All of lore.kernel.org
 help / color / mirror / Atom feed
From: Lu Baolu <baolu.lu@linux.intel.com>
To: Joerg Roedel <joro@8bytes.org>
Cc: iommu@lists.linux.dev, linux-kernel@vger.kernel.org
Subject: [PATCH 2/7] iommu/vt-d: Add device_block_translation() helper
Date: Tue, 22 Nov 2022 08:29:44 +0800	[thread overview]
Message-ID: <20221122002949.148140-3-baolu.lu@linux.intel.com> (raw)
In-Reply-To: <20221122002949.148140-1-baolu.lu@linux.intel.com>

If domain attaching to device fails, the IOMMU driver should bring the
device to blocking DMA state. The upper layer is expected to recover it
by attaching a new domain. Use device_block_translation() in the error
path of dev_attach to make the behavior specific.

The difference between device_block_translation() and the previous
dmar_remove_one_dev_info() is that, in the scalable mode, it is the
RID2PASID entry instead of context entry being cleared. As a result,
enabling PCI capabilities is moved up.

Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
Reviewed-by: Kevin Tian <kevin.tian@intel.com>
Link: https://lore.kernel.org/r/20221118132451.114406-3-baolu.lu@linux.intel.com
---
 drivers/iommu/intel/iommu.c | 44 ++++++++++++++++++++++++++++++++-----
 1 file changed, 38 insertions(+), 6 deletions(-)

diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index dea0190acc86..1c1a52d1a818 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -277,7 +277,7 @@ static LIST_HEAD(dmar_satc_units);
 #define for_each_rmrr_units(rmrr) \
 	list_for_each_entry(rmrr, &dmar_rmrr_units, list)
 
-static void dmar_remove_one_dev_info(struct device *dev);
+static void device_block_translation(struct device *dev);
 
 int dmar_disabled = !IS_ENABLED(CONFIG_INTEL_IOMMU_DEFAULT_ON);
 int intel_iommu_sm = IS_ENABLED(CONFIG_INTEL_IOMMU_SCALABLE_MODE_DEFAULT_ON);
@@ -1400,7 +1400,7 @@ static void iommu_enable_pci_caps(struct device_domain_info *info)
 {
 	struct pci_dev *pdev;
 
-	if (!info || !dev_is_pci(info->dev))
+	if (!dev_is_pci(info->dev))
 		return;
 
 	pdev = to_pci_dev(info->dev);
@@ -2045,7 +2045,6 @@ static int domain_context_mapping_one(struct dmar_domain *domain,
 	} else {
 		iommu_flush_write_buffer(iommu);
 	}
-	iommu_enable_pci_caps(info);
 
 	ret = 0;
 
@@ -2487,7 +2486,7 @@ static int domain_add_dev_info(struct dmar_domain *domain, struct device *dev)
 					dev, PASID_RID2PASID);
 		if (ret) {
 			dev_err(dev, "Setup RID2PASID failed\n");
-			dmar_remove_one_dev_info(dev);
+			device_block_translation(dev);
 			return ret;
 		}
 	}
@@ -2495,10 +2494,12 @@ static int domain_add_dev_info(struct dmar_domain *domain, struct device *dev)
 	ret = domain_context_mapping(domain, dev);
 	if (ret) {
 		dev_err(dev, "Domain context map failed\n");
-		dmar_remove_one_dev_info(dev);
+		device_block_translation(dev);
 		return ret;
 	}
 
+	iommu_enable_pci_caps(info);
+
 	return 0;
 }
 
@@ -4109,6 +4110,37 @@ static void dmar_remove_one_dev_info(struct device *dev)
 	info->domain = NULL;
 }
 
+/*
+ * Clear the page table pointer in context or pasid table entries so that
+ * all DMA requests without PASID from the device are blocked. If the page
+ * table has been set, clean up the data structures.
+ */
+static void device_block_translation(struct device *dev)
+{
+	struct device_domain_info *info = dev_iommu_priv_get(dev);
+	struct intel_iommu *iommu = info->iommu;
+	unsigned long flags;
+
+	iommu_disable_dev_iotlb(info);
+	if (!dev_is_real_dma_subdevice(dev)) {
+		if (sm_supported(iommu))
+			intel_pasid_tear_down_entry(iommu, dev,
+						    PASID_RID2PASID, false);
+		else
+			domain_context_clear(info);
+	}
+
+	if (!info->domain)
+		return;
+
+	spin_lock_irqsave(&info->domain->lock, flags);
+	list_del(&info->link);
+	spin_unlock_irqrestore(&info->domain->lock, flags);
+
+	domain_detach_iommu(info->domain, iommu);
+	info->domain = NULL;
+}
+
 static int md_domain_init(struct dmar_domain *domain, int guest_width)
 {
 	int adjust_width;
@@ -4230,7 +4262,7 @@ static int intel_iommu_attach_device(struct iommu_domain *domain,
 		struct device_domain_info *info = dev_iommu_priv_get(dev);
 
 		if (info->domain)
-			dmar_remove_one_dev_info(dev);
+			device_block_translation(dev);
 	}
 
 	ret = prepare_domain_attach_device(domain, dev);
-- 
2.34.1


  parent reply	other threads:[~2022-11-22  0:36 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-22  0:29 [PATCH 0/7] [PULL REQUEST] Intel IOMMU updates for Linux v6.2 Lu Baolu
2022-11-22  0:29 ` [PATCH 1/7] iommu/vt-d: Allocate pasid table in device probe path Lu Baolu
2022-11-22  0:29 ` Lu Baolu [this message]
2022-11-22  0:29 ` [PATCH 3/7] iommu/vt-d: Add blocking domain support Lu Baolu
2022-11-22  0:29 ` [PATCH 4/7] iommu/vt-d: Rename iommu_disable_dev_iotlb() Lu Baolu
2022-11-22  0:29 ` [PATCH 5/7] iommu/vt-d: Rename domain_add_dev_info() Lu Baolu
2022-11-22  0:29 ` [PATCH 6/7] iommu/vt-d: Remove unnecessary domain_context_mapped() Lu Baolu
2022-11-22  0:29 ` [PATCH 7/7] iommu/vt-d: Use real field for indication of first level Lu Baolu
2022-11-22  9:06 ` [PATCH 0/7] [PULL REQUEST] Intel IOMMU updates for Linux v6.2 Joerg Roedel
2022-11-22  9:41   ` Baolu Lu
2022-11-22 13:06     ` Joerg Roedel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20221122002949.148140-3-baolu.lu@linux.intel.com \
    --to=baolu.lu@linux.intel.com \
    --cc=iommu@lists.linux.dev \
    --cc=joro@8bytes.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.