docs: update xenstore migration design document
PATCH 1/1 docs: update xenstore migration design document
In the past there have been several attempts to make Xenstore restartable. This requires to transfer the internal Xenstore state to the new instance. With the Xenstore migration protocol added recently to Xen's documentation a first base has been defined to represent the state of Xenstore. This can be expanded a little bit in order to have a full state representation which is needed as a first step for live updating Xenstore.
Add some definitions to designs/xenstore-migration.md which are needed for live update of xenstored.
--- docs/designs/xenstore-migration.md | 90 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 87 insertions(+), 3 deletions(-)
docs/designs/xenstore-migration.md +87 -3
@@ -9,6 +9,10 @@ records must include details of registered xenstore watches as well as content; information that cannot currently be recovered from `xenstored`, and hence some extension to the xenstore protocol[2] will also be required. +As a similar set of data is needed for transferring xenstore data from +one instance to another when live updating xenstored the same definitions +are being used. + The *libxenlight Domain Image Format* specification[3] already defines a record type `EMULATOR_XENSTORE_DATA` but this is not suitable for transferring xenstore data pertaining to the domain directly as it is
@@ -48,7 +52,10 @@ where type is one of the following values | | 0x00000001: NODE_DATA | | | 0x00000002: WATCH_DATA | | | 0x00000003: TRANSACTION_DATA | -| | 0x00000004 - 0xFFFFFFFF: reserved for future use | +| | 0x00000004: TRANSACTION_NODE_DATA | +| | 0x00000005: GUEST_RING_DATA | +| | 0x00000006: DOMAIN_START (live update only) | +| | 0x00000007 - 0xFFFFFFFF: reserved for future use | and data is one of the record data formats described in the following
@@ -79,7 +86,7 @@ as follows: +-------------------------------+ | perm count (N) | +-------------------------------+ -| perm0 | +| perm1 | +-------------------------------+ ... +-------------------------------+
@@ -93,7 +100,7 @@ as follows: +-------------------------------+ ``` -where perm0..N are formatted as follows: +where perm1..N are formatted as follows: ```
@@ -164,6 +171,83 @@ as follows: where tx_id is the non-zero identifier values of an open transaction. +**TRANSACTION_NODE_DATA** + + +Each TRANSACTION_NODE_DATA record specifies a transaction local xenstore +node. Its is similar to the NODE_DATA record with the addition of a +transaction id: + +``` + 0 1 2 3 octet ++-------+-------+-------+-------+ +| TRANSACTION_NODE_DATA | ++-------------------------------+ +| tx_id | ++-------------------------------+ +| path length | ++-------------------------------+ +| path data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +| perm count (N) | ++-------------------------------+ +| perm1 | ++-------------------------------+ +... ++-------------------------------+ +| permN | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +where perm1..N are formatted as specified in the NODE_DATA record. A perm +count of 0 denotes a node having been deleted in the transaction. + + +**GUEST_RING_DATA** + + +The GUEST_RING_DATA record is used to transfer data which is pending to be +written to the guest's xenstore ring buffer. It si formatted as follows: + + +``` ++-------+-------+-------+-------+ +| GUEST_RING_DATA | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +**DOMAIN_START** + + +For live updating xenstored data of multiple domains needs to be transferred. +For this purpose the DOMAIN_START record is being used. All records of types +other than NODE_DATA always relate to the last DOMAIN_START record in the +stream. A DOMAIN_START record just contains a domain-id: + + +``` ++-------+-------+-------+-------+ +| DOMAIN_START | ++-------------------------------+ +| domid | pad | ++-------------------------------+ +``` + + ### Protocol Extension Before xenstore state is migrated it is necessary to wait for any pending
Signature · 1 line
2.16.4
Quoted 30 lines
-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Juergen Gross
Sent: 14 April 2020 17:00
To: [email protected]
Cc: Juergen Gross <[email protected]>; Stefano Stabellini <[email protected]>; Julien Grall <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Ian Jackson <[email protected]>; George Dunlap <[email protected]>; Jan Beulich <[email protected]>
Subject: [PATCH] docs: update xenstore migration design documentIn the past there have been several attempts to make Xenstore restartable. This requires to transfer the internal Xenstore state to the new instance. With the Xenstore migration protocol added recently to Xen's documentation a first base has been defined to represent the state of Xenstore. This can be expanded a little bit in order to have a full state representation which is needed as a first step for live updating Xenstore.
Add some definitions to
designs/xenstore-migration.mdwhich are needed for live update of xenstored.Signed-off-by Juergen Gross <[email protected]>--- docs/designs/xenstore-migration.md | 90 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 87 insertions(+), 3 deletions(-)docs/designs/xenstore-migration.md +0
-0
docs/designs/xenstore-migration.md records must include details of registered xenstore watches as well as · line 15
Quoted 7 lines
docs/designs/xenstore-migration.md +4
-0@@ -9,6 +9,10 @@ records must include details of registered xenstore watches as well as content; information that cannot currently be recovered from `xenstored`, and hence some extension to the xenstore protocol[2] will also be required. +As a similar set of data is needed for transferring xenstore data from +one instance to another when live updating xenstored the same definitions +are being used. +
That makes sense, but using an external entity to extract the state makes less sense in the context of live update so, going forward, I suggest dropping the section on extra commands.
Also, we should define a separate 'xenstore domain image format' which can be included as an opaque entity in the migration stream, in the same way as the qemu save image.
Quoted 66 lines
docs/designs/xenstore-migration.md +35
-3The *libxenlight Domain Image Format* specification[3] already defines a record type `EMULATOR_XENSTORE_DATA` but this is not suitable for transferring xenstore data pertaining to the domain directly as it is@@ -48,7 +52,10 @@ where type is one of the following values | | 0x00000001: NODE_DATA | | | 0x00000002: WATCH_DATA | | | 0x00000003: TRANSACTION_DATA | -| | 0x00000004 - 0xFFFFFFFF: reserved for future use | +| | 0x00000004: TRANSACTION_NODE_DATA | +| | 0x00000005: GUEST_RING_DATA | +| | 0x00000006: DOMAIN_START (live update only) | +| | 0x00000007 - 0xFFFFFFFF: reserved for future use | and data is one of the record data formats described in the following@@ -79,7 +86,7 @@ as follows: +-------------------------------+ | perm count (N) | +-------------------------------+ -| perm0 | +| perm1 | +-------------------------------+ +-------------------------------+@@ -93,7 +100,7 @@ as follows: +-------------------------------+ ``` -where perm0..N are formatted as follows: +where perm1..N are formatted as follows: ```@@ -164,6 +171,83 @@ as follows: where tx_id is the non-zero identifier values of an open transaction. +**TRANSACTION_NODE_DATA** + + +Each TRANSACTION_NODE_DATA record specifies a transaction local xenstore +node. Its is similar to the NODE_DATA record with the addition of a +transaction id: + +``` + 0 1 2 3 octet ++-------+-------+-------+-------+ +| TRANSACTION_NODE_DATA | ++-------------------------------+ +| tx_id | ++-------------------------------+ +| path length | ++-------------------------------+ +| path data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +| perm count (N) | ++-------------------------------+ +| perm1 | ++-------------------------------+ +... ++-------------------------------+ +| permN | ++-------------------------------+ +| value length |
docs/designs/xenstore-migration.md as follows: · line 212
Quoted 10 lines
docs/designs/xenstore-migration.md +10
-0++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +where perm1..N are formatted as specified in the NODE_DATA record. A perm +count of 0 denotes a node having been deleted in the transaction. +
Transferring this state means we can presumably drop the TRANSACTION_DATA, since we can infer open transactions from the presence of these records?
Quoted 17 lines
docs/designs/xenstore-migration.md +17
-0+ +**GUEST_RING_DATA** + + +The GUEST_RING_DATA record is used to transfer data which is pending to be +written to the guest's xenstore ring buffer. It si formatted as follows: + + +``` ++-------+-------+-------+-------+ +| GUEST_RING_DATA | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) |
docs/designs/xenstore-migration.md as follows: · line 239
Quoted 10 lines
docs/designs/xenstore-migration.md +10
-0++-------------------------------+ +``` + +**DOMAIN_START** + + +For live updating xenstored data of multiple domains needs to be transferred. +For this purpose the DOMAIN_START record is being used. All records of types +other than NODE_DATA always relate to the last DOMAIN_START record in the +stream. A DOMAIN_START record just contains a domain-id:
If we define a separate stream format as I suggest above, I'd expect the domain-id to be part of the header.
docs/designs/xenstore-migration.md as follows: · line 251
Quoted 12 lines
docs/designs/xenstore-migration.md +11
-0+ + +``` ++-------+-------+-------+-------+ +| DOMAIN_START | ++-------------------------------+ +| domid | pad | ++-------------------------------+ +``` + + ### Protocol Extension
As mentioned above, this section ought to be dropped for the moment. We should define the ways in which xenstored should dump state in various scenarios: e.g. for LU it will likely be serialize to memory (possibly dev/shmem?), for migration/save-restore it will probably be serialize to fd (socket/file). We should also define how dumping state will be triggered.
