From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ale.deltatee.com (ale.deltatee.com. [207.54.116.67]) by gmr-mx.google.com with ESMTPS id 23si458148ith.6.2017.06.19.13.27.48 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Mon, 19 Jun 2017 13:27:48 -0700 (PDT) References: <20170615203729.9009-1-logang@deltatee.com> <20170619200659.GA20437@kudzu.us> From: Logan Gunthorpe Message-ID: Date: Mon, 19 Jun 2017 14:27:41 -0600 MIME-Version: 1.0 In-Reply-To: <20170619200659.GA20437@kudzu.us> Content-Type: text/plain; charset=windows-1252 Content-Transfer-Encoding: 7bit Subject: Re: [RFC PATCH 00/13] Switchtec NTB Support To: Jon Mason Cc: linux-ntb@googlegroups.com, linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org, Dave Jiang , Allen Hubbe , Bjorn Helgaas , Greg Kroah-Hartman , Kurt Schwemmer , Stephen Bates List-ID: On 19/06/17 02:07 PM, Jon Mason wrote: > I think this code is of quality enough to go from an RFC to a standard > patch, and we can nit pick it to death there ;-) Thanks! > Please rebase on ntb-next (which I believe you are already doing), and > resbutmit. I'll try to get the rebase done and all the feedback so far applied by the end of the week and resend a v1. > I'm thinking that we'll want to keep this series all in one place. > So, #2 sounds like the best option. But, I need Bjorns $0.02 on this. I was thinking #2 was the best choice as well but really it's for you maintainers to decide. And, yes, we'd want to get Bjorn's ack. > FYI, I ran smatch on the patches and got this. Please correct them in > v2 (or v1 of the non-RFC...however you want to think of it). > drivers/pci/switch/switchtec.c:484 switchtec_dev_read() error: double unlock 'mutex:&stdev->mrpc_mutex' > drivers/pci/switch/switchtec.c:506 switchtec_dev_read() error: double unlock 'mutex:&stdev->mrpc_mutex' > drivers/pci/switch/switchtec.c:513 switchtec_dev_read() warn: inconsistent returns 'mutex:&stdev->mrpc_mutex'. This looks like a false positive to me. The code looks correct. smatch may have been confused by the fact that the lock is taken by two calls to the static function 'lock_mutex_and_test_alive'. This is also part of the switchtec management driver that's already in the kernel and not part of the NTB related patches I sent. Logan