From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00
	autolearn=unavailable autolearn_force=no version=3.4.4
Path: 
 eternal-september.org!reader01.eternal-september.org!reader02.eternal-september.org!news.eternal-september.org!feeder.eternal-september.org!aioe.org!.POSTED!not-for-mail
From: "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>
Newsgroups: comp.lang.ada
Subject: Re: Read-write mutex sometimes fails on deadlock
Date: Mon, 6 Nov 2017 22:01:38 +0100
Organization: Aioe.org NNTP Server
Message-ID: <otqijg$1qrv$1@gioia.aioe.org>
References: <4502cd60-b843-4ace-99da-d716fc455635@googlegroups.com>
 <14ca8f28-a689-4431-9122-81a1c504569c@googlegroups.com>
NNTP-Posting-Host: MajGvm9MbNtGBKE7r8NgYA.user.gioia.aioe.org
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 7bit
X-Complaints-To: abuse@aioe.org
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101
 Thunderbird/52.4.0
X-Notice: Filtered by postfilter v. 0.8.2
Content-Language: en-US
Xref: news.eternal-september.org comp.lang.ada:48742
Date: 2017-11-06T22:01:38+01:00
List-Id: <comp.lang.ada>

On 2017-11-06 19:24, Robert Eachus wrote:
> On Saturday, October 28, 2017 at 4:02:33 PM UTC-4, pascal....@gmail.com wrote:
>>
>> I have a package that implements several kinds of mutexes, including a
>> read-write mutex. Here is the full source (spec and body) with only the
>> read-write part.
> 
> You got started down the wrong road and never turned around. The
> right abstraction for this in Ada is a protected object. A typical
> mutex looks like:
> 
> protected type Mutex is
>     entry Write(D: in Data):
>     function Read return Data;
> private
>     The_Data: Data := No_Data;
> end Mutex;

This is not mutex. Mutex is a semaphore with k=1.

> protected body Mutex is
>     entry Write (D: in Data) is
>     begin
>        The_Data := D;
>        Some_Data := True;
>        Print(D);

This is illegal, because Print is potentially blocking.

With a proper mutex printing is done outside any protected action. Here 
is a proper (not re-entrant, not read-write) mutex is:

    protected type Mutex is
       entry Seize;
       procedure Release;
    private
       Owned : Boolean := False;
    end Mutex;

    protected body Mutex is
       entry Seize when not Owned is
       begin
          Owned := True;
       end Seize;
       procedure Release is
       begin
          if Owned then
             Owned := False;
          else
             raise Use_Error with "Releasing not owned mutex";
          end if;
       end Release;
    end Mutex;

Printing goes as follows (an elaborated example):

    Lock : Mutex;
    ...

    select
       Lock.Seize;
    or delay Default_Timeout;
       raise Time_Error with "Printer mutex deadlocked";
    end select;
    begin
       Print (Something);
       Lock.Release;
    exception
       when others =>
          Lock.Release;
          raise;
    end;

P.S. It is advisable to have a controlled holder object with 
initialization seizing and finalization releasing mutex. Then printing 
goes Like:

    declare
       Exclusion : Holder (Lock'Access); -- Seize Lock
    begin
       Print (Something);
    end; -- Release Lock

P.P.S. Re-entrant mutex is one the same task may seize consequently.

P.P.P.S. Read-write mutex is one with two levels of access. Read access 
is allowed to several tasks. Write access is to only one. Read-write 
mutexes have issues with pro-/demotion operations, that is when a task 
owning the mutext changes the access level without releasing it.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de