Paul
Quoted 33 lines · Paul Durrant
Quoted 33 lines
-----Original Message-----
From: Xen-devel <[email protected]> On Behalf Of Juergen Gross
Sent: 14 April 2020 17:00
To: [email protected]
Cc: Juergen Gross <[email protected]>; Stefano Stabellini <[email protected]>; Julien Grall <[email protected]>; Wei Liu <[email protected]>; Andrew Cooper <[email protected]>; Ian Jackson <[email protected]>; George Dunlap <[email protected]>; Jan Beulich <[email protected]>
Subject: [PATCH] docs: update xenstore migration design documentIn the past there have been several attempts to make Xenstore restartable. This requires to transfer the internal Xenstore state to the new instance. With the Xenstore migration protocol added recently to Xen's documentation a first base has been defined to represent the state of Xenstore. This can be expanded a little bit in order to have a full state representation which is needed as a first step for live updating Xenstore.
Add some definitions to
designs/xenstore-migration.mdwhich are needed for live update of xenstored.Signed-off-by Juergen Gross <[email protected]>--- docs/designs/xenstore-migration.md | 90 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 87 insertions(+), 3 deletions(-)docs/designs/xenstore-migration.md +0
-0@@ -9,6 +9,10 @@ records must include details of registered xenstore watches as well as content; information that cannot currently be recovered from `xenstored`, and hence some extension to the xenstore protocol[2] will also be required.
Quoted 7 lines
docs/designs/xenstore-migration.md records must include details of registered xenstore watches as well as · line 15
Quoted 4 lines
docs/designs/xenstore-migration.md +4
-0+As a similar set of data is needed for transferring xenstore data from +one instance to another when live updating xenstored the same definitions +are being used. +That makes sense, but using an external entity to extract the state makes less sense in the context of live update so, going forward, I suggest dropping the section on extra commands.
Right, this (if still needed) should rather go to docs/misc/xenstore.txt
Quoted 2 lines
Also, we should define a separate 'xenstore domain image format' which can be included as an opaque entity in the migration stream, in the same way as the qemu save image.
Fine with me.
Quoted 72 lines
Quoted 72 lines
docs/designs/xenstore-migration.md +40
-3The *libxenlight Domain Image Format* specification[3] already defines a record type `EMULATOR_XENSTORE_DATA` but this is not suitable for transferring xenstore data pertaining to the domain directly as it is@@ -48,7 +52,10 @@ where type is one of the following values | | 0x00000001: NODE_DATA | | | 0x00000002: WATCH_DATA | | | 0x00000003: TRANSACTION_DATA | -| | 0x00000004 - 0xFFFFFFFF: reserved for future use | +| | 0x00000004: TRANSACTION_NODE_DATA | +| | 0x00000005: GUEST_RING_DATA | +| | 0x00000006: DOMAIN_START (live update only) | +| | 0x00000007 - 0xFFFFFFFF: reserved for future use | and data is one of the record data formats described in the following@@ -79,7 +86,7 @@ as follows: +-------------------------------+ | perm count (N) | +-------------------------------+ -| perm0 | +| perm1 | +-------------------------------+ +-------------------------------+@@ -93,7 +100,7 @@ as follows: +-------------------------------+ ``` -where perm0..N are formatted as follows: +where perm1..N are formatted as follows: ```@@ -164,6 +171,83 @@ as follows: where tx_id is the non-zero identifier values of an open transaction. +**TRANSACTION_NODE_DATA** + + +Each TRANSACTION_NODE_DATA record specifies a transaction local xenstore +node. Its is similar to the NODE_DATA record with the addition of a +transaction id: + +``` + 0 1 2 3 octet ++-------+-------+-------+-------+ +| TRANSACTION_NODE_DATA | ++-------------------------------+ +| tx_id | ++-------------------------------+ +| path length | ++-------------------------------+ +| path data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +| perm count (N) | ++-------------------------------+ +| perm1 | ++-------------------------------+ +... ++-------------------------------+ +| permN | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+
Quoted 8 lines
docs/designs/xenstore-migration.md as follows: · line 212
Quoted 5 lines
docs/designs/xenstore-migration.md +5
-0+``` + +where perm1..N are formatted as specified in the NODE_DATA record. A perm +count of 0 denotes a node having been deleted in the transaction. +Transferring this state means we can presumably drop the
TRANSACTION_DATA, since we can infer open transactions from the presence of these records?
No, I don't think so. Imagine a case where just a transaction has been started without any further activity in this transaction having happened yet.
Quoted 22 lines
Quoted 22 lines
docs/designs/xenstore-migration.md +21
-0+ +**GUEST_RING_DATA** + + +The GUEST_RING_DATA record is used to transfer data which is pending to be +written to the guest's xenstore ring buffer. It si formatted as follows: + + +``` ++-------+-------+-------+-------+ +| GUEST_RING_DATA | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +**DOMAIN_START**
Quoted 8 lines
docs/designs/xenstore-migration.md as follows: · line 239
Quoted 6 lines
docs/designs/xenstore-migration.md +6
-0+ + +For live updating xenstored data of multiple domains needs to be transferred. +For this purpose the DOMAIN_START record is being used. All records of types +other than NODE_DATA always relate to the last DOMAIN_START record in the +stream. A DOMAIN_START record just contains a domain-id:If we define a separate stream format as I suggest above, I'd expect the domain-id to be part of the header.
Yes.
Quoted 17 lines
docs/designs/xenstore-migration.md as follows: · line 251
Quoted 13 lines
docs/designs/xenstore-migration.md +11
-0+ + +``` ++-------+-------+-------+-------+ +| DOMAIN_START | ++-------------------------------+ +| domid | pad | ++-------------------------------+ +``` + + ### Protocol ExtensionAs mentioned above, this section ought to be dropped for the moment. We should define the ways in which xenstored should dump state in various scenarios: e.g. for LU it will likely be serialize to memory (possibly dev/shmem?), for migration/save-restore it will probably be serialize to fd (socket/file). We should also define how dumping state will be triggered.
I was planning to use XS_CONTROL for this purpose. But I'm open for other solutions, too. And even using XS_CONTROL might want us to have a library routine for that purpose encapsulating the way how the extracted state is being transferred.
In the LU case the state dumping will need to be handled Xenstore internally (especially in stubdom this will just be a memory blob inside the stubdom, while in the daemon case it could easily be a file).
Juergen
Quoted 115 lines · Juergen Gross
In the past there have been several attempts to make Xenstore restartable. This requires to transfer the internal Xenstore state to the new instance. With the Xenstore migration protocol added recently to Xen's documentation a first base has been defined to represent the state of Xenstore. This can be expanded a little bit in order to have a full state representation which is needed as a first step for live updating Xenstore.
Add some definitions to
designs/xenstore-migration.mdwhich are needed
for live update of xenstored.Signed-off-by Juergen Gross <[email protected]>--- docs/designs/xenstore-migration.md | 90 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 87 insertions(+), 3 deletions(-)docs/designs/xenstore-migration.md +0
-0@@ -9,6 +9,10 @@ records must include details of registered xenstorewatches as well as
content; information that cannot currently be recovered fromxenstored,
and hence some extension to the xenstore protocol[2] will also be
required.+As a similar set of data is needed for transferring xenstore data from
+one instance to another when live updating xenstored the same definitionsdocs/designs/xenstore-migration.md +2
-0+are being used. + The *libxenlight Domain Image Format* specification[3] alreadydefines a
record typeEMULATOR_XENSTORE_DATAbut this is not suitable for transferring xenstore data pertaining to the domain directly as it
isdocs/designs/xenstore-migration.md +4
-1@@ -48,7 +52,10 @@ where type is one of the following values | | 0x00000001: NODE_DATA | | | 0x00000002: WATCH_DATA | | | 0x00000003: TRANSACTION_DATA | -| | 0x00000004 - 0xFFFFFFFF: reserved for future use | +| | 0x00000004: TRANSACTION_NODE_DATA | +| | 0x00000005: GUEST_RING_DATA | +| | 0x00000006: DOMAIN_START (live update only) | +| | 0x00000007 - 0xFFFFFFFF: reserved for future use | and data is one of the record data formats described in thefollowing
docs/designs/xenstore-migration.md +2
-2@@ -79,7 +86,7 @@ as follows: +-------------------------------+ | perm count (N) | +-------------------------------+ -| perm0 | +| perm1 | +-------------------------------+ +-------------------------------+@@ -93,7 +100,7 @@ as follows: +-------------------------------+ ``` -where perm0..N are formatted as follows: +where perm1..N are formatted as follows: ```@@ -164,6 +171,83 @@ as follows: where tx_id is the non-zero identifier values of an opentransaction.
