From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from gate.crashing.org (gate.crashing.org [63.228.1.57]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by ozlabs.org (Postfix) with ESMTPS id 761F22C0323 for ; Tue, 12 Feb 2013 19:53:47 +1100 (EST) Message-ID: <1360659217.2035.5.camel@pasglop> Subject: Re: [BUG] irq_dispose_mapping after irq request failure From: Benjamin Herrenschmidt To: Michael Ellerman Date: Tue, 12 Feb 2013 19:53:37 +1100 In-Reply-To: <20130212061822.GA8977@concordia> References: <20130211053100.GB18462@sapphire.tkos.co.il> <20130211061949.GA5561@concordia> <20130211205255.72B513E3530@localhost> <1360630273.2035.2.camel@pasglop> <20130212061822.GA8977@concordia> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 Cc: Baruch Siach , linuxppc-dev@lists.ozlabs.org, linux-kernel@vger.kernel.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Tue, 2013-02-12 at 17:18 +1100, Michael Ellerman wrote: > > I don't think you can, "active" is not well defined. Other code may have > done nothing other than create the mapping and remembered the virq, > which will break if you destroy the mapping. Or? Active as in "requested". Yes there's a potential problems with multiple requests for mappings & shared interrupts. This is not a problem for PCI on powerpc because we don't free those mappings afaik. > I agree refcounting is not fun. It'll end up with the same mess as > of_node_get/put() where practically every 2nd piece of code leaks > references. > > I guess we can't go the other way, and say that mapping the same hwirq > twice is an error. Might be worth it, and force the sharing case to be handled at some kind of upper level (bus or platform). Ben.