On 24.05.19 14:36, David Hildenbrand wrote: > On 24.05.19 14:29, Christian Borntraeger wrote: >> >> >> On 24.05.19 14:17, Christian Borntraeger wrote: >>> >>> >>> On 24.05.19 13:11, Christian Borntraeger wrote: >>>> I do get >>>> >>>> [10400.440298] kvm-s390: failed to commit memory region >>>> [10400.508723] kvm-s390: failed to commit memory region >>>> >>>> when running the tests. Will have a look. >>> >>> It comes from kvm_vm_free. This calls KVM_SET_USER_MEMORY_REGION with size 0, >>> which the s390 code does not like. >>> >> >> The doc says about KVM_SET_USER_MEMORY_REGION: >> >> This ioctl allows the user to create or modify a guest physical memory >> slot. When changing an existing slot, it may be moved in the guest >> physical memory space, or its flags may be modified. --> It may not be >> resized. <---- > > Size 0 is deleting, not resizing AFAIK. Right this seems to translate to KVM_MR_DELETE, which the s390 code does not handle (we will simply deliver a page fault as we share the last page table level). I will have a look at implementing KVM_MR_DELETE and KVM_MR_MOVE. In fact, we should have a testcase for that as well.
