AES DV document

Goals

  • DV
    • Verify all AES IP features by running dynamic simulations with a SV/UVM based testbench
    • Develop and run all tests based on the testplan below towards closing code and functional coverage on the IP.
  • FPV
    • Verify TileLink device protocol compliance with an SVA based testbench

Current status

Design features

For detailed information on AES design features, please see the AES HWIP Technical Specification.

Testbench architecture

AES testbench has been constructed based on the CIP testbench architecture.

Block diagram

Block diagram

Top level testbench

Top level testbench is located at hw/ip/aes/dv/tb/tb.sv. It instantiates the AES DUT module hw/ip/aes/rtl/aes.sv. In addition, it instantiates the following interfaces, connects them to the DUT and sets their handle into uvm_config_db:

Common DV utility components

The following utilities provide generic helper tasks and functions to perform activities that are common across the project:

Global types & methods

All common types and methods defined at the package level can be found in aes_env_pkg. Some of them in use are:

parameter uint NUM_ALERTS = 2;

TL_agent

AES instantiates (already handled in CIP base env) tl_agent which provides the ability to drive and independently monitor random traffic via TL host interface into AES device.

EDN agent

AES instantiates (already handles in the CIP base env) edn_agent which provides the ability to drive and monitor edn traffic via the edn interface.

UVM RAL model

The AES RAL model is created with the ralgen FuseSoC generator script automatically when the simulation is at the build stage.

It can be created manually by invoking regtool:

Stimulus strategy

Test sequences

All test sequences reside in hw/ip/aes/dv/env/seq_lib. The aes_base_vseq virtual sequence is extended from cip_base_vseq and serves as a starting point. All test sequences are extended from aes_base_vseq. aes_base_vseq provides commonly used handles, variables, functions and tasks that the test sequences can simply use / call. The tasks can be split into two groups and those that provide more complex functionality. Simple tasks include:

  • aes_init: Initialize the AES module from the randomized environment variables in the config.
  • set_op: Set AES operation to encrypt or decrypt.
  • write_key: Write initial key to AES init key registers.
  • add_data: Add the next 128 block to the input registers.
  • read_output: Poll the status register for data ready bit and read the result from AES output registers.
  • clear_reg: Based on the input this function clears data input-, data output- or key-registers or any combination of these.
  • set_manual_trigger: Chooses between AES auto start and manual start.
  • trigger_start: Set the start bit to trigger a new encryption/decryption.

More complex tasks include: These are the ones used by the higher level sequences and the ones that should be used to create new tests from:

  • generate_message_queue: Generate a queue of randomized message items. Each item will describe the parameters of a message but not hold any data, the data will be added later. This function does not call any sub-functions.
  • send_msg_queue: Take the queue of messages items and process them one by one. Send_msg_queue converts each message item into a queue of message transactions each called an aes_items by calling generate aes_item_queue(). Then each message now described by a queue of items is processed by calling send_msg().
  • generate_aes_item_queue: Expands a message into a queue of a configuration item and N Data items where N = Message_length/block size. Data is randomized based on the constraints.
  • send_msg: Take a queue of configuration and data items and pass them to the AES for processing. This task will handle configuration of the core based on the configuration item. Then based on the test configuration it will do either a balanced or unbalanced processing of the data items. A balanced processing implies that for each input the task will wait until the resulting output have been read before attempting to write the next input. The unbalanced version will write the next data item as soon as the AES status register indicates that a new input will be accepted. In unbalanced mode the send_msg task also handles reading the output register. Knobs are available to control the balance between reads and writes. In either case the status_fsm task is called.
  • status_fsm: Read the status of the AES IP and based on the inputs return the status. When the task is in control of reading the output register it will poll the status until it indicates that the output is valid. It will also track the progress of processing, detecting if something has gone differently than expected. If this happens it will try to recover.

Using these higher level functions and tasks one can build a highly customized constraint random test without low level knowledge of the test environment and the DUT.

Most tests use the aes_stress_vseq sequence as test sequence, and achieves different tests scenarios by using the constrained knobs to generate different behavior.

Functional coverage'

To ensure high quality constrained random stimulus, it is necessary to develop a functional coverage model. The following cover groups have been developed to prove that the test intent has been adequately met.

