Mailing-List: contact dev-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Received-SPF: pass (herse.apache.org: domain of gshimansky@gmail.com
 designates 64.233.182.186 as permitted sender)
DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
        s=beta; d=gmail.com;
        h=received:from:to:subject:date:user-agent:references:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:message-id;
        b=ecGrne0GA25KNeq79iO4b3DE2mra203GMDUCtwHEtpon0o5WOrvfK0Lr3p6UXtIRbljkdWvqAQfnKAKu7cOxAo/LM9ps35XKKLUAXmi1+pt8tKewr0/h/lfBPk02qCN2GUT4D5c8jMq2ShTplSFJ/p3sPNOa0prdag6jR5pyWm4=
From: Gregory Shimansky <gshimansky@gmail.com>
To: dev@harmony.apache.org
Subject: Re: svn commit: r479181 -
 /harmony/enhanced/drlvm/trunk/vm/vmcore/src/class_support/classloader.cpp
Date: Sun, 26 Nov 2006 05:08:16 +0300
User-Agent: KMail/1.9.1
References: <20061125195939.5167A1A9846@eris.apache.org>
 <4568E034.6000008@pobox.com>
In-Reply-To: <4568E034.6000008@pobox.com>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200611260508.17208.gshimansky@gmail.com>

On Sunday 26 November 2006 03:30 Geir Magnusson Jr. wrote:
> 1) Why add the SuspendDisabledChecker if not using it?
>
> 2) exactly where did you add the assertion?  :)

It is a hidden assertion class :)

Look at the file vm/vmcore/include/suspend_checker.h. There are 2 classes 
SuspendEnabledChecker and SuspendDisabledChecker. When declared, such 
variable in constructor checks for suspend status, in destructor it checks 
that the status is the same. It is often convenient to write just this one 
line to make sure that all returns from the function have the same suspend 
status because local variable destructor is executed on every function exit.

In release, when assert is a noop, these constructors/destructors are 
optimized into noop as well.

I saw that this function uses raw ManagedObject pointers. This is dangerous in 
case when suspend is enabled (equal to GC being enabled) as the object may be 
moved at any time. So I decided to add this assertion. If it fails some time 
it will signal that this function is called in a wrong unsafe mode.

> gshimansky@apache.org wrote:
> > Author: gshimansky
> > Date: Sat Nov 25 11:59:38 2006
> > New Revision: 479181
> >
> > URL: http://svn.apache.org/viewvc?view=rev&rev=479181
> > Log:
> > Fixed 32-bitness in classloader tracing. Added assertion before using a
> > raw heap object pointer
> >
> >
> > Modified:
> >    
> > harmony/enhanced/drlvm/trunk/vm/vmcore/src/class_support/classloader.cpp
> >
> > Modified:
> > harmony/enhanced/drlvm/trunk/vm/vmcore/src/class_support/classloader.cpp
> > URL:
> > http://svn.apache.org/viewvc/harmony/enhanced/drlvm/trunk/vm/vmcore/src/c
> >lass_support/classloader.cpp?view=diff&rev=479181&r1=479180&r2=479181
> > =========================================================================
> >===== ---
> > harmony/enhanced/drlvm/trunk/vm/vmcore/src/class_support/classloader.cpp
> > (original) +++
> > harmony/enhanced/drlvm/trunk/vm/vmcore/src/class_support/classloader.cpp
> > Sat Nov 25 11:59:38 2006 @@ -589,11 +589,13 @@
> >
> >  ClassLoader* ClassLoader::AddClassLoader( ManagedObject* loader )
> >  {
> > +    SuspendDisabledChecker sdc;
> > +
> >      LMAutoUnlock aulock( &(ClassLoader::m_tableLock) );
> >      ClassLoader* cl = new UserDefinedClassLoader();
> >      TRACE2("classloader.unloading.add", "Adding class loader "
> >          << cl << " (" << loader << " : "
> > -        << ((VTable*)(*(unsigned**)(loader)))->clss->get_name()->bytes
> > << ")"); +        << loader->vt()->clss->get_name()->bytes << ")");
> >      cl->Initialize( loader );
> >      if( m_capacity <= m_nextEntry )
> >          ReallocateTable( m_capacity?(2*m_capacity):32 );

-- 
Gregory