On Wed, Jan 15, 2014 at 8:23 AM, Richard Henderson <r...@redhat.com> wrote: > IMO the test case is invalid, but as always an ICE is unacceptable. > > I can imagine there are some legitimate uses for absconding with > the hard frame pointer, under -fomit-frame-pointer plus knowledge > that the code is simple enough. > > Previously when we got into this situation, we'd silently smash the > hard frame pointer. The user would find out the problem only through > testing. But the dwarf2cfi pass has consistency checks that trigger. > > This patch preserves the "simple enough" case while generating an > error early enough that the dwarf2cfi pass doesn't run. > > Tested on i386 and ppc64. >
We should add a testcase to verify this. -- H.J.
