6d55cd2d55
|
||
|---|---|---|
| .. | ||
| .gitignore | ||
| Makefile | ||
| Readme.md | ||
| amo_smoke.c | ||
| amo_smoke.expected_output | ||
| amo_smoke.gtkw | ||
| amo_timer_irq.c | ||
| amo_timer_irq.gtkw | ||
| cleantests | ||
| ecall_simple.c | ||
| ecall_simple.expected_output | ||
| ecall_simple.gtkw | ||
| hellow.c | ||
| hellow.expected_output | ||
| illegal_instr.c | ||
| illegal_instr.expected_output | ||
| lr_sc_smoke.c | ||
| lr_sc_smoke.gtkw | ||
| runtests | ||
| wfi_loop.c | ||
| wfi_loop.expected_output | ||
Readme.md
Software Testcases
A smorgasbord of software testcases for various features and cases that aren't well-covered by upstream tests such as riscv-arch-test, the riscv-test end-to-end debug tests or riscv-formal. Each test consists of one C file.
Some tests have an expected text output associated with them -- the test passes if this text output matches, and main() exits with a zero return code. Other tests are completely self-checking, reporting pass/fail only with the return code from main(). This means there is no point running these tests if the processor is in a fundamentally broken state (e.g. doesn't pass ISA compliance) and can't be trusted to check itself.
To run the tests:
./runtests
This will first rebuild the simulator (../tb_cxxrtl/) if needed, then build and run all the software testcases, then print out a summary of test pass/fail status. The ./run_tests executable itself returns a successful exit code if and only if all tests passed. A VCD trace and printf log will be created for each test, with the same name as the test, for debugging failures.
To clean up the junk:
./cleantests