Github_Repos/cvw
1
0
Fork 1
mirror of https://github.com/openhwgroup/cvw synced 2025-02-11 06:05:49 +00:00
Code Issues Packages Projects Releases Wiki Activity

Merge branch 'main' of https://github.com/openhwgroup/cvw

Browse Source
This commit is contained in:
Ross Thompson 2023-06-15 14:57:23 -05:00
commit d2219023c3
98 changed files with 1848 additions and 1762 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

36
Makefile
View File

@ -4,7 +4,11 @@
all:
make install
make regression
make testfloat
make riscof
make verify
make coverage
make benchmarks
# install copies over the Makefile.include from riscv-isa-sim
# And corrects the TARGETDIR path and the RISCV_PREFIX
@ -18,9 +22,37 @@ install:
##rm tests/imperas-riscv-tests/riscv-ovpsim-plus/bin/Linux64/riscvOVPsimPlus.exe
##ln -s ${RISCV}/imperas-riscv-tests/riscv-ovpsim-plus/bin/Linux64/riscvOVPsimPlus.exe tests/imperas-riscv-tests/riscv-ovpsim-plus/bin/Linux64/riscvOVPsimPlus.exe
regression:
riscof:
make -C sim
testfloat:
cd ${WALLY}/addins/SoftFloat-3e/build/Linux-x86_64-GCC; make
cd ${WALLY}/addins/TestFloat-3e/build/Linux-x86_64-GCC; make
cd ${WALLY}/tests/fp; ./create_all_vectors.sh
verify:
cd ${WALLY}/sim; ./regression-wally
cd ${WALLY}/sim; ./sim-testfloat-batch all
make imperasdv
imperasdv:
iter-elf.bash --search ${WALLY}/tests/riscof/work/wally-riscv-arch-test/rv64i_m
iter-elf.bash --search ${WALLY}/tests/riscof/work/riscv-arch-test/rv64i_m
coverage:
cd ${WALLY}/sim; ./regresssion-wally -coverage -fp
benchmarks:
make coremark
make embench
coremark:
cd ${WALLY}/benchmarks/coremark; make; make run
embench:
cd ${WALLY}/benchmarks/embench; make; make run
clean:
make clean -C sim

6
benchmarks/coremark/Makefile
View File

@ -11,7 +11,11 @@ sources=$(cmbase)/core_main.c $(cmbase)/core_list_join.c $(cmbase)/coremark.h \
$(PORT_DIR)/core_portme.h $(PORT_DIR)/core_portme.c $(PORT_DIR)/core_portme.mak \
$(PORT_DIR)/crt.S $(PORT_DIR)/encoding.h $(PORT_DIR)/util.h $(PORT_DIR)/syscalls.c
ABI := $(if $(findstring "64","$(XLEN)"),lp64,ilp32)
ARCH := rv$(XLEN)gc
#ARCH := rv$(XLEN)gc_zba_zbb_zbc_zbs
#ARCH := rv$(XLEN)gc
ARCH := rv$(XLEN)imc_zicsr
#ARCH := rv$(XLEN)im_zicsr
#ARCH := rv$(XLEN)i_zicsr
PORT_CFLAGS = -g -mabi=$(ABI) -march=$(ARCH) -static -falign-functions=16 \
-mbranch-cost=1 -DSKIP_DEFAULT_MEMSET -mtune=sifive-3-series -O3 -finline-functions -falign-jumps=4 \
-fno-delete-null-pointer-checks -fno-rename-registers --param=loop-max-datarefs-for-datadeps=0 \

19
bin/wally-tool-chain-install.sh
View File

@ -45,7 +45,7 @@ sudo mkdir -p $RISCV
# Update and Upgrade tools (see https://itsfoss.com/apt-update-vs-upgrade/)
sudo apt update -y
sudo apt upgrade -y
sudo apt install -y git gawk make texinfo bison flex build-essential python3 libz-dev libexpat-dev autoconf device-tree-compiler ninja-build libpixman-1-dev ncurses-base ncurses-bin libncurses5-dev dialog curl wget ftp libgmp-dev libglib2.0-dev python3-pip pkg-config opam z3 zlib1g-dev verilator automake autotools-dev libmpc-dev libmpfr-dev gperf libtool patchutils bc
sudo apt install -y git gawk make texinfo bison flex build-essential python3 libz-dev libexpat-dev autoconf device-tree-compiler ninja-build libpixman-1-dev ncurses-base ncurses-bin libncurses5-dev dialog curl wget ftp libgmp-dev libglib2.0-dev python3-pip pkg-config opam z3 zlib1g-dev automake autotools-dev libmpc-dev libmpfr-dev gperf libtool patchutils bc
# Other python libraries used through the book.
sudo pip3 install matplotlib scipy scikit-learn adjustText lief
@ -113,6 +113,23 @@ cd ../arch_test_target/spike/device
sed -i 's/--isa=rv32ic/--isa=rv32iac/' rv32i_m/privilege/Makefile.include
sed -i 's/--isa=rv64ic/--isa=rv64iac/' rv64i_m/privilege/Makefile.include
# Wally needs Verilator 5.0 or later.
# Verilator needs to be built from scratch to get the latest version
# apt-get install verilator installs version 4.028 as of 6/8/23
sudo apt-get install -y perl g++ ccache help2man libgoogle-perftools-dev numactl perl-doc zlibc zlib1g
sudo apt-get install -y libfl2 libfl-dev # Ubuntu only (ignore if gives error)
cd $RISCV
git clone https://github.com/verilator/verilator # Only first time
unsetenv VERILATOR_ROOT # For csh; ignore error if on bash
unset VERILATOR_ROOT # For bash
cd verilator
git pull # Make sure git repository is up-to-date
git checkout master # Use development branch (e.g. recent bug fixes)
autoconf # Create ./configure script
./configure # Configure and create Makefile
make -j NUM_THREADS # Build Verilator itself (if error, try just 'make')
sudo make install
# Sail (https://github.com/riscv/sail-riscv)
# Sail is the new golden reference model for RISC-V. Sail is written in OCaml, which
# is an object-oriented extension of ML, which in turn is a functional programming

8
config/buildroot/config.vh
View File

@ -40,10 +40,16 @@ localparam IEEE754 = 0;
localparam MISA = (32'h0014112D);
localparam ZICSR_SUPPORTED = 1;
localparam ZIFENCEI_SUPPORTED = 1;
localparam ZICOUNTERS_SUPPORTED = 1;
localparam ZICNTR_SUPPORTED = 1;
localparam ZIHPM_SUPPORTED = 1;
localparam COUNTERS = 12'd32;
localparam ZFH_SUPPORTED = 0;
localparam SSTC_SUPPORTED = 0;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 1;

7
config/rv32e/config.vh
View File

@ -41,9 +41,14 @@ localparam MISA = (32'h00000010);
localparam ZICSR_SUPPORTED = 0;
localparam ZIFENCEI_SUPPORTED = 0;
localparam COUNTERS = 12'd0;
localparam ZICOUNTERS_SUPPORTED = 0;
localparam ZICNTR_SUPPORTED = 0;
localparam ZIHPM_SUPPORTED = 0;
localparam ZFH_SUPPORTED = 0;
localparam SSTC_SUPPORTED = 0;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 1;

7
config/rv32gc/config.vh
View File

@ -42,9 +42,14 @@ localparam MISA = (32'h00000104 | 1 << 20 | 1 << 18 | 1 << 12 | 1 << 0 | 1 <<3 |
localparam ZICSR_SUPPORTED = 1;
localparam ZIFENCEI_SUPPORTED = 1;
localparam COUNTERS = 12'd32;
localparam ZICOUNTERS_SUPPORTED = 1;
localparam ZICNTR_SUPPORTED = 1;
localparam ZIHPM_SUPPORTED = 1;
localparam ZFH_SUPPORTED = 0;
localparam SSTC_SUPPORTED = 1;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 1;

9
config/rv32i/config.vh
View File

@ -40,10 +40,15 @@ localparam IEEE754 = 0;
localparam MISA = (32'h00000104);
localparam ZICSR_SUPPORTED = 0;
localparam ZIFENCEI_SUPPORTED = 0;
localparam COUNTERS = 12'd32;
localparam ZICOUNTERS_SUPPORTED = 0;
localparam COUNTERS = 0;
localparam ZICNTR_SUPPORTED = 0;
localparam ZIHPM_SUPPORTED = 0;
localparam ZFH_SUPPORTED = 0;
localparam SSTC_SUPPORTED = 0;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 0;

7
config/rv32imc/config.vh
View File

@ -40,9 +40,14 @@ localparam MISA = (32'h00000104 | 1 << 20 | 1 << 18 | 1 << 12);
localparam ZICSR_SUPPORTED = 1;
localparam ZIFENCEI_SUPPORTED = 1;
localparam COUNTERS = 12'd32;
localparam ZICOUNTERS_SUPPORTED = 1;
localparam ZICNTR_SUPPORTED = 1;
localparam ZIHPM_SUPPORTED = 1;
localparam ZFH_SUPPORTED = 0;
localparam SSTC_SUPPORTED = 0;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 1;

7
config/rv64fpquad/config.vh
View File

@ -41,9 +41,14 @@ localparam MISA = (32'h00000104 | 1 << 5 | 1 << 3 | 1 << 16 | 1 << 18 | 1 << 20
localparam ZICSR_SUPPORTED = 1;
localparam ZIFENCEI_SUPPORTED = 1;
localparam COUNTERS = 12'd32;
localparam ZICOUNTERS_SUPPORTED = 1;
localparam ZICNTR_SUPPORTED = 1;
localparam ZIHPM_SUPPORTED = 1;
localparam ZFH_SUPPORTED = 1;
localparam SSTC_SUPPORTED = 0;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 1;

7
config/rv64gc/config.vh
View File

@ -44,9 +44,14 @@ localparam MISA = (32'h00000104 | 1 << 5 | 1 << 3 | 1 << 18 | 1 << 20 | 1 << 12
localparam ZICSR_SUPPORTED = 1;
localparam ZIFENCEI_SUPPORTED = 1;
localparam COUNTERS = 12'd32;
localparam ZICOUNTERS_SUPPORTED = 1;
localparam ZICNTR_SUPPORTED = 1;
localparam ZIHPM_SUPPORTED = 1;
localparam ZFH_SUPPORTED = 0;
localparam SSTC_SUPPORTED = 1;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 1;

9
config/rv64i/config.vh
View File

@ -40,10 +40,15 @@ localparam IEEE754 = 0;
localparam MISA = (32'h00000104);
localparam ZICSR_SUPPORTED = 0;
localparam ZIFENCEI_SUPPORTED = 0;
localparam COUNTERS = 12'd32;
localparam ZICOUNTERS_SUPPORTED = 0;
localparam COUNTERS = 0;
localparam ZICNTR_SUPPORTED = 0;
localparam ZIHPM_SUPPORTED = 0;
localparam ZFH_SUPPORTED = 0;
localparam SSTC_SUPPORTED = 0;
localparam ZICBOM_SUPPORTED = 0;
localparam ZICBOZ_SUPPORTED = 0;
localparam ZICBOP_SUPPORTED = 0;
localparam SVPBMT_SUPPORTED = 0;
// LSU microarchitectural Features
localparam BUS_SUPPORTED = 0;

27
config/shared/parameter-defs.vh
View File

@ -11,20 +11,25 @@ parameter cvw_t P = '{
ZICSR_SUPPORTED : ZICSR_SUPPORTED,
ZIFENCEI_SUPPORTED : ZIFENCEI_SUPPORTED,
COUNTERS : COUNTERS,
ZICOUNTERS_SUPPORTED : ZICOUNTERS_SUPPORTED,
ZICNTR_SUPPORTED : ZICNTR_SUPPORTED,
ZIHPM_SUPPORTED : ZIHPM_SUPPORTED,
ZFH_SUPPORTED : ZFH_SUPPORTED,
SSTC_SUPPORTED : SSTC_SUPPORTED,
VIRTMEM_SUPPORTED : VIRTMEM_SUPPORTED,
VECTORED_INTERRUPTS_SUPPORTED : VECTORED_INTERRUPTS_SUPPORTED,
BIGENDIAN_SUPPORTED : BIGENDIAN_SUPPORTED,
SVADU_SUPPORTED : SVADU_SUPPORTED,
ZMMUL_SUPPORTED : ZMMUL_SUPPORTED,
VIRTMEM_SUPPORTED : VIRTMEM_SUPPORTED,
VECTORED_INTERRUPTS_SUPPORTED : VECTORED_INTERRUPTS_SUPPORTED,
BIGENDIAN_SUPPORTED : BIGENDIAN_SUPPORTED,
SVADU_SUPPORTED : SVADU_SUPPORTED,
ZMMUL_SUPPORTED : ZMMUL_SUPPORTED,
ZICBOM_SUPPORTED : ZICBOM_SUPPORTED,
ZICBOZ_SUPPORTED : ZICBOZ_SUPPORTED,
ZICBOP_SUPPORTED : ZICBOP_SUPPORTED,
SVPBMT_SUPPORTED : SVPBMT_SUPPORTED,
BUS_SUPPORTED : BUS_SUPPORTED,
DCACHE_SUPPORTED : DCACHE_SUPPORTED,
ICACHE_SUPPORTED : ICACHE_SUPPORTED,
ITLB_ENTRIES : ITLB_ENTRIES,
DTLB_ENTRIES : DTLB_ENTRIES,
DCACHE_NUMWAYS : DCACHE_NUMWAYS,
DCACHE_SUPPORTED : DCACHE_SUPPORTED,
ICACHE_SUPPORTED : ICACHE_SUPPORTED,
ITLB_ENTRIES : ITLB_ENTRIES,
DTLB_ENTRIES : DTLB_ENTRIES,
DCACHE_NUMWAYS : DCACHE_NUMWAYS,
DCACHE_WAYSIZEINBYTES : DCACHE_WAYSIZEINBYTES,
DCACHE_LINELENINBITS : DCACHE_LINELENINBITS,
ICACHE_NUMWAYS : ICACHE_NUMWAYS,

60
sim/FPbuild.txt Normal file
View File

@ -0,0 +1,60 @@
Procedure for Runnning SoftFloat/TestFloat with Wally
1.) First, compile SoftFloat and TestFloat by going to the addins
directory and finding the specific build directory (e.g.,
Linux_x86_64-GCC. Currently, we are using v3e of
SoftFloat/TestFloat. I am not sure of the order, but I always compile
SoftFloat first as I believe TestFloat uses the static library
SoftFloat creates.
2.) Once compiled both, go to the tests/fp directory and run the
create_vectors.sh Linux script. In the past, we have automated this,
but I believe this has fallen into more of a manual state lately.
3.) Then, run remove_spaces.sh which will remove spaces from the
output and put underscores between vectors (this helps differentiate
the vectors that are generated). Again, this can be combined with
Step 2.
4.) TestFloat is run from wally/cvw/sim and sim-testfloat-batch with
its respective test. The format is ./sim-testfloat-add add. All of
the tests are listed below. This can be augmented or added to for
other FP tests given by the great SoftFloat/TestFloat output.
cvtint - test integer conversion unit (fcvtint)
cvtfp - test floating-point conversion unit (fcvtfp)
cmp - test comparison unit's LT, LE, EQ opperations (fcmp)
add - test addition
fma - test fma
mul - test mult with fma
sub - test subtraction
div - test division
sqrt - test square root
all - test everything
4a.) Each test will test all its vectors - if you want to test a
subset of the vectors (e.g., only binary16), you should modify the
cvw/testbench/tests-fp.h and comment out the tests you do not want to
test. The best way to do this is to comment out each item out with
the // comment option in SV. For example,
string f128div[] = '{
// "f128_div_rne.tv",
// "f128_div_rz.tv",
// "f128_div_ru.tv",
// "f128_div_rd.tv",
// "f128_div_rnm.tv"
};

3
src/cache/cache.sv vendored
View File

@ -63,7 +63,7 @@ module cache #(parameter PA_BITS, XLEN, LINELEN, NUMLINES, NUMWAYS, LOGBWPL, W
localparam OFFSETLEN = $clog2(LINEBYTELEN); // Number of bits in offset field
localparam SETLEN = $clog2(NUMLINES); // Number of set bits
localparam SETTOP = SETLEN+OFFSETLEN; // Number of set plus offset bits
localparam TAGLEN = PA_BITS - SETTOP; // Number of tag bits
localparam TAGLEN = PA_BITS - SETTOP; // Number of tag bits
localparam CACHEWORDSPERLINE = LINELEN/WORDLEN;// Number of words in cache line
localparam LOGCWPL = $clog2(CACHEWORDSPERLINE);// Log2 of ^
localparam FLUSHADRTHRESHOLD = NUMLINES - 1; // Used to determine when flush is complete
@ -182,6 +182,7 @@ module cache #(parameter PA_BITS, XLEN, LINELEN, NUMLINES, NUMWAYS, LOGBWPL, W
assign LineWriteData = FetchBuffer;
assign LineByteMask = '1;
end
/////////////////////////////////////////////////////////////////////////////////////////////
// Flush logic
/////////////////////////////////////////////////////////////////////////////////////////////

3
src/cache/cacheLRU.sv vendored
View File

@ -104,8 +104,7 @@ module cacheLRU
if (node == NUMWAYS-2) begin
assign LRUUpdate[lchild] = ~WayEncoded[r];
assign LRUUpdate[rchild] = WayEncoded[r];
end
else begin
end else begin
assign LRUUpdate[lchild] = LRUUpdate[node] & ~WayEncoded[r];
assign LRUUpdate[rchild] = LRUUpdate[node] & WayEncoded[r];
end

14
src/cache/cacheway.sv vendored
View File

@ -84,8 +84,7 @@ module cacheway #(parameter PA_BITS, XLEN, NUMLINES=512, LINELEN = 256, TAGLEN =
// nonzero ways will never see SelFlush=0 while FlushWay=1 since FlushWay only advances on a subset of SelFlush assertion cases.
assign FlushWayEn = FlushWay & SelFlush;
assign SelNonHit = FlushWayEn | SetValid | SelWriteback;
end
else begin:flushlogic // no flush operation for read-only caches.
end else begin:flushlogic // no flush operation for read-only caches.
assign SelTag = VictimWay;
assign SelNonHit = SetValid;
end
@ -135,8 +134,7 @@ module cacheway #(parameter PA_BITS, XLEN, NUMLINES=512, LINELEN = 256, TAGLEN =
.dout(ReadDataLine[SRAMLEN*(words+1)-1:SRAMLEN*words]),
.din(LineWriteData[SRAMLEN*(words+1)-1:SRAMLEN*words]),
.we(SelectedWriteWordEn), .bwe(FinalByteMask[SRAMLENINBYTES*(words+1)-1:SRAMLENINBYTES*words]));
end
else begin:wordram // no byte-enable needed for i$.
end else begin:wordram // no byte-enable needed for i$.
ram1p1rwe #(.DEPTH(NUMLINES), .WIDTH(SRAMLEN)) CacheDataMem(.clk, .ce(CacheEn), .addr(CacheSet),
.dout(ReadDataLine[SRAMLEN*(words+1)-1:SRAMLEN*words]),
.din(LineWriteData[SRAMLEN*(words+1)-1:SRAMLEN*words]),
@ -154,8 +152,8 @@ module cacheway #(parameter PA_BITS, XLEN, NUMLINES=512, LINELEN = 256, TAGLEN =
always_ff @(posedge clk) begin // Valid bit array,
if (reset) ValidBits <= #1 '0;
if(CacheEn) begin
ValidWay <= #1 ValidBits[CacheSet];
if(InvalidateCache) ValidBits <= #1 '0; // exclusion-tag: dcache invalidateway
ValidWay <= #1 ValidBits[CacheSet];
if(InvalidateCache) ValidBits <= #1 '0; // exclusion-tag: dcache invalidateway
else if (SetValidEN) ValidBits[CacheSet] <= #1 SetValidWay;
end
end
@ -175,8 +173,4 @@ module cacheway #(parameter PA_BITS, XLEN, NUMLINES=512, LINELEN = 256, TAGLEN =
end
end
end else assign Dirty = 1'b0;
endmodule

5
src/cache/subcachelineread.sv vendored
View File

@ -43,9 +43,8 @@ module subcachelineread #(parameter LINELEN, WORDLEN,
logic [LINELEN+(WORDLEN-MUXINTERVAL)-1:0] ReadDataLinePad;
logic [WORDLEN-1:0] ReadDataLineSets [(LINELEN/MUXINTERVAL)-1:0];
if (PADLEN > 0) begin
assign ReadDataLinePad = {{PADLEN{1'b0}}, ReadDataLine};
end else assign ReadDataLinePad = ReadDataLine;
if (PADLEN > 0) assign ReadDataLinePad = {{PADLEN{1'b0}}, ReadDataLine};
else assign ReadDataLinePad = ReadDataLine;
genvar index;
for (index = 0; index < WORDSPERLINE; index++) begin:readdatalinesetsmux

22
src/ebu/ahbcacheinterface.sv
View File

@ -45,14 +45,14 @@ module ahbcacheinterface #(
output logic [2:0] HSIZE, // AHB transaction width
output logic [2:0] HBURST, // AHB burst length
// bus interface buses
input logic [AHBW-1:0] HRDATA, // AHB read data
output logic [PA_BITS-1:0] HADDR, // AHB address
output logic [AHBW-1:0] HWDATA, // AHB write data
output logic [AHBW/8-1:0] HWSTRB, // AHB byte mask
input logic [AHBW-1:0] HRDATA, // AHB read data
output logic [PA_BITS-1:0] HADDR, // AHB address
output logic [AHBW-1:0] HWDATA, // AHB write data
output logic [AHBW/8-1:0] HWSTRB, // AHB byte mask
// cache interface
input logic [PA_BITS-1:0] CacheBusAdr, // Address of cache line
input logic [LLEN-1:0] CacheReadDataWordM, // One word of cache line during a writeback
input logic [PA_BITS-1:0] CacheBusAdr, // Address of cache line
input logic [LLEN-1:0] CacheReadDataWordM, // One word of cache line during a writeback
input logic CacheableOrFlushCacheM, // Memory operation is cacheable or flushing D$
input logic Cacheable, // Memory operation is cachable
input logic [1:0] CacheBusRW, // Cache bus operation, 01: writeback, 10: fetch
@ -62,8 +62,8 @@ module ahbcacheinterface #(
output logic SelBusBeat, // Tells the cache to select the word from ReadData or WriteData from BeatCount rather than PAdr
// uncached interface
input logic [PA_BITS-1:0] PAdr, // Physical address of uncached memory operation
input logic [LLEN-1:0] WriteDataM, // IEU write data for uncached store
input logic [PA_BITS-1:0] PAdr, // Physical address of uncached memory operation
input logic [LLEN-1:0] WriteDataM, // IEU write data for uncached store
input logic [1:0] BusRW, // Uncached memory operation read/write control: 10: read, 01: write
input logic [2:0] Funct3, // Size of uncached memory operation
@ -75,11 +75,11 @@ module ahbcacheinterface #(
localparam BeatCountThreshold = BEATSPERLINE - 1; // Largest beat index
logic [PA_BITS-1:0] LocalHADDR; // Address after selecting between cached and uncached operation
logic [PA_BITS-1:0] LocalHADDR; // Address after selecting between cached and uncached operation
logic [AHBWLOGBWPL-1:0] BeatCountDelayed; // Beat within the cache line in the second (Data) cache stage
logic CaptureEn; // Enable updating the Fetch buffer with valid data from HRDATA
logic [AHBW/8-1:0] BusByteMaskM; // Byte enables within a word. For cache request all 1s
logic [AHBW-1:0] PreHWDATA; // AHB Address phase write data
logic [AHBW/8-1:0] BusByteMaskM; // Byte enables within a word. For cache request all 1s
logic [AHBW-1:0] PreHWDATA; // AHB Address phase write data
genvar index;

12
src/ebu/ahbinterface.sv
View File

@ -36,19 +36,19 @@ module ahbinterface #(
input logic HREADY, // AHB peripheral ready
output logic [1:0] HTRANS, // AHB transaction type, 00: IDLE, 10 NON_SEQ, 11 SEQ
output logic HWRITE, // AHB 0: Read operation 1: Write operation
input logic [XLEN-1:0] HRDATA, // AHB read data
output logic [XLEN-1:0] HWDATA, // AHB write data
output logic [XLEN/8-1:0] HWSTRB, // AHB byte mask
input logic [XLEN-1:0] HRDATA, // AHB read data
output logic [XLEN-1:0] HWDATA, // AHB write data
output logic [XLEN/8-1:0] HWSTRB, // AHB byte mask
// lsu/ifu interface
input logic Stall, // Core pipeline is stalled
input logic Flush, // Pipeline stage flush. Prevents bus transaction from starting
input logic [1:0] BusRW, // Memory operation read/write control: 10: read, 01: write
input logic [XLEN/8-1:0] ByteMask, // Bytes enables within a word
input logic [XLEN-1:0] WriteData, // IEU write data for a store
input logic [XLEN/8-1:0] ByteMask, // Bytes enables within a word
input logic [XLEN-1:0] WriteData, // IEU write data for a store
output logic BusStall, // Bus is busy with an in flight memory operation
output logic BusCommitted, // Bus is busy with an in flight memory operation and it is not safe to take an interrupt
output logic [(LSU ? XLEN : 32)-1:0] FetchBuffer // Register to hold HRDATA after arriving from the bus
output logic [(LSU ? XLEN : 32)-1:0] FetchBuffer // Register to hold HRDATA after arriving from the bus
);
logic CaptureEn;

9
src/ebu/controllerinput.sv
View File

@ -46,14 +46,14 @@ module controllerinput #(
input logic HWRITEIn, // Manager input. AHB 0: Read operation 1: Write operation
input logic [2:0] HSIZEIn, // Manager input. AHB transaction width
input logic [2:0] HBURSTIn, // Manager input. AHB burst length
input logic [PA_BITS-1:0] HADDRIn, // Manager input. AHB address
input logic [PA_BITS-1:0] HADDRIn, // Manager input. AHB address
output logic HREADYOut, // Indicate to manager the peripheral is not busy and another manager does not have priority
// controller output
output logic [1:0] HTRANSOut, // Arbitrated manager transaction. AHB transaction type, 00: IDLE, 10 NON_SEQ, 11 SEQ
output logic HWRITEOut, // Arbitrated manager transaction. AHB 0: Read operation 1: Write operation
output logic [2:0] HSIZEOut, // Arbitrated manager transaction. AHB transaction width
output logic [2:0] HBURSTOut, // Arbitrated manager transaction. AHB burst length
output logic [PA_BITS-1:0] HADDROut, // Arbitrated manager transaction. AHB address
output logic [PA_BITS-1:0] HADDROut, // Arbitrated manager transaction. AHB address
input logic HREADYIn // Peripheral ready
);
@ -61,7 +61,7 @@ module controllerinput #(
logic [2:0] HSIZESave;
logic [2:0] HBURSTSave;
logic [1:0] HTRANSSave;
logic [PA_BITS-1:0] HADDRSave;
logic [PA_BITS-1:0] HADDRSave;
if (SAVE_ENABLED) begin
flopenr #(1+3+3+2+PA_BITS) SaveReg(HCLK, ~HRESETn, Save,
@ -83,6 +83,3 @@ module controllerinput #(
assign HREADYOut = HREADYIn & ~Disable;
endmodule

18
src/ebu/ebu.sv
View File

@ -37,25 +37,25 @@ module ebu #(parameter XLEN, PA_BITS, AHBW)(
input logic [1:0] IFUHTRANS, // IFU AHB transaction request
input logic [2:0] IFUHSIZE, // IFU AHB transaction size
input logic [2:0] IFUHBURST, // IFU AHB burst length
input logic [PA_BITS-1:0] IFUHADDR, // IFU AHB address
input logic [PA_BITS-1:0] IFUHADDR, // IFU AHB address
output logic IFUHREADY, // AHB peripheral ready gated by possible non-grant
// Signals from LSU
input logic [1:0] LSUHTRANS, // LSU AHB transaction request
input logic LSUHWRITE, // LSU AHB transaction direction. 1: write, 0: read
input logic [2:0] LSUHSIZE, // LSU AHB size
input logic [2:0] LSUHBURST, // LSU AHB burst length
input logic [PA_BITS-1:0] LSUHADDR, // LSU AHB address
input logic [XLEN-1:0] LSUHWDATA, // initially support AHBW = XLEN
input logic [XLEN/8-1:0] LSUHWSTRB, // AHB byte mask
input logic [PA_BITS-1:0] LSUHADDR, // LSU AHB address
input logic [XLEN-1:0] LSUHWDATA, // initially support AHBW = XLEN
input logic [XLEN/8-1:0] LSUHWSTRB, // AHB byte mask
output logic LSUHREADY, // AHB peripheral. Never gated as LSU always has priority
// AHB-Lite external signals
output logic HCLK, HRESETn,
input logic HREADY, // AHB peripheral ready
input logic HRESP, // AHB peripheral response. 0: OK 1: Error
output logic [PA_BITS-1:0] HADDR, // AHB address to peripheral after arbitration
output logic [AHBW-1:0] HWDATA, // AHB Write data after arbitration
output logic [XLEN/8-1:0] HWSTRB, // AHB byte write enables after arbitration
output logic [PA_BITS-1:0] HADDR, // AHB address to peripheral after arbitration
output logic [AHBW-1:0] HWDATA, // AHB Write data after arbitration
output logic [XLEN/8-1:0] HWSTRB, // AHB byte write enables after arbitration
output logic HWRITE, // AHB transaction direction after arbitration
output logic [2:0] HSIZE, // AHB transaction size after arbitration
output logic [2:0] HBURST, // AHB burst length after arbitration
@ -71,13 +71,13 @@ module ebu #(parameter XLEN, PA_BITS, AHBW)(
logic IFUDisable;
logic IFUSelect;
logic [PA_BITS-1:0] IFUHADDROut;
logic [PA_BITS-1:0] IFUHADDROut;
logic [1:0] IFUHTRANSOut;
logic [2:0] IFUHBURSTOut;
logic [2:0] IFUHSIZEOut;
logic IFUHWRITEOut;
logic [PA_BITS-1:0] LSUHADDROut;
logic [PA_BITS-1:0] LSUHADDROut;
logic [1:0] LSUHTRANSOut;
logic [2:0] LSUHBURSTOut;
logic [2:0] LSUHSIZEOut;

2
src/fpu/fclassify.sv
View File

@ -33,7 +33,7 @@ module fclassify import cvw::*; #(parameter cvw_t P) (
input logic XSubnorm, // is Subnormal
input logic XZero, // is zero
input logic XInf, // is infinity
output logic [P.XLEN-1:0] ClassRes // classify result
output logic [P.XLEN-1:0] ClassRes // classify result
);
logic PInf, PZero, PNorm, PSubnorm; // is the input a positive infinity/zero/normal/subnormal

26
src/fpu/fcmp.sv
View File

@ -1,4 +1,3 @@
///////////////////////////////////////////
// fcmp.sv
//
@ -36,23 +35,23 @@
module fcmp import cvw::*; #(parameter cvw_t P) (
input logic [P.FMTBITS-1:0] Fmt, // format of fp number
input logic [2:0] OpCtrl, // see above table
input logic Xs, Ys, // input signs
input logic [2:0] OpCtrl, // see above table
input logic Xs, Ys, // input signs
input logic [P.NE-1:0] Xe, Ye, // input exponents
input logic [P.NF:0] Xm, Ym, // input mantissa
input logic XZero, YZero, // is zero
input logic XNaN, YNaN, // is NaN
input logic XSNaN, YSNaN, // is signaling NaN
input logic XZero, YZero, // is zero
input logic XNaN, YNaN, // is NaN
input logic XSNaN, YSNaN, // is signaling NaN
input logic [P.FLEN-1:0] X, Y, // original inputs (before unpacker)
output logic CmpNV, // invalid flag
output logic CmpNV, // invalid flag
output logic [P.FLEN-1:0] CmpFpRes, // compare floating-point result
output logic [P.XLEN-1:0] CmpIntRes // compare integer result
);
logic LTabs, LT, EQ; // is X < or > or = Y
logic LTabs, LT, EQ; // is X < or > or = Y
logic [P.FLEN-1:0] NaNRes; // NaN result
logic BothZero; // are both inputs zero
logic EitherNaN, EitherSNaN; // are either input a (signaling) NaN
logic BothZero; // are both inputs zero
logic EitherNaN, EitherSNaN; // are either input a (signaling) NaN
assign LTabs= {1'b0, Xe, Xm} < {1'b0, Ye, Ym}; // unsigned comparison, treating FP as integers
assign LT = (Xs & ~Ys) | (Xs & Ys & ~LTabs & ~EQ) | (~Xs & ~Ys & LTabs); // signed comparison
@ -62,7 +61,6 @@ module fcmp import cvw::*; #(parameter cvw_t P) (
assign EitherNaN = XNaN|YNaN;
assign EitherSNaN = XSNaN|YSNaN;
// flags
// Min/Max - if an input is a signaling NaN set invalid flag
// LT/LE - signaling - sets invalid if NaN input
@ -85,11 +83,11 @@ module fcmp import cvw::*; #(parameter cvw_t P) (
// select the NaN result
if (P.FPSIZES == 1)
if(P.IEEE754) assign NaNRes = {Xs, {P.NE{1'b1}}, 1'b1, Xm[P.NF-2:0]};
else assign NaNRes = {1'b0, {P.NE{1'b1}}, 1'b1, {P.NF-1{1'b0}}};
else assign NaNRes = {1'b0, {P.NE{1'b1}}, 1'b1, {P.NF-1{1'b0}}};
else if (P.FPSIZES == 2)
if(P.IEEE754) assign NaNRes = Fmt ? {Xs, {P.NE{1'b1}}, 1'b1, Xm[P.NF-2:0]} : {{P.FLEN-P.LEN1{1'b1}}, Xs, {P.NE1{1'b1}}, 1'b1, Xm[P.NF-2:P.NF-P.NF1]};
else assign NaNRes = Fmt ? {1'b0, {P.NE{1'b1}}, 1'b1, {P.NF-1{1'b0}}} : {{P.FLEN-P.LEN1{1'b1}}, 1'b0, {P.NE1{1'b1}}, 1'b1, (P.NF1-1)'(0)};
else assign NaNRes = Fmt ? {1'b0, {P.NE{1'b1}}, 1'b1, {P.NF-1{1'b0}}} : {{P.FLEN-P.LEN1{1'b1}}, 1'b0, {P.NE1{1'b1}}, 1'b1, (P.NF1-1)'(0)};
else if (P.FPSIZES == 3)
always_comb
@ -123,7 +121,6 @@ module fcmp import cvw::*; #(parameter cvw_t P) (
else NaNRes = {{P.FLEN-P.H_LEN{1'b1}}, 1'b0, {P.H_NE{1'b1}}, 1'b1, (P.H_NF-1)'(0)};
endcase
// Min/Max
// - outputs the min/max of X and Y
// - -0 < 0
@ -154,5 +151,4 @@ module fcmp import cvw::*; #(parameter cvw_t P) (
// - inf = inf and -inf = -inf
// - return 0 if comparison with NaN (unordered)
assign CmpIntRes = {(P.XLEN-1)'(0), (((EQ|BothZero)&OpCtrl[1])|(LT&OpCtrl[0]&~BothZero))&~EitherNaN};
endmodule

90
src/fpu/fctrl.sv
View File

@ -27,57 +27,57 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module fctrl import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic reset,
input logic clk,
input logic reset,
// input control signals
input logic StallE, StallM, StallW, // stall signals
input logic FlushE, FlushM, FlushW, // flush signals
input logic IntDivE, // is inteteger division
input logic [2:0] FRM_REGW, // rounding mode from CSR
input logic [1:0] STATUS_FS, // is FPU enabled?
input logic FDivBusyE, // is the divider busy
input logic StallE, StallM, StallW, // stall signals
input logic FlushE, FlushM, FlushW, // flush signals
input logic IntDivE, // is inteteger division
input logic [2:0] FRM_REGW, // rounding mode from CSR
input logic [1:0] STATUS_FS, // is FPU enabled?
input logic FDivBusyE, // is the divider busy
// intruction
input logic [31:0] InstrD, // the full instruction
input logic [6:0] Funct7D, // bits 31:25 of instruction - may contain percision
input logic [6:0] OpD, // bits 6:0 of instruction
input logic [4:0] Rs2D, // bits 24:20 of instruction
input logic [2:0] Funct3D, Funct3E, // bits 14:12 of instruction - may contain rounding mode
input logic [31:0] InstrD, // the full instruction
input logic [6:0] Funct7D, // bits 31:25 of instruction - may contain percision
input logic [6:0] OpD, // bits 6:0 of instruction
input logic [4:0] Rs2D, // bits 24:20 of instruction
input logic [2:0] Funct3D, Funct3E, // bits 14:12 of instruction - may contain rounding mode
// input mux selections
output logic XEnD, YEnD, ZEnD, // enable inputs
output logic XEnE, YEnE, ZEnE, // enable inputs
output logic XEnD, YEnD, ZEnD, // enable inputs
output logic XEnE, YEnE, ZEnE, // enable inputs
// opperation mux selections
output logic FCvtIntE, FCvtIntW, // convert to integer opperation
output logic [2:0] FrmM, // FP rounding mode
output logic FCvtIntE, FCvtIntW, // convert to integer opperation
output logic [2:0] FrmM, // FP rounding mode
output logic [P.FMTBITS-1:0] FmtE, FmtM, // FP format
output logic [2:0] OpCtrlE, OpCtrlM, // Select which opperation to do in each component
output logic FpLoadStoreM, // FP load or store instruction
output logic [1:0] PostProcSelE, PostProcSelM, // select result in the post processing unit
output logic [1:0] FResSelE, FResSelM, FResSelW, // Select one of the results that finish in the memory stage
output logic [2:0] OpCtrlE, OpCtrlM, // Select which opperation to do in each component
output logic FpLoadStoreM, // FP load or store instruction
output logic [1:0] PostProcSelE, PostProcSelM, // select result in the post processing unit
output logic [1:0] FResSelE, FResSelM, FResSelW, // Select one of the results that finish in the memory stage
// register control signals
output logic FRegWriteE, FRegWriteM, FRegWriteW, // FP register write enable
output logic FWriteIntE, FWriteIntM, // Write to integer register
output logic [4:0] Adr1D, Adr2D, Adr3D, // adresses of each input
output logic [4:0] Adr1E, Adr2E, Adr3E, // adresses of each input
output logic FRegWriteE, FRegWriteM, FRegWriteW, // FP register write enable
output logic FWriteIntE, FWriteIntM, // Write to integer register
output logic [4:0] Adr1D, Adr2D, Adr3D, // adresses of each input
output logic [4:0] Adr1E, Adr2E, Adr3E, // adresses of each input
// other control signals
output logic IllegalFPUInstrD, // Is the instruction an illegal fpu instruction
output logic FDivStartE, IDivStartE // Start division or squareroot
output logic IllegalFPUInstrD, // Is the instruction an illegal fpu instruction
output logic FDivStartE, IDivStartE // Start division or squareroot
);
`define FCTRLW 12
logic [`FCTRLW-1:0] ControlsD; // control signals
logic FRegWriteD; // FP register write enable
logic FDivStartD; // start division/sqrt
logic FWriteIntD; // integer register write enable
logic [2:0] OpCtrlD; // Select which opperation to do in each component
logic [1:0] PostProcSelD; // select result in the post processing unit
logic [1:0] FResSelD; // Select one of the results that finish in the memory stage
logic [2:0] FrmD, FrmE; // FP rounding mode
logic [P.FMTBITS-1:0] FmtD; // FP format
logic [1:0] Fmt, Fmt2; // format - before possible reduction
logic SupportedFmt; // is the format supported
logic SupportedFmt2; // is the source format supported for fp -> fp
logic FCvtIntD, FCvtIntM; // convert to integer opperation
logic [`FCTRLW-1:0] ControlsD; // control signals
logic FRegWriteD; // FP register write enable
logic FDivStartD; // start division/sqrt
logic FWriteIntD; // integer register write enable
logic [2:0] OpCtrlD; // Select which opperation to do in each component
logic [1:0] PostProcSelD; // select result in the post processing unit
logic [1:0] FResSelD; // Select one of the results that finish in the memory stage
logic [2:0] FrmD, FrmE; // FP rounding mode
logic [P.FMTBITS-1:0] FmtD; // FP format
logic [1:0] Fmt, Fmt2; // format - before possible reduction
logic SupportedFmt; // is the format supported
logic SupportedFmt2; // is the source format supported for fp -> fp
logic FCvtIntD, FCvtIntM; // convert to integer opperation
// FPU Instruction Decoder
assign Fmt = Funct7D[1:0];
@ -97,7 +97,7 @@ module fctrl import cvw::*; #(parameter cvw_t P) (
ControlsD = `FCTRLW'b0_0_00_00_000_0_1_0; // for anything other than loads and stores, check for supported format
else begin
ControlsD = `FCTRLW'b0_0_00_00_000_0_1_0; // default: non-implemented instruction
/* verilator lint_off CASEINCOMPLETE */ // default value above has priority so no other default needed
/* verilator lint_off CASEINCOMPLETE */ // default value above has priority so no other default needed
case(OpD)
7'b0000111: case(Funct3D)
3'b010: ControlsD = `FCTRLW'b1_0_10_00_0xx_0_0_0; // flw
@ -232,8 +232,7 @@ module fctrl import cvw::*; #(parameter cvw_t P) (
logic [1:0] FmtTmp;
assign FmtTmp = ((Funct7D[6:3] == 4'b0100)&OpD[4]) ? Rs2D[1:0] : (~OpD[6]&(&OpD[2:0])) ? {~Funct3D[1], ~(Funct3D[1]^Funct3D[0])} : Funct7D[1:0];
assign FmtD = (P.FMT == FmtTmp);
end
else if (P.FPSIZES == 3|P.FPSIZES == 4)
end else if (P.FPSIZES == 3|P.FPSIZES == 4)
assign FmtD = ((Funct7D[6:3] == 4'b0100)&OpD[4]) ? Rs2D[1:0] : Funct7D[1:0];
// Enables indicate that a source register is used and may need stalls. Also indicate special cases for infinity or NaN.
@ -250,12 +249,9 @@ module fctrl import cvw::*; #(parameter cvw_t P) (
((FResSelD==2'b11)&(PostProcSelD==2'b00))| // mv float to int
((FResSelD==2'b01)&((PostProcSelD==2'b00)|((PostProcSelD==2'b01)&OpCtrlD[0])))); // cvt both or sqrt
// Z - fma ops only
assign ZEnD = (PostProcSelD==2'b10)&(~OpCtrlD[2]|OpCtrlD[1]); // fma, add, sub
// Final Res Sel:
// fp int
// 00 other cmp
@ -321,7 +317,7 @@ module fctrl import cvw::*; #(parameter cvw_t P) (
// Integer division on FPU divider
if (P.M_SUPPORTED & P.IDIV_ON_FPU) assign IDivStartE = IntDivE;
else assign IDivStartE = 0;
else assign IDivStartE = 0;
// E/M pipleine register
flopenrc #(13+int'(P.FMTBITS)) EMCtrlReg (clk, reset, FlushM, ~StallM,

4
src/fpu/fcvt.sv
View File

@ -1,4 +1,3 @@
///////////////////////////////////////////
// fcvt.sv
//
@ -96,7 +95,7 @@ module fcvt import cvw::*; #(parameter cvw_t P) (
// int -> fp : | positive integer | 00000... (if needed) |
// fp -> fp : | fraction | 00000... (if needed) |
assign LzcInFull = IntToFp ? {TrimInt, {P.CVTLEN-P.XLEN+1{1'b0}}} :
{Xm, {P.CVTLEN-P.NF{1'b0}}};
{Xm, {P.CVTLEN-P.NF{1'b0}}};
// used as shifter input in postprocessor
assign LzcIn = LzcInFull[P.CVTLEN-1:0];
@ -222,5 +221,4 @@ module fcvt import cvw::*; #(parameter cvw_t P) (
if(Int64) Cs = Int[P.XLEN-1]&Signed;
else Cs = Int[31]&Signed;
else Cs = Xs;
endmodule

7
src/fpu/fdivsqrt/fdivsqrtcycles.sv
View File

@ -50,9 +50,10 @@ module fdivsqrtcycles import cvw::*; #(parameter cvw_t P) (
else if (P.FPSIZES == 3)
always_comb
case (FmtE)
P.FMT: Nf = P.NF;
P.FMT1: Nf = P.NF1;
P.FMT2: Nf = P.NF2;
P.FMT: Nf = P.NF;
P.FMT1: Nf = P.NF1;
P.FMT2: Nf = P.NF2;
default: Nf = 'x; // shouldn't happen
endcase
else if (P.FPSIZES == 4)
always_comb

3
src/fpu/fhazard.sv
View File

@ -56,7 +56,6 @@ module fhazard(
// if the needed value is in the writeback stage
end else if ((Adr1E == RdW) & FRegWriteW) ForwardXE = 2'b01; // choose FResult64W
// if the needed value is in the memory stage - input 2
if ((Adr2E == RdM) & FRegWriteM) begin
// if the result will be FResM (can be taken from the memory stage)
@ -64,7 +63,6 @@ module fhazard(
// if the needed value is in the writeback stage
end else if ((Adr2E == RdW) & FRegWriteW) ForwardYE = 2'b01; // choose FResult64W
// if the needed value is in the memory stage - input 3
if ((Adr3E == RdM) & FRegWriteM) begin
// if the result will be FResM (can be taken from the memory stage)
@ -72,5 +70,4 @@ module fhazard(
// if the needed value is in the writeback stage
end else if ((Adr3E == RdW) & FRegWriteW) ForwardZE = 2'b01; // choose FResult64W
end
endmodule

181
src/fpu/fpu.sv
View File

@ -27,40 +27,40 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module fpu import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic reset,
input logic clk,
input logic reset,
// Hazards
input logic StallE, StallM, StallW, // stall signals (from HZU)
input logic FlushE, FlushM, FlushW, // flush signals (from HZU)
output logic FPUStallD, // Stall the decode stage (To HZU)
output logic FDivBusyE, // Is the divide/sqrt unit busy (stall execute stage) (to HZU)
input logic StallE, StallM, StallW, // stall signals (from HZU)
input logic FlushE, FlushM, FlushW, // flush signals (from HZU)
output logic FPUStallD, // Stall the decode stage (To HZU)
output logic FDivBusyE, // Is the divide/sqrt unit busy (stall execute stage) (to HZU)
// CSRs
input logic [1:0] STATUS_FS, // Is floating-point enabled? (From privileged unit)
input logic [2:0] FRM_REGW, // Rounding mode (from CSR)
input logic [1:0] STATUS_FS, // Is floating-point enabled? (From privileged unit)
input logic [2:0] FRM_REGW, // Rounding mode (from CSR)
// Decode stage
input logic [31:0] InstrD, // instruction (from IFU)
input logic [31:0] InstrD, // instruction (from IFU)
// Execute stage
input logic [2:0] Funct3E, // Funct fields of instruction specify type of operations
input logic IntDivE, W64E, // Integer division on FPU
input logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // Integer input for convert, move, and int div (from IEU)
input logic [4:0] RdE, // which FP register to write to (from IEU)
output logic FWriteIntE, // integer register write enable (to IEU)
output logic FCvtIntE, // Convert to int (to IEU)
input logic [2:0] Funct3E, // Funct fields of instruction specify type of operations
input logic IntDivE, W64E, // Integer division on FPU
input logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // Integer input for convert, move, and int div (from IEU)
input logic [4:0] RdE, // which FP register to write to (from IEU)
output logic FWriteIntE, // integer register write enable (to IEU)
output logic FCvtIntE, // Convert to int (to IEU)
// Memory stage
input logic [2:0] Funct3M, // Funct fields of instruction specify type of operations
input logic [4:0] RdM, // which FP register to write to (from IEU)
output logic FRegWriteM, // FP register write enable (to privileged unit)
output logic FpLoadStoreM, // Fp load instruction? (to LSU)
output logic [P.FLEN-1:0] FWriteDataM, // Data to be written to memory (to LSU)
output logic [P.XLEN-1:0] FIntResM, // data to be written to integer register (to IEU)
output logic IllegalFPUInstrD, // Is the instruction an illegal fpu instruction (to IFU)
output logic [4:0] SetFflagsM, // FPU flags (to privileged unit)
input logic [2:0] Funct3M, // Funct fields of instruction specify type of operations
input logic [4:0] RdM, // which FP register to write to (from IEU)
output logic FRegWriteM, // FP register write enable (to privileged unit)
output logic FpLoadStoreM, // Fp load instruction? (to LSU)
output logic [P.FLEN-1:0] FWriteDataM, // Data to be written to memory (to LSU)
output logic [P.XLEN-1:0] FIntResM, // data to be written to integer register (to IEU)
output logic IllegalFPUInstrD, // Is the instruction an illegal fpu instruction (to IFU)
output logic [4:0] SetFflagsM, // FPU flags (to privileged unit)
// Writeback stage
input logic [4:0] RdW, // which FP register to write to (from IEU)
input logic [P.FLEN-1:0] ReadDataW, // Read data (from LSU)
output logic [P.XLEN-1:0] FCvtIntResW, // convert result to to be written to integer register (to IEU)
output logic FCvtIntW, // select FCvtIntRes (to IEU)
output logic [P.XLEN-1:0] FIntDivResultW // Result from integer division (to IEU)
input logic [4:0] RdW, // which FP register to write to (from IEU)
input logic [P.FLEN-1:0] ReadDataW, // Read data (from LSU)
output logic [P.XLEN-1:0] FCvtIntResW, // convert result to to be written to integer register (to IEU)
output logic FCvtIntW, // select FCvtIntRes (to IEU)
output logic [P.XLEN-1:0] FIntDivResultW // Result from integer division (to IEU)
);
// RISC-V FPU specifics:
@ -68,86 +68,86 @@ module fpu import cvw::*; #(parameter cvw_t P) (
// - RISC-V detects underflow after rounding
// control signals
logic FRegWriteW; // FP register write enable
logic [2:0] FrmM; // FP rounding mode
logic [P.FMTBITS-1:0] FmtE, FmtM; // FP precision 0-single 1-double
logic FDivStartE, IDivStartE; // Start division or squareroot
logic FWriteIntM; // Write to integer register
logic [1:0] ForwardXE, ForwardYE, ForwardZE; // forwarding mux control signals
logic [2:0] OpCtrlE, OpCtrlM; // Select which opperation to do in each component
logic [1:0] FResSelE, FResSelM, FResSelW; // Select one of the results that finish in the memory stage
logic [1:0] PostProcSelE, PostProcSelM; // select result in the post processing unit
logic [4:0] Adr1D, Adr2D, Adr3D; // register adresses of each input
logic [4:0] Adr1E, Adr2E, Adr3E; // register adresses of each input
logic XEnD, YEnD, ZEnD; // X, Y, Z inputs used for current operation
logic XEnE, YEnE, ZEnE; // X, Y, Z inputs used for current operation
logic FRegWriteE; // Write floating-point register
logic FRegWriteW; // FP register write enable
logic [2:0] FrmM; // FP rounding mode
logic [P.FMTBITS-1:0] FmtE, FmtM; // FP precision 0-single 1-double
logic FDivStartE, IDivStartE; // Start division or squareroot
logic FWriteIntM; // Write to integer register
logic [1:0] ForwardXE, ForwardYE, ForwardZE; // forwarding mux control signals
logic [2:0] OpCtrlE, OpCtrlM; // Select which opperation to do in each component
logic [1:0] FResSelE, FResSelM, FResSelW; // Select one of the results that finish in the memory stage
logic [1:0] PostProcSelE, PostProcSelM; // select result in the post processing unit
logic [4:0] Adr1D, Adr2D, Adr3D; // register adresses of each input
logic [4:0] Adr1E, Adr2E, Adr3E; // register adresses of each input
logic XEnD, YEnD, ZEnD; // X, Y, Z inputs used for current operation
logic XEnE, YEnE, ZEnE; // X, Y, Z inputs used for current operation
logic FRegWriteE; // Write floating-point register
// regfile signals
logic [P.FLEN-1:0] FRD1D, FRD2D, FRD3D; // Read Data from FP register - decode stage
logic [P.FLEN-1:0] FRD1E, FRD2E, FRD3E; // Read Data from FP register - execute stage
logic [P.FLEN-1:0] XE; // Input 1 to the various units (after forwarding)
logic [P.XLEN-1:0] IntSrcXE; // Input 1 to the various units (after forwarding)
logic [P.FLEN-1:0] PreYE, YE; // Input 2 to the various units (after forwarding)
logic [P.FLEN-1:0] PreZE, ZE; // Input 3 to the various units (after forwarding)
logic [P.FLEN-1:0] FRD1D, FRD2D, FRD3D; // Read Data from FP register - decode stage
logic [P.FLEN-1:0] FRD1E, FRD2E, FRD3E; // Read Data from FP register - execute stage
logic [P.FLEN-1:0] XE; // Input 1 to the various units (after forwarding)
logic [P.XLEN-1:0] IntSrcXE; // Input 1 to the various units (after forwarding)
logic [P.FLEN-1:0] PreYE, YE; // Input 2 to the various units (after forwarding)
logic [P.FLEN-1:0] PreZE, ZE; // Input 3 to the various units (after forwarding)
// unpacking signals
logic XsE, YsE, ZsE; // input's sign - execute stage
logic XsM, YsM; // input's sign - memory stage
logic [P.NE-1:0] XeE, YeE, ZeE; // input's exponent - execute stage
logic [P.NE-1:0] ZeM; // input's exponent - memory stage
logic [P.NF:0] XmE, YmE, ZmE; // input's significand - execute stage
logic [P.NF:0] XmM, YmM, ZmM; // input's significand - memory stage
logic XNaNE, YNaNE, ZNaNE; // is the input a NaN - execute stage
logic XNaNM, YNaNM, ZNaNM; // is the input a NaN - memory stage
logic XSNaNE, YSNaNE, ZSNaNE; // is the input a signaling NaN - execute stage
logic XSNaNM, YSNaNM, ZSNaNM; // is the input a signaling NaN - memory stage
logic XSubnormE; // is the input subnormal
logic XZeroE, YZeroE, ZZeroE; // is the input zero - execute stage
logic XZeroM, YZeroM; // is the input zero - memory stage
logic XInfE, YInfE, ZInfE; // is the input infinity - execute stage
logic XInfM, YInfM, ZInfM; // is the input infinity - memory stage
logic XExpMaxE; // is the exponent all ones (max value)
logic [P.FLEN-1:0] XPostBoxE; // X after fixing bad NaN box. Needed for 1-input operations
logic XsE, YsE, ZsE; // input's sign - execute stage
logic XsM, YsM; // input's sign - memory stage
logic [P.NE-1:0] XeE, YeE, ZeE; // input's exponent - execute stage
logic [P.NE-1:0] ZeM; // input's exponent - memory stage
logic [P.NF:0] XmE, YmE, ZmE; // input's significand - execute stage
logic [P.NF:0] XmM, YmM, ZmM; // input's significand - memory stage
logic XNaNE, YNaNE, ZNaNE; // is the input a NaN - execute stage
logic XNaNM, YNaNM, ZNaNM; // is the input a NaN - memory stage
logic XSNaNE, YSNaNE, ZSNaNE; // is the input a signaling NaN - execute stage
logic XSNaNM, YSNaNM, ZSNaNM; // is the input a signaling NaN - memory stage
logic XSubnormE; // is the input subnormal
logic XZeroE, YZeroE, ZZeroE; // is the input zero - execute stage
logic XZeroM, YZeroM; // is the input zero - memory stage
logic XInfE, YInfE, ZInfE; // is the input infinity - execute stage
logic XInfM, YInfM, ZInfM; // is the input infinity - memory stage
logic XExpMaxE; // is the exponent all ones (max value)
logic [P.FLEN-1:0] XPostBoxE; // X after fixing bad NaN box. Needed for 1-input operations
// Fma Signals
logic FmaAddSubE; // Multiply by 1.0 when adding or subtracting
logic [1:0] FmaZSelE; // Select Z = Y when adding or subtracting, 0 when multiplying
logic [3*P.NF+3:0] SmE, SmM; // Sum significand
logic FmaAStickyE, FmaAStickyM; // FMA addend sticky bit output
logic [P.NE+1:0] SeE,SeM; // Sum exponent
logic InvAE, InvAM; // Invert addend
logic AsE, AsM; // Addend sign
logic PsE, PsM; // Product sign
logic SsE, SsM; // Sum sign
logic [$clog2(3*P.NF+5)-1:0] SCntE, SCntM; // LZA sum leading zero count
logic FmaAddSubE; // Multiply by 1.0 when adding or subtracting
logic [1:0] FmaZSelE; // Select Z = Y when adding or subtracting, 0 when multiplying
logic [3*P.NF+3:0] SmE, SmM; // Sum significand
logic FmaAStickyE, FmaAStickyM; // FMA addend sticky bit output
logic [P.NE+1:0] SeE,SeM; // Sum exponent
logic InvAE, InvAM; // Invert addend
logic AsE, AsM; // Addend sign
logic PsE, PsM; // Product sign
logic SsE, SsM; // Sum sign
logic [$clog2(3*P.NF+5)-1:0] SCntE, SCntM; // LZA sum leading zero count
// Cvt Signals
logic [P.NE:0] CeE, CeM; // convert intermediate expoent
logic [P.LOGCVTLEN-1:0] CvtShiftAmtE, CvtShiftAmtM; // how much to shift by
logic CvtResSubnormUfE, CvtResSubnormUfM; // does the result underflow or is subnormal
logic CsE, CsM; // convert result sign
logic IntZeroE, IntZeroM; // is the integer zero?
logic CvtResSubnormUfE, CvtResSubnormUfM; // does the result underflow or is subnormal
logic CsE, CsM; // convert result sign
logic IntZeroE, IntZeroM; // is the integer zero?
logic [P.CVTLEN-1:0] CvtLzcInE, CvtLzcInM; // input to the Leading Zero Counter (priority encoder)
logic [P.XLEN-1:0] FCvtIntResM; // fcvt integer result (for IEU)
// divide signals
logic [P.DIVb:0] QmM; // fdivsqrt signifcand
logic [P.NE+1:0] QeM; // fdivsqrt exponent
logic DivStickyM; // fdivsqrt sticky bit
logic FDivDoneE, IFDivStartE; // fdivsqrt control signals
logic DivStickyM; // fdivsqrt sticky bit
logic FDivDoneE, IFDivStartE; // fdivsqrt control signals
logic [P.XLEN-1:0] FIntDivResultM; // fdivsqrt integer division result (for IEU)
// result and flag signals
logic [P.XLEN-1:0] ClassResE; // classify result
logic [P.FLEN-1:0] CmpFpResE; // compare result to FPU (min/max)
logic [P.XLEN-1:0] CmpIntResE; // compare result to IEU (eq/lt/le)
logic CmpNVE; // compare invalid flag (Not Valid)
logic CmpNVE; // compare invalid flag (Not Valid)
logic [P.FLEN-1:0] SgnResE; // sign injection result
logic [P.XLEN-1:0] FIntResE; // FPU to IEU E-stage result (classify, compare, move)
logic [P.FLEN-1:0] PostProcResM; // Postprocessor output
logic [4:0] PostProcFlgM; // Postprocessor flags
logic PreNVE, PreNVM; // selected flag that is ready in the memory stage
logic [4:0] PostProcFlgM; // Postprocessor flags
logic PreNVE, PreNVM; // selected flag that is ready in the memory stage
logic [P.FLEN-1:0] FpResM, FpResW; // FPU preliminary result
logic [P.FLEN-1:0] PreFpResE, PreFpResM; // selected result that is ready in the memory stage
logic [P.FLEN-1:0] FResultW; // final FP result being written to the FP register
@ -156,9 +156,9 @@ module fpu import cvw::*; #(parameter cvw_t P) (
logic [P.FLEN-1:0] AlignedSrcAE; // align SrcA from IEU to the floating point format for fmv
logic [P.FLEN-1:0] BoxedZeroE; // Zero value for Z for multiplication, with NaN boxing if needed
logic [P.FLEN-1:0] BoxedOneE; // One value for Z for multiplication, with NaN boxing if needed
logic StallUnpackedM; // Stall unpacker outputs during multicycle fdivsqrt
logic StallUnpackedM; // Stall unpacker outputs during multicycle fdivsqrt
logic [P.FLEN-1:0] SgnExtXE; // Sign-extended X input for move to integer
logic mvsgn; // sign bit for extending move
logic mvsgn; // sign bit for extending move
//////////////////////////////////////////////////////////////////////////////////////////
// Decode Stage: fctrl decoder, read register file
@ -205,15 +205,15 @@ module fpu import cvw::*; #(parameter cvw_t P) (
mux2 #(P.FLEN) fonemux ({{P.FLEN-P.LEN1{1'b1}}, 2'b0, {P.NE1-1{1'b1}}, (P.NF1)'(0)}, {2'b0, {P.NE-1{1'b1}}, (P.NF)'(0)}, FmtE, BoxedOneE); // NaN boxing zeroes
else if(P.FPSIZES == 3 | P.FPSIZES == 4)
mux4 #(P.FLEN) fonemux ({{P.FLEN-P.S_LEN{1'b1}}, 2'b0, {P.S_NE-1{1'b1}}, (P.S_NF)'(0)},
{{P.FLEN-P.D_LEN{1'b1}}, 2'b0, {P.D_NE-1{1'b1}}, (P.D_NF)'(0)},
{{P.FLEN-P.H_LEN{1'b1}}, 2'b0, {P.H_NE-1{1'b1}}, (P.H_NF)'(0)},
{2'b0, {P.NE-1{1'b1}}, (P.NF)'(0)}, FmtE, BoxedOneE); // NaN boxing zeroes
{{P.FLEN-P.D_LEN{1'b1}}, 2'b0, {P.D_NE-1{1'b1}}, (P.D_NF)'(0)},
{{P.FLEN-P.H_LEN{1'b1}}, 2'b0, {P.H_NE-1{1'b1}}, (P.H_NF)'(0)},
{2'b0, {P.NE-1{1'b1}}, (P.NF)'(0)}, FmtE, BoxedOneE); // NaN boxing zeroes
assign FmaAddSubE = OpCtrlE[2]&OpCtrlE[1]&(FResSelE==2'b01)&(PostProcSelE==2'b10);
mux2 #(P.FLEN) fyaddmux (PreYE, BoxedOneE, FmaAddSubE, YE); // Force Y to be 1 for add/subtract
// Select NAN-boxed value of Z = 0.0 in proper format for FMA for multiply X*Y+Z
// For add and subtract, Z comes from second source operand
if(P.FPSIZES == 1) assign BoxedZeroE = 0;
if(P.FPSIZES == 1) assign BoxedZeroE = 0;
else if(P.FPSIZES == 2)
mux2 #(P.FLEN) fmulzeromux ({{P.FLEN-P.LEN1{1'b1}}, {P.LEN1{1'b0}}}, (P.FLEN)'(0), FmtE, BoxedZeroE); // NaN boxing zeroes
else if(P.FPSIZES == 3 | P.FPSIZES == 4)
@ -262,7 +262,6 @@ module fpu import cvw::*; #(parameter cvw_t P) (
.ToInt(FWriteIntE), .XZero(XZeroE), .Fmt(FmtE), .Ce(CeE), .ShiftAmt(CvtShiftAmtE),
.ResSubnormUf(CvtResSubnormUfE), .Cs(CsE), .IntZero(IntZeroE), .LzcIn(CvtLzcInE));
// NaN Box SrcA to convert integer to requested FP size
if(P.FPSIZES == 1) assign AlignedSrcAE = {{P.FLEN-P.XLEN{1'b1}}, ForwardedSrcAE};
else if(P.FPSIZES == 2)
@ -334,7 +333,7 @@ module fpu import cvw::*; #(parameter cvw_t P) (
.PostProcSel(PostProcSelM), .PostProcRes(PostProcResM), .PostProcFlg(PostProcFlgM), .FCvtIntRes(FCvtIntResM));
// FPU flag selection - to privileged
mux2 #(5) FPUFlgMux({PreNVM&~FResSelM[1], 4'b0}, PostProcFlgM, ~FResSelM[1]&FResSelM[0], SetFflagsM);
mux2 #(5) FPUFlgMux({PreNVM&~FResSelM[1], 4'b0}, PostProcFlgM, ~FResSelM[1]&FResSelM[0], SetFflagsM);
mux2 #(P.FLEN) FPUResMux(PreFpResM, PostProcResM, FResSelM[0], FpResM);
// M/W pipe registers

5
src/fpu/fregfile.sv
View File

@ -30,8 +30,8 @@ module fregfile #(parameter FLEN) (
input logic clk, reset,
input logic we4, // write enable
input logic [4:0] a1, a2, a3, a4, // adresses
input logic [FLEN-1:0] wd4, // write data
output logic [FLEN-1:0] rd1, rd2, rd3 // read data
input logic [FLEN-1:0] wd4, // write data
output logic [FLEN-1:0] rd1, rd2, rd3 // read data
);
logic [FLEN-1:0] rf[31:0];
@ -51,4 +51,3 @@ module fregfile #(parameter FLEN) (
assign #2 rd3 = rf[a3];
endmodule // regfile

5
src/fpu/fsgninj.sv
View File

@ -27,10 +27,10 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module fsgninj import cvw::*; #(parameter cvw_t P) (
input logic Xs, Ys, // X and Y sign bits
input logic Xs, Ys, // X and Y sign bits
input logic [P.FLEN-1:0] X, // X
input logic [P.FMTBITS-1:0] Fmt, // format
input logic [1:0] OpCtrl, // operation control
input logic [1:0] OpCtrl, // operation control
output logic [P.FLEN-1:0] SgnRes // result
);
@ -73,5 +73,4 @@ module fsgninj import cvw::*; #(parameter cvw_t P) (
endcase
assign SgnRes = {SgnBits[3], X[P.Q_LEN-2:P.D_LEN], SgnBits[2], X[P.D_LEN-2:P.S_LEN], SgnBits[1], X[P.S_LEN-2:P.H_LEN], SgnBits[0], X[P.H_LEN-2:0]};
end
endmodule

2
src/fpu/postproc/cvtshiftcalc.sv
View File

@ -82,7 +82,7 @@ module cvtshiftcalc import cvw::*; #(parameter cvw_t P) (
P.FMT: ResNegNF = -($clog2(P.NF)+1)'(P.NF);
P.FMT1: ResNegNF = -($clog2(P.NF)+1)'(P.NF1);
P.FMT2: ResNegNF = -($clog2(P.NF)+1)'(P.NF2);
default: ResNegNF = 1'bx;
default: ResNegNF = 'x;
endcase
end else if (P.FPSIZES == 4) begin

30
src/fpu/unpackinput.sv
View File

@ -28,24 +28,24 @@
module unpackinput import cvw::*; #(parameter cvw_t P) (
input logic [P.FLEN-1:0] In, // inputs from register file
input logic En, // enable the input
input logic En, // enable the input
input logic [P.FMTBITS-1:0] Fmt, // format signal 00 - single 01 - double 11 - quad 10 - half
output logic Sgn, // sign bits of the number
output logic Sgn, // sign bits of the number
output logic [P.NE-1:0] Exp, // exponent of the number (converted to largest supported precision)
output logic [P.NF:0] Man, // mantissa of the number (converted to largest supported precision)
output logic NaN, // is the number a NaN
output logic SNaN, // is the number a signaling NaN
output logic Zero, // is the number zero
output logic Inf, // is the number infinity
output logic ExpNonZero, // is the exponent not zero
output logic FracZero, // is the fraction zero
output logic ExpMax, // does In have the maximum exponent (NaN or Inf)
output logic Subnorm, // is the number subnormal
output logic NaN, // is the number a NaN
output logic SNaN, // is the number a signaling NaN
output logic Zero, // is the number zero
output logic Inf, // is the number infinity
output logic ExpNonZero, // is the exponent not zero
output logic FracZero, // is the fraction zero
output logic ExpMax, // does In have the maximum exponent (NaN or Inf)
output logic Subnorm, // is the number subnormal
output logic [P.FLEN-1:0] PostBox // Number reboxed correctly as a NaN
);
logic [P.NF-1:0] Frac; // Fraction of XYZ
logic BadNaNBox; // incorrectly NaN Boxed
logic [P.NF-1:0] Frac; // Fraction of XYZ
logic BadNaNBox; // incorrectly NaN Boxed
if (P.FPSIZES == 1) begin // if there is only one floating point format supported
assign BadNaNBox = 0;
@ -107,7 +107,6 @@ module unpackinput import cvw::*; #(parameter cvw_t P) (
// is the exponent all 1's
assign ExpMax = Fmt ? &In[P.FLEN-2:P.NF] : &In[P.LEN1-2:P.NF1];
end else if (P.FPSIZES == 3) begin // three floating point precsions supported
// largest format | larger format | smallest format
@ -246,7 +245,6 @@ module unpackinput import cvw::*; #(parameter cvw_t P) (
2'b00: Sgn = In[P.S_LEN-1];
2'b10: Sgn = In[P.H_LEN-1];
endcase
// extract the fraction
always_comb
@ -266,7 +264,6 @@ module unpackinput import cvw::*; #(parameter cvw_t P) (
2'b10: ExpNonZero = |In[P.H_LEN-2:P.H_NF];
endcase
// example double to single conversion:
// 1023 = 0011 1111 1111
// 127 = 0000 0111 1111 (subtract this)
@ -284,8 +281,7 @@ module unpackinput import cvw::*; #(parameter cvw_t P) (
2'b00: Exp = {In[P.S_LEN-2], {P.Q_NE-P.S_NE{~In[P.S_LEN-2]}}, In[P.S_LEN-3:P.S_NF+1], In[P.S_NF]|~ExpNonZero};
2'b10: Exp = {In[P.H_LEN-2], {P.Q_NE-P.H_NE{~In[P.H_LEN-2]}}, In[P.H_LEN-3:P.H_NF+1], In[P.H_NF]|~ExpNonZero};
endcase
// is the exponent all 1's
always_comb
case (Fmt)