docs/designs/xenstore-migration.md +4
-0+**TRANSACTION_NODE_DATA** + + +Each TRANSACTION_NODE_DATA record specifies a transaction localxenstore
docs/designs/xenstore-migration.md +33
-0+node. Its is similar to the NODE_DATA record with the addition of a +transaction id: + +``` + 0 1 2 3 octet ++-------+-------+-------+-------+ +| TRANSACTION_NODE_DATA | ++-------------------------------+ +| tx_id | ++-------------------------------+ +| path length | ++-------------------------------+ +| path data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +| perm count (N) | ++-------------------------------+ +| perm1 | ++-------------------------------+ +... ++-------------------------------+ +| permN | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +where perm1..N are formatted as specified in the NODE_DATA record. A
Quoted 2 lines
perm
+count of 0 denotes a node having been deleted in the transaction.
oxenstored also tracks the number of operations that a transaction has performed, which includes readonly operations AFAICT, which cannot be inferred from counting TRANSACTION_NODE_DATA entries.
I think the operation count would have to be serialized as part of TRANSACTION_DATA.
Quoted 8 lines
docs/designs/xenstore-migration.md +6
-0+ + +**GUEST_RING_DATA** + + +The GUEST_RING_DATA record is used to transfer data which is pendingto be
+written to the guest's xenstore ring buffer. It si formatted as
typo: s/si/is/
Quoted 24 lines
follows:
docs/designs/xenstore-migration.md +18
-0+ + +``` ++-------+-------+-------+-------+ +| GUEST_RING_DATA | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +**DOMAIN_START** + + +For live updating xenstored data of multiple domains needs to betransferred.
+For this purpose theDOMAIN_STARTrecord is being used. All records of types
+other thanNODE_DATAalways relate to the lastDOMAIN_STARTrecord in the
docs/designs/xenstore-migration.md as follows: · line 248
Quoted 10 lines
docs/designs/xenstore-migration.md +10
-0+stream. A DOMAIN_START record just contains a domain-id: + + +``` ++-------+-------+-------+-------+ +| DOMAIN_START | ++-------------------------------+ +| domid | pad | ++-------------------------------+ +```
There is some more information that might be useful here: mfn and remote_port. (based on the information that INTRODUCE needs)
Best regards,
--Edwin
Quoted 117 lines · Edwin Torok
Quoted 116 lines · Juergen Gross
In the past there have been several attempts to make Xenstore restartable. This requires to transfer the internal Xenstore state to the new instance. With the Xenstore migration protocol added recently to Xen's documentation a first base has been defined to represent the state of Xenstore. This can be expanded a little bit in order to have a full state representation which is needed as a first step for live updating Xenstore.
Add some definitions to
designs/xenstore-migration.mdwhich are needed
for live update of xenstored.Signed-off-by Juergen Gross <[email protected]>--- docs/designs/xenstore-migration.md | 90 ++++++++++++++++++++++++++++++++++++-- 1 file changed, 87 insertions(+), 3 deletions(-)docs/designs/xenstore-migration.md +0
-0@@ -9,6 +9,10 @@ records must include details of registered xenstorewatches as well as
content; information that cannot currently be recovered fromxenstored,
and hence some extension to the xenstore protocol[2] will also be
required.+As a similar set of data is needed for transferring xenstore data from
+one instance to another when live updating xenstored the same definitionsdocs/designs/xenstore-migration.md +2
-0+are being used. + The *libxenlight Domain Image Format* specification[3] alreadydefines a
record typeEMULATOR_XENSTORE_DATAbut this is not suitable for transferring xenstore data pertaining to the domain directly as it
isdocs/designs/xenstore-migration.md +4
-1@@ -48,7 +52,10 @@ where type is one of the following values | | 0x00000001: NODE_DATA | | | 0x00000002: WATCH_DATA | | | 0x00000003: TRANSACTION_DATA | -| | 0x00000004 - 0xFFFFFFFF: reserved for future use | +| | 0x00000004: TRANSACTION_NODE_DATA | +| | 0x00000005: GUEST_RING_DATA | +| | 0x00000006: DOMAIN_START (live update only) | +| | 0x00000007 - 0xFFFFFFFF: reserved for future use | and data is one of the record data formats described in thefollowing
docs/designs/xenstore-migration.md +2
-2@@ -79,7 +86,7 @@ as follows: +-------------------------------+ | perm count (N) | +-------------------------------+ -| perm0 | +| perm1 | +-------------------------------+ +-------------------------------+@@ -93,7 +100,7 @@ as follows: +-------------------------------+ ``` -where perm0..N are formatted as follows: +where perm1..N are formatted as follows: ```@@ -164,6 +171,83 @@ as follows: where tx_id is the non-zero identifier values of an opentransaction.
docs/designs/xenstore-migration.md +4
-0+**TRANSACTION_NODE_DATA** + + +Each TRANSACTION_NODE_DATA record specifies a transaction localxenstore
docs/designs/xenstore-migration.md +33
-0+node. Its is similar to the NODE_DATA record with the addition of a +transaction id: + +``` + 0 1 2 3 octet ++-------+-------+-------+-------+ +| TRANSACTION_NODE_DATA | ++-------------------------------+ +| tx_id | ++-------------------------------+ +| path length | ++-------------------------------+ +| path data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +| perm count (N) | ++-------------------------------+ +| perm1 | ++-------------------------------+ +... ++-------------------------------+ +| permN | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +where perm1..N are formatted as specified in the NODE_DATA record. Aperm
Quoted 8 lines
Quoted 1 line
+count of 0 denotes a node having been deleted in the transaction.
oxenstored also tracks the number of operations that a transaction has performed, which includes readonly operations AFAICT, which cannot be inferred from counting
TRANSACTION_NODE_DATAentries.
I think the operation count would have to be serialized as part ofTRANSACTION_DATA.
No, I don't think this is necessary. The read nodes can be included in the TRANSACTION_NODE_DATA entries, too, as long as the transaction is terminated with failure (EAGAIN). In case oxenstored is needing more, e.g. access types, we can include that.
The TRANSACTION_NODE_DATA entries are primarily needed to ensure returning consistent data in case of reads of nodes after having them accessed before in the same transaction.
Quoted 11 lines
Quoted 9 lines
docs/designs/xenstore-migration.md +6
-0+ + +**GUEST_RING_DATA** + + +The GUEST_RING_DATA record is used to transfer data which is pendingto be
+written to the guest's xenstore ring buffer. It si formatted astypo: s/si/is/
Thanks.
Quoted 30 lines
Quoted 30 lines
follows:
docs/designs/xenstore-migration.md +18
-0+ + +``` ++-------+-------+-------+-------+ +| GUEST_RING_DATA | ++-------------------------------+ +| value length | ++-------------------------------+ +| value data | +... +| pad (0 to 3 octets) | ++-------------------------------+ +``` + +**DOMAIN_START** + + +For live updating xenstored data of multiple domains needs to betransferred.
+For this purpose theDOMAIN_STARTrecord is being used. All records of types
+other thanNODE_DATAalways relate to the lastDOMAIN_STARTrecord in thedocs/designs/xenstore-migration.md +5
-0+stream. A DOMAIN_START record just contains a domain-id: + + +``` ++-------+-------+-------+-------+
Quoted 8 lines
Quoted 5 lines
+|
DOMAIN_START|
++-------------------------------+
+| domid | pad |
++-------------------------------+
+```There is some more information that might be useful here: mfn and
remote_port. (based on the information that INTRODUCE needs)
Oh yes, indeed. And additionally we need something like SOCKET_START with the file descriptor of a socket based connection, and a global entry with the main socket file descriptor used for connecting.
Juergen