Description
PR #152 adds reservation-aware converter overloads so HOST-target conversions can allocate from a caller-provided memory::reservation instead of double-counting capacity. However, the same overload currently accepts reservations for GPU and DISK targets even though those converters ignore the reservation.
Description
PR #152 adds reservation-aware converter overloads so HOST-target conversions can allocate from a caller-provided memory::reservation instead of double-counting capacity. However, the same overload currently accepts reservations for GPU and DISK targets even though those converters ignore the reservation.