2
src/generic/adder.sv
View File

@ -33,5 +33,3 @@ module adder #(parameter WIDTH=8) (
assign y = a + b;
endmodule

2
src/generic/counter.sv
View File

@ -36,5 +36,3 @@ module counter #(parameter WIDTH=8) (
assign qnext = q + 1;
flopenr #(WIDTH) cntrflop(clk, reset, en, qnext, q);
endmodule

1
src/generic/neg.sv
View File

@ -32,4 +32,3 @@ module neg #(parameter WIDTH = 8) (
assign y = ~a + 1;
endmodule

2
src/generic/or_rows.sv
View File

@ -48,5 +48,3 @@ module or_rows #(parameter ROWS = 8, COLS=2) (
/* verilator lint_on UNOPTFLAT */
end
endmodule

3
src/generic/prioritythermometer.sv
View File

@ -48,6 +48,3 @@ module prioritythermometer #(parameter N = 8) (
end
/* verilator lint_on UNOPTFLAT */
endmodule

60
src/ieu/datapath.sv
View File

@ -28,55 +28,55 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module datapath import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic clk, reset,
// Decode stage signals
input logic [2:0] ImmSrcD, // Selects type of immediate extension
input logic [31:0] InstrD, // Instruction in Decode stage
input logic [2:0] ImmSrcD, // Selects type of immediate extension
input logic [31:0] InstrD, // Instruction in Decode stage
// Execute stage signals
input logic [P.XLEN-1:0] PCE, // PC in Execute stage
input logic [P.XLEN-1:0] PCLinkE, // PC + 4 (of instruction in Execute stage)
input logic [2:0] Funct3E, // Funct3 field of instruction in Execute stage
input logic StallE, FlushE, // Stall, flush Execute stage
input logic [1:0] ForwardAE, ForwardBE, // Forward ALU operands from later stages
input logic W64E, // W64-type instruction
input logic SubArithE, // Subtraction or arithmetic shift
input logic ALUSrcAE, ALUSrcBE, // ALU operands
input logic ALUResultSrcE, // Selects result to pass on to Memory stage
input logic [2:0] ALUSelectE, // ALU mux select signal
input logic JumpE, // Is a jump (j) instruction
input logic BranchSignedE, // Branch comparison operands are signed (if it's a branch)
input logic [1:0] BSelectE, // One hot encoding of ZBA_ZBB_ZBC_ZBS instruction
input logic [2:0] ZBBSelectE, // ZBB mux select signal
input logic [2:0] BALUControlE, // ALU Control signals for B instructions in Execute Stage
output logic [1:0] FlagsE, // Comparison flags ({eq, lt})
input logic [2:0] Funct3E, // Funct3 field of instruction in Execute stage
input logic StallE, FlushE, // Stall, flush Execute stage
input logic [1:0] ForwardAE, ForwardBE, // Forward ALU operands from later stages
input logic W64E, // W64-type instruction
input logic SubArithE, // Subtraction or arithmetic shift
input logic ALUSrcAE, ALUSrcBE, // ALU operands
input logic ALUResultSrcE, // Selects result to pass on to Memory stage
input logic [2:0] ALUSelectE, // ALU mux select signal
input logic JumpE, // Is a jump (j) instruction
input logic BranchSignedE, // Branch comparison operands are signed (if it's a branch)
input logic [1:0] BSelectE, // One hot encoding of ZBA_ZBB_ZBC_ZBS instruction
input logic [2:0] ZBBSelectE, // ZBB mux select signal
input logic [2:0] BALUControlE, // ALU Control signals for B instructions in Execute Stage
output logic [1:0] FlagsE, // Comparison flags ({eq, lt})
output logic [P.XLEN-1:0] IEUAdrE, // Address computed by ALU
output logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // ALU sources before the mux chooses between them and PCE to put in srcA/B
// Memory stage signals
input logic StallM, FlushM, // Stall, flush Memory stage
input logic FWriteIntM, FCvtIntW, // FPU writes integer register file, FPU converts float to int
input logic StallM, FlushM, // Stall, flush Memory stage
input logic FWriteIntM, FCvtIntW, // FPU writes integer register file, FPU converts float to int
input logic [P.XLEN-1:0] FIntResM, // FPU integer result
output logic [P.XLEN-1:0] SrcAM, // ALU's Source A in Memory stage to privilege unit for CSR writes
output logic [P.XLEN-1:0] WriteDataM, // Write data in Memory stage
// Writeback stage signals
input logic StallW, FlushW, // Stall, flush Writeback stage
input logic RegWriteW, IntDivW, // Write register file, integer divide instruction
input logic SquashSCW, // Squash a store conditional when a conflict arose
input logic [2:0] ResultSrcW, // Select source of result to write back to register file
input logic StallW, FlushW, // Stall, flush Writeback stage
input logic RegWriteW, IntDivW, // Write register file, integer divide instruction
input logic SquashSCW, // Squash a store conditional when a conflict arose
input logic [2:0] ResultSrcW, // Select source of result to write back to register file
input logic [P.XLEN-1:0] FCvtIntResW, // FPU convert fp to integer result
input logic [P.XLEN-1:0] ReadDataW, // Read data from LSU
input logic [P.XLEN-1:0] CSRReadValW, // CSR read result
input logic [P.XLEN-1:0] MDUResultW, // MDU (Multiply/divide unit) result
input logic [P.XLEN-1:0] FIntDivResultW, // FPU's integer divide result
// Hazard Unit signals
output logic [4:0] Rs1D, Rs2D, Rs1E, Rs2E, // Register sources to read in Decode or Execute stage
output logic [4:0] RdE, RdM, RdW // Register destinations in Execute, Memory, or Writeback stage
output logic [4:0] Rs1D, Rs2D, Rs1E, Rs2E, // Register sources to read in Decode or Execute stage
output logic [4:0] RdE, RdM, RdW // Register destinations in Execute, Memory, or Writeback stage
);
// Fetch stage signals
// Decode stage signals
logic [P.XLEN-1:0] R1D, R2D; // Read data from Rs1 (RD1), Rs2 (RD2)
logic [P.XLEN-1:0] ImmExtD; // Extended immediate in Decode stage
logic [4:0] RdD; // Destination register in Decode stage
logic [4:0] RdD; // Destination register in Decode stage
// Execute stage signals
logic [P.XLEN-1:0] R1E, R2E; // Source operands read from register file
logic [P.XLEN-1:0] ImmExtE; // Extended immediate in Execute stage
@ -103,9 +103,9 @@ module datapath import cvw::*; #(parameter cvw_t P) (
flopenrc #(P.XLEN) RD1EReg(clk, reset, FlushE, ~StallE, R1D, R1E);
flopenrc #(P.XLEN) RD2EReg(clk, reset, FlushE, ~StallE, R2D, R2E);
flopenrc #(P.XLEN) ImmExtEReg(clk, reset, FlushE, ~StallE, ImmExtD, ImmExtE);
flopenrc #(5) Rs1EReg(clk, reset, FlushE, ~StallE, Rs1D, Rs1E);
flopenrc #(5) Rs2EReg(clk, reset, FlushE, ~StallE, Rs2D, Rs2E);
flopenrc #(5) RdEReg(clk, reset, FlushE, ~StallE, RdD, RdE);
flopenrc #(5) Rs1EReg(clk, reset, FlushE, ~StallE, Rs1D, Rs1E);
flopenrc #(5) Rs2EReg(clk, reset, FlushE, ~StallE, Rs2D, Rs2E);
flopenrc #(5) RdEReg(clk, reset, FlushE, ~StallE, RdD, RdE);
mux3 #(P.XLEN) faemux(R1E, ResultW, IFResultM, ForwardAE, ForwardedSrcAE);
mux3 #(P.XLEN) fbemux(R2E, ResultW, IFResultM, ForwardBE, ForwardedSrcBE);
@ -144,5 +144,5 @@ module datapath import cvw::*; #(parameter cvw_t P) (
// handle Store Conditional result if atomic extension supported
if (P.A_SUPPORTED) assign SCResultW = {{(P.XLEN-1){1'b0}}, SquashSCW};
else assign SCResultW = 0;
else assign SCResultW = 0;
endmodule

6
src/ieu/extend.sv
View File

@ -30,7 +30,7 @@
module extend #(parameter XLEN, A_SUPPORTED) (
input logic [31:7] InstrD, // All instruction bits except opcode (lower 7 bits)
input logic [2:0] ImmSrcD, // Select what kind of extension to perform
output logic [XLEN-1:0 ] ImmExtD); // Extended immediate
output logic [XLEN-1:0] ImmExtD); // Extended immediate
localparam [XLEN-1:0] undefined = {(XLEN){1'bx}}; // could change to 0 after debug
@ -45,10 +45,10 @@ module extend #(parameter XLEN, A_SUPPORTED) (
// J-type (jal)
3'b011: ImmExtD = {{(XLEN-20){InstrD[31]}}, InstrD[19:12], InstrD[20], InstrD[30:21], 1'b0};
// U-type (lui, auipc)
3'b100: ImmExtD = {{(XLEN-31){InstrD[31]}}, InstrD[30:12], 12'b0};
3'b100: ImmExtD = {{(XLEN-31){InstrD[31]}}, InstrD[30:12], 12'b0};
// Store Conditional: zero offset
3'b101: if (A_SUPPORTED) ImmExtD = 0;
else ImmExtD = undefined;
else ImmExtD = undefined;
default: ImmExtD = undefined; // undefined
endcase
endmodule

28
src/ieu/ieu.sv
View File

@ -26,7 +26,6 @@
// and limitations under the License.
////////////////////////////////////////////////////////////////////////////////////////////////
module ieu import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
// Decode stage signals
@ -34,36 +33,36 @@ module ieu import cvw::*; #(parameter cvw_t P) (
input logic IllegalIEUFPUInstrD, // Illegal instruction
output logic IllegalBaseInstrD, // Illegal I-type instruction, or illegal RV32 access to upper 16 registers
// Execute stage signals
input logic [P.XLEN-1:0] PCE, // PC
input logic [P.XLEN-1:0] PCLinkE, // PC + 4
input logic [P.XLEN-1:0] PCE, // PC
input logic [P.XLEN-1:0] PCLinkE, // PC + 4
output logic PCSrcE, // Select next PC (between PC+4 and IEUAdrE)
input logic FWriteIntE, FCvtIntE, // FPU writes to integer register file, FPU converts float to int
output logic [P.XLEN-1:0] IEUAdrE, // Memory address
output logic [P.XLEN-1:0] IEUAdrE, // Memory address
output logic IntDivE, W64E, // Integer divide, RV64 W-type instruction
output logic [2:0] Funct3E, // Funct3 instruction field
output logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // ALU src inputs before the mux choosing between them and PCE to put in srcA/B
output logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // ALU src inputs before the mux choosing between them and PCE to put in srcA/B
output logic [4:0] RdE, // Destination register
// Memory stage signals
input logic SquashSCW, // Squash store conditional, from LSU
output logic [1:0] MemRWM, // Read/write control goes to LSU
output logic [1:0] AtomicM, // Atomic control goes to LSU
output logic [P.XLEN-1:0] WriteDataM, // Write data to LSU
output logic [P.XLEN-1:0] WriteDataM, // Write data to LSU
output logic [2:0] Funct3M, // Funct3 (size and signedness) to LSU
output logic [P.XLEN-1:0] SrcAM, // ALU SrcA to Privileged unit and FPU
output logic [P.XLEN-1:0] SrcAM, // ALU SrcA to Privileged unit and FPU
output logic [4:0] RdM, // Destination register
input logic [P.XLEN-1:0] FIntResM, // Integer result from FPU (fmv, fclass, fcmp)
input logic [P.XLEN-1:0] FIntResM, // Integer result from FPU (fmv, fclass, fcmp)
output logic InvalidateICacheM, FlushDCacheM, // Invalidate I$, flush D$
output logic InstrValidD, InstrValidE, InstrValidM,// Instruction is valid
output logic InstrValidD, InstrValidE, InstrValidM, // Instruction is valid
output logic BranchD, BranchE,
output logic JumpD, JumpE,
// Writeback stage signals
input logic [P.XLEN-1:0] FIntDivResultW, // Integer divide result from FPU fdivsqrt)
input logic [P.XLEN-1:0] CSRReadValW, // CSR read value,
input logic [P.XLEN-1:0] MDUResultW, // multiply/divide unit result
input logic [P.XLEN-1:0] FCvtIntResW, // FPU's float to int conversion result
input logic [P.XLEN-1:0] FIntDivResultW, // Integer divide result from FPU fdivsqrt)
input logic [P.XLEN-1:0] CSRReadValW, // CSR read value,
input logic [P.XLEN-1:0] MDUResultW, // multiply/divide unit result
input logic [P.XLEN-1:0] FCvtIntResW, // FPU's float to int conversion result
input logic FCvtIntW, // FPU converts float to int
output logic [4:0] RdW, // Destination register
input logic [P.XLEN-1:0] ReadDataW, // LSU's read data
input logic [P.XLEN-1:0] ReadDataW, // LSU's read data
// Hazard unit signals
input logic StallD, StallE, StallM, StallW, // Stall signals from hazard unit
input logic FlushD, FlushE, FlushM, FlushW, // Flush signals
@ -118,4 +117,3 @@ module ieu import cvw::*; #(parameter cvw_t P) (
.FCvtIntE, .SCE, .ForwardAE, .ForwardBE,
.FCvtIntStallD, .LoadStallD, .MDUStallD, .CSRRdStallD);
endmodule

8
src/ieu/regfile.sv
View File

@ -31,10 +31,10 @@ module regfile #(parameter XLEN, E_SUPPORTED) (
input logic clk, reset,
input logic we3, // Write enable
input logic [4:0] a1, a2, a3, // Source registers to read (a1, a2), destination register to write (a3)
input logic [XLEN-1:0] wd3, // Write data for port 3
output logic [XLEN-1:0] rd1, rd2); // Read data for ports 1, 2
input logic [XLEN-1:0] wd3, // Write data for port 3
output logic [XLEN-1:0] rd1, rd2); // Read data for ports 1, 2
localparam NUMREGS = E_SUPPORTED ? 16 : 32; // only 16 registers in E mode
localparam NUMREGS = E_SUPPORTED ? 16 : 32; // only 16 registers in E mode
logic [XLEN-1:0] rf[NUMREGS-1:1];
integer i;
@ -50,7 +50,7 @@ module regfile #(parameter XLEN, E_SUPPORTED) (
always_ff @(negedge clk)
if (reset) for(i=1; i<NUMREGS; i++) rf[i] <= 0;
else if (we3) rf[a3] <= wd3;
else if (we3) rf[a3] <= wd3;
assign #2 rd1 = (a1 != 0) ? rf[a1] : 0;
assign #2 rd2 = (a2 != 0) ? rf[a2] : 0;

4
src/ieu/shifter.sv
View File

@ -41,7 +41,7 @@ module shifter (
assign Sign = A[`XLEN-1] & SubArith; // sign bit for sign extension
if (`XLEN==32) begin // rv32
if (`ZBB_SUPPORTED) begin: rotfunnel32 //rv32 shifter with rotates
if (`ZBB_SUPPORTED) begin: rotfunnel32 // rv32 shifter with rotates
always_comb // funnel mux
case({Right, Rotate})
2'b00: Z = {A[31:0], 31'b0};
@ -84,5 +84,3 @@ module shifter (
assign ZShift = Z >> Offset;
assign Y = ZShift[`XLEN-1:0];
endmodule

112
src/ifu/CodeAligner.py
View File

@ -1,112 +0,0 @@
import os
# Kevin Wan kewan@hmc.edu 10/27/2021
def read_input(filename): #1
"""Takes in a string filename and outputs the parsed verilog code by line into a list
such that each element of the list is one line of verilog code as a string."""
lineOfCode = []
input_file = open(filename, 'r')
for line in input_file:
lineOfCode.append(line)
return lineOfCode
###################################################################################
def ID_start(GiantString):#2
"""takes in the list of sv file lines, outputs the location that variable names should start"""
VarLoc = 0
VarLineNum = None
for lines in GiantString:
if ' logic ' in lines and (lines.find("//") == -1 or lines.find("//") > lines.find(' logic ')): # // logic does not proceed. logic proceeds. logic // proceeds.
if "[" in lines and "]" in lines:# need to account for these space
NowLoc = lines.find(']') + 3# column number in sv code when 1st char of the var name should appear.
if NowLoc>VarLoc:
VarLoc = NowLoc
VarLineNum = GiantString.index(lines) # Update this number if new record is made.
else:
NowLoc = lines.find('logic') + 7 # same as before.
if NowLoc>VarLoc:
VarLoc = NowLoc
VarLineNum = GiantString.index(lines)
#print("Furthest variable appears on line", VarLineNum + 1,VarLoc) # Disable this line after debugging.
return VarLoc
##################################################################################
def modified_logNew(GS,SOV): #3
Ind = SOV - 1 # SOV is for human readability, Ind is the character's index in computer, since computers count from 0's we need to correct it.
Out = []
for l in GS:
lines = l.replace('\t',' ')
if ' logic ' in lines and (lines.find("//") == -1 or lines.find("//") > lines.find(' logic ')): # // logic does not proceed. logic proceeds. logic // proceeds.
if "[" in lines and "]" in lines: # the line is an extended declaration.
EditLoc = lines.find("]") # Re-finds the string index number of ].
VarLoc = FindCharRel(lines[EditLoc+1::]) + EditLoc + 1 # Checks where variable declaration currently is at.
#print(VarLoc,lines[VarLoc])# VERIFIED
NewLine = Mod_Space_at(lines,VarLoc,VarLoc-Ind)
Out.append(NewLine)# Verified0957 10272021
else:
EditLoc1 = lines.find('c') # Hopefully sees the c in 'logic'
VarLoc1 = FindCharRel(lines[EditLoc1+1::]) + EditLoc1 + 1
NewLine1 = Mod_Space_at(lines,VarLoc1,VarLoc1-Ind)
Out.append(NewLine1)# Verified 1005 10272021
else:
Out.append(lines)
return Out
################################################################################
def write_to_output(filename,GiantString,OW=True,Lines_editted=None): #4
"""Filename is preferrably passed from the early function calls"""
"""GiantString has all the corrected features in the code, each line is a good verilog code line"""
newname = filename
if not OW or OW =='f': #which means no overwrite (create a new file)
Decomposed=filename.split('.')
newname = Decomposed[0] + "_AL." + Decomposed[1] # AL for aligned.
OutFile = open(newname,'w') # This step should create a new file.
OutFile.writelines(GiantString)
OutFile.close()
print("Success! " + newname + " Now contains an aligned file!")
return newname
#################################################################################
def FindCharRel(Ln):
#returns the computer location of a character's first occurence
for num in range(len(Ln)):
if Ln[num] != " ":
return num
def Mod_Space_at(Ln,loc,diff):
#loc is the varLoc from mln, diff is varLoc - Ind
if diff > 0: # to delete
NewString = Ln[:(loc-diff)] + Ln[loc:]
if diff < 0: # to add
NewString = Ln[:loc] + (-diff)*" " + Ln[loc:]
if diff == 0:
NewString = Ln
return NewString
'''def main_filehandler(overwrite=False):
for filename in os.listdir():
if ".sv" in filename:
GiantString = read_input(filename)
SOV = ID_start(GiantString)
ModifiedGS = modified_logNew(GiantString,SOV)
Newname = write_to_output(filename,ModifiedGS,overwrite)'''
def root_filehandler(path,overwrite=False):
for f in os.listdir(path):
if os.path.isdir(f):
root_filehandler(path+"/"+f)
else:
if ".sv" in f:
GiantString = read_input(f)
SOV = ID_start(GiantString)
ModifiedGS = modified_logNew(GiantString,SOV)
Newname = write_to_output(f,ModifiedGS,overwrite)
def driver(overwrite=False):
root_filehandler(os.getcwd())
driver(True)

2
src/ifu/bpred/bpred.sv
View File

@ -191,7 +191,7 @@ module bpred import cvw::*; #(parameter cvw_t P) (
if(`INSTR_CLASS_PRED) mux2 #(P.XLEN) pcmuxBPWrongInvalidateFlush(PCE, PCF, BPWrongM, NextValidPCE);
else assign NextValidPCE = PCE;
if(P.ZICOUNTERS_SUPPORTED) begin
if(P.ZIHPM_SUPPORTED) begin
logic [P.XLEN-1:0] RASPCD, RASPCE;
logic BTAWrongE, RASPredPCWrongE;
// performance counters

1
src/ifu/decompress.sv
View File

@ -181,4 +181,3 @@ module decompress #(parameter XLEN)(
endcase
end
endmodule

191
src/ifu/ifu.sv
View File

@ -26,116 +26,115 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module ifu import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic StallF, StallD, StallE, StallM, StallW,
input logic FlushD, FlushE, FlushM, FlushW,
output logic IFUStallF, // IFU stalsl pipeline during a multicycle operation
input logic clk, reset,
input logic StallF, StallD, StallE, StallM, StallW,
input logic FlushD, FlushE, FlushM, FlushW,
output logic IFUStallF, // IFU stalsl pipeline during a multicycle operation
// Command from CPU
input logic InvalidateICacheM, // Clears all instruction cache valid bits
input logic CSRWriteFenceM, // CSR write or fence instruction, PCNextF = the next valid PC (typically PCE)
input logic InstrValidD, InstrValidE, InstrValidM,
input logic BranchD, BranchE,
input logic JumpD, JumpE,
input logic InvalidateICacheM, // Clears all instruction cache valid bits
input logic CSRWriteFenceM, // CSR write or fence instruction, PCNextF = the next valid PC (typically PCE)
input logic InstrValidD, InstrValidE, InstrValidM,
input logic BranchD, BranchE,
input logic JumpD, JumpE,
// Bus interface
output logic [P.PA_BITS-1:0] IFUHADDR, // Bus address from IFU to EBU
input logic [P.XLEN-1:0] HRDATA, // Bus read data from IFU to EBU
input logic IFUHREADY, // Bus ready from IFU to EBU
output logic IFUHWRITE, // Bus write operation from IFU to EBU
output logic [2:0] IFUHSIZE, // Bus operation size from IFU to EBU
output logic [2:0] IFUHBURST, // Bus burst from IFU to EBU
output logic [1:0] IFUHTRANS, // Bus transaction type from IFU to EBU
input logic IFUHREADY, // Bus ready from IFU to EBU
output logic IFUHWRITE, // Bus write operation from IFU to EBU
output logic [2:0] IFUHSIZE, // Bus operation size from IFU to EBU
output logic [2:0] IFUHBURST, // Bus burst from IFU to EBU
output logic [1:0] IFUHTRANS, // Bus transaction type from IFU to EBU
output logic [P.XLEN-1:0] PCSpillF, // PCF with possible + 2 to handle spill to HPTW
// Execute
output logic [P.XLEN-1:0] PCLinkE, // The address following the branch instruction. (AKA Fall through address)
input logic PCSrcE, // Executation stage branch is taken
input logic PCSrcE, // Executation stage branch is taken
input logic [P.XLEN-1:0] IEUAdrE, // The branch/jump target address
input logic [P.XLEN-1:0] IEUAdrM, // The branch/jump target address
output logic [P.XLEN-1:0] PCE, // Execution stage instruction address
output logic BPWrongE, // Prediction is wrong
output logic BPWrongM, // Prediction is wrong
output logic BPWrongE, // Prediction is wrong
output logic BPWrongM, // Prediction is wrong
// Mem
output logic CommittedF, // I$ or bus memory operation started, delay interrupts
output logic CommittedF, // I$ or bus memory operation started, delay interrupts
input logic [P.XLEN-1:0] UnalignedPCNextF, // The next PCF, but not aligned to 2 bytes.
output logic [P.XLEN-1:0] PC2NextF, // Selected PC between branch prediction and next valid PC if CSRWriteFence
output logic [31:0] InstrD, // The decoded instruction in Decode stage
output logic [31:0] InstrM, // The decoded instruction in Memory stage
output logic [31:0] InstrOrigM, // Original compressed or uncompressed instruction in Memory stage for Illegal Instruction MTVAL
output logic [31:0] InstrD, // The decoded instruction in Decode stage
output logic [31:0] InstrM, // The decoded instruction in Memory stage
output logic [31:0] InstrOrigM, // Original compressed or uncompressed instruction in Memory stage for Illegal Instruction MTVAL
output logic [P.XLEN-1:0] PCM, // Memory stage instruction address
// branch predictor
output logic [3:0] InstrClassM, // The valid instruction class. 1-hot encoded as jalr, ret, jr (not ret), j, br
output logic BPDirPredWrongM, // Prediction direction is wrong
output logic BTAWrongM, // Prediction target wrong
output logic RASPredPCWrongM, // RAS prediction is wrong
output logic IClassWrongM, // Class prediction is wrong
output logic ICacheStallF, // I$ busy with multicycle operation
output logic [3:0] InstrClassM, // The valid instruction class. 1-hot encoded as jalr, ret, jr (not ret), j, br
output logic BPDirPredWrongM, // Prediction direction is wrong
output logic BTAWrongM, // Prediction target wrong
output logic RASPredPCWrongM, // RAS prediction is wrong
output logic IClassWrongM, // Class prediction is wrong
output logic ICacheStallF, // I$ busy with multicycle operation
// Faults
input logic IllegalBaseInstrD, // Illegal non-compressed instruction
input logic IllegalFPUInstrD, // Illegal FP instruction
output logic InstrPageFaultF, // Instruction page fault
output logic IllegalIEUFPUInstrD, // Illegal instruction including compressed & FP
output logic InstrMisalignedFaultM, // Branch target not aligned to 4 bytes if no compressed allowed (2 bytes if allowed)
input logic IllegalBaseInstrD, // Illegal non-compressed instruction
input logic IllegalFPUInstrD, // Illegal FP instruction
output logic InstrPageFaultF, // Instruction page fault
output logic IllegalIEUFPUInstrD, // Illegal instruction including compressed & FP
output logic InstrMisalignedFaultM, // Branch target not aligned to 4 bytes if no compressed allowed (2 bytes if allowed)
// mmu management
input logic [1:0] PrivilegeModeW, // Priviledge mode in Writeback stage
input logic [1:0] PrivilegeModeW, // Priviledge mode in Writeback stage
input logic [P.XLEN-1:0] PTE, // Hardware page table walker (HPTW) writes Page table entry (PTE) to ITLB
input logic [1:0] PageType, // Hardware page table walker (HPTW) writes PageType to ITLB
input logic ITLBWriteF, // Writes PTE and PageType to ITLB
input logic [1:0] PageType, // Hardware page table walker (HPTW) writes PageType to ITLB
input logic ITLBWriteF, // Writes PTE and PageType to ITLB
input logic [P.XLEN-1:0] SATP_REGW, // Location of the root page table and page table configuration
input logic STATUS_MXR, // Status CSR: make executable page readable
input logic STATUS_SUM, // Status CSR: Supervisor access to user memory
input logic STATUS_MPRV, // Status CSR: modify machine privilege
input logic [1:0] STATUS_MPP, // Status CSR: previous machine privilege level
input logic sfencevmaM, // Virtual memory address fence, invalidate TLB entries
output logic ITLBMissF, // ITLB miss causes HPTW (hardware pagetable walker) walk
output logic InstrUpdateDAF, // ITLB hit needs to update dirty or access bits
input var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration from privileged unit
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP address from privileged unit
output logic InstrAccessFaultF, // Instruction access fault
output logic ICacheAccess, // Report I$ read to performance counters
output logic ICacheMiss // Report I$ miss to performance counters
input logic STATUS_MXR, // Status CSR: make executable page readable
input logic STATUS_SUM, // Status CSR: Supervisor access to user memory
input logic STATUS_MPRV, // Status CSR: modify machine privilege
input logic [1:0] STATUS_MPP, // Status CSR: previous machine privilege level
input logic sfencevmaM, // Virtual memory address fence, invalidate TLB entries
output logic ITLBMissF, // ITLB miss causes HPTW (hardware pagetable walker) walk
output logic InstrUpdateDAF, // ITLB hit needs to update dirty or access bits
input var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration from privileged unit
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0],// PMP address from privileged unit
output logic InstrAccessFaultF, // Instruction access fault
output logic ICacheAccess, // Report I$ read to performance counters
output logic ICacheMiss // Report I$ miss to performance counters
);
localparam [31:0] nop = 32'h00000013; // instruction for NOP
localparam [31:0] nop = 32'h00000013; // instruction for NOP
logic [P.XLEN-1:0] PCNextF; // Next PCF, selected from Branch predictor, Privilege, or PC+2/4
logic BranchMisalignedFaultE; // Branch target not aligned to 4 bytes if no compressed allowed (2 bytes if allowed)
logic [P.XLEN-1:0] PCPlus2or4F; // PCF + 2 (CompressedF) or PCF + 4 (Non-compressed)
logic [P.XLEN-1:0] PCSpillNextF; // Next PCF after possible + 2 to handle spill
logic [P.XLEN-1:0] PCLinkD; // PCF2or4F delayed 1 cycle. This is next PC after a control flow instruction (br or j)
logic [P.XLEN-1:2] PCPlus4F; // PCPlus4F is always PCF + 4. Fancy way to compute PCPlus2or4F
logic [P.XLEN-1:0] PCD; // Decode stage instruction address
logic [P.XLEN-1:0] NextValidPCE; // The PC of the next valid instruction in the pipeline after csr write or fence
logic [P.XLEN-1:0] PCF; // Fetch stage instruction address
logic [P.PA_BITS-1:0] PCPF; // Physical address after address translation
logic [P.XLEN+1:0] PCFExt; //
logic [P.XLEN-1:0] PCNextF; // Next PCF, selected from Branch predictor, Privilege, or PC+2/4
logic BranchMisalignedFaultE; // Branch target not aligned to 4 bytes if no compressed allowed (2 bytes if allowed)
logic [P.XLEN-1:0] PCPlus2or4F; // PCF + 2 (CompressedF) or PCF + 4 (Non-compressed)
logic [P.XLEN-1:0] PCSpillNextF; // Next PCF after possible + 2 to handle spill
logic [P.XLEN-1:0] PCLinkD; // PCF2or4F delayed 1 cycle. This is next PC after a control flow instruction (br or j)
logic [P.XLEN-1:2] PCPlus4F; // PCPlus4F is always PCF + 4. Fancy way to compute PCPlus2or4F
logic [P.XLEN-1:0] PCD; // Decode stage instruction address
logic [P.XLEN-1:0] NextValidPCE; // The PC of the next valid instruction in the pipeline after csr write or fence
logic [P.XLEN-1:0] PCF; // Fetch stage instruction address
logic [P.PA_BITS-1:0] PCPF; // Physical address after address translation
logic [P.XLEN+1:0] PCFExt;
logic [31:0] IROMInstrF; // Instruction from the IROM
logic [31:0] ICacheInstrF; // Instruction from the I$
logic [31:0] InstrRawF; // Instruction from the IROM, I$, or bus
logic CompressedF; // The fetched instruction is compressed
logic CompressedD; // The decoded instruction is compressed
logic CompressedE; // The execution instruction is compressed
logic CompressedM; // The execution instruction is compressed
logic [31:0] PostSpillInstrRawF; // Fetch instruction after merge two halves of spill
logic [31:0] InstrRawD; // Non-decompressed instruction in the Decode stage
logic IllegalIEUInstrD; // IEU Instruction (regular or compressed) is not good
logic [31:0] IROMInstrF; // Instruction from the IROM
logic [31:0] ICacheInstrF; // Instruction from the I$
logic [31:0] InstrRawF; // Instruction from the IROM, I$, or bus
logic CompressedF; // The fetched instruction is compressed
logic CompressedD; // The decoded instruction is compressed
logic CompressedE; // The execution instruction is compressed
logic CompressedM; // The execution instruction is compressed
logic [31:0] PostSpillInstrRawF; // Fetch instruction after merge two halves of spill
logic [31:0] InstrRawD; // Non-decompressed instruction in the Decode stage
logic IllegalIEUInstrD; // IEU Instruction (regular or compressed) is not good
logic [1:0] IFURWF; // IFU alreays read IFURWF = 10
logic [31:0] InstrE; // Instruction in the Execution stage
logic [31:0] NextInstrD, NextInstrE; // Instruction into the next stage after possible stage flush
logic [1:0] IFURWF; // IFU alreays read IFURWF = 10
logic [31:0] InstrE; // Instruction in the Execution stage
logic [31:0] NextInstrD, NextInstrE; // Instruction into the next stage after possible stage flush
logic CacheableF; // PMA indicates instruction address is cacheable
logic SelSpillNextF; // In a spill, stall pipeline and gate local stallF
logic BusStall; // Bus interface busy with multicycle operation
logic IFUCacheBusStallF; // EIther I$ or bus busy with multicycle operation
logic GatedStallD; // StallD gated by selected next spill
logic CacheableF; // PMA indicates instruction address is cacheable
logic SelSpillNextF; // In a spill, stall pipeline and gate local stallF
logic BusStall; // Bus interface busy with multicycle operation
logic IFUCacheBusStallF; // EIther I$ or bus busy with multicycle operation
logic GatedStallD; // StallD gated by selected next spill
// branch predictor signal
logic [P.XLEN-1:0] PC1NextF; // Branch predictor next PCF
logic BusCommittedF; // Bus memory operation in flight, delay interrupts
logic CacheCommittedF; // I$ memory operation started, delay interrupts
logic SelIROM; // PMA indicates instruction address is in the IROM
logic [15:0] InstrRawE, InstrRawM;
logic [P.XLEN-1:0] PC1NextF; // Branch predictor next PCF
logic BusCommittedF; // Bus memory operation in flight, delay interrupts
logic CacheCommittedF; // I$ memory operation started, delay interrupts
logic SelIROM; // PMA indicates instruction address is in the IROM
logic [15:0] InstrRawE, InstrRawM;
assign PCFExt = {2'b00, PCSpillF};
@ -199,6 +198,7 @@ module ifu import cvw::*; #(parameter cvw_t P) (
////////////////////////////////////////////////////////////////////////////////////////////////
// Memory
////////////////////////////////////////////////////////////////////////////////////////////////
// CommittedM tells the CPU's privileged unit the current instruction
// in the memory stage is a memory operaton and that memory operation is either completed
// or is partially executed. Partially completed memory operations need to prevent an interrupts.
@ -206,7 +206,7 @@ module ifu import cvw::*; #(parameter cvw_t P) (
// delay the interrupt until the LSU is in a clean state.
assign CommittedF = CacheCommittedF | BusCommittedF;
logic IgnoreRequest;
logic IgnoreRequest;
assign IgnoreRequest = ITLBMissF | FlushD;
// The IROM uses untranslated addresses, so it is not compatible with virtual memory.
@ -223,12 +223,12 @@ module ifu import cvw::*; #(parameter cvw_t P) (
localparam WORDSPERLINE = P.ICACHE_SUPPORTED ? P.ICACHE_LINELENINBITS/P.XLEN : 1;
localparam LOGBWPL = P.ICACHE_SUPPORTED ? $clog2(WORDSPERLINE) : 1;
if(P.ICACHE_SUPPORTED) begin : icache
localparam LINELEN = P.ICACHE_SUPPORTED ? P.ICACHE_LINELENINBITS : P.XLEN;
localparam LLENPOVERAHBW = P.LLEN / P.AHBW; // Number of AHB beats in a LLEN word. AHBW cannot be larger than LLEN. (implementation limitation)
logic [LINELEN-1:0] FetchBuffer;
localparam LINELEN = P.ICACHE_SUPPORTED ? P.ICACHE_LINELENINBITS : P.XLEN;
localparam LLENPOVERAHBW = P.LLEN / P.AHBW; // Number of AHB beats in a LLEN word. AHBW cannot be larger than LLEN. (implementation limitation)
logic [LINELEN-1:0] FetchBuffer;
logic [P.PA_BITS-1:0] ICacheBusAdr;
logic ICacheBusAck;
logic [1:0] CacheBusRW, BusRW, CacheRWF;
logic ICacheBusAck;
logic [1:0] CacheBusRW, BusRW, CacheRWF;
assign BusRW = ~ITLBMissF & ~CacheableF & ~SelIROM ? IFURWF : '0;
assign CacheRWF = ~ITLBMissF & CacheableF & ~SelIROM ? IFURWF : '0;
@ -255,7 +255,7 @@ module ifu import cvw::*; #(parameter cvw_t P) (
.Flush(FlushD), .CacheBusRW, .HSIZE(IFUHSIZE), .HBURST(IFUHBURST), .HTRANS(IFUHTRANS), .HWSTRB(),
.Funct3(3'b010), .HADDR(IFUHADDR), .HREADY(IFUHREADY), .HWRITE(IFUHWRITE), .CacheBusAdr(ICacheBusAdr),
.BeatCount(), .Cacheable(CacheableF), .SelBusBeat(), .WriteDataM('0),
.CacheBusAck(ICacheBusAck), .HWDATA(), .CacheableOrFlushCacheM(1'b0), .CacheReadDataWordM('0),
.CacheBusAck(ICacheBusAck), .HWDATA(), .CacheableOrFlushCacheM(1'b0), .CacheReadDataWordM('0),
.FetchBuffer, .PAdr(PCPF),
.BusRW, .Stall(GatedStallD),
.BusStall, .BusCommitted(BusCommittedF));
@ -321,7 +321,6 @@ module ifu import cvw::*; #(parameter cvw_t P) (
else PCPlus2or4F = {PCF[P.XLEN-1:2], 2'b10};
else PCPlus2or4F = {PCPlus4F, PCF[1:0]}; // add 4
////////////////////////////////////////////////////////////////////////////////////////////////
// Branch and Jump Predictor
////////////////////////////////////////////////////////////////////////////////////////////////
@ -341,10 +340,10 @@ module ifu import cvw::*; #(parameter cvw_t P) (
assign NextValidPCE = PCE;
end
////////////////////////////////////////////////////////////////////////////////////////////////
// Decode stage pipeline register and compressed instruction decoding.
////////////////////////////////////////////////////////////////////////////////////////////////
// Decode stage pipeline register and logic
flopenrc #(P.XLEN) PCDReg(clk, reset, FlushD, ~StallD, PCF, PCD);
@ -374,10 +373,10 @@ module ifu import cvw::*; #(parameter cvw_t P) (
// Instruction and PC/PCLink pipeline registers
// Cannot use flopenrc for Instr(E/M) as it resets to NOP not 0.
mux2 #(32) FlushInstrEMux(InstrD, nop, FlushE, NextInstrD);
mux2 #(32) FlushInstrMMux(InstrE, nop, FlushM, NextInstrE);
flopenr #(32) InstrEReg(clk, reset, ~StallE, NextInstrD, InstrE);
flopenr #(32) InstrMReg(clk, reset, ~StallM, NextInstrE, InstrM);
mux2 #(32) FlushInstrEMux(InstrD, nop, FlushE, NextInstrD);
mux2 #(32) FlushInstrMMux(InstrE, nop, FlushM, NextInstrE);
flopenr #(32) InstrEReg(clk, reset, ~StallE, NextInstrD, InstrE);
flopenr #(32) InstrMReg(clk, reset, ~StallM, NextInstrE, InstrM);
flopenr #(P.XLEN) PCEReg(clk, reset, ~StallE, PCD, PCE);
flopenr #(P.XLEN) PCMReg(clk, reset, ~StallM, PCE, PCM);
//flopenr #(P.XLEN) PCPDReg(clk, reset, ~StallD, PCPlus2or4F, PCLinkD);

