From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga05.intel.com (mga05.intel.com [192.55.52.43]) by dpdk.org (Postfix) with ESMTP id 802941BE0 for ; Wed, 7 Nov 2018 06:49:10 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga002.jf.intel.com ([10.7.209.21]) by fmsmga105.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 06 Nov 2018 21:49:09 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.54,474,1534834800"; d="scan'208";a="106558163" Received: from jguo15x-mobl.ccr.corp.intel.com (HELO [10.67.68.85]) ([10.67.68.85]) by orsmga002.jf.intel.com with ESMTP; 06 Nov 2018 21:49:07 -0800 To: Matan Azrad , "konstantin.ananyev@intel.com" , "anatoly.burakov@intel.com" , Thomas Monjalon , "bernard.iremonger@intel.com" , "jingjing.wu@intel.com" , "wenzhuo.lu@intel.com" Cc: "ferruh.yigit@intel.com" , "dev@dpdk.org" , "helin.zhang@intel.com" , "shaopeng.he@intel.com" References: <1541484436-91320-1-git-send-email-jia.guo@intel.com> <1541484436-91320-2-git-send-email-jia.guo@intel.com> From: Jeff Guo Message-ID: Date: Wed, 7 Nov 2018 13:49:06 +0800 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101 Thunderbird/60.2.1 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 8bit Content-Language: en-US Subject: Re: [dpdk-dev] [PATCH 1/3] eal: fix lock issue for hot-unplug 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: Wed, 07 Nov 2018 05:49:11 -0000 hi matan On 11/6/2018 2:22 PM, Matan Azrad wrote: > Hi Jeff > > Can you detail more in the commit log that we can understand the deadlock scenario. And how does this commit fix it? Before i add more detail in the commit log of next version, i would explain to you here at first here. When the device be hot-unplugged,  the hot-unplug handler will be invoked and the device will be detached, at this time if the interrupt still not disable soon and the second remove event come again(kernel will sent pci remove event after sent uio remove event) , the bus->find_device will return null and return, at this place lack of an unlock. Without this unlock, it will block the next remove or add event detection. So it definitely need an unlock here to avoid dead lock. >> -----Original Message----- >> From: Jeff Guo >> Sent: Tuesday, November 6, 2018 8:07 AM >> To: konstantin.ananyev@intel.com; anatoly.burakov@intel.com; Thomas >> Monjalon ; bernard.iremonger@intel.com; >> jingjing.wu@intel.com; wenzhuo.lu@intel.com >> Cc: ferruh.yigit@intel.com; dev@dpdk.org; jia.guo@intel.com; >> helin.zhang@intel.com; Matan Azrad ; >> shaopeng.he@intel.com >> Subject: [PATCH 1/3] eal: fix lock issue for hot-unplug >> >> This patch will add missing unlock for hot-unplug handler, without this patch >> potential dead lock will occur when device be hotplug-in after device be hot- >> unplugged. >> >> Fixes: 0fc54536b14a ("eal: add failure handling for hot-unplug") >> Signed-off-by: Jeff Guo >> --- >> lib/librte_eal/linuxapp/eal/eal_dev.c | 12 ++++++++---- >> 1 file changed, 8 insertions(+), 4 deletions(-) >> >> diff --git a/lib/librte_eal/linuxapp/eal/eal_dev.c >> b/lib/librte_eal/linuxapp/eal/eal_dev.c >> index d589c69..2830c86 100644 >> --- a/lib/librte_eal/linuxapp/eal/eal_dev.c >> +++ b/lib/librte_eal/linuxapp/eal/eal_dev.c >> @@ -258,7 +258,7 @@ dev_uev_handler(__rte_unused void *param) >> if (bus == NULL) { >> RTE_LOG(ERR, EAL, "Cannot find bus (%s)\n", >> busname); >> - return; >> + goto failure_handle_err; >> } >> >> dev = bus->find_device(NULL, cmp_dev_name, @@ >> -266,19 +266,23 @@ dev_uev_handler(__rte_unused void *param) >> if (dev == NULL) { >> RTE_LOG(ERR, EAL, "Cannot find device (%s) >> on " >> "bus (%s)\n", uevent.devname, >> busname); >> - return; >> + goto failure_handle_err; >> } >> >> ret = bus->hot_unplug_handler(dev); >> - rte_spinlock_unlock(&failure_handle_lock); >> if (ret) { >> RTE_LOG(ERR, EAL, "Can not handle hot- >> unplug " >> "for device (%s)\n", dev->name); >> - return; >> } >> + rte_spinlock_unlock(&failure_handle_lock); >> } >> rte_dev_event_callback_process(uevent.devname, >> uevent.type); >> } >> + >> + return; >> + >> +failure_handle_err: >> + rte_spinlock_unlock(&failure_handle_lock); >> } >> >> int __rte_experimental >> -- >> 2.7.4