On Tue, Jul 7, 2020 at 12:48 PM Arnaldo Carvalho de Melo
<arnaldo.melo@xxxxxxxxx> wrote:
>
> Em Tue, Jul 07, 2020 at 11:54:07AM -0700, Andrii Nakryiko escreveu:
> > On Fri, Jul 3, 2020 at 1:13 AM Hao Luo <haoluo@xxxxxxxxxx> wrote:
> > > + sym_name = elf_sym__name(&sym, btfe->symtab);
> > > + if (!btf_name_valid(sym_name)) {
> > > + if (verbose)
> > > + printf("invalid symbol name '%s' for btf.\n",
> > > + sym_name);
> > > + continue;
> > > + }
> > > + name = strings__add(strings, sym_name);
> > > + type = var->ip.tag.type + type_id_off;
> > > + size = elf_sym__size(&sym);
> > > + if (!size) {
> > > + if (verbose)
> > > + printf("zero symbol size for symbol '%s'.\n",
> > > + sym_name);
> > > + continue;
> > > + }
> >
> > Let's please make all the above checks (size and name) hard errors,
> > they shouldn't happen. If they happen, we need to be aware and
> > understand why, before we start ignoring them. I think this is very
> > important.
>
> Perhaps we can make as you suggest but add a --force to allow for
> explicitely ignoring them while doing tests?
sure, that would be ok as well
>
> - Arnaldo
