Replace stateful loop with a nested comprehension that makes obvious:
* that self._states is a dict(cpu, [CpuidleState])
* the sysfs folder being used and the constraint applied to make use
of each level (i.e. which subfolder is used)
* that the states are sorted by index
As a side effect:
* Gracefully handle non-contiguous idle state names like "state0,
state2" without a state1 (not sure if that can happen)
* Remove some antipatterns while iterating over a dict and counting
iterations.
No longer recursively attempt to push a file to the target. If the
second attempt goes wrong we assume there is something else wrong and
therefore let the error propagate.
Also log the original error in case it is not the error we were
expecting.
logcat -v lets you specify the format for the logcat output. Add a
parameter to the LogcatCollector to allow us to pass that information
down to the logcat invocation.
Paramiko uses sftp for file transfer rather then scp as with the
previous implementation however not all targets support this.
Expose a parameter to the SSHConnection to allow falling back to
the scp implementation.
In the case of connecting to a system without a password the password
parameter needs to be set to an empty string which will currently still
cause ssh keys to be loaded. Only check for ssh keys when the password
is explicitly set to `None`.
Pipe the list of perf events via cat to ensure that a pager is not
used to display the output as this can cause some systems to hang
waiting for user input.
Allow for configuring the number of connection attempts that will be
made to the device before failing to connect. This allows for waiting longer
periods of time for the device to come online.
Unless we tell pexpect to expect something it will not read from the
process' buffer, or write anything to the logfile. If we follow the
collector instructions from devlib's documentation:
In [1]: from devlib import AndroidTarget, LogcatCollector
In [2]: t = AndroidTarget()
# Set up the collector on the Target.
In [3]: collector = LogcatCollector(t)
# Configure the output file path for the collector to use.
In [4]: collector.set_output('adb_log.txt')
# Reset the Collector to preform any required configuration or
# preparation.
In [5]: collector.reset()
# Start Collecting
In [6]: collector.start()
# Wait for some output to be generated
In [7]: sleep(10)
# Stop Collecting
In [8]: collector.stop()
# Retrieved the collected data
In [9]: output = collector.get_data()
adb_log.txt will be empty because between collector.start() and
collector.stop() there were no expect() calls to
LogcatMonitor._logcat. As the get_log() function already has code to
flush the log, abstract it to a function and call it in stop() before
terminating the pexpect.spawn().
By default, pexpect.spawn() is a bytes interface: its read methods
return bytes and its write/send and expect method expect bytes. Logcat
uses utf-8 strings, so put pexpect.spawn() in str mode.
This code can't have worked in python3 before, since the logcat file
is not opened in "b" mode.
If an error occurs while executing a command the `output` and `error`
variables may not get initialised, only attempted to decode their
contents if this is not the case.
check_output(combined_output=True) does not guarantee that stdout will
come before stderr, but the ordering is needed in case check_exit_code
is True, as we are expecting the exit code at the end of stdout.
Furthermore, the exceptions can't report what is stdout and what is
stderr as they are combined.
Partially revert 77a6de9453 ("utils/android: include stderr in adb_shell
output") and parse output and err independently. Return them combined
from adb_shell() to keep the functionality that 77a6de9453 was
implementing.
d4b0dedc2a ("utils/misc: add combined output option to
check_output") adds an option that combines stdout and stderr, but
their order is arbitrary (stdout may appear before or after
stderr). This leads to problems in adb_shell() when it tries to check
the error code. Now that adb_shell() doesn't use combined_output,
remove the option as there are no more users in devlib.
squash! utils/misc: Make the return of check_output consistent
By default parmaiko attempts to search for SSH keys even when connecting
with a password. Ensure this is disabled to prevent issues where
non-valid keys are found on the host when connecting using password
authentication.
Update the documentation to indicated which module each class is
located. This allows the documentation to be referenced from other
modules as well as enabling links to the source code directly from the
documentation.
* Unify the behavior of background commands in connections.BackgroundCommand().
This implements a subset of subprocess.Popen class, with a unified behavior
across all connection types
* Implement the SSH connection using paramiko rather than pxssh.
Update a command line to redirect standard streams as specified using the
parameters. This helper allows honoring streams specified in the same way as
subprocess.Popen, by doing it as much using shell redirections as possible.
This frees the connection to have to handle threading issues, since each thread
using the Target will have its own connection. The connection will be garbage
collected when the thread using it dies, avoiding connection leaks.
Similar to a regular property(), with the following differences:
* Values are memoized and are threadlocal
* The value returned by the property needs to be called (like a weakref) to get
the actual value. This level of indirection is needed to allow methods to be
implemented in the proxy object without clashing with the value's methods.
* If the above is too annoying, a "sub property" can be created with the regular
property() behavior (and therefore without the additional methods) using
tls_property.basic_property .
Use the timeout parameter added in Python 3.3, which removes the need for the
timer thread and avoids some weird issues in preexec_fn, as it's now documented
to sometimes not work when threads are involved.
Add the monotonic clock time to the energy measurements to help
correlate the measurement with those of other collectors, like
FtraceCollector or LogcatCollector.
Instead of calling _read_next_rows() before the while and at the end,
it's simpler to read the rows in a for loop and have _read_rows() be a
generator.
self.tempdir is initialized to None, and os.path.isdir() throws an
exception if you don't pass it a str. This can happen if teardown()
is called before get_data(), which WA sometimes does. Check that
self.tempdir has been initializing before calling os.path.isdir().
Some SoC vendors add several sha's to the kernel version string. This is
problematic for the KernelVersion class, which might identify the wrong one.
Fixing this issue by matching the following "git describe" pattern:
<version>.<major>.<minor>-<rc>-<commits>-g<sha1>
Where commits is the number of commits on top of the tag, which is now a
member of the class.
Prior to this patch:
>> KernelVersion("4.14.111-00001-gd913f26_audio-00003-g3ab4335").sha1
3ab4335
>>> KernelVersion("4.14.111_audio-00003-g3ab4335").sha1
3ab4335
With the modified regex:
>> KernelVersion("4.14.111-00001-gd913f26_audio-00003-g3ab4335").sha1
d913f26
>>> KernelVersion("4.14.111_audio-00003-g3ab4335").sha1
None
Make sure the target.modules list stays up to date when a new module is
installed, since behaviors like devlib_cpu_frequency event injection rely on
content of target.modules.
