[PATCH 11/25] Merge sapi/Makefile.am into cpukit/Makefile.am

Chris Johns chrisj at rtems.org
Tue Oct 2 04:54:34 UTC 2018

On 01/10/2018 15:52, Sebastian Huber wrote:
> On 25/09/2018 07:58, Chris Johns wrote:
>> On 25/09/2018 15:30, Sebastian Huber wrote:
>>> On 25/09/2018 00:27, Chris Johns wrote:
>>>>>>> We would have to go from the symbol to the ELF file to the DWARF info of the
>>>>>>> compilation unit.
>>>>>> All info should be directly via the ELF and DWARF format. I do not know if we
>>>>>> need to reference the ELF symbol table or we can just use the DWARF CU data.
>>>>>> The
>>>>>> use of objdump should be avoided.
>>>>> I only briefly look into this, but you have to go from the
>>>>> rld::symbols::symbol
>>>>> to the DWARF info somehow if DesiredSymbols.cc:95 is the right spot to add
>>>>> this
>>>>> stuff.
>>>> You could be right. The DWARF info will contain all symbols and more.
>>>> Should these files move to the rtems.git repo?
>>> The *.ini files or which files?
>> Just the INI files, the files which describe the data.
>>> I would leave the files where they are and fix
>>> the issues first. Afterwards I would think about moving files around across
>>> repositories.
>> Oh of course, that is fine and it makes sense. I would like to understand the
>> tasks and agree on them.
> This stalled my background work on general RTEMS improvements (in this case the
> build system). I think the path is clear. We use source tree path regular
> expressions to assign source files to categories. The source tree path is
> determined for each symbol by the DWARF compilation unit information. The
> categories are defined in an INI file in rtems-tools for now. Maybe later it is
> places somewhere else.


> I would like to continue now with the build system changes. This will break the
> coverage tool categories for some time.

Please continue. We will have to update the coverage tool as things breaks.

> Is there some documentation on how I can use the coverage tool? Which BSP is
> recommended for test runs?

I am only using leon3 and the special qemu what generates the needed trace data.


More information about the devel mailing list