From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga06.intel.com (mga06.intel.com [134.134.136.31]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ml01.01.org (Postfix) with ESMTPS id D7AE32115993B for ; Thu, 27 Sep 2018 08:27:56 -0700 (PDT) Subject: Re: [RFC workqueue/driver-core PATCH 3/5] driver core: Probe devices asynchronously instead of the driver References: <20180926214433.13512.30289.stgit@localhost.localdomain> <20180926215149.13512.51991.stgit@localhost.localdomain> From: Alexander Duyck Message-ID: <021d55fb-9f6a-0b52-3513-e9c5493bd7d7@linux.intel.com> Date: Thu, 27 Sep 2018 08:27:55 -0700 MIME-Version: 1.0 In-Reply-To: Content-Language: en-US List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="us-ascii"; Format="flowed" Errors-To: linux-nvdimm-bounces@lists.01.org Sender: "Linux-nvdimm" To: Dan Williams Cc: "Brown, Len" , Linux-pm mailing list , Greg KH , linux-nvdimm , jiangshanlai@gmail.com, Linux Kernel Mailing List , zwisler@kernel.org, Pavel Machek , Tejun Heo , Andrew Morton , "Rafael J. Wysocki" List-ID: On 9/26/2018 5:48 PM, Dan Williams wrote: > On Wed, Sep 26, 2018 at 2:51 PM Alexander Duyck > wrote: >> >> This change makes it so that we probe devices asynchronously instead of the >> driver. This results in us seeing the same behavior if the device is >> registered before the driver or after. This way we can avoid serializing >> the initialization should the driver not be loaded until after the devices >> have already been added. >> >> The motivation behind this is that if we have a set of devices that >> take a significant amount of time to load we can greatly reduce the time to >> load by processing them in parallel instead of one at a time. In addition, >> each device can exist on a different node so placing a single thread on one >> CPU to initialize all of the devices for a given driver can result in poor >> performance on a system with multiple nodes. >> >> One issue I can see with this patch is that I am using the >> dev_set/get_drvdata functions to store the driver in the device while I am >> waiting on the asynchronous init to complete. For now I am protecting it by >> using the lack of a dev->driver and the device lock. >> >> Signed-off-by: Alexander Duyck > [..] >> @@ -891,6 +914,25 @@ static int __driver_attach(struct device *dev, void *data) >> return ret; >> } /* ret > 0 means positive match */ >> >> + if (driver_allows_async_probing(drv)) { >> + /* >> + * Instead of probing the device synchronously we will >> + * probe it asynchronously to allow for more parallelism. >> + * >> + * We only take the device lock here in order to guarantee >> + * that the dev->driver and driver_data fields are protected >> + */ >> + dev_dbg(dev, "scheduling asynchronous probe\n"); >> + device_lock(dev); >> + if (!dev->driver) { >> + get_device(dev); >> + dev_set_drvdata(dev, drv); >> + async_schedule(__driver_attach_async_helper, dev); > > I'm not sure async drivers / sub-systems are ready for their devices > to show up in parallel. While userspace should not be relying on > kernel device names, people get upset when devices change kernel names > from one boot to the next, and I can see this change leading to that > scenario. The thing is the current async behavior already does this if the driver is loaded before the device is added. All I am doing is making the behavior with the driver loaded first the standard instead of letting it work the other way around. This way we get consistent behavior. > If a driver / sub-system wants more parallelism than what > driver_allows_async_probing() provides it should do it locally, for > example, like libata does. So where I actually saw this was with the pmem legacy setup I had. After doing all the work to parallelize things in the driver it had no effect. That was because the nd_pmem driver wasn't loaded yet so all the device_add calls did is add the device but didn't attach the nd_pmem driver. Then when the driver loaded it serialized the probe calls resulting in it taking twice as long as it needed to in order to initialize the memory. This seems to affect standard persistent memory as well. The only difference is that instead of probing the device on the first pass we kick it back and reprobe it in nd_pmem_probe/nd_pfn_probe in order to set the correct personality and that in turn allows us to asynchronously reschedule the work on the correct CPU and deserialize it. _______________________________________________ Linux-nvdimm mailing list Linux-nvdimm@lists.01.org https://lists.01.org/mailman/listinfo/linux-nvdimm From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-3.8 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,URIBL_BLOCKED autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id A7035C43382 for ; Thu, 27 Sep 2018 15:27:59 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 5FD8B215F0 for ; Thu, 27 Sep 2018 15:27:59 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 5FD8B215F0 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.intel.com Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-kernel-owner@vger.kernel.org Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727522AbeI0Vqo (ORCPT ); Thu, 27 Sep 2018 17:46:44 -0400 Received: from mga01.intel.com ([192.55.52.88]:28311 "EHLO mga01.