[Xen-devel] [PATCH v5 7/7] VT-d: Fix vt-d Device-TLB flush timeout issue.

[Quan Xu <quan.xu@xxxxxxxxx>]

If Device-TLB flush is timeout, we'll hide the target ATS
device and crash the domain owning this ATS device.

If impacted domain is hardware domain, just throw out a warning.

The hidden device will be disallowed to be further assigned to
any domain.

Signed-off-by: Quan Xu <quan.xu@xxxxxxxxx>
---
 xen/drivers/passthrough/pci.c         |  2 +-
 xen/drivers/passthrough/vtd/extern.h  |  8 +++-
 xen/drivers/passthrough/vtd/qinval.c  | 77 ++++++++++++++++++++++++++++++++++-
 xen/drivers/passthrough/vtd/x86/ats.c | 14 ++++++-
 4 files changed, 95 insertions(+), 6 deletions(-)

diff --git a/xen/drivers/passthrough/pci.c b/xen/drivers/passthrough/pci.c
index 27b3ca7..2d7dc59 100644
--- a/xen/drivers/passthrough/pci.c
+++ b/xen/drivers/passthrough/pci.c
@@ -407,7 +407,7 @@ static void _pci_hide_device(struct pci_dev *pdev)
     list_add(&pdev->domain_list, &dom_xen->arch.pdev_list);
 }
 
-int __init pci_hide_device(int bus, int devfn)
+int pci_hide_device(int bus, int devfn)
 {
     struct pci_dev *pdev;
     int rc = -ENOMEM;
diff --git a/xen/drivers/passthrough/vtd/extern.h 
b/xen/drivers/passthrough/vtd/extern.h
index ec9c513..a129460 100644
--- a/xen/drivers/passthrough/vtd/extern.h
+++ b/xen/drivers/passthrough/vtd/extern.h
@@ -56,8 +56,12 @@ struct acpi_drhd_unit * find_ats_dev_drhd(struct iommu 
*iommu);
 
 int ats_device(const struct pci_dev *, const struct acpi_drhd_unit *);
 
-int dev_invalidate_iotlb(struct iommu *iommu, u16 did,
-                         u64 addr, unsigned int size_order, u64 type);
+int dev_invalidate_iotlb(struct iommu *iommu, u16 did, u64 addr,
+                         unsigned int size_order, u64 type,
+                         unsigned int lock);
+int dev_invalidate_iotlb_sync(struct iommu *iommu, u16 did,
+                              u16 seg, u8 bus, u8 devfn,
+                              unsigned int lock);
 
 int qinval_device_iotlb(struct iommu *iommu,
                         u32 max_invs_pend, u16 sid, u16 size, u64 addr);
diff --git a/xen/drivers/passthrough/vtd/qinval.c 
b/xen/drivers/passthrough/vtd/qinval.c
index f2e7ffb..76046a7 100644
--- a/xen/drivers/passthrough/vtd/qinval.c
+++ b/xen/drivers/passthrough/vtd/qinval.c
@@ -229,6 +229,69 @@ int qinval_device_iotlb(struct iommu *iommu,
     return 0;
 }
 
+static void dev_invalidate_iotlb_timeout(struct iommu *iommu, u16 did,
+                                         u16 seg, u8 bus, u8 devfn,
+                                         unsigned int lock)
+{
+    struct domain *d = NULL;
+    struct pci_dev *pdev;
+
+    if ( test_bit(did, iommu->domid_bitmap) )
+        d = rcu_lock_domain_by_id(iommu->domid_map[did]);
+
+    if ( d == NULL )
+        return;
+
+    for_each_pdev(d, pdev)
+    {
+        if ( (pdev->seg == seg) &&
+             (pdev->bus == bus) &&
+             (pdev->devfn == devfn) )
+        {
+            ASSERT ( pdev->domain );
+            list_del(&pdev->domain_list);
+            pdev->domain = NULL;
+
+            if ( !(lock & PCIDEVS_LOCK) )
+                spin_lock(&pcidevs_lock);
+
+            if ( pci_hide_device(bus, devfn) )
+            {
+                printk(XENLOG_ERR
+                       "IOMMU hide device %04x:%02x:%02x.%02x error.",
+                       seg, bus, PCI_SLOT(devfn), PCI_FUNC(devfn));
+            }
+
+            if ( !(lock & PCIDEVS_LOCK) )
+                spin_unlock(&pcidevs_lock);
+
+            break;
+        }
+    }
+
+    if ( !is_hardware_domain(d) )
+        domain_crash(d);
+    rcu_unlock_domain(d);
+}
+
+int dev_invalidate_iotlb_sync(struct iommu *iommu, u16 did,
+                              u16 seg, u8 bus, u8 devfn,
+                              unsigned int lock)
+{
+    struct qi_ctrl *qi_ctrl = iommu_qi_ctrl(iommu);
+    int rc = 0;
+
+    if ( qi_ctrl->qinval_maddr )
+    {
+        rc = queue_invalidate_wait(iommu, 0, 1, 1);
+        if ( rc == -ETIMEDOUT )
+            dev_invalidate_iotlb_timeout(iommu, did,
+                                         seg, bus, devfn, lock);
+    }
+
+    return rc;
+}
+
 static void queue_invalidate_iec(struct iommu *iommu, u8 granu, u8 im, u16 
iidx)
 {
     unsigned long flags;
@@ -350,9 +413,19 @@ static int flush_iotlb_qi(
         queue_invalidate_iotlb(iommu,
                                type >> DMA_TLB_FLUSH_GRANU_OFFSET, dr,
                                dw, did, size_order, 0, addr);
-        if ( flush_dev_iotlb )
-            ret = dev_invalidate_iotlb(iommu, did, addr, size_order, type);
+
+        /*
+         * Before Device-TLB invalidation we need to synchronize
+         * invalidation completions with hardware.
+         */
         rc = invalidate_sync(iommu);
+        if ( rc )
+             return rc;
+
+        if ( flush_dev_iotlb )
+            ret = dev_invalidate_iotlb(iommu, did, addr, size_order,
+                                       type, lock);
+
         if ( !ret )
             ret = rc;
     }
diff --git a/xen/drivers/passthrough/vtd/x86/ats.c 
b/xen/drivers/passthrough/vtd/x86/ats.c
index 7c797f6..dcc3394 100644
--- a/xen/drivers/passthrough/vtd/x86/ats.c
+++ b/xen/drivers/passthrough/vtd/x86/ats.c
@@ -106,7 +106,7 @@ out:
 }
 
 int dev_invalidate_iotlb(struct iommu *iommu, u16 did,
-    u64 addr, unsigned int size_order, u64 type)
+    u64 addr, unsigned int size_order, u64 type, unsigned int lock)
 {
     struct pci_ats_dev *pdev;
     int ret = 0;
@@ -162,6 +162,18 @@ int dev_invalidate_iotlb(struct iommu *iommu, u16 did,
             return -EOPNOTSUPP;
         }
 
+        /*
+         * Synchronize with hardware for Device-TLB invalidate
+         * descriptor.
+         */
+        ret = dev_invalidate_iotlb_sync(iommu, did, pdev->seg,
+                                        pdev->bus, pdev->devfn, lock);
+        if ( ret )
+            printk(XENLOG_ERR
+                   "Flush error %d on device %04x:%02x:%02x.%02x \n",
+                   ret, pdev->seg, pdev->bus, PCI_SLOT(pdev->devfn),
+                   PCI_FUNC(pdev->devfn));
+
         if ( !ret )
             ret = rc;
     }
-- 
1.9.1


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel