Conversation
66dfb26 to
34137a7
Compare
|
Close #20 |
|
What's the ETA on this? |
34137a7 to
ab526e4
Compare
|
Only needs review |
robin-nitrokey
left a comment
There was a problem hiding this comment.
I think it is counter-intuitive that delete_all_pins also affects the pin-less application key. We should either rename the syscall or change it to only reset the PINs and introduce a separate syscall to clear all data.
|
Ok, I'll make
|
|
Hurh, trussed's filestore doesn't have a |
|
And now I'm finding bugs in the littlefs bindings and trussed |
|
See trussed-dev/littlefs2#36 (review), which is required to properly implement But this also mean we will have to have a new release of littlefs2 and merge trussed-dev/trussed#96 to benefit from it |
|
We’re still using a patched littlefs2 in nitrokey-3-firmware so we should be able to cherry-pick the fix for our fork: https://github.com/Nitrokey/littlefs2 |
|
@sosthene-nitrokey @robin-nitrokey |
|
Just tested and now it works, without changing anything additionally, specifically |
This method is meant to encrypt data stored on the external flash so that it can't be accessed just but plugging into it.
Multiple keys can be obtained with an
infoparameter.This adds one step after the
get_app_keystep described in #10 to add a per-application salt and an info parameter:HMAC(application_key, application_salt || len(info) || info).With the
application_keybeing the result ofget_app_keyand the salt being a per-application salt, that is deleted withdelete_all_pins, so that the keys change.