Message-ID: <MWHPR11MB1645726E6EB0D9248BC56DA78CA50@MWHPR11MB1645.namprd11.prod.outlook.com>
Date:   Thu, 7 May 2020 06:35:58 +0000
From:   "Tian, Kevin" <kevin.tian@...el.com>
To:     Lu Baolu <baolu.lu@...ux.intel.com>, Joerg Roedel <joro@...tes.org>
CC:     "Raj, Ashok" <ashok.raj@...el.com>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "iommu@...ts.linux-foundation.org" <iommu@...ts.linux-foundation.org>
Subject: RE: [PATCH v4 4/5] iommu/vt-d: Add page request draining support

> From: Lu Baolu
> Sent: Thursday, May 7, 2020 8:56 AM
> 
> When a PASID is stopped or terminated, there can be pending PRQs
> (requests that haven't received responses) in remapping hardware.
> This adds the interface to drain page requests and call it when a
> PASID is terminated.
> 
> Signed-off-by: Jacob Pan <jacob.jun.pan@...ux.intel.com>
> Signed-off-by: Liu Yi L <yi.l.liu@...el.com>
> Signed-off-by: Lu Baolu <baolu.lu@...ux.intel.com>
> ---
>  drivers/iommu/intel-svm.c   | 102 ++++++++++++++++++++++++++++++++++-
> -
>  include/linux/intel-iommu.h |   4 ++
>  2 files changed, 101 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/iommu/intel-svm.c b/drivers/iommu/intel-svm.c
> index 9561ba59a170..7256eb965cf6 100644
> --- a/drivers/iommu/intel-svm.c
> +++ b/drivers/iommu/intel-svm.c
> @@ -23,6 +23,7 @@
>  #include "intel-pasid.h"
> 
>  static irqreturn_t prq_event_thread(int irq, void *d);
> +static void intel_svm_drain_prq(struct device *dev, int pasid);
> 
>  #define PRQ_ORDER 0
> 
> @@ -66,6 +67,8 @@ int intel_svm_enable_prq(struct intel_iommu *iommu)
>  	dmar_writeq(iommu->reg + DMAR_PQT_REG, 0ULL);
>  	dmar_writeq(iommu->reg + DMAR_PQA_REG, virt_to_phys(iommu-
> >prq) | PRQ_ORDER);
> 
> +	init_completion(&iommu->prq_complete);
> +
>  	return 0;
>  }
> 
> @@ -403,12 +406,8 @@ int intel_svm_unbind_gpasid(struct device *dev, int
> pasid)
>  			list_del_rcu(&sdev->list);
>  			intel_pasid_tear_down_entry(iommu, dev,
>  						    svm->pasid, false);
> +			intel_svm_drain_prq(dev, svm->pasid);
>  			intel_flush_svm_range_dev(svm, sdev, 0, -1, 0);
> -			/* TODO: Drain in flight PRQ for the PASID since it
> -			 * may get reused soon, we don't want to
> -			 * confuse with its previous life.
> -			 * intel_svm_drain_prq(dev, pasid);
> -			 */
>  			kfree_rcu(sdev, rcu);
> 
>  			if (list_empty(&svm->devs)) {
> @@ -647,6 +646,7 @@ int intel_svm_unbind_mm(struct device *dev, int
> pasid)
>  			 * hard to be as defensive as we might like. */
>  			intel_pasid_tear_down_entry(iommu, dev,
>  						    svm->pasid, false);
> +			intel_svm_drain_prq(dev, svm->pasid);
>  			intel_flush_svm_range_dev(svm, sdev, 0, -1, 0);
>  			kfree_rcu(sdev, rcu);
> 
> @@ -725,6 +725,92 @@ static bool is_canonical_address(u64 addr)
>  	return (((saddr << shift) >> shift) == saddr);
>  }
> 
> +/**
> + * intel_svm_drain_prq:
> + *
> + * Drain all pending page requests and responses related to a specific
> + * pasid in both software and hardware.
> + */
> +static void intel_svm_drain_prq(struct device *dev, int pasid)
> +{
> +	struct device_domain_info *info;
> +	struct dmar_domain *domain;
> +	struct intel_iommu *iommu;
> +	struct qi_desc desc[3];
> +	struct pci_dev *pdev;
> +	int head, tail;
> +	u16 sid, did;
> +	int qdep;
> +
> +	info = get_domain_info(dev);
> +	if (WARN_ON(!info || !dev_is_pci(dev)))
> +		return;
> +
> +	if (!info->ats_enabled)
> +		return;

