Systems and methods for managing host virtual addresses in a hypercall are disclosed. In one implementation, a processing device may receive, by a hypervisor managing a virtual machine (VM), a hypercall initiated by the VM, wherein a first parameter of the hypercall specifies a guest physical address (GPA) of a memory buffer and a second parameter of the hypercall specifies a host virtual address (HVA) of the memory buffer. The processing device may also translate the GPA to a first host physical address (HPA) and may translate the HVA to a second HPA. The processing device may further compare the first HPA to the second HPA and responsive to determining that the first HPA matches the second HPA, the processing device may execute the hypercall using the HVA.