  • aes_ctrl_cg: this cover group checks that all parts of the control register have been exercised. A few crosses are also included to ensure that we have seen all both forward and inverse for all modes. Similarly that we have tried all possible key_sizes in all modes, and that all have been tried both in auto and manual mode. More importantly for AES which is a security critical IP it also verifies that all illegal states have been seen.
  • aes_status_cg: makes sure all possible states has been seen.
  • aes_trigger_cg: verify that all parts of the trigger register was exercised. A cross also checks that we tried to clear all fields possible at once.

(WIP) coverage points to come A temporary functional coverage plan can be found here coverage_plan

Self-checking strategy

Scoreboard

The aes_scoreboard is primarily used for end to end checking. It creates the following analysis FIFOs to retrieve the data monitored by corresponding interface agents:

  • tl_a_chan_fifo: tl address channel
  • tl_d_chan_fifo: tl data channel

These 2 FIFOs provide transaction items at the end of the address channel and data channel respectively. Each FIFO is monitored and incoming transactions are stored. Whenever a transaction is finished the sequence item is handed over to a reference model that will generate the expected response. At the same time the scoreboard is waiting for the result of the AES module to compute. Once complete the result is scored against the prediction made by the reference model.

The reference model is selected to be either a C-implementation or an SSL-library selected on a random basis with the default distribution of 80% OpenSSL/BoringSSL and 20% C-model.

The default behavior for the verification is that the scoreboard wait until the complete message has been encrypted/decrypted before checking the result against the reference model.

The scoreboard has a step through mode where the scoring is done after each 128bit block. This setting is only available when using the C-model as reference and is controlled with a knob.

Assertions

  • TLUL assertions: The tb/aes_bind.sv binds the tlul_assert assertions to the IP to ensure TileLink interface protocol compliance.
  • Unknown checks on DUT outputs: The RTL has assertions to ensure all outputs are initialized to known values after coming out of reset.

Building and running tests

We are using our in-house developed regression tool for both building and running our tests and regressions. Please take a look at the link for detailed information on the usage, capabilities, features and known issues. Here’s how to run a smoke test:

$ $REPO_TOP/util/dvsim/dvsim.py $REPO_TOP/hw/ip/aes/dv/aes_sim_cfg.hjson -i aes_smoke

Here’s how to run a basic test without DPI calls:

$ $REPO_TOP/util/dvsim/dvsim.py $REPO_TOP/hw/ip/aes/dv/aes_sim_cfg.hjson -i aes_wakeup

Testplan

Testpoints

Milestone Name Tests Description
V1 wake_up aes_wake_up

Basic hello world, encrypt a plain text read it back - decrypt and compare to input.

V1 smoke aes_smoke

Encrypt a plain text read it back - decrypt and compare to input but use reference model to compare after both encryption and decryption.

V1 csr_hw_reset aes_csr_hw_reset

Verify the reset values as indicated in the RAL specification.

  • Write all CSRs with a random value.
  • Apply reset to the DUT as well as the RAL model.
  • Read each CSR and compare it against the reset value. it is mandatory to replicate this test for each reset that affects all or a subset of the CSRs.
  • It is mandatory to run this test for all available interfaces the CSRs are accessible from.
  • Shuffle the list of CSRs first to remove the effect of ordering.
V1 csr_rw aes_csr_rw

Verify accessibility of CSRs as indicated in the RAL specification.

  • Loop through each CSR to write it with a random value.
  • Read the CSR back and check for correctness while adhering to its access policies.
  • It is mandatory to run this test for all available interfaces the CSRs are accessible from.
  • Shuffle the list of CSRs first to remove the effect of ordering.
V1 csr_bit_bash aes_csr_bit_bash

Verify no aliasing within individual bits of a CSR.

  • Walk a 1 through each CSR by flipping 1 bit at a time.
  • Read the CSR back and check for correctness while adhering to its access policies.
  • This verify that writing a specific bit within the CSR did not affect any of the other bits.
  • It is mandatory to run this test for all available interfaces the CSRs are accessible from.
  • Shuffle the list of CSRs first to remove the effect of ordering.
V1 csr_aliasing aes_csr_aliasing

Verify no aliasing within the CSR address space.

