> On Jan 30, 2017, at 7:35 PM, Andrew Pinski <apin...@cavium.com> wrote: > > On Mon, Jan 30, 2017 at 3:48 AM, Maxim Kuvyrkov > <maxim.kuvyr...@linaro.org> wrote: >> This patch port prefetch configuration from aarch32 backend to aarch64. >> There is no code-generation change from this patch. >> >> This patch also happens to address Kyrill's comment on Andrew's prefetching >> patch at https://gcc.gnu.org/ml/gcc-patches/2017-01/msg02133.html . >> >> This patch also fixes a minor bug in aarch64_override_options_internal(), >> which used "selected_cpu->tune" instead of "aarch64_tune_params". > > I am not a fan of the macro at all.
Are you referring to the AARCH64_PREFETCH_NON_BENEFICIAL macro? If so, then consider consider how it simplifies updates to prefetch tuning parameters for targets that don't care about prefetching. My guess is that growing number of ARM cores was the reason for this macro in aarch32 backend. Similarly, the number of aarch64 cores will only increase over time, and updating every single tuning structure will become a pain. Here is an updated patch that fixes references from global_options.foo to opts->foo. I spotted this mistake while looking at Andrew's patch. -- Maxim Kuvyrkov www.linaro.org
0004-Port-prefetch-configuration-from-aarch32-to-aarch64-.patch
Description: Binary data
