I recently started using qemu and it's a great tool when you don't have the required hardware to run your firmware (currently using it for cortex-m3).
Now what I want to do is to do some test coverage with it. I tried using GNUC ARM Eclipse, and I've been successfull compiling and executing the code in qemu, but whenever I add the -fprofile-arcs -ftest-coverage flags (for the project and then for the desired file to run coverage) I am able to create the .gcno file, which means that after executing my code it will generate a .gcda file and then I should be able to see the coverage.
Thtat's where everything goes wrong. I was able to generate a .gcda file but whenever I try to open any of them, eclipse tells me that it wasn't able to open the file because it was null. I've tried replicating the procedure in another computer, but I haven't been successful creating the gcda file (probably different binaries).
At this point I don't really know how to proceed. Should I abandon ARM Eclipse and stick to makefiles (is it possibll to run gcov this way?) or am I missing something really small that is fixable?
PS: I using windows 7 64 bits, the latest versions available on the GNU ARM Eclipse website. Also the idea of doing it via makefiles just occurred to me (it was a stresfull day, it's really late) so I haven't tried it yet, I've only tried executing the code, but without coverage.
As far as I know, qemu is not able to generate DWARF information. But there is a project with the proposal of code coverage with qemu: Couverture Project
When you use qemu as a user space emulator (see also the qemu documentation) you actually can measure the code coverage as usual. In this mode qemu has the full access to the host file system.
For a CMake project you can simply use the CROSSCOMPILING_EMULATOR property of your test executable, e.g.:
if(CMAKE_CROSSCOMPILING)
set_target_properties(mytest
PROPERTIES
CROSSCOMPILING_EMULATOR "qemu-${CMAKE_SYSTEM_PROCESSOR};-L;$ENV{SDKTARGETSYSROOT}"
)
endif()
With this setting ctest will use qemu for running the test and will write the .gcda files to the usual location in your build directory.
Related
Is there a way or steps to follow to integrate a yocto SDK (standard or extensible) with VSCode? I want to cross-compile, remote connect, and debug a C/C++ application within VSCode for target hardware using a yocto generated Linux image. Is this possible? I know of the bitbake extension but couldn't find one for the SDK. Thank you!
Conservatively, I would say it depends on the level of integration you want to achieve but I use regularly VS Code to edit and build, sometime to debug C applications using a Yocto toolchain, that's really easy for Makefile projects for example.
Assuming you do not ask for Yocto integration into VS Code (I don't know if something exists) but really to use the tools generated by the SDK from Yocto and that you already are familiar with Yocto toolchain usage.
I personally compile on Linux server remotely from a Windows PC. The server contains therefore my projects and the Yocto toolchain.
I use for that the nice Remote SSH extension from Microsoft on VS Code. From there, I can edit easily the files, compile and a terminal is available (that's out of the scope of your question however).
So if working as me or directly in Linux, you can create a Makefile/CMake project for example. The C/C++ VS Code extension is a must have.
Each time you start working, you source the Yocto SDK toolchain and compile directly using make from the terminal window of VSCode. If you want to automatize the build step, you can use the task feature of VS which allows you to launch build script for example.
Regarding the remote connect, the terminal window of VS can also have multiple sub-windows with various connections like SSH to the target. The build script can also use scp to send the generated binary directly to the target but your question is vague regarding what you want to do.
Finally for the debug aspect, GDB is well supported in VS Code and the official doc is a good start as well as the C++ debugging doc.
On the Yocto side, you need to add gdbserver to the image running on the target, it can be done by adding the following to your conf/local.conf:
EXTRA_IMAGE_FEATURES += "tools-debug"
If you want to have debug information for the shared libs on the target, you also need to add:
EXTRA_IMAGE_FEATURES += "dbg-pkgs"
Finally, the SDK must be generated with the same options as the image running on the target and will contain the cross-gdb tool like -gdb to be used on the host side.
So that's possible but requires some setup especially the debug part. As far as I know, there is not a VS Code extension managing all these steps for you automagically.
I have a simulink model (2016b with MC 2013 C/C++ and Mingw-64 compilers) that I'd like to generate a standalone executable for windows-64 bit.
I was able to run the grt executable but due to the fact that I need to read a mat file runtime as opposed to compile time, I am using rsim code generation for this purpose, however the executable that gets generated appears to need quite a bit of .dll, I provided the dll it was asking for however, the application still unable to run. This is the error that results
The application was unable to start correctly 0xc000007b. Click OK to
close the application
What am I missing ?
Your main program is compiled for x64 (64-bit) target, but the dll you provided is compiled for x86 (32-bit) target. Or vice versa.
If it is Mingw-64 stuff, you should be able to obtain all (or most) of them by using the official online installer. Link is here.
I'm trying to debug an open source package, called libprotoident in Eclipse, Kepler version, within Debian. As it has the Makefile, I choose to make an empty Makefile project, and then add all the sources into the workspace. So after that the source compiled and run successfully as in the command line using the Makefile.
