[erlang-questions] erlang nif resource & thread safety

Benoit Chesneau <>
Thu May 19 09:07:24 CEST 2016


thanks! the current behaviour is what i anted, though indeed having a way
to attach a resource to one owner would be useful :) Anyway thans again for
your answer.

Benoît


On Wed, 18 May 2016 at 14:49, Sverker Eriksson <
> wrote:

>
>
> On 05/18/2016 02:13 PM, Benoit Chesneau wrote:
>
>
>
> On Mon, May 2, 2016 at 4:24 PM Sverker Eriksson <
> > wrote:
>
>> A resource will be deallocated when
>>
>> * all terms created with enif_make_resource() have been garbage collected
>>
>> * AND the call to enif_alloc_resource() has been balanced by a call to
>> enif_release_resource()
>>
>> * AND all calls to enif_keep_resource() have been balanced by an equal
>> number of calls to enif_release_resource()
>>
>>
>>
>> On 05/02/2016 11:47 AM, Benoit Chesneau wrote:
>>
>> I had a look in the doc but I can't figure exactly how much a nif
>> resource could live outside the process that created it, ie. how others
>> processes can write/read to it and how to make sure the resource will be
>> accessible to them.
>>
>> So I created a nif resource that have the following properties:
>> - it  is ref counted, each time a call is made to it, the counter is
>> incremented and decremented when it return.
>>
>> A resource already has an internal ref counter, which is decremented by
>> the GC and by enif_release_resource().
>>
>> Adding your own ref counter is probably either redundant or just plain
>> wrong.
>> Every call to enif_release_resource() must be balanced by a prior call
>> to either enif_alloc_resource() or enif_keep_resource(). You cannot
>> release a resource that is still referred to by one of more terms.
>>
>>
>> - the process that create it will live until the vm is closed
>>
>> A resource it not tied to the process that created it,
>> it is tied to the terms that refer it.
>>
>
> Related to that I recently came with the following question:
> How to tie a nif resource to a owner process to make sure it's cleaned if
> the process crash?
>
> Looking at the source of ETS it seems to use unexposed functions from
> https://github.com/erlang/otp/blob/maint/erts/emulator/beam/erl_process.h
>
> I think I miss a C function like "nif_tie_resource(pid, somecalleback)" .
> Or is there any code that already does that?
>
> - benoît
>
>
>
> If the crashing process contains the last reference to your resource
> then it will be destructed as part of the process termination cleanup
> and you don't have to do anything.
>
> If you want your resource to be "cleaned" at process crash
> even though it is still referred by other processes or ETS tables,
> then there is no such callback today.
>
> You can accomplish that however by creating a second "owner" resource
> with a pointer to the "payload" resource. Keep the owner resource term
> in your owning process and let it clean the payload resource in its
> destructor.
> The destructor of the payload resource (if any) and all code doing
> enif_get_resource must recognize and handle an already cleaned resource.
>
>
> /Sverker, Erlang/OTP
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20160519/92a2dee8/attachment.html>


More information about the erlang-questions mailing list