> > But I can also hide the cfun->function_frequency trick in
> > DEFAULT_BRANCH_COST macro if it seems to help. (in longer term I hope
> > they will all go away as expansion needs to be aware of hotness info
> > anyway)
>
> Well, it definitly helps. I originally hoped there will be fewer places
> querying BRANCH_COST without profile info. I am testing updated patch.
>
> Honza
Grr
>
> ! if (GET_MODE_CLASS (mode) == MODE_INT && BRANCH_COST >= 2)
> {
> rtx extended = expand_shift (RSHIFT_EXPR, mode, op0,
> size_int (GET_MODE_BITSIZE (mode) - 1),
> --- 3425,3432 ----
> value of X as (((signed) x >> (W-1)) ^ x) - ((signed) x >> (W-1)),
> where W is the width of MODE. */
>
> ! if (GET_MODE_CLASS (mode) == MODE_INT
> ! && DEFAULT_BRANCH_COST)
Should be:
> ! if (GET_MODE_CLASS (mode) == MODE_INT
> ! && DEFAULT_BRANCH_COST >= 2)
Sorry for all the mess. I know I should re-read patches even after
trivial mechanical changes :( I am re-testing with this fixed.
Honza
