Traditional VT-10x terminals (and their emulators) [0] have a "magic
margins" feature that enables the last character position to be updated
without scrolling the screen: whenever a character is printed on the
last column, the cursor stays over the character, instead of moving to
the next line.
The Busybox shell, ash, attempts to defeat this feature by printing
CR,LF right after echoing a character to the last column.[1] This
doesn't play well with emulator.py. The run() method of the Emulator
class captures the output of the emulated system and assumes the first
line it reads is the echo of the command, and all subsequent lines are
the command's output. If the line made by the command + shell prompt is
longer than 80 characters, then it is echoed as two or more lines, and
all but the first one are mistaken for the command's output.
We fix this by telling the emulated system that we are using an
ultra-wide terminal with 29999 columns. Larger values would be ignored
and replaced by the default, namely 80 columns.[2]
[0] https://vt100.net/docs/vt100-ug/chapter3.html - DECAWM
[1] https://git.busybox.net/busybox/tree/libbb/lineedit.c?h=1_34_0#n412
[2] https://git.busybox.net/busybox/tree/libbb/xfuncs.c?h=1_34_0#n258
Signed-off-by: Edgar Bonet <bonet@grenoble.cnrs.fr>
Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Co-authored-by: Yann E. MORIN <yann.morin.1998@free.fr>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
When booting under EFI, grub2 will output a nice and shiny boot menu,
using extended ASCII characters (in the [0x80..0xFF] range), namely
CP437 [0], on the assumption that the VGA BIOS is a real one and has the
corresponding (and only!) font, as is the case on real hardware.
However, when run in our runtime test infrastructure, this triggers the
infamous python UnicodeDecodeError exception:
Traceback (most recent call last):
[...]
emulator.login()
File "[...]/buildroot/support/testing/infra/emulator.py", line 89, in login
index = self.qemu.expect(["buildroot login:", pexpect.TIMEOUT],
File "/usr/lib/python3/dist-packages/pexpect/spawnbase.py", line 340, in expect
return self.expect_list(compiled_pattern_list,
File "/usr/lib/python3/dist-packages/pexpect/spawnbase.py", line 369, in expect_list
return exp.expect_loop(timeout)
File "/usr/lib/python3/dist-packages/pexpect/expect.py", line 111, in expect_loop
incoming = spawn.read_nonblocking(spawn.maxread, timeout)
File "/usr/lib/python3/dist-packages/pexpect/pty_spawn.py", line 485, in read_nonblocking
return super(spawn, self).read_nonblocking(size)
File "/usr/lib/python3/dist-packages/pexpect/spawnbase.py", line 178, in read_nonblocking
s = self._decoder.decode(s, final=False)
File "/usr/lib/python3.8/codecs.py", line 322, in decode
(result, consumed) = self._buffer_decode(data, self.errors, final)
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xda in position 0: invalid continuation byte
Grub2 is not wrong in emitting those chars, and basically we should not
expect the packages we test to always emit correct UTF-8 sequences; at
the very least, this should not cause the test infra to fail.
We fix that by telling pexpect.spawn to "fix" such invalid sequences by
replacing them with the suitable Unicode character, U+FFFD REPLACEMENT
CHARACTER.
[0] https://en.wikipedia.org/wiki/Code_page_437
[1] https://docs.python.org/3/library/codecs.html#error-handlers
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com>
[yann.morin.1998@free.fr:
- don't change encoding, use codec_errors
- rewrite commit log accordingly
]
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
The pre-built vexpress kernel used by the testing infrastructure is a
4.0.0 kernel, which is getting old to be used with reasonably recent
toolchains.
This commit updates the pre-built kernels for both the versatile and
vexpress machines to 5.10.7 (they have already been put online).
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr>
Python 2.7 will not be maintained past 2020.
Many scripts on the tree are used during the build and should keep
Python 2 compatibility for a while.
This is not the case for the runtime test infra. It's meant to be run in
modern distros only, so it can safely switch to support Python 3 only.
An advantage of this approach is to have less scenarios to test in.
Otherwise every change to the test infra or runtime tests would need to
be tested against both versions of the interpreter, increasing the
effort of the developers, to ensure the compatibility to Python 2 was
not broken.
In order to accomplish the change to Python 3:
- change the shebang for run-tests;
- use Python 3 urllib as a drop-in replacement for Python 2 urllib2;
- when writing the downloaded binary files, explicitly open the output
file as binary;
- when subprocess is used to retrieve the text output from commands,
explicitly ask for text output. For this, use 'universal_newlines'
because 'text' was added only on Python 3.7;
- when pexpect is used to retrieve the text output from qemu or git,
explicitly ask for text output using 'encoding';
- the code using csv currently follows the example in the documentation
for the Python 2 module, change it to follow the example in the
documentation for the Python 3 module;
- fix the relative import for test_git.py to be Python 3 compliant.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Tested-by: Romain Naour <romain.naour@smile.fr>
Tested-by: Nicolas Carrier <nicolas.carrier@orolia.com>
Signed-off-by: Nicolas Carrier <nicolas.carrier@orolia.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
More and more packages being tested by the test infra, e.g. syslog-ng,
need entropy at startup, usually reading from /dev/random.
Some test cases can also depend on a kernel version newer than the
builtin ones already provided by the test infra:
- 3.11.0 for armv5;
- 4.0.0 for armv7.
Add a new builtin kernel to be used by such test cases.
Add it for armv5 so most test cases that switch to use this kernel can
keep using BASIC_TOOLCHAIN_CONFIG.
Use the same kernel version and kernel config as qemu_arm_versatile plus
HW_RANDOM_VIRTIO for VirtIORNG to be usable.
Copy the actual binary file from the syslog-ng runtime test at current
master @ 29e1cb8884.
Since there is already a 'kernel-versatile' file on autobuild.buildroot.net
and we must keep it with this name for reproducibility purposes, create a
simple naming convention for newer builtin kernel images and dtb files:
kernel-<defconfig>-<kernel_series_version>
<dtb_name>-<kernel_series_version>.dtb
Pass '-device virtio-rng-pci' to qemu when this kernel is used.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Peter Korsgaard <peter@korsgaard.com>
Cc: Romain Naour <romain.naour@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
[Peter: use this new kernel instead of the old builtin/armv5 kernel]
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
By default, Qemu emulates a system with 128 MB of RAM. This is not
sufficient for some test cases we have, such as TestPerlDBDmysql,
where the initramfs is quite large. Therefore, this commit extends the
RAM size emulated by Qemu to 256 MB.
Fixes:
https://gitlab.com/buildroot.org/buildroot/-/jobs/237108668
Thanks to Arnout for the analysis of the issue.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Acked-by: Francois Perrad <francois.perrad@gadz.org>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Add a parameter to run-tests to act as a multiplier for all timeouts of
emulator.
It can be used to avoid sporadic failures on slow host machines as well
in elastic runners on the cloud.
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
[Arnout: rename multiplier to timeout_multiplier everywhere]
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
The default audio backend for qemu is configured at compile time. It
generates annoying warning messages to qemu's stderr when running our
tests, like these:
pulseaudio: set_sink_input_volume() failed
pulseaudio: Reason: Invalid argument
pulseaudio: set_sink_input_mute() failed
pulseaudio: Reason: Invalid argument
Explicitly set the audio backend to "none" at runtime to remove those
messages from our logs. There is no command line argument for this, so
use an environment variable when starting qemu.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Some commands take more than 5 seconds to complete under QEMU, so add
provisions to allow individual unit-test to specify different duration
to avoid false negative test failures.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Reviewed-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Tested-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When running multiple instances of emulator in parallel, the login
prompt can take some time to appear.
Use a large timeout when waiting for the prompt to avoid random
failures.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Remove unused import.
Use 2 empty lines before a class.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When using pexpect there is no need for a helper function. Just use
expect() directly everywhere.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When using pexpect there is no need for a helper function. Just use
sendline() directly everywhere.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When the parameter logfile is passed to spawn(), pexpect sends both
stdin and stdout to the logfile and it creates a double echo effect.
One way to avoid the double echo in the logfile would be to disable the
echo on the terminal just after login ("stty -echo"), but double echo of
user and password would remain.
Instead of that, send only the stdout to the logfile using the
logfile_read property.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Instead of redirecting qemu serial to telnet, redirect it to stdio.
It allows to run testcases in parallel without random failing caused by
two emulators trying to use the same telnet port (1234).
'qemu -serial stdio' returns some extra <CR> characters, so remove them
from the log.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Replace subprocess + telnetlib with pexpect.
Use the telnet installed on the host machine instead of telnetlib, while
the serial from qemu is not yet redirected to stdio.
Signed-off-by: Ricardo Martincoski <ricardo.martincoski@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
We currently call infra.smart_open() to open log files each time we
need to write to them.
Opening the file once in the constructor of Builder and Emulator and
writing to it whenever needed is simpler and slightly more efficient.
Remove smart_open and instead create a new open_log_file() function
which just opens the logfile. Also let it compute the filename, in
order to simplify even further the Builder and Emulator code.
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This commit adds the core of a new testing infrastructure that allows to
perform runtime testing of Buildroot generated systems. This
infrastructure uses the Python unittest logic as its foundation.
This core infrastructure commit includes the following aspects:
- A base test class, called BRTest, defined in
support/testing/infra/basetest.py. This base test class inherited
from the Python provided unittest.TestCase, and must be subclassed by
all Buildroot test cases.
Its main purpose is to provide the Python unittest setUp() and
tearDown() methods. In our case, setUp() takes care of building the
Buildroot system described in the test case, and instantiate the
Emulator object in case runtime testing is needed. The tearDown()
method simply cleans things up (stop the emulator, remove the output
directory).
- A Builder class, defined in support/testing/infra/builder.py, simply
responsible for building the Buildroot system in each test case.
- An Emulator class, defined in support/testing/infra/emulator.py,
responsible for running the generated system under Qemu, allowing
each test case to run arbitrary commands inside the emulated system.
- A run-tests script, which is the entry point to start the tests.
Even though I wrote the original version of this small infrastructure, a
huge amount of rework and improvement has been done by Maxime
Hadjinlian, and squashed into this patch. So many thanks to Maxime for
cleaning up and improving my Python code!
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
