On Wed, Feb 19, 2020 at 3:40 AM Craig Blackmore
<craig.blackm...@embecosm.com> wrote:
> On 10/12/2019 18:28, Craig Blackmore wrote:
> Thank you for your review. I have posted an updated patch below which I think
> addresses your comments.
>
> Ping
>
> https://gcc.gnu.org/ml/gcc-patches/2019-12/msg00712.html
This looks OK. There are some minor issues.
(riscv_new_address_profitable_p): New function.
(TARGET_NEW_ADDRESS_PROFITABLE_P): Define.
These are actually in part2, and part2 already has changelog entries
for them, so these can just be dropped.
+ /* When optimizing for size, make uncompressible 32-bit addresses more
+ * expensive so that compressible 32-bit addresses are preferred. */
+ if (!speed && riscv_mshorten_memrefs && mode == SImode
+ && !riscv_compressed_lw_address_p (addr))
+ return riscv_address_insns (addr, mode, false) + 1;
I think that there should be a TARGET_RVC check here, just like in the gate
function for the new pass. But I also suspect that this probably
doesn't matter much.
> https://gcc.gnu.org/ml/gcc-patches/2019-12/msg00713.html
This also looks OK to me, but adding a new target hook is something
that should really wait for stage1 which should only be a few weeks
away now. That will also give us more time to test the patch before
it appears in a gcc release. So this is OK once stage1 opens.
Jim
