On Wed, 10 Jan 2018, David Gibson wrote:
> On Tue, Jan 09, 2018 at 01:27:48PM +0100, Julia Lawall wrote:
> > > > +struct srcpos *
> > > > +srcpos_copy_all(struct srcpos *pos)
> > > > +{
> > > > + struct srcpos *pos_new;
> > > > + struct srcfile_state *srcfile_state;
> > > > +
> > > > + if (!pos)
> > > > + return NULL;
> > > > +
> > > > + pos_new = srcpos_copy(pos);
> > > > +
> > > > + if (pos_new) {
> > > > + /* allocate without free */
> > > > + srcfile_state = xmalloc(sizeof(struct srcfile_state));
> > > > + memcpy(srcfile_state, pos->file, sizeof(struct srcfile_state));
> > > > +
> > > > + pos_new->file = srcfile_state;
> > > > + }
> > > > +
> > > > + return pos_new;
> > > > +}
> > >
> > > I don't really see a reason we'd need both a deep and a shallow copy.
> > > If you need a deep copy, I'd suggest just changing srcpos_copy() to do
> > > that.
> >
> > The deep copy is needed due to the treatment of #includes. The following
> > function overwrites the file name information:
> >
> > void srcpos_set_line(char *f, int l)
> > {
> > current_srcfile->name = f;
> > current_srcfile->lineno = l;
> > }
> >
> > srcpos_combine doesn't need to use the deep copy, because the result gets
> > copied again. Maybe this is not a big issue or there is a better way to
> > solve this.
>
> Well, dtbs are generally so small that performance really isn't an
> issue for dtc. So I think simplifying the code to only have one
> (deep) copy routine is worth more than the minor speedup from avoiding
> deep copies in some places they're not necessary.
OK, I will do this.
julia
--
To unsubscribe from this list: send the line "unsubscribe devicetree-compiler" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
