From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga12.intel.com (mga12.intel.com [192.55.52.136]) by dpdk.org (Postfix) with ESMTP id D1A471BE82 for ; Thu, 5 Jul 2018 08:23:05 +0200 (CEST) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga003.fm.intel.com ([10.253.24.29]) by fmsmga106.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 04 Jul 2018 23:23:05 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.51,310,1526367600"; d="scan'208,217";a="62397440" Received: from jguo15x-mobl3.ccr.corp.intel.com (HELO [10.67.68.63]) ([10.67.68.63]) by FMSMGA003.fm.intel.com with ESMTP; 04 Jul 2018 23:23:01 -0700 To: Thomas Monjalon References: <1498711073-42917-1-git-send-email-jia.guo@intel.com> <6033435.uqaPCSQc1q@xps> <2407710.IcmoihaDi7@xps> Cc: dev@dpdk.org, stephen@networkplumber.org, bruce.richardson@intel.com, ferruh.yigit@intel.com, konstantin.ananyev@intel.com, gaetan.rivet@6wind.com, jingjing.wu@intel.com, motih@mellanox.com, matan@mellanox.com, harry.van.haaren@intel.com, qi.z.zhang@intel.com, shaopeng.he@intel.com, bernard.iremonger@intel.com, jblunck@infradead.org, shreyansh.jain@nxp.com, helin.zhang@intel.com From: "Guo, Jia" Message-ID: <2f9bb7bb-0647-5601-7e1f-1eba44890843@intel.com> Date: Thu, 5 Jul 2018 14:23:00 +0800 User-Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:45.0) Gecko/20100101 Thunderbird/45.4.0 MIME-Version: 1.0 In-Reply-To: <2407710.IcmoihaDi7@xps> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit X-Content-Filtered-By: Mailman/MimeDel 2.1.15 Subject: Re: [dpdk-dev] [PATCH V4 1/9] bus: introduce hotplug failure handler X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 05 Jul 2018 06:23:06 -0000 On 7/4/2018 3:55 PM, Thomas Monjalon wrote: > 04/07/2018 09:16, Guo, Jia: >> On 7/4/2018 6:21 AM, Thomas Monjalon wrote: >>> 29/06/2018 12:30, Jeff Guo: >>>> /** >>>> + * Implementation a specific hot plug handler, which is responsible >>>> + * for handle the failure when hot remove the device, guaranty the system >>>> + * would not crash in the case. >>>> + * @param dev >>>> + * Pointer of the device structure. >>>> + * >>>> + * @return >>>> + * 0 on success. >>>> + * !0 on error. >>>> + */ >>>> +typedef int (*rte_bus_hotplug_handler_t)(struct rte_device *dev); >>> [...] >>>> @@ -211,6 +224,8 @@ struct rte_bus { >>>> rte_bus_parse_t parse; /**< Parse a device name */ >>>> struct rte_bus_conf conf; /**< Bus configuration */ >>>> rte_bus_get_iommu_class_t get_iommu_class; /**< Get iommu class */ >>>> + rte_bus_hotplug_handler_t hotplug_handler; >>>> + /**< handle hot plug on bus */ >>> The name is misleading. >>> It is to handle unplugging but is called "hotplug". >> ok, so i prefer hotplug_failure_handler than hot_unplug_handler, since >> it is more explicit for failure handle, and more clearly. >> >>> In order to demonstrate how the handler is used, you should >>> introduce the code using this handler in the same patch. >>> >> sorry, i check the history of rte_bus.h, and the way is introduce ops at >> first, second implement in specific bus, then come across the usage. >> I think that way clear and make sense. what do you think? >> Anyway, i will check the commit log if is there any misleading. > I think it is better to call ops when they are introduced, > and implement the ops in second step. > Hi, Thomas sorry but i want to detail the relationship of the ops and api as bellow to try if we can get the better sequence. Patch num: 1: introduce ops hotplug_failure_handler 2: implement ops hotplug_failure_handler 3:introduce ops sigbus_handler. 4:implement ops sigbus_handler 5: introduce helper rte_bus_sigbus_handler to call the ops sigbus_handler 6: introduce the mechanism to call helper rte_bus_sigbus_handler and call hotplug_failure_handler. If per you said , could I modify the sequence like 6->5->3->4->1->2? I don't think it will make sense, and might be more confused. And I think should be better that introduce each ops just say item, then when introduce the caller patch, the functional is ready to use by the patch. if i did not got your point and you have other better sequence about that please explicit to let me know. Thanks.