As it has 4 apps you can use, I choose to run lpi_protoident package in the run configuration window, as the following image shown.
So the Program ran successfully. Now I'm trying to debug it but it generates the following error.
How can I solve this error and debug the Project?
The file you are trying to debug is most likely a shell script created by automake that acts as a wrapper around the real executable, which has been built in a hidden directory.
Instead of telling Eclipse that tools/protoident/lpi_protoident is your application, try using tools/protoident/.libs/lpi_protoident instead.
General Answer about the error you are getting
What not in executable format: File format not reconized error means is that lpi_protoident is not an executable on the platform you are working on.
Are you sure that is an executable you can run (E.g. from the command line)?
There is also the small chance that the GDB you are using is somehow incompatible with the executable, but that is less likely.
Building libprotoident from source
(Assuming you are trying to build https://github.com/wanduow/libprotoident)
You are trying to build an automake project. The normal way to do that is by configuring to create Makefile, you shouldn't be making your own makefile. Please refer to the README in the project, but the key parts you need to do are:
Installation
After having installed the required libraries, running the following series of commands should install libprotoident
./bootstrap.sh (only if you've cloned the source from GitHub)
./configure
make
make install
By default, libprotoident installs to /usr/local - this can be changed
by appending the --prefix= option to ./configure.
The libprotoident tools are built by default - this can be changed by
using the
--with-tools=no option with ./configure.
I was wondering if there is a way to get Bazel to list, output, display, etc., all of the commands that can be executed from a command line that are run during a build after a clean. I do not care if the output is to the screen, in a file, etc. I will massage it into a usable form if necessary.
I have captured the screen output during a run of Bazel which gives me an idea of what is being done, however it does not give me a command I can execute on the command line. The command would have to include all of the command options and not display variables.
If this is not possible, since Bazel is open source, where in the code is/are the lines that represent the commands to be run so that I can modify Bazel to output the executable commands.
I am aware of the query command within Bazel, and used it generate the dependency diagram. If this could be done as a query command it would be even better.
TLDR;
My goal is to build TensorFlow using Bazel on Windows. Yes I know of all of the problems and reasons NOT to do it and have successfully installed TensorFlow on Windows via a Virtual Machine or Docker. I did take a shot at building Bazel on Windows starting with Cygwin, but that started to get out of hand as I am use to installing with packages and Cygwin doesn't play nice with packages, so then I started trying to build Bazel by hand and that was turning into a quagmire. So I am now trying to just build TensorFlow by hand on Windows by duplicating what Bazel would do to build TensorFlow on Linux.
You are correct, you can use the -s (--subcommands) option:
bazel build -s //foo
See https://docs.bazel.build/versions/master/user-manual.html#flag--subcommands.
For your use case, you'd probably want to redirect the output to a file and then global replace any library/binary paths to the Windows equivalents.
You might want to track https://github.com/bazelbuild/bazel/issues/276 (Windows support), although it'll probably be a while.
(Disclaimer: This solution does not print the commands that currently get executed but the commands that would get or got executed.)
I'd use aquery (action graph query) (forget about "graph"):
bazel aquery //foo
Advantages:
It's very fast, because it prints the actions without executing the build.
It's a query. It does not have side effects.
You don't have to do a bazel clean before in order to find out the build steps for a library that has already been built.
It prints information about the specific build step that you request. It does not print all the build commands required for the dependencies.
I'm having problems debugging a JNI application. I've read several threads in StackOverflow, like this one, this one or this one. I've also tried to start gdb in a separated shell and attach it to the running java process. In both cases, the problem is the same: GDB can't find the sources to debug. Things tried
Add "dir" line to gdbinit, pointing to C++ sources folder
Adding the C++ sources folder to the GDB debbuging configuration in Eclipse, in the "Sources" tab.
Adding set environment LD_LIBRARY_PATH=/path/to/library.so, being library.so the library file built from C++ source files
Attach ddd to the java process, but then I get an error because pthread_join.c is not found in the working directory. I don't have this file in my hard disk. I don't know what is this about.
Nothing worked. I've spent several days on this. I know my bug is in the C++ code called by the JNI wrapper, but I can't debug it. Any hints? If helps, I'm running Eclipse Juno in Debian 7 under a Parallels VM on Mac OS.
Many thanks in advance,
You need to have debug information in your native library. You should pass -g to your compiler and linker to have this information in the executable. You may also want to add -O0.
As an alternative to attaching to the Java process, you can create a C++ app and debug it directly. You just need to link in the functions you want to test. In the main function, create the VM, register the functions with RegisterNatives, and kick off a Java test class the uses them.
Hopefully, the debugger has no problem finding the sources since it is just part of the normal compile/link/debug loop of a C++ app.
I would suggest to start with the latest ADT bundle. You can even download the Mac version, so you will not even need Parallels (see a detailed instructions). Then, choose Debug Android Native Application in launch menu.