Raptor Engineering Development Team [Tue, 29 Mar 2022 01:11:17 +0000 (20:11 -0500)]
Add initial integration for OpenCores 10/100 Ethernet MAC
Cesar Strauss [Sun, 27 Mar 2022 17:23:39 +0000 (14:23 -0300)]
Finish the SRAM formal proof by implementing induction
Add a combinatorial (no delay) debug read port to the SRAM model,
for turning an unreacheable state (memory and holding register differ)
into an illegal state.
Just two clock cycles are needed for a working proof.
Cesar Strauss [Sat, 26 Mar 2022 21:09:12 +0000 (18:09 -0300)]
Add formal verification of the single port memory block
We test a single, random memory location.
If any memory location, chosen at random, passes, then the whole memory
is correct.
Note that an induction proof is still missing, since it requires
additional assertions to avoid illegal states.
Luke Kenneth Casson Leighton [Sat, 26 Mar 2022 20:18:53 +0000 (20:18 +0000)]
rename PLRU modules to avoid conflict in microwatt
Luke Kenneth Casson Leighton [Fri, 18 Mar 2022 10:45:59 +0000 (10:45 +0000)]
whitespace cleanup (80 char limit, pep8)
Luke Kenneth Casson Leighton [Fri, 18 Mar 2022 10:00:13 +0000 (10:00 +0000)]
turn CompALU/CompLDST latches synchronous
to cut down combinatorial loops
Raptor Engineering Development Team [Mon, 14 Mar 2022 00:33:04 +0000 (19:33 -0500)]
Add initial Tercel integration
NOTE: Still needs testing on physical hardware,
waiting for Arctic Tern support.
Cesar Strauss [Sun, 13 Mar 2022 19:13:04 +0000 (16:13 -0300)]
Simulate some read/write/modify operations on the SRAM model
Cesar Strauss [Sun, 13 Mar 2022 11:56:24 +0000 (08:56 -0300)]
Add a Single R/W Port SRAM model
Begin making unit tests by checking with Yosys.
Luke Kenneth Casson Leighton [Sat, 12 Mar 2022 16:08:29 +0000 (16:08 +0000)]
add extra pipeline stages to ALU FU to make timing
Luke Kenneth Casson Leighton [Sat, 12 Mar 2022 15:45:16 +0000 (15:45 +0000)]
introduce extra register of delay to split combinatorial loops
by making fetch_ic_i_valid sync
Luke Kenneth Casson Leighton [Sat, 12 Mar 2022 14:43:19 +0000 (14:43 +0000)]
Revert "read last row from r.wb.adr not r.req_adr in icache"
This reverts commit
1bcb90676398c8fa0c309c3c7ed3c9ef306c577e.
Luke Kenneth Casson Leighton [Sat, 12 Mar 2022 14:13:15 +0000 (14:13 +0000)]
Revert "store cur_state.pc+4 in separate register to help reduce"
This reverts commit
d9b58d594ede33db2b0290d7edf8e062f5e91460.
Luke Kenneth Casson Leighton [Sat, 12 Mar 2022 13:22:39 +0000 (13:22 +0000)]
store cur_state.pc+4 in separate register to help reduce
combinatorial chains
Luke Kenneth Casson Leighton [Sat, 12 Mar 2022 13:22:15 +0000 (13:22 +0000)]
read last row from r.wb.adr not r.req_adr in icache
Luke Kenneth Casson Leighton [Tue, 8 Mar 2022 16:47:27 +0000 (16:47 +0000)]
remove stbs_done in icache.py
Luke Kenneth Casson Leighton [Tue, 8 Mar 2022 15:46:15 +0000 (15:46 +0000)]
remove ld_stbs_done from dcache: not needed
Luke Kenneth Casson Leighton [Tue, 8 Mar 2022 15:45:49 +0000 (15:45 +0000)]
work-in-progress on sdram opencores wrapper
Cesar Strauss [Sun, 6 Mar 2022 18:00:37 +0000 (15:00 -0300)]
Copy the startup delay from issuer.py to inorder.py
Fixes TestIssuerInternalInOrder hanging at startup.