14
src/ifu/irom.sv
View File

@ -25,10 +25,10 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module irom import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic ce, // Chip Enable. 0: Holds IROMInstrF constant
input logic clk,
input logic ce, // Chip Enable. 0: Holds IROMInstrF constant
input logic [P.XLEN-1:0] Adr, // PCNextFSpill
output logic [31:0] IROMInstrF // Instruction read data
output logic [31:0] IROMInstrF // Instruction read data
);
localparam XLENBYTES = {{P.PA_BITS-32{1'b0}}, P.XLEN/8}; // XLEN/8, adjusted for width
@ -36,14 +36,13 @@ module irom import cvw::*; #(parameter cvw_t P) (
localparam OFFSET = $clog2(XLENBYTES);
logic [P.XLEN-1:0] IROMInstrFFull;
logic [31:0] RawIROMInstrF;
logic [1:0] AdrD;
logic [31:0] RawIROMInstrF;
logic [1:0] AdrD;
flopen #(2) AdrReg(clk, ce, Adr[2:1], AdrD);
rom1p1r #(ADDR_WDITH, P.XLEN) rom(.clk, .ce, .addr(Adr[ADDR_WDITH+OFFSET-1:OFFSET]), .dout(IROMInstrFFull));
if (P.XLEN == 32) assign RawIROMInstrF = IROMInstrFFull;
else begin
else begin
// IROM is aligned to XLEN words, but instructions are 32 bits. Select between the two
// haves. Adr is the Next PCF not PCF so we delay 1 cycle.
assign RawIROMInstrF = AdrD[1] ? IROMInstrFFull[63:32] : IROMInstrFFull[31:0];
@ -52,4 +51,3 @@ module irom import cvw::*; #(parameter cvw_t P) (
// The spill logic will handle merging the two together.
assign IROMInstrF = AdrD[0] ? {16'b0, RawIROMInstrF[31:16]} : RawIROMInstrF;
endmodule

37
src/ifu/spill.sv
View File

@ -32,33 +32,33 @@
`include "wally-config.vh"
module spill import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic reset,
input logic StallD, FlushD,
input logic clk,
input logic reset,
input logic StallD, FlushD,
input logic [P.XLEN-1:0] PCF, // 2 byte aligned PC in Fetch stage
input logic [P.XLEN-1:2] PCPlus4F, // PCF + 4
input logic [P.XLEN-1:0] PCNextF, // The next PCF
input logic [31:0] InstrRawF, // Instruction from the IROM, I$, or bus. Used to check if the instruction if compressed
input logic IFUCacheBusStallF, // I$ or bus are stalled. Transition to second fetch of spill after the first is fetched
input logic ITLBMissF, // ITLB miss, ignore memory request
input logic InstrUpdateDAF, // Ignore memory request if the hptw support write and a DA page fault occurs (hptw is still active)
input logic [31:0] InstrRawF, // Instruction from the IROM, I$, or bus. Used to check if the instruction if compressed
input logic IFUCacheBusStallF, // I$ or bus are stalled. Transition to second fetch of spill after the first is fetched
input logic ITLBMissF, // ITLB miss, ignore memory request
input logic InstrUpdateDAF, // Ignore memory request if the hptw support write and a DA page fault occurs (hptw is still active)
output logic [P.XLEN-1:0] PCSpillNextF, // The next PCF for one of the two memory addresses of the spill
output logic [P.XLEN-1:0] PCSpillF, // PCF for one of the two memory addresses of the spill
output logic SelSpillNextF, // During the transition between the two spill operations, the IFU should stall the pipeline
output logic [31:0] PostSpillInstrRawF,// The final 32 bit instruction after merging the two spilled fetches into 1 instruction
output logic CompressedF); // The fetched instruction is compressed
output logic SelSpillNextF, // During the transition between the two spill operations, the IFU should stall the pipeline
output logic [31:0] PostSpillInstrRawF,// The final 32 bit instruction after merging the two spilled fetches into 1 instruction
output logic CompressedF); // The fetched instruction is compressed
// Spill threshold occurs when all the cache offset PC bits are 1 (except [0]). Without a cache this is just PCF[1]
typedef enum logic [1:0] {STATE_READY, STATE_SPILL} statetype;
localparam SPILLTHRESHOLD = P.ICACHE_SUPPORTED ? P.ICACHE_LINELENINBITS/32 : 1;
typedef enum logic [1:0] {STATE_READY, STATE_SPILL} statetype;
localparam SPILLTHRESHOLD = P.ICACHE_SUPPORTED ? P.ICACHE_LINELENINBITS/32 : 1;
statetype CurrState, NextState;
statetype CurrState, NextState;
logic [P.XLEN-1:0] PCPlus2F;
logic TakeSpillF;
logic SpillF;
logic SelSpillF;
logic SpillSaveF;
logic [15:0] InstrFirstHalfF;
logic TakeSpillF;
logic SpillF;
logic SelSpillF;
logic SpillSaveF;
logic [15:0] InstrFirstHalfF;
////////////////////////////////////////////////////////////////////////////////////////////////////
// PC logic
@ -71,7 +71,6 @@ module spill import cvw::*; #(parameter cvw_t P) (
// select between PCF and PCF+2
mux2 #(P.XLEN) pcspillmux(.d0(PCF), .d1(PCPlus2F), .s(SelSpillF), .y(PCSpillF));
////////////////////////////////////////////////////////////////////////////////////////////////////
// Detect spill
////////////////////////////////////////////////////////////////////////////////////////////////////

5
src/lsu/amoalu.sv
View File

@ -30,8 +30,8 @@
module amoalu import cvw::*; #(parameter cvw_t P) (
input logic [P.XLEN-1:0] ReadDataM, // LSU's ReadData
input logic [P.XLEN-1:0] IHWriteDataM, // LSU's WriteData
input logic [6:0] LSUFunct7M, // ALU Operation
input logic [2:0] LSUFunct3M, // Memoy access width
input logic [6:0] LSUFunct7M, // ALU Operation
input logic [2:0] LSUFunct3M, // Memoy access width
output logic [P.XLEN-1:0] AMOResultM // ALU output
);
@ -72,4 +72,3 @@ module amoalu import cvw::*; #(parameter cvw_t P) (
end
end
endmodule

22
src/lsu/atomic.sv
View File

@ -28,24 +28,24 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module atomic import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic reset,
input logic StallW,
input logic clk,
input logic reset,
input logic StallW,
input logic [P.XLEN-1:0] ReadDataM, // LSU ReadData XLEN because FPU does not issue atomic memory operation from FPU registers
input logic [P.XLEN-1:0] IHWriteDataM, // LSU WriteData XLEN because FPU does not issue atomic memory operation from FPU registers
input logic [P.PA_BITS-1:0] PAdrM, // Physical memory address
input logic [6:0] LSUFunct7M, // AMO alu operation gated by HPTW
input logic [2:0] LSUFunct3M, // IEU or HPTW memory operation size
input logic [1:0] LSUAtomicM, // 10: AMO operation, select AMOResultM as the writedata output, 01: LR/SC operation
input logic [1:0] PreLSURWM, // IEU or HPTW Read/Write signal
input logic IgnoreRequest, // On FlushM or TLB miss ignore memory operation
input logic [6:0] LSUFunct7M, // AMO alu operation gated by HPTW
input logic [2:0] LSUFunct3M, // IEU or HPTW memory operation size
input logic [1:0] LSUAtomicM, // 10: AMO operation, select AMOResultM as the writedata output, 01: LR/SC operation
input logic [1:0] PreLSURWM, // IEU or HPTW Read/Write signal
input logic IgnoreRequest, // On FlushM or TLB miss ignore memory operation
output logic [P.XLEN-1:0] IMAWriteDataM, // IEU, HPTW, or AMO write data
output logic SquashSCW, // Store conditional failed disable write to GPR
output logic [1:0] LSURWM // IEU or HPTW Read/Write signal gated by LR/SC
output logic SquashSCW, // Store conditional failed disable write to GPR
output logic [1:0] LSURWM // IEU or HPTW Read/Write signal gated by LR/SC
);
logic [P.XLEN-1:0] AMOResultM;
logic MemReadM;
logic MemReadM;
amoalu #(P) amoalu(.ReadDataM, .IHWriteDataM, .LSUFunct7M, .LSUFunct3M, .AMOResultM);

11
src/lsu/dtim.sv
View File

@ -28,17 +28,17 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module dtim import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic FlushW,
input logic ce, // Chip Enable. 0: Holds ReadDataWordM
input logic [1:0] MemRWM, // Read/Write control
input logic clk,
input logic FlushW,
input logic ce, // Chip Enable. 0: Holds ReadDataWordM
input logic [1:0] MemRWM, // Read/Write control
input logic [P.PA_BITS-1:0] DTIMAdr, // No stall: Execution stage memory address. Stall: Memory stage memory address
input logic [P.LLEN-1:0] WriteDataM, // Write data from IEU
input logic [P.LLEN/8-1:0] ByteMaskM, // Selects which bytes within a word to write
output logic [P.LLEN-1:0] ReadDataWordM // Read data before subword selection
);
logic we;
logic we;
localparam LLENBYTES = P.LLEN/8;
// verilator lint_off WIDTH
@ -52,4 +52,3 @@ module dtim import cvw::*; #(parameter cvw_t P) (
ram1p1rwbe #(.DEPTH(DEPTH), .WIDTH(P.LLEN))
ram(.clk, .ce, .we, .bwe(ByteMaskM), .addr(DTIMAdr[ADDR_WDITH+OFFSET-1:OFFSET]), .dout(ReadDataWordM), .din(WriteDataM));
endmodule

22
src/lsu/lrsc.sv
View File

@ -29,23 +29,23 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module lrsc import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic reset,
input logic StallW,
input logic MemReadM, // Memory read
input logic [1:0] PreLSURWM, // Memory operation from the HPTW or IEU [1]: read, [0]: write
output logic [1:0] LSURWM, // Memory operation after potential squash of SC
input logic [1:0] LSUAtomicM, // Atomic memory operaiton
input logic clk,
input logic reset,
input logic StallW,
input logic MemReadM, // Memory read
input logic [1:0] PreLSURWM, // Memory operation from the HPTW or IEU [1]: read, [0]: write
output logic [1:0] LSURWM, // Memory operation after potential squash of SC
input logic [1:0] LSUAtomicM, // Atomic memory operaiton
input logic [P.PA_BITS-1:0] PAdrM, // Physical memory address
output logic SquashSCW // Squash the store conditional by not allowing rf write
output logic SquashSCW // Squash the store conditional by not allowing rf write
);
// possible bug: *** double check if PreLSURWM needs to be flushed by ignorerequest.
// Handle atomic load reserved / store conditional
logic [P.PA_BITS-1:2] ReservationPAdrW;
logic ReservationValidM, ReservationValidW;
logic lrM, scM, WriteAdrMatchM;
logic SquashSCM;
logic ReservationValidM, ReservationValidW;
logic lrM, scM, WriteAdrMatchM;
logic SquashSCM;
assign lrM = MemReadM & LSUAtomicM[0];
assign scM = PreLSURWM[0] & LSUAtomicM[0];

170
src/lsu/lsu.sv
View File

@ -30,63 +30,63 @@
/////////////////////////////////////////////////////////////////////////////////////////////////////////
module lsu import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic StallM, FlushM, StallW, FlushW,
output logic LSUStallM, // LSU stalls pipeline during a multicycle operation
input logic clk, reset,
input logic StallM, FlushM, StallW, FlushW,
output logic LSUStallM, // LSU stalls pipeline during a multicycle operation
// connected to cpu (controls)
input logic [1:0] MemRWM, // Read/Write control
input logic [2:0] Funct3M, // Size of memory operation
input logic [6:0] Funct7M, // Atomic memory operation function
input logic [1:0] AtomicM, // Atomic memory operation
input logic FlushDCacheM, // Flush D cache to next level of memory
output logic CommittedM, // Delay interrupts while memory operation in flight
output logic SquashSCW, // Store conditional failed disable write to GPR
output logic DCacheMiss, // D cache miss for performance counters
output logic DCacheAccess, // D cache memory access for performance counters
input logic [1:0] MemRWM, // Read/Write control
input logic [2:0] Funct3M, // Size of memory operation
input logic [6:0] Funct7M, // Atomic memory operation function
input logic [1:0] AtomicM, // Atomic memory operation
input logic FlushDCacheM, // Flush D cache to next level of memory
output logic CommittedM, // Delay interrupts while memory operation in flight
output logic SquashSCW, // Store conditional failed disable write to GPR
output logic DCacheMiss, // D cache miss for performance counters
output logic DCacheAccess, // D cache memory access for performance counters
// address and write data
input logic [P.XLEN-1:0] IEUAdrE, // Execution stage memory address
output logic [P.XLEN-1:0] IEUAdrM, // Memory stage memory address
input logic [P.XLEN-1:0] WriteDataM, // Write data from IEU
output logic [P.LLEN-1:0] ReadDataW, // Read data to IEU or FPU
input logic [P.XLEN-1:0] IEUAdrE, // Execution stage memory address
output logic [P.XLEN-1:0] IEUAdrM, // Memory stage memory address
input logic [P.XLEN-1:0] WriteDataM, // Write data from IEU
output logic [P.LLEN-1:0] ReadDataW, // Read data to IEU or FPU
// cpu privilege
input logic [1:0] PrivilegeModeW, // Current privilege mode
input logic BigEndianM, // Swap byte order to big endian
input logic sfencevmaM, // Virtual memory address fence, invalidate TLB entries
output logic DCacheStallM, // D$ busy with multicycle operation
input logic [1:0] PrivilegeModeW, // Current privilege mode
input logic BigEndianM, // Swap byte order to big endian
input logic sfencevmaM, // Virtual memory address fence, invalidate TLB entries
output logic DCacheStallM, // D$ busy with multicycle operation
// fpu
input logic [P.FLEN-1:0] FWriteDataM, // Write data from FPU
input logic FpLoadStoreM, // Selects FPU as store for write data
input logic [P.FLEN-1:0] FWriteDataM, // Write data from FPU
input logic FpLoadStoreM, // Selects FPU as store for write data
// faults
output logic LoadPageFaultM, StoreAmoPageFaultM, // Page fault exceptions
output logic LoadMisalignedFaultM, // Load address misaligned fault
output logic LoadAccessFaultM, // Load access fault (PMA)
output logic HPTWInstrAccessFaultF, // HPTW generated access fault during instruction fetch
output logic LoadPageFaultM, StoreAmoPageFaultM, // Page fault exceptions
output logic LoadMisalignedFaultM, // Load address misaligned fault
output logic LoadAccessFaultM, // Load access fault (PMA)
output logic HPTWInstrAccessFaultF, // HPTW generated access fault during instruction fetch
// cpu hazard unit (trap)
output logic StoreAmoMisalignedFaultM, // Store or AMO address misaligned fault
output logic StoreAmoAccessFaultM, // Store or AMO access fault
output logic StoreAmoMisalignedFaultM, // Store or AMO address misaligned fault
output logic StoreAmoAccessFaultM, // Store or AMO access fault
// connect to ahb
output logic [P.PA_BITS-1:0] LSUHADDR, // Bus address from LSU to EBU
input logic [P.XLEN-1:0] HRDATA, // Bus read data from LSU to EBU
output logic [P.XLEN-1:0] LSUHWDATA, // Bus write data from LSU to EBU
input logic LSUHREADY, // Bus ready from LSU to EBU
output logic LSUHWRITE, // Bus write operation from LSU to EBU
output logic [2:0] LSUHSIZE, // Bus operation size from LSU to EBU
output logic [2:0] LSUHBURST, // Bus burst from LSU to EBU
output logic [1:0] LSUHTRANS, // Bus transaction type from LSU to EBU
output logic [P.XLEN/8-1:0] LSUHWSTRB, // Bus byte write enables from LSU to EBU
output logic [P.PA_BITS-1:0] LSUHADDR, // Bus address from LSU to EBU
input logic [P.XLEN-1:0] HRDATA, // Bus read data from LSU to EBU
output logic [P.XLEN-1:0] LSUHWDATA, // Bus write data from LSU to EBU
input logic LSUHREADY, // Bus ready from LSU to EBU
output logic LSUHWRITE, // Bus write operation from LSU to EBU
output logic [2:0] LSUHSIZE, // Bus operation size from LSU to EBU
output logic [2:0] LSUHBURST, // Bus burst from LSU to EBU
output logic [1:0] LSUHTRANS, // Bus transaction type from LSU to EBU
output logic [P.XLEN/8-1:0] LSUHWSTRB, // Bus byte write enables from LSU to EBU
// page table walker
input logic [P.XLEN-1:0] SATP_REGW, // SATP (supervisor address translation and protection) CSR
input logic STATUS_MXR, STATUS_SUM, STATUS_MPRV, // STATUS CSR bits: make executable readable, supervisor user memory, machine privilege
input logic [1:0] STATUS_MPP, // Machine previous privilege mode
input logic [P.XLEN-1:0] PCSpillF, // Fetch PC
input logic ITLBMissF, // ITLB miss causes HPTW (hardware pagetable walker) walk
input logic InstrUpdateDAF, // ITLB hit needs to update dirty or access bits
output logic [P.XLEN-1:0] PTE, // Page table entry write to ITLB
output logic [1:0] PageType, // Type of page table entry to write to ITLB
output logic ITLBWriteF, // Write PTE to ITLB
output logic SelHPTW, // During a HPTW walk the effective privilege mode becomes S_MODE
input var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration from privileged unit
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0] // PMP address from privileged unit
input logic [P.XLEN-1:0] SATP_REGW, // SATP (supervisor address translation and protection) CSR
input logic STATUS_MXR, STATUS_SUM, STATUS_MPRV, // STATUS CSR bits: make executable readable, supervisor user memory, machine privilege
input logic [1:0] STATUS_MPP, // Machine previous privilege mode
input logic [P.XLEN-1:0] PCSpillF, // Fetch PC
input logic ITLBMissF, // ITLB miss causes HPTW (hardware pagetable walker) walk
input logic InstrUpdateDAF, // ITLB hit needs to update dirty or access bits
output logic [P.XLEN-1:0] PTE, // Page table entry write to ITLB
output logic [1:0] PageType, // Type of page table entry to write to ITLB
output logic ITLBWriteF, // Write PTE to ITLB
output logic SelHPTW, // During a HPTW walk the effective privilege mode becomes S_MODE
input var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration from privileged unit
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0] // PMP address from privileged unit
);
logic [P.XLEN+1:0] IEUAdrExtM; // Memory stage address zero-extended to PA_BITS or XLEN whichever is longer
@ -94,20 +94,20 @@ module lsu import cvw::*; #(parameter cvw_t P) (
logic [P.PA_BITS-1:0] PAdrM; // Physical memory address
logic [P.XLEN+1:0] IHAdrM; // Either IEU or HPTW memory address
logic [1:0] PreLSURWM; // IEU or HPTW Read/Write signal
logic [1:0] LSURWM; // IEU or HPTW Read/Write signal gated by LR/SC
logic [2:0] LSUFunct3M; // IEU or HPTW memory operation size
logic [6:0] LSUFunct7M; // AMO function gated by HPTW
logic [1:0] LSUAtomicM; // AMO signal gated by HPTW
logic [1:0] PreLSURWM; // IEU or HPTW Read/Write signal
logic [1:0] LSURWM; // IEU or HPTW Read/Write signal gated by LR/SC
logic [2:0] LSUFunct3M; // IEU or HPTW memory operation size
logic [6:0] LSUFunct7M; // AMO function gated by HPTW
logic [1:0] LSUAtomicM; // AMO signal gated by HPTW
logic GatedStallW; // Hazard unit StallW gated when SelHPTW = 1
logic GatedStallW; // Hazard unit StallW gated when SelHPTW = 1
logic BusStall; // Bus interface busy with multicycle operation
logic HPTWStall; // HPTW busy with multicycle operation
logic BusStall; // Bus interface busy with multicycle operation
logic HPTWStall; // HPTW busy with multicycle operation
logic CacheableM; // PMA indicates memory address is cacheable
logic BusCommittedM; // Bus memory operation in flight, delay interrupts
logic DCacheCommittedM; // D$ memory operation started, delay interrupts
logic CacheableM; // PMA indicates memory address is cacheable
logic BusCommittedM; // Bus memory operation in flight, delay interrupts
logic DCacheCommittedM; // D$ memory operation started, delay interrupts
logic [P.LLEN-1:0] DTIMReadDataWordM; // DTIM read data
logic [P.LLEN-1:0] DCacheReadDataWordM; // D$ read data
@ -123,16 +123,15 @@ module lsu import cvw::*; #(parameter cvw_t P) (
logic [P.LLEN-1:0] LSUWriteDataM; // Final write data
logic [(P.LLEN-1)/8:0] ByteMaskM; // Selects which bytes within a word to write
logic DTLBMissM; // DTLB miss causes HPTW walk
logic DTLBWriteM; // Writes PTE and PageType to DTLB
logic DataUpdateDAM; // DTLB hit needs to update dirty or access bits
logic LSULoadAccessFaultM; // Load acces fault
logic LSUStoreAmoAccessFaultM; // Store access fault
logic IgnoreRequestTLB; // On either ITLB or DTLB miss, ignore miss so HPTW can handle
logic IgnoreRequest; // On FlushM or TLB miss ignore memory operation
logic SelDTIM; // Select DTIM rather than bus or D$
logic DTLBMissM; // DTLB miss causes HPTW walk
logic DTLBWriteM; // Writes PTE and PageType to DTLB
logic DataUpdateDAM; // DTLB hit needs to update dirty or access bits
logic LSULoadAccessFaultM; // Load acces fault
logic LSUStoreAmoAccessFaultM; // Store access fault
logic IgnoreRequestTLB; // On either ITLB or DTLB miss, ignore miss so HPTW can handle
logic IgnoreRequest; // On FlushM or TLB miss ignore memory operation
logic SelDTIM; // Select DTIM rather than bus or D$
/////////////////////////////////////////////////////////////////////////////////////////////
// Pipeline for IEUAdr E to M
// Zero-extend address to 34 bits for XLEN=32
@ -222,7 +221,7 @@ module lsu import cvw::*; #(parameter cvw_t P) (
if (P.DTIM_SUPPORTED) begin : dtim
logic [P.PA_BITS-1:0] DTIMAdr;
logic [1:0] DTIMMemRWM;
logic [1:0] DTIMMemRWM;
// The DTIM uses untranslated addresses, so it is not compatible with virtual memory.
mux2 #(P.PA_BITS) DTIMAdrMux(IEUAdrExtE[P.PA_BITS-1:0], IEUAdrExtM[P.PA_BITS-1:0], MemRWM[0], DTIMAdr);
@ -237,23 +236,23 @@ module lsu import cvw::*; #(parameter cvw_t P) (
if (P.BUS_SUPPORTED) begin : bus
if(P.DCACHE_SUPPORTED) begin : dcache
localparam LLENWORDSPERLINE = P.DCACHE_LINELENINBITS/P.LLEN; // Number of LLEN words in cacheline
localparam LLENLOGBWPL = $clog2(LLENWORDSPERLINE); // Log2 of ^
localparam LLENLOGBWPL = $clog2(LLENWORDSPERLINE); // Log2 of ^
localparam BEATSPERLINE = P.DCACHE_LINELENINBITS/P.AHBW; // Number of AHBW words (beats) in cacheline
localparam AHBWLOGBWPL = $clog2(BEATSPERLINE); // Log2 of ^
localparam LINELEN = P.DCACHE_LINELENINBITS; // Number of bits in cacheline
localparam AHBWLOGBWPL = $clog2(BEATSPERLINE); // Log2 of ^
localparam LINELEN = P.DCACHE_LINELENINBITS; // Number of bits in cacheline
localparam LLENPOVERAHBW = P.LLEN / P.AHBW; // Number of AHB beats in a LLEN word. AHBW cannot be larger than LLEN. (implementation limitation)
logic [LINELEN-1:0] FetchBuffer; // Temporary buffer to hold partially fetched cacheline
logic [P.PA_BITS-1:0] DCacheBusAdr; // Cacheline address to fetch or writeback.
logic [AHBWLOGBWPL-1:0] BeatCount; // Position within a cacheline. ahbcacheinterface to cache
logic DCacheBusAck; // ahbcacheinterface completed fetch or writeback
logic SelBusBeat; // ahbcacheinterface selects postion in cacheline with BeatCount
logic [1:0] CacheBusRW; // Cache sends request to ahbcacheinterface
logic [1:0] BusRW; // Uncached bus memory access
logic CacheableOrFlushCacheM; // Memory address is cacheable or operation is a cache flush
logic [1:0] CacheRWM; // Cache read (10), write (01), AMO (11)
logic [1:0] CacheAtomicM; // Cache AMO
logic FlushDCache; // Suppress d cache flush if there is an ITLB miss.
logic [LINELEN-1:0] FetchBuffer; // Temporary buffer to hold partially fetched cacheline
logic [P.PA_BITS-1:0] DCacheBusAdr; // Cacheline address to fetch or writeback.
logic [AHBWLOGBWPL-1:0] BeatCount; // Position within a cacheline. ahbcacheinterface to cache
logic DCacheBusAck; // ahbcacheinterface completed fetch or writeback
logic SelBusBeat; // ahbcacheinterface selects postion in cacheline with BeatCount
logic [1:0] CacheBusRW; // Cache sends request to ahbcacheinterface
logic [1:0] BusRW; // Uncached bus memory access
logic CacheableOrFlushCacheM; // Memory address is cacheable or operation is a cache flush
logic [1:0] CacheRWM; // Cache read (10), write (01), AMO (11)
logic [1:0] CacheAtomicM; // Cache AMO
logic FlushDCache; // Suppress d cache flush if there is an ITLB miss.
assign BusRW = ~CacheableM & ~IgnoreRequestTLB & ~SelDTIM ? LSURWM : '0;
assign CacheableOrFlushCacheM = CacheableM | FlushDCacheM;
@ -320,6 +319,7 @@ module lsu import cvw::*; #(parameter cvw_t P) (
/////////////////////////////////////////////////////////////////////////////////////////////
// Atomic operations
/////////////////////////////////////////////////////////////////////////////////////////////
if (P.A_SUPPORTED) begin:atomic
atomic #(P) atomic(.clk, .reset, .StallW, .ReadDataM(ReadDataM[P.XLEN-1:0]), .IHWriteDataM, .PAdrM,
.LSUFunct7M, .LSUFunct3M, .LSUAtomicM, .PreLSURWM, .IgnoreRequest,
@ -335,6 +335,7 @@ module lsu import cvw::*; #(parameter cvw_t P) (
/////////////////////////////////////////////////////////////////////////////////////////////
// Subword Accesses
/////////////////////////////////////////////////////////////////////////////////////////////
subwordread #(P.LLEN) subwordread(.ReadDataWordMuxM(LittleEndianReadDataWordM), .PAdrM(PAdrM[2:0]), .BigEndianM,
.FpLoadStoreM, .Funct3M(LSUFunct3M), .ReadDataM);
subwordwrite #(P.LLEN) subwordwrite(.LSUFunct3M, .IMAFWriteDataM, .LittleEndianWriteDataM);
@ -361,5 +362,4 @@ module lsu import cvw::*; #(parameter cvw_t P) (
assign LSUWriteDataM = LittleEndianWriteDataM;
assign LittleEndianReadDataWordM = ReadDataWordMuxM;
end
endmodule

6
src/lsu/subwordread.sv
View File

@ -29,12 +29,12 @@
module subwordread #(parameter LLEN)
(
input logic [LLEN-1:0] ReadDataWordMuxM,
input logic [LLEN-1:0] ReadDataWordMuxM,
input logic [2:0] PAdrM,
input logic [2:0] Funct3M,
input logic FpLoadStoreM,
input logic BigEndianM,
output logic [LLEN-1:0] ReadDataM
output logic [LLEN-1:0] ReadDataM
);
logic [7:0] ByteM;
@ -89,7 +89,7 @@ module subwordread #(parameter LLEN)
//3'b100: ReadDataM = FpLoadStoreM ? ReadDataWordMuxM : {{LLEN-8{1'b0}}, ByteM[7:0]}; // lbu/flq - only needed when LLEN=128
3'b101: ReadDataM = {{LLEN-16{1'b0}}, HalfwordM[15:0]}; // lhu
3'b110: ReadDataM = {{LLEN-32{1'b0}}, WordM[31:0]}; // lwu
default: ReadDataM = ReadDataWordMuxM; // Shouldn't happen
default: ReadDataM = ReadDataWordMuxM; // Shouldn't happen
endcase
end else begin:swrmux // 32-bit

20
src/lsu/subwordwrite.sv
View File

@ -28,8 +28,8 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module subwordwrite #(parameter LLEN) (
input logic [2:0] LSUFunct3M,
input logic [LLEN-1:0] IMAFWriteDataM,
input logic [2:0] LSUFunct3M,
input logic [LLEN-1:0] IMAFWriteDataM,
output logic [LLEN-1:0] LittleEndianWriteDataM
);
@ -46,18 +46,18 @@ module subwordwrite #(parameter LLEN) (
end else if (LLEN == 64) begin:sww
always_comb
case(LSUFunct3M[1:0])
2'b00: LittleEndianWriteDataM = {8{IMAFWriteDataM[7:0]}}; // sb
2'b01: LittleEndianWriteDataM = {4{IMAFWriteDataM[15:0]}}; // sh
2'b10: LittleEndianWriteDataM = {2{IMAFWriteDataM[31:0]}}; // sw
2'b11: LittleEndianWriteDataM = IMAFWriteDataM; // sd
2'b00: LittleEndianWriteDataM = {8{IMAFWriteDataM[7:0]}}; // sb
2'b01: LittleEndianWriteDataM = {4{IMAFWriteDataM[15:0]}}; // sh
2'b10: LittleEndianWriteDataM = {2{IMAFWriteDataM[31:0]}}; // sw
2'b11: LittleEndianWriteDataM = IMAFWriteDataM; // sd
endcase
end else begin:sww // 32-bit
always_comb
case(LSUFunct3M[1:0])
2'b00: LittleEndianWriteDataM = {4{IMAFWriteDataM[7:0]}}; // sb
2'b01: LittleEndianWriteDataM = {2{IMAFWriteDataM[15:0]}}; // sh
2'b10: LittleEndianWriteDataM = IMAFWriteDataM; // sw
default: LittleEndianWriteDataM = IMAFWriteDataM; // shouldn't happen
2'b00: LittleEndianWriteDataM = {4{IMAFWriteDataM[7:0]}}; // sb
2'b01: LittleEndianWriteDataM = {2{IMAFWriteDataM[15:0]}}; // sh
2'b10: LittleEndianWriteDataM = IMAFWriteDataM; // sw
default: LittleEndianWriteDataM = IMAFWriteDataM; // shouldn't happen
endcase
end
endmodule

6
src/lsu/swbytemask.sv
View File

@ -28,9 +28,9 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module swbytemask #(parameter WORDLEN)(
input logic [2:0] Size,
input logic [$clog2(WORDLEN/8)-1:0] Adr,
output logic [WORDLEN/8-1:0] ByteMask
input logic [2:0] Size,
input logic [$clog2(WORDLEN/8)-1:0] Adr,
output logic [WORDLEN/8-1:0] ByteMask
);
assign ByteMask = ((2**(2**Size))-1) << Adr;

50
src/mdu/div.sv
View File

@ -27,38 +27,38 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module div import cvw::*; #(parameter cvw_t P) (
input logic clk,
input logic reset,
input logic StallM,
input logic FlushE,
input logic IntDivE, // integer division/remainder instruction of any type
input logic DivSignedE, // signed division
input logic W64E, // W-type instructions (divw, divuw, remw, remuw)
input logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE,// Forwarding mux outputs for Source A and B
output logic DivBusyE, // Divide is busy - stall pipeline
output logic [P.XLEN-1:0] QuotM, RemM // Quotient and remainder outputs
input logic clk,
input logic reset,
input logic StallM,
input logic FlushE,
input logic IntDivE, // integer division/remainder instruction of any type
input logic DivSignedE, // signed division
input logic W64E, // W-type instructions (divw, divuw, remw, remuw)
input logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // Forwarding mux outputs for Source A and B
output logic DivBusyE, // Divide is busy - stall pipeline
output logic [P.XLEN-1:0] QuotM, RemM // Quotient and remainder outputs
);
localparam STEPBITS = $clog2(P.XLEN/P.IDIV_BITSPERCYCLE); // Number of steps
typedef enum logic [1:0] {IDLE, BUSY, DONE} statetype; // division FSM state
typedef enum logic [1:0] {IDLE, BUSY, DONE} statetype; // division FSM state
statetype state;
logic [P.XLEN-1:0] W[P.IDIV_BITSPERCYCLE:0]; // Residual for each of k steps
logic [P.XLEN-1:0] XQ[P.IDIV_BITSPERCYCLE:0]; // dividend/quotient for each of k steps
logic [P.XLEN-1:0] WNext, XQNext; // initialized W and XQ going into registers
logic [P.XLEN-1:0] DinE, XinE; // divisor & dividend, possibly truncated to 32 bits
logic [P.XLEN-1:0] DnE; // DnE = ~DinE
logic [P.XLEN-1:0] DAbsBE; // absolute value of D
logic [P.XLEN-1:0] DAbsB; // registered absolute value of D, constant during division
logic [P.XLEN-1:0] XnE; // DXnE = ~XinE
logic [P.XLEN-1:0] XInitE; // |X|, or original X for divide by 0
logic [P.XLEN-1:0] WnM, XQnM; // negated residual W and quotient XQ for postprocessing sign correction
logic [STEPBITS:0] step; // division step
logic Div0E, Div0M; // divide by 0
logic DivStartE; // start integer division
logic SignXE, SignDE; // sign of dividend and divisor
logic NegQE, NegWM, NegQM; // negate quotient or residual during postprocessing
logic [P.XLEN-1:0] WNext, XQNext; // initialized W and XQ going into registers
logic [P.XLEN-1:0] DinE, XinE; // divisor & dividend, possibly truncated to 32 bits
logic [P.XLEN-1:0] DnE; // DnE = ~DinE
logic [P.XLEN-1:0] DAbsBE; // absolute value of D
logic [P.XLEN-1:0] DAbsB; // registered absolute value of D, constant during division
logic [P.XLEN-1:0] XnE; // DXnE = ~XinE
logic [P.XLEN-1:0] XInitE; // |X|, or original X for divide by 0
logic [P.XLEN-1:0] WnM, XQnM; // negated residual W and quotient XQ for postprocessing sign correction
logic [STEPBITS:0] step; // division step
logic Div0E, Div0M; // divide by 0
logic DivStartE; // start integer division
logic SignXE, SignDE; // sign of dividend and divisor
logic NegQE, NegWM, NegQM; // negate quotient or residual during postprocessing
//////////////////////////////
// Execute Stage: prepare for division calculation with control logic, W logic and absolute values, initialize W and XQ
@ -131,7 +131,7 @@ module div import cvw::*; #(parameter cvw_t P) (
step <= 1;
if (Div0E) state <= DONE;
else state <= BUSY;
end else if (state == BUSY) begin // pause one cycle at beginning of signed operations for absolute value
end else if (state == BUSY) begin // pause one cycle at beginning of signed operations for absolute value
if (step[STEPBITS] | (P.XLEN==64) & W64E & step[STEPBITS-1]) begin // complete in half the time for W-type instructions
state <= DONE;
end

30
src/mdu/mdu.sv
View File

@ -27,21 +27,21 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module mdu import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic StallM, StallW,
input logic FlushE, FlushM, FlushW,
input logic clk, reset,
input logic StallM, StallW,
input logic FlushE, FlushM, FlushW,
input logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // inputs A and B from IEU forwarding mux output
input logic [2:0] Funct3E, Funct3M, // type of MDU operation
input logic IntDivE, W64E, // Integer division/remainder, and W-type instrutions
input logic [2:0] Funct3E, Funct3M, // type of MDU operation
input logic IntDivE, W64E, // Integer division/remainder, and W-type instrutions
output logic [P.XLEN-1:0] MDUResultW, // multiply/divide result
output logic DivBusyE // busy signal to stall pipeline in Execute stage
output logic DivBusyE // busy signal to stall pipeline in Execute stage
);
logic [P.XLEN*2-1:0] ProdM; // double-width product from mul
logic [P.XLEN-1:0] QuotM, RemM; // quotient and remainder from intdivrestoring
logic [P.XLEN-1:0] PrelimResultM; // selected result before W truncation
logic [P.XLEN-1:0] MDUResultM; // result after W truncation
logic W64M; // W-type instruction
logic W64M; // W-type instruction
// Multiplier
mul #(P.XLEN) mul(.clk, .reset, .StallM, .FlushM, .ForwardedSrcAE, .ForwardedSrcBE, .Funct3E, .ProdM);
@ -64,13 +64,13 @@ module mdu import cvw::*; #(parameter cvw_t P) (
always_comb
case (Funct3M)
3'b000: PrelimResultM = ProdM[P.XLEN-1:0]; // mul
3'b001: PrelimResultM = ProdM[P.XLEN*2-1:P.XLEN]; // mulh
3'b010: PrelimResultM = ProdM[P.XLEN*2-1:P.XLEN]; // mulhsu
3'b011: PrelimResultM = ProdM[P.XLEN*2-1:P.XLEN]; // mulhu
3'b100: PrelimResultM = QuotM; // div
3'b101: PrelimResultM = QuotM; // divu
3'b110: PrelimResultM = RemM; // rem
3'b111: PrelimResultM = RemM; // remu
3'b001: PrelimResultM = ProdM[P.XLEN*2-1:P.XLEN]; // mulh
3'b010: PrelimResultM = ProdM[P.XLEN*2-1:P.XLEN]; // mulhsu
3'b011: PrelimResultM = ProdM[P.XLEN*2-1:P.XLEN]; // mulhu
3'b100: PrelimResultM = QuotM; // div
3'b101: PrelimResultM = QuotM; // divu
3'b110: PrelimResultM = RemM; // rem
3'b111: PrelimResultM = RemM; // remu
endcase
// Handle sign extension for W-type instructions
@ -84,5 +84,3 @@ module mdu import cvw::*; #(parameter cvw_t P) (
// Writeback stage pipeline register
flopenrc #(P.XLEN) MDUResultWReg(clk, reset, FlushW, ~StallW, MDUResultM, MDUResultW);
endmodule // mdu

13
src/mdu/mul.sv
View File

@ -29,9 +29,9 @@
module mul #(parameter XLEN) (
input logic clk, reset,
input logic StallM, FlushM,
input logic [XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // source A and B from after Forwarding mux
input logic [XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE, // source A and B from after Forwarding mux
input logic [2:0] Funct3E, // type of multiply
output logic [XLEN*2-1:0] ProdM // double-widthproduct
output logic [XLEN*2-1:0] ProdM // double-widthproduct
);
// Number systems
@ -48,12 +48,12 @@ module mul #(parameter XLEN) (
// Signed * Unsigned = P' + ( PA - PB)*2^(XLEN-1) - PP*2^(2XLEN-2)
// Unsigned * Unsigned = P' + ( PA + PB)*2^(XLEN-1) + PP*2^(2XLEN-2)
logic [XLEN-1:0] Aprime, Bprime; // lower bits of source A and B
logic [XLEN-1:0] Aprime, Bprime; // lower bits of source A and B
logic MULH, MULHSU; // type of multiply
logic [XLEN-2:0] PA, PB; // product of msb and lsbs
logic [XLEN-2:0] PA, PB; // product of msb and lsbs
logic PP; // product of msbs
logic [XLEN*2-1:0] PP1E, PP2E, PP3E, PP4E; // partial products
logic [XLEN*2-1:0] PP1M, PP2M, PP3M, PP4M; // registered partial proudcts
logic [XLEN*2-1:0] PP1E, PP2E, PP3E, PP4E; // partial products
logic [XLEN*2-1:0] PP1M, PP2M, PP3M, PP4M; // registered partial proudcts
//////////////////////////////
// Execute Stage: Compute partial products
@ -90,4 +90,3 @@ module mul #(parameter XLEN) (
// add up partial products; this multi-input add implies CSAs and a final CPA
assign ProdM = PP1M + PP2M + PP3M + PP4M; //ForwardedSrcAE * ForwardedSrcBE;
endmodule

5
src/mmu/adrdec.sv
View File

@ -27,8 +27,8 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module adrdec #(parameter PA_BITS) (
input logic [PA_BITS-1:0] PhysicalAddress, // Physical address to decode
input logic [PA_BITS-1:0] Base, Range, // Base and range of peripheral addresses
input logic [PA_BITS-1:0] PhysicalAddress, // Physical address to decode
input logic [PA_BITS-1:0] Base, Range, // Base and range of peripheral addresses
input logic Supported, // Is this peripheral supported?
input logic AccessValid, // Is the access type valid?
input logic [1:0] Size, // Size of access
@ -50,4 +50,3 @@ module adrdec #(parameter PA_BITS) (
// Select this peripheral if the address matches, the peripheral is supported, and the type and size of access is ok
assign Sel = Match & Supported & AccessValid & SizeValid;
endmodule

8
src/mmu/adrdecs.sv
View File

@ -30,12 +30,12 @@
module adrdecs import cvw::*; #(parameter cvw_t P) (
input logic [P.PA_BITS-1:0] PhysicalAddress,
input logic AccessRW, AccessRX, AccessRWX,
input logic [1:0] Size,
output logic [10:0] SelRegions
input logic AccessRW, AccessRX, AccessRWX,
input logic [1:0] Size,
output logic [10:0] SelRegions
);
localparam logic [3:0] SUPPORTED_SIZE = (P.LLEN == 32 ? 4'b0111 : 4'b1111);
localparam logic [3:0] SUPPORTED_SIZE = (P.LLEN == 32 ? 4'b0111 : 4'b1111);
// Determine which region of physical memory (if any) is being accessed
adrdec #(P.PA_BITS) dtimdec(PhysicalAddress, P.DTIM_BASE[P.PA_BITS-1:0], P.DTIM_RANGE[P.PA_BITS-1:0], P.DTIM_SUPPORTED, AccessRW, Size, SUPPORTED_SIZE, SelRegions[10]);
adrdec #(P.PA_BITS) iromdec(PhysicalAddress, P.IROM_BASE[P.PA_BITS-1:0], P.IROM_RANGE[P.PA_BITS-1:0], P.IROM_SUPPORTED, AccessRX, Size, SUPPORTED_SIZE, SelRegions[9]);

118
src/mmu/hptw.sv
View File

@ -30,39 +30,39 @@
///////////////////////////////////////////
module hptw import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic clk, reset,
input logic [P.XLEN-1:0] SATP_REGW, // includes SATP.MODE to determine number of levels in page table
input logic [P.XLEN-1:0] PCSpillF, // addresses to translate
input logic [P.XLEN+1:0] IEUAdrExtM, // addresses to translate
input logic [1:0] MemRWM, AtomicM,
input logic [1:0] MemRWM, AtomicM,
// system status
input logic STATUS_MXR, STATUS_SUM, STATUS_MPRV,
input logic [1:0] STATUS_MPP,
input logic [1:0] PrivilegeModeW,
input logic STATUS_MXR, STATUS_SUM, STATUS_MPRV,
input logic [1:0] STATUS_MPP,
input logic [1:0] PrivilegeModeW,
input logic [P.XLEN-1:0] ReadDataM, // page table entry from LSU
input logic [P.XLEN-1:0] WriteDataM,
input logic DCacheStallM, // stall from LSU
input logic [2:0] Funct3M,
input logic [6:0] Funct7M,
input logic ITLBMissF,
input logic DTLBMissM,
input logic FlushW,
input logic InstrUpdateDAF,
input logic DataUpdateDAM,
input logic DCacheStallM, // stall from LSU
input logic [2:0] Funct3M,
input logic [6:0] Funct7M,
input logic ITLBMissF,
input logic DTLBMissM,
input logic FlushW,
input logic InstrUpdateDAF,
input logic DataUpdateDAM,
output logic [P.XLEN-1:0] PTE, // page table entry to TLBs
output logic [1:0] PageType, // page type to TLBs
output logic ITLBWriteF, DTLBWriteM, // write TLB with new entry
output logic [1:0] PreLSURWM,
output logic [1:0] PageType, // page type to TLBs
output logic ITLBWriteF, DTLBWriteM, // write TLB with new entry
output logic [1:0] PreLSURWM,
output logic [P.XLEN+1:0] IHAdrM,
output logic [P.XLEN-1:0] IHWriteDataM,
output logic [1:0] LSUAtomicM,
output logic [2:0] LSUFunct3M,
output logic [6:0] LSUFunct7M,
output logic IgnoreRequestTLB,
output logic SelHPTW,
output logic HPTWStall,
input logic LSULoadAccessFaultM, LSUStoreAmoAccessFaultM,
output logic LoadAccessFaultM, StoreAmoAccessFaultM, HPTWInstrAccessFaultF
output logic [1:0] LSUAtomicM,
output logic [2:0] LSUFunct3M,
output logic [6:0] LSUFunct7M,
output logic IgnoreRequestTLB,
output logic SelHPTW,
output logic HPTWStall,
input logic LSULoadAccessFaultM, LSUStoreAmoAccessFaultM,
output logic LoadAccessFaultM, StoreAmoAccessFaultM, HPTWInstrAccessFaultF
);
typedef enum logic [3:0] {L0_ADR, L0_RD,
@ -72,39 +72,39 @@ module hptw import cvw::*; #(parameter cvw_t P) (
LEAF, IDLE, UPDATE_PTE,
FAULT} statetype;
logic DTLBWalk; // register TLBs translation miss requests
logic DTLBWalk; // register TLBs translation miss requests
logic [P.PPN_BITS-1:0] BasePageTablePPN;
logic [P.PPN_BITS-1:0] CurrentPPN;
logic Executable, Writable, Readable, Valid, PTE_U;
logic Misaligned, MegapageMisaligned;
logic ValidPTE, LeafPTE, ValidLeafPTE, ValidNonLeafPTE;
logic StartWalk;
logic TLBMiss;
logic PRegEn;
logic [1:0] NextPageType;
logic Executable, Writable, Readable, Valid, PTE_U;
logic Misaligned, MegapageMisaligned;
logic ValidPTE, LeafPTE, ValidLeafPTE, ValidNonLeafPTE;
logic StartWalk;
logic TLBMiss;
logic PRegEn;
logic [1:0] NextPageType;
logic [P.SVMODE_BITS-1:0] SvMode;
logic [P.XLEN-1:0] TranslationVAdr;
logic [P.XLEN-1:0] NextPTE;
logic UpdatePTE;
logic HPTWUpdateDA;
logic UpdatePTE;
logic HPTWUpdateDA;
logic [P.PA_BITS-1:0] HPTWReadAdr;
logic SelHPTWAdr;
logic SelHPTWAdr;
logic [P.XLEN+1:0] HPTWAdrExt;
logic ITLBMissOrUpdateDAF;
logic DTLBMissOrUpdateDAM;
logic LSUAccessFaultM;
logic ITLBMissOrUpdateDAF;
logic DTLBMissOrUpdateDAM;
logic LSUAccessFaultM;
logic [P.PA_BITS-1:0] HPTWAdr;
logic [1:0] HPTWRW;
logic [2:0] HPTWSize; // 32 or 64 bit access
statetype WalkerState, NextWalkerState, InitialWalkerState;
logic HPTWLoadAccessFault, HPTWStoreAmoAccessFault, HPTWInstrAccessFault;
logic HPTWLoadAccessFaultDelay, HPTWStoreAmoAccessFaultDelay, HPTWInstrAccessFaultDelay;
logic HPTWAccessFaultDelay;
logic TakeHPTWFault, TakeHPTWFaultDelay;
logic [1:0] HPTWRW;
logic [2:0] HPTWSize; // 32 or 64 bit access
statetype WalkerState, NextWalkerState, InitialWalkerState;
logic HPTWLoadAccessFault, HPTWStoreAmoAccessFault, HPTWInstrAccessFault;
logic HPTWLoadAccessFaultDelay, HPTWStoreAmoAccessFaultDelay, HPTWInstrAccessFaultDelay;
logic HPTWAccessFaultDelay;
logic TakeHPTWFault, TakeHPTWFaultDelay;
// map hptw access faults onto either the original LSU load/store fault or instruction access fault
assign LSUAccessFaultM = LSULoadAccessFaultM | LSUStoreAmoAccessFaultM;
assign HPTWLoadAccessFault = LSUAccessFaultM & DTLBWalk & MemRWM[1] & ~MemRWM[0];
assign LSUAccessFaultM = LSULoadAccessFaultM | LSUStoreAmoAccessFaultM;
assign HPTWLoadAccessFault = LSUAccessFaultM & DTLBWalk & MemRWM[1] & ~MemRWM[0];
assign HPTWStoreAmoAccessFault = LSUAccessFaultM & DTLBWalk & MemRWM[0];
assign HPTWInstrAccessFault = LSUAccessFaultM & ~DTLBWalk;
@ -140,16 +140,16 @@ module hptw import cvw::*; #(parameter cvw_t P) (
assign ValidNonLeafPTE = ValidPTE & ~LeafPTE;
if(P.SVADU_SUPPORTED) begin : hptwwrites
logic ReadAccess, WriteAccess;
logic InvalidRead, InvalidWrite, InvalidOp;
logic UpperBitsUnequal;
logic OtherPageFault;
logic [1:0] EffectivePrivilegeMode;
logic ImproperPrivilege;
logic SaveHPTWAdr, SelHPTWWriteAdr;
logic ReadAccess, WriteAccess;
logic InvalidRead, InvalidWrite, InvalidOp;
logic UpperBitsUnequal;
logic OtherPageFault;
logic [1:0] EffectivePrivilegeMode;
logic ImproperPrivilege;
logic SaveHPTWAdr, SelHPTWWriteAdr;
logic [P.PA_BITS-1:0] HPTWWriteAdr;
logic SetDirty;
logic Dirty, Accessed;
logic SetDirty;
logic Dirty, Accessed;
logic [P.XLEN-1:0] AccessedPTE;
assign AccessedPTE = {PTE[P.XLEN-1:8], (SetDirty | PTE[7]), 1'b1, PTE[5:0]}; // set accessed bit, conditionally set dirty bit
@ -194,8 +194,8 @@ module hptw import cvw::*; #(parameter cvw_t P) (
end
// Enable and select signals based on states
assign StartWalk = (WalkerState == IDLE) & TLBMiss;
assign HPTWRW[1] = (WalkerState == L3_RD) | (WalkerState == L2_RD) | (WalkerState == L1_RD) | (WalkerState == L0_RD);
assign StartWalk = (WalkerState == IDLE) & TLBMiss;
assign HPTWRW[1] = (WalkerState == L3_RD) | (WalkerState == L2_RD) | (WalkerState == L1_RD) | (WalkerState == L0_RD);
assign DTLBWriteM = (WalkerState == LEAF & ~HPTWUpdateDA) & DTLBWalk;
assign ITLBWriteF = (WalkerState == LEAF & ~HPTWUpdateDA) & ~DTLBWalk;
@ -310,7 +310,7 @@ module hptw import cvw::*; #(parameter cvw_t P) (
// multiplex the outputs to LSU
if(P.XLEN == 64) assign HPTWAdrExt = {{(P.XLEN+2-P.PA_BITS){1'b0}}, HPTWAdr}; // extend to 66 bits
else assign HPTWAdrExt = HPTWAdr;
else assign HPTWAdrExt = HPTWAdr;
mux2 #(2) rwmux(MemRWM, HPTWRW, SelHPTW, PreLSURWM);
mux2 #(3) sizemux(Funct3M, HPTWSize, SelHPTW, LSUFunct3M);
mux2 #(7) funct7mux(Funct7M, 7'b0, SelHPTW, LSUFunct7M);

80
src/mmu/mmu.sv
View File

@ -28,48 +28,48 @@
module mmu import cvw::*; #(parameter cvw_t P,
parameter TLB_ENTRIES = 8, IMMU = 0) (
input logic clk, reset,
input logic [P.XLEN-1:0] SATP_REGW, // Current value of satp CSR (from privileged unit)
input logic STATUS_MXR, // Status CSR: make executable page readable
input logic STATUS_SUM, // Status CSR: Supervisor access to user memory
input logic STATUS_MPRV, // Status CSR: modify machine privilege
input logic [1:0] STATUS_MPP, // Status CSR: previous machine privilege level
input logic [1:0] PrivilegeModeW, // Current privilege level of the processeor
input logic DisableTranslation, // virtual address translation disabled during D$ flush and HPTW walk that use physical addresses
input logic [P.XLEN+1:0] VAdr, // virtual/physical address from IEU or physical address from HPTW
input logic [1:0] Size, // access size: 00 = 8 bits, 01 = 16 bits, 10 = 32 bits , 11 = 64 bits
input logic [P.XLEN-1:0] PTE, // page table entry
input logic [1:0] PageTypeWriteVal, // page type
input logic TLBWrite, // write TLB entry
input logic TLBFlush, // Invalidate all TLB entries
input logic clk, reset,
input logic [P.XLEN-1:0] SATP_REGW, // Current value of satp CSR (from privileged unit)
input logic STATUS_MXR, // Status CSR: make executable page readable
input logic STATUS_SUM, // Status CSR: Supervisor access to user memory
input logic STATUS_MPRV, // Status CSR: modify machine privilege
input logic [1:0] STATUS_MPP, // Status CSR: previous machine privilege level
input logic [1:0] PrivilegeModeW, // Current privilege level of the processeor
input logic DisableTranslation, // virtual address translation disabled during D$ flush and HPTW walk that use physical addresses
input logic [P.XLEN+1:0] VAdr, // virtual/physical address from IEU or physical address from HPTW
input logic [1:0] Size, // access size: 00 = 8 bits, 01 = 16 bits, 10 = 32 bits , 11 = 64 bits
input logic [P.XLEN-1:0] PTE, // page table entry
input logic [1:0] PageTypeWriteVal, // page type
input logic TLBWrite, // write TLB entry
input logic TLBFlush, // Invalidate all TLB entries
output logic [P.PA_BITS-1:0] PhysicalAddress, // PAdr when no translation, or translated VAdr (TLBPAdr) when there is translation
output logic TLBMiss, // Miss TLB
output logic Cacheable, // PMA indicates memory address is cachable
output logic Idempotent, // PMA indicates memory address is idempotent
output logic SelTIM, // Select a tightly integrated memory
output logic TLBMiss, // Miss TLB
output logic Cacheable, // PMA indicates memory address is cachable
output logic Idempotent, // PMA indicates memory address is idempotent
output logic SelTIM, // Select a tightly integrated memory
// Faults
output logic InstrAccessFaultF, LoadAccessFaultM, StoreAmoAccessFaultM, // access fault sources
output logic InstrPageFaultF, LoadPageFaultM, StoreAmoPageFaultM, // page fault sources
output logic UpdateDA, // page fault due to setting dirty or access bit
output logic LoadMisalignedFaultM, StoreAmoMisalignedFaultM, // misaligned fault sources
output logic InstrAccessFaultF, LoadAccessFaultM, StoreAmoAccessFaultM, // access fault sources
output logic InstrPageFaultF, LoadPageFaultM, StoreAmoPageFaultM, // page fault sources
output logic UpdateDA, // page fault due to setting dirty or access bit
output logic LoadMisalignedFaultM, StoreAmoMisalignedFaultM, // misaligned fault sources
// PMA checker signals
input logic AtomicAccessM, ExecuteAccessF, WriteAccessM, ReadAccessM, // access type
input var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0] // PMP addresses
input logic AtomicAccessM, ExecuteAccessF, WriteAccessM, ReadAccessM, // access type
input var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0] // PMP addresses
);
logic [P.PA_BITS-1:0] TLBPAdr; // physical address for TLB
logic PMAInstrAccessFaultF; // Instruction access fault from PMA
logic PMPInstrAccessFaultF; // Instruction access fault from PMP
logic PMALoadAccessFaultM; // Load access fault from PMA
logic PMPLoadAccessFaultM; // Load access fault from PMP
logic PMAStoreAmoAccessFaultM; // Store or AMO access fault from PMA
logic PMPStoreAmoAccessFaultM; // Store or AMO access fault from PMP
logic DataMisalignedM; // load or store misaligned
logic Translate; // Translation occurs when virtual memory is active and DisableTranslation is off
logic TLBHit; // Hit in TLB
logic TLBPageFault; // Page fault from TLB
logic ReadNoAmoAccessM; // Read that is not part of atomic operation causes Load faults. Otherwise StoreAmo faults
logic [P.PA_BITS-1:0] TLBPAdr; // physical address for TLB
logic PMAInstrAccessFaultF; // Instruction access fault from PMA
logic PMPInstrAccessFaultF; // Instruction access fault from PMP
logic PMALoadAccessFaultM; // Load access fault from PMA
logic PMPLoadAccessFaultM; // Load access fault from PMP
logic PMAStoreAmoAccessFaultM; // Store or AMO access fault from PMA
logic PMPStoreAmoAccessFaultM; // Store or AMO access fault from PMP
logic DataMisalignedM; // load or store misaligned
logic Translate; // Translation occurs when virtual memory is active and DisableTranslation is off
logic TLBHit; // Hit in TLB
logic TLBPageFault; // Page fault from TLB
logic ReadNoAmoAccessM; // Read that is not part of atomic operation causes Load faults. Otherwise StoreAmo faults
// only instantiate TLB if Virtual Memory is supported
if (P.VIRTMEM_SUPPORTED) begin:tlb
@ -86,9 +86,9 @@ module mmu import cvw::*; #(parameter cvw_t P,
.TLBWrite, .TLBFlush, .TLBPAdr, .TLBMiss, .TLBHit,
.Translate, .TLBPageFault, .UpdateDA);
end else begin:tlb // just pass address through as physical
assign Translate = 0;
assign TLBMiss = 0;
assign TLBHit = 1; // *** is this necessary
assign Translate = 0;
assign TLBMiss = 0;
assign TLBHit = 1; // *** is this necessary
assign TLBPageFault = 0;
end

37
src/mmu/pmachecker.sv
View File

@ -30,26 +30,26 @@
module pmachecker import cvw::*; #(parameter cvw_t P) (
input logic [P.PA_BITS-1:0] PhysicalAddress,
input logic [1:0] Size,
input logic AtomicAccessM, // Atomic access
input logic ExecuteAccessF, // Execute access
input logic WriteAccessM, // Write access
input logic ReadAccessM, // Read access
output logic Cacheable, Idempotent, SelTIM,
output logic PMAInstrAccessFaultF,
output logic PMALoadAccessFaultM,
output logic PMAStoreAmoAccessFaultM
input logic [1:0] Size,
input logic AtomicAccessM, // Atomic access
input logic ExecuteAccessF, // Execute access
input logic WriteAccessM, // Write access
input logic ReadAccessM, // Read access
output logic Cacheable, Idempotent, SelTIM,
output logic PMAInstrAccessFaultF,
output logic PMALoadAccessFaultM,
output logic PMAStoreAmoAccessFaultM
);
logic PMAAccessFault;
logic AccessRW, AccessRWX, AccessRX;
logic [10:0] SelRegions;
logic AtomicAllowed;
logic PMAAccessFault;
logic AccessRW, AccessRWX, AccessRX;
logic [10:0] SelRegions;
logic AtomicAllowed;
// Determine what type of access is being made
assign AccessRW = ReadAccessM | WriteAccessM;
assign AccessRW = ReadAccessM | WriteAccessM;
assign AccessRWX = ReadAccessM | WriteAccessM | ExecuteAccessF;
assign AccessRX = ReadAccessM | ExecuteAccessF;
assign AccessRX = ReadAccessM | ExecuteAccessF;
// Determine which region of physical memory (if any) is being accessed
adrdecs #(P) adrdecs(PhysicalAddress, AccessRW, AccessRX, AccessRWX, Size, SelRegions);
@ -65,9 +65,8 @@ module pmachecker import cvw::*; #(parameter cvw_t P) (
assign SelTIM = SelRegions[10] | SelRegions[9]; // exclusion-tag: unused-tim
// Detect access faults
assign PMAAccessFault = (SelRegions[0]) & AccessRWX | AtomicAccessM & ~AtomicAllowed;
assign PMAInstrAccessFaultF = ExecuteAccessF & PMAAccessFault;
assign PMALoadAccessFaultM = ReadAccessM & PMAAccessFault;
assign PMAAccessFault = (SelRegions[0]) & AccessRWX | AtomicAccessM & ~AtomicAllowed;
assign PMAInstrAccessFaultF = ExecuteAccessF & PMAAccessFault;
assign PMALoadAccessFaultM = ReadAccessM & PMAAccessFault;
assign PMAStoreAmoAccessFaultM = WriteAccessM & PMAAccessFault;
endmodule

1
src/mmu/pmpadrdec.sv
View File

@ -87,4 +87,3 @@ module pmpadrdec import cvw::*; #(parameter cvw_t P) (
// attempts an 8-byte access to 0x8, the access should fail (see page 60 of privileged specification 20211203). This
// implementation will not detect the failure.
endmodule

12
src/mmu/pmpchecker.sv
View File

@ -40,7 +40,7 @@ module pmpchecker import cvw::*; #(parameter cvw_t P) (
// keyword, the compiler warns us that it's interpreting the signal as a var,
// which we might not intend.
input var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0],
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW [P.PMP_ENTRIES-1:0],
input var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW [P.PMP_ENTRIES-1:0],
input logic ExecuteAccessF, WriteAccessM, ReadAccessM,
output logic PMPInstrAccessFaultF,
output logic PMPLoadAccessFaultM,
@ -48,11 +48,11 @@ module pmpchecker import cvw::*; #(parameter cvw_t P) (
);
// Bit i is high when the address falls in PMP region i
logic EnforcePMP; // should PMP be checked in this privilege level
logic [P.PMP_ENTRIES-1:0] Match; // physical address matches one of the pmp ranges
logic [P.PMP_ENTRIES-1:0] FirstMatch; // onehot encoding for the first pmpaddr to match the current address.
logic [P.PMP_ENTRIES-1:0] L, X, W, R; // PMP matches and has flag set
logic [P.PMP_ENTRIES-1:0] PAgePMPAdr; // for TOR PMP matching, PhysicalAddress > PMPAdr[i]
logic EnforcePMP; // should PMP be checked in this privilege level
logic [P.PMP_ENTRIES-1:0] Match; // physical address matches one of the pmp ranges
logic [P.PMP_ENTRIES-1:0] FirstMatch; // onehot encoding for the first pmpaddr to match the current address.
logic [P.PMP_ENTRIES-1:0] L, X, W, R; // PMP matches and has flag set
logic [P.PMP_ENTRIES-1:0] PAgePMPAdr; // for TOR PMP matching, PhysicalAddress > PMPAdr[i]
if (P.PMP_ENTRIES > 0) begin: pmp // prevent complaints about array of no elements when PMP_ENTRIES = 0
pmpadrdec #(P) pmpadrdecs[P.PMP_ENTRIES-1:0](

