PuTTY .ppk v3 support

[elipsion]

• Extends putty2john with support for reading Argon fields
• Extends putty_fmt_plug to consume Argon related data
• Formatting with indent and astyle

[solardiz]

Thank you!

We need sample input file(s) to test this with, and to be able to re-test it later. Can you please contribute some to https://github.com/openwall/john-samples/tree/main/PuTTY?

The format should start reporting tunable costs - algorithm and its parameters. Looks like up to 4 total in this case.

For hex parsing, we should probably use shared code from common.[ch]. Some other *2john programs depend on these files. In Makefile.in, we have:

../run/hccap2john@EXE_EXT@: hccap2john.o common.o jumbo.o
        $(LD) $(LDFLAGS) @PTHREAD_CFLAGS@ @PTHREAD_LIBS@ common.o hccap2john.o jumbo.o @OPENMP_CFLAGS@ -o $@

../run/putty2john@EXE_EXT@: putty2john.o jumbo.o
        $(LD) $(LDFLAGS) @PTHREAD_CFLAGS@ @PTHREAD_LIBS@ putty2john.o jumbo.o @OPENMP_CFLAGS@ -o $@

so this is where the extra dependency will go.

Please also add a doc/NEWS item to the end of the post-1.9.0 list, currently around line 490.

[elipsion]

Thanks for the feedback! Addressed in this PR and john-samples#47

[solardiz]

Thank you! This is moving in the right direction, but I think there are a few more things to fix here. Also, please roll all of the review feedback changes into your first commit (and force-push), so that we can preserve your commits as they are when merging this yet not have the PR-temporary review/change history stay in the project (such micro-history is of no long-term relevance).

[solardiz]

The "32/" part becomes wrong for Argon2, which can use more of the word width... or actually also SIMD. Need to check how we report this in other formats where Argon2 is also optional and do it consistently.

Should also list SHA256.

[elipsion]

I feel like this is above my pay grade. Tell me what to type and I'll drop the letters.

[solardiz]

Great, but we also need to be able to request realistic benchmark if possible. What are the normal parameters?

We may also want to have 2+ test vectors per KDF type, so that things such as salt switching are tested. Those extra tests may in fact use weaker settings.

[elipsion]

The standard options are argon2id-8192-13-1 which (at least on my machine) takes several minutes to establish a performance baseline before starting useful hashing operations.

[solardiz]

Oh, that's probably because this format also supports the faster varieties, for which we need high keys per crypt. And it does not "establish a performance baseline", but rather tests many of the key indices for proper operation. However, I am puzzled - you seem to also have a test vector with "8192131" at the end, doesn't it cause the same problem?

[elipsion]

My bad, it was left in when collecting the timing info.

brick:/mnt/c/Users/elips/source/john# time run/john hashes.txt
Using default input encoding: UTF-8
Loaded 10 password hashes with 8 different salts (1.3x same-salt boost) (PuTTY, Private Key (RSA/DSA/ECDSA/ED25519) [SHA1/AES/Argon2 32/64])
Loaded hashes with cost 1 (type [0:none 1:Argon2d 2:Argon2i 3:Argon2id]) varying from 1 to 3
Loaded hashes with cost 2 (t) varying from 1 to 13
Loaded hashes with cost 3 (m) varying from 8 to 8192
Cost 4 (p) is 1 for all loaded hashes
Will run 14 OpenMP threads
Note: Passwords longer than 10 [worst case UTF-8] to 32 [ASCII] rejected
Proceeding with single, rules:Single
Press Ctrl-C to abort, or send SIGUSR1 to john process for status
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
10g 0:00:00:00 DONE 1/3 (2026-08-13 07:17) 62.50g/s 700.0p/s 700.0c/s 875.0C/s openwall..openwall123
Use the "--show" option to display all of the cracked passwords reliably
Session completed
real    0m 13.58s
user    0m 28.14s
sys     0m 35.61s

This is the timing with that vector in place:

brick:/mnt/c/Users/elips/source/john# time run/john hashes.txt
Using default input encoding: UTF-8
Loaded 10 password hashes with 8 different salts (1.3x same-salt boost) (PuTTY, Private Key (RSA/DSA/ECDSA/ED25519) [SHA1/AES/Argon2 32/64])
Loaded hashes with cost 1 (type [0:none 1:Argon2d 2:Argon2i 3:Argon2id]) varying from 1 to 3
Loaded hashes with cost 2 (t) varying from 1 to 13
Loaded hashes with cost 3 (m) varying from 8 to 8192
Cost 4 (p) is 1 for all loaded hashes
Will run 14 OpenMP threads
Note: Passwords longer than 10 [worst case UTF-8] to 32 [ASCII] rejected
Proceeding with single, rules:Single
Press Ctrl-C to abort, or send SIGUSR1 to john process for status
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
openwall         (openwall)
10g 0:00:00:00 DONE 1/3 (2026-08-13 07:21) 45.45g/s 509.1p/s 509.1c/s 636.4C/s openwall..openwall123
Use the "--show" option to display all of the cracked passwords reliably
Session completed
real    3m 42.44s
user    44m 21.71s
sys     3m 29.30s
brick:/mnt/c/Users/elips/source/john#

[solardiz]

please roll all of the review feedback changes into your first commit (and force-push), so that we can preserve your commits as they are when merging this yet not have the PR-temporary review/change history stay in the project (such micro-history is of no long-term relevance).

Actually, given that you also reformatted the code I understand it may be significant effort for you to merge changes into the earlier commit. Nevermind. We can accept not so clean commits history this time.

[solardiz]

@elipsion Are you going to address my feedback here, or should we merge this as-is and then fix it? Also, there's a merge conflict here for doc/NEWS with another PR we merged first (since that one became fully ready first).

[elipsion]

Sorry for leaving you hanging. I vent on vacation and promptly forgot about this.

Will give it a shot when I'm back.