Hi Ioannis,
I am writing to you to say that I am really considering your problem.
Actually the VM does look into a strange state where it eats all the CPU
without going further anymore (though I am not sure yet at the moment I
have to analyze the dynamic of what it is doing). I am experiencing the
problem on both pthreads and jthreads. I will try to fix this asap. Give
me something like one week to know what happens. I am really sorry to
have missed your previous meesage in december but sometimes I parse the
ML quite quickly.
I keep you informed.
Regards,
Guilhem.
Ioannis Liverezas wrote:
Hi,
In my previous post
(http://www.kaffe.org/pipermail/kaffe/2005-December/103736.html), while
working with an arm processor based system, I have found a bug in the
threads system of kaffe-1.1.6. The bug exists in the combined use of
Thread.interrupt() and Thread.sleep(). The bug also exists and is worse
in the x86 architecture.
After having tried the test many times in both architectures, (and with
various VMs also), the results are:
--- the x86 architecture demonstrates the problem very fast in about
less than a minute.
--- ARM based systems also fail to pass the test, but it may take from 5
minutes to one day to hang.
In my previous post I have also included the source code of my test.
Yesterday, 9/1/2006, while trying the cvs version on x86, after running the same test
for half an hour, I took the following error:
Thread: Time1000 :sync :1 Alive:09/01/2006, 18:48:06.567
Test started at 09/01/2006, 18:14:26.086
kaffe-bin: jthread.c:2452: jmutex_unlock: Assertion `tid->status != 1' failed.
Aborted.
The two first lines are printouts from my test program. In previous versions of
kaffe, 1.1.6 or cvs, the test would hang with all threads sleeping.
- iliverez
_______________________________________________
kaffe mailing list
[email protected]
http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
_______________________________________________
kaffe mailing list
[email protected]
http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
