Age | Commit message (Collapse) | Author |
|
The target_dumper code is basically broken. It has been reading binary files
over the text base serial communication and runs at every command failure which
makes no sense. Each run might overwrite files from the previous run and the
output appears corrupted due to confusion from the binary data.
For now, remove the commands and the target dumper code as the command
and execution point are problematic. Also remove the same pieces of the monitor
code but leave the command list since in theory this can be moved to a more
useful place in the code.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The intent behind these functions was to dump the system state when issues occured
but it has never really worked as we'd planned. Regular monitoring as the build
runs has largely replaced this as that allows a trend to be seen rather than a spot
value which was never really useful. The code is bitrotting and not functioning
correctly so drop it.
[YOCTO #13872]
RP: Reword commit message
Signed-off-by: Thomas Roos <throos@amazon.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Enrico Jorns <ejo@pengutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Enrico Jorns <ejo@pengutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This is already set above, thus simply use the set value from there.
Signed-off-by: Enrico Jorns <ejo@pengutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Enrico Jorns <ejo@pengutronix.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Need to ensure that the dump_dir is created correctly and available
When command arguemnts are passed construct a filename if needed and
convert the arguements to a json object to pass to QMP.
Signed-off-by: Saul Wold <saul.wold@windriver.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This adds support for the Qemu Machine Protocol [0] extending
the current dump process for Host and Target. The commands are
added in the testimage.bbclass.
Currently, we setup qemu to stall until qmp gets connected and
sends the initialization and continue commands, this works
correctly. If the UNIX Socket does not exist, we wait an timeout
to ensure to socket file is created.
With this version, the monitor_dumper is created in OEQemuTarget
but then set in OESSHTarget as that's where we get the SSH failure
happens. Python's @property is used to create a setter/getter type
of setup in OESSHTarget to get overridden by OEQemuTarget.
By default the data is currently dumped to files for each command in
TMPDIR/log/runtime-hostdump/<date>_qmp/unknown_<seq>_qemu_monitor as
this is the naming convenstion in the dump.py code.
We use the qmp.py from qemu, which needs to get installed in the
recipe-sysroot-native of the target image.
[0] https://github.com/qemu/qemu/blob/master/docs/interop/qmp-spec.txt
Signed-off-by: Saul Wold <saul.wold@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We have a working theory that IO queues on the autobuilder are impacting
runtime testing under qemu, particularly async writes which inice does not
influence. We already pass the snapshot option to qemu which copies the
image and runs out of the copy. Add in the ability to copy the image to
a specificed location which can be a tmpfs. This means that writes to the
image would no longer be blocked by other writes to disk in the system.
Preliminary tests show that this does improve the qemu errors at the expense
of sometimes showing qemu startup timeouts as on a loaded system with a large
test image, it can take longer than 120s to copy the image to tmpfs. Having
a most consistent failure mode for loaded tests is probably desireable though.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We continue to see the warning message:
WARNING: lib/bb/daemonize.py:76: ResourceWarning: unclosed file <_io.TextIOWrapper
name='build/tmp/work/qemux86_64-poky-linux/oe-selftest-image/1.0-r0/testimage/qemurunner_log.20200703011821'
mode='a' encoding='UTF-8'>
I've been unable to reprodue this but believe its caused by garbage
collection of the FileHandler used in QemuTarget being delayed until
after a new tinfoil instance is created by a subseqent test.
Force the log file to be closed when we stop using it to avoid this.
[YOCTO #13961]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We're seeing:
WARNING: bitbake/lib/bb/cookerdata.py:136: ResourceWarning: unclosed file <_io.FileIO
name='tmp/work/qemux86_64-poky-linux/core-image-minimal/1.0-r0/testimage/qemurunner_log.20200601181912'
mode='ab' closefd=True
which can only be caused by the qemu.stop() method not being called.
Tweak the error handling to fix the blanket exception handler which
is likely meaning this function isn't getting called.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We had a mystery failure on the autobuilder where runqemu appeared to
be failing as a logfile directory no longer existed. The key to
reproducing was running a runqemu where the image was deleted (as
devtool does), then running another runqemu test. E.g.:
'oe-selftest -r devtool.DevtoolExtractTests.test_devtool_deploy_target wic.Wic2.test_qemu_efi'
This then tries to write to the logfile from the first test, the
image directory was deleted and we get strange failures.
The fix is to remove the logging handler when qemu is stopped.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Not all QEMU machines are capable of having more than one serial port,
this is due to the machine emulating a physical device/board.
Rework QemuRunner to handle machines that only have 1 serial port, where
the serial port shares output of the kernel log buffer and a login
console. In this case the output is mixed but enables the machine to
boot and have QemuRunner detect the login prompt.
QemuTarget uses SERIAL_CONSOLES to determine the number of available
serial ports.
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Whilst seemingly a good idea, this exception doesn't really serve any purpose
that bb.fatal() doesn't cover. Wrapping exceptions within exceptions isn't
pythonic.
Its not used in many places, lets clean up those and remove usage of it
entirely. It may ultimately be dropped form bitbake entirely.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
As runqemu with launch_cmd means directly run the command, don't need set
rootfs or env vars.
Since commit [a847dd7202 runqemu: Let qemuparams override default settings]
applied in oe-core, if launch_cmd contains "qemuparams='***'", it does not
work, which is overridden by latter qemuparams="-serial tcp:127.0.0.1" in
QemuRunner.launch();
So we set qemuparams as a parameter in runqemu, the fix makes it work
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This adds SPDX license headers in place of the wide assortment of things
currently in our script headers. We default to GPL-2.0-only except for the
oeqa code where it was clearly submitted and marked as MIT on the most part
or some scripts which had the "or later" GPL versioning.
The patch also drops other obsolete bits of file headers where they were
encoountered such as editor modelines, obsolete maintainer information or
the phrase "All rights reserved" which is now obsolete and not required in
copyright headers (in this case its actually confusing for licensing as all
rights were not reserved).
More work is needed for OE-Core but this takes care of the bulk of the scripts
and meta/lib directories.
The top level LICENSE files are tweaked to match the new structure and the
SPDX naming.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There are failures when multiple users run oe-selftest on the same
host:
PermissionError: [Errno 13] Permission denied: '/tmp/oe-saved-
tests/201812250324_qemu'
This is because /tmp/oe-saved-tests was created by user A, while user B tries
to write data in it, then the error will happen. This patch can fix the
problem.
Move the dumped data to ${LOG_DIR}/runtime-hostdump/ rather than
/tmp/oe-saved-tests/ to fix the problem.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Fixed:
MACHINE = "qemux86"
QEMU_USE_KVM = "qemux86"
IMAGE_CLASSES += "testimage"
$ oe-selftest -r runqemu.RunqemuTests.test_boot_rootfs
[snip]
File "/buildarea1/lyang1/poky/meta/lib/oe/types.py", line 122, in boolean
raise ValueError("Invalid boolean value '%s'" % value)
ValueError: Invalid boolean value 'qemux86'
Now QEMU_USE_KVM can only be boolean, can not contain MACHINE any more, kvm
will be enabled if target_arch == build_arch or both of them are x86 archs.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The 5s timeout for non-kvm is too short, especially when the load is high,
which leads to unexpected errors, so set timeout to 60s by default.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
So that both QEMU_USE_KVM = "True" and "1" will work.
[YOCTO #12343]
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
The logger handling in oeqa was confused at best. This patch:
a) Passes in a logger through various qemu runner pieces
b) Uses that logger consistently in the code
c) Creates a logger for QemuRunner outside the bitbake namespace
meaning we don't conflict with the tinfoil logging changes
The result of this is more consistency. For runtime tests in testimage,
the logs always contain the debug info, nothing is shwon on the console.
For the oe-selftests, logs are intercepted and only shown if the test
fails.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This funciton appears completely unused, drop it.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
By default, QemuRunner avoids modifying the image files that it boots
into by enabling the qemu snapshot mode. However, some tests may want
to test changes that must persists across reboots, so this mode
should be optional.
This can be combined by copying the image file to a temporary location
first and then booting with that copy. It's also useful when testing
with additional drives attached to a virtual machine.
QemuTinyRunner doesn't use the snapshot parameter and therefore ignores
the new parameter.
Long term, a better way of passing these various configuration
parameters should be used, and perhaps QemuRunner and QemuTinyRunner
can be merged into one again to avoid code duplication. But for now
the patch follows the exiting style.
Also beware that QemuTarget.start() now acts in two different modes
(with or without explicit launch command), and depending on that mode
parameters like discard_writes must be ignored, i.e. not get passed to
launch().
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Previously, QEMU_USE_KVM=True enabled the use of kvm only when "x86"
was in the MACHINE name. That is too limiting, because for example
intel-corei7-64 can also use kvm but it wasn't possible to enable that
without changing OE-core.
That traditional usage is still supported. In addition, QEMU_USE_KVM
can be set to a list of space-separated MACHINE names for which kvm is
to be enabled.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The bb.note prints multiple same lines when invoke this class again, but
if we set mainlogger.propagate = False, nothing would be printed,
according to logging's document:
https://docs.python.org/3/library/logging.html
Note
If you attach a handler to a logger and one or more of its ancestors, it
may emit the same record multiple times. In general, you should not need
to attach a handler to more than one logger - if you just attach it to
the appropriate logger which is highest in the logger hierarchy, then it
will see all events logged by all descendant loggers, provided that
their propagate setting is left set to True. A common scenario is to
attach handlers only to the root logger, and to let propagation take
care of the rest.
We may need avoid using bb.note or bb.warn in oeqa since it attaches
multiple log handlers which may cause confusions
This patch only sets "mainlogger.propagate = False" in
selftest/runqemu.py and use logger.info to replace bb.note in
targetcontrol.py to minimize the impact.
[YOCTO #10249]
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Modify the following files to test runqemu:
targetcontrol.py
utils/commands.py
utils/qemurunner.py
We need simulate how "runqemu" works in command line, so when test
"runqemu", the targetcontrol.py, utils/commands.py and
utils/qemurunner.py don't have to find the rootfs or set env vars.
[YOCTO #10249]
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Some commands might need to run longer than the default timeout of
five seconds. If that occurred, run_serial() returned with a status
code of zero (sic!) and no other indication of what went wrong.
Now the timeout is configurable (with five still the default) and
an explicit warning ("<<< run_serial(): command timed out after 5 seconds without output >>>")
gets appended at the end of the data returned to the caller.
While at it, the logic for checking for the timeout was updated a bit
because both implementations could overshoot the timeout when entering
select() right before the final deadline.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
qemu runner picks up first fsimage type from the hard-coded
list of supported types. This makes it impossible to test
particular image type unless it's not ext4(first type in
the hardcoded list of types).
Added image_fstypes argument to commands.runqemu and QemuTarget
__init__ to specify type of the image to run qemu with.
This will be used to pass wic image type to test efi wic images.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Added possibility to pass additional runqemu parameters
down the stack of APIs:
commands.runqemu -> QemuTarget.start -> QemuRunner.start
This will be used to pass ovmf parameter in testing of
efi wic images under qemu.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
getVar() now defaults to expanding by default, thus remove the True
option from getVar() calls with a regex search and replace.
Search made with the following regex: getVar ?\(( ?[^,()]*), True\)
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Rather than copying images, use the snapshot option to qemu. This fixes a regression
caused by the recent runqemu changes where the wrong images were being testes since
the image is copied without the qemuboot.conf file. This means the latest image is
found by runqemu rather than the specified one, leading to various confused testing
results.
It could be fixed by copying more files but use snapshot mode instead.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Using kvm can provide significant speedups when running qemux86* machines
on an x86* host. Enabled by using the new QEMU_USE_KVM variable.
[YOCTO #9298]
Signed-off-by: Bill Randle <william.c.randle@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Now Runner's support extra_bootargs for the kernel so add
extra_bootparams to the start() methods to avoid exception.
Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
cmdline
Add ability to specify extra_bootargs (kernel cmdline) in order to enable systemd
debug log in images that enables systemd init.
[YOCTO #9299]
Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
This patch contains all the other misc pieces of the transition to
python3 which didn't make sense to be broken into individual patches.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Added 'wic' to the list image types supported by targetcontrol.
This is a preparation for booting and testing wic images
with oe-selftest.
[YOCTO #8498]
(From OE-Core rev: 7dda053fbd1ea1354b7720cfa691470ba88ef5b9)
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Added new parameter 'ssh' to targetcontrol 'start' method
to be able to test images without running ssh server.
[YOCTO #8498]
(From OE-Core rev: 1c3c66aadd43092bc19242b0651ee810cc31fe7c)
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently, if qemu segfaults, the tests merrily continue trying to execute
which takes time for them to timeout and is a bit silly. Worse, no logs about
the segfault are shown to the user, its silent!
This patch tries to unravel the tangled web of issues and ensures that we:
* install a SIGCHLD handler which tells the user qemu exited
* check if qemu is running, if it isn't fail the test outright
* don't leave processes behind in sshcontrol which would hold
bitbake.lock and block shutdown
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This adds an instance of HostDumper to qemurunner,
with this instance now is possible to get dumps
from the host when there is an error.
This adds dump points in the next cases:
- runqemu exits before seeing qemu pid
- Fail to get qemu process arguments
- Not reach login banner before timeout
- qemu pid never appears
This also modifies the constructors of BaseDumper,
HostDumper and TargetDumper, they don't require
the datastore anymore, but the feature to replace
datastore variables has been lost (never used)
[YOCTO #8118]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It makes sense to separate the dump commands from the
oeRuntimeTest class, this way it can be used in all
the test context.
These are the changes included in this patch:
- Created classes: BaseDumper, HostDumper, TargetDumper
- Create an instance of HostDumper in imagetest.bbclass
and add it to TestContext class, this way any class
that have access to the TestContext would be able
to dump logs from the host
- Create an instance of TargetDumper in QemuTarget
class after get the runner, this way it is
accessible during the tests.
[YOCTO #8118]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
|
|
This patch modify three files altought two of them
are minimal modifications. This version includes
the changes proposed by Paul.
testimage.bbclass:
Create new vars for easy modification of the dump
directory and commands to be run on host and target
when a test fails
TESTIMAGE_DUMP_DIR: Directory to save the dumps
testimage_dump_target: Commands to run on target
testimage_dump_host: Commands to run on host
oetest.py:
- Allow to use the vars defined in testimage class
- Now able to run commands in the host and dump the
results
- Fix an issue with the condition where to run the
dump commands (Before it run the commands every
test after a failure, now it runs the commands only
in tests that failed)
- Fix the output to stdout
[YOCTO #8118]
(From OE-Core rev: 26fe645457633f90bb5ddbb12f5f7b9ca4a06cc5)
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We've had a few cases on the autobuilder where we've lost logs of
the boot and been unable to debug it further. Show this information
onto the console to be more useful.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If we use this outside of testimage we don't have a task log; so let's
just explicitly write the log output to a file all the time so it's
always there to look at (particularly useful when runqemu exits
immediately with an error.)
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The test directory might not exist at this point so just go ahead and
create it.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Rather than just use d.getVar(X), use the more explict d.getVar(X, False)
since at some point in the future, having the default of expansion would
be nice. This is the first step towards that.
This patch was mostly made using the command:
sed -e 's:\(getVar([^,()]*\)\s*):\1, False):g' -i `grep -ril getVar *`
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Lucian Musat <george.l.musat@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
There is no good reason not to use ext4 at this point, it has advantages
and few drawbacks. Therefore switch the qemu machines over (and the default
runqemu script options).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
the resulting value check.
Because we used a bb.fatal call inside the get_image_fstype classmethod, this caused problems when accessed without instantiating the object with a valid bb environment.
Separating the matching part of the classmethod(that is usable by outside scripts) from the check of the resulting value.
The matching is done within a new classmethod and the latter keeps the old method name and internal functionality, this way we don't have to change any other target controllers code.
Signed-off-by: Corneliu Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
controllers
YB: #6254
Add a new classmethod that can be used by outside scripts to get the extra files needed by the target controllers.
An outside script can predict rootfs, manifest and kernel files needed by a target controller, but sometimes there are other files needed.
Signed-off-by: Corneliu Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
YB: #6375
Added a new method get_image_fstype() that autodetermines what fstype to use for the rootfs file.
This method uses a new list variable 'supported_image_fstypes' that contains image fstypes supported by the target controller.
This method is also a classmethod which means outside scripts can get the image fstype.
Signed-off-by: Corneliu Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|