This will produce a more useful hint for those setting
up or testing virgl headless:
runqemu - ERROR - No render nodes found in /dev/dri: ['by-path', 'card0']. If /dev/dri/renderD* is absent due to lack of suitable GPU, 'modprobe vgem' will create one suitable for mesa llvmpipe software renderer.
as qemu itself isn't helpful:
alex@Zen2:/srv/storage/alex/yocto/build-64-alt$ qemu-system-x86_64 -display egl-headless
qemu-system-x86_64: egl: no drm render node available
qemu-system-x86_64: egl: render node init failed
(From OE-Core rev: cbbada6a6c9b0a2e97f7395117dad986555f2db9)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
runqemu (which the test also runs) already performs this check,
and does it better, as it additionally unsets a number of environment
variables that can confuse and misdirect host's pkg-config.
(From OE-Core rev: ad179728a6978d929e2a876abe7e1cbd6b76a502)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Add some tests for new yocto_testresults_query.py helper. First test is taken
from yocto-autobuilder-helper feature which has moved in yocto_testresults_query
(From OE-Core rev: 9ce28f685e4282d81f179877cbafd0a52fa887bd)
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Many stored results TEST_TYPE are set to "oeselftest", however all those
tests are not run with the same sets of parameters, so those tests results may
not be comparable.
Attach relevant parameters as tests metadata to allow identifying tests
configuration so we can compare tests only when they are run with the same
parameters.
(From OE-Core rev: 406ec0a87e80d1ee0dd1d246adfe9507c10450ff)
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We keep seeing this failure on the autobuilder but the output amounts
to "False is not True". Improve the debug message on the chance it may
make the issue clearer.
(From OE-Core rev: d03f4cf19c2cc96e9d942252a451521dfec42ebc)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When qemu machine hangs, the ssh commands done by tests
are not timing out. do_testimage() task has last logs like this:
DEBUG: time: 1673531086.3155053, endtime: 1673531686.315502
The test process is stuck for hours, or for ever if the
executing command or test case did not set a timeout correctly.
The default 300 second timeout is not working when target hangs.
Note that timeout is really a "inactive timeout" since data returned
by the process will reset the timeout.
Make the process stdout non-blocking so read() will always return
right away using os.set_blocking() available in python 3.5 and later.
Then change from python codec reader to plain read() and make
the ssh subprocess stdout non-blocking. Even with select()
making sure the file had input to be read, the codec reader was
trying to find more stuff and blocking for ever when process hangs.
While at it, add a small timeout to read data in larger chunks if
possible. This avoids reading data one or few characters at a time
and makes the debug logs more readable.
close() the stdout file in all cases after read loop is complete.
Then make sure to wait or kill the ssh subprocess in all cases.
Just reading the output stream and receiving EOF there does not mean
that the process exited, and wait() needs a timeout if the process
is hanging. In the end kill the process and return the return value
and captured output utf-8 encoded, just like before these changes.
This fixes ssh run() related deadlocks when a qemu target hangs
completely.
(From OE-Core rev: 9c63970fce3a3d6029745252a6ec2bf9b9da862d)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Read from serial console with a small delay to bundle data to e.g.
full lines. Reading one character at a time is not needed and causes
busy looping.
(From OE-Core rev: 0049f6757f6f956fb4cc77b3df6a672c20b53cf4)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When a qemu machine hangs, the QMP calls can hang for ever
too, and when this happens any failing test commands from ssh
runner may be followed by dump_monitor() calls which
then also hang. Hangs followed by hangs.
Use runqemutime at setup and run_monitor() specific timeout
for later calls.
(From OE-Core rev: 3a07bdf77dc6ecbf4c620b051dd032abaaf1e4ff)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Use a short sleep to bundle serial console reads so that
we are not reading one character at a time which reduces busy
looping.
(From OE-Core rev: cafe65d8cf7544edbd387f7f5f6d77c64c6b18fa)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If test target qemu machine hangs completely, dump_target() calls
over serial console are taking a long time to time out, possibly
for every failing ssh command execution and a lot of test cases,
and same with dump_monitor().
Instead of trying for ever, count errors and after 5 stop trying
to dump_target() and dump_monitor() completely.
These help to end testing earlier when a test target is completely
deadlocked and all ssh, serial and QMP communication with it are
failing.
(From OE-Core rev: d9ad0a055abba983c6cee1dca4d2f0a8a3c48782)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Configure ssh client to test that connection with server is up.
If the server does not respond within a minute then the connection,
target machine or sshd daemon are stuck and it's better to exit
the command execution with errors.
Some tests can execute a long time without returning stdout/stderror
data and it's difficult to adjust timers for those cases if
connection to target machine or the target machine itself hangs
and output is not expected in minutes or even hours.
(From OE-Core rev: ba68ff04c5786eca7cd8dd44056705867dea8ac4)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The output from is garbled otherwise and it's not
easy to remove debug output form real command output on target.
(From OE-Core rev: 917a70cbc43ac1c70c477b220c4115735457ef04)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Providing ssh port number is supported too with
"--target-ip 192.168.0.10:22".
(From OE-Core rev: 637919b9df0abc06da5b2f9b389cf25376bd6b7c)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Similarly to the recently added test for binary generated locales, add
a version to test on target locale generation. This was broken but should
be fixed now so we can add the test sharing code from the previous test.
(From OE-Core rev: 50a29c167eb9fe9fa96aa53a379ae7597cefd1cc)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
[YOCTO #9070]
Add a new selftest to validate locale generation. This selftest builds a
complete target with GLIBC_GENERATE_LOCALES, IMAGE_LINGUAS,
ENABLE_BINARY_LOCALE_GENERATION set.
(From OE-Core rev: 53258fd810bea6475af9f908f7b712a13a02b628)
Signed-off-by: Louis Rannou <lrannou@baylibre.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Call logger.info() in the log_message handler so that we get request
logging, and hopefully even error messages.
Create a child logger to be neat and compartmentalise the logging.
Add a __main__ entrypoint so this class can be exercised outside of oeqa.
Remove unused traceback import.
(From OE-Core rev: 8fc939d942cec1662ca386b87c63f167bb10fd05)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Pass our logger to the HTTPService instance so we can see the requests
and any errors.
(From OE-Core rev: 3485df2ff61143aac03d92300b7bac4e5d6b2427)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This does not actually guarantee that the child runqemu process has completely exited:
poll() may return prematurely while the SIGTERM handler in runqemu is still running.
This thwarts the rest of the processing, and may terminate the handler before
it completes.
Use Popen.communicate() instead: this is what python documentation recommends as well:
https://docs.python.org/3/library/subprocess.html#subprocess.Popen.communicate
(From OE-Core rev: cd3e55606c427287f37585c5d7cde936471e52f4)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In recent kernel versions the string "Linux" moved to a header,
'include/linux/uts.h' instead of init/version.c. Allow the test
to work with both situations.
(From OE-Core rev: c15c59c88d229e35eeac1ed948c84168633e7cb1)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The check needs to report dri location on the host machine,
so pkg-config binary needs to be capable of finding the needed
dri.pc file on the host, and therefore needs to know where
host .pc files are located.
This may not be the case when using pkg-config from buildtools,
so this forces usage of host pkg-config.
runqemu already does the same PATH tweak, so this simply brings
the two in sync.
(From OE-Core rev: f0521f8a3ba7e15482756529ee7b0a95b3d53e7d)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Primarily, before running the debuginfod-find tool, check that the
debuginfod server has finished sweeping the deploy directory. If we
make the request too soon then there's a rare chance that we run the
client before it has scanned the right packages, and the log gets
swamped with warnings from sqlite due to a race.
Also:
- unset DEBUGINFOD_URLS so the debuginfod doesn't proxy to an upstream
server provided by the host distro
- Lower concurrency to reduce system load and handle systems with lower
maximum open file counts but lots of cores (as the concurrency means
cores*2*2 open files)
- Set the refresh times to 0 so we never rescan during the test
- Only scan the packages for the format which the image is using
- Log the commands that are being invoked
(From OE-Core rev: d65729748253eaa640333198ca8aec05946cb9e8)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
[Yocto #14346]
Systemd may be slow in killing pam session sometimes [1][2]. It may cause rpm
test to fail because there's process (sd_pam) running and own by "test1" user
after timeout.
