Skip to contentSkip to conversation

random: use memmove instead of memcpy for remaining 32 bytes

PATCH In review Jason A. Donenfeld 3 participants 4 messages Apr 13, 2022 – Apr 18, 2022 Random number driver

lore
Share & tools
lore

Jason A. Donenfeld

author maintainer
#0 lore raw

PATCH 1/1 random: use memmove instead of memcpy for remaining 32 bytes

In order to immediately overwrite the old key on the stack, before servicing a userspace request for bytes, we use the remaining 32 bytes of block 0 as the key. This means moving indices 8,9,a,b,c,d,e,f -> 4,5,6,7,8,9,a,b. Since 4 < 8, for the kernel implementations of memcpy(), this doesn't actually appear to be a problem in practice. But relying on that characteristic seems a bit brittle. So let's change that to a proper memmove(), which is the by-the-books way of handling overlapping memory copies.

---
 drivers/char/random.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/random.c +1 -1
index 6b01b2be9dd4..3a293f919af9 100644
@@ -333,7 +333,7 @@ static void crng_fast_key_erasure(u8 key[CHACHA_KEY_SIZE],
 	chacha20_block(chacha_state, first_block);
 
 	memcpy(key, first_block, CHACHA_KEY_SIZE);
-	memcpy(random_data, first_block + CHACHA_KEY_SIZE, random_data_len);
+	memmove(random_data, first_block + CHACHA_KEY_SIZE, random_data_len);
 	memzero_explicit(first_block, sizeof(first_block));
 }
 
Signature · 1 line
2.35.1
Reviewed-by on v1 1/1
#1 lore raw

Am Thu, Apr 14, 2022 at 01:56:49AM +0200 schrieb Jason A. Donenfeld:

Quoted 11 lines

In order to immediately overwrite the old key on the stack, before servicing a userspace request for bytes, we use the remaining 32 bytes of block 0 as the key. This means moving indices 8,9,a,b,c,d,e,f -> 4,5,6,7,8,9,a,b. Since 4 < 8, for the kernel implementations of memcpy(), this doesn't actually appear to be a problem in practice. But relying on that characteristic seems a bit brittle. So let's change that to a proper memmove(), which is the by-the-books way of handling overlapping memory copies.

Thanks,
Dominik

#2 lore raw
In reply to Jason A. Donenfeld
Quoted 20 lines · Jason A. Donenfeld

In order to immediately overwrite the old key on the stack, before servicing a userspace request for bytes, we use the remaining 32 bytes of block 0 as the key. This means moving indices 8,9,a,b,c,d,e,f -> 4,5,6,7,8,9,a,b. Since 4 < 8, for the kernel implementations of memcpy(), this doesn't actually appear to be a problem in practice. But relying on that characteristic seems a bit brittle. So let's change that to a proper memmove(), which is the by-the-books way of handling overlapping memory copies.

---
 drivers/char/random.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/random.c +0 -0
index 6b01b2be9dd4..3a293f919af9 100644

drivers/char/random.c static void crng_fast_key_erasure(u8 key[CHACHA_KEY_SIZE], · line 338

Quoted 7 lines
drivers/char/random.c +1 -1
@@ -333,7 +333,7 @@ static void crng_fast_key_erasure(u8 key[CHACHA_KEY_SIZE],
 	chacha20_block(chacha_state, first_block);
 
 	memcpy(key, first_block, CHACHA_KEY_SIZE);
-	memcpy(random_data, first_block + CHACHA_KEY_SIZE, random_data_len);
+	memmove(random_data, first_block + CHACHA_KEY_SIZE, random_data_len);
 	memzero_explicit(first_block, sizeof(first_block));
 }

first_block is on the stack, so this is never an overlapping copy.

It would be more important to document the fact that random_data can point into chacha_state, as this is not obvious.

- Eric

author maintainer
#3 lore raw

Hey Eric,

Quoted 23 lines · Eric Biggers
Quoted 22 lines · Jason A. Donenfeld

In order to immediately overwrite the old key on the stack, before servicing a userspace request for bytes, we use the remaining 32 bytes of block 0 as the key. This means moving indices 8,9,a,b,c,d,e,f -> 4,5,6,7,8,9,a,b. Since 4 < 8, for the kernel implementations of memcpy(), this doesn't actually appear to be a problem in practice. But relying on that characteristic seems a bit brittle. So let's change that to a proper memmove(), which is the by-the-books way of handling overlapping memory copies.

---
 drivers/char/random.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/random.c +0 -0
index 6b01b2be9dd4..3a293f919af9 100644
@@ -333,7 +333,7 @@ static void crng_fast_key_erasure(u8 key[CHACHA_KEY_SIZE],
 	chacha20_block(chacha_state, first_block);
 
Quoted 7 lines

drivers/char/random.c static void crng_fast_key_erasure(u8 key[CHACHA_KEY_SIZE], · line 338

Quoted 5 lines
drivers/char/random.c +1 -1
 	memcpy(key, first_block, CHACHA_KEY_SIZE);
-	memcpy(random_data, first_block + CHACHA_KEY_SIZE, random_data_len);
+	memmove(random_data, first_block + CHACHA_KEY_SIZE, random_data_len);
 	memzero_explicit(first_block, sizeof(first_block));
 }

first_block is on the stack, so this is never an overlapping copy.

Oh, grrr, yes you're right. I clearly over thought this into non-sense. Will revert.

Quoted 3 lines

It would be more important to document the fact that random_data can point into chacha_state, as this is not obvious.

Good idea. I'll do that.

Jason