This is the mail archive of the
java@gcc.gnu.org
mailing list for the Java project.
Re: IA-64 non-call exceptions
Richard Henderson writes:
> On Tue, Jun 19, 2001 at 04:22:41PM +0100, Andrew Haley wrote:
> > The biggest issue is that (unlike the DWARF 2 unwinder) the IA-64
> > unwinder can't cope with unwinding leaf functions. There are several
> > reasons for this, but one is that such functions don't save a copy of
> > the return link but leave it in br0; the unwinder can't cope with
> > that. In any case, the compiler/assembler outputs no unwind
> > information for leaf functions so the unwinder aborts.
>
> Hmm. I _think_ we're supposed to assume that in the absense of
> unwind information that b0 contains the return address, and that
> there is no register frame or stack frame. I'll have to look at
> the abi again...
Yeah, but 'twould be nice in that case at least to emit a basic
unwinder frame. As it is we emit zilch.
Andrew.