Skip to contentSkip to conversation

vfio: fix potential deadlock on vfio group lock

PATCH Superseded Matthew Rosato 3 participants 6 messages Jan 13, 2023

Subsystems (3)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 vfio: fix potential deadlock on vfio group lock

Currently it is possible that the final put of a KVM reference comes from vfio during its device close operation. This occurs while the vfio group lock is held; however, if the vfio device is still in the kvm device list, then the following call chain could result in a deadlock:

kvm_put_kvm
-> kvm_destroy_vm
-> kvm_destroy_devices
-> kvm_vfio_destroy

    -> kvm_vfio_file_set_kvm
     -> vfio_file_set_kvm
      -> group->group_lock/group_rwsem

Avoid this scenario by having vfio core code acquire a KVM reference the first time a device is opened and hold that reference until the device fd is closed, at a point after the group lock has been released.

Fixes 421cfe6596f6 ("vfio: remove VFIO_GROUP_NOTIFY_SET_KVM")

---
Changes from v2:
* Re-arrange vfio_kvm_set_kvm_safe error path to still trigger device_open with device->kvm=NULL (Alex)
* get device->dev_set->lock when checking device->open_count (Alex)
* but don't hold it over the kvm_put_kvm (Jason)
* get kvm_put symbol upfront and stash it in device until close (Jason)
* check CONFIG_HAVE_KVM to avoid build errors on architectures without KVM support

Changes from v1:
* Re-write using symbol get logic to get kvm ref during first device open, release the ref during device fd close after group lock is released
* Drop kvm get/put changes to drivers; now that vfio core holds a kvm ref until sometime after the device_close op is called, it should be fine for drivers to get and put their own references to it.

---
 drivers/vfio/group.c     |  6 ++--
 drivers/vfio/vfio_main.c | 78 +++++++++++++++++++++++++++++++++++++---
 include/linux/vfio.h     |  2 +-
 3 files changed, 77 insertions(+), 9 deletions(-)
drivers/vfio/group.c +3 -3
index bb24b2f0271e..2b0da82f82f4 100644
@@ -165,9 +165,9 @@ static int vfio_device_group_open(struct vfio_device *device)
 	}
 
 	/*
-	 * Here we pass the KVM pointer with the group under the lock.  If the
-	 * device driver will use it, it must obtain a reference and release it
-	 * during close_device.
+	 * Here we pass the KVM pointer with the group under the lock.  A
+	 * reference will be obtained the first time the device is opened and
+	 * will be held until the device fd is closed.
 	 */
 	ret = vfio_device_open(device, device->group->iommufd,
 			       device->group->kvm);
drivers/vfio/vfio_main.c +73 -5

117 lines not on this page · raw message

include/linux/vfio.h +1 -1
index 35be78e9ae57..87ff862ff555 100644
@@ -46,7 +46,6 @@ struct vfio_device {
 	struct vfio_device_set *dev_set;
 	struct list_head dev_set_list;
 	unsigned int migration_flags;
-	/* Driver must reference the kvm during open_device or never touch it */
 	struct kvm *kvm;
 
 	/* Members below here are private, not for driver use */
@@ -58,6 +57,7 @@ struct vfio_device {
 	struct list_head group_next;
 	struct list_head iommu_entry;
 	struct iommufd_access *iommufd_access;
+	void (*put_kvm)(struct kvm *kvm);
 #if IS_ENABLED(CONFIG_IOMMUFD)
 	struct iommufd_device *iommufd_device;
 	struct iommufd_ctx *iommufd_ictx;
Signature · 1 line
2.39.0
#1 lore raw

[Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for vfio: fix potential deadlock on vfio group lock (rev2)

== Series Details ==

== Series Details ==

Series: vfio: fix potential deadlock on vfio group lock (rev2)
URL : https://patchwork.freedesktop.org/series/112759/
State : warning

== Summary ==

Error: dim checkpatch failed
3ddf9fe96184 vfio: fix potential deadlock on vfio group lock -:109: CHECK:BRACES: Blank lines aren't necessary after an open brace '{' #109: FILE: drivers/vfio/vfio_main.c:397:
+{
+

