1
Changes to Xen in support of Intel(R) Vanderpool Technology
2
-------------------------------------------------------------
4
Our VT extensions to the Xen hypervisor provide full platform
5
virtualization, including CPU(s), memory, and I/O infrastructure. The
6
generic code in Xen handles and schedules those virtual machines as it
7
does for the existing para-virtualized domains.
9
Full virtualization required by the OS guests requires full device
10
virtualization as well. The device models in BOCHS
11
(http://bochs.sourceforge.net/) were decoupled from the CPU
12
virtualization, and are used to virtualize the legacy devices (such as
13
keyboard, mouse, VGA, IDE) in the PC platform. At this point, the
14
device models run in user mode on domain 0, not in the Xen hypervisor.
16
We would like to thank Ian Pratt and Keir Fraser for reviewing our
17
design and code intensively, and for providing numerous useful
18
suggestions to improve the architecture and code.
20
We have a list of Intel team members who take credit for making this
21
release happen: Yunhong Jiang, Nitin Kamble, Chengyuan Li, Xin Li,
22
Xiaofeng Ling, Benjamin Liu, Asit Mallick, Jun Nakajima, Sunil Saxena,
23
Arun Sharma, Edwin Zhai, Jeff Zheng, and Louis Zhuang. We'll continue
24
to add more features to complete full virtualization in Xen using VT.
26
The notes document the changes to the Xen hypervisor in order to add
27
VT support. The changes to other areas, such as Control Panel will be
28
added as we deliver the code.
30
Summary of changes for the first release
31
----------------------------------------
34
* VT specific event handling and domain management were added.
36
* Shadow mode was extended to support full 32-bit guests
38
* Domain switching code was extended to support VT domain
40
* I/O request handling was added to communicate with the device model
42
* Domain builder was extended to provide the environment when the
43
guest enters the protected mode, including E820 memory and VGA
44
info, typically obtained by BIOS calls.
48
VT (Vanderpool Technology) is based on the new VMX (Virtual
49
Machine Extensions) architecture. The current release of the
50
software supports 32-bit only.
52
* arch/x86/vmx.[ch] and arch/x86/vmx_*.[ch]: created to handle
53
VMX-specific events in order to provide virtual machine.
55
* arch/x86/x86_32/entry.S: new code path was added to have the
56
first-level handler from VM exits. The first-level handler calls
57
the second-level handler in arch/x86/vmx.c.
59
* arch/x86/setup.c: new function start_vmx() to init_intel() to
62
* include/asm-x86/config.h: #ifdef CONFIG_VMX was added.
64
* arch/x86/domain.c: new code patch was added to create a VMX
65
domain given the flag from the control panel.
67
* include/public/io/ioreq.h: A new data structure was added to
68
define the I/O requests between the Xen hypervisor and the
71
Changes to the existing code:
72
-----------------------------
74
* arch/x86/shadow.[ch]: new mode SHM_full_32 was added to support
75
full virtualization. The current Xen code assumes that the guest
76
page directory and tables have _machine_ (or host) physical page
77
frame numbers, and the new code allows to support _guest_
78
physical page frame numbers
80
* include/asm-x86/processor.h: struct arch_vmx_struct arch_vmx has
81
been added to the thread_struct data structure. The arch_vmx has
82
the additional VMX-related CPU context.
84
* arch/x86/io_apic.c: reverse mapping between vector and irq has
85
been added. We will revisit this code when considering MSI