|
|
|
|
|
|
by DamonHD
318 days ago
|
|
|
The 100kloc thing had no malloc/free, ie most significant persistent data objects were static, and the inlining with LTO (and some help) was pretty intense, so the stack was used, but not as much as it seemed on the surface. We did (re)invent some carefully managed buffers shared through several levels of call, since we were doing, for example, AES-GCM: https://github.com/opentrv/OTAESGCM |
|
|