jmb
/
hardened_malloc
mirror of https://github.com/GrapheneOS/hardened_malloc.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
hardened_malloc/random.h

26 lines
731 B
C
Raw Permalink Normal View History

initial commit
2018-08-21 21:23:22 +02:00
#ifndef RANDOM_H
#define RANDOM_H
use ChaCha8-based CSPRNG
2018-08-26 05:00:00 +02:00
#include "chacha.h"
use a consistent style for fixed-size int types The stdint.h types don't cover 128-bit integers and the underscore makes them ill suited to usage in function suffixes. Instead, use the common naming style in the Linux kernel and elsewhere including the ChaCha8 implementation included here.
2018-10-04 20:25:16 +02:00
#include "util.h"
use ChaCha8-based CSPRNG
2018-08-26 05:00:00 +02:00
use smaller integers for random cache state
2018-11-04 20:23:12 +01:00
#define RANDOM_CACHE_SIZE 256U
#define RANDOM_RESEED_SIZE (256U * 1024)
implement a cache for the CSPRNG
2018-08-23 22:30:44 +02:00
initial commit
2018-08-21 21:23:22 +02:00
struct random_state {
use smaller integers for random cache state
2018-11-04 20:23:12 +01:00
unsigned index;
unsigned reseed;
use ChaCha8-based CSPRNG
2018-08-26 05:00:00 +02:00
chacha_ctx ctx;
use a consistent style for fixed-size int types The stdint.h types don't cover 128-bit integers and the underscore makes them ill suited to usage in function suffixes. Instead, use the common naming style in the Linux kernel and elsewhere including the ChaCha8 implementation included here.
2018-10-04 20:25:16 +02:00
u8 cache[RANDOM_CACHE_SIZE];
initial commit
2018-08-21 21:23:22 +02:00
};
void random_state_init(struct random_state *state);
initialize size class CSPRNGs from init CSPRNG This avoids making a huge number of getrandom system calls during initialization. The init CSPRNG is unmapped before initialization finishes and these are still reseeded from the OS. The purpose of the independent CSPRNGs is simply to avoid the massive performance hit of synchronization and there's no harm in doing it this way. Keeping around the init CSPRNG and reseeding from it would defeat the purpose of reseeding, and it isn't a measurable performance issue since it can just be tuned to reseed less often.
2019-04-15 12:29:57 +02:00
void random_state_init_from_random_state(struct random_state *state, struct random_state *source);
add generic get_random_bytes function
2019-04-15 10:52:18 +02:00
void get_random_bytes(struct random_state *state, void *buf, size_t size);
use a consistent style for fixed-size int types The stdint.h types don't cover 128-bit integers and the underscore makes them ill suited to usage in function suffixes. Instead, use the common naming style in the Linux kernel and elsewhere including the ChaCha8 implementation included here.
2018-10-04 20:25:16 +02:00
u16 get_random_u16(struct random_state *state);
u16 get_random_u16_uniform(struct random_state *state, u16 bound);
u64 get_random_u64(struct random_state *state);
u64 get_random_u64_uniform(struct random_state *state, u64 bound);
initial commit
2018-08-21 21:23:22 +02:00
#endif