This avoids leaving behind an empty file when an error occurs before we
write the header (for example, because the passphrase is invalid). Do a
best-effort check before taking user input for whether the file exists
so we don't waste user effort. An error might still happen after user
input if other kind of open errors happen (for example, a permission
issue, or disk full).
Fixes#159Fixes#57Closes#169
Most writes in the cmd/age Writer stack are chunk-sized, so
approximately 64KiB. However, the newlineWriter, which splits lines at
64 columns, was doing a Write on the underlying Writer for each line,
making chunks effectively 48 bytes (before base64). There is no
buffering underneath it, so it was resulting in a lot of write syscalls.
Add a reusable bytes.Buffer to buffer the output of each
(*newlineWriter).Write call, and Write it all at once on the
destination.
This makes --armor just 50% slower than plain, instead of 10x.
Fixes#167
It's possible to craft ChaCha20Poly1305 ciphertexts that decrypt under
multiple keys. (I know, it's wild.)
The impact is different for different recipients, but in general only
applies to Chosen Ciphertext Attacks against online decryption oracles:
* With the scrypt recipient, it lets the attacker make a recipient
stanza that decrypts with multiple passwords, speeding up a bruteforce
in terms of oracle queries (but not scrypt work, which can be
precomputed) to logN by binary search.
Limiting the ciphertext size limits the keys to two, which makes this
acceptable: it's a loss of only one bit of security in a scenario
(online decryption oracles) that is not recommended.
* With the X25519 recipient, it lets the attacker search for accepted
public keys without using multiple recipient stanzas in the message.
That lets the attacker bypass the 20 recipients limit (which was not
actually intended to defend against deanonymization attacks).
This is not really in the threat model for age: we make no attempt to
provide anonymity in an online CCA scenario.
Anyway, limiting the keys to two by enforcing short ciphertexts
mitigates the attack: it only lets the attacker test 40 keys per
message instead of 20.
* With the ssh-ed25519 recipient, the attack should be irrelevant, since
the recipient stanza includes a 32-bit hash of the public key, making
it decidedly not anonymous.
Also to avoid breaking the abstraction in the agessh package, we don't
mitigate the attack for this recipient, but we document the lack of
anonymity.
This was reported by Paul Grubbs in the context of the upcoming paper
"Partitioning Oracle Attacks", USENIX Security 2021 (to appear), by
Julia Len, Paul Grubbs, and Thomas Ristenpart.
Move the SSH recipient types out of the main package to declutter the
godoc. This also allows us to drop the x/crypto/ssh build dependency
entirely from the age package import tree.
The caller can take care of the armor. For consistency move the
responsibility to close the armor to the caller, and make the stream
Writer not propagate Close.
This also will also allow us to spin the armor implementation out into
its won package that imports format, without getting an import loop from
format.Parse magically invoking armor decoding.
Less magic in the API, more magic in the CLI.
