On Wed, 27 Jan 2021 12:02:06 GMT, Aleksey Shipilev <sh...@openjdk.org> wrote:
>> Then we would free as well for the realloc code path >> } else { >> /* do we have enough space? */ >> ... >> } >> Is this okay? > > Yes, I think that is better: `realloc` can fail as well, and we don't want to > leak those? Please also reset the free'd pointers to NULL in GPData* before returning. ------------- PR: https://git.openjdk.java.net/jdk/pull/2250