160
src/privileged/csr.sv
View File

@ -29,100 +29,101 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module csr import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic FlushM, FlushW,
input logic StallE, StallM, StallW,
input logic [31:0] InstrM, // current instruction
input logic [31:0] InstrOrigM, // Original compressed or uncompressed instruction in Memory stage for Illegal Instruction MTVAL
input logic [P.XLEN-1:0] PCM, PC2NextF, // program counter, next PC going to trap/return logic
input logic [P.XLEN-1:0] SrcAM, IEUAdrM, // SrcA and memory address from IEU
input logic CSRReadM, CSRWriteM, // read or write CSR
input logic TrapM, // trap is occurring
input logic mretM, sretM, wfiM, // return or WFI instruction
input logic IntPendingM, // at least one interrupt is pending and could occur if enabled
input logic InterruptM, // interrupt is occurring
input logic ExceptionM, // interrupt is occurring
input logic MTimerInt, // timer interrupt
input logic MExtInt, SExtInt, // external interrupt (from PLIC)
input logic MSwInt, // software interrupt
input logic [63:0] MTIME_CLINT, // TIME value from CLINT
input logic InstrValidM, // current instruction is valid
input logic FRegWriteM, // writes to floating point registers change STATUS.FS
input logic [4:0] SetFflagsM, // Set floating point flag bits in FCSR
input logic [1:0] NextPrivilegeModeM, // STATUS bits updated based on next privilege mode
input logic [1:0] PrivilegeModeW, // current privilege mode
input logic [3:0] CauseM, // Trap cause
input logic SelHPTW, // hardware page table walker active, so base endianness on supervisor mode
input logic clk, reset,
input logic FlushM, FlushW,
input logic StallE, StallM, StallW,
input logic [31:0] InstrM, // current instruction
input logic [31:0] InstrOrigM, // Original compressed or uncompressed instruction in Memory stage for Illegal Instruction MTVAL
input logic [P.XLEN-1:0] PCM, PC2NextF, // program counter, next PC going to trap/return logic
input logic [P.XLEN-1:0] SrcAM, IEUAdrM, // SrcA and memory address from IEU
input logic CSRReadM, CSRWriteM, // read or write CSR
input logic TrapM, // trap is occurring
input logic mretM, sretM, wfiM, // return or WFI instruction
input logic IntPendingM, // at least one interrupt is pending and could occur if enabled
input logic InterruptM, // interrupt is occurring
input logic ExceptionM, // interrupt is occurring
input logic MTimerInt, // timer interrupt
input logic MExtInt, SExtInt, // external interrupt (from PLIC)
input logic MSwInt, // software interrupt
input logic [63:0] MTIME_CLINT, // TIME value from CLINT
input logic InstrValidM, // current instruction is valid
input logic FRegWriteM, // writes to floating point registers change STATUS.FS
input logic [4:0] SetFflagsM, // Set floating point flag bits in FCSR
input logic [1:0] NextPrivilegeModeM, // STATUS bits updated based on next privilege mode
input logic [1:0] PrivilegeModeW, // current privilege mode
input logic [3:0] CauseM, // Trap cause
input logic SelHPTW, // hardware page table walker active, so base endianness on supervisor mode
// inputs for performance counters
input logic LoadStallD,
input logic StoreStallD,
input logic ICacheStallF,
input logic DCacheStallM,
input logic BPDirPredWrongM,
input logic BTAWrongM,
input logic RASPredPCWrongM,
input logic IClassWrongM,
input logic BPWrongM, // branch predictor is wrong
input logic [3:0] InstrClassM,
input logic DCacheMiss,
input logic DCacheAccess,
input logic ICacheMiss,
input logic ICacheAccess,
input logic sfencevmaM,
input logic InvalidateICacheM,
input logic DivBusyE, // integer divide busy
input logic FDivBusyE, // floating point divide busy
input logic LoadStallD,
input logic StoreStallD,
input logic ICacheStallF,
input logic DCacheStallM,
input logic BPDirPredWrongM,
input logic BTAWrongM,
input logic RASPredPCWrongM,
input logic IClassWrongM,
input logic BPWrongM, // branch predictor is wrong
input logic [3:0] InstrClassM,
input logic DCacheMiss,
input logic DCacheAccess,
input logic ICacheMiss,
input logic ICacheAccess,
input logic sfencevmaM,
input logic InvalidateICacheM,
input logic DivBusyE, // integer divide busy
input logic FDivBusyE, // floating point divide busy
// outputs from CSRs
output logic [1:0] STATUS_MPP,
output logic STATUS_SPP, STATUS_TSR, STATUS_TVM,
output logic [15:0] MEDELEG_REGW,
output logic [P.XLEN-1:0] SATP_REGW,
output logic [11:0] MIP_REGW, MIE_REGW, MIDELEG_REGW,
output logic STATUS_MIE, STATUS_SIE,
output logic STATUS_MXR, STATUS_SUM, STATUS_MPRV, STATUS_TW,
output logic [1:0] STATUS_FS,
output var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0],
output logic [1:0] STATUS_MPP,
output logic STATUS_SPP, STATUS_TSR, STATUS_TVM,
output logic [15:0] MEDELEG_REGW,
output logic [P.XLEN-1:0] SATP_REGW,
output logic [11:0] MIP_REGW, MIE_REGW, MIDELEG_REGW,
output logic STATUS_MIE, STATUS_SIE,
output logic STATUS_MXR, STATUS_SUM, STATUS_MPRV, STATUS_TW,
output logic [1:0] STATUS_FS,
output var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0],
output var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0],
output logic [2:0] FRM_REGW,
output logic [2:0] FRM_REGW,
//
output logic [P.XLEN-1:0] CSRReadValW, // value read from CSR
output logic [P.XLEN-1:0] UnalignedPCNextF, // Next PC, accounting for traps and returns
output logic IllegalCSRAccessM, // Illegal CSR access: CSR doesn't exist or is inaccessible at this privilege level
output logic BigEndianM // memory access is big-endian based on privilege mode and STATUS register endian fields
output logic [P.XLEN-1:0] CSRReadValW, // value read from CSR
output logic [P.XLEN-1:0] UnalignedPCNextF, // Next PC, accounting for traps and returns
output logic IllegalCSRAccessM, // Illegal CSR access: CSR doesn't exist or is inaccessible at this privilege level
output logic BigEndianM // memory access is big-endian based on privilege mode and STATUS register endian fields
);
localparam MIP = 12'h344;
localparam SIP = 12'h144;
logic [P.XLEN-1:0] CSRMReadValM, CSRSReadValM, CSRUReadValM, CSRCReadValM;
logic [P.XLEN-1:0] CSRReadValM;
logic [P.XLEN-1:0] CSRSrcM;
logic [P.XLEN-1:0] CSRRWM, CSRRSM, CSRRCM;
logic [P.XLEN-1:0] CSRWriteValM;
logic [P.XLEN-1:0] MSTATUS_REGW, SSTATUS_REGW, MSTATUSH_REGW;
logic [P.XLEN-1:0] STVEC_REGW, MTVEC_REGW;
logic [P.XLEN-1:0] MEPC_REGW, SEPC_REGW;
logic [P.XLEN-1:0] CSRMReadValM, CSRSReadValM, CSRUReadValM, CSRCReadValM;
logic [P.XLEN-1:0] CSRReadValM;
logic [P.XLEN-1:0] CSRSrcM;
logic [P.XLEN-1:0] CSRRWM, CSRRSM, CSRRCM;
logic [P.XLEN-1:0] CSRWriteValM;
logic [P.XLEN-1:0] MSTATUS_REGW, SSTATUS_REGW, MSTATUSH_REGW;
logic [P.XLEN-1:0] STVEC_REGW, MTVEC_REGW;
logic [P.XLEN-1:0] MEPC_REGW, SEPC_REGW;
logic [31:0] MCOUNTINHIBIT_REGW, MCOUNTEREN_REGW, SCOUNTEREN_REGW;
logic WriteMSTATUSM, WriteMSTATUSHM, WriteSSTATUSM;
logic CSRMWriteM, CSRSWriteM, CSRUWriteM;
logic UngatedCSRMWriteM;
logic WriteFRMM, WriteFFLAGSM;
logic [P.XLEN-1:0] UnalignedNextEPCM, NextEPCM, NextMtvalM;
logic [P.XLEN-1:0] UnalignedNextEPCM, NextEPCM, NextMtvalM;
logic [4:0] NextCauseM;
logic [11:0] CSRAdrM;
logic IllegalCSRCAccessM, IllegalCSRMAccessM, IllegalCSRSAccessM, IllegalCSRUAccessM;
logic InsufficientCSRPrivilegeM;
logic IllegalCSRMWriteReadonlyM;
logic [P.XLEN-1:0] CSRReadVal2M;
logic [P.XLEN-1:0] CSRReadVal2M;
logic [11:0] MIP_REGW_writeable;
logic [P.XLEN-1:0] TVecM, TrapVectorM, NextFaultMtvalM;
logic [P.XLEN-1:0] TVecM, TrapVectorM, NextFaultMtvalM;
logic MTrapM, STrapM;
logic [P.XLEN-1:0] EPC;
logic [P.XLEN-1:0] EPC;
logic RetM;
logic SelMtvecM;
logic [P.XLEN-1:0] TVecAlignedM;
logic [P.XLEN-1:0] TVecAlignedM;
logic InstrValidNotFlushedM;
logic STimerInt;
logic MENVCFG_STCE;
// only valid unflushed instructions can access CSRs
assign InstrValidNotFlushedM = InstrValidM & ~StallW & ~FlushW;
@ -179,13 +180,13 @@ module csr import cvw::*; #(parameter cvw_t P) (
else CSRReadVal2M = CSRReadValM;
// Compute AND/OR modification
CSRRWM = CSRSrcM;
CSRRSM = CSRReadVal2M | CSRSrcM;
CSRRCM = CSRReadVal2M & ~CSRSrcM;
CSRRWM = CSRSrcM;
CSRRSM = CSRReadVal2M | CSRSrcM;
CSRRCM = CSRReadVal2M & ~CSRSrcM;
case (InstrM[13:12])
2'b01: CSRWriteValM = CSRRWM;
2'b10: CSRWriteValM = CSRRSM;
2'b11: CSRWriteValM = CSRRCM;
2'b01: CSRWriteValM = CSRRWM;
2'b10: CSRWriteValM = CSRRSM;
2'b11: CSRWriteValM = CSRRCM;
default: CSRWriteValM = CSRReadValM;
endcase
end
@ -213,7 +214,7 @@ module csr import cvw::*; #(parameter cvw_t P) (
csri #(P) csri(.clk, .reset,
.CSRMWriteM, .CSRSWriteM, .CSRWriteValM, .CSRAdrM,
.MExtInt, .SExtInt, .MTimerInt, .STimerInt, .MSwInt,
.MIDELEG_REGW, .MIP_REGW, .MIE_REGW, .MIP_REGW_writeable);
.MIDELEG_REGW, .MENVCFG_STCE, .MIP_REGW, .MIE_REGW, .MIP_REGW_writeable);
csrsr #(P) csrsr(.clk, .reset, .StallW,
.WriteMSTATUSM, .WriteMSTATUSHM, .WriteSSTATUSM,
@ -231,7 +232,8 @@ module csr import cvw::*; #(parameter cvw_t P) (
.MEPC_REGW, .MCOUNTEREN_REGW, .MCOUNTINHIBIT_REGW,
.MEDELEG_REGW, .MIDELEG_REGW,.PMPCFG_ARRAY_REGW, .PMPADDR_ARRAY_REGW,
.MIP_REGW, .MIE_REGW, .WriteMSTATUSM, .WriteMSTATUSHM,
.IllegalCSRMAccessM, .IllegalCSRMWriteReadonlyM);
.IllegalCSRMAccessM, .IllegalCSRMWriteReadonlyM,
.MENVCFG_STCE);
if (P.S_SUPPORTED) begin:csrs
@ -242,7 +244,7 @@ module csr import cvw::*; #(parameter cvw_t P) (
.CSRWriteValM, .PrivilegeModeW,
.CSRSReadValM, .STVEC_REGW, .SEPC_REGW,
.SCOUNTEREN_REGW,
.SATP_REGW, .MIP_REGW, .MIE_REGW, .MIDELEG_REGW, .MTIME_CLINT,
.SATP_REGW, .MIP_REGW, .MIE_REGW, .MIDELEG_REGW, .MTIME_CLINT, .MENVCFG_STCE,
.WriteSSTATUSM, .IllegalCSRSAccessM, .STimerInt);
end else begin
assign WriteSSTATUSM = 0;
@ -266,7 +268,7 @@ module csr import cvw::*; #(parameter cvw_t P) (
assign IllegalCSRUAccessM = 1;
end
if (P.ZICOUNTERS_SUPPORTED) begin:counters
if (P.ZICNTR_SUPPORTED) begin:counters
csrc #(P) counters(.clk, .reset, .StallE, .StallM, .FlushM,
.InstrValidNotFlushedM, .LoadStallD, .StoreStallD, .CSRWriteM, .CSRMWriteM,
.BPDirPredWrongM, .BTAWrongM, .RASPredPCWrongM, .IClassWrongM, .BPWrongM,

146
src/privileged/csrc.sv
View File

@ -1,5 +1,3 @@
///////////////////////////////////////////
// csrc.sv
//
@ -31,57 +29,57 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module csrc import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic StallE, StallM,
input logic FlushM,
input logic InstrValidNotFlushedM, LoadStallD, StoreStallD,
input logic CSRMWriteM, CSRWriteM,
input logic BPDirPredWrongM,
input logic BTAWrongM,
input logic RASPredPCWrongM,
input logic IClassWrongM,
input logic BPWrongM, // branch predictor is wrong
input logic [3:0] InstrClassM,
input logic DCacheMiss,
input logic DCacheAccess,
input logic ICacheMiss,
input logic ICacheAccess,
input logic ICacheStallF,
input logic DCacheStallM,
input logic sfencevmaM,
input logic InterruptM,
input logic ExceptionM,
input logic InvalidateICacheM,
input logic DivBusyE, // integer divide busy
input logic FDivBusyE, // floating point divide busy
input logic [11:0] CSRAdrM,
input logic [1:0] PrivilegeModeW,
input logic clk, reset,
input logic StallE, StallM,
input logic FlushM,
input logic InstrValidNotFlushedM, LoadStallD, StoreStallD,
input logic CSRMWriteM, CSRWriteM,
input logic BPDirPredWrongM,
input logic BTAWrongM,
input logic RASPredPCWrongM,
input logic IClassWrongM,
input logic BPWrongM, // branch predictor is wrong
input logic [3:0] InstrClassM,
input logic DCacheMiss,
input logic DCacheAccess,
input logic ICacheMiss,
input logic ICacheAccess,
input logic ICacheStallF,
input logic DCacheStallM,
input logic sfencevmaM,
input logic InterruptM,
input logic ExceptionM,
input logic InvalidateICacheM,
input logic DivBusyE, // integer divide busy
input logic FDivBusyE, // floating point divide busy
input logic [11:0] CSRAdrM,
input logic [1:0] PrivilegeModeW,
input logic [P.XLEN-1:0] CSRWriteValM,
input logic [31:0] MCOUNTINHIBIT_REGW, MCOUNTEREN_REGW, SCOUNTEREN_REGW,
input logic [63:0] MTIME_CLINT,
input logic [31:0] MCOUNTINHIBIT_REGW, MCOUNTEREN_REGW, SCOUNTEREN_REGW,
input logic [63:0] MTIME_CLINT,
output logic [P.XLEN-1:0] CSRCReadValM,
output logic IllegalCSRCAccessM
output logic IllegalCSRCAccessM
);
localparam MHPMCOUNTERBASE = 12'hB00;
localparam MTIME = 12'hB01; // this is a memory-mapped register; no such CSR exists, and access should faul;
localparam MHPMCOUNTERBASE = 12'hB00;
localparam MTIME = 12'hB01; // this is a memory-mapped register; no such CSR exists, and access should faul;
localparam MHPMCOUNTERHBASE = 12'hB80;
localparam MTIMEH = 12'hB81; // this is a memory-mapped register; no such CSR exists, and access should fault
localparam MHPMEVENTBASE = 12'h320;
localparam HPMCOUNTERBASE = 12'hC00;
localparam HPMCOUNTERHBASE = 12'hC80;
localparam TIME = 12'hC01;
localparam TIMEH = 12'hC81;
localparam MTIMEH = 12'hB81; // this is a memory-mapped register; no such CSR exists, and access should fault
localparam MHPMEVENTBASE = 12'h320;
localparam HPMCOUNTERBASE = 12'hC00;
localparam HPMCOUNTERHBASE = 12'hC80;
localparam TIME = 12'hC01;
localparam TIMEH = 12'hC81;
logic [4:0] CounterNumM;
logic [P.XLEN-1:0] HPMCOUNTER_REGW[P.COUNTERS-1:0];
logic [P.XLEN-1:0] HPMCOUNTERH_REGW[P.COUNTERS-1:0];
logic [P.XLEN-1:0] HPMCOUNTER_REGW[P.COUNTERS-1:0];
logic [P.XLEN-1:0] HPMCOUNTERH_REGW[P.COUNTERS-1:0];
logic LoadStallE, LoadStallM;
logic StoreStallE, StoreStallM;
logic [P.COUNTERS-1:0] WriteHPMCOUNTERM;
logic [P.COUNTERS-1:0] CounterEvent;
logic [P.COUNTERS-1:0] WriteHPMCOUNTERM;
logic [P.COUNTERS-1:0] CounterEvent;
logic [63:0] HPMCOUNTERPlusM[P.COUNTERS-1:0];
logic [P.XLEN-1:0] NextHPMCOUNTERM[P.COUNTERS-1:0];
logic [P.XLEN-1:0] NextHPMCOUNTERM[P.COUNTERS-1:0];
genvar i;
// Interface signals
@ -89,38 +87,38 @@ module csrc import cvw::*; #(parameter cvw_t P) (
flopenrc #(2) LoadStallMReg(.clk, .reset, .clear(FlushM), .en(~StallM), .d({StoreStallE, LoadStallE}), .q({StoreStallM, LoadStallM}));
// Determine when to increment each counter
assign CounterEvent[0] = 1'b1; // MCYCLE always increments
assign CounterEvent[1] = 1'b0; // Counter 1 doesn't exist
assign CounterEvent[2] = InstrValidNotFlushedM; // MINSTRET instructions retired
if(P.QEMU) begin: cevent // No other performance counters in QEMU
assign CounterEvent[P.COUNTERS-1:3] = 0;
end else begin: cevent // User-defined counters
assign CounterEvent[3] = InstrClassM[0] & InstrValidNotFlushedM; // branch instruction
assign CounterEvent[4] = InstrClassM[1] & ~InstrClassM[2] & InstrValidNotFlushedM; // jump and not return instructions
assign CounterEvent[5] = InstrClassM[2] & InstrValidNotFlushedM; // return instructions
assign CounterEvent[6] = BPWrongM & InstrValidNotFlushedM; // branch predictor wrong
assign CounterEvent[7] = BPDirPredWrongM & InstrValidNotFlushedM; // Branch predictor wrong direction
assign CounterEvent[8] = BTAWrongM & InstrValidNotFlushedM; // branch predictor wrong target
assign CounterEvent[9] = RASPredPCWrongM & InstrValidNotFlushedM; // return address stack wrong address
assign CounterEvent[10] = IClassWrongM & InstrValidNotFlushedM; // instruction class predictor wrong
assign CounterEvent[11] = LoadStallM; // Load Stalls. don't want to suppress on flush as this only happens if flushed.
assign CounterEvent[12] = StoreStallM; // Store Stall
assign CounterEvent[13] = DCacheAccess & InstrValidNotFlushedM; // data cache access
assign CounterEvent[14] = DCacheMiss; // data cache miss. Miss asserted 1 cycle at start of cache miss
assign CounterEvent[15] = DCacheStallM; // d cache miss cycles
assign CounterEvent[16] = ICacheAccess & InstrValidNotFlushedM; // instruction cache access
assign CounterEvent[17] = ICacheMiss; // instruction cache miss. Miss asserted 1 cycle at start of cache miss
assign CounterEvent[18] = ICacheStallF; // i cache miss cycles
assign CounterEvent[19] = CSRWriteM & InstrValidNotFlushedM; // CSR writes
assign CounterEvent[20] = InvalidateICacheM & InstrValidNotFlushedM; // fence.i
assign CounterEvent[21] = sfencevmaM & InstrValidNotFlushedM; // sfence.vma
assign CounterEvent[22] = InterruptM; // interrupt, InstrValidNotFlushedM will be low
assign CounterEvent[23] = ExceptionM; // exceptions, InstrValidNotFlushedM will be low
assign CounterEvent[0] = 1'b1; // MCYCLE always increments
assign CounterEvent[1] = 1'b0; // Counter 1 doesn't exist
assign CounterEvent[2] = InstrValidNotFlushedM; // MINSTRET instructions retired
if (P.ZIHPM_SUPPORTED) begin: cevent // User-defined counters
assign CounterEvent[3] = InstrClassM[0] & InstrValidNotFlushedM; // branch instruction
assign CounterEvent[4] = InstrClassM[1] & ~InstrClassM[2] & InstrValidNotFlushedM; // jump and not return instructions
assign CounterEvent[5] = InstrClassM[2] & InstrValidNotFlushedM; // return instructions
assign CounterEvent[6] = BPWrongM & InstrValidNotFlushedM; // branch predictor wrong
assign CounterEvent[7] = BPDirPredWrongM & InstrValidNotFlushedM; // Branch predictor wrong direction
assign CounterEvent[8] = BTAWrongM & InstrValidNotFlushedM; // branch predictor wrong target
assign CounterEvent[9] = RASPredPCWrongM & InstrValidNotFlushedM; // return address stack wrong address
assign CounterEvent[10] = IClassWrongM & InstrValidNotFlushedM; // instruction class predictor wrong
assign CounterEvent[11] = LoadStallM; // Load Stalls. don't want to suppress on flush as this only happens if flushed.
assign CounterEvent[12] = StoreStallM; // Store Stall
assign CounterEvent[13] = DCacheAccess & InstrValidNotFlushedM; // data cache access
assign CounterEvent[14] = DCacheMiss; // data cache miss. Miss asserted 1 cycle at start of cache miss
assign CounterEvent[15] = DCacheStallM; // d cache miss cycles
assign CounterEvent[16] = ICacheAccess & InstrValidNotFlushedM; // instruction cache access
assign CounterEvent[17] = ICacheMiss; // instruction cache miss. Miss asserted 1 cycle at start of cache miss
assign CounterEvent[18] = ICacheStallF; // i cache miss cycles
assign CounterEvent[19] = CSRWriteM & InstrValidNotFlushedM; // CSR writes
assign CounterEvent[20] = InvalidateICacheM & InstrValidNotFlushedM; // fence.i
assign CounterEvent[21] = sfencevmaM & InstrValidNotFlushedM; // sfence.vma
assign CounterEvent[22] = InterruptM; // interrupt, InstrValidNotFlushedM will be low
assign CounterEvent[23] = ExceptionM; // exceptions, InstrValidNotFlushedM will be low
// coverage off
// DivBusyE will never be assert high since this configuration uses the FPU to do integer division
assign CounterEvent[24] = DivBusyE | FDivBusyE; // division cycles *** RT: might need to be delay until the next cycle
assign CounterEvent[24] = DivBusyE | FDivBusyE; // division cycles *** RT: might need to be delay until the next cycle
// coverage on
assign CounterEvent[P.COUNTERS-1:25] = 0; // eventually give these sources, including FP instructions, I$/D$ misses, branches and mispredictions
end else begin: cevent
assign CounterEvent[P.COUNTERS-1:3] = 0;
end
// Counter update and write logic
@ -158,14 +156,14 @@ module csrc import cvw::*; #(parameter cvw_t P) (
/* verilator lint_on WIDTH */
else if (CSRAdrM >= MHPMCOUNTERBASE & CSRAdrM < MHPMCOUNTERBASE+P.COUNTERS & CSRAdrM != MTIME)
CSRCReadValM = HPMCOUNTER_REGW[CounterNumM];
else if (CSRAdrM >= HPMCOUNTERBASE & CSRAdrM < HPMCOUNTERBASE+P.COUNTERS)
else if (CSRAdrM >= HPMCOUNTERBASE & CSRAdrM < HPMCOUNTERBASE+P.COUNTERS)
CSRCReadValM = HPMCOUNTER_REGW[CounterNumM];
else begin
CSRCReadValM = 0;
IllegalCSRCAccessM = 1; // requested CSR doesn't exist
end
end else begin // 32-bit counter reads
// Veri lator doesn't realize this only occurs for XLEN=32
// Veril ator doesn't realize this only occurs for XLEN=32
/* verilator lint_off WIDTH */
if (CSRAdrM == TIME) CSRCReadValM = MTIME_CLINT[31:0];// TIME register is a shadow of the memory-mapped MTIME from the CLINT
else if (CSRAdrM == TIMEH) CSRCReadValM = MTIME_CLINT[63:32];
@ -189,4 +187,4 @@ module csrc import cvw::*; #(parameter cvw_t P) (
end
endmodule
// mounteren should only exist if u-mode exists
// mounteren should only exist if u-mode exists

6
src/privileged/csri.sv
View File

@ -30,10 +30,11 @@
module csri import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic CSRMWriteM, CSRSWriteM,
input logic [P.XLEN-1:0] CSRWriteValM,
input logic [P.XLEN-1:0] CSRWriteValM,
input logic [11:0] CSRAdrM,
input logic MExtInt, SExtInt, MTimerInt, STimerInt, MSwInt,
input logic [11:0] MIDELEG_REGW,
input logic MENVCFG_STCE,
output logic [11:0] MIP_REGW, MIE_REGW,
output logic [11:0] MIP_REGW_writeable // only SEIP, STIP, SSIP are actually writeable; the rest are hardwired to 0
);
@ -60,7 +61,7 @@ module csri import cvw::*; #(parameter cvw_t P) (
if (P.S_SUPPORTED) begin:mask
if (P.SSTC_SUPPORTED) begin
assign MIP_WRITE_MASK = 12'h202; // SEIP and SSIP are writable, but STIP is not writable when STIMECMP is implemented (see SSTC spec)
assign STIP = STimerInt;
assign STIP = MENVCFG_STCE ? STimerInt : MIP_REGW_writeable[5];
end else begin
assign MIP_WRITE_MASK = 12'h222; // SEIP, STIP, SSIP are writeable in MIP (20210108-draft 3.1.9)
assign STIP = MIP_REGW_writeable[5];
@ -81,7 +82,6 @@ module csri import cvw::*; #(parameter cvw_t P) (
else if (WriteMIEM) MIE_REGW <= (CSRWriteValM[11:0] & MIE_WRITE_MASK); // MIE controls M and S fields
else if (WriteSIEM) MIE_REGW <= (CSRWriteValM[11:0] & 12'h222 & MIDELEG_REGW) | (MIE_REGW & 12'h888); // only S fields
assign MIP_REGW = {MExtInt, 1'b0, SExtInt|MIP_REGW_writeable[9], 1'b0,
MTimerInt, 1'b0, STIP, 1'b0,
MSwInt, 1'b0, MIP_REGW_writeable[1], 1'b0};

195
src/privileged/csrm.sv
View File

@ -32,67 +32,74 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module csrm import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic UngatedCSRMWriteM, CSRMWriteM, MTrapM,
input logic [11:0] CSRAdrM,
input logic clk, reset,
input logic UngatedCSRMWriteM, CSRMWriteM, MTrapM,
input logic [11:0] CSRAdrM,
input logic [P.XLEN-1:0] NextEPCM, NextMtvalM, MSTATUS_REGW, MSTATUSH_REGW,
input logic [4:0] NextCauseM,
input logic [4:0] NextCauseM,
input logic [P.XLEN-1:0] CSRWriteValM,
input logic [11:0] MIP_REGW, MIE_REGW,
input logic [11:0] MIP_REGW, MIE_REGW,
output logic [P.XLEN-1:0] CSRMReadValM, MTVEC_REGW,
output logic [P.XLEN-1:0] MEPC_REGW,
output logic [31:0] MCOUNTEREN_REGW, MCOUNTINHIBIT_REGW,
output logic [15:0] MEDELEG_REGW,
output logic [11:0] MIDELEG_REGW,
output var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0],
output logic [31:0] MCOUNTEREN_REGW, MCOUNTINHIBIT_REGW,
output logic [15:0] MEDELEG_REGW,
output logic [11:0] MIDELEG_REGW,
output var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0],
output var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW [P.PMP_ENTRIES-1:0],
output logic WriteMSTATUSM, WriteMSTATUSHM,
output logic IllegalCSRMAccessM, IllegalCSRMWriteReadonlyM
output logic WriteMSTATUSM, WriteMSTATUSHM,
output logic IllegalCSRMAccessM, IllegalCSRMWriteReadonlyM,
output logic MENVCFG_STCE
);
logic [P.XLEN-1:0] MISA_REGW, MHARTID_REGW;
logic [P.XLEN-1:0] MSCRATCH_REGW, MTVAL_REGW, MCAUSE_REGW;
logic WriteMTVECM, WriteMEDELEGM, WriteMIDELEGM;
logic WriteMSCRATCHM, WriteMEPCM, WriteMCAUSEM, WriteMTVALM;
logic WriteMCOUNTERENM, WriteMCOUNTINHIBITM;
logic [63:0] MENVCFG_REGW;
logic [P.XLEN-1:0] MENVCFGH_REGW;
logic [63:0] MENVCFG_PreWriteValM, MENVCFG_WriteValM;
logic WriteMTVECM, WriteMEDELEGM, WriteMIDELEGM;
logic WriteMSCRATCHM, WriteMEPCM, WriteMCAUSEM, WriteMTVALM;
logic WriteMCOUNTERENM, WriteMCOUNTINHIBITM;
logic WriteMENVCFGM;
// Machine CSRs
localparam MVENDORID = 12'hF11;
localparam MARCHID = 12'hF12;
localparam MIMPID = 12'hF13;
localparam MHARTID = 12'hF14;
localparam MCONFIGPTR = 12'hF15;
localparam MSTATUS = 12'h300;
localparam MISA_ADR = 12'h301;
localparam MEDELEG = 12'h302;
localparam MIDELEG = 12'h303;
localparam MIE = 12'h304;
localparam MTVEC = 12'h305;
localparam MCOUNTEREN = 12'h306;
localparam MSTATUSH = 12'h310;
localparam MVENDORID = 12'hF11;
localparam MARCHID = 12'hF12;
localparam MIMPID = 12'hF13;
localparam MHARTID = 12'hF14;
localparam MCONFIGPTR = 12'hF15;
localparam MSTATUS = 12'h300;
localparam MISA_ADR = 12'h301;
localparam MEDELEG = 12'h302;
localparam MIDELEG = 12'h303;
localparam MIE = 12'h304;
localparam MTVEC = 12'h305;
localparam MCOUNTEREN = 12'h306;
localparam MENVCFG = 12'h30A;
localparam MSTATUSH = 12'h310;
localparam MENVCFGH = 12'h31A;
localparam MCOUNTINHIBIT = 12'h320;
localparam MSCRATCH = 12'h340;
localparam MEPC = 12'h341;
localparam MCAUSE = 12'h342;
localparam MTVAL = 12'h343;
localparam MIP = 12'h344;
localparam MTINST = 12'h34A;
localparam PMPCFG0 = 12'h3A0;
localparam MSCRATCH = 12'h340;
localparam MEPC = 12'h341;
localparam MCAUSE = 12'h342;
localparam MTVAL = 12'h343;
localparam MIP = 12'h344;
localparam MTINST = 12'h34A;
localparam PMPCFG0 = 12'h3A0;
// .. up to 15 more at consecutive addresses
localparam PMPADDR0 = 12'h3B0;
localparam PMPADDR0 = 12'h3B0;
// ... up to 63 more at consecutive addresses
localparam TSELECT = 12'h7A0;
localparam TDATA1 = 12'h7A1;
localparam TDATA2 = 12'h7A2;
localparam TDATA3 = 12'h7A3;
localparam DCSR = 12'h7B0;
localparam DPC = 12'h7B1;
localparam DSCRATCH0 = 12'h7B2;
localparam DSCRATCH1 = 12'h7B3;
localparam TSELECT = 12'h7A0;
localparam TDATA1 = 12'h7A1;
localparam TDATA2 = 12'h7A2;
localparam TDATA3 = 12'h7A3;
localparam DCSR = 12'h7B0;
localparam DPC = 12'h7B1;
localparam DSCRATCH0 = 12'h7B2;
localparam DSCRATCH1 = 12'h7B3;
// Constants
localparam ZERO = {(P.XLEN){1'b0}};
localparam MEDELEG_MASK = 16'hB3FF;
localparam MIDELEG_MASK = 12'h222; // we choose to not make machine interrupts delegable
localparam MEDELEG_MASK = 16'hB3FF;
localparam MIDELEG_MASK = 12'h222; // we choose to not make machine interrupts delegable
// There are PMP_ENTRIES = 0, 16, or 64 PMPADDR registers, each of which has its own flop
genvar i;
@ -130,16 +137,17 @@ module csrm import cvw::*; #(parameter cvw_t P) (
assign MHARTID_REGW = 0;
// Write machine Mode CSRs
assign WriteMSTATUSM = CSRMWriteM & (CSRAdrM == MSTATUS);
assign WriteMSTATUSHM = CSRMWriteM & (CSRAdrM == MSTATUSH)& (P.XLEN==32);
assign WriteMTVECM = CSRMWriteM & (CSRAdrM == MTVEC);
assign WriteMEDELEGM = CSRMWriteM & (CSRAdrM == MEDELEG);
assign WriteMIDELEGM = CSRMWriteM & (CSRAdrM == MIDELEG);
assign WriteMSCRATCHM = CSRMWriteM & (CSRAdrM == MSCRATCH);
assign WriteMEPCM = MTrapM | (CSRMWriteM & (CSRAdrM == MEPC));
assign WriteMCAUSEM = MTrapM | (CSRMWriteM & (CSRAdrM == MCAUSE));
assign WriteMTVALM = MTrapM | (CSRMWriteM & (CSRAdrM == MTVAL));
assign WriteMCOUNTERENM = CSRMWriteM & (CSRAdrM == MCOUNTEREN);
assign WriteMSTATUSM = CSRMWriteM & (CSRAdrM == MSTATUS);
assign WriteMSTATUSHM = CSRMWriteM & (CSRAdrM == MSTATUSH) & (P.XLEN==32);
assign WriteMTVECM = CSRMWriteM & (CSRAdrM == MTVEC);
assign WriteMEDELEGM = CSRMWriteM & (CSRAdrM == MEDELEG);
assign WriteMIDELEGM = CSRMWriteM & (CSRAdrM == MIDELEG);
assign WriteMSCRATCHM = CSRMWriteM & (CSRAdrM == MSCRATCH);
assign WriteMEPCM = MTrapM | (CSRMWriteM & (CSRAdrM == MEPC));
assign WriteMCAUSEM = MTrapM | (CSRMWriteM & (CSRAdrM == MCAUSE));
assign WriteMTVALM = MTrapM | (CSRMWriteM & (CSRAdrM == MTVAL));
assign WriteMCOUNTERENM = CSRMWriteM & (CSRAdrM == MCOUNTEREN);
assign WriteMENVCFGM = CSRMWriteM & (CSRAdrM == MENVCFG);
assign WriteMCOUNTINHIBITM = CSRMWriteM & (CSRAdrM == MCOUNTINHIBIT);
assign IllegalCSRMWriteReadonlyM = UngatedCSRMWriteM & (CSRAdrM == MVENDORID | CSRAdrM == MARCHID | CSRAdrM == MIMPID | CSRAdrM == MHARTID);
@ -161,6 +169,39 @@ module csrm import cvw::*; #(parameter cvw_t P) (
flopenr #(32) MCOUNTERENreg(clk, reset, WriteMCOUNTERENM, CSRWriteValM[31:0], MCOUNTEREN_REGW);
end else assign MCOUNTEREN_REGW = '0;
// MENVCFG is always 64 bits even for RV32
assign MENVCFG_WriteValM = {
MENVCFG_PreWriteValM[63] & P.SSTC_SUPPORTED,
MENVCFG_PreWriteValM[62] & P.SVPBMT_SUPPORTED,
54'b0,
MENVCFG_PreWriteValM[7] & P.ZICBOZ_SUPPORTED,
MENVCFG_PreWriteValM[6:4] & {3{P.ZICBOM_SUPPORTED}},
3'b0,
MENVCFG_PreWriteValM[0] & P.S_SUPPORTED & P.VIRTMEM_SUPPORTED
};
if (P.XLEN == 64) begin
assign MENVCFG_PreWriteValM = CSRWriteValM;
flopenr #(P.XLEN) MENVCFGreg(clk, reset, WriteMENVCFGM, MENVCFG_WriteValM, MENVCFG_REGW);
assign MENVCFGH_REGW = 0;
end else begin
logic WriteMENVCFGHM;
assign MENVCFG_PreWriteValM = {CSRWriteValM, CSRWriteValM};
assign WriteMENVCFGHM = CSRMWriteM & (CSRAdrM == MENVCFGH) & (P.XLEN==32);
flopenr #(P.XLEN) MENVCFGreg(clk, reset, WriteMENVCFGM, MENVCFG_WriteValM[31:0], MENVCFG_REGW[31:0]);
flopenr #(P.XLEN) MENVCFGHreg(clk, reset, WriteMENVCFGHM, MENVCFG_WriteValM[63:32], MENVCFG_REGW[63:32]);
assign MENVCFGH_REGW = MENVCFG_REGW[63:32];
end
// Extract bit fields
assign MENVCFG_STCE = MENVCFG_REGW[63];
// Uncomment these other fields when they are defined
// assign MENVCFG_PBMTE = MENVCFG_REGW[62];
// assign MENVCFG_CBZE = MENVCFG_REGW[7];
// assign MENVCFG_CBCFE = MENVCFG_REGW[6];
// assign MENVCFG_CBIE = MENVCFG_REGW[5:4];
// assign MENVCFG_FIOM = MENVCFG_REGW[0];
// Read machine mode CSRs
// verilator lint_off WIDTH
logic [5:0] entry;
@ -181,26 +222,28 @@ module csrm import cvw::*; #(parameter cvw_t P) (
end
end
else case (CSRAdrM)
MISA_ADR: CSRMReadValM = MISA_REGW;
MVENDORID: CSRMReadValM = 0;
MARCHID: CSRMReadValM = 0;
MIMPID: CSRMReadValM = {{P.XLEN-12{1'b0}}, 12'h100}; // pipelined implementation
MHARTID: CSRMReadValM = MHARTID_REGW; // hardwired to 0
MCONFIGPTR: CSRMReadValM = 0; // hardwired to 0
MSTATUS: CSRMReadValM = MSTATUS_REGW;
MSTATUSH: CSRMReadValM = MSTATUSH_REGW;
MTVEC: CSRMReadValM = MTVEC_REGW;
MEDELEG: CSRMReadValM = {{(P.XLEN-16){1'b0}}, MEDELEG_REGW};
MIDELEG: CSRMReadValM = {{(P.XLEN-12){1'b0}}, MIDELEG_REGW};
MIP: CSRMReadValM = {{(P.XLEN-12){1'b0}}, MIP_REGW};
MIE: CSRMReadValM = {{(P.XLEN-12){1'b0}}, MIE_REGW};
MSCRATCH: CSRMReadValM = MSCRATCH_REGW;
MEPC: CSRMReadValM = MEPC_REGW;
MCAUSE: CSRMReadValM = MCAUSE_REGW;
MTVAL: CSRMReadValM = MTVAL_REGW;
MTINST: CSRMReadValM = 0; // implemented as trivial zero
MCOUNTEREN:CSRMReadValM = {{(P.XLEN-32){1'b0}}, MCOUNTEREN_REGW};
MCOUNTINHIBIT:CSRMReadValM = {{(P.XLEN-32){1'b0}}, MCOUNTINHIBIT_REGW};
MISA_ADR: CSRMReadValM = MISA_REGW;
MVENDORID: CSRMReadValM = 0;
MARCHID: CSRMReadValM = 0;
MIMPID: CSRMReadValM = {{P.XLEN-12{1'b0}}, 12'h100}; // pipelined implementation
MHARTID: CSRMReadValM = MHARTID_REGW; // hardwired to 0
MCONFIGPTR: CSRMReadValM = 0; // hardwired to 0
MSTATUS: CSRMReadValM = MSTATUS_REGW;
MSTATUSH: CSRMReadValM = MSTATUSH_REGW;
MTVEC: CSRMReadValM = MTVEC_REGW;
MEDELEG: CSRMReadValM = {{(P.XLEN-16){1'b0}}, MEDELEG_REGW};
MIDELEG: CSRMReadValM = {{(P.XLEN-12){1'b0}}, MIDELEG_REGW};
MIP: CSRMReadValM = {{(P.XLEN-12){1'b0}}, MIP_REGW};
MIE: CSRMReadValM = {{(P.XLEN-12){1'b0}}, MIE_REGW};
MSCRATCH: CSRMReadValM = MSCRATCH_REGW;
MEPC: CSRMReadValM = MEPC_REGW;
MCAUSE: CSRMReadValM = MCAUSE_REGW;
MTVAL: CSRMReadValM = MTVAL_REGW;
MTINST: CSRMReadValM = 0; // implemented as trivial zero
MCOUNTEREN: CSRMReadValM = {{(P.XLEN-32){1'b0}}, MCOUNTEREN_REGW};
MENVCFG: CSRMReadValM = MENVCFG_REGW[P.XLEN-1:0];
MENVCFGH: CSRMReadValM = MENVCFGH_REGW;
MCOUNTINHIBIT: CSRMReadValM = {{(P.XLEN-32){1'b0}}, MCOUNTINHIBIT_REGW};
default: begin
CSRMReadValM = 0;
@ -209,4 +252,4 @@ module csrm import cvw::*; #(parameter cvw_t P) (
endcase
end
// verilator lint_on WIDTH
endmodule
endmodule

