linux-kernel - Re: [PATCHv2 1/1] [tools]: android/ion: userspace test utility for ion buffer sharing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <10929681-0816-18d5-1bcb-5616462eecb2@redhat.com>
Date:   Wed, 4 Oct 2017 15:07:21 -0700
From:   Laura Abbott <labbott@...hat.com>
To:     Pintu Kumar <pintu.ping@...il.com>
Cc:     linux-kernel@...r.kernel.org,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Pintu Kumar <pintu_agarwal@...oo.com>
Subject: Re: [PATCHv2 1/1] [tools]: android/ion: userspace test utility for
 ion buffer sharing

On 10/04/2017 04:29 AM, Pintu Kumar wrote:
> The whole thing about this test is to share an FD over 2 independent processes.
> I think sharing an FD using fork() [parent/child] is not a real use
> case scenarios.
> Some people may not like the fork example.
> Initially when I started with ION, I also needed an FD sharing mechanism between
> 2 different process. Thus I came up with this framework using ipcsocket.
> Later, if required, we can even replace this with binder_ipc for
> android use cases.
> Anyways, binder_ipc also internally uses the same concept as this ipcsocket.
> To reduce the pain, we can invoke both the tests from a single shell scripts.
> I will try to include the same in kselftests, if possible.
> 
> If fork example is really required, we can add another test for it.
> This is my opinion.

You don't have to treat them as parent/child. Once they've forked
the processes are independent so you can use the same socket behavior
you've implemented here. I prefer the fork behavior over the shell
script because it means I only need to copy one file vs. three if
I'm testing in other environments but I'm not that opposed.

Thanks,
Laura