From patchwork Mon Jul 4 08:39:29 2016 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Christoph Hellwig X-Patchwork-Id: 9211969 X-Patchwork-Delegate: bhelgaas@google.com Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id CD9066086B for ; Mon, 4 Jul 2016 08:41:54 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id BC9AA28475 for ; Mon, 4 Jul 2016 08:41:54 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id B07C927D64; Mon, 4 Jul 2016 08:41:54 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=2.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=unavailable version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 0B52B27D64 for ; Mon, 4 Jul 2016 08:41:54 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932521AbcGDIkN (ORCPT ); Mon, 4 Jul 2016 04:40:13 -0400 Received: from bombadil.infradead.org ([198.137.202.9]:57141 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932487AbcGDIkK (ORCPT ); Mon, 4 Jul 2016 04:40:10 -0400 Received: from i60-34-157-115.s41.a006.ap.plala.or.jp ([60.34.157.115] helo=localhost) by bombadil.infradead.org with esmtpsa (Exim 4.85_2 #1 (Red Hat Linux)) id 1bJzQ9-00052E-H8; Mon, 04 Jul 2016 08:40:01 +0000 From: Christoph Hellwig To: tglx@linutronix.de, axboe@fb.com Cc: agordeev@redhat.com, linux-block@vger.kernel.org, linux-pci@vger.kernel.org, linux-nvme@lists.infradead.org, linux-kernel@vger.kernel.org Subject: [PATCH 08/13] pci: spread interrupt vectors in pci_alloc_irq_vectors Date: Mon, 4 Jul 2016 17:39:29 +0900 Message-Id: <1467621574-8277-9-git-send-email-hch@lst.de> X-Mailer: git-send-email 2.1.4 In-Reply-To: <1467621574-8277-1-git-send-email-hch@lst.de> References: <1467621574-8277-1-git-send-email-hch@lst.de> X-SRS-Rewrite: SMTP reverse-path rewritten from by bombadil.infradead.org. See http://www.infradead.org/rpr.html Sender: linux-pci-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-pci@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP Set the affinity_mask in the PCI device before allocating vectors so that the affinity can be propagated through the MSI descriptor structures to the core IRQ code. Add a new helper __pci_enable_msi_range which is similar to __pci_enable_msix_range introduced in the last patch so that we can allocate the affinity mask in a self-contained fashion and for the right number of vectors. A new PCI_IRQ_NOAFFINITY flag is added to pci_alloc_irq_vectors so that this function can also be used by drivers that don't wish to use the automatic affinity assignment. Signed-off-by: Christoph Hellwig --- Documentation/PCI/MSI-HOWTO.txt | 3 +- drivers/pci/msi.c | 72 ++++++++++++++++++++++++++++++++++++++--- include/linux/pci.h | 2 ++ 3 files changed, 72 insertions(+), 5 deletions(-) diff --git a/Documentation/PCI/MSI-HOWTO.txt b/Documentation/PCI/MSI-HOWTO.txt index 35d1326..dcd3f6d 100644 --- a/Documentation/PCI/MSI-HOWTO.txt +++ b/Documentation/PCI/MSI-HOWTO.txt @@ -95,7 +95,8 @@ argument set to this limit, and the PCI core will return -ENOSPC if it can't meet the minimum number of vectors. The flags argument should normally be set to 0, but can be used to pass the PCI_IRQ_NOMSI and PCI_IRQ_NOMSIX flag in case a device claims -to support MSI or MSI-X, but the support is broken. +to support MSI or MSI-X, but the support is broken, or to PCI_IRQ_NOAFFINITY +if the driver does not wish to use the automatic affinity assignment feature. To get the Linux IRQ numbers passed to request_irq and free_irq and the vectors use the following function: diff --git a/drivers/pci/msi.c b/drivers/pci/msi.c index 6b0834d..7f38e07 100644 --- a/drivers/pci/msi.c +++ b/drivers/pci/msi.c @@ -568,6 +568,7 @@ static struct msi_desc *msi_setup_entry(struct pci_dev *dev, int nvec) entry->msi_attrib.multi_cap = (control & PCI_MSI_FLAGS_QMASK) >> 1; entry->msi_attrib.multiple = ilog2(__roundup_pow_of_two(nvec)); entry->nvec_used = nvec; + entry->affinity = dev->irq_affinity; if (control & PCI_MSI_FLAGS_64BIT) entry->mask_pos = dev->msi_cap + PCI_MSI_MASK_64; @@ -679,10 +680,18 @@ static void __iomem *msix_map_region(struct pci_dev *dev, unsigned nr_entries) static int msix_setup_entries(struct pci_dev *dev, void __iomem *base, struct msix_entry *entries, int nvec) { + const struct cpumask *mask = NULL; struct msi_desc *entry; - int i; + int cpu = -1, i; for (i = 0; i < nvec; i++) { + if (dev->irq_affinity) { + cpu = cpumask_next(cpu, dev->irq_affinity); + if (cpu >= nr_cpu_ids) + cpu = cpumask_first(dev->irq_affinity); + mask = cpumask_of(cpu); + } + entry = alloc_msi_entry(&dev->dev); if (!entry) { if (!i) @@ -699,6 +708,7 @@ static int msix_setup_entries(struct pci_dev *dev, void __iomem *base, entry->msi_attrib.default_irq = dev->irq; entry->mask_base = base; entry->nvec_used = 1; + entry->affinity = mask; list_add_tail(&entry->list, dev_to_msi_list(&dev->dev)); } @@ -1121,8 +1131,53 @@ int pci_enable_msix_range(struct pci_dev *dev, struct msix_entry *entries, } EXPORT_SYMBOL(pci_enable_msix_range); +static int __pci_enable_msi_range(struct pci_dev *dev, int min_vecs, int max_vecs, + unsigned int flags) +{ + int vecs, ret; + + if (!pci_msi_supported(dev, min_vecs)) + return -EINVAL; + + vecs = pci_msi_vec_count(dev); + if (vecs < 0) + return vecs; + if (vecs < min_vecs) + return -EINVAL; + if (vecs > max_vecs) + vecs = max_vecs; + +retry: + if (vecs < min_vecs) + return -ENOSPC; + + if (!(flags & PCI_IRQ_NOAFFINITY)) { + dev->irq_affinity = irq_create_affinity_mask(&vecs); + if (vecs < min_vecs) { + ret = -ERANGE; + goto out_fail; + } + } + + ret = msi_capability_init(dev, vecs); + if (ret) + goto out_fail; + + return vecs; + +out_fail: + kfree(dev->irq_affinity); + if (ret >= 0) { + /* retry with the actually supported number of vectors */ + vecs = ret; + goto retry; + } + + return ret; +} + static int __pci_enable_msix_range(struct pci_dev *dev, unsigned int min_vecs, - unsigned int max_vecs) + unsigned int max_vecs, unsigned int flags) { int vecs = max_vecs, ret, i; @@ -1138,6 +1193,13 @@ retry: for (i = 0; i < vecs; i++) dev->msix_vectors[i].entry = i; + if (!(flags & PCI_IRQ_NOAFFINITY)) { + dev->irq_affinity = irq_create_affinity_mask(&vecs); + ret = -ENOSPC; + if (vecs < min_vecs) + goto out_fail; + } + ret = pci_enable_msix(dev, dev->msix_vectors, vecs); if (ret) goto out_fail; @@ -1147,6 +1209,8 @@ retry: out_fail: kfree(dev->msix_vectors); dev->msix_vectors = NULL; + kfree(dev->irq_affinity); + dev->irq_affinity = NULL; if (ret >= 0) { /* retry with the actually supported number of vectors */ @@ -1180,13 +1244,13 @@ int pci_alloc_irq_vectors(struct pci_dev *dev, unsigned int min_vecs, int vecs; if (!(flags & PCI_IRQ_NOMSIX)) { - vecs = __pci_enable_msix_range(dev, min_vecs, max_vecs); + vecs = __pci_enable_msix_range(dev, min_vecs, max_vecs, flags); if (vecs > 0) return vecs; } if (!(flags & PCI_IRQ_NOMSI)) { - vecs = pci_enable_msi_range(dev, min_vecs, max_vecs); + vecs = __pci_enable_msi_range(dev, min_vecs, max_vecs, flags); if (vecs > 0) return vecs; } diff --git a/include/linux/pci.h b/include/linux/pci.h index 129871f..6a64c54 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -321,6 +321,7 @@ struct pci_dev { */ unsigned int irq; struct msix_entry *msix_vectors; + struct cpumask *irq_affinity; struct resource resource[DEVICE_COUNT_RESOURCE]; /* I/O and memory regions + expansion ROMs */ bool match_driver; /* Skip attaching driver */ @@ -1240,6 +1241,7 @@ int pci_set_vga_state(struct pci_dev *pdev, bool decode, #define PCI_IRQ_NOMSI (1 << 0) /* don't try to use MSI interrupts */ #define PCI_IRQ_NOMSIX (1 << 1) /* don't try to use MSI-X interrupts */ +#define PCI_IRQ_NOAFFINITY (1 << 2) /* don't auto-assign affinity */ /* kmem_cache style wrapper around pci_alloc_consistent() */