Re: [PATCH] arm: move _end to linker symbols

[...]

...

...

•         *(.__end)
  

• } >.sram
  

• _end = .;
  

Does this have any kind of consequence in terms of checking the offset of .end vs the SRAM size ? (.sram)

The value of _end does grow by 16b on the SPL for the kria boards, but it doesn't seem to cause issues. I can only find a single assert on those checks in arch/microblaze/cpu/u-boot-spl.lds. Anything else you spotted?

Keeping in mind that the code used to be like that before that linker bug -- defining a linker symbol instead of a section, I don't think anything will blow up

Also if you compile with -O2 instead of -Os the final binary and the _end memory address remain unchanged. So I assume the changes in the _end offset is due to compiler optimizations since the .end section is removed

Cheers /Ilias

/Ilias

...

...

  _image_binary_end = .;

  .bss :