From mboxrd@z Thu Jan 1 00:00:00 1970 From: sagig@dev.mellanox.co.il (Sagi Grimberg) Date: Thu, 31 Dec 2015 19:19:37 +0200 Subject: [PATCH 3/5] NVMe: Remove queue freezing on resets In-Reply-To: <568441A4.3030007@dev.mellanox.co.il> References: <1451496471-29370-1-git-send-email-keith.busch@intel.com> <1451496471-29370-4-git-send-email-keith.busch@intel.com> <568441A4.3030007@dev.mellanox.co.il> Message-ID: <568563A9.1080901@dev.mellanox.co.il> >> NVMe submits all commands through the block layer now. This means we >> can let requests queue at the blk-mq hardware context since there is no >> path that bypasses this anymore. This means we don't need to freeze the >> queues anymore and simply stop the h/w queues from running during a >> reset. > > Does this have any dependencies? I tried this at some point but it > didn't work as I expected it to. OK, I figured out what I did wrong. >> This also fixes a WARN in percpu_ref_reinit when the queue was unfrozen >> with requeued requests. > > Oh yes, I've seen this one too! I'll test it for sure. This works great for me, Tested-by: Sagi Grimberg