Skip to content

Latest commit

 

History

History
76 lines (57 loc) · 3.14 KB

trace32.rst

File metadata and controls

76 lines (57 loc) · 3.14 KB
Raw

Generating Coverage Information from Lauterbach Trace32

Lauterbach Trace32 BranchFlow trace data can be produced while running an executable on a processor supporting the needed execution trace output.

In environments where a Lauterbach probe is available for a target board with a trace port, the Trace32 software can be used to control program loading and execution on target hardware, and configure program trace data collection and output. Trace32 APIs are provided for several languages (including Python and C), and these are useful when not working in interactive mode.

Trace32 runs on a host computer that is connected via USB or Ethernet to a hardware probe, which in turn is connected to a target processor.

Coverage analysis from Trace32 data involves four steps, requiring the specific ability to export BranchFlow traces from Trace32:

  • Run the executable on the target using Lauterbach probe and Trace32;
  • Export BranchFlow trace with Trace32;
  • Invoke |gcvcnv| on the BranchFlow trace data file, producing an intermediate file;
  • Invoke |gcvcov| on this intermediate file.

While the program is running, trace data is sent over the processor's trace interface (depending on the architecture) to the probe. The trace data is then sent to Trace32 on the host computer to be exported to a file suitable for |gcvcnv|.

Enabling Trace Data

For Trace32 to produce BranchFlow execution traces, the target hardware and the probe have to be configured properly. This configuration depends on the target architecture and the specific device.

We recommend starting from one of the examples provided for your hardware in the Trace32 demo folder, for instance demo/arm/hardware/stm32/stm32f4/sieve/stm32f4_discovery_offchip_trace.cmm, and to contact Lauterbach support for advice on the configuration for your target.

Running the executable

The executable can be loaded on the target using the Trace32 command Data.LOAD.auto <ELF_executable>.

Before starting the program, a termination point has to be defined using breakpoint. We recommend setting a breakpoint on the _exit function using the Trace32 command Break.Set _exit.

To run the program, use the Go.direct Trace32 command.

Exporting the Trace Data

Once the program reached the termination breakpoint, the trace data can be exported using the Trace.export.BranchFlow <filename> /NOSYMBOL /CALLER command. The last two options are required to produce a trace file in a format that |gcvcnv| can translate.

The BranchFlow trace should then be converted to |gcp| format using |gcvcnv| command (see :ref:`gnatcov_convert-commandline`).

Automation

Trace32 provides an external API in C and Python to control the execution of commands. This interface can be used to automate the process described above: enable trace, run executable, export trace. Please refer to the Trace32 documentation for more information, for instance Controlling TRACE32 via Python 3.