In the Linux kernel, the following vulnerability has been resolved:<\/p>\n
s390\/vfio-ap: fix stale pqap_hook pointer on error in vfio_ap_mdev_set_kvm()<\/p>\n
In vfio_ap_mdev_set_kvm(), kvm->arch.crypto.pqap_hook is set to
\n&matrix_mdev->pqap_hook before the update locks are acquired and the
\nmdev list is checked for a conflicting assignment. If another mdev is
\nalready attached to the same KVM instance, the function returns -EPERM
\nwithout restoring the hook pointer, leaving kvm->arch.crypto.pqap_hook
\npointing at the failing matrix_mdev instead of the mdev that legitimately
\nowns the KVM.<\/p>\n
Since matrix_mdev->kvm is never set on this error path,
\nvfio_ap_mdev_unset_kvm() will not clean up the hook when matrix_mdev
\nis later closed. If matrix_mdev is subsequently freed, any PQAP
\ninstruction executed by the guest will dereference the stale pointer
\nthrough pqap_hook_rwsem, resulting in a use-after-free.<\/p>\n
Since kvm->arch.crypto.pqap_hook is only set in the vfio_ap_mdev_set_kvm()
\nfunction and is cleared in the vfio_ap_mdev_unset_kvm() function, a check
\nfor ‘kvm->arch.crypto.pqap_hook != NULL’ is all that is needed to determine
\nwhether it belongs to another mdev. This will alleviate the need to iterate
\nthe matrix_dev->mdev_list list to see if the kvm object is assigned to
\nanother mdev.This was introduced in v3 to alleviate the need to take the
\nmdevs_lock while iterating the list; however, this did not prevent a
\npotential race condition.<\/p>\n
The pqap_hook_rwsem(write) is now performed inside
\nget_update_locks_for_kvm(), which is updated to acquire
\npqap_hook_rwsem(write) between kvm->lock and mdevs_lock. This ordering
\nis consistent with the PQAP intercept path, which acquires pqap_hook_rwsem
\nin read mode while srcu is held under vcpu->mutex, establishing the
\ndependency: kvm->lock -> vcpu->mutex -> srcu -> pqap_hook_rwsem(read).<\/p>\n
The pqap_hook_rwsem is now released inside the
\nrelease_update_locks_for_kvm(), which is updated to release
\npqap_hook_rwsem(write) between mdevs_lock and kvm->lock.<\/p>\n
Additionally, kvm_put_kvm() in vfio_ap_mdev_unset_kvm() is moved
\nafter release_update_locks_for_kvm(). Previously it was called while
\nkvm->lock was held; if it were ever the last reference, kvm_destroy_vm()
\nwould run under kvm->lock, which would deadlock.<\/p>\n","protected":false},"excerpt":{"rendered":"
In the Linux kernel, the following vulnerability has been resolved:<\/p>\n
s390\/vfio-ap: fix stale pqap_hook pointer on error in vfio_ap_mdev_set_kvm()<\/p>\n
In vfio_ap_mdev_set_kvm(), kvm->arch.crypto.pqap_hook<\/p>\n","protected":false},"template":"","meta":[],"categories":[],"tags":[],"class_list":["post-24647","vuln","type-vuln","status-publish","hentry"],"_links":{"self":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=\/wp\/v2\/vuln\/24647","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=\/wp\/v2\/vuln"}],"about":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=\/wp\/v2\/types\/vuln"}],"wp:attachment":[{"href":"https:\/\/www.54iter.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=24647"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.54iter.cn\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=24647"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.54iter.cn\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=24647"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}