feature: add regex to inst.assembly schema

[kbroch-rivosinc]

relates to #807

[kbroch-rivosinc]

This regex doesn't try list the exact operand names. (just alphanumeric)
There are still these failures that I'm wondering if they are actual errors:

❯ pre-commit run --all-files check-jsonschema-inst
Validate instruction files with jsonschema...............................Failed
- hook id: check-jsonschema
- exit code: 1

ok -- validation done
ok -- validation done
ok -- validation done
Schema validation errors were encountered.
  arch/inst/Zilsd/sd.yaml::$.assembly: 'xs2, offset(xs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'
ok -- validation done
ok -- validation done
ok -- validation done
ok -- validation done
Schema validation errors were encountered.
  arch/inst/Zfh/fsh.yaml::$.assembly: 'fs2, imm12(xs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'
Schema validation errors were encountered.
  arch/inst/C/c.ebreak.yaml::$.assembly: ' ' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'
  arch/inst/Zfh/flh.yaml::$.assembly: 'xd, imm12(xs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'
Schema validation errors were encountered.
  arch/inst/Zcmp/cm.push.yaml::$.assembly: 'reg_list, -stack_adj' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'
Schema validation errors were encountered.
  arch/inst/Zilsd/ld.yaml::$.assembly: 'rd, offset(rs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'

[ThinkOpenly]

  arch/inst/Zilsd/sd.yaml::$.assembly: 'xs2, offset(xs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'
  arch/inst/Zilsd/ld.yaml::$.assembly: 'rd, offset(rs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'

These files shouldn't exist, actually, since we have arch/inst/I/sd.yaml and arch/inst/I/ld.yaml. The Zilsd code should be merged into the I code, like what was done for Zclsd. Need an issue or PR for that.

  arch/inst/Zfh/fsh.yaml::$.assembly: 'fs2, imm12(xs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'
  arch/inst/Zfh/flh.yaml::$.assembly: 'xd, imm12(xs1)' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'

These will be fixed when #822 is merged.

  arch/inst/C/c.ebreak.yaml::$.assembly: ' ' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'

This is using a space instead of an empty string. Issue/PR warranted.

  arch/inst/Zcmp/cm.push.yaml::$.assembly: 'reg_list, -stack_adj' does not match '^$|^\\(?[a-zA-Z][a-zA-Z0-9_]*\\)?(?:, (imm\\(|\\()?[a-zA-Z][a-zA-Z0-9_]*\\)?)*$'

❓

[kbroch-rivosinc]

Thanks @ThinkOpenly for quick review. I fixed the extra space in ebreak for now.

[dhower-qc]

arch/inst/Zcmp/cm.push.yaml::$.assembly: 'reg_list, -stack_adj' does not match '^$|^\(?[a-zA-Z][a-zA-Z0-9_]\)?(?:, (imm\(|\()?[a-zA-Z][a-zA-Z0-9_]\)?)*$'

That's valid. But it's also a case highlighting that our simple string is insufficient. The syntax for push/pop includes a register list that determined by a 16-entry LUT.