Re: [Libjit-developers] Patch to handle constants in jit-insn.c

[Gopal V]

Avinash Atreya wrote:

> On a brighter note, it does seem like I have a start on arrays for dpas. 
> Multidimensional arrays are also working. Need a little bit more work on 
> that ( bounds checking and all that). Should hopefully have a patch soon.

Trying to debug the segfault there.

a[2] := 10;

gets converted to -

i3 = 2
i14 = &s1
i16 = i3 * 4
i17 = i14 + i16
store_relative_int(i17, 10, 0)

So far so good ... But when run in the interpreter
this screws up badly.

From what I can see it is doing.

stacktop[0].int_value = 2;
stacktop--;
stacktop[0].ptr_value = <ptr>;
stacktop--;
stacktop[0].int_value = 4;
stacktop--;
stacktop[1].int_value = stacktop[0].int_value * stacktop[1].int_value;
stacktop++;
stacktop[1].int_value = stacktop[0].int_value + stacktop[1].int_value;
stacktop++;
*stacktop[0].ptr_value = 10
stacktop++;

The problem seems to be that it is doing

addr = (base_addr * size) + offset;

by doing the above operations. I cannot seem to find a real way in which
this can be fixed properly (after all both the modes should work
exactly the same in libjit). Any ideas ?

http://pastebin.com/465192

Cheers,
Gopal