From: hesobreira <herrsobreira@gmail.com>
Subject: Re: Protected Objects - eggshell model
Date: Thu, 25 Dec 2008 06:29:01 -0800 (PST)
Date: 2008-12-25T06:29:01-08:00 [thread overview]
Message-ID: <d80f3b06-3602-4930-841e-a18928c5ff86@z6g2000pre.googlegroups.com> (raw)
In-Reply-To: 734a696c-f69f-4dc4-8c66-d01a3a468a19@z28g2000prd.googlegroups.com
It is all clear now.
Thank you (Adam and Bob) for the answers and thanks for the references
to the RM.
Hugo
On Dec 24, 1:43 am, Adam Beneschan <a...@irvine.com> wrote:
> On Dec 23, 2:08 pm, hesobreira <herrsobre...@gmail.com> wrote:
>
> > This was not clear to me while reading Barnes' book about protected
> > types.
>
> > There are two conditions that must be satisfied in order to execute
> > the entry call: no other task accessing the object at the same time
> > and the condition barrier evaluated to TRUE (eggshell, or two level
> > protection model). My question is: which one of these conditions is
> > evaluated first, the access lock or the condition barrier?
>
> > Thank you all in advance,
> > Hugo
>
> The access lock has to be obtained first, and in fact the task calling
> the protected entry has to have the protected object all to itself
> while it's evaluating the entry barrier. If the barrier is closed
> (False), the entry call is added to a queue, and *then* the lock is
> relinquished.
>
> In the RM, the concept of access locking is expressed by the phrase
> "protected actions". See 9.5.1(4-7) and 9.5.3(8-12).
>
> -- Adam
next prev parent reply other threads:[~2008-12-25 14:29 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-12-23 22:08 Protected Objects - eggshell model hesobreira
2008-12-24 0:43 ` Adam Beneschan
2008-12-25 14:29 ` hesobreira [this message]
2008-12-24 0:49 ` Robert A Duff
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox