Enable testing of wireless stack using following essential CDRouter
tests:
wifi_1 Graceful wireless LAN client restart
wifi_2 Restart wireless LAN client without releasing address
wifi_3 Restart wireless LAN client without releasing or obtaining a new address
wifi_10 Verify connectivity using all wifi modes advertised by the DUT
wifi_11 Verify DUT prevents use of TKIP cipher when using HT/VHT/HE
wifi_20 Verify DUT's wifi beacons contain expected mode information
wifi_30 WiFi association stress test
wifi_40 WiFi SSID scan test
Run those 8 testcases, in a loop 2 times and repeat each test 2 times.
References: PCF-883
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit eb74a59a99925f040813d364ca19b8b91260ad6c)
So the upcoming changes into the wireless configuration are better
visible.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 95d961f05c40d30f0127f7bc2dbfef7b7a5dd95f)
Use latest labgrid with following fixes and additions:
* modbus: allow exclusive management of a single coil
* shelldriver: make run wrapper injecting more reliable
* shelldriver: make kernel silencing more reliable
* shelldriver: fix missing kernel silencing on passwordless consoles
* shelldriver: make post_login_settle_time generally available
* ubootdriver: allow handling of console output decoding errors
* ubootdriver: use codec and decodeerrors args for decoding
* smallubootdriver: use codec and decodeerrors args for decoding
* ubootdriver: add possibility to delay boot_command execution
* ubootdriver: add possibility to extend init commands timeout
Mainly doing that in a hope, that it should hopefully improve DUT
bringup at least for Turris Omnia and URX851 reference targets.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Recent miniupnpd versions shipped with prplos-next likely introduced
some changes which result in a lot of noisy logging output directly into
the console, mainly during the startup:
miniupnpd[5549]: Cannot get IP address for ext interface . Network is down
miniupnpd[5549]: HTTP listening on port 5000
miniupnpd[5549]: no HTTP IPv6 address, disabling IPv6
miniupnpd[5549]: Listening for NAT-PMP/PCP traffic on port 5351
miniupnpd[5549]: shutting down MiniUPnPd
miniupnpd[6161]: HTTP listening on port 5000
miniupnpd[6161]: no HTTP IPv6 address, disabling IPv6
miniupnpd[6161]: Listening for NAT-PMP/PCP traffic on port 5351
Those errors are expected and can’t be fixed, as in the time when
miniupnpd starts the network is not ready yet.
So lets silence those console log splats by moving the console logging
under `-d` debug flag, so basically restoring previous logging
experience.
References: PCF-924
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Update mxl_wlan_hostap_ng profile for below changes:
* bump to version updk_9.1.20.1
* enable Mxl vendor module package for wireless hardware manager
* remove prplmesh files from mxl-wlan-feed so that prplmesh is
enabled from feed-prpl
* disable existing wifi manager from mxl-wlan-feed so that pwhm
controls wifi interfaces, wifi config and manages the daemons
mxl_x86 profile bump to version updk_9.1.20.1
mxl_x86_osp_tb341 profile bump to version updk_9.1.20.1
mxl_x86_sec profile bump to version updk_9.1.20.1
Bump feed-prpl to bring in following changes:
* prplMesh: update libswla, libswlc and pwhm version
* prplMesh: platform: Fix prplMesh compilation failure for prplOS UGW platform
* prplMesh: platform: Enable pwhm for intel_x86 platform
Signed-off-by: Bhavesh Bharadiya <bnbharadiya@maxlinear.com>
libamxp: fix blocked SIGINT in child processes like dropbear PCF-926
Its currently not possible to stop foreground processes with CTRL+C in
dropbear as ssh-server via libamxp is blocking SIGINT signal delivery.
So lets fix it by backporting fix delivered in commit 7abf0cd1 ("Issue:
HOP-2538 [SSH][AMX] Processes spawned by dropbear instance managed SSH
Manager ignores SIGINT [fix]").
Various fixes making dots in netdev names usable PCF-885
netdev names containing dots like `eth0.1` are causing various issues,
so lets backport a patch which attempts to fix this issues:
* tr181-ppp: fix handling of dots in netdev names
* tr181-bridging: fix handling of dots in netdev names
* ethernet-manager: fix handling of dots in netdev names
* libamxo: backport fix for sending events with aliases containing a dot
Fixes: PCF-926, PCF-885
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit ff8b9ec650db241cc27da6808e041f5c103507f6)
It seems like current 90s timeout for LAN network readiness checking is
not long enough, likely due to some netifd issues connected to PCF-860,
so lets increase it to 180s and see how it goes.
References: PCF-860
Signed-off-by: Petr Štetiar <ynezz@true.cz>
- Generic: remove some WiFi object/parameter to the ubus-list test since
we move those objects/parameters under the number of Radio, AccessPoint
or SSID to be TR-181 compliant.
- NEC-WX3000HP: same as above
- Turris-omnia: adapt the output after wireless interface fix
- Glinet-b1300: adapt the output after wireless interface fix
Signed-off-by: Marouan MCHARFI <marouan.mcharfi_ext@softathome.com>
This reverts commit c8ffba119520d1e1036110c9feba3c81e2869e83 since there
is a plan, that prplMesh should be used with pwhm in all upcoming prplOS
releases.
Signed-off-by: Andrey Erokhin <a.erokhin@inango-systems.com>
There is no plan to use those as reference targets, so lets don't spend
time on those and remove them from the ci.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
In commit 5b84bd0aec29 ("profiles: remove outdated, unused or
incompatible profiles") I've removed tplinkC6 profile, but missed the
fact, that its being used for testing of builder container.
So lets fix it by replacing the obsolete tplinkC6 profile with x86_64
profile.
Fixes: 5b84bd0aec29 ("profiles: remove outdated, unused or incompatible profiles")
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Added secureboot features
- LED indication for boot stages
- LED indication for error scenarios
Signed-off-by: Ramulu K's avatarRamulu Komme <rkomme@maxlinear.com>
After addition of prplMesh+pwhm it seems, that rootfs partition
occupation has grown by 30MB and thus current 104M rootfs partition size
is too small for this default configuration and build ends with:
error: ext4_allocate_best_fit_partial: failed to allocate 38 blocks, out of space?
Squashfs reports following:
Filesystem size 28126.08 Kbytes (27.47 Mbytes)
27.50% of uncompressed filesystem size (102276.84 Kbytes)
so the filesystem is apparently at roughly 103MB, so fix it by
increasing the default rootfs partition size to 130MB which should
hopefully last for some time.
References: https://gitlab.com/prpl-foundation/prplos/prplos/-/jobs/3942299487
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Bring back debug profile as it seems to be actively used & thus
confirmed to be still working.
Signed-off-by: Andrey Erokhin <a.erokhin@inango-systems.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz> [commit facelift]
pwhm bump to version v2.23.2 that include:
- remove initial arbitrary timer for dm conf loading
- turris-omnia: random: pwhm failed to start
- make configurable pwhm coredump generation
- deauth connected stations when disabling host rad/vap
- fix disabled endpoint interface name and status on boot
- Add Operating Classes for United States and Japan
- split wld datamodel definition into multi files
- handle actions on vendor object hierarchy
- Add Operating Classes for United States and Japan
- add more options for IEEE80211v BTM request
- IEEE80211k capabilities per station
- Remove the set of bridge interfaces
- [pwhm] fix splitted defaults loading
- adding bssid,ssid,channels extra filters in the pwhm side
- Add disassociation event
- [pwhm] fix iphone device authentication issue in wpa3 modes
- fix cleanup of obsolete assocDevs with invalid objs
- cleanup wld: replace wld securityMode with swl type
- adding the bssid as an argument to the scan function(pwhm+Nl802) Also adjusted some functions names to meet our naming rules.
