Assuming you already:
make simenv SIM=RTL make verify-<dv-test> # OR SIM=GL make verify-<dv-test>
<dv-test>: io_ports, mprj_stimulus, la_test1, la_test2 or wb_port.
There are two options for setting up the simulation environment:
There is an available docker setup with the needed tools at efabless/dockerized-verification-setup
Run the following to pull the image:
docker pull efabless/dv_setup:latest
First, you will need to export a number of environment variables:
export PDK_PATH=<pdk-location/sky130A> export CARAVEL_ROOT=<caravel_root> export TARGET_PATH=<caravel_user_project>
Then, run the following command to start the docker container :
docker run -it -v ${TARGET_PATH}:${TARGET_PATH} -v ${PDK_ROOT}:${PDK_ROOT} \
-v ${CARAVEL_ROOT}:${CARAVEL_ROOT} \
-e TARGET_PATH=${TARGET_PATH} -e PDK_ROOT=${PDK_ROOT} \
-e CARAVEL_ROOT=${CARAVEL_ROOT} \
-e TOOLS=/foss/tools/riscv-gnu-toolchain-rv32i/411d134 \
-e DESIGNS=$(TARGET_PATH) \
-e CORE_VERILOG_PATH=$(TARGET_PATH)/mgmt_core_wrapper/verilog \
-e MCW_ROOT=$(MCW_ROOT) \
efabless/dv_setup:latest
Then, navigate to the directory where the DV tests reside :
cd $TARGET_PATH/verilog/dv/
To run any simulation, you need to be on the top level or caravel_user_project.
To run RTL simulation for one of the DV tests,
SIM=RTL make verify-<dv-test>
To run gate level simulation for one of the DV tests,
SIM=GL make verify-<dv-test>
A discription of the example testbenches
The directory includes five tests for the counter user-project example:
This test is meant to verify that we can configure the pads for the user project area. The firmware configures the lower 8 IO pads in the user space as outputs:
reg_mprj_io_0 = GPIO_MODE_USER_STD_OUTPUT; reg_mprj_io_1 = GPIO_MODE_USER_STD_OUTPUT; ..... reg_mprj_io_7 = GPIO_MODE_USER_STD_OUTPUT;
Then, the firmware applies the pad configuration by enabling the serial transfer on the shift register responsible for configuring the pads and waits until the transfer is done.
reg_mprj_xfer = 1; while (reg_mprj_xfer == 1);
The testbench success criteria is that we can observe the counter value on the lower 8 I/O pads. This criteria is checked by the testbench through observing the values on the I/O pads as follows:
wait(mprj_io_0 == 8'h01); wait(mprj_io_0 == 8'h02); wait(mprj_io_0 == 8'h03); .... wait(mprj_io_0 == 8'hFF);
If the testbench fails, it will print a timeout message to the terminal.
This test is meant to verify that we can use the logic analyzer to monitor and write signals in the user project from the management SoC. Firstly, the firmware configures the upper 16 of the first 32 GPIO pads as outputs from the managent SoC, applies the configuration by initiating the serial transfer on the shift register, and writes a value on the pads to indicate the end of pad configuration and the start of the test.
reg_mprj_io_31 = GPIO_MODE_MGMT_STD_OUTPUT; reg_mprj_io_30 = GPIO_MODE_MGMT_STD_OUTPUT; ..... reg_mprj_io_16 = GPIO_MODE_MGMT_STD_OUTPUT; reg_mprj_io_15 = GPIO_MODE_USER_STD_OUTPUT; reg_mprj_io_14 = GPIO_MODE_USER_STD_OUTPUT; ..... reg_mprj_io_0 = GPIO_MODE_USER_STD_OUTPUT; reg_mprj_io_6 = GPIO_MODE_MGMT_STD_OUTPUT; reg_mprj_xfer = 1; while (reg_mprj_xfer == 1); // Flag start of the test reg_mprj_datal = 0xAB400000;
This is done to flag the start/success/end of the simulation by writing a certain value to the I/Os which is then checked by the testbench to know whether the test started/ended/succeeded. For example, the testbench checks on the value of the upper 16 of 32 I/Os, if it is equal to 16'hAB40, then we know that the test started.
