Re: [PATCH v2 02/19] gendwarfksyms: Add symbol list handling

[Sami Tolvanen <samitolvanen@google.com>]

On Thu, Aug 29, 2024 at 03:16:21AM +0900, Masahiro Yamada wrote:
> On Fri, Aug 16, 2024 at 2:39 AM Sami Tolvanen <samitolvanen@google.com> wrote:
> > @@ -105,6 +105,8 @@ int main(int argc, const char **argv)
> >         if (parse_options(argc, argv) < 0)
> >                 return usage();
> >
> > +       check(symbol_read_exports(stdin));
> 
> 
> 
> symbol_read_exports() is only called from main().
> 
> Do you need to make symbol_read_exports() return
> the error code all the way back to the main()
> function?
> 
> Personally, I'd like to make the program bail out as early as
> possible if there is no point in continuing running.

That's a valid point. The current error handling prints out a short
trace of exactly where something failed as the error propagates
through the call stack, but bailing out after printing the first
error is probably informative enough. I'll look into cleaning this
up.

> See also this patchset.
> 
> https://lore.kernel.org/linux-kbuild/20240812124858.2107328-1-masahiroy@kernel.org/T/#m5c0f795b57588a2c313cd2cc6e24ac95169fd225

Thanks for the link. In general I prefer to print out an error to
indicate what went wrong, but I suppose memory allocation errors
should be rare enough that it's not necessary. I'll switch to these
in the next version.

> > +int symbol_read_exports(FILE *file)
> > +{
> > +       struct symbol *sym;
> > +       char *line = NULL;
> > +       char *name = NULL;
> > +       size_t size = 0;
> > +       int nsym = 0;
> > +
> > +       while (getline(&line, &size, file) > 0) {
> > +               if (sscanf(line, "%ms\n", &name) != 1) {
> > +                       error("malformed input line: %s", line);
> > +                       return -1;
> > +               }
> > +
> > +               free(line);
> > +               line = NULL;
> > +
> > +               if (is_exported(name))
> > +                       continue; /* Ignore duplicates */
> > +
> > +               sym = malloc(sizeof(struct symbol));
> > +               if (!sym) {
> > +                       error("malloc failed");
> > +                       return -1;
> > +               }
> > +
> > +               sym->name = name;
> > +               name = NULL;
> 
> Is this necessary?

Here, no, but in Petr's cleaned up version it is again necessary, so
you'll see this in v3 still.

Sami