How to determine when/where pointers in .ctors get mapped?

Question

So I am learning about ELF, and am looking through a binary in Ghidra as I do. I've made sense of the ELF header, and now I am looking through the program header table.

My binary has a bunch of entries in the program header table, but I am hung up on one in particular...

From referencing...

http://www.sco.com/developers/gabi/latest/ch5.pheader.html#p_flags

...I can see what the different flags mean. Of relevance:

p_offset = 0xABCDEF
p_vaddr = 0x1BCDEF
p_filesz = <number>
p_memsz = <bigger number>

I am able to go to the p_vaddr value in the binary, and it brings me to the .ctors section. Where I do see what appears to be a list of pointers, but currently those pointers do not represent valid virtual addresses (by currently I mean they are not virtual addresses that I can "go" to in Ghidra). When I run the program dynamically in gdb though, I can run x addr_of_interest and it succeeds and says it is pointing to something in a library blah blah.

I found a relevant link from GCC too, but it did not answer my question as far as I could tell...

https://gcc.gnu.org/onlinedocs/gccint/Initialization.html

My question therefore is this: When/how are these pointers mapped to valid memory, and where in the ELF file is the information that would tell me how this happens?

Answer 1

Duh. Stupid.

Importantly, the binary I am looking at is not a PIE (Position Independent Executable). I had Ghidra's image base set to 0x0, rather than 0x400000 (which is the p_vaddr of the loadable segment containing all of the code, and notably, the canonical value used here for x86_64 binaries). This is the only reason the function pointers listed in .ctors were not addresses Ghidra could go to. As soon as I set the image base to 0x400000 they were legitimate.

:facepalm: