How to obtain thread stack information WITHOUT pstack and gdb

Question

Is there a way of obtaining information about why program got stuck, when not being able to insert any logs and not having pstack and gdb, on some very tiny linux?

Answer 1

My suggestion is to do the following:

1. Be able to retrieve manipulate and print a stack trace within your own program
2. Install a signal handler (say, for SIGUSR1) for your process which would do that
3. Send the signal to your program when it's stuck - control will be moved to the signal handler (assuming you haven't masked out signals).

Now (2.) and (3.) and pretty simple. For (3.) it's kill -USR1 1234 (for a process with ID 1234). For (2.), it's:

#include <signal.h>
typedef void (*sighandler_t)(int);
sighandler_t sigset(int sig, sighandler_t disp);

(read man sigset for details).

As for (1.), this used to be quite difficult and tricky, but now it's been essentially resolved:

Boost StackTrace: Code | Documentation

It will do the best possible to get as much stack trace info as is possible. A simple example of use:

#include <iostream>
#include <boost/stacktrace.hpp>

// ... somewhere inside the `bar(int)` function that is called recursively:
std::cout << boost::stacktrace::stacktrace();

may result in the following output:

0# bar(int) at /path/to/source/file.cpp:70
1# bar(int) at /path/to/source/file.cpp:70
2# bar(int) at /path/to/source/file.cpp:70
3# bar(int) at /path/to/source/file.cpp:70
4# main at /path/to/main.cpp:93
5# __libc_start_main in /lib/x86_64-linux-gnu/libc.so.6
6# _start

Note: StackTrace is a very recent addition to Boost (as of August 2017); you might consider getting it independently if your Boost version is not recent enough.