  • Loop through each CSR to write it with a random value
  • Shuffle and read ALL CSRs back.
  • All CSRs except for the one that was written in this iteration should read back the previous value.
  • The CSR that was written in this iteration is checked for correctness while adhering to its access policies.
  • It is mandatory to run this test for all available interfaces the CSRs are accessible from.
  • Shuffle the list of CSRs first to remove the effect of ordering.
V1 csr_mem_rw_with_rand_reset aes_csr_mem_rw_with_rand_reset

Verify random reset during CSR/memory access.

  • Run csr_rw sequence to randomly access CSRs
  • If memory exists, run mem_partial_access in parallel with csr_rw
  • Randomly issue reset and then use hw_reset sequence to check all CSRs are reset to default value
  • It is mandatory to run this test for all available interfaces the CSRs are accessible from.
V1 shadow_reg_update_error aes_shadow_reg_errors

Verify shadowed registers' update error.

  • Randomly pick a shadowed register in the DUT.
  • Write it twice with different values.
  • Verify that the update error alert is triggered and the register value remains unchanged.
  • Verify the update_error status register field is set to 1.
  • Repeat the above steps a bunch of times.
V1 shadow_reg_read_clear_staged_value aes_shadow_reg_errors

Verify reading a shadowed register will clear its staged value.

  • Randomly pick a shadowed register in the DUT.
  • Write it once and read it back to clear the staged value.
  • Then write it twice with the same new value (but different from the previous step).
  • Read it back to verify the new value and ensure that the update error alert did not trigger.
  • Verify the update_error status register field remains the same value.
  • Repeat the above steps a bunch of times.
V1 shadow_reg_storage_error aes_shadow_reg_errors

Verify shadowed registers' storage error.

  • Randomly pick a shadowed register in the DUT.
  • Backdoor write to shadowed or committed flops to create a storage fatal alert.
  • Check if fatal alert continuously fires until reset.
  • Verify that all other frontdoor write attempts are blocked during the storage error.
  • Verify that storage_error status register field is set to 1.
  • Reset the DUT.
  • Read all CSRs to ensure the DUT is properly reset.
  • Repeat the above steps a bunch of times.
V1 shadowed_reset_glitch aes_shadow_reg_errors

Verify toggle shadowed_rst_n pin can trigger storage error.

  • Randomly drive shadowed_rst_n pin to low or rst_n pin to low.
  • check if any registers have been written before the reset. If so check if storage error fatal alert is triggered.
  • Check status register.
  • Drive shadowed_rst_n pin or rst_n pin back to high.
  • If fatal alert is triggered, reset the DUT.
  • Read all CSRs to ensure the DUT is properly reset.
  • Repeat the above steps a bunch of times.
V1 shadow_reg_update_error_with_csr_rwaes_shadow_reg_errors_with_csr_rw

Run shadow_reg_update_error sequence in parallel with csr_rw sequence.

  • Randomly select one of the above sequences.
  • Apply csr_rw sequence in parallel but disable the csr_access_abort to ensure all shadowed registers' write/read to be executed without aborting.
  • Repeat the above steps a bunch of times.
V2 algorithm aes_smoke
aes_config_error

Compare cypher text from DUT with the output of a C model using same key and data.

V2 key_length aes_stress
aes_smoke
aes_config_error

Randomly select key length to verify all supported key lengths are working.

V2 back2back aes_b2b
aes_stress

Randomly select the spacing between consecutive messages and blocks from 0 - n clock cycles. The distribution will be weighted toward no and small gaps (0-10 cycles) but will also cover larger gaps.

V2 backpressure aes_stress

Try to write data to registers without offloading the DUT output to verify Stall functionality.

V2 multi_message aes_stress
aes_smoke
aes_config_error
aes_alert_reset

Run multiple messages in a random mix of encryption / decryption. Each message should select its mode randomly.

