On 10/22/20 2:08 PM, Peter Maydell wrote: > Commit ef96e3ae9698d6 in January 2019 removed the last user of the > VMSTATE_FLOAT64* macros. These were used by targets which defined > their floating point register file as an array of 'float64'.
Similar candidate: VMSTATE_CPUDOUBLE_ARRAY() > We used to try to maintain a stricter distinction between > 'float64' (a type for holding an integer representing an IEEE float) > and 'uint64_t', including having a debug option for 'float64' being > a struct and supposedly mandatory macros for converting between > float64 and uint64_t. We no longer think that's a usefully > strong distinction to draw and we allow ourselves to freely > assume that float64 really is just a 64-bit integer type, so > for new targets we would simply recommend use of the uint64_t type > for a floating point register file. The float64 type remains > as a useful way of documenting in the type signature of helper > functions and the like that they expect to receive an IEEE float > from the TCG generated code rather than an arbitrary integer. > > Since the VMSTATE_FLOAT64* macros have no remaining users and > we don't recommend new code uses them, delete them. > > Signed-off-by: Peter Maydell <peter.mayd...@linaro.org> > --- > include/migration/vmstate.h | 13 ------------- > migration/vmstate-types.c | 26 -------------------------- > 2 files changed, 39 deletions(-)