Increasing timeout to 2 mins and assert earlier with debug output if
there's such process(es). If increasing of timeout doesn't help we may
want to force deletion of the user as [2] suggests.
[1] https://github.com/systemd/systemd/issues/8598
[2] https://access.redhat.com/solutions/6969188
(From OE-Core rev: 972fcc0ed1e0d36c3470071a9c667c5327c1ef78)
Signed-off-by: Pavel Zhukov <pavel@zhukoff.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For now, there's no easy way to skip some specified tests when
using 'oe-selftest -r'. This is because '-R' and '-r' are in the
mutually exclusive group. Looking at the codes, the codes have already
allowed running specified tests and skipping specified tests at
the same time. So move '-R' out of the mutually exclusive group
and change its help message.
After this change, the following command could be used.
oe-selftest -r A -R A.B.C
This does introduce a behavior regression.
Previous 'oe-selftest -R xxx' needs to now be changed to
'oe-selftest -a -R xxx'.
(From OE-Core rev: df9a2b69748d8a24c3390f812225231e9e9acb66)
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When running oe-selftest and seeing the end of a running log, it is
extremely helpful to know if there have been failures or not to save
looking at the rest of the log. Add the number of failures to the summary
line so that people monitoring builds have an easier time before the end
totals are printed.
(From OE-Core rev: 6b23996911d91f7f99774646c6db9f3490b4cb62)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It seems some layers want to subvert the intent of LAYERSERIES_COMPAT
so bitbake is going to have to become stricter about the values there.
To work with this, use LAYERSERIES_CORENAMES to generate the entries in
LAYERSERIES_COMPAT instead of the current magic LAYERSERIES_COMPAT_core
value which may not continue to work.
The downside to this is when migating between releases, people would
need to update devtool workspace layer.conf files. I guess you could
argue this is a feature!
(From OE-Core rev: 96ff9baa8ead57504f40f362ed3a4aaa776d1b58)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We've seen two different regressions in this API since it is used by
layer-index but not be the core code. Add a test for it to try and
ensure we don't break it again.
(From OE-Core rev: b07de5de43ec9c9a2c5d496a64940ccdc5b47cf8)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
With the previous fixes the test can be run again,
and it doesn't need all those extra steps. Runqemu
takes care of everything automatically now.
(From OE-Core rev: 99083d01bd576eee7c2c569c66042d064c193a9a)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is the first release in 13 years;
I have reviewed the accumulated patches, and dropped some of them
where purpose or issue being fixed is not clear. Specifically:
0001-Add-listen-action-for-a-tcp-socket.patch
0001-daemon.c-Libtirpc-porting-fixes.patch
fixed upstream in
84ab475f93
0001-attr-fix-utime-for-symlink.patch
addresses an open issue in
https://github.com/unfs3/unfs3/issues/4
please rebase and re-submit as a PR if the problem is still present.
alternate_rpc_ports.patch
unnecessary as of
https://git.yoctoproject.org/poky/commit/?id=6bb9860ef7ba9c84fe9bd3a81aa6555f67ebd38e
Command line options introduced by the patch no longer used anywhere.
fix_compile_warning.patch
merged upstream.
fix_pid_race_parent_writes_child_pid.patch
rebased and re-submitted upstream.
no-yywrap.patch
dropped as backport.
relative_max_socket_path_len.patch
needs to be re-submitted by the original author, purpose and reproducer scenario unclear.
rename_fh_cache.patch
merged upstream.
tcp_no_delay.patch
purpose and use case for oe unclear.
unfs3_parallel_build.patch
fixed upstream in
987d32ca12a39a78995c
Drop -N option from oeqa nfs helper and runqemu helper;
the option was provided by tcp_no_delay.patch
and is not needed for the tests or qemu.
Drop ad hoc libtirpc support; upstream supports it directly now.
Drop the check for portmap/rpcbind, it is unnecessary as of
https://git.yoctoproject.org/poky/commit/?id=6bb9860ef7ba9c84fe9bd3a81aa6555f67ebd38e
(From OE-Core rev: fa2f7cf545137b071db97015bca5b70d77566cd8)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Qemurunner should not pass rootfs to runqemu in case
rootfs is not a filesystem itself. Some images could
be built into some disk format like vmdk and this
commit makes qemurunner handle such images properly.
(From OE-Core rev: 5aea74046cf4c1aa7fa9e2402788d662268ccf53)
Signed-off-by: Sergei Zhmylev <s.zhmylev@yadro.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
By default host side IP address is not set and qemu listens
on all IP addresses on the host machine which is not a good
idea when images have root login enabled without password.
It make sense to listen only on localhost IP address 127.0.0.1 using
config change like:
QB_SLIRP_OPT = "-netdev user,id=net0,hostfwd=tcp:127.0.0.1:2222-:22"
This config works for qemu itself, but breaks runqemu which tries to
parse the host side port number from qemu process command line arguments.
So change the runqemu side hostfwd parsing for port number to ignore
the host IP address field.
Reviewed-by: Quentin Schulz <foss+yocto@0leil.net>
(From OE-Core rev: bdbd52082eb26f418000eb4e424baae9babc272c)
Signed-off-by: Mikko Rapeli <mikko.rapeli@linaro.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
test_rawcopy_plugin_qemu and test_wic_image_type are x86-specific
currently as the .wks uses x86-specific bootloaders.
This can be fixed, but that can come later.
(From OE-Core rev: 93525809a1ecb01ae7218558c0d6c1b0344606c5)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
These tests should be able to work on qemuarm64, but this is untested
and the runners will need configuration.
(From OE-Core rev: 09b9558e20e58b473154895b93cff16261c7f561)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Don't hardcode qemux86-64. This has some complications: the
IMAGE_FSTYPES needs to be constructed to reflect what the machine can
do (only x86 machines can build ISO images), and several tests which need
a wic file are currently limited to qemux86-64.
(From OE-Core rev: a30680a869ff3be63d26468f6365751c56bbb006)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This test was failing very oddly in qemuarm64 runs. Rewriting the test
to be clearer and less fragile fixed it.
(From OE-Core rev: a26fc7c2119df12468b0a834de6fe67aa9c86085)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When the trivial test binary, which just calls printf(), is compiled for
aarch64 with -O2 -D_FORTIFY=2 (as is the default configuration), gdb
resolves main() to the inlined printf() wrapper in stdio2.h instead of
main.c, so the test fails.
Presumably, this is due to debugging being unreliable with -O2. Solve
this problem by not caring where the main() breakpoint resolves to, just
check that it was resolved at all.
(From OE-Core rev: c51c12154851d04a81c8fbe190e712b3cd8dc941)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For now, not all machines have a default wks file, so mark the tests
which need a wks as being specific to qemux86-64.
(From OE-Core rev: 152f1a8cdf698b71c956e9910911dcb141a1f5be)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Set WKS_FILE to wictestdisk.wks, which is a very simple Wic file that
simply contains the root filesystem. It may not actually boot but this
test doesn't care, and it does exercise the wic image construction on
all machines.
(From OE-Core rev: b66a94896193f8d8eeff43b66e9daeb9a74bfed9)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There's no need to build a wic image in this test as not all machines
(such as qemuarm64 currently) have wic images out of the box.
We can simply build ext4 images to work on more machines and save some time.
(From OE-Core rev: cdcf858d00eaf54814e23f550f83f3646bf83a24)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
There's no need for this to be built for genericx86-64, we want to test
the current MACHINE.
(From OE-Core rev: b5a7ebe9627b28b207ccccba4f26c6d4a937d6a8)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This test was overly complex with a setUp() method for a single test case,
which was marked as a class function for no good reason.
Generalise the test so that it has the possibility of working on more
machines in the future, add a decorator so that it only runs on
qemux86-64, and respect QEMU_USE_KVM to speed up test execution.
(From OE-Core rev: 2bc2ee171f976807053b7da44c1eedbb07c10949)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
qemuarm64 currently uses SERIAL_CONSOLES_CHECK but this is incompatible
with read-only rootfs under sysvinit. Until qemuarm64 doesn't use this,
skip the test on qemuarm64.
(From OE-Core rev: c46be833efc5abea577251ef7e87ef90f08c4de3)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
