[Libre-soc-bugs] [Bug 432] mulld pseudocode overflow calculation incorrect

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=432

--- Comment #10 from Jacob Lifshay <programmerjake at gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #7)
> (In reply to Jacob Lifshay from comment #5)
> > (In reply to Luke Kenneth Casson Leighton from comment #3)
> > > look reasonable?
> > 
> > looks good to me!
> 
> ok great.  unit tests pass too.  however they're simulator-HDL not
> simulator-qemu.  if similar overflow unit tests had been added to run against
> qemu this would have been caught earlier.

qemu is known to not match POWER9, so some tests should fail. I added the
Python  bindings to power-instruction-simulator specifically so we could write
unit tests against its instruction models, since they are likely to catch cases
where POWER9 and qemu disagree. So, to use it:

import power_instruction_analyzer as pia

fn unit_test_mulldo_(self, ra, rb, rt, so, ov, ov32, cr0):
    # ra, rb, and rt are u64
    # so, ov, and ov32 are bool

    # includes rc for madd*, didn't yet set a default value
    # so rc=<something> required for now:
    inputs = pia.InstructionInputs(ra=ra, rb=rb, rc=0)
    # all functions assumes SO flag starts as 0
    outputs = pia.mulldo_(inputs) # the mulldo. instruction
    self.assertEqual(outputs.rt, rt)
    self.assertEqual(outputs.overflow.so, so)
    self.assertEqual(outputs.overflow.ov, ov)
    self.assertEqual(outputs.overflow.ov32, ov32)
    self.assertEqual(outputs.cr0.lt, bool(cr0 & 8))
    self.assertEqual(outputs.cr0.gt, bool(cr0 & 4))
    self.assertEqual(outputs.cr0.eq, bool(cr0 & 2))
    self.assertEqual(outputs.cr0.so, bool(cr0 & 1))

-- 
You are receiving this mail because:
You are on the CC list for the bug.