caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Goswin von Brederlow <goswin-v-b@web.de>
To: lyn HONG <lynxiamen@gmail.com>
Cc: caml-list@yquem.inria.fr
Subject: Re: [Caml-list] What should the "size" in "caml_alloc_custom" be?
Date: Tue, 18 May 2010 19:33:35 +0200	[thread overview]
Message-ID: <8739xp3yj4.fsf@frosties.localdomain> (raw)
In-Reply-To: <AANLkTimvPmK_QGMlfNt578bxCknDk7ZprA0iNoolC91x@mail.gmail.com> (lyn HONG's message of "Tue, 18 May 2010 10:58:39 -0400")

lyn HONG <lynxiamen@gmail.com> writes:

> On Tue, May 18, 2010 at 2:58 AM, Goswin von Brederlow <goswin-v-b@web.de>
> wrote:
>
>     lyn HONG <lynxiamen@gmail.com> writes:
>
>     > Hi all,
>     >
>     > I have a question about "allocating custom blocks" in "iterfacing C with
>     object
>     > Ocaml". when we call function "caml_alloc_custom(ops, size, used, max)"
>     in the
>     > C side, if the structure we want to allocate has a pointer, is the
>     "size" 
>     > going to be size of the structure itself only, or should we also include
>     the
>     > memory block that pointer points to?
>     >
>     > Thanks,
>     > Lin
>
>     The size is the number of ocaml words in the structure itself. The
>     pointer then points outside the ocaml heap to some C memory allocated by
>     malloc(). You should account for that size in the used/max pair.
>
>
> Thank you so much for the reply.
> So when we free the memory, what do we do with the block that pointer points
> to? Free it in the 'finalization function' associated to 'ops'?
>
> Best,
> Lin
>  
>
>     MfG
>            Goswin

That depends on who allocted the memory and who is supposed to free
it.

In some cases a library will free the chunk when some cleanup function
is called and your binding for cleanup() should then invalidate the
pointer and other bindings should verify the pointer is still valid
before using it. In that case the 'finalization function' can either
call cleanup() when the pointer is still valid when ocaml declares the
block as unused, give a warning that cleanup() wasn't called and call it
or even give an error. I like the warning best.

In other cases the chunk is something you are supposed to free when you
are done with it. Then you just free() it.

I really depends on the situation. Make damn sure that neigther the
ocaml nor C code accesses the chunk after it has been freed, ever.
Also make sure it doesn't get leaked.

MfG
        Goswin


  reply	other threads:[~2010-05-18 17:33 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-05-17 18:00 lyn HONG
2010-05-18  6:58 ` [Caml-list] " Goswin von Brederlow
2010-05-18 14:58   ` lyn HONG
2010-05-18 17:33     ` Goswin von Brederlow [this message]
2010-05-20 22:04 ` Richard Jones

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=8739xp3yj4.fsf@frosties.localdomain \
    --to=goswin-v-b@web.de \
    --cc=caml-list@yquem.inria.fr \
    --cc=lynxiamen@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).