From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnd Bergmann Subject: Re: [PATCH 2/3] net: msm_ipc: Add user-space support for MSM_IPC Router Date: Thu, 9 Jun 2011 22:15:49 +0200 Message-ID: <201106092215.49079.arnd@arndb.de> References: <1307555467-1530-1-git-send-email-kramasub@codeaurora.org> <4DF0FD8A.1000304@codeaurora.org> <20110609.123553.784125057506332397.davem@davemloft.net> Mime-Version: 1.0 Content-Type: Text/Plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Return-path: Received: from moutng.kundenserver.de ([212.227.17.10]:52218 "EHLO moutng.kundenserver.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755512Ab1FIUP6 (ORCPT ); Thu, 9 Jun 2011 16:15:58 -0400 In-Reply-To: <20110609.123553.784125057506332397.davem@davemloft.net> Sender: linux-arm-msm-owner@vger.kernel.org List-Id: linux-arm-msm@vger.kernel.org To: David Miller Cc: kramasub@codeaurora.org, netdev@vger.kernel.org, linux-arm-msm@vger.kernel.org On Thursday 09 June 2011 21:35:53 David Miller wrote: > From: Karthikeyan Ramasubramanian > Date: Thu, 09 Jun 2011 11:06:18 -0600 > > > Please let me know if you need any other information. > > I don't agree with creating all of this non-trivial infrastructure > for what seems to be an obscure protocol stack. > > You can do all of this in userspace, it seems. I can see some good use cases for having a socket to do IPC between OS images that have a shared memory interconnect. What I see as most problematic is to have it defined in a way that is tightly connected to specific hardware from one vendor. If the protocol was defined in a way that is useful for communication between virtual machines (e.g. on top of virtio) and tranparent PCI bridges, it would probably be much more acceptable. Arnd