Adds mising dependencies and uses correct formula for enforcing same version

Signed-off-by: Pau Ruiz Safont <pau.ruizsafont@cloud.com>

Add the missing dependency on xapi-stdext-threads

Signed-off-by: Pau Ruiz Safont <pau.ruizsafont@cloud.com>

With newer dune versions, unused bindings for binaries are also checked.

The biggest chunks of code dropped are from tests:
- the ones in stdext-encodings were missed when removing the tests, this was
  done because they were unused by xapi.
- the export in rrd was never run, it's not trivial to get working due to
  changes that have happened in the meantime, and there were tests introduced
  in that same commit to ensure the data contains sane values.

Signed-off-by: Pau Ruiz Safont <pau.ruizsafont@cloud.com>

This enables more warnings and fixes opam metadata generation

Signed-off-by: Pau Ruiz Safont <pau.ruizsafont@cloud.com>

Calls `varstore-sb-state` to edit the uefi mode of a VM
Takes in input the uuid of a VM and a mode (`setup` or `user`)

Returns the output of the script calls

Signed-off-by: Benjamin Reis <benjamin.reis@vates.tech>

Returns the SecureBoot status of a VM:
- `not_supported`: VM's firmware is not UEFI
- `disabled`: Secureboot is disabled on this VM
- `first_boot`: Secured boot is enabled on this VM and its NVRAM.EFI-variables is empty
- `ready`: Secured boot is enabled on this VM and PK, KEK, db and dbx are defined in its EFI variables
- `ready_no_dbx`: Secured boot is enabled on this VM and PK, KEK, db but not dbx are defined in its EFI variables
- `setup_mode`: Secured boot is enabled on this VM and PK is not defined in its EFI variables
- `certs_incomplete`: Secured boot is enabled on this VM and the certificates defines in its EFI variables are incomplete

Signed-off-by: Benjamin Reis <benjamin.reis@vates.tech>

Returns a pool's state for guest SecureBoot:
- `ready`: the active pool UEFI certificates (custom ones first, default ones if no custom ones) contain PK, KEK, db and dbx
- `ready_no_dbx`: the active pool UEFI certificates contain PK, KEK and db but not dbx
- `not_ready`: otherwise

Signed-off-by: Benjamin Reis <benjamin.reis@vates.tech>

Currently, the storage layer tracks two extra SR health statuses that XAPI is
unaware of. These are "unreachable" and "unavailable". At present, the storage
side simply maps these to "recovering" when sending information to the
toolstack. This change will allow the storage layer to track more accurate SR
health statuses.

Signed-off-by: Colin James <colin.barr@cloud.com>

Signed-off-by: Rob Hoes <rob.hoes@citrix.com>

Signed-off-by: Rob Hoes <rob.hoes@citrix.com>

Signed-off-by: Rob Hoes <rob.hoes@citrix.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Signed-off-by: Christian Lindig <christian.lindig@cloud.com>

Adds the possibility of `xe` commands to pass a `traceparent`.

This can be done through the following ways:

- a command arg: `xe <cmd> --traceparent <traceparent>`;
- a environment variable: `TRACEPARENT=<traceparent>`:
- specifying `traceparent=<traceparent>` in `.xe`.

Priority is given in the same order as above.

This should enable creating a single trace from the `xe-cli` caller
through `xapi`.

Signed-off-by: Gabriel Buica <danutgabriel.buica@cloud.com>

Signed-off-by: Rob Hoes <rob.hoes@citrix.com>

Signed-off-by: Rob Hoes <rob.hoes@citrix.com>

Design docs

Instrument `XenAPI.py` to submit the current traceparent back into xapi
if it can import `opentelemetry`.

Currently, we don't see the traces of `sm` calling back to `xapi` using
`XenAPI.py`. This will instrument `XenAPI.py` to pass a traceparent into
`xapi` when opentelemetry is available.

Signed-off-by: Gabriel Buica <danutgabriel.buica@cloud.com>

Signed-off-by: Colin James <colin.barr@cloud.com>

Implements the changes required to forward CBT-related functionality to the
SMAPIv3 storage plugins, which now support CBT for XFS and GFS2 storage
repositories.

The length of the changed blocks requested is hardcoded as a negative number,
which is interpreted by the related plugins as requesting the changes for the
entire length of the VDIs being compared. SMAPIv1 does not support this
functionality but we retain it here with this behaviour in case we want to make
the functionality explicit in future.

Signed-off-by: Colin James <colin.barr@cloud.com>

Signed-off-by: Danilo Del Busso <danilo.delbusso@cloud.com>

Add a `@default` attribute to the `cluster_stack` field in the cluster
config. This field was introduced in #c1bd0e31a but causes RPU to fail
since it is not optional, and while xapi-clusterd was unmarshalling an
old db (which contains cluster config), it cannot find this field, and
throws an exception.

Adding `@default` can make the rpc library to fill this field in when it
is missing, solving the above problem.

Signed-off-by: Vincent Liu <shuntian.liu2@cloud.com>

…/CP-48995

CP-48995: Instrument `XenAPI.py` to submit a `traceparent`

…/CP-45235

…P-49249

CP-49249: Implement SMAPIv3 CBT Forwarding

…luster-stack-default

CA-393507: Default cluster_stack value

The last reference to fix_firewall.sh was removed in e696eb8 back
in 2011 and the script doesn't appear to be used anywhere else so remove
it.

Signed-off-by: Ross Lagerwall <ross.lagerwall@citrix.com>

…y_CA-393866

CA-393866: Add support for Infinity in Java SDK parser

…a/firewall

Remove fix_firewall.sh