A lot of utility code was just being imported willy-nilly from one
package to another. This moves the common code into util to make things
more sensible. The code moved were
* The AsyncQueue and AsyncDecoupledCrossing from junctions.
* All of the code in rocket's util.scala
* The BlackBox asynchronous reset registers from uncore.tilelink2
* The implicit definitions from uncore.util
The NastiConverterTest, PCIeMockupTest, and DirectGroundtest
configurations have been broken by recent changes.
The NastiConverterTest has been superseded by a unit test and the
other two were only created for an attempt at FPGA debugging.
They weren't actually very useful for that purpose, so might as well get
rid of them.
The mask of a Get should also be converted.
This manifested as a bug when going from 32=>64 bits. A large Get
could end up with mask that was not full.
This would allow, for instance, putting the coreplex on a separate clock
domain and crossing the IOs over through asynchronous queues.
The ExampleMultiClockTop* classes are removed since they no longer fit
into the class hierarchy.
We have a handful of TileLink-related helper objects
(wrappers, unwrappers, width adapters, and enqueuers). Previously, using
them could be error-prone, because you had to make sure the implicit
parameters they took in had the same TLId as the TileLinkIO bundles
passed in as inputs. This is rather silly, we should just use the
parameters in the bundle.
* tilelink2 Edges: add accessor methods for address and addr_{hi,lo}
* tilelink2: use addr_lo instead of relying on truncation
Truncation can mess up if the width should be 0, but IS 1.
* rocketchip: rename identically names devices with _%d
If you connect two devices with the same name in TL2 (totally ok there),
when they get put into the TL1 addrmap, one gets silently overwritten.
This renames the second occurance as _1, third as _2, and so on...
* junctions: blow if duplicates add to addrmap
* debug: Clean up Debug TransportModule synchronizer
With async reset async queues, I feel its safe/cleaner
to remove the one-off "AsyncMailbox verilog black-box
and use the common primitive.
I also added some comments about correct usage of this
block. Probably the 'TRST' signal should be renamed
to make it less confusing, as it requires some processing
of the real JTAG 'TRST' signal.
* crossing: use async reset
* crossings: asyncqueue needs Asynchronous reset.
* crossing: Actually enable the head of the synchronizer flop chain
* crossing: remove reset from logic. This flop will no longer be written during reset because valid will be low.
* crossing: Tidy up code & comments
If a device has configurable bus-width, we need a stable way of
enumerating registers. The byte offset stays unchanged.
This change also makes it possible to put an arbitrary number of RegFields
starting at some address which are then chopped up into appropriately bus-
sized registers.
The implementation unconditionally drove the register.
This made it incompatible with drivers from the device itself.
Besides, writing only parts of a register at a time is ultra-shady.
This API makes it much more readable when you have multiple adapters
combined into a single line. The arguments for each adapter stay
beside the adapter.
For example, this:
peripheryBus.node := TLWidthWidget(TLBuffer(TLAtomicAutomata()(TLHintHandler(legacy.node))), legacy.tlDataBytes)
becomes this:
peripheryBus.node := TLWidthWidget(legacy.tlDataBytes)(TLBuffer()(TLAtomicAutomata()(TLHintHandler()(legacy.node))))
