mshr: Prevent assertion misfire during reset

[paulsc96]

While the cache is held in reset, the one_command_assert assertion in hpdcache_mshr will always fail, raising a false-positive error any time the cache is being reset:

# ** Error: MSHR: ack with concurrent alloc or check

This should not happen as the reported error is not an actual issue. Furthermore, it presents a problem for testing setups designed to fail on assertion failures, as any meaningful simulation will trigger this assertion.

The proposed fix masks the assertion during reset with the disable iff clause.

[cfuguet]

Hello @paulsc96,

Yes, thank you for this PR. I accept it.

Could you please accept the Eclipse agreement before I can merge your contribution ?

Thank you

[khandelwaltanuj]

Hi @cfuguet

I just looked at this issue and I am wondering why we never caught it in the regression run by CEA. Is it a recent change or there is a scenerio that we need to add to reproduce this issue ?

Thanks and regards
Tanuj Khandelwal

[cfuguet]

Yes, I don't know why this was not triggered in the UVM testbench.

I've never saw it neither in the Verilator testbench, but probably because Verilator only models 1 and 0 in bit vectors (not X or Z).