-:110: CHECK:BRACES: Blank lines aren't necessary before a close brace '}' #110: FILE: drivers/vfio/vfio_main.c:398:
+
+}

total: 0 errors, 0 warnings, 2 checks, 143 lines checked

#2 lore raw
In reply to Matthew Rosato
Quoted 4 lines · Matthew Rosato
drivers/vfio/vfio_main.c +0 -0
@@ -462,9 +520,19 @@ static inline void vfio_device_pm_runtime_put(struct vfio_device *device)
 static int vfio_device_fops_release(struct inode *inode, struct file *filep)
 {
 	struct vfio_device *device = filep->private_data;

drivers/vfio/vfio_main.c static inline void vfio_device_pm_runtime_put(struct vfio_device *device) · line 532

Quoted 10 lines
drivers/vfio/vfio_main.c +7 -0
+	struct kvm *kvm = NULL;
 
 	vfio_device_group_close(device);
 
+	mutex_lock(&device->dev_set->lock);
+	if (device->open_count == 0 && device->kvm) {
+		kvm = device->kvm;
+		device->kvm = NULL;
+	}
+	mutex_unlock(&device->dev_set->lock);

This still doesn't seem right, another thread could have incr'd the open_count already

This has to be done at the moment open_count is decremented to zero, while still under the original lock.

Jason

#3 lore raw
Quoted 19 lines · Jason Gunthorpe
Quoted 4 lines · Matthew Rosato
drivers/vfio/vfio_main.c +0 -0
@@ -462,9 +520,19 @@ static inline void vfio_device_pm_runtime_put(struct vfio_device *device)
 static int vfio_device_fops_release(struct inode *inode, struct file *filep)
 {
 	struct vfio_device *device = filep->private_data;

drivers/vfio/vfio_main.c static inline void vfio_device_pm_runtime_put(struct vfio_device *device) · line 532

Quoted 10 lines
drivers/vfio/vfio_main.c +7 -0
+	struct kvm *kvm = NULL;
 
 	vfio_device_group_close(device);
 
+	mutex_lock(&device->dev_set->lock);
+	if (device->open_count == 0 && device->kvm) {
+		kvm = device->kvm;
+		device->kvm = NULL;
+	}
+	mutex_unlock(&device->dev_set->lock);

This still doesn't seem right, another thread could have incr'd the open_count already

Quoted 2 lines

This has to be done at the moment open_count is decremented to zero, while still under the original lock.

Hmm.. Fair. Well, we can go back to clearing of device->kvm in vfio_device_last_close but the group lock is held then so we can't immediately do the kvm_put at that time -- unless we go back to the notion of stacking the kvm_put on a workqueue, but now from vfio. If we do that, I think we also have to scrap the idea of putting the kvm_put_kvm function pointer into device->put_kvm too (or otherwise stash it along with the kvm value to be picked up by the scheduled work).

Another thought would be passing the device->open_count that was read while holding the dev_set->lock back on vfio_close_device() / vfio_device_group_close() as an indicator of whether vfio_device_last_close() was called - then you could use the stashed kvm value because it doesn't matter what's currently in device->kvm or what the current device->open_count is, you know that kvm reference needs to be put.

e.g.:
struct *kvm = device->kvm;
void (*put)(struct kvm *kvm) = device->put_kvm;
opened = vfio_device_group_close(device);
if (opened == 0 && kvm)

	put(kvm);
#4 lore raw
Quoted 6 lines · Matthew Rosato
Quoted 5 lines

This still doesn't seem right, another thread could have incr'd the open_count already

This has to be done at the moment open_count is decremented to zero, while still under the original lock.

Quoted 8 lines

Hmm.. Fair. Well, we can go back to clearing of device->kvm in vfio_device_last_close but the group lock is held then so we can't immediately do the kvm_put at that time -- unless we go back to the notion of stacking the kvm_put on a workqueue, but now from vfio. If we do that, I think we also have to scrap the idea of putting the kvm_put_kvm function pointer into device->put_kvm too (or otherwise stash it along with the kvm value to be picked up by the scheduled work).

Well, you have to keep the same sort of design, the vfio_device_last_close() has to put the kvm on the stack until the group lock is unlocked.

It is messy due to how the functions are nested, but not hard.

