Output Sections

da cheng

Hi all,

I am not sure about the meaning of output sections, I know they collect all input section to form an output section.

For example, all .const sections are collected into a big .const, right?

But to me, it is not that necessary, as the two .const input sections are totally independent, right?

For .bss, I can understand, as it is not initialized, thus an complete area is provided to the program as an output section .bss.

So my question is, is there any benefit that the linker use output sections to organize the input sections? Take the example of .const, .text, etc.

-da

over 15 years ago

0 George Mock over 15 years ago

TI__Guru**** 253170 points

da cheng said:
For example, all .const sections are collected into a big .const, right?

That's true if you use the short hand syntax ...

.const > MEMORY_RANGE

In theory, you could explicitly say where every input .const section goes ...

const_one { file1.obj(.const) } > MEMORY_RANGE
const_two { file2.obj(.const) } > MEMORY_RANGE
/* and so on */

da cheng said:
But to me, it is not that necessary, as the two .const input sections are totally independent, right?

Correct.

da cheng said:
So my question is, is there any benefit that the linker use output sections to organize the input sections? Take the example of .const, .text, etc.

The combination of output sections and short hand syntax make it easy to specify the most common ways input sections are finally placed in memory. Where that doesn't serve your purpose, use the longer form syntax to take control. You might find this useful: http://processors.wiki.ti.com/index.php/Code_Generation_Tools_FAQ#Q:_How_can_I_get_the_linker_to_place_a_piece_of_code_or_data_so_that_it_comes_before_all_the_rest.3F

Thanks and regards,

-George