From: MITSUNARI Shigeo <[email protected]> Thanks for the detailed review.
I have posted a v3 series in a new thread and addressed the points you raised: - clarified the expmed comments/wording - renamed the local flag to did_wide_mulh_path - switched the magic constant computation to uint64_t - added a non-BMI2 x86_64 testcase - added Signed-off-by lines for DCO compliance New thread: [PATCH v3 0/3] expmed/i386: optimize u32 constant division and improve BMI2 codegen
