After extensive public beta testing, Percepio has launched version 4.4 of its Tracealyzer development tool with support for embedded Linux alongside its real time operating systems.
This gives more insight during software debugging and verification at the system level by enabling visual top-down exploratory analysis. This makes it easy for developers to spot issues, also during full system testing, and drill down into the details to find the cause. The new version adds several views optimized for Linux tracing to the vast set of visualizations already present in Tracealyzer, and leverages CTF, the Common Trace Format, and the widely supported LTTng open source tracing framework.
This version adds a Signals and Syscalls Explorer, which operates as an index over the trace, showing how each thread, process and process tree interacts with the Linux kernel through syscalls, and how signals are generated and delivered.
The Communication Flow view has also been optimized for Linux and shows a visual graph over the process interactions with respect to file descriptors, signals and pipes. An Actor Tree field in the main trace view gives visibility of how processes and threads are spawned over time, including their parent/child relations.
“Tracealyzer is firmly established as the leading solution for visual trace diagnostics in the RTOS space. Linux is the single largest platform for embedded and IoT systems today and has an even greater need for better debugging support at system level. We are therefore thrilled to release an even better version of Tracealyzer that is now also optimized for the needs of embedded Linux developers,” said Dr Johan Kraft, founder and CEo of Percepio.
Other additions to the tool include Quick Zoom, a feature that allows users to quickly zoom in by holding down the Ctrl key while dragging the mouse pointer over an interval as well as a set of high-level overviews for top-down exploratory analysis, including process interactions, process forking, CPU usage, RAM usage, I/O usage, file usage, state machines and user-defined metrics while the trace view shows the details, scalable for large Linux traces with