From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jean-Philippe Brucker Subject: [PATCH v2 09/40] iommu/sva: Register page fault handler Date: Fri, 11 May 2018 20:06:10 +0100 Message-ID: <20180511190641.23008-10-jean-philippe.brucker@arm.com> References: <20180511190641.23008-1-jean-philippe.brucker@arm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20180511190641.23008-1-jean-philippe.brucker-5wv7dgnIgG8@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: iommu-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org Errors-To: iommu-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org To: linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, linux-pci-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-acpi-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, kvm-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org Cc: xuzaibo-hv44wF8Li93QT0dZR+AlfA@public.gmane.org, will.deacon-5wv7dgnIgG8@public.gmane.org, okaya-sgV2jX0FEOL9JmXXK+q4OQ@public.gmane.org, ashok.raj-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org, bharatku-gjFFaj9aHVfQT0dZR+AlfA@public.gmane.org, rfranz-YGCgFSpz5w/QT0dZR+AlfA@public.gmane.org, rgummal-gjFFaj9aHVfQT0dZR+AlfA@public.gmane.org, ilias.apalodimas-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org, dwmw2-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org, christian.koenig-5C7GfCeVMHo@public.gmane.org List-Id: iommu@lists.linux-foundation.org Let users call iommu_sva_device_init() with the IOMMU_SVA_FEAT_IOPF flag, that enables the I/O Page Fault queue. The IOMMU driver checks is the device supports a form of page fault, in which case they add the device to a fault queue. If the device doesn't support page faults, the IOMMU driver aborts iommu_sva_device_init(). The fault queue must be flushed before any io_mm is freed, to make sure that its PASID isn't used in any fault queue, and can be reallocated. Add iopf_queue_flush() calls in a few strategic locations. Signed-off-by: Jean-Philippe Brucker --- v1->v2: new --- drivers/iommu/iommu-sva.c | 36 ++++++++++++++++++++++++++++++++---- drivers/iommu/iommu.c | 6 +++--- include/linux/iommu.h | 2 ++ 3 files changed, 37 insertions(+), 7 deletions(-) diff --git a/drivers/iommu/iommu-sva.c b/drivers/iommu/iommu-sva.c index 5abe0f0b445c..e98b994c15f1 100644 --- a/drivers/iommu/iommu-sva.c +++ b/drivers/iommu/iommu-sva.c @@ -441,6 +441,8 @@ static void iommu_notifier_release(struct mmu_notifier *mn, struct mm_struct *mm dev_WARN(bond->dev, "possible leak of PASID %u", io_mm->pasid); + iopf_queue_flush_dev(bond->dev); + spin_lock(&iommu_sva_lock); next = list_next_entry(bond, mm_head); @@ -518,6 +520,9 @@ static struct mmu_notifier_ops iommu_mmu_notifier = { * description. Setting @max_pasid to a non-zero value smaller than this limit * overrides it. * + * If the device should support recoverable I/O Page Faults (e.g. PCI PRI), the + * IOMMU_SVA_FEAT_IOPF feature must be requested. + * * If the driver intends to share process address spaces, it should pass a valid * @mm_exit handler. Otherwise @mm_exit can be NULL. After @mm_exit returns, the * device must not issue any more transaction with the PASID given as argument. @@ -546,12 +551,21 @@ int iommu_sva_device_init(struct device *dev, unsigned long features, if (!domain || !domain->ops->sva_device_init) return -ENODEV; - if (features) + if (features & ~IOMMU_SVA_FEAT_IOPF) return -EINVAL; + if (features & IOMMU_SVA_FEAT_IOPF) { + ret = iommu_register_device_fault_handler(dev, iommu_queue_iopf, + dev); + if (ret) + return ret; + } + param = kzalloc(sizeof(*param), GFP_KERNEL); - if (!param) - return -ENOMEM; + if (!param) { + ret = -ENOMEM; + goto err_remove_handler; + } param->features = features; param->max_pasid = max_pasid; @@ -584,6 +598,9 @@ int iommu_sva_device_init(struct device *dev, unsigned long features, err_free_param: kfree(param); +err_remove_handler: + iommu_unregister_device_fault_handler(dev); + return ret; } EXPORT_SYMBOL_GPL(iommu_sva_device_init); @@ -593,7 +610,8 @@ EXPORT_SYMBOL_GPL(iommu_sva_device_init); * @dev: the device * * Disable SVA. Device driver should ensure that the device isn't performing any - * DMA while this function is running. + * DMA while this function is running. In addition all faults should have been + * flushed to the IOMMU. */ int iommu_sva_device_shutdown(struct device *dev) { @@ -617,6 +635,8 @@ int iommu_sva_device_shutdown(struct device *dev) kfree(param); + iommu_unregister_device_fault_handler(dev); + return 0; } EXPORT_SYMBOL_GPL(iommu_sva_device_shutdown); @@ -694,6 +714,12 @@ int __iommu_sva_unbind_device(struct device *dev, int pasid) if (!param || WARN_ON(!domain)) return -EINVAL; + /* + * Caller stopped the device from issuing PASIDs, now make sure they are + * out of the fault queue. + */ + iopf_queue_flush_dev(dev); + /* spin_lock_irq matches the one in wait_event_lock_irq */ spin_lock_irq(&iommu_sva_lock); list_for_each_entry(bond, ¶m->mm_list, dev_head) { @@ -721,6 +747,8 @@ void __iommu_sva_unbind_dev_all(struct device *dev) struct iommu_sva_param *param; struct iommu_bond *bond, *next; + iopf_queue_flush_dev(dev); + /* * io_mm_detach_locked might wait, so we shouldn't call it with the dev * param lock held. It's fine to read sva_param outside the lock because diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c index 333801e1519c..13f705df0725 100644 --- a/drivers/iommu/iommu.c +++ b/drivers/iommu/iommu.c @@ -2278,9 +2278,9 @@ EXPORT_SYMBOL_GPL(iommu_fwspec_add_ids); * iommu_sva_device_init() must be called first, to initialize the required SVA * features. @flags is a subset of these features. * - * The caller must pin down using get_user_pages*() all mappings shared with the - * device. mlock() isn't sufficient, as it doesn't prevent minor page faults - * (e.g. copy-on-write). + * If IOMMU_SVA_FEAT_IOPF isn't requested, the caller must pin down using + * get_user_pages*() all mappings shared with the device. mlock() isn't + * sufficient, as it doesn't prevent minor page faults (e.g. copy-on-write). * * On success, 0 is returned and @pasid contains a valid ID. Otherwise, an error * is returned. diff --git a/include/linux/iommu.h b/include/linux/iommu.h index fad3a60e1c14..933100678f64 100644 --- a/include/linux/iommu.h +++ b/include/linux/iommu.h @@ -64,6 +64,8 @@ typedef int (*iommu_fault_handler_t)(struct iommu_domain *, typedef int (*iommu_dev_fault_handler_t)(struct iommu_fault_event *, void *); typedef int (*iommu_mm_exit_handler_t)(struct device *dev, int pasid, void *); +#define IOMMU_SVA_FEAT_IOPF (1 << 0) + struct iommu_domain_geometry { dma_addr_t aperture_start; /* First address that can be mapped */ dma_addr_t aperture_end; /* Last address that can be mapped */ -- 2.17.0