drm/i915/gvt: Fix possible recursive locking issue
authorChuanxiao Dong <chuanxiao.dong@intel.com>
Mon, 26 Jun 2017 07:20:49 +0000 (15:20 +0800)
committerZhenyu Wang <zhenyuw@linux.intel.com>
Mon, 26 Jun 2017 08:31:49 +0000 (16:31 +0800)
vfio_unpin_pages will hold a read semaphore however it is already hold
in the same thread by vfio ioctl. It will cause below warning:

[ 5102.127454] ============================================
[ 5102.133379] WARNING: possible recursive locking detected
[ 5102.139304] 4.12.0-rc4+ #3 Not tainted
[ 5102.143483] --------------------------------------------
[ 5102.149407] qemu-system-x86/1620 is trying to acquire lock:
[ 5102.155624]  (&container->group_lock){++++++}, at: [<ffffffff817768c6>] vfio_unpin_pages+0x96/0xf0
[ 5102.165626]
but task is already holding lock:
[ 5102.172134]  (&container->group_lock){++++++}, at: [<ffffffff8177728f>] vfio_fops_unl_ioctl+0x5f/0x280
[ 5102.182522]
other info that might help us debug this:
[ 5102.189806]  Possible unsafe locking scenario:

[ 5102.196411]        CPU0
[ 5102.199136]        ----
[ 5102.201861]   lock(&container->group_lock);
[ 5102.206527]   lock(&container->group_lock);
[ 5102.211191]
*** DEADLOCK ***

[ 5102.217796]  May be due to missing lock nesting notation

[ 5102.225370] 3 locks held by qemu-system-x86/1620:
[ 5102.230618]  #0:  (&container->group_lock){++++++}, at: [<ffffffff8177728f>] vfio_fops_unl_ioctl+0x5f/0x280
[ 5102.241482]  #1:  (&(&iommu->notifier)->rwsem){++++..}, at: [<ffffffff810de775>] __blocking_notifier_call_chain+0x35/0x70
[ 5102.253713]  #2:  (&vgpu->vdev.cache_lock){+.+...}, at: [<ffffffff8157b007>] intel_vgpu_iommu_notifier+0x77/0x120
[ 5102.265163]
stack backtrace:
[ 5102.270022] CPU: 5 PID: 1620 Comm: qemu-system-x86 Not tainted 4.12.0-rc4+ #3
[ 5102.277991] Hardware name: Intel Corporation S1200RP/S1200RP, BIOS S1200RP.86B.03.01.APER.061220151418 06/12/2015
[ 5102.289445] Call Trace:
[ 5102.292175]  dump_stack+0x85/0xc7
[ 5102.295871]  validate_chain.isra.21+0x9da/0xaf0
[ 5102.300925]  __lock_acquire+0x405/0x820
[ 5102.305202]  lock_acquire+0xc7/0x220
[ 5102.309191]  ? vfio_unpin_pages+0x96/0xf0
[ 5102.313666]  down_read+0x2b/0x50
[ 5102.317259]  ? vfio_unpin_pages+0x96/0xf0
[ 5102.321732]  vfio_unpin_pages+0x96/0xf0
[ 5102.326024]  intel_vgpu_iommu_notifier+0xe5/0x120
[ 5102.331283]  notifier_call_chain+0x4a/0x70
[ 5102.335851]  __blocking_notifier_call_chain+0x4d/0x70
[ 5102.341490]  blocking_notifier_call_chain+0x16/0x20
[ 5102.346935]  vfio_iommu_type1_ioctl+0x87b/0x920
[ 5102.351994]  vfio_fops_unl_ioctl+0x81/0x280
[ 5102.356660]  ? __fget+0xf0/0x210
[ 5102.360261]  do_vfs_ioctl+0x93/0x6a0
[ 5102.364247]  ? __fget+0x111/0x210
[ 5102.367942]  SyS_ioctl+0x41/0x70
[ 5102.371542]  entry_SYSCALL_64_fastpath+0x1f/0xbe

put the vfio_unpin_pages in a workqueue can fix this.

v2:
- use for style instead of do{}while(1). (Zhenyu)
v3:
- rename gvt_cache_mark to gvt_cache_mark_remove. (Zhenyu)

Fixes: 659643f7d814 ("drm/i915/gvt/kvmgt: add vfio/mdev support to KVMGT")
Signed-off-by: Chuanxiao Dong <chuanxiao.dong@intel.com>
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Cc: stable@vger.kernel.org # v4.10+
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
drivers/gpu/drm/i915/gvt/gvt.h
drivers/gpu/drm/i915/gvt/kvmgt.c

index 930732e..8b5876d 100644 (file)
@@ -183,6 +183,9 @@ struct intel_vgpu {
                struct kvm *kvm;
                struct work_struct release_work;
                atomic_t released;
+               struct work_struct unpin_work;
+               spinlock_t unpin_lock; /* To protect unpin_list */
+               struct list_head unpin_list;
        } vdev;
 #endif
 };
index 1ae0b40..56a8d6a 100644 (file)
@@ -78,6 +78,7 @@ struct gvt_dma {
        struct rb_node node;
        gfn_t gfn;
        unsigned long iova;
+       struct list_head list;
 };
 
 static inline bool handle_valid(unsigned long handle)
@@ -166,6 +167,7 @@ static void gvt_cache_add(struct intel_vgpu *vgpu, gfn_t gfn,
 
        new->gfn = gfn;
        new->iova = iova;
+       INIT_LIST_HEAD(&new->list);
 
        mutex_lock(&vgpu->vdev.cache_lock);
        while (*link) {
@@ -197,26 +199,52 @@ static void __gvt_cache_remove_entry(struct intel_vgpu *vgpu,
        kfree(entry);
 }
 
-static void gvt_cache_remove(struct intel_vgpu *vgpu, gfn_t gfn)
+static void intel_vgpu_unpin_work(struct work_struct *work)
 {
+       struct intel_vgpu *vgpu = container_of(work, struct intel_vgpu,
+                                              vdev.unpin_work);
        struct device *dev = mdev_dev(vgpu->vdev.mdev);
        struct gvt_dma *this;
-       unsigned long g1;
-       int rc;
+       unsigned long gfn;
+
+       for (;;) {
+               spin_lock(&vgpu->vdev.unpin_lock);
+               if (list_empty(&vgpu->vdev.unpin_list)) {
+                       spin_unlock(&vgpu->vdev.unpin_lock);
+                       break;
+               }
+               this = list_first_entry(&vgpu->vdev.unpin_list,
+                                       struct gvt_dma, list);
+               list_del(&this->list);
+               spin_unlock(&vgpu->vdev.unpin_lock);
+
+               gfn = this->gfn;
+               vfio_unpin_pages(dev, &gfn, 1);
+               kfree(this);
+       }
+}
+
+static bool gvt_cache_mark_remove(struct intel_vgpu *vgpu, gfn_t gfn)
+{
+       struct gvt_dma *this;
 
        mutex_lock(&vgpu->vdev.cache_lock);
        this  = __gvt_cache_find(vgpu, gfn);
        if (!this) {
                mutex_unlock(&vgpu->vdev.cache_lock);
-               return;
+               return false;
        }
-
-       g1 = gfn;
        gvt_dma_unmap_iova(vgpu, this->iova);
-       rc = vfio_unpin_pages(dev, &g1, 1);
-       WARN_ON(rc != 1);
-       __gvt_cache_remove_entry(vgpu, this);
+       /* remove this from rb tree */
+       rb_erase(&this->node, &vgpu->vdev.cache);
        mutex_unlock(&vgpu->vdev.cache_lock);
+
+       /* put this to the unpin_list */
+       spin_lock(&vgpu->vdev.unpin_lock);
+       list_move_tail(&this->list, &vgpu->vdev.unpin_list);
+       spin_unlock(&vgpu->vdev.unpin_lock);
+
+       return true;
 }
 
 static void gvt_cache_init(struct intel_vgpu *vgpu)
@@ -453,6 +481,9 @@ static int intel_vgpu_create(struct kobject *kobj, struct mdev_device *mdev)
        }
 
        INIT_WORK(&vgpu->vdev.release_work, intel_vgpu_release_work);
+       INIT_WORK(&vgpu->vdev.unpin_work, intel_vgpu_unpin_work);
+       spin_lock_init(&vgpu->vdev.unpin_lock);
+       INIT_LIST_HEAD(&vgpu->vdev.unpin_list);
 
        vgpu->vdev.mdev = mdev;
        mdev_set_drvdata(mdev, vgpu);
@@ -482,6 +513,7 @@ static int intel_vgpu_iommu_notifier(struct notifier_block *nb,
        struct intel_vgpu *vgpu = container_of(nb,
                                        struct intel_vgpu,
                                        vdev.iommu_notifier);
+       bool sched_unmap = false;
 
        if (action == VFIO_IOMMU_NOTIFY_DMA_UNMAP) {
                struct vfio_iommu_type1_dma_unmap *unmap = data;
@@ -491,7 +523,10 @@ static int intel_vgpu_iommu_notifier(struct notifier_block *nb,
                end_gfn = gfn + unmap->size / PAGE_SIZE;
 
                while (gfn < end_gfn)
-                       gvt_cache_remove(vgpu, gfn++);
+                       sched_unmap |= gvt_cache_mark_remove(vgpu, gfn++);
+
+               if (sched_unmap)
+                       schedule_work(&vgpu->vdev.unpin_work);
        }
 
        return NOTIFY_OK;