There is really no point in faulting in memory regions page by page
if they are not backed by demand paged system RAM but by a linear
passthrough mapping of a host MMIO region.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
---
I have omitted the other 5 patches of the series of which this was #6, as
Christoffer had indicated they could be merged separately.
Changes since v1:
- move this logic to kvm_arch_prepare_memory_region() so it can be invoked
when moving memory regions as well as when creating memory regions
- as we are reasoning about memory regions now instead of memslots, all data
is retrieved from the 'mem' argument which points to a struct
kvm_userspace_memory_region
- minor tweaks to the logic flow
Again, compile tested only, due to lack of test cases.
arch/arm/kvm/mmu.c | 52 +++++++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 51 insertions(+), 1 deletion(-)
diff --git a/arch/arm/kvm/mmu.c b/arch/arm/kvm/mmu.c
index fe53c3a30383..1403d9dc1190 100644
--- a/arch/arm/kvm/mmu.c
+++ b/arch/arm/kvm/mmu.c
@@ -1151,7 +1151,57 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm,
struct kvm_userspace_memory_region *mem,
enum kvm_mr_change change)
{
- return 0;
+ hva_t hva = mem->userspace_addr;
+ hva_t reg_end = hva + mem->memory_size;
+ phys_addr_t gpa = mem->guest_phys_addr;
+ int ret = 0;
+
+ if (change != KVM_MR_CREATE && change != KVM_MR_MOVE)
+ return 0;
+
+ /*
+ * A memory region could potentially cover multiple VMAs, so iterate
+ * over all of them to find out if we can map any of them right now.
+ *
+ * +--------------------------------------------+
+ * +---+---------+-------------------+--------------+----+
+ * | : VMA 1 | VMA 2 | VMA 3 : |
+ * +---+---------+-------------------+--------------+----+
+ * | memory region |
+ * +--------------------------------------------+
+ */
+ do {
+ struct vm_area_struct *vma = find_vma(current->mm, hva);
+ hva_t vm_end;
+
+ if (!vma || vma->vm_start > hva) {
+ ret = -EFAULT;
+ break;
+ }
+
+ vm_end = min(reg_end, vma->vm_end);
+
+ if (vma->vm_flags & VM_PFNMAP) {
+ phys_addr_t pa = (vma->vm_pgoff << PAGE_SHIFT) + hva -
+ vma->vm_start;
+ bool writable = (vma->vm_flags & VM_WRITE) &&
+ !(mem->flags & KVM_MEM_READONLY);
+
+ ret = kvm_phys_addr_ioremap(kvm, gpa, pa, vm_end - hva,
+ writable);
+ if (ret)
+ break;
+ }
+ gpa += vm_end - hva;
+ hva = vm_end;
+ } while (hva < reg_end);
+
+ if (ret) {
+ spin_lock(&kvm->mmu_lock);
+ unmap_stage2_range(kvm, mem->guest_phys_addr, mem->memory_size);
+ spin_unlock(&kvm->mmu_lock);
+ }
+ return ret;
}
void kvm_arch_free_memslot(struct kvm *kvm, struct kvm_memory_slot *free,
--
1.8.3.2
_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