107
src/privileged/csrs.sv
View File

@ -29,61 +29,69 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module csrs import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic CSRSWriteM, STrapM,
input logic [11:0] CSRAdrM,
input logic clk, reset,
input logic CSRSWriteM, STrapM,
input logic [11:0] CSRAdrM,
input logic [P.XLEN-1:0] NextEPCM, NextMtvalM, SSTATUS_REGW,
input logic [4:0] NextCauseM,
input logic STATUS_TVM,
input logic MCOUNTEREN_TM, // TM bit (1) of MCOUNTEREN; cause illegal instruction when trying to access STIMECMP if clear
input logic [4:0] NextCauseM,
input logic STATUS_TVM,
input logic MCOUNTEREN_TM, // TM bit (1) of MCOUNTEREN; cause illegal instruction when trying to access STIMECMP if clear
input logic [P.XLEN-1:0] CSRWriteValM,
input logic [1:0] PrivilegeModeW,
input logic [1:0] PrivilegeModeW,
output logic [P.XLEN-1:0] CSRSReadValM, STVEC_REGW,
output logic [P.XLEN-1:0] SEPC_REGW,
output logic [31:0] SCOUNTEREN_REGW,
output logic [31:0] SCOUNTEREN_REGW,
output logic [P.XLEN-1:0] SATP_REGW,
input logic [11:0] MIP_REGW, MIE_REGW, MIDELEG_REGW,
input logic [63:0] MTIME_CLINT,
output logic WriteSSTATUSM,
output logic IllegalCSRSAccessM,
output logic STimerInt
input logic [11:0] MIP_REGW, MIE_REGW, MIDELEG_REGW,
input logic [63:0] MTIME_CLINT,
input logic MENVCFG_STCE,
output logic WriteSSTATUSM,
output logic IllegalCSRSAccessM,
output logic STimerInt
);
// Supervisor CSRs
localparam SSTATUS = 12'h100;
localparam SIE = 12'h104;
localparam STVEC = 12'h105;
localparam SSTATUS = 12'h100;
localparam SIE = 12'h104;
localparam STVEC = 12'h105;
localparam SCOUNTEREN = 12'h106;
localparam SSCRATCH = 12'h140;
localparam SEPC = 12'h141;
localparam SCAUSE = 12'h142;
localparam STVAL = 12'h143;
localparam SIP= 12'h144;
localparam STIMECMP = 12'h14D;
localparam STIMECMPH = 12'h15D;
localparam SATP = 12'h180;
localparam SENVCFG = 12'h10A;
localparam SSCRATCH = 12'h140;
localparam SEPC = 12'h141;
localparam SCAUSE = 12'h142;
localparam STVAL = 12'h143;
localparam SIP = 12'h144;
localparam STIMECMP = 12'h14D;
localparam STIMECMPH = 12'h15D;
localparam SATP = 12'h180;
// Constants
localparam ZERO = {(P.XLEN){1'b0}};
localparam ZERO = {(P.XLEN){1'b0}};
localparam SEDELEG_MASK = ~(ZERO | {{P.XLEN-3{1'b0}}, 3'b111} << 9);
logic WriteSTVECM;
logic WriteSSCRATCHM, WriteSEPCM;
logic WriteSCAUSEM, WriteSTVALM, WriteSATPM, WriteSCOUNTERENM;
logic WriteSTIMECMPM, WriteSTIMECMPHM;
logic [P.XLEN-1:0] SSCRATCH_REGW, STVAL_REGW, SCAUSE_REGW;
logic WriteSENVCFGM;
logic [P.XLEN-1:0] SSCRATCH_REGW, STVAL_REGW, SCAUSE_REGW;
logic [P.XLEN-1:0] SENVCFG_REGW;
logic [P.XLEN-1:0] SENVCFG_WriteValM;
logic [63:0] STIMECMP_REGW;
// write enables
assign WriteSSTATUSM = CSRSWriteM & (CSRAdrM == SSTATUS);
assign WriteSTVECM = CSRSWriteM & (CSRAdrM == STVEC);
assign WriteSSCRATCHM = CSRSWriteM & (CSRAdrM == SSCRATCH);
assign WriteSEPCM = STrapM | (CSRSWriteM & (CSRAdrM == SEPC));
assign WriteSCAUSEM = STrapM | (CSRSWriteM & (CSRAdrM == SCAUSE));
assign WriteSTVALM = STrapM | (CSRSWriteM & (CSRAdrM == STVAL));
assign WriteSATPM = CSRSWriteM & (CSRAdrM == SATP) & (PrivilegeModeW == P.M_MODE | ~STATUS_TVM);
assign WriteSSTATUSM = CSRSWriteM & (CSRAdrM == SSTATUS);
assign WriteSTVECM = CSRSWriteM & (CSRAdrM == STVEC);
assign WriteSSCRATCHM = CSRSWriteM & (CSRAdrM == SSCRATCH);
assign WriteSEPCM = STrapM | (CSRSWriteM & (CSRAdrM == SEPC));
assign WriteSCAUSEM = STrapM | (CSRSWriteM & (CSRAdrM == SCAUSE));
assign WriteSTVALM = STrapM | (CSRSWriteM & (CSRAdrM == STVAL));
assign WriteSATPM = CSRSWriteM & (CSRAdrM == SATP) & (PrivilegeModeW == P.M_MODE | ~STATUS_TVM);
assign WriteSCOUNTERENM = CSRSWriteM & (CSRAdrM == SCOUNTEREN);
assign WriteSTIMECMPM = CSRSWriteM & (CSRAdrM == STIMECMP) & (PrivilegeModeW == P.M_MODE | MCOUNTEREN_TM);
assign WriteSTIMECMPHM = CSRSWriteM & (CSRAdrM == STIMECMPH) & (PrivilegeModeW == P.M_MODE | MCOUNTEREN_TM) & (P.XLEN == 32);
assign WriteSENVCFGM = CSRSWriteM & (CSRAdrM == SENVCFG);
assign WriteSTIMECMPM = CSRSWriteM & (CSRAdrM == STIMECMP) & (PrivilegeModeW == P.M_MODE | (MCOUNTEREN_TM & MENVCFG_STCE));
assign WriteSTIMECMPHM = CSRSWriteM & (CSRAdrM == STIMECMPH) & (PrivilegeModeW == P.M_MODE | (MCOUNTEREN_TM & MENVCFG_STCE)) & (P.XLEN == 32);
// CSRs
flopenr #(P.XLEN) STVECreg(clk, reset, WriteSTVECM, {CSRWriteValM[P.XLEN-1:2], 1'b0, CSRWriteValM[0]}, STVEC_REGW);
@ -108,9 +116,27 @@ module csrs import cvw::*; #(parameter cvw_t P) (
// Supervisor timer interrupt logic
// Spec is a bit peculiar - Machine timer interrupts are produced in CLINT, while Supervisor timer interrupts are in CSRs
if (P.SSTC_SUPPORTED)
assign STimerInt = ({1'b0, MTIME_CLINT} >= {1'b0, STIMECMP_REGW}); // unsigned comparison
assign STimerInt = ({1'b0, MTIME_CLINT} >= {1'b0, STIMECMP_REGW}); // unsigned comparison
else
assign STimerInt = 0;
assign SENVCFG_WriteValM = {
{(P.XLEN-8){1'b0}},
CSRWriteValM[7] & P.ZICBOZ_SUPPORTED,
CSRWriteValM[6:4] & {3{P.ZICBOM_SUPPORTED}},
3'b0,
CSRWriteValM[0] & P.S_SUPPORTED & P.VIRTMEM_SUPPORTED
};
flopenr #(P.XLEN) SENVCFGreg(clk, reset, WriteSENVCFGM, SENVCFG_WriteValM, SENVCFG_REGW);
// Extract bit fields
// Uncomment these other fields when they are defined
// assign SENVCFG_PBMTE = SENVCFG_REGW[62];
// assign SENVCFG_CBZE = SENVCFG_REGW[7];
// assign SENVCFG_CBCFE = SENVCFG_REGW[6];
// assign SENVCFG_CBIE = SENVCFG_REGW[5:4];
// assign SENVCFG_FIOM = SENVCFG_REGW[0];
// CSR Reads
always_comb begin:csrr
@ -130,12 +156,15 @@ module csrs import cvw::*; #(parameter cvw_t P) (
IllegalCSRSAccessM = 1;
end
SCOUNTEREN:CSRSReadValM = {{(P.XLEN-32){1'b0}}, SCOUNTEREN_REGW};
STIMECMP: if (P.SSTC_SUPPORTED & (PrivilegeModeW == P.M_MODE | MCOUNTEREN_TM)) CSRSReadValM = STIMECMP_REGW[P.XLEN-1:0];
SENVCFG: CSRSReadValM = SENVCFG_REGW;
STIMECMP: if (P.SSTC_SUPPORTED & (PrivilegeModeW == P.M_MODE | (MCOUNTEREN_TM && MENVCFG_STCE)))
CSRSReadValM = STIMECMP_REGW[P.XLEN-1:0];
else begin
CSRSReadValM = 0;
IllegalCSRSAccessM = 1;
end
STIMECMPH: if (P.SSTC_SUPPORTED & (P.XLEN == 32) & (PrivilegeModeW == P.M_MODE | MCOUNTEREN_TM)) CSRSReadValM[31:0] = STIMECMP_REGW[63:32];
STIMECMPH: if (P.SSTC_SUPPORTED & (P.XLEN == 32) & (PrivilegeModeW == P.M_MODE | (MCOUNTEREN_TM && MENVCFG_STCE)))
CSRSReadValM[31:0] = STIMECMP_REGW[63:32];
else begin // not supported for RV64
CSRSReadValM = 0;
IllegalCSRSAccessM = 1;
@ -146,4 +175,4 @@ module csrs import cvw::*; #(parameter cvw_t P) (
end
endcase
end
endmodule
endmodule

190
src/privileged/csrsr.sv
View File

@ -4,7 +4,7 @@
// Written: David_Harris@hmc.edu 9 January 2021
// Modified:
//
// Purpose: Status register
// Purpose: Status register (and environment configuration register and others shared across modes)
// See RISC-V Privileged Mode Specification 20190608
//
// Documentation: RISC-V System on Chip Design Chapter 5
@ -28,22 +28,22 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module csrsr import cvw::*; #(parameter cvw_t P) (
input logic clk, reset, StallW,
input logic WriteMSTATUSM, WriteMSTATUSHM, WriteSSTATUSM,
input logic TrapM, FRegWriteM,
input logic [1:0] NextPrivilegeModeM, PrivilegeModeW,
input logic mretM, sretM,
input logic WriteFRMM, WriteFFLAGSM,
input logic clk, reset, StallW,
input logic WriteMSTATUSM, WriteMSTATUSHM, WriteSSTATUSM,
input logic TrapM, FRegWriteM,
input logic [1:0] NextPrivilegeModeM, PrivilegeModeW,
input logic mretM, sretM,
input logic WriteFRMM, WriteFFLAGSM,
input logic [P.XLEN-1:0] CSRWriteValM,
input logic SelHPTW,
input logic SelHPTW,
output logic [P.XLEN-1:0] MSTATUS_REGW, SSTATUS_REGW, MSTATUSH_REGW,
output logic [1:0] STATUS_MPP,
output logic STATUS_SPP, STATUS_TSR, STATUS_TW,
output logic STATUS_MIE, STATUS_SIE,
output logic STATUS_MXR, STATUS_SUM,
output logic STATUS_MPRV, STATUS_TVM,
output logic [1:0] STATUS_FS,
output logic BigEndianM
output logic [1:0] STATUS_MPP,
output logic STATUS_SPP, STATUS_TSR, STATUS_TW,
output logic STATUS_MIE, STATUS_SIE,
output logic STATUS_MXR, STATUS_SUM,
output logic STATUS_MPRV, STATUS_TVM,
output logic [1:0] STATUS_FS,
output logic BigEndianM
);
logic STATUS_SD, STATUS_TW_INT, STATUS_TSR_INT, STATUS_TVM_INT, STATUS_MXR_INT, STATUS_SUM_INT, STATUS_MPRV_INT;
@ -56,27 +56,27 @@ module csrsr import cvw::*; #(parameter cvw_t P) (
// Lower privilege status registers are a subset of the full status register
// *** consider adding MBE, SBE, UBE fields, parameterized to be fixed or adjustable
if (P.XLEN==64) begin: csrsr64 // RV64
assign MSTATUS_REGW = {STATUS_SD, 25'b0, STATUS_MBE, STATUS_SBE, STATUS_SXL, STATUS_UXL, 9'b0,
STATUS_TSR, STATUS_TW, STATUS_TVM, STATUS_MXR, STATUS_SUM, STATUS_MPRV,
STATUS_XS, STATUS_FS, STATUS_MPP, 2'b0,
STATUS_SPP, STATUS_MPIE, STATUS_UBE, STATUS_SPIE, 1'b0,
STATUS_MIE, 1'b0, STATUS_SIE, 1'b0};
assign SSTATUS_REGW = {STATUS_SD, /*27'b0, */ 29'b0, /*STATUS_SXL, */ {P.QEMU ? 2'b0 : STATUS_UXL}, /*9'b0, */ 12'b0,
assign MSTATUS_REGW = {STATUS_SD, 25'b0, STATUS_MBE, STATUS_SBE, STATUS_SXL, STATUS_UXL, 9'b0,
STATUS_TSR, STATUS_TW, STATUS_TVM, STATUS_MXR, STATUS_SUM, STATUS_MPRV,
STATUS_XS, STATUS_FS, STATUS_MPP, 2'b0,
STATUS_SPP, STATUS_MPIE, STATUS_UBE, STATUS_SPIE, 1'b0,
STATUS_MIE, 1'b0, STATUS_SIE, 1'b0};
assign SSTATUS_REGW = {STATUS_SD, /*27'b0, */ 29'b0, /*STATUS_SXL, */ {P.QEMU ? 2'b0 : STATUS_UXL}, /*9'b0, */ 12'b0,
/*STATUS_TSR, STATUS_TW, STATUS_TVM, */STATUS_MXR, STATUS_SUM, /* STATUS_MPRV, */ 1'b0,
STATUS_XS, STATUS_FS, /*STATUS_MPP, 2'b0*/ 4'b0,
STATUS_SPP, /*STATUS_MPIE*/ 1'b0, STATUS_UBE, STATUS_SPIE,
STATUS_XS, STATUS_FS, /*STATUS_MPP, 2'b0*/ 4'b0,
STATUS_SPP, /*STATUS_MPIE*/ 1'b0, STATUS_UBE, STATUS_SPIE,
/*1'b0, STATUS_MIE, 1'b0*/ 3'b0, STATUS_SIE, 1'b0};
assign MSTATUSH_REGW = '0; // *** does not exist when XLEN=64, but don't want it to have an undefined value. Spec is not clear what it should be.
end else begin: csrsr32 // RV32
assign MSTATUS_REGW = {STATUS_SD, 8'b0,
STATUS_TSR, STATUS_TW, STATUS_TVM, STATUS_MXR, STATUS_SUM, STATUS_MPRV,
STATUS_XS, STATUS_FS, STATUS_MPP, 2'b0,
STATUS_SPP, STATUS_MPIE, STATUS_UBE, STATUS_SPIE, 1'b0, STATUS_MIE, 1'b0, STATUS_SIE, 1'b0};
assign MSTATUS_REGW = {STATUS_SD, 8'b0,
STATUS_TSR, STATUS_TW, STATUS_TVM, STATUS_MXR, STATUS_SUM, STATUS_MPRV,
STATUS_XS, STATUS_FS, STATUS_MPP, 2'b0,
STATUS_SPP, STATUS_MPIE, STATUS_UBE, STATUS_SPIE, 1'b0, STATUS_MIE, 1'b0, STATUS_SIE, 1'b0};
assign MSTATUSH_REGW = {26'b0, STATUS_MBE, STATUS_SBE, 4'b0};
assign SSTATUS_REGW = {STATUS_SD, 11'b0,
assign SSTATUS_REGW = {STATUS_SD, 11'b0,
/*STATUS_TSR, STATUS_TW, STATUS_TVM, */STATUS_MXR, STATUS_SUM, /* STATUS_MPRV, */ 1'b0,
STATUS_XS, STATUS_FS, /*STATUS_MPP, 2'b0*/ 4'b0,
STATUS_SPP, /*STATUS_MPIE*/ 1'b0, STATUS_UBE, STATUS_SPIE,
STATUS_XS, STATUS_FS, /*STATUS_MPP, 2'b0*/ 4'b0,
STATUS_SPP, /*STATUS_MPIE*/ 1'b0, STATUS_UBE, STATUS_SPIE,
/*1'b0, STATUS_MIE, 1'b0*/ 3'b0, STATUS_SIE, 1'b0};
end
@ -90,21 +90,21 @@ module csrsr import cvw::*; #(parameter cvw_t P) (
end
// harwired STATUS bits
assign STATUS_TSR = P.S_SUPPORTED & STATUS_TSR_INT; // override reigster with 0 if supervisor mode not supported
assign STATUS_TW = (P.S_SUPPORTED | P.U_SUPPORTED) & STATUS_TW_INT; // override register with 0 if only machine mode supported
assign STATUS_TVM = P.S_SUPPORTED & STATUS_TVM_INT; // override reigster with 0 if supervisor mode not supported
assign STATUS_MXR = P.S_SUPPORTED & STATUS_MXR_INT; // override reigster with 0 if supervisor mode not supported
/* assign STATUS_UBE = 0; // little-endian
assign STATUS_SBE = 0; // little-endian
assign STATUS_MBE = 0; // little-endian */
assign STATUS_TSR = P.S_SUPPORTED & STATUS_TSR_INT; // override reigster with 0 if supervisor mode not supported
assign STATUS_TW = (P.S_SUPPORTED | P.U_SUPPORTED) & STATUS_TW_INT; // override register with 0 if only machine mode supported
assign STATUS_TVM = P.S_SUPPORTED & STATUS_TVM_INT; // override reigster with 0 if supervisor mode not supported
assign STATUS_MXR = P.S_SUPPORTED & STATUS_MXR_INT; // override reigster with 0 if supervisor mode not supported
/* assign STATUS_UBE = 0; // little-endian
assign STATUS_SBE = 0; // little-endian
assign STATUS_MBE = 0; // little-endian */
// SXL and UXL bits only matter for RV64. Set to 10 for RV64 if mode is supported, or 0 if not
assign STATUS_SXL = P.S_SUPPORTED ? 2'b10 : 2'b00; // 10 if supervisor mode supported
assign STATUS_UXL = P.U_SUPPORTED ? 2'b10 : 2'b00; // 10 if user mode supported
assign STATUS_SUM = P.S_SUPPORTED & P.VIRTMEM_SUPPORTED & STATUS_SUM_INT; // override reigster with 0 if supervisor mode not supported
assign STATUS_SXL = P.S_SUPPORTED ? 2'b10 : 2'b00; // 10 if supervisor mode supported
assign STATUS_UXL = P.U_SUPPORTED ? 2'b10 : 2'b00; // 10 if user mode supported
assign STATUS_SUM = P.S_SUPPORTED & P.VIRTMEM_SUPPORTED & STATUS_SUM_INT; // override reigster with 0 if supervisor mode not supported
assign STATUS_MPRV = P.U_SUPPORTED & STATUS_MPRV_INT; // override with 0 if user mode not supported
assign STATUS_FS = (P.S_SUPPORTED & (P.F_SUPPORTED | P.D_SUPPORTED)) ? STATUS_FS_INT : 2'b00; // off if no FP
assign STATUS_SD = (STATUS_FS == 2'b11) | (STATUS_XS == 2'b11); // dirty state logic
assign STATUS_XS = 2'b00; // No additional user-mode state to be dirty
assign STATUS_FS = (P.S_SUPPORTED & (P.F_SUPPORTED | P.D_SUPPORTED)) ? STATUS_FS_INT : 2'b00; // off if no FP
assign STATUS_SD = (STATUS_FS == 2'b11) | (STATUS_XS == 2'b11); // dirty state logic
assign STATUS_XS = 2'b00; // No additional user-mode state to be dirty
always_comb
if (CSRWriteValM[12:11] == P.U_MODE & P.U_SUPPORTED) STATUS_MPP_NEXT = P.U_MODE;
@ -122,14 +122,14 @@ module csrsr import cvw::*; #(parameter cvw_t P) (
if (SelHPTW) EndiannessPrivMode = P.S_MODE;
//coverage off -item c 1 -feccondrow 1
// status.MPRV always gets reset upon leaving machine mode, so MPRV will never be high when out of machine mode
else if (PrivilegeModeW == P.M_MODE & STATUS_MPRV) EndiannessPrivMode = STATUS_MPP;
else if (PrivilegeModeW == P.M_MODE & STATUS_MPRV) EndiannessPrivMode = STATUS_MPP;
//coverage on
else EndiannessPrivMode = PrivilegeModeW;
case (EndiannessPrivMode)
P.M_MODE: BigEndianM = STATUS_MBE;
P.S_MODE: BigEndianM = STATUS_SBE;
default: BigEndianM = STATUS_UBE;
default: BigEndianM = STATUS_UBE;
endcase
end
end else begin: endianmux
@ -140,22 +140,22 @@ module csrsr import cvw::*; #(parameter cvw_t P) (
// complex register with reset, write enable, and the ability to update other bits in certain cases
always_ff @(posedge clk) //, posedge reset)
if (reset) begin
STATUS_TSR_INT <= #1 0;
STATUS_TW_INT <= #1 0;
STATUS_TVM_INT <= #1 0;
STATUS_MXR_INT <= #1 0;
STATUS_SUM_INT <= #1 0;
STATUS_TSR_INT <= #1 0;
STATUS_TW_INT <= #1 0;
STATUS_TVM_INT <= #1 0;
STATUS_MXR_INT <= #1 0;
STATUS_SUM_INT <= #1 0;
STATUS_MPRV_INT <= #1 0; // Per Priv 3.3
STATUS_FS_INT <= #1 P.F_SUPPORTED ? 2'b00 : 2'b00; // leave floating-point off until activated, even if F_SUPPORTED
STATUS_MPP <= #1 0;
STATUS_SPP <= #1 0;
STATUS_MPIE <= #1 0;
STATUS_SPIE <= #1 0;
STATUS_MIE <= #1 0;
STATUS_SIE <= #1 0;
STATUS_MBE <= #1 0;
STATUS_SBE <= #1 0;
STATUS_UBE <= #1 0;
STATUS_FS_INT <= #1 P.F_SUPPORTED ? 2'b00 : 2'b00; // leave floating-point off until activated, even if F_SUPPORTED
STATUS_MPP <= #1 0;
STATUS_SPP <= #1 0;
STATUS_MPIE <= #1 0;
STATUS_SPIE <= #1 0;
STATUS_MIE <= #1 0;
STATUS_SIE <= #1 0;
STATUS_MBE <= #1 0;
STATUS_SBE <= #1 0;
STATUS_UBE <= #1 0;
end else if (~StallW) begin
if (TrapM) begin
// Update interrupt enables per Privileged Spec p. 21
@ -164,54 +164,54 @@ module csrsr import cvw::*; #(parameter cvw_t P) (
// Modes: 11 = Machine, 01 = Supervisor, 00 = User
if (NextPrivilegeModeM == P.M_MODE) begin
STATUS_MPIE <= #1 STATUS_MIE;
STATUS_MIE <= #1 0;
STATUS_MPP <= #1 PrivilegeModeW;
STATUS_MIE <= #1 0;
STATUS_MPP <= #1 PrivilegeModeW;
end else begin // supervisor mode
STATUS_SPIE <= #1 STATUS_SIE;
STATUS_SIE <= #1 0;
STATUS_SPP <= #1 PrivilegeModeW[0];
STATUS_SIE <= #1 0;
STATUS_SPP <= #1 PrivilegeModeW[0];
end
end else if (mretM) begin // Privileged 3.1.6.1
STATUS_MIE <= #1 STATUS_MPIE; // restore global interrupt enable
STATUS_MPIE <= #1 1; //
STATUS_MPP <= #1 P.U_SUPPORTED ? P.U_MODE : P.M_MODE; // set MPP to lowest supported privilege level
STATUS_MPRV_INT <= #1 STATUS_MPRV_INT & (STATUS_MPP == P.M_MODE); // page 21 of privileged spec.
STATUS_MIE <= #1 STATUS_MPIE; // restore global interrupt enable
STATUS_MPIE <= #1 1; //
STATUS_MPP <= #1 P.U_SUPPORTED ? P.U_MODE : P.M_MODE; // set MPP to lowest supported privilege level
STATUS_MPRV_INT <= #1 STATUS_MPRV_INT & (STATUS_MPP == P.M_MODE); // page 21 of privileged spec.
end else if (sretM) begin
STATUS_SIE <= #1 STATUS_SPIE; // restore global interrupt enable
STATUS_SPIE <= #1 P.S_SUPPORTED;
STATUS_SPP <= #1 0; // set SPP to lowest supported privilege level to catch bugs
STATUS_SIE <= #1 STATUS_SPIE; // restore global interrupt enable
STATUS_SPIE <= #1 P.S_SUPPORTED;
STATUS_SPP <= #1 0; // set SPP to lowest supported privilege level to catch bugs
STATUS_MPRV_INT <= #1 0; // always clear MPRV
end else if (WriteMSTATUSM) begin
STATUS_TSR_INT <= #1 CSRWriteValM[22];
STATUS_TW_INT <= #1 CSRWriteValM[21];
STATUS_TVM_INT <= #1 CSRWriteValM[20];
STATUS_MXR_INT <= #1 CSRWriteValM[19];
STATUS_SUM_INT <= #1 CSRWriteValM[18];
STATUS_TSR_INT <= #1 CSRWriteValM[22];
STATUS_TW_INT <= #1 CSRWriteValM[21];
STATUS_TVM_INT <= #1 CSRWriteValM[20];
STATUS_MXR_INT <= #1 CSRWriteValM[19];
STATUS_SUM_INT <= #1 CSRWriteValM[18];
STATUS_MPRV_INT <= #1 CSRWriteValM[17];
STATUS_FS_INT <= #1 CSRWriteValM[14:13];
STATUS_MPP <= #1 STATUS_MPP_NEXT;
STATUS_SPP <= #1 P.S_SUPPORTED & CSRWriteValM[8];
STATUS_MPIE <= #1 CSRWriteValM[7];
STATUS_SPIE <= #1 P.S_SUPPORTED & CSRWriteValM[5];
STATUS_MIE <= #1 CSRWriteValM[3];
STATUS_SIE <= #1 P.S_SUPPORTED & CSRWriteValM[1];
STATUS_UBE <= #1 CSRWriteValM[6] & P.U_SUPPORTED & P.BIGENDIAN_SUPPORTED;
STATUS_MBE <= #1 nextMBE;
STATUS_SBE <= #1 nextSBE;
STATUS_FS_INT <= #1 CSRWriteValM[14:13];
STATUS_MPP <= #1 STATUS_MPP_NEXT;
STATUS_SPP <= #1 P.S_SUPPORTED & CSRWriteValM[8];
STATUS_MPIE <= #1 CSRWriteValM[7];
STATUS_SPIE <= #1 P.S_SUPPORTED & CSRWriteValM[5];
STATUS_MIE <= #1 CSRWriteValM[3];
STATUS_SIE <= #1 P.S_SUPPORTED & CSRWriteValM[1];
STATUS_UBE <= #1 CSRWriteValM[6] & P.U_SUPPORTED & P.BIGENDIAN_SUPPORTED;
STATUS_MBE <= #1 nextMBE;
STATUS_SBE <= #1 nextSBE;
// coverage off
// MSTATUSH only exists in 32-bit configurations, will not be hit on rv64gc
end else if (WriteMSTATUSHM) begin
STATUS_MBE <= #1 CSRWriteValM[5] & P.BIGENDIAN_SUPPORTED;
STATUS_SBE <= #1 CSRWriteValM[4] & P.S_SUPPORTED & P.BIGENDIAN_SUPPORTED;
STATUS_MBE <= #1 CSRWriteValM[5] & P.BIGENDIAN_SUPPORTED;
STATUS_SBE <= #1 CSRWriteValM[4] & P.S_SUPPORTED & P.BIGENDIAN_SUPPORTED;
// coverage on
end else if (WriteSSTATUSM) begin // write a subset of the STATUS bits
STATUS_MXR_INT <= #1 CSRWriteValM[19];
STATUS_SUM_INT <= #1 CSRWriteValM[18];
STATUS_FS_INT <= #1 CSRWriteValM[14:13];
STATUS_SPP <= #1 P.S_SUPPORTED & CSRWriteValM[8];
STATUS_SPIE <= #1 P.S_SUPPORTED & CSRWriteValM[5];
STATUS_SIE <= #1 P.S_SUPPORTED & CSRWriteValM[1];
STATUS_UBE <= #1 CSRWriteValM[6] & P.U_SUPPORTED & P.BIGENDIAN_SUPPORTED;
STATUS_MXR_INT <= #1 CSRWriteValM[19];
STATUS_SUM_INT <= #1 CSRWriteValM[18];
STATUS_FS_INT <= #1 CSRWriteValM[14:13];
STATUS_SPP <= #1 P.S_SUPPORTED & CSRWriteValM[8];
STATUS_SPIE <= #1 P.S_SUPPORTED & CSRWriteValM[5];
STATUS_SIE <= #1 P.S_SUPPORTED & CSRWriteValM[1];
STATUS_UBE <= #1 CSRWriteValM[6] & P.U_SUPPORTED & P.BIGENDIAN_SUPPORTED;
end else if (FRegWriteM | WriteFRMM | WriteFFLAGSM) STATUS_FS_INT <= #1 2'b11;
end
endmodule

