Wip/detect boot #2
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Codecov ReportAttention: Patch coverage is
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## master #2 +/- ##
=========================================
- Coverage 63.2% 53.2% -10.0%
=========================================
Files 152 163 +11
Lines 11330 12037 +707
=========================================
- Hits 7169 6414 -755
- Misses 4161 5623 +1462 ☔ View full report in Codecov by Sentry. |
…rol/disk paths
When a udev event is retrieved from the UdevManager's queue, it is
matched against each resource. If a match is found, the resource's
device is updated (on "add"/"change"/"move" events) or removed (on
"unbind"/"remove" events).
Normally the `avail` attribute is set afterwards, but not for the
USBSDMuxDevice and USBSDWireDevice: those devices prevent their
USBResource super class from setting the `avail` attribute. Instead, the
devices are available depending on the availability of their disk and
control paths. Then the device's `update()` method is called, which is a
no-op for those devices.
On `poll()`, their disk and control paths is set to None if their device
is None. Otherwise, these paths are set if the device is not already
available.
In case of a USB reset (or fast replug) of a big USB tree, loads of udev
events need to be processed. Cases were observed where `poll()` did not
run between processing of a "remove" and "add" event of the same device.
This leads to a race condition, following the description from above:
The USBSDMuxDevice's or USBSDWireDevice's underlying device is set to
None on "remove" and then set to the new device again, on "add". The
`avail` attribute is not updated as intended. After that, the `poll()`
method runs: the device is set and the `avail` attribute evaluates to
True because the disk and control paths are still set from the time
before the reset replug. If this situation happens, these paths stay
invalid until the next reset/replug or restart of the labgrid exporter.
These invalid paths might point to non-existent /dev/sg* and dev/sd*
devices, rendering interaction with the Mux and its SD card impossible.
Or, even worse, the paths point to valid /dev/sg* and dev/sd* devices of
another Mux connected as observed here (note different USB devnums):
$ labgrid-client -vv resources rlab/usb-2-p3/NetworkUSBSDMuxDevice
Exporter 'rlab':
Group 'usb-2-p3' (rlab/usb-2-p3/*):
Resource 'USBSDMuxDevice' (rlab/usb-2-p3/NetworkUSBSDMuxDevice[/USBSDMuxDevice]):
{'acquired': 'board1',
'avail': True,
'cls': 'NetworkUSBSDMuxDevice',
'params': {'busnum': 2,
'control_path': '/dev/sg6',
'devnum': 42,
'extra': {'proxy': '[...]',
'proxy_required': False},
'host': 'rlab',
'model_id': 16449,
'path': '/dev/sdg',
'vendor_id': 1060}}
$ labgrid-client -vv r rlab/usb-1-p4/NetworkUSBSDMuxDevice
Exporter 'rlab':
Group 'usb-1-p4' (rlabC-srv/c-usb-1-p4/*):
Resource 'USBSDMuxDevice' (rlab/usb-1-p4/NetworkUSBSDMuxDevice[/USBSDMuxDevice]):
{'acquired': 'board2',
'avail': True,
'cls': 'NetworkUSBSDMuxDevice',
'params': {'busnum': 2,
'control_path': '/dev/sg6',
'devnum': 43,
'extra': {'proxy': '[...]',
'proxy_required': False},
'host': 'rlab',
'model_id': 16449,
'path': '/dev/sdg',
'vendor_id': 1060}}
Fix this race condition by implementing the disk and control paths reset
in the USBSDMuxDevice's or USBSDWireDevice's update() method, in case the
underlying device is gone. The update() method is called during the udev
event processing, so it does no longer matter when the poll() method is
called.
