Hi, On Thu, 13 Jul 2017, Martin Liška wrote:
> Hopefully following patch will fix that. I returned to the first version > and saved/restored static_chain register before/after > __asan_stack_malloc. It should also work if you emit the parm_birth_note after the static chain is set up (not before it), but before you store into the nonlocal_goto_save_area. With that you don't need to worry about clobbering the incoming static chain with the asan setup. Can you test that? It would better reflect the intent of this note (the static chain being an implicit parameter, but the nonlocal_goto_save_area setup not being such). Ciao, Michael.
