From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from bombadil.infradead.org ([65.50.211.133]:58798 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750937AbdA1Jta (ORCPT ); Sat, 28 Jan 2017 04:49:30 -0500 Date: Fri, 27 Jan 2017 23:36:05 -0800 From: Christoph Hellwig To: Keith Busch Cc: Greg Kroah-Hartman , Lukas Wunner , linux-pci@vger.kernel.org, Bjorn Helgaas , Wei Zhang , Austin Bolen Subject: Re: [PATCHv4 next 0/3] Limiting pci access Message-ID: <20170128073605.GA20114@infradead.org> References: <20161208193253.GK25959@localhost.localdomain> <20161212234226.GA7973@bhelgaas-glaptop.roam.corp.google.com> <20161213005547.GA12844@localhost.localdomain> <20161213205012.GA29950@bhelgaas-glaptop.roam.corp.google.com> <20161213231840.GD12113@localhost.localdomain> <20170120213550.GA16618@localhost.localdomain> <20170121073140.GA26396@wunner.de> <20170121084243.GA24706@kroah.com> <20170123160452.GA10771@localhost.localdomain> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20170123160452.GA10771@localhost.localdomain> Sender: linux-pci-owner@vger.kernel.org List-ID: Or to put it the other way around - why even start an operation that will fail? We have a pretty similar issue in file system land: A file system must handle failure from the storage device at any time to provide data integrity. But that doesn't mean we can't optimize for the case where a device is toast (host removed, fails any I/O, etc). In that case we mark the file system as shutdown and stop submitting more I/O, potentially avoiding up to hours of recovery time. It would be helpful for everyone up and down the stack to do these don't do anyting stupid if you can avoid it starting from the lowest layers.