Jason

#5 lore raw

[Intel-gfx] ✓ Fi.CI.BAT: success for vfio: fix potential deadlock on vfio group lock (rev2)

In reply to Matthew Rosato
== Series Details ==

== Series Details ==

Series: vfio: fix potential deadlock on vfio group lock (rev2)
URL : https://patchwork.freedesktop.org/series/112759/
State : success

== Summary ==

CI Bug Log - changes from CI_DRM_12583 -> Patchwork_112759v2
====================================================

Summary
-------

**SUCCESS**

No regressions found.

External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/index.html

Participating hosts (44 -> 43)
------------------------------

Additional (1): fi-kbl-8809g
Missing (2): fi-kbl-soraka fi-snb-2520m

Known issues
------------

Here are the changes found in Patchwork_112759v2 that come from known issues:

### IGT changes ###

#### Issues hit ####

* igt@gem_huc_copy@huc-copy:
- fi-kbl-8809g: NOTRUN -> [SKIP][1] ([fdo#109271] / [i915#2190])
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/fi-kbl-8809g/igt@[email protected]

* igt@i915_module_load@reload:
- fi-kbl-8809g: NOTRUN -> [DMESG-WARN][2] ([i915#6899])
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/fi-kbl-8809g/igt@[email protected]

* igt@i915_pm_backlight@basic-brightness:
- fi-kbl-8809g: NOTRUN -> [SKIP][3] ([fdo#109271]) +32 similar issues
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/fi-kbl-8809g/igt@[email protected]

* igt@kms_cursor_legacy@basic-busy-flip-before-cursor@atomic-transitions:
- fi-bsw-kefka: [PASS][4] -> [FAIL][5] ([i915#6298])
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12583/fi-bsw-kefka/igt@kms_cursor_legacy@[email protected]
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/fi-bsw-kefka/igt@kms_cursor_legacy@[email protected]

* igt@runner@aborted:
- fi-kbl-8809g: NOTRUN -> [FAIL][6] ([i915#4312] / [i915#4991])
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/fi-kbl-8809g/igt@[email protected]

#### Possible fixes ####

* igt@i915_selftest@live@slpc:
- {bat-adlp-9}: [DMESG-FAIL][7] ([i915#6367]) -> [PASS][8]
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12583/bat-adlp-9/igt@i915_selftest@[email protected]
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/bat-adlp-9/igt@i915_selftest@[email protected]

* igt@kms_cursor_legacy@basic-busy-flip-before-cursor@atomic-transitions:
- fi-bsw-n3050: [FAIL][9] ([i915#6298]) -> [PASS][10]
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_12583/fi-bsw-n3050/igt@kms_cursor_legacy@[email protected]
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/fi-bsw-n3050/igt@kms_cursor_legacy@[email protected]

{name}: This element is suppressed. This means it is ignored when computing
the status of the difference (SUCCESS, WARNING, or FAILURE).

[fdo#109271]: https://bugs.freedesktop.org/show_bug.cgi?id=109271
[i915#2190]: https://gitlab.freedesktop.org/drm/intel/issues/2190
[i915#4258]: https://gitlab.freedesktop.org/drm/intel/issues/4258
[i915#4312]: https://gitlab.freedesktop.org/drm/intel/issues/4312
[i915#4991]: https://gitlab.freedesktop.org/drm/intel/issues/4991
[i915#6298]: https://gitlab.freedesktop.org/drm/intel/issues/6298
[i915#6367]: https://gitlab.freedesktop.org/drm/intel/issues/6367
[i915#6899]: https://gitlab.freedesktop.org/drm/intel/issues/6899

Build changes
-------------

* Linux: CI_DRM_12583 -> Patchwork_112759v2

CI-20190529: 20190529
CI_DRM_12583: cb00fa40d334dc6cd0b4894cea5488a4491f5cda @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_7119: 1e6d24e6dfa42b22f950f7d5e436b8f9acf8747f @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
Patchwork_112759v2: cb00fa40d334dc6cd0b4894cea5488a4491f5cda @ git://anongit.freedesktop.org/gfx-ci/linux

### Linux commits

8f01896ce126 vfio: fix potential deadlock on vfio group lock

== Logs ==

For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_112759v2/index.html