Re: [Libjit-developers] Clobbered stack for jit-interp.c

[Gopal V]

Gopal V wrote:

>         return_reg(i9, r0)

As it turns out, I was really stupid there. The i9
was being spilled into location #0 in the code about
7 lines later.

But the real problem is that the register spill occurs
after the brfalse which got me thinking. So let me
clarify my understanding of the code, every jump
or jump offset is supposed to generate a basic block
of type jit_block_t ?

If that is the case, then the spill should actually
happen before the brfalse ?

Or am I completely off the map, this time as well ?

Cheers,
Gopal
--
If I learn from my mistakes, pretty soon I'll know everything