wait(checkbits == 16'hAB40); $display("LA Test 1 started");
Then, the firmware configures the logic analyzer (LA) probes [31:0] as inputs to the management SoC to monitor the counter value, and configure the logic analyzer probes [63:32] as outputs from the management SoC (inputs to the user_project) to set the counter initial value. This is done by writing to the LA probes enable registers. Note that the output enable is active low, while the input enable is active high. Every channel can be configured for input, output, or both independently.
reg_la0_oenb = reg_la0_iena = 0x00000000; // [31:0] reg_la1_oenb = reg_la1_iena = 0xFFFFFFFF; // [63:32] reg_la2_oenb = reg_la2_iena = 0x00000000; // [95:64] reg_la3_oenb = reg_la3_iena = 0x00000000; // [127:96]
Then, the firmware writes an initial value to the counter through the LA1 data register. Afte writing the counter value, the LA probes are disabled to prevent the counter write signal from being always set to one.
reg_la1_data = 0x00000000; // Write zero to count register reg_la1_oenb = reg_la1_iena = 0x00000000; // Disable probes
The firmware then waits until the count value exceeds 500 and flags the success of the test by writing 0xAB41 to pads 16 to 31. The firmware reads the count value through the logic analyzer probes [31:0]
while (1) { if (reg_la0_data > 0x1F4) { // Read current count value through LA reg_mprj_datal = 0xAB410000; // Flag success of the test break; } }
This test is meant to verify that we can drive the clock and reset signals for the user project example through the logic analyzer. In the user_project RTL, the clock can either be supplied from the wb_clk_i or from the logic analyzer through bit [64]. Similarly, the reset signal can be supplied from the wb_rst_i or through LA[65]. The firmware configures the clk and reset LA probes as outputs from the management SoC by writing to the LA2 enable register.
reg_la0_oenb = reg_la0_iena = 0x00000000; // [31:0] reg_la1_oenb = reg_la1_iena = 0x00000000; // [63:32] reg_la2_oenb = reg_la2_iena = 0x00000000; // [95:64] reg_la3_oenb = reg_la3_iena = 0x00000000; // [127:96]
Then, the firmware supplies both clock reset signals through LA2 data register. First, both are set to one. Then, reset is driven to zero and the clock is toggled for 6 clock cycles.
reg_la2_data = 0x00000003; // Write one to LA[64] and LA[65] for (i=0; i<11; i=i+1) { // Toggle clk & de-assert reset clk = !clk; reg_la2_data = 0x00000000 | clk; }
The testbench success criteria is that the firmware reads a count value of five through the LA probes.
while (1){ if (reg_la0_data == 0x05) { reg_mprj_datal = 0xAB610000; // FLag success of the test } }
This test is meant to verify that we can read data generated from the user project on the gpio outputs. This is done by configuring the LA probes as inputs to the cpu, except for LA probes [63:32] as output from the cpu
reg_la0_oenb = reg_la0_iena = 0x00000000; // [31:0] reg_la1_oenb = reg_la1_iena = 0xFFFFFFFF; // [63:32] reg_la2_oenb = reg_la2_iena = 0x00000000; // [95:64] reg_la3_oenb = reg_la3_iena = 0x00000000; // [127:96]
Then tests the ability to force data on the gpio through channel 37
reg_mprj_datah = 0x0f0f0fc0; reg_mprj_datah = 0x00000000; reg_mprj_datah = 0x0f0f0fca; reg_mprj_datah = 0x0000000a; reg_mprj_datah = 0x0f0f0fc0; reg_mprj_datah = 0x00000000; reg_mprj_datah = 0x0f0f0fc5; reg_mprj_datah = 0x00000005;
It then tests the ability to read back the data generated by the user project on the gpio
testval = reg_mprj_datal; reg_mprj_datal = (testval << 16); testval = reg_mprj_datal; reg_mprj_datal = (testval << 16);
0x2710 to the count register, then reads back the count value after some time. The read and write transactions happen through the management SoC wishbone bus and are initiated by either writing or reading from the user project address on the wishbone bus. The reg_wb_enable needs to be set to 1 in order to enable the wishbone bus.