https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81926
--- Comment #38 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot
Uni-Bielefeld.DE> ---
> --- Comment #15 from Richard Biener <rguenth at gcc dot gnu.org> ---
> Note that dwarf2out.c causes the section to be emitted via
>
> static void
> dwarf2out_assembly_start (void)
> {
> #ifndef DWARF2_LINENO_DEBUGGING_INFO
> ASM_GENERATE_INTERNAL_LABEL (text_section_label, TEXT_SECTION_LABEL, 0);
> ASM_GENERATE_INTERNAL_LABEL (text_end_label, TEXT_END_LABEL, 0);
> ASM_GENERATE_INTERNAL_LABEL (cold_text_section_label,
> COLD_TEXT_SECTION_LABEL, 0);
> ASM_GENERATE_INTERNAL_LABEL (cold_end_label, COLD_END_LABEL, 0);
>
> switch_to_section (text_section);
> ASM_OUTPUT_LABEL (asm_out_file, text_section_label);
> #endif
> ...
>
> part of the code refering to the label is guarded by a text_section_used
> flag btw., not sure if that works reliably enough but I guess it is
> only properly initialized after the above. Thus the middle-end would
> need to support sth like 'give me a label at the start of section FOO'.
>
> The easiest workaround is indeed to always emit .text (which indeed gas does).
> And I'd rather do that where the assemby_start debug hook is called instead
> of in target specific hooks.
The newly attached patch does that and worked for me on
sparc-sun-solaris2.11.
I've left the assembly_start call in vmsdbgout.c
(vmsdbgout_assembly_start) alone since I can neither test it nor does it
seem affected: it's only called for VMS_AND_DWARF2_DEBUG which isn't
problematic.
Rainer
