From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Xie, Huawei" Subject: Re: [PATCH v4 1/5] vhost: eventfd_link: moving ioctl to a function Date: Thu, 18 Jun 2015 03:08:30 +0000 Message-ID: References: <1427123731-15654-1-git-send-email-pboldin@mirantis.com> <1427994080-10163-1-git-send-email-pboldin@mirantis.com> <1427994080-10163-2-git-send-email-pboldin@mirantis.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Cc: "dev@dpdk.org" To: Pavel Boldin Return-path: Received: from mga11.intel.com (mga11.intel.com [192.55.52.93]) by dpdk.org (Postfix) with ESMTP id BC6B0C450 for ; Thu, 18 Jun 2015 05:08:56 +0200 (CEST) Content-Language: en-US List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Two opens here, the trivial one is i think it is not good practice to=0A= explicitly inline non performance critical functions in c file, even if=0A= it will be done by compiler anyway.=0A= The critical one i have concern is whether it will introduce=0A= inconsistency if we call fd_install on a fd that is just closed by=0A= sys_close, because that fd will be set to next-to-be-allocated fd. I=0A= prefer to keep the original logic in patch 4/5 if we are not clear.=0A= =0A= As we actually don't need to use the eventfd that is allocated in user=0A= space at all, future patch would be: directly allocate a new fd in the=0A= kernel and call fd_install on it.=0A=