Fixes: 27ff02f ("udev: correct availability for USBSDMuxDevice")
Fixes: 9324d32 ("Add sdwire driver")
Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Nicolas Labriet <[email protected]>
…p-update pyproject.toml: move pysnmp -> pysnmp-lextudio, drop unnecessary pysnmp-mibs
…mux-race udev/resource: fix mux device race condition leading to outdated control/disk paths
Starting with version 6.1.0, the QEMU command line argument to enable virtio with VirGL support was changed from "-vga virtio" to "-device virtio-vgal-gl". To correctly handle this, scrape the QEMU version number from the command output and use it to pass the correct arguments. Signed-off-by: Joshua Watt <[email protected]>
Signed-off-by: Jan Vermaete <[email protected]>
Signed-off-by: Jan Vermaete <[email protected]>
doc: do not open if already closed -> open
Exposes the place tags in the RemotePlace. This allows persistent data to be stored in the tags and retrieved by tests, for example unique MAC address that need to be programmed to a device after flashing Signed-off-by: Joshua Watt <[email protected]>
The option to use sftp was added multiple times when using SSHDriver::put() and SSHDriver::get() causing them to only work one time. Also adds the sftp option to SSHDriver::scp(). Signed-off-by: Nikolaj Rahbek <[email protected]>
No functional changes, a following commit will ensure that logging is only configured when the logging plugin is available. Signed-off-by: Rouven Czerwinski <[email protected]>
Check if the logging plugin is available and only configure logging if it is. Fixes an internal pytest error during loading when the logging plugin is explicitly disabled via "-p no:logging". Signed-off-by: Rouven Czerwinski <[email protected]>
In case the logging plugin is not loaded, the option "log_cli_level" is
never registered and getoption fails. Catch the exception and return
from the function since there is nothing for us to configure in this
case. Fixes the following traceback:
Traceback (most recent call last):
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/_pytest/config/__init__.py", line 1542, in getoption
val = getattr(self.option, name)
AttributeError: 'Namespace' object has no attribute 'log_cli_level'
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/home/phoenix/work/ptx/labgrid/venv/bin/pytest", line 8, in <module>
sys.exit(console_main())
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/_pytest/config/__init__.py", line 190, in console_main
code = main()
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/_pytest/config/__init__.py", line 167, in main
ret: Union[ExitCode, int] = config.hook.pytest_cmdline_main(
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/pluggy/_hooks.py", line 265, in __call__
return self._hookexec(self.name, self.get_hookimpls(), kwargs, firstresult)
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/pluggy/_manager.py", line 80, in _hookexec
return self._inner_hookexec(hook_name, methods, kwargs, firstresult)
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/pluggy/_callers.py", line 60, in _multicall
return outcome.get_result()
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/pluggy/_result.py", line 60, in get_result
raise ex[1].with_traceback(ex[2])
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/pluggy/_callers.py", line 39, in _multicall
res = hook_impl.function(*args)
File "/home/phoenix/work/ptx/labgrid/labgrid/pytestplugin/hooks.py", line 32, in pytest_cmdline_main
set_cli_log_level(logging.DEBUG)
File "/home/phoenix/work/ptx/labgrid/labgrid/pytestplugin/hooks.py", line 16, in set_cli_log_level
current_level = config.getoption("log_cli_level") or config.getini("log_cli_level")
File "/home/phoenix/work/ptx/labgrid/venv/lib/python3.10/site-packages/_pytest/config/__init__.py", line 1553, in getoption
raise ValueError(f"no option named {name!r}") from e
ValueError: no option named 'log_cli_level'
Signed-off-by: Rouven Czerwinski <[email protected]>
Add a test which uses the pytestplugin with -p no:logging set. Signed-off-by: Rouven Czerwinski <[email protected]>
Fix handling disabled pytest logging plugin
fix sftp option for scp
…resource Signed-off-by: Andreas Martinsson <[email protected]> Signed-off-by: Jan Luebbe <[email protected]>
…bugger Add SEGGER J-Link vendor and model IDs to USBDebugger resource
remote: Expose tags in RemotePlace
qemudriver: Handle Virtio GL change in 6.1.0
Signed-off-by: James Puderer <[email protected]> Signed-off-by: Jan Luebbe <[email protected]>
This avoids the overhead on startup. Signed-off-by: Jan Luebbe <[email protected]>
sync-places: add named match support
