Fix inline assembler constraints for vmcall_xen and vmcall_kvm

Author: JaredWright

devpal/linux/x64/intel/vmcall_kvm.c

@@ -29,8 +29,8 @@ long handle_devpal_ioctl_vmcall_kvm(struct vmcall_kvm_operands * user_ops)
         "vmcall;"
         "mov %%rax, %[aout];"
         : [aout] "=r"(rax_out)
-        : [rax] "r"(rax_in), [rbx] "r"(rbx_in), [rcx] "r"(rcx_in), [rdx] "r"(rdx_in),
-          [rsi] "r"(rsi_in)
+        : [rax] "g"(rax_in), [rbx] "g"(rbx_in), [rcx] "g"(rcx_in), [rdx] "g"(rdx_in),
+          [rsi] "g"(rsi_in)
         : "rax", "rbx", "rcx", "rdx", "rsi"
     );
 

devpal/linux/x64/intel/vmcall_xen.c

@@ -35,8 +35,8 @@ long handle_devpal_ioctl_vmcall_xen(struct vmcall_xen_operands * user_ops)
         "vmcall;"
         "mov %%rax, %[aout];"
         : [aout] "=r"(rax_out)
-        : [rax] "r"(rax_in), [rdi] "r"(rdi_in), [rsi] "r"(rsi_in), [rdx] "r"(rdx_in),
-          [r10] "r"(r10_in), [r8] "r"(r8_in), [r9] "r"(r9_in)
+        : [rax] "g"(rax_in), [rdi] "g"(rdi_in), [rsi] "g"(rsi_in), [rdx] "g"(rdx_in),
+          [r10] "g"(r10_in), [r8] "g"(r8_in), [r9] "g"(r9_in)
         : "rax", "rcx", "rdx", "rdi", "rsi", "r8", "r9", "r10", "r11"
     );