Luke Kenneth Casson Leighton [Mon, 28 Feb 2022 17:43:39 +0000 (17:43 +0000)]
attempting to introduce an extra few clock cycles delay on power-up
this may help with initialisation of I-Cache SRAM which is combinatorial.
maybe
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 21:23:12 +0000 (21:23 +0000)]
for lulz make I-Cache possible to set to 32-bit (XLEN=32)
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 19:50:33 +0000 (19:50 +0000)]
bit_length is 1 more than needed: subtract 1 from XLEN first
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 19:34:05 +0000 (19:34 +0000)]
fix up shift_rot test_pipe_caller to new regspeckls style
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 19:25:54 +0000 (19:25 +0000)]
convert shift_rot pipeline to XLEN=32/64
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 19:16:35 +0000 (19:16 +0000)]
fix up Logical pipeline to produce HDL with XLEN=32
fix Logical test_pipe_caller.py to use new regspeckls style
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 18:48:28 +0000 (18:48 +0000)]
whoops ALU common output target must be XLEN-bit,
cannot set to length of output (o) using Signal.like
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 18:41:43 +0000 (18:41 +0000)]
set up dummy parent_pspec to pass XLEN=64 in
ALU test_pipe_spec.py
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 18:37:37 +0000 (18:37 +0000)]
start on converting MUL and DIV pipelines to XLEN
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 18:17:08 +0000 (18:17 +0000)]
convert from public static functions/properties for regspecs
to member functions to obtain regspecs
this allows pspec (containing XLEN) to be passed to the regspecs,
which in turn allows them to be dynamically set by issuer_verilog.py
and unit tests
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 18:09:13 +0000 (18:09 +0000)]
fix ALU with XLEN=32, carry and overflow
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 18:01:54 +0000 (18:01 +0000)]
use XLEN in Function Units (starting with ALU)
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 17:31:51 +0000 (17:31 +0000)]
add XLEN to issuer_verilog.py defaults to 64
Luke Kenneth Casson Leighton [Sun, 27 Feb 2022 17:17:07 +0000 (17:17 +0000)]
add XLEN option to regfiles via pspec
Jacob Lifshay [Thu, 24 Feb 2022 03:33:13 +0000 (19:33 -0800)]
add running instructions
Jacob Lifshay [Thu, 24 Feb 2022 03:28:02 +0000 (19:28 -0800)]
add formal proof for shift/rot o.ok
Jacob Lifshay [Thu, 24 Feb 2022 03:16:17 +0000 (19:16 -0800)]
clean up code
Jacob Lifshay [Thu, 24 Feb 2022 03:12:49 +0000 (19:12 -0800)]
add formal proof for OP_RLCR
Jacob Lifshay [Thu, 24 Feb 2022 03:01:49 +0000 (19:01 -0800)]
add formal proof for OP_RLCL
Jacob Lifshay [Thu, 24 Feb 2022 02:40:48 +0000 (18:40 -0800)]
add formal proof for OP_RLC
Luke Kenneth Casson Leighton [Wed, 23 Feb 2022 17:17:50 +0000 (17:17 +0000)]
forgot to pass cix (cache-inhibited) through to LD/ST which was
causing D-Cache to make too many read/writes and overrun wishbone
bus addressing on peripherals.
D-Cache always (for non-cache-inhibited) makes 8 64-bit requests:
if a memory-mapped peripheral has only say 3 32-bit CSRs then the
additional requests beyond the range of the peripheral will cause
a permanent lock-up
Jacob Lifshay [Tue, 22 Feb 2022 09:33:16 +0000 (01:33 -0800)]
speed up shift/rot formal proof by running stuff in parallel
it now runs in about 1m30s in `pytest -n auto <file>.py`
also start work on rewriting proofs to hopefully work better, all
of grev, ternlog, extswsli, shl, and shr are completed.
