Multi-thread seems not supported by `crt.S`

[CaiTH0618]

I'm using riscv-tests with Chipyard and I failed to start a multi-thread program on a multi-core Chipyard config.

I noticed in the file benchmarks/common/crt.S, the startup code currently handles core entry as follows. This implementation puts all cores with mhartid>=1 into a deadloop and never calls the entry function for those cores.

  # get core id
  csrr a0, mhartid
  # for now, assume only 1 core
  li a1, 1
1:bgeu a0, a1, 1b

A straightforward way is to replace these lines with the following ones. And add a -DNUM_CORES=4 (according to actual core numbers) when compiling.

  # get core id
  csrr a0, mhartid
  li a1, NUM_CORES
1:bgeu a0, a1, 1b

Would be great if riscv-tests provide a complete solution for multi-thread support.