Conversation
Will take a look.
This is inevitable, and we decided to not prematurely optimize any allocations before getting evidence that there's a performance regression.
1
Which is fine, but will there be evidence presented for anything if no one is looking for it? At least providing a flag to avoid it seems to be bare minimum, and it makes no sense e.g. for the ~1500 kmemdup/kstrdup callers (and other similar APIs) to be clearing before replacing
1
Show replies