threads branch status update
Andy Dougherty
doughera at lafayette.edu
Mon Apr 23 14:17:40 UTC 2012
On Mon, 23 Apr 2012, Stefan Seifert wrote:
> On Monday 23 April 2012 07:50:41 you wrote:
>
> > In this test, the main thread is stuck waiting in the pause() command in
> > Parrot_cx_outer_runloop(PARROT_INTERP). alarm_count = 1.
>
> Ah, could be the same thing that hit me this weekend. There's a comment in
> parrotinterpreter.pmc predicting exactly this situation but for some unknown
> reason it never hit me till now. As so often with threads it may just me a
> timing problem. Could you try the attached patch if it fixes the test for you?
>
> > Each of the 4 worker threads is stuck in an infinte loop here in
> > threads.c:
> >
> > Here's the function. The call to VTABLE_get_integer(interp, scheduler)
> > always returns 0. The next statement,
> > alarm_count = VTABLE_get_integer(interp, sched->alarms);
> > also returns 0, and the process repeats indefinitely.
>
> I should just leave out the if (alarm_count > 0). If a child thread has no
> active tasks, it should simply wait for new ones. But as it is, it's harmless
> except for uselessly burning CPU.
>
> > I didn't have time to investigate all the other failures; I figure I
> > should start with the simplest one first. If there's a simpler test
> > to look at, I could do so, but since this is my first foray into
> > debugging threads on Solaris, and since it's also my first foray into
> > threads on Parrot, it takes a long time to make any progress.
>
> Was the same for me :) But let's fix the simple ones, maybe the others will
> disappear as well.
I changed your patch slightly (<> instead of "", and used the symbolic
name SIGALRM instead of the non-portable number 14).
However, it didn't make any difference. This code is never reached
anyway.
diff --git a/src/pmc/parrotinterpreter.pmc b/src/pmc/parrotinterpreter.pmc
index be1efe9..9243c05 100644
--- a/src/pmc/parrotinterpreter.pmc
+++ b/src/pmc/parrotinterpreter.pmc
@@ -32,6 +32,8 @@ These are the vtable functions for the ParrotInterpreter base class
#include "pmc/pmc_sub.h"
#include "pmc/pmc_proxy.h"
#include "pmc/pmc_task.h"
+#include <signal.h>
+#include <unistd.h>
#define PMC_interp(x) ((Parrot_ParrotInterpreter_attributes *)PMC_data(x))->interp
#define PMC_args(x) ((Parrot_ParrotInterpreter_attributes *)PMC_data(x))->args
@@ -816,6 +818,11 @@ Schedules the given task on the proxy's interpreter.
Parrot_unblock_GC_mark_locked(proxied_interp);
/*TODO: investigate why it's not neccessary to send a signal to the process here
in theory the main thread could be pause()ing but in practice it seems to work anyway */
+#ifdef _WIN32
+ /* TODO: Implement on Windows */
+#else
+ kill(getpid(), SIGALRM);
+#endif
}
}
--
Andy Dougherty doughera at lafayette.edu
More information about the parrot-dev
mailing list