Merge tag 'vfio-v6.3-rc1' of https://github.com/awilliam/linux-vfio

Pull VFIO updates from Alex Williamson:

 - Remove redundant resource check in vfio-platform (Angus Chen)

 - Use GFP_KERNEL_ACCOUNT for persistent userspace allocations, allowing
   removal of arbitrary kernel limits in favor of cgroup control (Yishai
   Hadas)

 - mdev tidy-ups, including removing the module-only build restriction
   for sample drivers, Kconfig changes to select mdev support,
   documentation movement to keep sample driver usage instructions with
   sample drivers rather than with API docs, remove references to
   out-of-tree drivers in docs (Christoph Hellwig)

 - Fix collateral breakages from mdev Kconfig changes (Arnd Bergmann)

 - Make mlx5 migration support match device support, improve source and
   target flows to improve pre-copy support and reduce downtime (Yishai
   Hadas)

 - Convert additional mdev sysfs case to use sysfs_emit() (Bo Liu)

 - Resolve copy-paste error in mdev mbochs sample driver Kconfig (Ye
   Xingchen)

 - Avoid propagating missing reset error in vfio-platform if reset
   requirement is relaxed by module option (Tomasz Duszynski)

 - Range size fixes in mlx5 variant driver for missed last byte and
   stricter range calculation (Yishai Hadas)

 - Fixes to suspended vaddr support and locked_vm accounting, excluding
   mdev configurations from the former due to potential to indefinitely
   block kernel threads, fix underflow and restore locked_vm on new mm
   (Steve Sistare)

 - Update outdated vfio documentation due to new IOMMUFD interfaces in
   recent kernels (Yi Liu)

 - Resolve deadlock between group_lock and kvm_lock, finally (Matthew
   Rosato)

 - Fix NULL pointer in group initialization error path with IOMMUFD (Yan
   Zhao)

* tag 'vfio-v6.3-rc1' of https://github.com/awilliam/linux-vfio: (32 commits)
  vfio: Fix NULL pointer dereference caused by uninitialized group->iommufd
  docs: vfio: Update vfio.rst per latest interfaces
  vfio: Update the kdoc for vfio_device_ops
  vfio/mlx5: Fix range size calculation upon tracker creation
  vfio: no need to pass kvm pointer during device open
  vfio: fix deadlock between group lock and kvm lock
  vfio: revert "iommu driver notify callback"
  vfio/type1: revert "implement notify callback"
  vfio/type1: revert "block on invalid vaddr"
  vfio/type1: restore locked_vm
  vfio/type1: track locked_vm per dma
  vfio/type1: prevent underflow of locked_vm via exec()
  vfio/type1: exclude mdevs from VFIO_UPDATE_VADDR
  vfio: platform: ignore missing reset if disabled at module init
  vfio/mlx5: Improve the target side flow to reduce downtime
  vfio/mlx5: Improve the source side flow upon pre_copy
  vfio/mlx5: Check whether VF is migratable
  samples: fix the prompt about SAMPLE_VFIO_MDEV_MBOCHS
  vfio/mdev: Use sysfs_emit() to instead of sprintf()
  vfio-mdev: add back CONFIG_VFIO dependency
  ...

1 files changed, 59 insertions, 11 deletions

diff --git a/drivers/vfio/vfio_main.c b/drivers/vfio/vfio_main.c
index 90541fc94988..43bd6b76e2b6 100644
--- a/drivers/vfio/vfio_main.c
+++ b/drivers/vfio/vfio_main.c
@@ -16,6 +16,9 @@
 #include <linux/fs.h>
 #include <linux/idr.h>
 #include <linux/iommu.h>
+#ifdef CONFIG_HAVE_KVM
+#include <linux/kvm_host.h>
+#endif
 #include <linux/list.h>
 #include <linux/miscdevice.h>
 #include <linux/module.h>
@@ -345,6 +348,55 @@ void vfio_unregister_group_dev(struct vfio_device *device)
 }
 EXPORT_SYMBOL_GPL(vfio_unregister_group_dev);
 
