vfio-pci: Don't use device_lock around AER interrupt setup
Subsystems (2)
PATCH 1/1 vfio-pci: Don't use device_lock around AER interrupt setup
device_lock is much too prone to lockups. For instance if we have a pending .remove then device_lock is already held. If userspace attempts to modify AER signaling after that point, a deadlock occurs. eventfd setup/teardown is already protected in vfio with the igate mutex. AER is not a high performance interrupt, so we can also use the same mutex to protect signaling versus setup races.
Signed-off-by Alex Williamson <[email protected]>
--- drivers/vfio/pci/vfio_pci.c | 4 ++++ drivers/vfio/pci/vfio_pci_intrs.c | 17 ----------------- 2 files changed, 4 insertions(+), 17 deletions(-)
drivers/vfio/pci/vfio_pci.c +4 -0
@@ -872,9 +872,13 @@ static pci_ers_result_t vfio_pci_aer_err_detected(struct pci_dev *pdev,return PCI_ERS_RESULT_DISCONNECT ; } +mutex_lock (&vdev->igate); +if (vdev->err_trigger)eventfd_signal (vdev->err_trigger,1 ); +mutex_unlock (&vdev->igate); +vfio_device_put (device);return PCI_ERS_RESULT_CAN_RECOVER ;
drivers/vfio/pci/vfio_pci_intrs.c +0 -17
@@ -749,54 +749,37 @@ static int vfio_pci_set_err_trigger(struct vfio_pci_device *vdev, unsigned count, uint32_t flags, void *data) { int32_t fd = *(int32_t *)data; -struct pci_dev *pdev = vdev->pdev;if ((index !=VFIO_PCI_ERR_IRQ_INDEX ) || !(flags &VFIO_IRQ_SET_DATA_TYPE_MASK ))return -EINVAL ; -/* -* device_lock synchronizes setting and checking of -* err_trigger. The vfio_pci_aer_err_detected() is also -* called with device_lock held. -*/ -/* DATA_NONE/DATA_BOOL enables loopback testing */ -if (flags &VFIO_IRQ_SET_DATA_NONE ) { -device_lock (&pdev->dev);if (vdev->err_trigger)eventfd_signal (vdev->err_trigger,1 ); -device_unlock (&pdev->dev);return 0 ; }else if (flags &VFIO_IRQ_SET_DATA_BOOL ) {uint8_t trigger = *(uint8_t *)data; -device_lock (&pdev->dev);if (trigger && vdev->err_trigger)eventfd_signal (vdev->err_trigger,1 ); -device_unlock (&pdev->dev);return 0 ; }/* Handle SET_DATA_EVENTFD */ -if (fd ==-1 ) { -device_lock (&pdev->dev);if (vdev->err_trigger)eventfd_ctx_put (vdev->err_trigger); vdev->err_trigger =NULL ; -device_unlock (&pdev->dev);return 0 ; }else if (fd >=0 ) {struct eventfd_ctx *efdctx; efdctx =eventfd_ctx_fdget (fd);if (IS_ERR (efdctx))return PTR_ERR (efdctx); -device_lock (&pdev->dev);if (vdev->err_trigger)eventfd_ctx_put (vdev->err_trigger); vdev->err_trigger = efdctx; -device_unlock (&pdev->dev);return 0 ; }else return -EINVAL ;