AES-XTS : Tweak Value Increment when dataUnitLenMatchesPayload == false

[afazio2]

environment

Demo

testSessionId

464670

vsId

2025897

Algorithm registration

[
  {
    "acvVersion":"1.0"
  },
  {
    "isSample":true,
    "operation":"register",
    "certificateRequest":"no",
    "debugRequest":"yes",
    "production":"no",
    "encryptAtRest":"yes",
    "algorithms":[
      {
        "revision":"2.0",
        "algorithm":"ACVP-AES-XTS",
        "prereqVals":[
          {
            "algorithm":"AES",
            "valValue":"same"
          },
          {
            "algorithm":"DRBG",
            "valValue":"same"
          }
        ],
        "direction":[
          "encrypt",
          "decrypt"
        ],
        "keyLen":[
          128,
          256
        ],
        "payloadLen":[
          {
            "min":128,
            "max":65536,
            "increment":128
          }
        ],
        "tweakMode":[
          "hex"
        ],
        "dataUnitLen":[
          {
            "min":1024,
            "max":65536,
            "increment":128
          }
        ],
        "dataUnitLenMatchesPayload":false
      },
      ...
    ]
  }
]

Expected behavior

Process the vectors and obtain the same results as the expected answers.

Actual behavior

Vectors results differ from the expected answers when dataUnitLen differs from payloadLen .

We kindly ask to confirm the ACVP server logic behind the tweak value increment when dataUnitLen can differ from payloadLen. We provide below an example:

Data Unit	IUT Tweak Value Increment (little endian)	ACVP Server Tweak Value Increment (little endian)
Data Unit #1	DF60B25ABDEC7163C531DBFF858C85FF	DF60B25ABDEC7163C531DBFF858C85FF
Data Unit #2	E060B25ABDEC7163C531DBFF858C85FF	E060B25ABDEC7163C531DBFF858C8500
Data Unit #3	E160B25ABDEC7163C531DBFF858C85FF	E160B25ABDEC7163C531DBFF858C8500
Data Unit #4	E260B25ABDEC7163C531DBFF858C85FF	E260B25ABDEC7163C531DBFF858C8500
Data Unit #5	E360B25ABDEC7163C531DBFF858C85FF	E360B25ABDEC7163C531DBFF858C8500

It seems there's a similar issue #1475

[k-c-w]

When would a fix for this issue be rolled out?

[jbrock24]

Currently working on it, a patch would be 2-3 weeks out from going to demo. I'll reply here earlier than that to let you know when it's done and when the expected release is going to be.

[jbrock24]

Hi everyone, I have found the issue out and will be implementing a fix that will go out with the next release. Sorry for the inconvenience, and we really appreciate all the helpful info, thanks again. Once the release is out, we'll comment here that it's ready for testing. It will move to prod a week or two later.

[livebe01]

The fix for this is on Demo in release v1.1.0.33

[livebe01]

The fix for this is on Prod in release v1.1.0.33