Is akcipher ready for userspace?
Hello Herbert,
back in 2019 you wrote that akcipher is still in a state of flux and not ready to be exposed to userspace via AF_ALG [1].
Has this changed since then?
I am asking because I am implementing another driver [2] for a device which allows for signing messages with an ECDSA private key securely stored inside the device, and Greg asks again [3] for this to be exposed to userspace via a dedicated kernel API, instead of debugfs.
Back in 2019 when we needed this for the turris-mox-rwtm driver, I implemented it via debugfs because akcipher was not ready.
Thanks.
Marek
[1] https://www.spinics.net/lists/linux-crypto/msg38388.html
[2] https://lore.kernel.org/soc/[email protected]/T/
[3] https://lore.kernel.org/soc/2024051042-unbuckled-barometer-1099@gregkh/
Herbert Xu
maintainerQuoted 7 lines · Marek Behún
Hello Herbert,
back in 2019 you wrote that akcipher is still in a state of flux and not ready to be exposed to userspace via
AF_ALG[1].Has this changed since then?
Quoted 8 lines
I am asking because I am implementing another driver [2] for a device which allows for signing messages with an ECDSA private key securely stored inside the device, and Greg asks again [3] for this to be exposed to userspace via a dedicated kernel API, instead of debugfs.
Back in 2019 when we needed this for the turris-mox-rwtm driver, I implemented it via debugfs because akcipher was not ready.
No I don't think akcipher is quite ready yet, given that the recent change to kernel pointers from SG lists is still incomplete.
However, akcipher algorithms are already partially exposed to user-space through the keyring subsystem.
Cheers,
Signature · 3 lines
Email: Herbert Xu <[email protected]> Home Page: http://gondor.apana.org.au/~herbert/ PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt