3D3DA0 BEGIN<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 (* Walk activations and unlock all =
>threads, condi=3D
>>> tions. *)<br=3D3D<br>
>>> >><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 act :=3D3D3D me;<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 REPEAT<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 cond :=3D3D3D =
>slots[act.slot].join;<b=3D
>>> r><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 IF cond # NIL THEN =
>PThreadUnlockMutex(=3D
>>> cond.mutex, This=3D3D<br>
>>> >Line()) END;<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =
>PThreadUnlockMutex(act.mutex, ThisLine=3D
>>> ());<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0 act :=3D3D3D =
>act.next;<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 UNTIL act =3D3D3D me;<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadUnlockMutex(activeMu, =
>ThisLine());<br&g=3D
>>> t;<br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadUnlockMutex(heapMu, =
>ThisLine());<br>=3D
>>> <br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadUnlockMutex(initMu, =
>ThisLine());<br>=3D
>>> <br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadUnlockMutex(perfMu, =
>ThisLine());<br>=3D
>>> <br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 PThreadUnlockMutex(slotsMu, =
>ThisLine());<br>=3D
>>> ;<br>
>>> >=3D3DC2=3D3DA0 END AtForkParent;<br><br>
>>> ><br><br>
>>> >We can see by inspecting the code that a necessary precondition =
>for<=3D
>>> br><br>
>>> >this routine is that heapMu is locked! =3D3DC2=3D3DA0(Since =
>it's go=3D
>>> ing to unloc=3D3D<br>
>>> >k it,<br><br>
>>> >it had BETTER be locked on entry.)<br><br>
>>> ><br><br>
>>> >But the cond :=3D3D3D ... causes a =
>RTHooks.CheckLoadTracedRef<br><b=3D
>>> r>
>>> ><br><br>
>>> >which causes an RTOS.LockHeap<br><br>
>>> ><br><br>
>>> >the code of which we just saw:<br><br>
>>> ><br><br>
>>> >PROCEDURE LockHeap () =3D3D3D<br><br>
>>> >=3D3DC2=3D3DA0 VAR self :=3D3D3D pthread_self();<br><br>
>>> >=3D3DC2=3D3DA0 BEGIN<br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 WITH r =3D3D3D =
>pthread_mutex_lock(heapMu,ThisLine()=3D
>>> ) DO <*ASSE=3D3D<br>
>>> >RT r=3D3D3D0*> END;<br><br>
>>> >...<br><br>
>>> ><br><br>
>>> >we try to lock heapMu. =3D3DC2=3D3DA0kaboom! =3D3DC2=3D3DA0No =
>surprise there, r=3D
>>> eally?<br><br>
>>> ><br><br>
>>> >Am I going about this totally the wrong way? =3D3DC2=3D3DA0Other =
>people are=3D
>>> running=3D3D<br>
>>> > Modula-3<br><br>
>>> >with pthreads, right? =3D3DC2=3D3DA0Right?? =3D3DC2=3D3DA0Somewhere=
> out there i=3D
>>> n m3devel-la=3D3D<br>
>>> >nd?<br><br>
>>> ><span><font =
>color=3D3D3D"#888888"><br><br>
>>> >=3D3DC2=3D3DA0 =3D3DC2=3D3DA0 =3D3DC2=3D3DA0Mika<br><br>
>>> ><br><br>
>>> =
>></font></span></blockquote></div><br><=
>;=3D
>>> /div></div><br>
>>> ><br>
>>> >--001a11c2ced4471a2e050079db82--<br>
>>> </blockquote></div><br></div>
>>> </div></div></blockquote></div><br></div>
>>>=20
>>> --089e01183a1813188705007c3498--
More information about the M3devel
mailing list