From owner-freebsd-java  Mon Aug 27 12:44:17 2001
Delivered-To: freebsd-java@freebsd.org
Received: from matsulab.is.titech.ac.jp (matsulab.is.titech.ac.jp [131.112.35.129])
	by hub.freebsd.org (Postfix) with ESMTP id 1E42D37B401
	for <java@FreeBSD.ORG>; Mon, 27 Aug 2001 12:44:12 -0700 (PDT)
	(envelope-from fuyuhik8@is.titech.ac.jp)
Received: from tripper.private by matsulab.is.titech.ac.jp (8.8.8+Sun/3.7W) id EAA21585; Tue, 28 Aug 2001 04:44:02 +0900 (JST)
Date: Tue, 28 Aug 2001 04:46:28 +0900
Message-ID: <55itf9cmln.wl@tripper.private>
From: Fuyuhiko Maruyama <fuyuhik8@is.titech.ac.jp>
To: Arun Sharma <arun@sharmas.dhs.org>
Cc: Fuyuhiko Maruyama <fuyuhik8@is.titech.ac.jp>, java@FreeBSD.ORG
Subject: Re: JVM and native threads
In-Reply-To: <20010827095624.B14875@sharmas.dhs.org>
References: <20010824233304.A32099@sharmas.dhs.org>
	<55vgjcqv87.wl@tripper.private>
	<20010825095746.A1015@sharmas.dhs.org>
	<55heuv0xtl.wl@tripper.private>
	<20010826171010.A6742@gnuppy>
	<20010827143244.A12305@misty.eyesbeyond.com>
	<55k7zqnd1x.wl@tripper.private>
	<20010827095624.B14875@sharmas.dhs.org>
User-Agent: Wanderlust/2.6.0 (Twist And Shout) on XEmacs/21.5.1 (anise)
MIME-Version: 1.0 (generated by SEMI 1.14.3 - "Ushinoya")
Content-Type: text/plain; charset=US-ASCII
Sender: owner-freebsd-java@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-java.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20freebsd-java>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20freebsd-java>
X-Loop: FreeBSD.org

At Mon, 27 Aug 2001 09:56:24 -0700,
Arun Sharma wrote:
> 
> On Mon, Aug 27, 2001 at 05:04:58PM +0900, Fuyuhiko Maruyama wrote:
> > What APIs are needed for ORP, Arun?  How do you use NGPT for ORP?
> 
> ORP was written for NT first and then ported to Linux. The approach
> they've taken is to emulate NT calls using the POSIX API:
> 
> WaitForSingleObject
> CreateEvent
> 
> See <orp>/os/<platform>/os_wrapper.c
> 
> However, the suspend and resume APIs haven't been implemented for 
> POSIX platforms.
> 
> #if defined(ORP_NT)
> 
> #define SUSPEND_THREAD(orpthread) SuspendThread((orpthread)->thread_handle)
> #define RESUME_THREAD(orpthread) ResumeThread((orpthread)->thread_handle)
> 
> #elif defined(ORP_POSIX)
> 
> // Linux asserts if these are called, signal logic TBD
> inline DWORD _enum_bad_call(){assert(0); return 1;}
> #define SUSPEND_THREAD(orpthread) _enum_bad_call()
> #define RESUME_THREAD(orpthread) _enum_bad_call()
> 
Thanks, it is interesting.

In fact, it seems quite curious for me because you say nothing about
thread's context.  So I dig the orp-20010727 sources and found that
current ORP doesn't well support Linux platform.  Current ORP may
cause disaster when GC is happen on Linux, possible senarios are:
1. GC system collects live object refered only by register.
2. GC system moves live objects refered by register and it doesn't
   treat pointers on register to catch up moving of objects.
3. GC system cannot ensure all threads are suspended at GC safepoint.

Do I misunderstand?

--
Fuyuhiko MARUYAMA <fuyuhik8@is.titech.ac.jp>
Matsuoka laboratory,
Department of Mathematical and Computing Sciences,
Graduate School of Information Science and Engineering,
Tokyo Institute of Technology.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-java" in the body of the message