The following error is seen:
ERROR: Nothing PROVIDES 'llvm-native' (but virtual:native:/builds/jonmason00/meta-arm/work/build/../poky/meta/recipes-graphics/mesa/mesa_23.1.3.bb DEPENDS on or otherwise requires it). Close matches:
ovmf-native
rust-llvm-native
vim-native
Work around this by using the llvm-native from meta-clang
Signed-off-by: Jon Mason <jon.mason@arm.com>
Upstream change 26d97acc71379ab6702fa54a23b6542a3f51779c changed the
IMAGE_LINK_NAME to have IMAGE_NAME_SUFFIX and breaks the automatic
finding of the image by name. Work around here until upstream fix can
be added.
Signed-off-by: Jon Mason <jon.mason@arm.com>
Remove IMAGE_NAME_SUFFIX from image path in
wic_nopt.bblass
Signed-off-by: Ziad Elhanafy <ziad.elhanafy@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
The fvp_devices test suite can be used to verify the following
functionality at runtime, common to most FVPs:
* CPU hotplug
* virtio-net device presence and functionality
* virtio-rng device presence and functionality
* PL031 RTC device presence and functionality
* SP805 watchdog device presence
The list of devices to be tested can be configured by a BSP using the
variable TEST_FVP_DEVICES.
Add this test suite for fvp-base and fvp-baser-aemv8r64.
Signed-off-by: Peter Hoyes <Peter.Hoyes@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
The linuxboot test case is already FVP-specific due to the use of the
OEFVPTarget pexpect interface. Clarify this by renaming to fvp_boot.
So that fvp_boot can be used alongside other OEQA test cases (e.g. those
in OE-core):
* Call self.target.transition("off") at the start of the test to
ensure the model starts from reset
* Call self.target.transition("linux") to reuse the "wait for boot"
logic in OEFVPTarget.
Additionally, minimally validate the firmware boot by checking for
common error patterns in all console logs. Expose the runfvp config in
OEFVPTarget to support this.
Align the list of test cases executed on both fvp-base and
fvp-baser-aemv8r64 by using TEST_CASES:append = " fvp_boot" for both.
Signed-off-by: Peter Hoyes <Peter.Hoyes@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
To better support firmware testing alongside Linux runtime testing,
introduce model state support to OEFVPTarget. The following states are
supported using self.target.transition(state):
* off
* on
* linux
Instead of assuming a specific state in OEFVPTarget.start,
responsibility is delegated to test cases to lazily put the model in
the required state. But to support OE-core test cases, OEFVPTarget.run
automatically puts the model in the "linux" state for running the
command. Firmware and Linux tests can subsequently run alongside each
other without introducing complex test dependencies.
The concept is inspired by Labgrid strategies [1], albeit simplified.
Tweak log file handling so that output is collected across (possibly)
multiple model processes.
[1] https://labgrid.readthedocs.io/en/latest/overview.html#strategies
Signed-off-by: Peter Hoyes <Peter.Hoyes@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
The differences between OEFVPTarget, OEFVPSSHTarget and
OEFVPSerialTarget are not obvious and there is a lot of duplication.
Merge all the logic into one OEFVPTarget (again). This has the following
features:
* Run SSH commands
* Run serial console assertions
* Lazily await a Linux login prompt while running test cases (only when
self.target.run is called).
Signed-off-by: Peter Hoyes <Peter.Hoyes@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Using runfvp without explicitly specifying the fvpconf path currently
fails due to a missing fvp.conffile include.
Signed-off-by: Peter Hoyes <Peter.Hoyes@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Update to the latest version of OP-TEE, and move the legacy versions
that are still in use by meta-arm-bsp to that directory
Signed-off-by: Jon Mason <jon.mason@arm.com>
By default, optee-test is using an ancient version of openssl (1.2.0o)
in binary form, located in the optee-test build tree. musl is already
working around this. So, use those defaults for everyone.
Signed-off-by: Jon Mason <jon.mason@arm.com>
scp-firmware passes -I/core/include to the compiler which doesn't exist,
and sometimes gcc emits a fatal error. It's unclear why this doesn't
happen for everyone, but this workaround appears to be the correct
solution.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
SCP_COMPILER isn't used anymore, so remove it.
Explicitly set SCP_TOOLCHAIN=GNU for clarity.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Now that this defaults to MACHINE, explicit SCP_PLATFORM assignments can
be removed.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Many platform names are the same as the machine name, so this saves
some typing.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
It is more difficult to update optee recipes to bleeding edge
3.21 when patches are added to SRC_URI via :append and thus they
need to be explicitly removed with :remove and name of the patch file.
For our boards we know 3.21 will work without patches but we still want
to keep using meta-arm side base recipe and just update the SRC_URI
to remove patches and update SRCREV and PV.
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Jon Mason <jon.mason@arm.com>
This tool makes it easy to lock a build to a known good configuration,
for example by locking the SHAs to the last good build of master.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
In the update-repos job, generate a Kas lock file to pin the SHAs of the
repositories being build during the build. This ensures that commits to
the repositories _during_ the build doesn't cause differing builds. All
of the kas calls use this lock file to ensure that their builds are
identical.
This lockfile is also added to the artifacts, so that it can be reused
afterwards to replicate the build: either as a known good base or to
replicate failures.
This lock file is only generated if it doesn't exist, so that
development branches can temporarily add a lockfile.yml if for example
master is too unstable to develop on.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Kas 3.3 uses branch and commit keys instead of a general-purpose
refspec. Change the base to use branch:mickledore and remove the
explicit use of master for meta-virtualization now that a mickledore
branch exists there too now.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Kas 3.3 has the following new features:
- kas: Introduce commit and branch as alternative to refspec key
- kas: Warn if a repo uses legacy refspec
- kas: add support for lock files via dump plugin
- kas: track root repo dir config files of menu plugin
- kas: add support for --log-level argument
- kas: add GIT_CREDENTIAL_USEHTTPPATH environment variable
- kas: improve error reporting
- kas: drop support for Python 3.5
- kas-container: fix invocations with --isar for some layers
- kas-container: Purge tmp* on clean
- kas-container: enable colored logging
4.0 is basically the same but the Kas container uses Debian 12, which is
not yet supported as a tested platform. Until Debian 12 is tested we
should continue to use 3.3.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
This will still emit the diagnostic but it wont break the build,
additionally pass it to CXXFLAGS since thats the right subset to
disable it for.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Same as for OP-TEE client, this eases debugging and is required to use
the OP-TEE symbolize.py script.
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Jon Mason <jon.mason@arm.com>
The TA builds always create ELF files, add them to the deploy dir to
ease debugging via the OP-TEE symbolize.py script.
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Jon Mason <jon.mason@arm.com>
The optee-os compilation provides not only the core files, TAs are also
usually build. Create a separate package which contains them, so they
can be installed.
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Jon Mason <jon.mason@arm.com>
TF-A version 2.8.6 hangs on boot for corstone1000 non-deterministically.
This sets TF-A version to v2.8.0 which is a tested working version for
corstone1000.
Signed-off-by: Emekcan Aras <emekcan.aras@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Adds missing update service definitions for using stateless platform
services and initializes the capsule udpate provider in se-proxy-sp
for corstone1000.
Signed-off-by: Emekcan Aras <emekcan.aras@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Adds missing compilation option to fix psa_raw_key_aggrement test for
corstone1000.
Signed-off-by: Emekcan Aras <emekcan.aras@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
From: Gowtham Suresh Kumar <gowtham.sureshkumar@arm.com>
These 2 patches causes the secure world to enter into an infinite loop
when the PSA arch tests are triggered. This is a temporary fix and the
issue needs to be investigated before the patches can be enabled.
Signed-off-by: Gowtham Suresh Kumar <gowtham.sureshkumar@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
This patch is required to handle one of the corner cases of the
GetNextVariableName EFI service as specified in the UEFI spec.
Signed-off-by: Gowtham Suresh Kumar <gowtham.sureshkumar@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
From: Gowtham Suresh Kumar <gowtham.sureshkumar@arm.com>
This patch adds the required configs to the corstone1000 u-boot
defconfig to enable the EFI services. This is done to fix the SCT
failure reported by the SetTime_Conf and SetTime_Func.
Signed-off-by: Gowtham Suresh Kumar <gowtham.sureshkumar@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
From: Gowtham Suresh Kumar <gowtham.sureshkumar@arm.com>
The patch fixes the ACS InstallMultipleProtocolInterfaces_Conf failures
in corstone1000 platform by dropping a workaround u-boot patch. The NVMXIP
initialization had some issues during u-boot boot stage which led to the
workaround patch.
Upstream-Status: Pending [Not submitted to upstream yet]
Signed-off-by: Emekcan Aras <emekcan.aras@arm.com>
Signed-off-by: Gowtham Suresh Kumar <gowtham.sureshkumar@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
It's useful to have known good logs when debugging problems, so always
preserve the logs.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
The previous commit fixed the build of GN with GCC 13, but broke it for
any other version.
Remove the patch and simply disable the fatal warning that causes the
breakage. Interestingly, this warning is already disabled for Windows
builds.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Jon Mason <jon.mason@arm.com>
EXTERNAL_TOOLCHAIN variable should provide absolute path to external Arm
toolchain install directory. So make that absolute path check explicit.
Signed-off-by: Sumit Garg <sumit.garg@linaro.org>
Signed-off-by: Jon Mason <jon.mason@arm.com>
Jon Mason
Ziad Elhanafy
Ross Burton
Peter Hoyes
Mikko Rapeli
Abdellatif El Khlifi
Khem Raj
Rouven Czerwinski
Tomás González
Emekcan Aras
Sumit Garg