From mboxrd@z Thu Jan 1 00:00:00 1970 From: jianchao.w.wang@oracle.com (jianchao.wang) Date: Wed, 18 Apr 2018 10:26:28 +0800 Subject: [PATCH 2/3] nvme: Sync queues on controller resets In-Reply-To: <20180212214601.GB20962@localhost.localdomain> References: <20180209174127.7224-1-keith.busch@intel.com> <20180209174127.7224-2-keith.busch@intel.com> <8a34b67a-1889-46d7-0658-b3834f22720f@oracle.com> <20180212214601.GB20962@localhost.localdomain> Message-ID: <2e46f418-ea62-eebf-f2a6-0627645dcfff@oracle.com> Hi Keith On 02/13/2018 05:46 AM, Keith Busch wrote: > On Sun, Feb 11, 2018@09:53:03AM +0800, jianchao.wang wrote: >> On 02/10/2018 09:55 AM, jianchao.wang wrote: >>> There could be a circular pattern here. Please consider the following scenario: >>> >>> timeout_work context reset_work context >>> nvme_timeout nvme_reset_work >>> -> nvme_dev_disable -> nvme_sync_queues // hold namespace_mutex >>> -> nvme_stop_queues -> blk_sync_queue >>> -> require namespaces_mutex -> cancel_work_sync(&q->timeout_work) >>> >> >> Looks like we could use rwsem to replace namespaces_mutex. > > Looks like rwsem is queued up for 4.17. I'll send an update based on > that. I guess this one and 3/3 can wait for 4.17, but 1/3 should still > go in 4.16. > Would you please queue this patch for next ? I incurred this issue when NVMe card died with a lot of in-flight requests. Thanks Jianchao