2011/1/7 Chung-Lin Tang <clt...@codesourcery.com>:
> I analyzed this testcase regression a while earlier; the direct cause of
> this is due to mips_order_regs_for_local_alloc(), which now serves as
> MIPS' ADJUST_REG_ALLOC_ORDER macro.
>
> The mips_order_regs_for_local_alloc() function seems to be written for
> the old local-alloc.c, still left as the deprecated
> ORDER_REGS_FOR_LOCAL_ALLOC macro after the transition to IRA (actually
> not called at all during then), and relatively recently 'revived' after
> a patch by Bernd that created the ADJUST_REG_ALLOC_ORDER macro went in.
>
> So you have a local-alloc.c heuristic working in IRA, which seemed to
> cause these regressions.
>
> Removing mips_order_regs_for_local_alloc() will let this testcase pass;
> of course the real fix should be to review the MIPS reg-ordering logic,
> left for you MIPS people...
>
> Chung-Lin
>
As I can see, mips_order_regs_for_local_alloc() is only used to
reorder $24 (T_REG) for MIPS16. Since current definition of
REG_ALLOC_ORDER for IRA is not {0,1,2,...} any more, the old loop
code,
for (i = 0; i < FIRST_PSEUDO_REGISTER; i++)
reg_alloc_order[i] = i;
in mips_order_regs_for_local_alloc() seems inadequate. Because this
will override the definition of REG_ALLOC_ORDER.
Anyway, when I tried to rewrite the function to keep the register
order, the testcase is passed, but I can't see the speedup on my port.
Thanks,
Mingjie