- Testing issue
- [pwhm] use default dir iso one default file [change]
- cleanup and add testing
- fix calling scan rpcs from datamodel
- use generic frame parser to get associated device capabilities
- add scan action support in generic plugin
- add apis for nl80211 ap scan actions
- RssiMonitoring issues fixed
- load datamodel extension definition and defaults
- disable config opt check
- Fix pwhm integration into feeds
- Fix rpc argument types
- [prplMesh WHM] wps connection credentials are empty and parsing issue
- [prplMesh WHM] endpoint ProfileReference && RadioReference setting issue
- [prplMesh WHM] endpoint ConnectionStatus dm change notification is not sent
libswlc bump to version v3.12.1 include:
- define 80211 management action wnm frame
- [prplMesh WHM] Exposing VHT,HT and HE radio capabilities through the datamodel.
- Add Operating Classes for United States and Japan
- test issue due to uninitialized memory
- add api to locate IEs in 80211 management frame
- Add Operating Classes for United States and Japan
- SecurityModesSupported not filled, impacting band steering in WPA2-WPA3 Add management frame subtype type
- add radStd to Mcs & checkMatches function
- add reproting detail element id
- add certificate validation
- add certificate validation
- add more options for IEEE80211v BTM request
- add RSSI from RCPI function
- add more options for IEEE80211v BTM request
- Fix testing dependencies
- Add MFP required constraint for OWE security mode
- add apis to check apMode type to deduce mfp mode
- add apis to convert mfp modes to/from string
- add apis to convert security modes mask to/from string
- add SWL prefix to avoid redefinition
- add api to count differences in two strings
- add apis to map wps config method IDs and names
- RRM request fields and report parsing
- add new swl fileUtils function
- improve management frame IEs parsing
- Add RRM Capabilties string flags
- Fix swlc debug when empty type array
- histogram can not compile.
- IEEE80211k capabilities per station
libswla bump to version v2.3.1 include:
- restore swla exec apis
- Fix variant printing
- Header files not exported
- add Tuple Type array map functions
Fixes: PCF-781
Signed-off-by: Marouan MCHARFI <marouan.mcharfi_ext@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz> [commit message]
Fixes random issue where rule for allowing DNS traffic from LAN would be
missing.
Fixes: PCF-864
Signed-off-by: Mouna Aloui <mouna.aloui_ext@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz> [commit message]
Contains following improvements and fixes:
* ip-manager: 15_migrate_netmask_to_cidr: make the migration more robust
* ip-manager: 15_migrate_netmask_to_cidr: fix shellcheck warnings
* ip-manager: 15_migrate_netmask_to_cidr: cleanup whitespace issues
It seems, that currently in some corner cases we can produce following
output:
option ipaddr '192.168.1.1 192.168.1.1/24/7'
Unfortunately we don't have exact source network configuration, so we
don't know what has caused this, but lets try to prevent it anyway.
References: PCF-656
Signed-off-by: Petr Štetiar <ynezz@true.cz>
This rule in FORWARD_class is expected according to the `lansubnet1`
default settings in tr181-qos_defaults.odl (tr181-qos v3.1.0-1).
Signed-off-by: Petr Štetiar <ynezz@true.cz>
We're currently facing some occasional test failures mainly on Turris
Omnia, so lets try to fix it by relaxing current probably tight 1s wait
time to 2s and thus give a system more time to propagate the requested
changes.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
It seems, that providing proper testbed/CDRouter configuration needs
more time, so until it's fixed properly, lets admit it and allow the
test to fail.
References: PCF-841
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Currently CDrouter TR069 tests are failing:
FAIL: Expected to find Device.LANConfigSecurity.ConfigPassword in Values list
FAIL: Expected to find Device.LANConfigSecurity. in Names list
FAIL: Expected to find Device.LANConfigSecurity.ConfigPassword in Names list
Those failures are caused by a recent TR069 updates, where the mapping
was done internally by the TR069 stack, but its not doing that mapping
anymore, so until we get the fix of the mapping in the proper TR181
component, we should simply provide it manually via ProxyManager.
Suggested-by: Vincent Harle <vincent.harle@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
* network: add libqosmodule (A helper library to ease the implementation of QoS modules)
* network: add mod-autosensing (SoC module wan-manager select the active wan-mode)
* network: add mod-fw-host (Host Module for Firewall)
* tr69: bump packages
- libtr69-engine: bump to v0.7.10
- tr069-manager: bump to v1.21.2
* netmodel: bump packages
- libnetmodel: bump to v1.5.10
- netmodel-ip: bump to v1.2.1
- netmodel: bump to v2.5.4
* network: bump packages
- libdhcpoptions: bump to v0.5.0
- libfwinterface: bump to v1.3.12
- libfwrules: bump to v2.0.1
- libimtp: bump to v1.6.0
- libqosnode: bump to v1.3.0
- mod-dmext: bump to v0.5.0
- mod-fw-amx: bump to v0.3.1
- mod-pcm-svc: bump to v0.3.5
- mod-qos-tc: bump to v2.3.0
- mod-vlan-uci: bump to v0.1.2
- mod-xpon-prpl: bump to v1.0.5
- deviceinfo-manager: bump to v2.5.1
- dhcpv4-manager: bump to v1.9.2
- dhcpv6s-manager: bump to v0.10.1
- ethernet-manager: bump to v1.9.0
- hosts-manager: bump to v0.1.16
- ip-manager: bump to v1.32.10
- moca-manager: bump to v0.3.0
- multisettings: bump to v0.7.2
- packet-interception: bump to v0.0.4
- pcm-manager: bump to v0.6.4
- routing-manager: bump to v1.10.2
- ssh-server: bump to v0.3.18
- tr181-bridging: bump to v2.0.2
- tr181-device: bump to v0.7.2
- tr181-dhcpv4client: bump to v1.9.6
- tr181-dhcpv6client: bump to v0.10.0
- tr181-dns: bump to v1.1.2
- tr181-dnssd: bump to v0.3.11
- tr181-dslite: bump to v0.5.6
- tr181-dynamicdns: bump to v0.1.9
- tr181-firewall: bump to v4.10.1
- tr181-httpaccess: bump to v0.2.7
- tr181-logical: bump to v0.3.2
- tr181-neighbordiscovery: bump to v0.6.3
- tr181-pcp: bump to v0.4.0
- tr181-ppp: bump to v0.9.5
- tr181-qos: bump to v3.1.0
- tr181-routeradvertisement: bump to v0.12.1
- tr181-upnp: bump to v0.2.2
- tr181-usermanagement: bump to v0.7.3
- tr181-xpon: bump to v1.0.5
- wan-autosensing: bump to v0.3.0
- wan-manager: bump to v0.15.0
At the same time remove the wan-autosensing package and replace it with mod-autosensing.
The mod-fw-amx package is introduced as well.
Signed-off-by: Matthias FRANCK <matthias.franck_ext@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz> [tests, version bump, commit message]
During the last WG meeting it was decided, that we're going to use
`autoconf` WAN mode in CDRouter tests, where CDRouter expects the DUT to
listen for a Router Advertisement on the WAN interface and to
self-assign an IPv6 address using a prefix learned during this process.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Bump feed-prpl to include following latest fixes:
* lua-amx: bump to v0.3.9
* odl-generator: fix prplMesh+pwhm by bumping to proj_prpl_M2-2022_v0.5.1 version again
In commit 918c179d755d ("network: bump packages") we've accidentally
downgraded package from fixed version v0.5.1 to broken version v0.5.0,
so lets fix it, by bumping the package version to v0.5.1 again.
Lua powered modules are now working, so lets enable them in the builds,
while at it add some simple test cases to QA those new Lua components.