Luke Kenneth Casson Leighton [Mon, 21 Feb 2022 00:26:14 +0000 (00:26 +0000)]
again reduce combinatorial chains, similar to Trap pipeline,
introduce dummy passthrough stage in Branch pipeline
Luke Kenneth Casson Leighton [Sun, 20 Feb 2022 23:58:58 +0000 (23:58 +0000)]
add syn_ramstyle "block_ram" attributes and reduce i/d-cache sizes again
Luke Kenneth Casson Leighton [Sun, 20 Feb 2022 23:51:32 +0000 (23:51 +0000)]
same as shiftrot, split out separate pipelines for logical
stages in order to meet FPGA timing
Luke Kenneth Casson Leighton [Sun, 20 Feb 2022 23:38:13 +0000 (23:38 +0000)]
put LDST go-store on a 1-clock delay to help with combinatorial chains
Luke Kenneth Casson Leighton [Sun, 20 Feb 2022 23:28:50 +0000 (23:28 +0000)]
name core_stop and terminated_o synchronous to potentially help
cut down on combinatorial chains
Luke Kenneth Casson Leighton [Sun, 20 Feb 2022 22:20:21 +0000 (22:20 +0000)]
nope, it's perfectly fine
Revert "weird exception, oe not found in the shiftrot input record"
This reverts commit
264cc7fd7d7547e1e19424b8f8fd0fbfea29cec5.
Luke Kenneth Casson Leighton [Sun, 20 Feb 2022 22:17:18 +0000 (22:17 +0000)]
weird exception, oe not found in the shiftrot input record
Revert "separate out shiftrot stages due to size of main stage being so large"
This reverts commit
87d6b84e96fb62cda16cc9f335e34fe15ad6cd97.
Luke Kenneth Casson Leighton [Sun, 20 Feb 2022 22:09:42 +0000 (22:09 +0000)]
separate out shiftrot stages due to size of main stage being so large
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 20:50:22 +0000 (20:50 +0000)]
add blockram style to regfile Memory
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 20:49:36 +0000 (20:49 +0000)]
use block_ram attribute for FPGA synthesis
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 19:42:44 +0000 (19:42 +0000)]
reduce number of d-cache lines in microwatt fpga mode
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 19:41:42 +0000 (19:41 +0000)]
couple of adjustments to reduce gate count in i/d-cache
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 19:41:23 +0000 (19:41 +0000)]
add SDRAM Configuration Record
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 11:53:06 +0000 (11:53 +0000)]
reduce TLB set size from 64 to 16 to get FPGA resource utilisation down
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 11:41:29 +0000 (11:41 +0000)]
drastically reduce I-Cache size in microwatt-compat mode
should really be called "small FPGA mode" or something
Luke Kenneth Casson Leighton [Fri, 18 Feb 2022 11:36:19 +0000 (11:36 +0000)]
parameterise I-Cache similar to D-Cache. lots of "self."
all over the place. yuk.
Jacob Lifshay [Fri, 18 Feb 2022 06:01:41 +0000 (22:01 -0800)]
add grev
Luke Kenneth Casson Leighton [Thu, 17 Feb 2022 17:09:21 +0000 (17:09 +0000)]
add opencores SDRAM verilog wrapper
Luke Kenneth Casson Leighton [Wed, 16 Feb 2022 16:18:05 +0000 (16:18 +0000)]
oof. big update to DCache to accept config parameters
Luke Kenneth Casson Leighton [Wed, 16 Feb 2022 16:17:22 +0000 (16:17 +0000)]
connect UART16550 pins if given
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 20:07:35 +0000 (20:07 +0000)]
for *write* the counter-address on downconvert was correct
but for *read* it has to be pre-advanced (if that makes any sense)
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 19:35:14 +0000 (19:35 +0000)]
add wishbone downconvert "skip" of slave sel so that action
is not taken
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 19:24:37 +0000 (19:24 +0000)]
add SysCon reg_info, has uart and has large SYSCON
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 19:20:36 +0000 (19:20 +0000)]
sigh, stall was not working but actually turns out that
the downconvert counter was not being set properly (one cycle late)
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 15:28:48 +0000 (15:28 +0000)]
add option to specify UART16550 width (32/8)
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 13:51:53 +0000 (13:51 +0000)]
add beginnings of syscon bus peripheral
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 11:42:26 +0000 (11:42 +0000)]
update comments
Luke Kenneth Casson Leighton [Tue, 15 Feb 2022 01:35:06 +0000 (01:35 +0000)]
resolve WBDownConvert ack issues when stall is active
Luke Kenneth Casson Leighton [Mon, 14 Feb 2022 14:22:48 +0000 (14:22 +0000)]
strip first 3 bits of WB address from microwatt d/i-bus, bug in microwatt
Luke Kenneth Casson Leighton [Mon, 14 Feb 2022 14:22:14 +0000 (14:22 +0000)]
slave sends stall signal, master receives, in
WBDownConvert
Luke Kenneth Casson Leighton [Mon, 14 Feb 2022 14:02:59 +0000 (14:02 +0000)]
sort out ExternalCore signal names
Luke Kenneth Casson Leighton [Mon, 14 Feb 2022 14:02:40 +0000 (14:02 +0000)]
add wishbone slave signal to downconvert if present
Luke Kenneth Casson Leighton [Mon, 14 Feb 2022 13:08:56 +0000 (13:08 +0000)]
add external core verilog wrapper, ironically around Libre-SOC
(as well as Microwatt)
Luke Kenneth Casson Leighton [Sun, 13 Feb 2022 14:24:58 +0000 (14:24 +0000)]
bugfixing for ls2 imports of uart16550
Luke Kenneth Casson Leighton [Thu, 10 Feb 2022 18:06:20 +0000 (18:06 +0000)]
Revert "remove dummy trap pipeline"
This reverts commit
de3765400532be229ed4dd8a1d9fdcf1b4bca0ef.