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727212AbeI0Vqn (ORCPT ); Thu, 27 Sep 2018 17:46:43 -0400 X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from orsmga003.jf.intel.com ([10.7.209.27]) by fmsmga101.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 27 Sep 2018 08:27:56 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.54,311,1534834800"; d="scan'208";a="87166124" Received: from unknown (HELO [10.7.198.153]) ([10.7.198.153]) by orsmga003.jf.intel.com with ESMTP; 27 Sep 2018 08:27:56 -0700 Subject: Re: [RFC workqueue/driver-core PATCH 3/5] driver core: Probe devices asynchronously instead of the driver To: Dan Williams Cc: linux-nvdimm , Greg KH , Linux-pm mailing list , Linux Kernel Mailing List , Tejun Heo , Andrew Morton , "Brown, Len" , Dave Jiang , "Rafael J. Wysocki" , Vishal L Verma , jiangshanlai@gmail.com, Pavel Machek , zwisler@kernel.org References: <20180926214433.13512.30289.stgit@localhost.localdomain> <20180926215149.13512.51991.stgit@localhost.localdomain> From: Alexander Duyck Message-ID: <021d55fb-9f6a-0b52-3513-e9c5493bd7d7@linux.intel.com> Date: Thu, 27 Sep 2018 08:27:55 -0700 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101 Thunderbird/60.0 MIME-Version: 1.0 In-Reply-To: Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 9/26/2018 5:48 PM, Dan Williams wrote: > On Wed, Sep 26, 2018 at 2:51 PM Alexander Duyck > wrote: >> >> This change makes it so that we probe devices asynchronously instead of the >> driver. This results in us seeing the same behavior if the device is >> registered before the driver or after. This way we can avoid serializing >> the initialization should the driver not be loaded until after the devices >> have already been added. >> >> The motivation behind this is that if we have a set of devices that >> take a significant amount of time to load we can greatly reduce the time to >> load by processing them in parallel instead of one at a time. In addition, >> each device can exist on a different node so placing a single thread on one >> CPU to initialize all of the devices for a given driver can result in poor >> performance on a system with multiple nodes. >> >> One issue I can see with this patch is that I am using the >> dev_set/get_drvdata functions to store the driver in the device while I am >> waiting on the asynchronous init to complete. For now I am protecting it by >> using the lack of a dev->driver and the device lock. >> >> Signed-off-by: Alexander Duyck > [..] >> @@ -891,6 +914,25 @@ static int __driver_attach(struct device *dev, void *data) >> return ret; >> } /* ret > 0 means positive match */ >> >> + if (driver_allows_async_probing(drv)) { >> + /* >> + * Instead of probing the device synchronously we will >> + * probe it asynchronously to allow for more parallelism. >> + * >> + * We only take the device lock here in order to guarantee >> + * that the dev->driver and driver_data fields are protected >> + */ >> + dev_dbg(dev, "scheduling asynchronous probe\n"); >> + device_lock(dev); >> + if (!dev->driver) { >> + get_device(dev); >> + dev_set_drvdata(dev, drv); >> + async_schedule(__driver_attach_async_helper, dev); > > I'm not sure async drivers / sub-systems are ready for their devices > to show up in parallel. While userspace should not be relying on > kernel device names, people get upset when devices change kernel names > from one boot to the next, and I can see this change leading to that > scenario. The thing is the current async behavior already does this if the driver is loaded before the device is added. All I am doing is making the behavior with the driver loaded first the standard instead of letting it work the other way around. This way we get consistent behavior. > If a driver / sub-system wants more parallelism than what > driver_allows_async_probing() provides it should do it locally, for > example, like libata does. So where I actually saw this was with the pmem legacy setup I had. After doing all the work to parallelize things in the driver it had no effect. That was because the nd_pmem driver wasn't loaded yet so all the device_add calls did is add the device but didn't attach the nd_pmem driver. Then when the driver loaded it serialized the probe calls resulting in it taking twice as