+#ifdef CONFIG_HAVE_KVM
+void _vfio_device_get_kvm_safe(struct vfio_device *device, struct kvm *kvm)
+{
+	void (*pfn)(struct kvm *kvm);
+	bool (*fn)(struct kvm *kvm);
+	bool ret;
+
+	lockdep_assert_held(&device->dev_set->lock);
+
+	pfn = symbol_get(kvm_put_kvm);
+	if (WARN_ON(!pfn))
+		return;
+
+	fn = symbol_get(kvm_get_kvm_safe);
+	if (WARN_ON(!fn)) {
+		symbol_put(kvm_put_kvm);
+		return;
+	}
+
+	ret = fn(kvm);
+	symbol_put(kvm_get_kvm_safe);
+	if (!ret) {
+		symbol_put(kvm_put_kvm);
+		return;
+	}
+
+	device->put_kvm = pfn;
+	device->kvm = kvm;
+}
+
+void vfio_device_put_kvm(struct vfio_device *device)
+{
+	lockdep_assert_held(&device->dev_set->lock);
+
+	if (!device->kvm)
+		return;
+
+	if (WARN_ON(!device->put_kvm))
+		goto clear;
+
+	device->put_kvm(device->kvm);
+	device->put_kvm = NULL;
+	symbol_put(kvm_put_kvm);
+
+clear:
+	device->kvm = NULL;
+}
+#endif
+
 /* true if the vfio_device has open_device() called but not close_device() */
 static bool vfio_assert_device_open(struct vfio_device *device)
 {
@@ -352,7 +404,7 @@ static bool vfio_assert_device_open(struct vfio_device *device)
 }
 
 static int vfio_device_first_open(struct vfio_device *device,
-				  struct iommufd_ctx *iommufd, struct kvm *kvm)
+				  struct iommufd_ctx *iommufd)
 {
 	int ret;
 
@@ -368,7 +420,6 @@ static int vfio_device_first_open(struct vfio_device *device,
 	if (ret)
 		goto err_module_put;
 
-	device->kvm = kvm;
 	if (device->ops->open_device) {
 		ret = device->ops->open_device(device);
 		if (ret)
@@ -377,7 +428,6 @@ static int vfio_device_first_open(struct vfio_device *device,
 	return 0;
 
 err_unuse_iommu:
-	device->kvm = NULL;
 	if (iommufd)
 		vfio_iommufd_unbind(device);
 	else
@@ -394,7 +444,6 @@ static void vfio_device_last_close(struct vfio_device *device,
 
 	if (device->ops->close_device)
 		device->ops->close_device(device);
-	device->kvm = NULL;
 	if (iommufd)
 		vfio_iommufd_unbind(device);
 	else
@@ -402,19 +451,18 @@ static void vfio_device_last_close(struct vfio_device *device,
 	module_put(device->dev->driver->owner);
 }
 
-int vfio_device_open(struct vfio_device *device,
-		     struct iommufd_ctx *iommufd, struct kvm *kvm)
+int vfio_device_open(struct vfio_device *device, struct iommufd_ctx *iommufd)
 {
 	int ret = 0;
 
-	mutex_lock(&device->dev_set->lock);
+	lockdep_assert_held(&device->dev_set->lock);
+
 	device->open_count++;
 	if (device->open_count == 1) {
-		ret = vfio_device_first_open(device, iommufd, kvm);
+		ret = vfio_device_first_open(device, iommufd);
 		if (ret)
 			device->open_count--;
 	}
-	mutex_unlock(&device->dev_set->lock);
 
 	return ret;
 }
@@ -422,12 +470,12 @@ int vfio_device_open(struct vfio_device *device,
 void vfio_device_close(struct vfio_device *device,
 		       struct iommufd_ctx *iommufd)
 {
-	mutex_lock(&device->dev_set->lock);
+	lockdep_assert_held(&device->dev_set->lock);
+
 	vfio_assert_device_open(device);
 	if (device->open_count == 1)
 		vfio_device_last_close(device, iommufd);
 	device->open_count--;
-	mutex_unlock(&device->dev_set->lock);
 }
 
 /*