device-types: add mt8195-cherry-tomato-r2

See merge request !127

Add device type template for the Acer Chromebook Spin 513 tomato
Chromebook.

Upstream MR: https://git.lavasoftware.org/lava/lava/-/merge_requests/1938



Signed-off-by: Laura Nao <laura.nao@collabora.com>

lava_dispatcher: Rework ShellLogger and add unit test for it

See merge request !120

Apply patch to simplify error detection for testing

See merge request !125

Rémi Duraffort authored Dec 07, 2022 and Laura Nao committed Jan 05, 2023

Signed-off-by: Rémi Duraffort <remi.duraffort@linaro.org>

Merge branch 'collabora/staging-lava-dispatcher-mirror-timeout-exception-to-last-exception-class' into 'collabora/staging'

lava_dispatcher: RetryAction use last exception class for final exception

See merge request !124

User APIs for automated management

See merge request !121

This adds two new XMLRPC user management APIs:
- system.get_all_users to find out which users are present on this
  LAVA instance.
- system.set_user_active to lock and unlock accounts on the LAVA
  instance.

Both APIs require a token from a user with staff access.

The background here is our ongoing attempt to integrate LAVA more
fully into a SSO environment, where OIDC only takes us so far in terms
of things like account locking and access by use of user tokens.

Signed-off-by: Ed Smith <ed.smith@collabora.com>

For example, if last try raised InfrastructureError the old code
would convert it to JobError which could have undesired side effects.
(like canceling health checks)

The new code will raise the same exception class as the last
exception caught during last try.

Add partial support for docker login in lava test actions

See merge request !115

lava_rest_app: Reworked API v0.3

See merge request !118

device-types: add hp-x360-14a-cb0001xx-zork

See merge request !122

The rework has following properties:

1). Add support for \r\n and \r line separators. Some software
uses \r as line separators and before that LAVA would simply
log entire output as one big line if the software used \r.
LAVA would also stall all output until it could read the new
line.

2). Make logs preserve unprintable characters such as new lines,
carriage return and ANSI terminal escape sequences. This allows
for better compatibility with other software that relies on those
characters such as Gitlab. The unprintable characters are stripped
from output before the logs are displayed in the LAVA's web UI.

Unit test will check various logger inputs with different new
line separators and fragmented inputs. (for example, if \r\n gets
split between the write calls)

Add device type template for the HP Chromebook x360 14a Chromebook.

MR upstream: https://git.lavasoftware.org/lava/lava/-/merge_requests/1898



Signed-off-by: Laura Nao <laura.nao@collabora.com>

This reverts commit 9ac1f101.

Run django migrations as the correct target user

See merge request !109

Parse YAML from environment variable settings

See merge request !108

Add login support to containers, but there's no single location to
update the scheme, the container mixin just plucks it out of the
standard parameters.

This uses the standard utilities, so is simply a matter
of updating the parameters used.

This action uses the standard utilities, so this amounts to passing
amended parameters to enable the necessary stanza in the schema, and
to parse it.

This does not use the standard utility, so support for invoking
`docker login` appropriately must be coded separately. This also
enables the necessary stanza in the schema.

This change means that all LAVA actions that rely on this utility to
interact with docker can add support for `docker login` by
initializing the utility object appropriately. Actions must separately
enable the schema support for the login stanza.

Most actions rely on the same utility function to describe the docker
parameters they take in their schema. This extends that function so
that, optionally, an extra stanza for login details is recognised. By
default, the stanza is not recognised, it must be opted into.

The reason this stanza is optional is that it will take time to
implement docker login support in all the necessary places. This
allows us to implement it for a subset of LAVA actions, whilst keeping
the schema checking accurate for actions which do not yet have login
support.

Add an XMLRPC API for setting the groups a user belongs to

See merge request !107

Running them as root leaves blank unwritable logs if they don't
already exist, as is the case for the docker image. It seemed safer to
run the commands as the correct user (to avoid future permission
problems) than to simply bughunt the files that are currently created
that have the wrong owners.

This means you can set values to something other than strings, for
example boolean, numeric and even dictionary settings. The failover
is so that invalid values are treated as they were before.

Lava uses XMLRPC for its system management endpoints, which is secured
via HTTP basic authentication. This adds `system.set_user_groups`
taking a string email address for the user, and an array of string
group names which will be the exact set of groups of which the user is
a member.

This endpoint requires a token from a user with staff permissions in
order to access it.

Add support for definining individual action timeouts in the action blocks

See merge request !116

Allow defining a timeout for an individual action in the block section.
This allows to override all the actions in a block at once, while still
being able to control single action timeouts within the same block.

Signed-off-by: Laura Nao <laura.nao@collabora.com>

Allow definining individual action timeouts in the block section, by
adding a "timeouts" section.

Signed-off-by: Laura Nao <laura.nao@collabora.com>

No longer is a copy paste of v0.2
Now is a very limited API that only has the new methods.
Previous API version was difficult to update and keep in sync with
v0.2

Fix remote code execution at /api/v0.2/devices/validate/ and
/api/v0.3-experimental/devices/validate/ endpoints.

Jinja2 templates created from a user input are open to code
injection. By switching to sandboxed jinja2 environment only
a restricted set of function calls will be available
when rendering a template.

This way "lava_upload_rules" are used in both "upload to obs" and "obs"
jobs.

Signed-off-by: Paweł Wieczorek <pawiecz@collabora.com>

Sjoerd Simons authored Nov 02, 2022 and Paweł Wieczorek committed Nov 09, 2022

Signed-off-by: Sjoerd Simons <sjoerd@collabora.com>

Igor Ponomarev authored Oct 03, 2022 and Paweł Wieczorek committed Nov 09, 2022

Equal treatment of newlines will fix issues with software
that uses carriage returns as newlines.
Allowing escape character \x1b will let Gitlab print colored logs

Laura Nao authored Jul 20, 2022 and Paweł Wieczorek committed Nov 09, 2022

Refine pattern used to match Depthcharge TFTP errors. This prevents
LAVA from wrongly matching similar errors in Coreboot's late phase
output.

MR upstream: https://git.lavasoftware.org/lava/lava/-/merge_requests/1783



Signed-off-by: Laura Nao <laura.nao@collabora.com>

Sjoerd Simons authored Jul 27, 2022 and Paweł Wieczorek committed Nov 09, 2022

python3-bpfcc will pull in libelf1 from backports as well; Which will
later conflict with libdw1 which the lava packages want to install,
making a mess.

Install libdw1 directly from backports as well to avoid this

Signed-off-by: Sjoerd Simons <sjoerd@collabora.com>