Files
xen/docs/specs/libxl-migration-stream.pandoc
T
Andrew Cooper 5020c170e2 docs/pandoc: Don't escape underscores in the middle of text
Pandoc deliberately (and contrary to markdown) doesn't treat underscores in
the middle of normal text as emphasis markers, as this is almost always the
unhelpful interpretation.

For text which is emphasised using _, an underscore in the middle is
interpreted, but the emphasis marker can be switched to * instead.

One problem case is where we use {} globbing with identifier names, as it
counts as a word break.  Therefore, we do need to retain the escaped
underscore immediately following a closing brace.

No change in content - only formatting.  This results in the text being easier
to read and grep.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Acked-by: Ian Jackson <ian.jackson@eu.citrix.com>
2019-01-02 17:50:36 +00:00

320 lines
10 KiB
Plaintext

% LibXenLight Domain Image Format
% Andrew Cooper <<andrew.cooper3@citrix.com>>
Wen Congyang <<wency@cn.fujitsu.com>>
Yang Hongyang <<hongyang.yang@easystack.cn>>
% Revision 2
Introduction
============
For the purposes of this document, `xl` is used as a representation of any
implementer of the `libxl` API. `xl` should be considered completely
interchangeable with alternates, such as `libvirt` or `xenopsd-xl`.
Purpose
-------
The _domain image format_ is the context of a running domain used for
snapshots of a domain or for transferring domains between hosts during
migration.
There are a number of problems with the domain image format used in Xen 4.5
and earlier (the _legacy format_)
* There is no `libxl` context information. `xl` is required to send certain
pieces of `libxl` context itself.
* The contents of the stream is passed directly through `libxl` to `libxc`.
The legacy `libxc` format contained some information which belonged at the
`libxl` level, resulting in awkward layer violation to return the
information back to `libxl`.
* The legacy `libxc` format was inextensible, causing inextensibility in the
legacy `libxl` handling.
This design addresses the above points, allowing for a completely
self-contained, extensible stream with each layer responsible for its own
appropriate information.
Not Yet Included
----------------
The following features are not yet fully specified and will be
included in a future draft.
* ARM
Overview
========
The image format consists of a _Header_, followed by 1 or more _Records_.
Each record consists of a type and length field, followed by any type-specific
data.
\clearpage
Header
======
The header identifies the stream as a `libxl` stream, including the version of
this specification that it complies with.
All fields in this header shall be in _big-endian_ byte order, regardless of
the setting of the endianness bit.
0 1 2 3 4 5 6 7 octet
+-------------------------------------------------+
| ident |
+-----------------------+-------------------------+
| version | options |
+-----------------------+-------------------------+
--------------------------------------------------------------------
Field Description
----------- --------------------------------------------------------
ident 0x4c6962786c466d74 ("LibxlFmt" in ASCII).
version 0x00000002. The version of this specification.
options bit 0: Endianness. 0 = little-endian, 1 = big-endian.
bit 1: Legacy Format. If set, this stream was created by
the legacy conversion tool.
bits 2-31: Reserved.
--------------------------------------------------------------------
The endianness shall be 0 (little-endian) for images generated on an
i386, x86_64, or arm host.
\clearpage
Record Overview
===============
A record has a record header, type specific data and a trailing footer. If
`length` is not a multiple of 8, the body is padded with zeroes to align the
end of the record on an 8 octet boundary.
0 1 2 3 4 5 6 7 octet
+-----------------------+-------------------------+
| type | body_length |
+-----------+-----------+-------------------------+
| body... |
...
| | padding (0 to 7 octets) |
+-----------+-------------------------------------+
--------------------------------------------------------------------
Field Description
----------- -------------------------------------------------------
type 0x00000000: END
0x00000001: LIBXC_CONTEXT
0x00000002: EMULATOR_XENSTORE_DATA
0x00000003: EMULATOR_CONTEXT
0x00000004: CHECKPOINT_END
0x00000005: CHECKPOINT_STATE
0x00000006 - 0x7FFFFFFF: Reserved for future _mandatory_
records.
0x80000000 - 0xFFFFFFFF: Reserved for future _optional_
records.
body_length Length in octets of the record body.
body Content of the record.
padding 0 to 7 octets of zeros to pad the whole record to a multiple
of 8 octets.
--------------------------------------------------------------------
\clearpage
Emulator Records
----------------
Several records are specifically for emulators, and have a common sub header.
0 1 2 3 4 5 6 7 octet
+------------------------+------------------------+
| emulator_id | index |
+------------------------+------------------------+
| record specific data |
...
+-------------------------------------------------+
--------------------------------------------------------------------
Field Description
------------ ---------------------------------------------------
emulator_id 0x00000000: Unknown (In the case of a legacy stream)
0x00000001: Qemu Traditional
0x00000002: Qemu Upstream
0x00000003 - 0xFFFFFFFF: Reserved for future emulators.
index Index of this emulator for the domain.
--------------------------------------------------------------------
\clearpage
Records
=======
END
----
A end record marks the end of the image, and shall be the final record
in the stream.
0 1 2 3 4 5 6 7 octet
+-------------------------------------------------+
The end record contains no fields; its body_length is 0.
LIBXC_CONTEXT
-------------
A libxc context record is a marker, indicating that the stream should be
handed to `xc_domain_restore()`. `libxc` shall be responsible for reading its
own image format from the stream.
0 1 2 3 4 5 6 7 octet
+-------------------------------------------------+
The libxc context record contains no fields; its body_length is 0[^1].
[^1]: The sending side cannot calculate ahead of time how much data `libxc`
might write into the stream, especially for live migration where the quantity
of data is partially proportional to the elapsed time.
EMULATOR_XENSTORE_DATA
----------------------
A set of xenstore key/value pairs for a specific emulator associated with the
domain.
0 1 2 3 4 5 6 7 octet
+------------------------+------------------------+
| emulator_id | index |
+------------------------+------------------------+
| xenstore key/value data |
...
+-------------------------------------------------+
Xenstore key/value data are encoded as a packed sequence of (key, value)
tuples. Each (key, value) tuple is a packed pair of NUL terminated octets,
conforming to xenstore protocol character encoding (keys strictly as
alphanumeric ASCII and `-/_@`, values expected to be human-readable ASCII).
Keys shall be relative to to the device models xenstore tree for the new
domain. At the time of writing, keys are relative to the path
> `/local/domain/$dm_domid/device-model/$domid/`
although this path is free to change moving forward, thus should not be
assumed.
EMULATOR_CONTEXT
----------------
A context blob for a specific emulator associated with the domain.
0 1 2 3 4 5 6 7 octet
+------------------------+------------------------+
| emulator_id | index |
+------------------------+------------------------+
| emulator_ctx |
...
+-------------------------------------------------+
The *emulator_ctx* is a binary blob interpreted by the emulator identified by
*emulator_id*. Its format is unspecified.
CHECKPOINT_END
--------------
A checkpoint end record marks the end of a checkpoint in the image.
0 1 2 3 4 5 6 7 octet
+-------------------------------------------------+
The end record contains no fields; its body_length is 0.
CHECKPOINT_STATE
----------------
A checkpoint state record contains the control information for checkpoint. It
is only used by COLO, more detail please reference README.colo.
0 1 2 3 4 5 6 7 octet
+------------------------+------------------------+
| control_id | padding |
+------------------------+------------------------+
--------------------------------------------------------------------
Field Description
------------ ---------------------------------------------------
control_id 0x00000000: Secondary VM is out of sync, start a new checkpoint
(Primary -> Secondary)
0x00000001: Secondary VM is suspended (Secondary -> Primary)
0x00000002: Secondary VM is ready (Secondary -> Primary)
0x00000003: Secondary VM is resumed (Secondary -> Primary)
--------------------------------------------------------------------
In COLO, Primary is running in below loop:
1. Suspend primary vm
a. Suspend primary vm
b. Read *CHECKPOINT_SVM_SUSPENDED* sent by secondary
2. Checkpoint
3. Resume primary vm
a. Read *CHECKPOINT_SVM_READY* from secondary
b. Resume primary vm
c. Read *CHECKPOINT_SVM_RESUMED* from secondary
4. Wait a new checkpoint
a. Send *CHECKPOINT_NEW* to secondary
While Secondary is running in below loop:
1. Resume secondary vm
a. Send *CHECKPOINT_SVM_READY* to primary
b. Resume secondary vm
c. Send *CHECKPOINT_SVM_RESUMED* to primary
2. Wait a new checkpoint
a. Read *CHECKPOINT_NEW* from primary
3. Suspend secondary vm
a. Suspend secondary vm
b. Send *CHECKPOINT_SVM_SUSPENDED* to primary
4. Checkpoint
Future Extensions
=================
All changes to this specification should bump the revision number in
the title block.
All changes to the header require the header version to be increased.
The format may be extended by adding additional record types.
Extending an existing record type must be done by adding a new record
type. This allows old images with the old record to still be
restored.