ats_enabled -> pri_enabled

> +
> +	iommu = info->iommu;
> +	domain = info->domain;
> +	pdev = to_pci_dev(dev);
> +	sid = PCI_DEVID(info->bus, info->devfn);
> +	did = domain->iommu_did[iommu->seq_id];
> +	qdep = pci_ats_queue_depth(pdev);
> +
> +	memset(desc, 0, sizeof(desc));
> +	desc[0].qw0 = QI_IWD_STATUS_DATA(QI_DONE) |
> +			QI_IWD_FENCE |
> +			QI_IWD_TYPE;
> +	desc[1].qw0 = QI_EIOTLB_PASID(pasid) |
> +			QI_EIOTLB_DID(did) |
> +			QI_EIOTLB_GRAN(QI_GRAN_NONG_PASID) |
> +			QI_EIOTLB_TYPE;
> +	desc[2].qw0 = QI_DEV_EIOTLB_PASID(pasid) |
> +			QI_DEV_EIOTLB_SID(sid) |
> +			QI_DEV_EIOTLB_QDEP(qdep) |
> +			QI_DEIOTLB_TYPE |
> +			QI_DEV_IOTLB_PFSID(info->pfsid);
> +
> +	/*
> +	 * Submit an invalidation wait descriptor with fence and page request
> +	 * drain flags set to invalidation queue. This ensures that all requests
> +	 * submitted to the invalidation queue ahead of this wait descriptor
> are
> +	 * processed and completed, and all already issued page requests
> from
> +	 * the device are put in the page request queue.
> +	 */

I feel this comment is better moved earlier since it explains the overall 
flow including all 3 descriptors. Also it is not one wait descriptor which
gets both fence and drain flags set. There are two wait descriptors with
one setting fence and the other setting drain.

> +	qi_submit_sync(iommu, desc, 1, QI_OPT_WAIT_DRAIN);

the count is '3' instead of '1'.

> +
> +	/*
> +	 * Check and wait until all pending page requests in the queue are
> +	 * handled by the intr thread.
> +	 */
> +prq_retry:
> +	tail = dmar_readq(iommu->reg + DMAR_PQT_REG) &
> PRQ_RING_MASK;
> +	head = dmar_readq(iommu->reg + DMAR_PQH_REG) &
> PRQ_RING_MASK;
> +	while (head != tail) {
> +		struct page_req_dsc *req;
> +
> +		req = &iommu->prq[head / sizeof(*req)];
> +		if (!req->pasid_present || req->pasid != pasid) {
> +			head = (head + sizeof(*req)) & PRQ_RING_MASK;
> +			continue;
> +		}
> +
> +		wait_for_completion_timeout(&iommu->prq_complete, HZ);
> +		goto prq_retry;
> +	}
> +
> +	/*
> +	 * Perform steps described in VT-d spec CH7.10 to drain page
> +	 * requests and responses in hardware.
> +	 */
> +qi_retry:
> +	qi_submit_sync(iommu, desc, 3, QI_OPT_WAIT_DRAIN);
> +	if (readl(iommu->reg + DMAR_PRS_REG) & DMA_PRS_PRO) {
> +		wait_for_completion_timeout(&iommu->prq_complete, HZ);
> +		goto qi_retry;
> +	}
> +}
> +
>  static irqreturn_t prq_event_thread(int irq, void *d)
>  {
>  	struct intel_iommu *iommu = d;
> @@ -859,6 +945,12 @@ static irqreturn_t prq_event_thread(int irq, void *d)
>  	}
> 
>  	dmar_writeq(iommu->reg + DMAR_PQH_REG, tail);
> +	/*
> +	 * Clear the page request overflow bit and wake up all threads that
> +	 * are waiting for the completion of this handling.
> +	 */
> +	writel(DMA_PRS_PRO, iommu->reg + DMAR_PRS_REG);
> +	complete(&iommu->prq_complete);
> 
>  	return IRQ_RETVAL(handled);
>  }
> diff --git a/include/linux/intel-iommu.h b/include/linux/intel-iommu.h
> index cca1e5f9aeaa..a0512b401a59 100644
> --- a/include/linux/intel-iommu.h
> +++ b/include/linux/intel-iommu.h
> @@ -292,6 +292,8 @@
> 
>  /* PRS_REG */
>  #define DMA_PRS_PPR	((u32)1)
> +#define DMA_PRS_PRO	((u32)2)
> +
>  #define DMA_VCS_PAS	((u64)1)
> 
>  #define IOMMU_WAIT_OP(iommu, offset, op, cond, sts)
> 	\
> @@ -333,6 +335,7 @@ enum {
> 
>  #define QI_IWD_STATUS_DATA(d)	(((u64)d) << 32)
>  #define QI_IWD_STATUS_WRITE	(((u64)1) << 5)
> +#define QI_IWD_FENCE		(((u64)1) << 6)
>  #define QI_IWD_PRQ_DRAIN	(((u64)1) << 7)
> 
>  #define QI_IOTLB_DID(did) 	(((u64)did) << 16)
> @@ -590,6 +593,7 @@ struct intel_iommu {
>  #ifdef CONFIG_INTEL_IOMMU_SVM
>  	struct page_req_dsc *prq;
>  	unsigned char prq_name[16];    /* Name for PRQ interrupt */
> +	struct completion prq_complete;
>  	struct ioasid_allocator_ops pasid_allocator; /* Custom allocator for
> PASIDs */
>  #endif
>  	struct q_inval  *qi;            /* Queued invalidation info */
> --
> 2.17.1
> 
> _______________________________________________
> iommu mailing list
> iommu@...ts.linux-foundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/iommu

Powered by blists - more mailing lists