> This part of the patch series fixes problems with bad code being emitted > for unaligned bitfield accesses, as reported in PRs 48784, 56341, and > 56997. A secondary goal of this patch was making the bitfield store and > extract code follow similar logic, at least for the parts relating to > -fstrict-volatile-bitfield handling.
Thanks for tackling this. I think we all agree that making the store and extract paths follow a similar logic would be a step forward. Some remarks: - the extract_fixed_bit_field changes look good to me (nice cleanup!) modulo s/type/mode/ in the added comment, - the store_fixed_bit_field changes are less immediate, as the fallback isn't the default behavior (contrary to what the comment says). I presume that this is intended? - the extract_bit_field_1 changes look good to me (modulo a superfluous "of the field" in the first added comment) but I don't really understand the ChangeLog entry; I'd just write "do not short-circuit direct extraction for flag_strict_volatile_bitfields handling" or something along these lines. -- Eric Botcazou
