From mboxrd@z Thu Jan 1 00:00:00 1970 From: Roland Dreier Subject: Re: [ofa-general][PATCH] mlx4_core: Synch catastrophic flow with module unload Date: Mon, 13 Jul 2009 12:45:32 -0700 Message-ID: References: <4A5B5274.2020801@mellanox.co.il> <20090713.111407.196601373.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: yevgenyp@mellanox.co.il, general@lists.openfabrics.org, netdev@vger.kernel.org To: David Miller Return-path: Received: from sj-iport-6.cisco.com ([171.71.176.117]:37611 "EHLO sj-iport-6.cisco.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754906AbZGMTpd (ORCPT ); Mon, 13 Jul 2009 15:45:33 -0400 In-Reply-To: <20090713.111407.196601373.davem@davemloft.net> (David Miller's message of "Mon, 13 Jul 2009 11:14:07 -0700 (PDT)") Sender: netdev-owner@vger.kernel.org List-ID: > Applied, thanks. Dave, please don't apply mlx4_core patches without giving me a chance to review them. In this case the patch looks buggy to me: I don't see how it handles, say, hot remove of one device -- it only handles module removal. And I would hope we could fix this without adding a global symbol as namespace polluting as "drv_mutex". Yevgeny didn't even send this patch to you; he just cc'ed netdev as a courtesy. However I understand that the physical location of mlx4_core in drivers/net makes it easy to do this. Maybe this is the best argument in favor of moving the mlx4_core stuff to drivers/shared? - R.