User Space
vcpu_fd = ioctl(vm_fd, KVM_CREATE_VCPU, 0);
kvm_run = (struct kvm_run *)mmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_SHARED, vcpu_fd, 0);
kvm_run 是经常要用到的数据,copy来copy去,效率不高。在这里User Space 是直接使用kernel 的 vcpu->run
我们看看是如何实现的,其实本质就是mmap
Kernel 部分
KVM_CREATE_VCPU
就是调用以下函数
static int kvm_vm_ioctl_create_vcpu(struct kvm *kvm, u32 id)
...
page = alloc_page(GFP_KERNEL | __GFP_ZERO);
if (!page) {
r = -ENOMEM;
goto vcpu_free;
}
vcpu->run = page_address(page); //为vcpu->run申请一个page
...
r = create_vcpu_fd(vcpu); //看名字关键是这里
...
return r;
}
很简单的函数,即使创建见匿名fd
static int create_vcpu_fd(struct kvm_vcpu *vcpu)
{
char name[8 + 1 + ITOA_MAX_LEN + 1];
snprintf(name, sizeof(name), "kvm-vcpu:%d", vcpu->vcpu_id);
return anon_inode_getfd(name, &kvm_vcpu_fops, vcpu, O_RDWR | O_CLOEXEC);
}
注意这个:
kvm_vcpu_fops
static struct file_operations kvm_vcpu_fops = {
.release = kvm_vcpu_release,
.unlocked_ioctl = kvm_vcpu_ioctl,
.mmap = kvm_vcpu_mmap,
.llseek = noop_llseek,
KVM_COMPAT(kvm_vcpu_compat_ioctl),
};
当User Space 调用mmap映射这个fd时,即以下函数, mmap 就会调用
kvm_run = (struct kvm_run *)mmap(NULL, mmap_size, PROT_READ | PROT_WRITE, MAP_SHARED, vcpu_fd, 0);
static int kvm_vcpu_mmap(struct file *file, struct vm_area_struct *vma)
{
vma->vm_ops = &kvm_vcpu_vm_ops;
return 0;
}
static const struct vm_operations_struct kvm_vcpu_vm_ops = {
.fault = kvm_vcpu_fault,
};
static vm_fault_t kvm_vcpu_fault(struct vm_fault *vmf)
{
struct kvm_vcpu *vcpu = vmf->vma->vm_file->private_data;
struct page *page;
if (vmf->pgoff == 0)
page = virt_to_page(vcpu->run); //这个就是KVM_CREATE_VCPU 时,Kernel 申请的1个page
...
get_page(page);
vmf->page = page;
return 0;
}
Kernel 在这个vma 发生abort 时,会将此page 映射到User Space
这样User Space 和Kernel 就Share 这个关键的状态信息。
我在调试时Kernel log
[ 57.044046] kvm_vm_ioctl_create_vcpu page 0xefd140 //Create 时vcpu->run的page
[ 57.044059] kvm_vm_ioctl_create_vcpu vcpu->run 0x43f45000
[ 57.044097] kvm_vcpu_mmap
[ 57.470751] kvm_vcpu_gethpfar_el2(vcpu) 0x1000
[ 57.470767] fault_ipa 0x100000
[ 57.470950] kvm_vcpu_gethpfar_el2(vcpu) 0x80
[ 57.470960] fault_ipa 0x8000
[ 57.470970] kvm_vcpu_get_hfar(vcpu) 0x8128
[ 57.470978] fault_ipa 0x8128 before io_mem_abort
[ 57.471007] kvm_vcpu_fault
[ 57.471017] page 0xefd140 //把这个page 映射到User Space