Luke Kenneth Casson Leighton [Thu, 10 Feb 2022 18:06:10 +0000 (18:06 +0000)]
Andrey Miroshnikov [Thu, 10 Feb 2022 15:50:20 +0000 (15:50 +0000)]
Added optional reverse arg to send TDI data MSB-first
Luke Kenneth Casson Leighton [Wed, 9 Feb 2022 12:22:37 +0000 (12:22 +0000)]
add opencores uart16550 instance wrapper
Tobias Platen [Tue, 1 Feb 2022 18:13:33 +0000 (18:13 +0000)]
correct path for make target microwatt_external_core
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 20:38:45 +0000 (20:38 +0000)]
fix bug in itlb_valid SRLatch set/reset, a bit weird but it works
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 16:09:12 +0000 (16:09 +0000)]
whoops tlb_valids in ICache is a combinatorial-get/set
set SRLatch sync=False
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 15:57:56 +0000 (15:57 +0000)]
convert TLBValidArray in ICache to SRLatch
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 15:47:06 +0000 (15:47 +0000)]
add microwatt external core build target to Makefile
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 15:29:44 +0000 (15:29 +0000)]
use an SRLatch for cache_valids, at least it reduces graphviz size
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 15:06:56 +0000 (15:06 +0000)]
use Memory for cache tags in dcache
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 15:06:36 +0000 (15:06 +0000)]
use Memory for cache_tags in icache
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 15:05:37 +0000 (15:05 +0000)]
doh
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 15:02:37 +0000 (15:02 +0000)]
remove dummy trap pipeline
Luke Kenneth Casson Leighton [Mon, 31 Jan 2022 14:59:45 +0000 (14:59 +0000)]
remove combinatorial loop from MultiCompUnit
actually not a loop due to an SRLatch but synth tools still think it is
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 22:18:26 +0000 (22:18 +0000)]
break out cache_tags and cache_valids (again) this time in dcache
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 22:08:38 +0000 (22:08 +0000)]
remove CacheTagArray in icache.py
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 22:03:24 +0000 (22:03 +0000)]
create Memory for Cache Tags in I-Cache
another huge reduction in number of LUT4s, uses (again) a
combinatorial-read sync-write
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 21:30:34 +0000 (21:30 +0000)]
remove unneeded parameter
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 21:28:52 +0000 (21:28 +0000)]
add Array of CacheValids back in, so as to reduce LUT4 usage
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 21:11:09 +0000 (21:11 +0000)]
tagset is a local Signal in ICache
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 21:10:55 +0000 (21:10 +0000)]
identify combinatorial loop signals in MultiCompUnit, TODO resolve
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 18:16:37 +0000 (18:16 +0000)]
use nmigen Memory in I-Cache for TLB Lookups
surprisingly this makes the Libre-SOC core *50% faster* than microwatt
when running under verilator, despite only being a FSM
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 17:49:17 +0000 (17:49 +0000)]
put itlb_valid back, ready for conversion to Memory, in ICache
Luke Kenneth Casson Leighton [Sun, 30 Jan 2022 16:51:41 +0000 (16:51 +0000)]
convert CacheRAM to Memory, acts much faster now