From mboxrd@z Thu Jan 1 00:00:00 1970 From: ming.lei@redhat.com (Ming Lei) Date: Fri, 28 Dec 2018 06:16:33 +0800 Subject: [PATCH 2/2] nvme pci: try to allocate multiple irq vectors again in case of -EINVAL In-Reply-To: <20181227130834.GA22967@lst.de> References: <20181226103755.2101-1-ming.lei@redhat.com> <20181226103755.2101-3-ming.lei@redhat.com> <20181226182027.GA5866@lst.de> <20181227082136.GA14423@ming.t460p> <20181227130834.GA22967@lst.de> Message-ID: <20181227221631.GA22073@ming.t460p> On Thu, Dec 27, 2018@02:08:34PM +0100, Christoph Hellwig wrote: > On Thu, Dec 27, 2018@04:21:38PM +0800, Ming Lei wrote: > > On Wed, Dec 26, 2018@07:20:27PM +0100, Christoph Hellwig wrote: > > > On Wed, Dec 26, 2018@06:37:55PM +0800, Ming Lei wrote: > > > > It is observed on QEMU that pci_alloc_irq_vectors_affinity() may > > > > returns -EINVAL when the requested number is too big(such as 64). > > > > > > Which is not how this API is supposed to work and documented to work. > > > > > > We need to fix pci_alloc_irq_vectors_affinity to not return a spurious > > > error and just return the allocated number of vectors instead of > > > hacking around that in drivers. > > > > Yeah, you are right. > > > > The issue is that QEMU nvme-pci is MSIX-capable only, and hasn't MSI > > capability. > > > > __pci_enable_msix_range() actually returns -ENOSPC, but __pci_enable_msi_range() > > returns -EINVAL because dev->msi_cap is zero. > > > > Maybe we need the following fix? > > Should it matter? We still get a negative vecs back, and still fall > back to the next option. Unless ther are no irqs available at all > for the selected types pci_alloc_irq_vectors_affinity should never > return an error. The patch in last email does fix this issue. In this case, the number of NVMe PCI's MSI-X table entries is 64, so __pci_enable_msix_range() return -ENOSPC when we ask for 65. However, the following __pci_enable_msi_range() returns -EINVAL because the NVMe PCI isn't capable of MSI, then this error is returned from pci_alloc_irq_vectors_affinity() finally to NVMe driver. Of course, -EINVAL makes a difference because the current code only tries to assign one irq vector in this case, and it shouldn't be returned from pci_alloc_irq_vectors_affinity(), given there is enough msix entries for fallback, right? Thanks, Ming From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-2.5 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, MAILING_LIST_MULTI,SPF_PASS,USER_AGENT_MUTT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C105EC43387 for ; Thu, 27 Dec 2018 22:16:47 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 8460020811 for ; Thu, 27 Dec 2018 22:16:47 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1733138AbeL0WQr (ORCPT ); Thu, 27 Dec 2018 17:16:47 -0500 Received: from mx1.redhat.com ([209.132.183.28]:59996 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1733136AbeL0WQq (ORCPT ); Thu, 27 Dec 2018 17:16:46 -0500 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.phx2.redhat.com [10.5.11.14]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 7BA852D2BFE; Thu, 27 Dec 2018 22:16:46 +0000 (UTC) Received: from ming.t460p (ovpn-8-19.pek2.redhat.com [10.72.8.19]) by smtp.corp.redhat.com (Postfix) with ESMTPS id F40B55DAA0; Thu, 27 Dec 2018 22:16:39 +0000 (UTC) Date: Fri, 28 Dec 2018 06:16:33 +0800 From: Ming Lei To: Christoph Hellwig Cc: Jens Axboe , linux-pci@vger.kernel.org, linux-nvme@lists.infradead.org, Keith Busch , Bjorn Helgaas Subject: Re: [PATCH 2/2] nvme pci: try to allocate multiple irq vectors again in case of -EINVAL Message-ID: <20181227221631.GA22073@ming.t460p> References: <20181226103755.2101-1-ming.lei@redhat.com> <20181226103755.2101-3-ming.lei@redhat.com> <20181226182027.GA5866@lst.de> <20181227082136.GA14423@ming.t460p> <20181227130834.GA22967@lst.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20181227130834.GA22967@lst.de> User-Agent: Mutt/1.9.1 (2017-09-22) X-Scanned-By: MIMEDefang 2.79 on 10.5.11.14 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.39]); Thu, 27 Dec 2018 22:16:46 +0000 (UTC) Sender: linux-pci-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-pci@vger.kernel.org On Thu, Dec 27, 2018 at 02:08:34PM +0100, Christoph Hellwig wrote: > On Thu, Dec 27, 2018 at 04:21:38PM +0800, Ming Lei wrote: > > On Wed, Dec 26, 2018 at 07:20:27PM +0100, Christoph Hellwig wrote: > > > On Wed, Dec 26, 2018 at 06:37:55PM +0800, Ming Lei wrote: > > > > It is observed on QEMU that pci_alloc_irq_vectors_affinity() may > > > > returns -EINVAL when the requested number is too big(such as 64). > > > > > > Which is not how this API is supposed to work and documented to work. > > > > > > We need to fix pci_alloc_irq_vectors_affinity to not return a spurious > > > error and just return the allocated number of vectors instead of > > > hacking around that in drivers. > > > > Yeah, you are right. > > > > The issue is that QEMU nvme-pci is MSIX-capable only, and hasn't MSI > > capability. > > > > __pci_enable_msix_range() actually returns -ENOSPC, but __pci_enable_msi_range() > > returns -EINVAL because dev->msi_cap is zero. > > > > Maybe we need the following fix? > > Should it matter? We still get a negative vecs back, and still fall > back to the next option. Unless ther are no irqs available at all > for the selected types pci_alloc_irq_vectors_affinity should never > return an error. The patch in last email does fix this issue. In this case, the number of NVMe PCI's MSI-X table entries is 64, so __pci_enable_msix_range() return -ENOSPC when we ask for 65. However, the following __pci_enable_msi_range() returns -EINVAL because the NVMe PCI isn't capable of MSI, then this error is returned from pci_alloc_irq_vectors_affinity() finally to NVMe driver. Of course, -EINVAL makes a difference because the current code only tries to assign one irq vector in this case, and it shouldn't be returned from pci_alloc_irq_vectors_affinity(), given there is enough msix entries for fallback, right? Thanks, Ming