That effectively increased the miss latency by 5 cycles when there was
a store hit followed by a load miss. Since pending stores are drained
when releaseInFlight, the check I removed was redundant.
When we first implemented TL, we thought this was helpful, because
it made WidthWidgets stateless in all cases. However, it put too
much burden on all other masters and slaves, none of which benefitted
from this signal. Furthermore, even with addr_lo, WidthWidgets were
information lossy because when they widen, they have no information
about what to fill in the new high bits of addr_lo.
- split the data based on the address, not the mask
(the first version of TileLink did not have low address bits)
- the dependency on addr_lo is now exposed and easy to replace
- errors are properly OR reduced
- registers latched only as needed (was previously a shift register)
- combines beats without inspecting address (removes addr_lo dependency)
We are overloading the BTB-hit signal to mean that any part of the frontend
changed the control-flow, not just the BTB. That's the right thing to do for
most of the control logic, but it means the BTB sometimes won't get refilled
when we'd like it to. This commit makes the frontend use an invalid BTB entry
number when it, rather than the BTB, changes the control flow. Since the
entry number is invalid, the BTB will treat it as a miss and refill itself.
This is kind of a hack, but a more palatable fix requires reworking the RVC
IBuf, which I don't have time for right now.
Consider the following waveform for two 4-beat bursts:
---A----A------------
-------D-----DDD-DDDD
Under TL rules, the second A can use the same source as the first A,
because the source is released for reuse on the first response beat.
However, if we fragment the requests, it looks like this:
---3210-3210---------
-------3-----210-3210
... now we've broken the rules because 210 are twice inflight.
To solve this, we alternate an a.source bit every time D completes a txn.
* coreplex collapse: peripherals now in coreplex
* coreplex: better factoring of TLBusWrapper attachement points
* diplomacy: allow monitorless :*= and :=*
* rocket: don't connect monitors to tile tim slave ports
* rename chip package to system
* coreplex: only sbus has a splitter
* TLFragmenter: Continuing my spot battles on requires without explanatory strings
* pbus: toFixedWidthSingleBeatSlave
* tilelink: more verbose requires
* use the new system package for regression
* sbus: add more explicit FIFO attachment points
* delete leftover top-level utils
* cleanup ResetVector and RTC