long as it needed to in order to initialize the memory. This seems to affect standard persistent memory as well. The only difference is that instead of probing the device on the first pass we kick it back and reprobe it in nd_pmem_probe/nd_pfn_probe in order to set the correct personality and that in turn allows us to asynchronously reschedule the work on the correct CPU and deserialize it. From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Duyck Subject: Re: [RFC workqueue/driver-core PATCH 3/5] driver core: Probe devices asynchronously instead of the driver Date: Thu, 27 Sep 2018 08:27:55 -0700 Message-ID: <021d55fb-9f6a-0b52-3513-e9c5493bd7d7@linux.intel.com> References: <20180926214433.13512.30289.stgit@localhost.localdomain> <20180926215149.13512.51991.stgit@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; Format="flowed" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Content-Language: en-US List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: linux-nvdimm-bounces-hn68Rpc1hR1g9hUCZPvPmw@public.gmane.org Sender: "Linux-nvdimm" To: Dan Williams Cc: "Brown, Len" , Linux-pm mailing list , Greg KH , linux-nvdimm , jiangshanlai-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org, Linux Kernel Mailing List , zwisler-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org, Pavel Machek , Tejun Heo , Andrew Morton , "Rafael J. Wysocki" List-Id: linux-pm@vger.kernel.org On 9/26/2018 5:48 PM, Dan Williams wrote: > On Wed, Sep 26, 2018 at 2:51 PM Alexander Duyck > wrote: >> >> This change makes it so that we probe devices asynchronously instead of the >> driver. This results in us seeing the same behavior if the device is >> registered before the driver or after. This way we can avoid serializing >> the initialization should the driver not be loaded until after the devices >> have already been added. >> >> The motivation behind this is that if we have a set of devices that >> take a significant amount of time to load we can greatly reduce the time to >> load by processing them in parallel instead of one at a time. In addition, >> each device can exist on a different node so placing a single thread on one >> CPU to initialize all of the devices for a given driver can result in poor >> performance on a system with multiple nodes. >> >> One issue I can see with this patch is that I am using the >> dev_set/get_drvdata functions to store the driver in the device while I am >> waiting on the asynchronous init to complete. For now I am protecting it by >> using the lack of a dev->driver and the device lock. >> >> Signed-off-by: Alexander Duyck > [..] >> @@ -891,6 +914,25 @@ static int __driver_attach(struct device *dev, void *data) >> return ret; >> } /* ret > 0 means positive match */ >> >> + if (driver_allows_async_probing(drv)) { >> + /* >> + * Instead of probing the device synchronously we will >> + * probe it asynchronously to allow for more parallelism. >> + * >> + * We only take the device lock here in order to guarantee >> + * that the dev->driver and driver_data fields are protected >> + */ >> + dev_dbg(dev, "scheduling asynchronous probe\n"); >> + device_lock(dev); >> + if (!dev->driver) { >> + get_device(dev); >> + dev_set_drvdata(dev, drv); >> + async_schedule(__driver_attach_async_helper, dev); > > I'm not sure async drivers / sub-systems are ready for their devices > to show up in parallel. While userspace should not be relying on > kernel device names, people get upset when devices change kernel names > from one boot to the next, and I can see this change leading to that > scenario. The thing is the current async behavior already does this if the driver is loaded before the device is added. All I am doing is making the behavior with the driver loaded first the standard instead of letting it work the other way around. This way we get consistent behavior. > If a driver / sub-system wants more parallelism than what > driver_allows_async_probing() provides it should do it locally, for > example, like libata does. So where I actually saw this was with the pmem legacy setup I had. After doing all the work to parallelize things in the driver it had no effect. That was because the nd_pmem driver wasn't loaded yet so all the device_add calls did is add the device but didn't attach the nd_pmem driver. Then when the driver loaded it serialized the probe calls resulting in it taking twice as long as it needed to in order to initialize the memory. This seems to affect standard persistent memory as well. The only difference is that instead of probing the device on the first pass we kick it back and reprobe it in nd_pmem_probe/nd_pfn_probe in order to set the correct personality and that in turn allows us to asynchronously reschedule the work on the correct CPU and deserialize it.