V2 failure_test aes_config_error
aes_alert_reset
  • Tests what happens if a register is written a the wrong time? If a key does not match the key setting etc. Will the DUT ignore or fail gracefully.
  • Enter a 256bit key but set DUT to use 128bit for encryption. Then enter the 128bit of the key and use for decryption. Will result match plain text and vice.
  • Write unsupported configurations (Key length and mode are 1 hot, what happens if more than one bit is set.)
V2 trigger_clear_test aes_clear

Exercise trigger and clear registers at random times to make sure we handle the different cornercases correctly. Example of a cornercases clearing data input or data output before the data is consumed or the DUT finishes an operation.

V2 nist_test_vectors

Verify that the DUT handles the NIST test vectors correctly.

V2 performance

Verify that the DUT performs as specified for each key length in terms of latency and throughput. This testpoint will use automode (this will feed input data and offload output data as fast as the DUT can support it.)

V2 reset_recovery aes_alert_reset

Pull reset at random times, make sure DUT recover/resets correctly and there is no residual data left in the registers.

V2 stress aes_stress

This will combine the other individual testpoints to ensure we stress test everything across the board.

V2 deinitialization

Make sure that there is no residual data from latest operation.

V2 alert_test aes_alert_test

Verify common alert_test CSR that allows SW to mock-inject alert requests.

  • Enable a random set of alert requests by writing random value to alert_test CSR.
  • Check each alert_tx.alert_p pin to verify that only the requested alerts are triggered.
  • During alert_handshakes, write alert_test CSR again to verify that: If alert_test writes to current ongoing alert handshake, the alert_test request will be ignored. If alert_test writes to current idle alert handshake, a new alert_handshake should be triggered.
  • Wait for the alert handshakes to finish and verify alert_tx.alert_p pins all sets back to 0.
  • Repeat the above steps a bunch of times.
V2 tl_d_oob_addr_access aes_tl_errors

Access out of bounds address and verify correctness of response / behavior

V2 tl_d_illegal_access aes_tl_errors

Drive unsupported requests via TL interface and verify correctness of response / behavior. Below error cases are tested bases on the [TLUL spec]({{< relref "hw/ip/tlul/doc/_index.md#explicit-error-cases" >}})

  • TL-UL protocol error cases
    • invalid opcode
    • some mask bits not set when opcode is PutFullData
    • mask does not match the transfer size, e.g. a_address = 0x00, a_size = 0, a_mask = 'b0010
    • mask and address misaligned, e.g. a_address = 0x01, a_mask = 'b0001
    • address and size aren't aligned, e.g. a_address = 0x01, a_size != 0
    • size is greater than 2
  • OpenTitan defined error cases
    • access unmapped address, expect d_error = 1 when devmode_i == 1
    • write a CSR with unaligned address, e.g. a_address[1:0] != 0
    • write a CSR less than its width, e.g. when CSR is 2 bytes wide, only write 1 byte
    • write a memory with a_mask != '1 when it doesn't support partial accesses
    • read a WO (write-only) memory
    • write a RO (read-only) memory
V2 tl_d_outstanding_access aes_csr_hw_reset
aes_csr_rw
aes_csr_aliasing
aes_same_csr_outstanding

Drive back-to-back requests without waiting for response to ensure there is one transaction outstanding within the TL device. Also, verify one outstanding when back- to-back accesses are made to the same address.

V2 tl_d_partial_access aes_csr_hw_reset
aes_csr_rw
aes_csr_aliasing
aes_same_csr_outstanding

Access CSR with one or more bytes of data. For read, expect to return all word value of the CSR. For write, enabling bytes should cover all CSR valid fields.

V2S tl_intg_err aes_tl_intg_err

Verify that the data integrity check violation generates an alert.

Randomly inject errors on the control, data, or the ECC bits during CSR accesses. Verify that triggers the correct fatal alert.

Covergroups

Name Description
tl_errors_cg

Cover the following error cases on TL-UL bus:

  • TL-UL protocol error cases.
  • OpenTitan defined error cases, refer to testpoint tl_d_illegal_access.
tl_intg_err_cg

Cover all kinds of integrity errors (command, data or both) and cover number of error bits on each integrity check.

Cover the kinds of integrity errors with byte enabled write on memory if applicable: Some memories store the integrity values. When there is a subword write, design re-calculate the integrity with full word data and update integrity in the memory. This coverage ensures that memory byte write has been issued and the related design logic has been verfied.