Commit Graph

106 Commits

Author SHA1 Message Date
David Anderson aa048537ef vram/MemArbiter: rewrite to use client/server idioms
In preparation for making the two Connectable and defining an
arbitrated memory client/server for VRAM access.
2024-09-07 16:58:11 -07:00
David Anderson b2b2c14009 vram/VRAM: finish the top-level VRAM module
Well, for now at least. It can build 112KiB and 128KiB memories that
seem to synthesize to something reasonable.
2024-09-07 16:04:21 -07:00
David Anderson dd551ce09b tasks.py: fix typo in synth process 2024-09-07 16:03:52 -07:00
David Anderson fc4b212139 compile all tests before running them, so the outputs are grouped 2024-09-07 12:15:33 -07:00
David Anderson 6e84e9c689 adjust 'inv genclk' to allow per-hardware PLLs 2024-09-07 12:11:27 -07:00
David Anderson 77b772c7ee tasks.py: improve handling of Bluespec libdirs
To help divvy up the sources better, the build now makes a libpath
consisting of all directories that have bsv files in them, with a few
exceptions: hardware subdirs are target-specific so only get used if
they're the current build target. Experiments are random crap so get
the same treatment. And the 'sim' dir is only test helpers, so they
only get pulled in by tests.
2024-09-07 10:31:32 -07:00
David Anderson 25d1806590 lib/ClockOut: hack module to export a clock as an ordinary signal
Used to output a clock signal from an FPGA pin. The resultant output
signal is unclocked, so can be presented to any output at will.
2024-09-07 10:06:50 -07:00
David Anderson 5e997201db hardware/sentinel65x: move top-level hw module to hw subdir 2024-09-07 10:04:36 -07:00
David Anderson 6fd040565c Grab the inout port fixer from bsc tree, wire it in
Yosys doesn't understand Verilog-2001 port aliases. Unfortunately
bsc uses those to represent inout ports because it's the only way
to represent a particular kind of shared bus in Verilog source code.