When using `power` command, one can use `-n` to choose to which resource on a target the command will apply. However, when looping over resources to find a compatible driver, it ignores the named resource. This will cause an exception if another resource is chosen due another driver being selected on the loop. This patch fixes it by ignoring resources whose name do not match the resource given in the command line. Signed-off-by: Ederson de Souza <[email protected]>
A new NetworkYKUSHPowerPort is created to represent a remote ykush hub. In order to control it via ssh, YKUSHPowerDriver now uses the `ykushcmd` command instead of python module - it must be installed on the host where the ykush is connected. One can get it from https://github.com/Yepkit/ykush A YKUSHPowerPortExport is also created to allow automatic exporting the remote resource from a local one. Signed-off-by: Chen Peng1 <[email protected]> Signed-off-by: Ederson de Souza <[email protected]>
So that targets which expose YKUSHPowerPort instances can be controlled via the command line client. Signed-off-by: Ederson de Souza <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
Other network resourced do not list the drivers using them, neither. Signed-off-by: Bastian Krause <[email protected]>
Initially, all driver bindings were documented with their respective binding key. This wasn't documented consistently since then. Especially when using multiple instances of a single driver class, the bindings are needed to specify on which resource or other driver the driver should bind on. Until now, most binding keys had to be looked up in the source code. So improve this situation by fixing and adding binding keys as well as also adding all Remote/Network resource variants a driver is able to bind to. Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
Explicitly quote string top prevent strange yaml string auto dection issues. Also use better readable dict notation everywhere. Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
…yaml snippet In the previous form, it looks like the environemnt's "drivers" section also contains the "images" section, which is not the case. Make that clear by splitting both sections into their own snippet. Signed-off-by: Bastian Krause <[email protected]>
…busRTUDriver Signed-off-by: Bastian Krause <[email protected]>
Signed-off-by: Bastian Krause <[email protected]>
…tions-20240213 github/workflows: update GitHub actions to latest versions
Add USB ID for i.MX93 SoC to IMXUSBLoader. Signed-off-by: Christian Hemp <[email protected]>
resource: udev: add new USB ID for IMXUSBLoader
The new major release 6.0 introduces breaking changes (functions are now async), limit the dependency until we fix this properly. Signed-off-by: Rouven Czerwinski <[email protected]>
pyproject.yaml: limit pysnmp to versions < 6
Decoding via VA-API fails for the C920 model and the playbin element on AMD graphic cards, playbin3 has no problems though - so lets switch to that. Signed-off-by: Stefan Kerkmann <[email protected]>
…e-decoding driver/usbvideodriver: use playbin3 element
…pdate doc/configuration: update/fix bindings, protocols and resource usage
A USB serial device often provides multiple interfaces for the same
serial number. In these cases it is not possible to distinguish
the interfaces solely based on the serial number (ID_SERIAL_SHORT).
However, if additionally the interface number (ID_USB_INTERFACE_NUM)
is provided, it is possible to distinguish all interfaces without
the need of using the full device path (ID_PATH).
Example:
(...)
=== suggested matches ===
(...)
USBSerialPort:
match:
ID_SERIAL_SHORT: ABCDEF00001
ID_USB_INTERFACE_NUM: '00'
(...)
=== suggested matches ===
(...)
USBSerialPort:
match:
ID_SERIAL_SHORT: ABCDEF00001
ID_USB_INTERFACE_NUM: '01'
Signed-off-by: Joerg Hofrichter <[email protected]>
Signed-off-by: Joerg Hofrichter <[email protected]>
…t-interface-num resource: udev: also suggest ID_USB_INTERFACE_NUM for USBResource
jremmet
force-pushed
the
WIP/detect_boot
branch
2 times, most recently
from
16bbdcd to
23c8e98
Compare
Sets self.boot_detected in _await_prompt. This is used in reset to check if a reboot has occurred. Can also be used in strategies. Signed-off-by: Jan Remmet <[email protected]>
Sets self.boot_detected in _await_prompt. This is used in reset to check if a reboot has occurred. Can also be used in strategies. Direct checking was deprecated with: 7b55a19 ("driver/ubootdriver: deprecate boot_expression attribute") Signed-off-by: Jan Remmet <[email protected]>
jremmet
force-pushed
the
WIP/detect_boot
branch
from
23c8e98 to
226b3eb
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Checklist