Possible bug; code-block vanishes when adding one assembler-instruction.

[Richard Sandiford]

Jens Bauer <jens-lists@gpio.dk> writes:
>> Normally you'd have a fake "section" like:
>> 
>>   /DISCARD/ : { *(.stack) }
>> 
>> I don't know whether there is something special about .stack on ARM
>> that stops that from working though.
>
> Does this differ from removing the section from the linker script completely ?

Yeah.  By default, all input sections are linked, even those not listed
in the linker script.  If you omit a section from the linker script,
the linker will try to find a logical place for it, based on things like
section type.  /DISCARD/ says you explicitly don't want the section to
be linked.

>>> <http://scratch.gpio.dk/Fcode-Test.zip>
>> 
>> I'm afraid I don't have a full EABI toolchain to hand.  Could you upload
>> the .elf and .hex files?  That ought to be enough to see what objcopy
>> is doing.
>
> Definitely, here they are:
>
> <http://scratch.gpio.dk/Fcode-test-results.zip>
>
> -I've also included the .map, the .dis and the .bin files, just in case
> they can provide additional information.
> I just made a .diff on the .dis files; it seems that the label of the
> interrupt routine differs.
> In one file, it's named _edata, in the other, it's named _sfastcode.
> -So after all, there might be a difference in the .elf file too...
>
> Fastcode-test1.dis:
> Disassembly of section .fastcode:
>
> 10000000 <_edata>:
> 10000000:	6001      	str	r1, [r0, #0]
> 10000002:	f241 2034 	movw	r0, #4660	; 0x1234
> 10000006:	f2c5 6078 	movt	r0, #22136	; 0x5678
> 1000000a:	4770      	bx	lr
>
> Fastcode-test2.dis:
> Disassembly of section .fastcode:
>
> 10000000 <_sfastcode>:
> 10000000:	f241 2034 	movw	r0, #4660	; 0x1234
> 10000004:	f2c5 6078 	movt	r0, #22136	; 0x5678
> 10000008:	4770      	bx	lr
> 	...
>
> ...It may just be a side-effect, though.

Thanks, I'll take a look (might not be till the weekend).

Richard