From patchwork Fri Jan 3 00:08:08 2014 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Yinghai Lu X-Patchwork-Id: 3428171 X-Patchwork-Delegate: bhelgaas@google.com Return-Path: X-Original-To: patchwork-linux-pci@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.19.201]) by patchwork1.web.kernel.org (Postfix) with ESMTP id 13CD89F374 for ; Fri, 3 Jan 2014 00:09:11 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id D3BCB20170 for ; Fri, 3 Jan 2014 00:09:09 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 7FB5F20158 for ; Fri, 3 Jan 2014 00:09:08 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754356AbaACAIU (ORCPT ); Thu, 2 Jan 2014 19:08:20 -0500 Received: from userp1040.oracle.com ([156.151.31.81]:23840 "EHLO userp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754220AbaACAIP (ORCPT ); Thu, 2 Jan 2014 19:08:15 -0500 Received: from acsinet21.oracle.com (acsinet21.oracle.com [141.146.126.237]) by userp1040.oracle.com (Sentrion-MTA-4.3.1/Sentrion-MTA-4.3.1) with ESMTP id s0307as9031308 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 3 Jan 2014 00:07:37 GMT Received: from aserz7021.oracle.com (aserz7021.oracle.com [141.146.126.230]) by acsinet21.oracle.com (8.14.4+Sun/8.14.4) with ESMTP id s0307Ztv021895 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 3 Jan 2014 00:07:36 GMT Received: from abhmp0002.oracle.com (abhmp0002.oracle.com [141.146.116.8]) by aserz7021.oracle.com (8.14.4+Sun/8.14.4) with ESMTP id s0307ZQe021892; Fri, 3 Jan 2014 00:07:35 GMT Received: from linux-siqj.site (/10.132.126.191) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Thu, 02 Jan 2014 16:07:35 -0800 From: Yinghai Lu To: "H. Peter Anvin" , Tony Luck , Bjorn Helgaas , "Rafael J. Wysocki" , x86 Cc: linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org, linux-acpi@vger.kernel.org, Yinghai Lu , David Woodhouse , Vinod Koul , Dan Williams , Donald Dutile , iommu@lists.linux-foundation.org Subject: [PATCH v2 01/10] IOMMU: Update dmar units devices list during hotplug Date: Thu, 2 Jan 2014 16:08:08 -0800 Message-Id: <1388707697-16800-2-git-send-email-yinghai@kernel.org> X-Mailer: git-send-email 1.8.4 In-Reply-To: <1388707697-16800-1-git-send-email-yinghai@kernel.org> References: <1388707697-16800-1-git-send-email-yinghai@kernel.org> X-Source-IP: acsinet21.oracle.com [141.146.126.237] Sender: linux-pci-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-pci@vger.kernel.org X-Spam-Status: No, score=-7.4 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP When do pci remove/rescan on system that have more iommus, got [ 894.089745] Set context mapping for c4:00.0 [ 894.110890] mpt2sas3: Allocated physical memory: size(4293 kB) [ 894.112556] mpt2sas3: Current Controller Queue Depth(1883), Max Controller Queue Depth(2144) [ 894.127278] mpt2sas3: Scatter Gather Elements per IO(128) [ 894.361295] DRHD: handling fault status reg 2 [ 894.364053] DMAR:[DMA Read] Request device [c4:00.0] fault addr fffbe000 [ 894.364056] DMAR:[fault reason 02] Present bit in context entry is cl It turns out when remove/rescan, pci dev will be freed and will get another new dev. But drhd units still keep old one... so dmar_find_matched_drhd_unit will return wrong drhd and iommu for the device that is not on first iommu. So need to update devices in drhd_units during pci remove/rescan. Could save domain/bus/device/function aside in the list and according that info restore new dev to drhd_units later. Then dmar_find_matched_drdh_unit and device_to_iommu could return right drhd and iommu. Add remove_dev_from_drhd/restore_dev_to_drhd functions to do the real work. call them in device ADD_DEVICE and UNBOUND_DRIVER Need to do the samething to atsr. (expose dmar_atsr_units and add atsru->segment) After patch, will have right iommu for the new dev and will not get DMAR error anymore. -v2: add pci_dev_put/pci_dev_get to make refcount consistent. -v3: fix one left over CONFIG_DMAR -v4: pass pci_dev *dev in save_dev_dmaru()/get_dev_dmaru() according to Bjorn. -v5: fix case only have intr-remap enabled. -v6: skip handling of vf adding/removing. Signed-off-by: Yinghai Lu Cc: David Woodhouse Cc: Vinod Koul Cc: Dan Williams Cc: Donald Dutile Cc: iommu@lists.linux-foundation.org --- drivers/iommu/dmar.c | 138 ++++++++++++++++++++++++++++++++++++ drivers/iommu/intel-iommu.c | 32 +++++---- drivers/iommu/intel_irq_remapping.c | 33 ++++++++- include/linux/dmar.h | 10 +++ 4 files changed, 200 insertions(+), 13 deletions(-) diff --git a/drivers/iommu/dmar.c b/drivers/iommu/dmar.c index f4eaa50..a14867c 100644 --- a/drivers/iommu/dmar.c +++ b/drivers/iommu/dmar.c @@ -450,6 +450,144 @@ fail: return ret; } +struct dev_dmaru { + struct list_head list; + void *dmaru; + int index; + int segment; + unsigned char bus; + unsigned int devfn; +}; + +static int save_dev_dmaru(struct pci_dev *dev, void *dmaru, + int index, struct list_head *lh) +{ + struct dev_dmaru *m; + + m = kzalloc(sizeof(*m), GFP_KERNEL); + if (!m) + return -ENOMEM; + + m->segment = pci_domain_nr(dev->bus); + m->bus = dev->bus->number; + m->devfn = dev->devfn; + m->dmaru = dmaru; + m->index = index; + + list_add(&m->list, lh); + + return 0; +} +static void *get_dev_dmaru(struct pci_dev *dev, int *index, + struct list_head *lh) +{ + int segment = pci_domain_nr(dev->bus); + unsigned char bus = dev->bus->number; + unsigned int devfn = dev->devfn; + struct dev_dmaru *m; + void *dmaru = NULL; + + list_for_each_entry(m, lh, list) { + if (m->segment == segment && + m->bus == bus && m->devfn == devfn) { + *index = m->index; + dmaru = m->dmaru; + list_del(&m->list); + kfree(m); + break; + } + } + + return dmaru; +} + +static LIST_HEAD(saved_dev_drhd_list); +void remove_dev_from_drhd(struct pci_dev *dev) +{ + struct dmar_drhd_unit *drhd = NULL; + int segment = pci_domain_nr(dev->bus); + int i; + + if (dev->is_virtfn) + return; + + for_each_drhd_unit(drhd) { + if (drhd->ignored) + continue; + if (segment != drhd->segment) + continue; + + for (i = 0; i < drhd->devices_cnt; i++) { + if (drhd->devices[i] == dev) { + /* save it at first if it is in drhd */ + save_dev_dmaru(dev, drhd, i, + &saved_dev_drhd_list); + /* always remove it */ + pci_dev_put(dev); + drhd->devices[i] = NULL; + return; + } + } + } +} +void restore_dev_to_drhd(struct pci_dev *dev) +{ + struct dmar_drhd_unit *drhd = NULL; + int i; + + if (dev->is_virtfn) + return; + + /* find the stored drhd */ + drhd = get_dev_dmaru(dev, &i, &saved_dev_drhd_list); + /* restore that into drhd */ + if (drhd) + drhd->devices[i] = pci_dev_get(dev); +} + +#ifdef CONFIG_INTEL_IOMMU +static LIST_HEAD(saved_dev_atsr_list); +void remove_dev_from_atsr(struct pci_dev *dev) +{ + struct dmar_atsr_unit *atsr = NULL; + int segment = pci_domain_nr(dev->bus); + int i; + + if (dev->is_virtfn) + return; + + for_each_atsr_unit(atsr) { + if (segment != atsr->segment) + continue; + + for (i = 0; i < atsr->devices_cnt; i++) { + if (atsr->devices[i] == dev) { + /* save it at first if it is in drhd */ + save_dev_dmaru(dev, atsr, i, + &saved_dev_atsr_list); + /* always remove it */ + pci_dev_put(dev); + atsr->devices[i] = NULL; + return; + } + } + } +} +void restore_dev_to_atsr(struct pci_dev *dev) +{ + struct dmar_atsr_unit *atsr = NULL; + int i; + + if (dev->is_virtfn) + return; + + /* find the stored atsr */ + atsr = get_dev_dmaru(dev, &i, &saved_dev_atsr_list); + /* restore that into atsr */ + if (atsr) + atsr->devices[i] = pci_dev_get(dev); +} +#endif /* CONFIG_INTEL_IOMMU */ int __init dmar_table_init(void) { diff --git a/drivers/iommu/intel-iommu.c b/drivers/iommu/intel-iommu.c index 43b9bfe..a3eb689 100644 --- a/drivers/iommu/intel-iommu.c +++ b/drivers/iommu/intel-iommu.c @@ -3528,7 +3528,7 @@ rmrr_parse_dev(struct dmar_rmrr_unit *rmrru) return ret; } -static LIST_HEAD(dmar_atsr_units); +LIST_HEAD(dmar_atsr_units); int __init dmar_parse_one_atsr(struct acpi_dmar_header *hdr) { @@ -3542,6 +3542,7 @@ int __init dmar_parse_one_atsr(struct acpi_dmar_header *hdr) atsru->hdr = hdr; atsru->include_all = atsr->flags & 0x1; + atsru->segment = atsr->segment; list_add(&atsru->list, &dmar_atsr_units); @@ -3573,16 +3574,13 @@ int dmar_find_matched_atsr_unit(struct pci_dev *dev) { int i; struct pci_bus *bus; - struct acpi_dmar_atsr *atsr; struct dmar_atsr_unit *atsru; dev = pci_physfn(dev); - list_for_each_entry(atsru, &dmar_atsr_units, list) { - atsr = container_of(atsru->hdr, struct acpi_dmar_atsr, header); - if (atsr->segment == pci_domain_nr(dev->bus)) + list_for_each_entry(atsru, &dmar_atsr_units, list) + if (atsru->segment == pci_domain_nr(dev->bus)) goto found; - } return 0; @@ -3642,20 +3640,30 @@ static int device_notifier(struct notifier_block *nb, struct pci_dev *pdev = to_pci_dev(dev); struct dmar_domain *domain; - if (iommu_no_mapping(dev)) - return 0; + switch (action) { + case BUS_NOTIFY_UNBOUND_DRIVER: + if (iommu_no_mapping(dev) || iommu_pass_through) + break; - domain = find_domain(pdev); - if (!domain) - return 0; + domain = find_domain(pdev); + if (!domain) + break; - if (action == BUS_NOTIFY_UNBOUND_DRIVER && !iommu_pass_through) { domain_remove_one_dev_info(domain, pdev); if (!(domain->flags & DOMAIN_FLAG_VIRTUAL_MACHINE) && !(domain->flags & DOMAIN_FLAG_STATIC_IDENTITY) && list_empty(&domain->devices)) domain_exit(domain); + break; + case BUS_NOTIFY_DEL_DEVICE: + remove_dev_from_drhd(pdev); + remove_dev_from_atsr(pdev); + break; + case BUS_NOTIFY_ADD_DEVICE: + restore_dev_to_drhd(pdev); + restore_dev_to_atsr(pdev); + break; } return 0; diff --git a/drivers/iommu/intel_irq_remapping.c b/drivers/iommu/intel_irq_remapping.c index 07ab6e8..8471f40 100644 --- a/drivers/iommu/intel_irq_remapping.c +++ b/drivers/iommu/intel_irq_remapping.c @@ -807,12 +807,43 @@ int __init parse_ioapics_under_ir(void) return 1; } +static int device_notifier(struct notifier_block *nb, + unsigned long action, void *data) +{ + struct device *dev = data; + struct pci_dev *pdev = to_pci_dev(dev); + + switch (action) { + case BUS_NOTIFY_DEL_DEVICE: + remove_dev_from_drhd(pdev); + break; + case BUS_NOTIFY_ADD_DEVICE: + restore_dev_to_drhd(pdev); + break; + } + + return 0; +} + +static struct notifier_block device_nb = { + .notifier_call = device_notifier, +}; + int __init ir_dev_scope_init(void) { + int ret; + if (!irq_remapping_enabled) return 0; - return dmar_dev_scope_init(); + ret = dmar_dev_scope_init(); + if (ret < 0) + return ret; + + if (intel_iommu_enabled != 1) + bus_register_notifier(&pci_bus_type, &device_nb); + + return ret; } rootfs_initcall(ir_dev_scope_init); diff --git a/include/linux/dmar.h b/include/linux/dmar.h index b029d1a..0a3dd71 100644 --- a/include/linux/dmar.h +++ b/include/linux/dmar.h @@ -131,6 +131,11 @@ extern int dmar_set_interrupt(struct intel_iommu *iommu); extern irqreturn_t dmar_fault(int irq, void *dev_id); extern int arch_setup_dmar_msi(unsigned int irq); +void remove_dev_from_drhd(struct pci_dev *dev); +void restore_dev_to_drhd(struct pci_dev *dev); +void remove_dev_from_atsr(struct pci_dev *dev); +void restore_dev_to_atsr(struct pci_dev *dev); + #ifdef CONFIG_INTEL_IOMMU extern int iommu_detected, no_iommu; extern struct list_head dmar_rmrr_units; @@ -146,14 +151,19 @@ struct dmar_rmrr_unit { #define for_each_rmrr_units(rmrr) \ list_for_each_entry(rmrr, &dmar_rmrr_units, list) +extern struct list_head dmar_atsr_units; struct dmar_atsr_unit { struct list_head list; /* list of ATSR units */ struct acpi_dmar_header *hdr; /* ACPI header */ struct pci_dev **devices; /* target devices */ int devices_cnt; /* target device count */ + u16 segment; /* PCI domain */ u8 include_all:1; /* include all ports */ }; +#define for_each_atsr_unit(atsr) \ + list_for_each_entry(atsr, &dmar_atsr_units, list) + int dmar_parse_rmrr_atsr_dev(void); extern int dmar_parse_one_rmrr(struct acpi_dmar_header *header); extern int dmar_parse_one_atsr(struct acpi_dmar_header *header);