Re: Strange symbols_lookup() behaviour in test-symbols on arm64 CI

[Jason Andryuk <jason.andryuk@xxxxxxx>]

On 2026-01-15 11:57, Jan Beulich wrote:
> On 15.01.2026 16:33, Jan Beulich wrote:
> > On 14.01.2026 07:00, Mykola Kvach wrote:
> > > On Mon, Dec 15, 2025 at 1:27 PM Jan Beulich <jbeulich@xxxxxxxx> wrote:
> > > > On 15.12.2025 12:00, Mykola Kvach wrote:
> > > > > On Thu, Dec 11, 2025 at 6:40 PM Jan Beulich <jbeulich@xxxxxxxx> wrote:
> > > > > > On 11.12.2025 17:30, Mykola Kvach wrote:
> > > > > > > I have now attached the corresponding build log.
> > > > > > Okay, so indeed not a table size change issue here. Then I fear some 
> > > > > > instrumenting
> > > > > > will be needed to at least know what exactly is going wrong. Alternatively you 
> > > > > > could
> > > > > > arrange for the intermediate binaries to not be deleted, and make them available
> > > > > > somehow / somewhere for me to see whether by inspection I can gain some clue.
> > > > > I prepared a small patch to keep the intermediate artifacts instead of
> > > > > deleting them.
> > > > >
> > > > > It removes two cleanup commands:
> > > > >      xen/arch/arm/Makefile: drops rm -f $(@D)/.$(@F).[0-9]* (keeps
> > > > > .xen-syms.* intermediates)
> > > > This alone should be sufficient.
> > > Understood. I have rerun the build with the cleanup line removed
> > > so the intermediate .xen-syms.* files are kept.
> > >
> > > The build artifacts are available here:
> > > https://gitlab.com/xen-project/people/mykola_kvach/xen/-/jobs/12707528457/artifacts/browse/xen/
> > Apart from the intermediate files there's a file named xen there, but xen-syms
> > is missing. I'll see what I can do without that.
> Actually, can you give the patch below a try? That would explain the 24-byte
> difference (albeit I'm struggling with some other aspects of a proper
> explanation).
I had qemu-smoke-dom0less-arm32-gcc-debug-staticmem fail in CI with:
test_symbols: non-zero offset (0x48c) unexpected

https://gitlab.com/xen-project/people/jandryuk-amd/xen/-/jobs/12787541697

With your patch added on top, it passed:
https://gitlab.com/xen-project/people/jandryuk-amd/xen/-/jobs/12788189122

Tested-by: Jason Andryuk <jason.andryuk@xxxxxxx>

The branch includes --gc-sections and associated changes, fwiw.

Thanks,
Jason

> --- a/xen/arch/arm/Makefile
> +++ b/xen/arch/arm/Makefile
> @@ -87,13 +87,13 @@ endif
>   $(TARGET)-syms: $(objtree)/prelink.o $(obj)/xen.lds
>         $(objtree)/tools/symbols $(all_symbols) --empty > $(dot-target).0.S
>         $(MAKE) $(build)=$(@D) $(dot-target).0.o
> -       $(LD) $(XEN_LDFLAGS) -T $(obj)/xen.lds $< \
> +       $(LD) $(XEN_LDFLAGS) -T $(obj)/xen.lds $< $(build_id_linker) \
>               $(dot-target).0.o -o $(dot-target).0
>         $(NM) -pa --format=sysv $(dot-target).0 \
>                 | $(objtree)/tools/symbols $(all_symbols) --sysv --sort \
>                 > $(dot-target).1.S
>         $(MAKE) $(build)=$(@D) $(dot-target).1.o
> -       $(LD) $(XEN_LDFLAGS) -T $(obj)/xen.lds $< \
> +       $(LD) $(XEN_LDFLAGS) -T $(obj)/xen.lds $< $(build_id_linker) \
>             $(dot-target).1.o -o $(dot-target).1
>         $(NM) -pa --format=sysv $(dot-target).1 \
>                 | $(objtree)/tools/symbols $(all_symbols) --sysv --sort \