Thankfully, a kind soul at Bluespec Inc made a perl script that
transforms the port alias construct into regular verilog-1995, which
works fine in cases like mine where the only user of the inout port
is a TriState module which tears it apart into separate
input/output/enable signals for the rest of bsc to work with.
2024-09-06 21:26:39 -07:00
David Anderson 0005ad6fe5 sentinel65x/PLL: update PLL settings, fix generator 2024-09-06 21:18:00 -07:00
David Anderson d960eee973 flake.lock: update tools 2024-09-06 21:17:32 -07:00
David Anderson f61328dac4 experiments/primitive_ram: clean up old testing code 2024-09-06 16:23:49 -07:00
David Anderson 930f9f7078 sentinel65x: start of Sentinel 65X top-level glue, with a PLL module
The PLL generates a 100MHz clock for GARY, 10MHz for the CPU and
system bus, and 25MHz for VGA out.
2024-09-06 16:21:25 -07:00
David Anderson f7cb4b6ba2 vram/VRAM: early VRAM implementation
Only checked up to mkByteRAMArray, main VRAM still WIP
2024-09-06 16:11:55 -07:00
David Anderson 7560199251 tasks: support running nextpnr with a generic ulx3s pin map
This is handy for quick experiments, rather than having to define a
bespoke pin map for each experiment.
2024-09-06 16:11:03 -07:00
David Anderson 2953106ec7 lib/ECP5_RAM: mark interfaces always_ready
Elides a couple of unnecessary signals from the design
2024-09-06 16:10:22 -07:00
David Anderson ab20db44f4 vram: rename MemoryArbiter to something shorter
As I implement the whole transaction-level modeling thing with
arbitration, the interface names are getting longer and longer.
2024-09-06 10:04:20 -07:00
David Anderson 7f10694371 vram: get rid of old commented code 2024-09-06 00:13:42 -07:00
David Anderson f010693b9b vram: compact the test output even more 2024-09-06 00:08:14 -07:00
David Anderson 71f422b8c1 vram: fix up documentation for MemoryArbiterWriter 2024-09-06 00:02:23 -07:00
David Anderson e0d156cdd1 vram: a little more formatting of the vram test
Make it a bit more compact, for large test runs
2024-09-06 00:00:25 -07:00
David Anderson d41ca7daae vram: tidy up formatting 2024-09-05 23:51:21 -07:00
David Anderson 60348fefd2 vram: one more arbiter test for port conflicts 2024-09-05 23:46:01 -07:00
David Anderson 5e22d03e15 vram: implement a MemoryArbiter for VRAM 2024-09-05 23:39:21 -07:00
David Anderson 2cd172cc73 tasks.py: add sim directory when running tests
So that bluesim can find the DP16KD simulation model.
2024-08-31 00:00:59 -07:00
David Anderson efb5327f53 lib: clean up ECP5_RAM.v a bit 2024-08-31 00:00:41 -07:00
David Anderson 119eeceaef Add some early testing harness for the sim DP16KD 2024-08-30 22:14:10 -07:00
David Anderson a4a10becbf tasks.py: print one more intermediate output path 2024-08-30 20:06:36 -07:00
David Anderson a0892fefcd sim/tb: add "bad" signals to make it easier to see wrong outputs 2024-08-30 20:00:11 -07:00
David Anderson b913afd416 sim/tb: add wanted output signal, to help see incorrect outputs 2024-08-30 19:56:08 -07:00
David Anderson 0b384c6619 sim: implementation of a simulation model DP16KD
Not fully verified with tests yet, WIP
2024-08-30 18:54:54 -07:00
David Anderson e57f7e05b0 lib/ECP5_RAM: fixups based on checking synth output 2024-08-23 00:22:48 -07:00
David Anderson febd840be0 Add svlint config 2024-08-23 00:21:15 -07:00
David Anderson 6666c1ca2d flake.nix: fix up xdot, add systemverilog linter 2024-08-23 00:20:45 -07:00
David Anderson 5d16ce23aa tasks.py: adjust yosys script 2024-08-23 00:20:31 -07:00
David Anderson b73a211ec4 lib/ECP5_RAM: clean up old core, fix error in module docstring 2024-08-20 19:25:04 -07:00
David Anderson 1ccd1b0072 lib/DelayLine: add a bit more documentation 2024-08-20 09:15:52 -07:00
David Anderson 2a8689564c tasks.py: remove debug message 2024-08-20 09:13:36 -07:00
David Anderson 5df41d4b94 lib: use DelayLine in ECP5_RAM
Cleans up the code nicely, and still produces the correct logic.
2024-08-20 00:54:50 -07:00
David Anderson f1e705fd31 lib: add more documentation 2024-08-20 00:29:32 -07:00
David Anderson 85e27554ec lib: add a DelayLine module
A delay line takes a write and echoes it back N cycles later,
with N fixed at compile time. It's a handy primitive to have
when wrapping Verilog blackbox modules because the blackbox
often specifies something like having 2 cycles of latency,
and so you need to bubble the fact that a write occurred 2
cycles ago through to the output so that you can wire up the
right implicit conditions.
2024-08-19 23:00:15 -07:00
David Anderson 27da4958d2 experiments/rmw_ram: document failed/paused memory trickery experiments 2024-08-19 15:39:43 -07:00
David Anderson da6ea4cf42 lib: flesh out the ECP5 EBR modules, write copious documentation 2024-08-18 16:12:57 -07:00
David Anderson a69cc878ce experiments/primitive_ram: customize the clock/reset of one of the RAM ports 2024-08-17 16:41:24 -07:00
David Anderson a23661a449 lib: default clocks and resets to the ambient ones from context
Callers can still specify whacky cross-domain RAMs in the cfg, but the
default is what you usually want: a dual-port RAM with both ports in the
caller's clock/reset domain.
2024-08-17 16:39:45 -07:00
David Anderson e64b990f80 lib: fix port B reset wiring for ECP5_RAM 2024-08-17 16:38:52 -07:00
David Anderson 8d2261e245 lib: initial implementation of an ECP5 EBR primitive
Only the core unconditioned primitive right now, and still needs refining.
2024-08-17 15:43:36 -07:00
David Anderson db30e4a23f tasks.py: prettify output, support running partial synthesis
Partial synth is handy when writing gnarly Bluespec modules, because it
lets you inspect the Verilog output of the Bluespec compiler as well as
Yosys's compile output at various stages of synthesis, to see if things
are being produced the way you expect.
2024-08-17 15:41:21 -07:00
David Anderson e6fa717507 Experiment comparing bsc-contrib's video timing generator with brute force
Brute force is a naively written state machine that combines both horizontal
and vertical timings into one, in a way that unrolls comically badly. It's
obviously uncompetitive as-is, but I wanted to use that as a starting point
to see how much bsc and yosys would still be able to cope with it.

The result: the worse code takes much longer for bluespec to evaluate, and it
consumes ~4x the amount of logic elements after synthesis. Less terrible than
I expected, to be honest!
2024-08-15 00:22:51 -07:00
David Anderson dad128b56b Fix up some bugs in the Invoke script. 2024-08-15 00:22:30 -07:00