Signed-off-by: Matthias FRANCK <matthias.franck_ext@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz> [pwhm fix and Lua test cases]
In commit 51042914df66 ("ci: cdrouter: fully enable IPv6 testing") I've
missed addition of a rule which prevents execution of those CI jobs
during Coverity weekly testing, so lets fix it now.
Fixes: 51042914df66 ("ci: cdrouter: fully enable IPv6 testing")
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit efbf57ca686d3af20d1f1d85463a070761785586)
In commit 51042914df66 ("ci: cdrouter: fully enable IPv6 testing") I've
missed addition of a rule which actually enables the autmatic IPv6
testing, so lets fix it now.
Fixes: 51042914df66 ("ci: cdrouter: fully enable IPv6 testing")
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 88421c2ad028af28d178adb26dbc5ecaa0a9afaf)
It seems, that this service is mostly not available when the test is
being run and since it's allowed, whitelisted service we can fix this
flapping test by simply excluding the 5000/tcp port from the nmap scan.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit de0337b9bac8bd8117e91e75489a753e1ba1b2e7)
PPPoE IPv6 WAN is currently not supported, so lets disable the test.
References: PCF-843
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 5156beae26cbd79ad67f5e8d0eec825915e5ca1d)
IPv6 tests were currently allowed to fail, but it seems, that most of
the remaining issues were fixed, thus lets remove `allow_failure`
temporary workaround and make any CDRouter IPv6 test failure hard CI error.
CDRouter-UPnP-IPv46 is the exception as IPv6 support wasn't in the scope
and we still need to sort out the mandatory test cases and remove the
rest etc.
References: PCF-767
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 51042914df667fe77f84cb2ab4679f504f63f8c2)
This CDRouter test seems to be failing on current 4.14.284 kernel,
likely to the kernel bug related to the gre IPv6 connection tracking. So
lets enable this test on latest 5.15 kernel in prplOS v3 and check how
it goes there.
References: PCF-743, PCF-842
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 520c2c3823cf5e1d45c43a43e9fd9a2bbd15454f)
Contains changes needed to make DSLite configuration actually working on
the DUT.
References: PCF-791
Signed-off-by: Vincent Harle <vincent.harle@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 9eb7cab05cf14c8d83702fa3daeeb46cc36c71c1)
cdrouter-3484 and cdrouter-3499 tests are checking if CPE is answering
ping6 on WAN and as this is our expected behavior, we need to set
cdrouter conf `ipv6WanPingRespond=yes` on various IPv6 test
configurations.
References: PCF-671
Suggested-by: Vincent Harle <vincent.harle@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 2c3881ad312d0089c545d85d2c503c3c07538863)
There are hundreds of the NTP servers behind those NTP pool hostnames
and it seems, that not all of them are configured to allow ping checks,
which leads to occasional CI test failures, so lets just remove those
unreliable checks.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 76d0026770d18c58d97256b3d3d60fb7bb075868)
It seems, that latest updates has changed behavior of SSH connections
during initial DUT bringup, so lets try to workaround those issues by
moving the initial sleep after the boot just before the initial SSH
connection. While at it, increase duration of initial DUT sleep after
boot in CDRouter CI test jobs to the same default value of 90s as used
for cram CI test jobs.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit b6776ee9aad7499ab3c87b891502bd1af0d703e9)
We're currently seeing a test run instabilities as the emitter and
receiver are handled by the same amxrt instance.
To apply a config change, multisetting currently restarts the receiver
plugin. In that case it will also restart the emitter plugin. This one
not being setup to have persitency, it would reload the default settings
and thus cause multisetting to react again and restart it with original
profile.
So lets fix it by splitting receiver and emitter into separate amxrt
instances.
References: PCF-583
Reviewed-by: Vincent Harle <vincent.harle@softathome.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 3efc0748cc17f3e57de09776c1c3601dd2abad52)
