babylog

easy as pie debug-message logger, intended for bootcode or other minimalist environments

Intention and idea of babylog

The intention of babylog is to ease debugging of very minimalistic applications (e.g. bootloaders / boot-code).

The idea for babylog came when deugging the SPL (Secondary Program Loader) of the bootloader "Das U-Boot". The SPL is U-Boot's very early bootcode that must be super small to fit in every SoC's (System-On-Chip) oboard S-RAM. Therefore:

• SPL has just minimal C standard lib support
• A significant portion of the SPL runs in a very early boot-state, where no interfaces are available to print debug messages (e.g. serial UART)

All that makes debugging very hard, especially for the inital system/board bring-up.

How babylog ease deubgging

Babylog allows to put various debug messages into a static buffer and print them later to some interface (e.g. UART) whenever that interface was initialized.

Therefore babylog has to important elements:

• use of as less as possible stdlib functionality
  • babylog's basic functions (babylog_putNamed_i(), babylog_putNamed_s(), ...) have almost no C stdlib dependency; they can be used to create a log message with a single value (e.g. int in decimal format, int in hex format, string, ...) and a name-string;
  • optionally the function babylog_printf() allows to create log message with the printf-like C-formatting style; babylog_printf() can be disabled if the target environmanet does not support vsnprintf();
• easy way to integrate babylog into the target application
  • the idea is to have all babylog's prototypes and implementation in just one c header-file (i.E. there is no *.c file at all); that header-file is used to embed (or include) the babylog's implementation into exactly one c-file of the target appliction.
  • So that babylog.h file is used for two purposes:
    • 1. Include that file in exactly one of your target application's c-files, to have all babylog's prototypes available in that file. Then set the preprocessor macro BABYLOG_HAVE_IMPLEMENTATION in the line before the include directive. That macro tells babylog.h to activate babylog's function-bodies and so embeds the implementation of babylog into that target application's c-file.
    • 2. When (a) above applied, then optionally include the same babylog.h also in any other c-file of the target application, but without the macro BABYLOG_HAVE_IMPLEMENTATION. So for those c-file babylog.h does not activate the function-bodies but provides all babylog's prototypes.

How to use babylog

Example use for babylog is available with the "test application".

Build the "test application" with command line:

gcc test.c test-another-file.c -o test.out

Then run with:

./test.out

The test application shows:

• how to embed babylog into one c-file, just with include (i.E. prototypes AND code implementation)
• how to use babylog also on other c-file, just with include (on those others just prototypes)
• how to create log-messages
• how to get and output log-buffer
• how detection of log-buffer overflow works