38
src/privileged/csru.sv
View File

@ -27,40 +27,40 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module csru import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic InstrValidNotFlushedM,
input logic CSRUWriteM,
input logic [11:0] CSRAdrM,
input logic clk, reset,
input logic InstrValidNotFlushedM,
input logic CSRUWriteM,
input logic [11:0] CSRAdrM,
input logic [P.XLEN-1:0] CSRWriteValM,
input logic [1:0] STATUS_FS,
input logic [1:0] STATUS_FS,
output logic [P.XLEN-1:0] CSRUReadValM,
input logic [4:0] SetFflagsM,
output logic [2:0] FRM_REGW,
output logic WriteFRMM, WriteFFLAGSM,
output logic IllegalCSRUAccessM
input logic [4:0] SetFflagsM,
output logic [2:0] FRM_REGW,
output logic WriteFRMM, WriteFFLAGSM,
output logic IllegalCSRUAccessM
);
localparam FFLAGS = 12'h001;
localparam FRM = 12'h002;
localparam FCSR = 12'h003;
localparam FRM = 12'h002;
localparam FCSR = 12'h003;
logic [4:0] FFLAGS_REGW;
logic [2:0] NextFRMM;
logic [4:0] NextFFLAGSM;
logic SetOrWriteFFLAGSM;
logic [4:0] FFLAGS_REGW;
logic [2:0] NextFRMM;
logic [4:0] NextFFLAGSM;
logic SetOrWriteFFLAGSM;
// Write enables
assign WriteFRMM = CSRUWriteM & (STATUS_FS != 2'b00) & (CSRAdrM == FRM | CSRAdrM == FCSR);
assign WriteFRMM = CSRUWriteM & (STATUS_FS != 2'b00) & (CSRAdrM == FRM | CSRAdrM == FCSR);
assign WriteFFLAGSM = CSRUWriteM & (STATUS_FS != 2'b00) & (CSRAdrM == FFLAGS | CSRAdrM == FCSR);
// Write Values
assign NextFRMM = (CSRAdrM == FCSR) ? CSRWriteValM[7:5] : CSRWriteValM[2:0];
assign NextFFLAGSM = WriteFFLAGSM ? CSRWriteValM[4:0] : FFLAGS_REGW | SetFflagsM;
assign NextFRMM = (CSRAdrM == FCSR) ? CSRWriteValM[7:5] : CSRWriteValM[2:0];
assign NextFFLAGSM = WriteFFLAGSM ? CSRWriteValM[4:0] : FFLAGS_REGW | SetFflagsM;
assign SetOrWriteFFLAGSM = WriteFFLAGSM | (|SetFflagsM & InstrValidNotFlushedM);
// CSRs
flopenr #(3) FRMreg(clk, reset, WriteFRMM, NextFRMM, FRM_REGW);
flopenr #(5) FFLAGSreg(clk, reset, SetOrWriteFFLAGSM, NextFFLAGSM, FFLAGS_REGW);
flopenr #(5) FFLAGSreg(clk, reset, SetOrWriteFFLAGSM, NextFFLAGSM, FFLAGS_REGW);
// CSR Reads
always_comb begin

140
src/privileged/privileged.sv
View File

@ -28,91 +28,90 @@
///////////////////////////////////////////
module privileged import cvw::*; #(parameter cvw_t P) (
input logic clk, reset,
input logic StallD, StallE, StallM, StallW,
input logic FlushD, FlushE, FlushM, FlushW,
input logic clk, reset,
input logic StallD, StallE, StallM, StallW,
input logic FlushD, FlushE, FlushM, FlushW,
// CSR Reads and Writes, and values needed for traps
input logic CSRReadM, CSRWriteM, // Read or write CSRs
input logic CSRReadM, CSRWriteM, // Read or write CSRs
input logic [P.XLEN-1:0] SrcAM, // GPR register to write
input logic [31:0] InstrM, // Instruction
input logic [31:0] InstrOrigM, // Original compressed or uncompressed instruction in Memory stage for Illegal Instruction MTVAL
input logic [31:0] InstrM, // Instruction
input logic [31:0] InstrOrigM, // Original compressed or uncompressed instruction in Memory stage for Illegal Instruction MTVAL
input logic [P.XLEN-1:0] IEUAdrM, // address from IEU
input logic [P.XLEN-1:0] PCM, PC2NextF, // program counter, next PC going to trap/return PC logic
// control signals
input logic InstrValidM, // Current instruction is valid (not flushed)
input logic CommittedM, CommittedF, // current instruction is using bus; don't interrupt
input logic PrivilegedM, // privileged instruction
input logic InstrValidM, // Current instruction is valid (not flushed)
input logic CommittedM, CommittedF, // current instruction is using bus; don't interrupt
input logic PrivilegedM, // privileged instruction
// processor events for performance counter logging
input logic FRegWriteM, // instruction will write floating-point registers
input logic LoadStallD, // load instruction is stalling
input logic StoreStallD, // store instruction is stalling
input logic ICacheStallF, // I cache stalled
input logic DCacheStallM, // D cache stalled
input logic BPDirPredWrongM, // branch predictor guessed wrong direction
input logic BTAWrongM, // branch predictor guessed wrong target
input logic RASPredPCWrongM, // return adddress stack guessed wrong target
input logic IClassWrongM, // branch predictor guessed wrong instruction class
input logic BPWrongM, // branch predictor is wrong
input logic [3:0] InstrClassM, // actual instruction class
input logic DCacheMiss, // data cache miss
input logic DCacheAccess, // data cache accessed (hit or miss)
input logic ICacheMiss, // instruction cache miss
input logic ICacheAccess, // instruction cache access
input logic DivBusyE, // integer divide busy
input logic FDivBusyE, // floating point divide busy
input logic FRegWriteM, // instruction will write floating-point registers
input logic LoadStallD, // load instruction is stalling
input logic StoreStallD, // store instruction is stalling
input logic ICacheStallF, // I cache stalled
input logic DCacheStallM, // D cache stalled
input logic BPDirPredWrongM, // branch predictor guessed wrong direction
input logic BTAWrongM, // branch predictor guessed wrong target
input logic RASPredPCWrongM, // return adddress stack guessed wrong target
input logic IClassWrongM, // branch predictor guessed wrong instruction class
input logic BPWrongM, // branch predictor is wrong
input logic [3:0] InstrClassM, // actual instruction class
input logic DCacheMiss, // data cache miss
input logic DCacheAccess, // data cache accessed (hit or miss)
input logic ICacheMiss, // instruction cache miss
input logic ICacheAccess, // instruction cache access
input logic DivBusyE, // integer divide busy
input logic FDivBusyE, // floating point divide busy
// fault sources
input logic InstrAccessFaultF, // instruction access fault
input logic LoadAccessFaultM, StoreAmoAccessFaultM, // load or store access fault
input logic HPTWInstrAccessFaultF, // hardware page table access fault while fetching instruction PTE
input logic InstrPageFaultF, // page faults
input logic LoadPageFaultM, StoreAmoPageFaultM, // page faults
input logic InstrMisalignedFaultM, // misaligned instruction fault
input logic LoadMisalignedFaultM, StoreAmoMisalignedFaultM, // misaligned data fault
input logic IllegalIEUFPUInstrD, // illegal instruction from IEU or FPU
input logic MTimerInt, MExtInt, SExtInt, MSwInt, // interrupt sources
input logic [63:0] MTIME_CLINT, // timer value from CLINT
input logic [4:0] SetFflagsM, // set FCSR flags from FPU
input logic SelHPTW, // HPTW in use. Causes system to use S-mode endianness for accesses
input logic InstrAccessFaultF, // instruction access fault
input logic LoadAccessFaultM, StoreAmoAccessFaultM, // load or store access fault
input logic HPTWInstrAccessFaultF, // hardware page table access fault while fetching instruction PTE
input logic InstrPageFaultF, // page faults
input logic LoadPageFaultM, StoreAmoPageFaultM, // page faults
input logic InstrMisalignedFaultM, // misaligned instruction fault
input logic LoadMisalignedFaultM, StoreAmoMisalignedFaultM, // misaligned data fault
input logic IllegalIEUFPUInstrD, // illegal instruction from IEU or FPU
input logic MTimerInt, MExtInt, SExtInt, MSwInt, // interrupt sources
input logic [63:0] MTIME_CLINT, // timer value from CLINT
input logic [4:0] SetFflagsM, // set FCSR flags from FPU
input logic SelHPTW, // HPTW in use. Causes system to use S-mode endianness for accesses
// CSR outputs
output logic [P.XLEN-1:0] CSRReadValW, // Value read from CSR
output logic [1:0] PrivilegeModeW, // current privilege mode
output logic [1:0] PrivilegeModeW, // current privilege mode
output logic [P.XLEN-1:0] SATP_REGW, // supervisor address translation register
output logic STATUS_MXR, STATUS_SUM, STATUS_MPRV, // status register bits
output logic [1:0] STATUS_MPP, STATUS_FS, // status register bits
output var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration entries to MMU
output var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW [P.PMP_ENTRIES-1:0], // PMP address entries to MMU
output logic [2:0] FRM_REGW, // FPU rounding mode
output logic STATUS_MXR, STATUS_SUM, STATUS_MPRV, // status register bits
output logic [1:0] STATUS_MPP, STATUS_FS, // status register bits
output var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0], // PMP configuration entries to MMU
output var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW [P.PMP_ENTRIES-1:0], // PMP address entries to MMU
output logic [2:0] FRM_REGW, // FPU rounding mode
// PC logic output in privileged unit
output logic [P.XLEN-1:0] UnalignedPCNextF, // Next PC from trap/return PC logic
// control outputs
output logic RetM, TrapM, // return instruction, or trap
output logic sfencevmaM, // sfence.vma instruction
input logic InvalidateICacheM, // fence instruction
output logic BigEndianM, // Use big endian in current privilege mode
output logic RetM, TrapM, // return instruction, or trap
output logic sfencevmaM, // sfence.vma instruction
input logic InvalidateICacheM, // fence instruction
output logic BigEndianM, // Use big endian in current privilege mode
// Fault outputs
output logic BreakpointFaultM, EcallFaultM, // breakpoint and Ecall traps should retire
output logic wfiM, IntPendingM // Stall in Memory stage for WFI until interrupt pending or timeout
output logic BreakpointFaultM, EcallFaultM, // breakpoint and Ecall traps should retire
output logic wfiM, IntPendingM // Stall in Memory stage for WFI until interrupt pending or timeout
);
logic [3:0] CauseM; // trap cause
logic [15:0] MEDELEG_REGW; // exception delegation CSR
logic [11:0] MIDELEG_REGW; // interrupt delegation CSR
logic sretM, mretM; // supervisor / machine return instruction
logic IllegalCSRAccessM; // Illegal access to CSR
logic IllegalIEUFPUInstrM; // Illegal IEU or FPU instruction, delayed to Mem stage
logic InstrPageFaultM; // Instruction page fault, delayed to Mem stage
logic InstrAccessFaultM; // Instruction access fault, delayed to Mem stages
logic IllegalInstrFaultM; // Illegal instruction fault
logic STATUS_SPP, STATUS_TSR, STATUS_TW, STATUS_TVM; // Status bits needed within privileged unit
logic STATUS_MIE, STATUS_SIE; // status bits: interrupt enables
logic [11:0] MIP_REGW, MIE_REGW; // interrupt pending and enable bits
logic [1:0] NextPrivilegeModeM; // next privilege mode based on trap or return
logic DelegateM; // trap should be delegated
logic InterruptM; // interrupt occuring
logic ExceptionM; // Memory stage instruction caused a fault
logic HPTWInstrAccessFaultM; // Hardware page table access fault while fetching instruction PTE
logic [3:0] CauseM; // trap cause
logic [15:0] MEDELEG_REGW; // exception delegation CSR
logic [11:0] MIDELEG_REGW; // interrupt delegation CSR
logic sretM, mretM; // supervisor / machine return instruction
logic IllegalCSRAccessM; // Illegal access to CSR
logic IllegalIEUFPUInstrM; // Illegal IEU or FPU instruction, delayed to Mem stage
logic InstrPageFaultM; // Instruction page fault, delayed to Mem stage
logic InstrAccessFaultM; // Instruction access fault, delayed to Mem stages
logic IllegalInstrFaultM; // Illegal instruction fault
logic STATUS_SPP, STATUS_TSR, STATUS_TW, STATUS_TVM; // Status bits needed within privileged unit
logic STATUS_MIE, STATUS_SIE; // status bits: interrupt enables
logic [11:0] MIP_REGW, MIE_REGW; // interrupt pending and enable bits
logic [1:0] NextPrivilegeModeM; // next privilege mode based on trap or return
logic DelegateM; // trap should be delegated
logic InterruptM; // interrupt occuring
logic ExceptionM; // Memory stage instruction caused a fault
logic HPTWInstrAccessFaultM; // Hardware page table access fault while fetching instruction PTE
// track the current privilege level
privmode #(P) privmode(.clk, .reset, .StallW, .TrapM, .mretM, .sretM, .DelegateM,
.STATUS_MPP, .STATUS_SPP, .NextPrivilegeModeM, .PrivilegeModeW);
@ -156,8 +155,3 @@ module privileged import cvw::*; #(parameter cvw_t P) (
.InstrValidM, .CommittedM, .CommittedF,
.TrapM, .RetM, .wfiM, .InterruptM, .ExceptionM, .IntPendingM, .DelegateM, .CauseM);
endmodule

8
src/privileged/privmode.sv
View File

@ -43,10 +43,10 @@ module privmode import cvw::*; #(parameter cvw_t P) (
always_comb begin
if (TrapM) begin // Change privilege based on DELEG registers (see 3.1.8)
if (P.S_SUPPORTED & DelegateM) NextPrivilegeModeM = P.S_MODE;
else NextPrivilegeModeM = P.M_MODE;
end else if (mretM) NextPrivilegeModeM = STATUS_MPP;
else if (sretM) NextPrivilegeModeM = {1'b0, STATUS_SPP};
else NextPrivilegeModeM = PrivilegeModeW;
else NextPrivilegeModeM = P.M_MODE;
end else if (mretM) NextPrivilegeModeM = STATUS_MPP;
else if (sretM) NextPrivilegeModeM = {1'b0, STATUS_SPP};
else NextPrivilegeModeM = PrivilegeModeW;
end
flopenl #(2) privmodereg(clk, reset, ~StallW, NextPrivilegeModeM, P.M_MODE, PrivilegeModeW);

16
src/privileged/trap.sv
View File

@ -63,13 +63,13 @@ module trap import cvw::*; #(parameter cvw_t P) (
assign MIntGlobalEnM = (PrivilegeModeW != P.M_MODE) | STATUS_MIE; // if M ints enabled or lower priv 3.1.9
assign SIntGlobalEnM = (PrivilegeModeW == P.U_MODE) | ((PrivilegeModeW == P.S_MODE) & STATUS_SIE); // if in lower priv mode, or if S ints enabled and not in higher priv mode 3.1.9
assign PendingIntsM = MIP_REGW & MIE_REGW;
assign IntPendingM = |PendingIntsM;
assign Committed = CommittedM | CommittedF;
assign EnabledIntsM = ({12{MIntGlobalEnM}} & PendingIntsM & ~MIDELEG_REGW | {12{SIntGlobalEnM}} & PendingIntsM & MIDELEG_REGW);
assign ValidIntsM = {12{~Committed}} & EnabledIntsM;
assign InterruptM = (|ValidIntsM) & InstrValidM; // suppress interrupt if the memory system has partially processed a request.
assign DelegateM = P.S_SUPPORTED & (InterruptM ? MIDELEG_REGW[CauseM] : MEDELEG_REGW[CauseM]) &
assign PendingIntsM = MIP_REGW & MIE_REGW;
assign IntPendingM = |PendingIntsM;
assign Committed = CommittedM | CommittedF;
assign EnabledIntsM = ({12{MIntGlobalEnM}} & PendingIntsM & ~MIDELEG_REGW | {12{SIntGlobalEnM}} & PendingIntsM & MIDELEG_REGW);
assign ValidIntsM = {12{~Committed}} & EnabledIntsM;
assign InterruptM = (|ValidIntsM) & InstrValidM; // suppress interrupt if the memory system has partially processed a request.
assign DelegateM = P.S_SUPPORTED & (InterruptM ? MIDELEG_REGW[CauseM] : MEDELEG_REGW[CauseM]) &
(PrivilegeModeW == P.U_MODE | PrivilegeModeW == P.S_MODE);
///////////////////////////////////////////
@ -88,7 +88,7 @@ module trap import cvw::*; #(parameter cvw_t P) (
LoadAccessFaultM | StoreAmoAccessFaultM;
// coverage on
assign TrapM = ExceptionM | InterruptM;
assign RetM = mretM | sretM;
assign RetM = mretM | sretM;
///////////////////////////////////////////
// Cause priority defined in table 3.7 of 20190608 privileged spec

33
src/uncore/ahbapbbridge.sv
View File

@ -27,27 +27,27 @@
module ahbapbbridge import cvw::*; #(parameter cvw_t P,
parameter PERIPHS = 2) (
input logic HCLK, HRESETn,
input logic [PERIPHS-1:0] HSEL,
input logic HCLK, HRESETn,
input logic [PERIPHS-1:0] HSEL,
input logic [P.PA_BITS-1:0] HADDR,
input logic [P.XLEN-1:0] HWDATA,
input logic [P.XLEN/8-1:0] HWSTRB,
input logic HWRITE,
input logic [1:0] HTRANS,
input logic HREADY,
input logic HWRITE,
input logic [1:0] HTRANS,
input logic HREADY,
// input logic [3:0] HPROT, // not used
output logic [P.XLEN-1:0] HRDATA,
output logic HRESP, HREADYOUT,
output logic PCLK, PRESETn,
output logic [PERIPHS-1:0] PSEL,
output logic PWRITE,
output logic PENABLE,
output logic [31:0] PADDR,
output logic HRESP, HREADYOUT,
output logic PCLK, PRESETn,
output logic [PERIPHS-1:0] PSEL,
output logic PWRITE,
output logic PENABLE,
output logic [31:0] PADDR,
output logic [P.XLEN-1:0] PWDATA,
// output logic [2:0] PPROT, // not used
output logic [P.XLEN/8-1:0] PSTRB,
// output logic PWAKEUP // not used
input logic [PERIPHS-1:0] PREADY,
input logic [PERIPHS-1:0] PREADY,
input var [PERIPHS-1:0][P.XLEN-1:0] PRDATA
);
@ -56,15 +56,15 @@ module ahbapbbridge import cvw::*; #(parameter cvw_t P,
logic PREADYOUT;
// convert AHB to APB signals
assign PCLK = HCLK;
assign PCLK = HCLK;
assign PRESETn = HRESETn;
// identify start of a transaction
assign initTrans = HTRANS[1] & HREADY; // start a transaction when the bus is ready and an active transaction is requested
assign initTrans = HTRANS[1] & HREADY; // start a transaction when the bus is ready and an active transaction is requested
assign initTransSel = initTrans & |HSEL; // capture data and address if any of the peripherals are selected
// delay AHB Address phase signals to align with AHB Data phase because APB expects them at the same time
flopen #(32) addrreg(HCLK, HREADY, HADDR[31:0], PADDR);
flopen #(32) addrreg(HCLK, HREADY, HADDR[31:0], PADDR);
flopenr #(1) writereg(HCLK, ~HRESETn, HREADY, HWRITE, PWRITE);
flopenr #(PERIPHS) selreg(HCLK, ~HRESETn, HREADY, HSEL & {PERIPHS{initTrans}}, PSEL);
// PPROT[2:0] = {Data/InstrB, Secure, Privileged};
@ -73,7 +73,7 @@ module ahbapbbridge import cvw::*; #(parameter cvw_t P,
// AHB Data phase signal doesn't need delay. Note that they are guaranteed to remain stable until READY is asserted
assign PWDATA = HWDATA;
assign PSTRB = HWSTRB;
assign PSTRB = HWSTRB;
// enable logic: goes high a cycle after initTrans, then back low on cycle after desired PREADY is asserted
// cycle1: AHB puts HADDR, HWRITE, HSEL on bus. initTrans is 1, and these are captured
@ -103,4 +103,3 @@ assign HREADYOUT = PREADYOUT & ~initTransSelD; // don't raise HREADYOUT before a
// resp logic
assign HRESP = 0; // bridge never indicates errors
endmodule

32
src/uncore/clint_apb.sv
View File

@ -31,11 +31,11 @@ module clint_apb import cvw::*; #(parameter cvw_t P) (
input logic PCLK, PRESETn,
input logic PSEL,
input logic [15:0] PADDR,
input logic [P.XLEN-1:0] PWDATA,
input logic [P.XLEN/8-1:0] PSTRB,
input logic [P.XLEN-1:0] PWDATA,
input logic [P.XLEN/8-1:0] PSTRB,
input logic PWRITE,
input logic PENABLE,
output logic [P.XLEN-1:0] PRDATA,
output logic [P.XLEN-1:0] PRDATA,
output logic PREADY,
output logic [63:0] MTIME,
output logic MTimerInt, MSwInt
@ -48,11 +48,11 @@ module clint_apb import cvw::*; #(parameter cvw_t P) (
integer i, j;
assign memwrite = PWRITE & PENABLE & PSEL; // only write in access phase
assign PREADY = 1'b1; // CLINT never takes >1 cycle to respond
assign PREADY = 1'b1; // CLINT never takes >1 cycle to respond
// word aligned reads
if (P.XLEN==64) assign #2 entry = {PADDR[15:3], 3'b000};
else assign #2 entry = {PADDR[15:2], 2'b00};
else assign #2 entry = {PADDR[15:2], 2'b00};
// DH 2/20/21: Eventually allow MTIME to run off a separate clock
// This will require synchronizing MTIME to the system clock
@ -150,36 +150,36 @@ module clint_apb import cvw::*; #(parameter cvw_t P) (
endmodule
module timeregsync import cvw::*; #(parameter cvw_t P) (
input logic clk, resetn,
input logic we0, we1,
input logic clk, resetn,
input logic we0, we1,
input logic [P.XLEN-1:0] wd,
output logic [63:0] q);
output logic [63:0] q);
if (P.XLEN==64)
always_ff @(posedge clk or negedge resetn)
if (~resetn) q <= 0;
if (~resetn) q <= 0;
else if (we0) q <= wd;
else q <= q + 1;
else
always_ff @(posedge clk or negedge resetn)
if (~resetn) q <= 0;
if (~resetn) q <= 0;
else if (we0) q[31:0] <= wd;
else if (we1) q[63:32] <= wd;
else q <= q + 1;
endmodule
module timereg import cvw::*; #(parameter cvw_t P) (
input logic PCLK, PRESETn, TIMECLK,
input logic we0, we1,
input logic PCLK, PRESETn, TIMECLK,
input logic we0, we1,
input logic [P.XLEN-1:0] PWDATA,
output logic [63:0] MTIME,
output logic done);
output logic [63:0] MTIME,
output logic done);
// if (P.TIMEBASE_SYNC) begin:timereg // use PCLK for MTIME
if (1) begin:timereg // use PCLK for MTIME
timregsync timeregsync(.clk(PCLK), .resetn(PRESETn), .we0, .we1, .wd(PWDATA), .q(MTIME));
assign done = 1; // immediately completes
end else begin // use asynchronous TIMECLK
assign done = 1; // immediately completes
end else begin // use asynchronous TIMECLK
// TIME counter runs on TIMECLK but bus interface runs on PCLK
// Need to synchronize reads and writes
// This is subtle because synchronizing a binary counter on a per-bit basis could give a mix of old and new bits

131
src/uncore/gpio_apb.sv
View File

@ -29,78 +29,78 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module gpio_apb import cvw::*; #(parameter cvw_t P) (
input logic PCLK, PRESETn,
input logic PSEL,
input logic [7:0] PADDR,
input logic [P.XLEN-1:0] PWDATA,
input logic PCLK, PRESETn,
input logic PSEL,
input logic [7:0] PADDR,
input logic [P.XLEN-1:0] PWDATA,
input logic [P.XLEN/8-1:0] PSTRB,
input logic PWRITE,
input logic PENABLE,
output logic [P.XLEN-1:0] PRDATA,
output logic PREADY,
input logic [31:0] iof0, iof1,
input logic [31:0] GPIOIN,
output logic [31:0] GPIOOUT, GPIOEN,
output logic GPIOIntr
input logic PWRITE,
input logic PENABLE,
output logic [P.XLEN-1:0] PRDATA,
output logic PREADY,
input logic [31:0] iof0, iof1,
input logic [31:0] GPIOIN,
output logic [31:0] GPIOOUT, GPIOEN,
output logic GPIOIntr
);
logic [31:0] input0d, input1d, input2d, input3d;
logic [31:0] input_val, input_en, output_en, output_val;
logic [31:0] rise_ie, rise_ip, fall_ie, fall_ip, high_ie, high_ip, low_ie, low_ip;
logic [31:0] out_xor, iof_en, iof_sel, iof_out, gpio_out;
logic [7:0] entry;
logic [31:0] Din, Dout;
logic memwrite;
logic [31:0] input0d, input1d, input2d, input3d;
logic [31:0] input_val, input_en, output_en, output_val;
logic [31:0] rise_ie, rise_ip, fall_ie, fall_ip, high_ie, high_ip, low_ie, low_ip;
logic [31:0] out_xor, iof_en, iof_sel, iof_out, gpio_out;
logic [7:0] entry;
logic [31:0] Din, Dout;
logic memwrite;
// APB I/O
assign entry = {PADDR[7:2],2'b00}; // 32-bit word-aligned accesses
assign entry = {PADDR[7:2],2'b00}; // 32-bit word-aligned accesses
assign memwrite = PWRITE & PENABLE & PSEL; // only write in access phase
assign PREADY = 1'b1; // GPIO never takes >1 cycle to respond
assign PREADY = 1'b1; // GPIO never takes >1 cycle to respond
// account for subword read/write circuitry
// -- Note GPIO registers are 32 bits no matter what; access them with LW SW.
// (At least that's what I think when FE310 spec says "only naturally aligned 32-bit accesses are supported")
if (P.XLEN == 64) begin
assign Din = entry[2] ? PWDATA[63:32] : PWDATA[31:0];
assign Din = entry[2] ? PWDATA[63:32] : PWDATA[31:0];
assign PRDATA = entry[2] ? {Dout,32'b0} : {32'b0,Dout};
end else begin // 32-bit
assign Din = PWDATA[31:0];
assign Din = PWDATA[31:0];
assign PRDATA = Dout;
end
// register access
always_ff @(posedge PCLK, negedge PRESETn)
if (~PRESETn) begin // asynch reset
input_en <= 0;
input_en <= 0;
output_en <= 0;
// *** synch reset not yet implemented [DH: can we delete this comment? Check if a sync reset is required]
output_val <= #1 0;
rise_ie <= #1 0;
rise_ip <= #1 0;
fall_ie <= #1 0;
fall_ip <= #1 0;
high_ie <= #1 0;
high_ip <= #1 0;
low_ie <= #1 0;
low_ip <= #1 0;
iof_en <= #1 0;
iof_sel <= #1 0;
out_xor <= #1 0;
rise_ie <= #1 0;
rise_ip <= #1 0;
fall_ie <= #1 0;
fall_ip <= #1 0;
high_ie <= #1 0;
high_ip <= #1 0;
low_ie <= #1 0;
low_ip <= #1 0;
iof_en <= #1 0;
iof_sel <= #1 0;
out_xor <= #1 0;
end else begin // writes
// According to FE310 spec: Once the interrupt is pending, it will remain set until a 1 is written to the *_ip register at that bit.
/* verilator lint_off CASEINCOMPLETE */
if (memwrite)
case(entry)
8'h04: input_en <= #1 Din;
8'h08: output_en <= #1 Din;
8'h04: input_en <= #1 Din;
8'h08: output_en <= #1 Din;
8'h0C: output_val <= #1 Din;
8'h18: rise_ie <= #1 Din;
8'h20: fall_ie <= #1 Din;
8'h28: high_ie <= #1 Din;
8'h30: low_ie <= #1 Din;
8'h38: iof_en <= #1 Din;
8'h3C: iof_sel <= #1 Din;
8'h40: out_xor <= #1 Din;
8'h18: rise_ie <= #1 Din;
8'h20: fall_ie <= #1 Din;
8'h28: high_ie <= #1 Din;
8'h30: low_ie <= #1 Din;
8'h38: iof_en <= #1 Din;
8'h3C: iof_sel <= #1 Din;
8'h40: out_xor <= #1 Din;
endcase
/* verilator lint_on CASEINCOMPLETE */
@ -115,21 +115,21 @@ module gpio_apb import cvw::*; #(parameter cvw_t P) (
else low_ip <= low_ip | ~input3d;
case(entry) // flop to sample inputs
8'h00: Dout <= #1 input_val;
8'h04: Dout <= #1 input_en;
8'h08: Dout <= #1 output_en;
8'h0C: Dout <= #1 output_val;
8'h18: Dout <= #1 rise_ie;
8'h1C: Dout <= #1 rise_ip;
8'h20: Dout <= #1 fall_ie;
8'h24: Dout <= #1 fall_ip;
8'h28: Dout <= #1 high_ie;
8'h2C: Dout <= #1 high_ip;
8'h30: Dout <= #1 low_ie;
8'h34: Dout <= #1 low_ip;
8'h38: Dout <= #1 iof_en;
8'h3C: Dout <= #1 iof_sel;
8'h40: Dout <= #1 out_xor;
8'h00: Dout <= #1 input_val;
8'h04: Dout <= #1 input_en;
8'h08: Dout <= #1 output_en;
8'h0C: Dout <= #1 output_val;
8'h18: Dout <= #1 rise_ie;
8'h1C: Dout <= #1 rise_ip;
8'h20: Dout <= #1 fall_ie;
8'h24: Dout <= #1 fall_ip;
8'h28: Dout <= #1 high_ie;
8'h2C: Dout <= #1 high_ip;
8'h30: Dout <= #1 low_ie;
8'h34: Dout <= #1 low_ip;
8'h38: Dout <= #1 iof_en;
8'h3C: Dout <= #1 iof_sel;
8'h40: Dout <= #1 out_xor;
default: Dout <= #1 0;
endcase
end
@ -137,18 +137,17 @@ module gpio_apb import cvw::*; #(parameter cvw_t P) (
// chip i/o
// connect OUT to IN for loopback testing
if (P.GPIO_LOOPBACK_TEST) assign input0d = ((output_en & GPIOOUT) | (~output_en & GPIOIN)) & input_en;
else assign input0d = GPIOIN & input_en;
else assign input0d = GPIOIN & input_en;
// synchroninzer for inputs
flop #(32) sync1(PCLK,input0d,input1d);
flop #(32) sync2(PCLK,input1d,input2d);
flop #(32) sync3(PCLK,input2d,input3d);
assign input_val = input3d;
assign iof_out = iof_sel & iof1 | ~iof_sel & iof0; // per-bit mux between iof1 and iof0
assign gpio_out = iof_en & iof_out | ~iof_en & output_val; // per-bit mux between IOF and output_val
assign GPIOOUT = gpio_out ^ out_xor; // per-bit flip output polarity
assign GPIOEN = output_en;
assign iof_out = iof_sel & iof1 | ~iof_sel & iof0; // per-bit mux between iof1 and iof0
assign gpio_out = iof_en & iof_out | ~iof_en & output_val; // per-bit mux between IOF and output_val
assign GPIOOUT = gpio_out ^ out_xor; // per-bit flip output polarity
assign GPIOEN = output_en;
assign GPIOIntr = |{(rise_ip & rise_ie),(fall_ip & fall_ie),(high_ip & high_ie),(low_ip & low_ie)};
assign GPIOIntr = |{(rise_ip & rise_ie),(fall_ip & fall_ie),(high_ip & high_ie),(low_ip & low_ie)};
endmodule

50
src/uncore/plic_apb.sv
View File

@ -41,22 +41,22 @@
// hardcoded to 2 contexts for now; later upgrade to arbitrary (up to 15872) contexts
module plic_apb import cvw::*; #(parameter cvw_t P) (
input logic PCLK, PRESETn,
input logic PSEL,
input logic [27:0] PADDR,
input logic PCLK, PRESETn,
input logic PSEL,
input logic [27:0] PADDR,
input logic [P.XLEN-1:0] PWDATA,
input logic [P.XLEN/8-1:0] PSTRB,
input logic PWRITE,
input logic PENABLE,
input logic PWRITE,
input logic PENABLE,
output logic [P.XLEN-1:0] PRDATA,
output logic PREADY,
input logic UARTIntr,GPIOIntr,
output logic MExtInt, SExtInt
output logic PREADY,
input logic UARTIntr,GPIOIntr,
output logic MExtInt, SExtInt
);
logic memwrite, memread;
logic [23:0] entry;
logic [31:0] Din, Dout;
logic memwrite, memread;
logic [23:0] entry;
logic [31:0] Din, Dout;
// context-independent signals
logic [`N:1] requests;
@ -78,9 +78,9 @@ module plic_apb import cvw::*; #(parameter cvw_t P) (
// =======
assign memwrite = PWRITE & PENABLE & PSEL; // only write in access phase
assign memread = ~PWRITE & PSEL; // read at start of access phase. PENABLE hasn't set up before this
assign PREADY = 1'b1; // PLIC never takes >1 cycle to respond
assign entry = {PADDR[23:2],2'b0};
assign memread = ~PWRITE & PSEL; // read at start of access phase. PENABLE hasn't set up before this
assign PREADY = 1'b1; // PLIC never takes >1 cycle to respond
assign entry = {PADDR[23:2],2'b0};
// account for subword read/write circuitry
// -- Note PLIC registers are 32 bits no matter what; access them with LW SW.
@ -97,6 +97,7 @@ module plic_apb import cvw::*; #(parameter cvw_t P) (
// ==================
// Register Interface
// ==================
always @(posedge PCLK) begin
// resetting
if (~PRESETn) begin
@ -110,19 +111,19 @@ module plic_apb import cvw::*; #(parameter cvw_t P) (
casez(entry)
24'h0000??: intPriority[entry[7:2]] <= #1 Din[2:0];
`ifdef PLIC_NUM_SRC_LT_32 // eventually switch to a generate for loop so as to deprecate PLIC_NUM_SRC_LT_32 and allow up to 1023 sources
24'h002000: intEn[0][`N:1] <= #1 Din[`N:1];
24'h002080: intEn[1][`N:1] <= #1 Din[`N:1];
24'h002000: intEn[0][`N:1] <= #1 Din[`N:1];
24'h002080: intEn[1][`N:1] <= #1 Din[`N:1];
`endif
`ifndef PLIC_NUM_SRC_LT_32
24'h002000: intEn[0][31:1] <= #1 Din[31:1];
24'h002004: intEn[0][`N:32] <= #1 Din[31:0];
24'h002080: intEn[1][31:1] <= #1 Din[31:1];
24'h002084: intEn[1][`N:32] <= #1 Din[31:0];
24'h002000: intEn[0][31:1] <= #1 Din[31:1];
24'h002004: intEn[0][`N:32] <= #1 Din[31:0];
24'h002080: intEn[1][31:1] <= #1 Din[31:1];
24'h002084: intEn[1][`N:32] <= #1 Din[31:0];
`endif
24'h200000: intThreshold[0] <= #1 Din[2:0];
24'h200004: intInProgress <= #1 intInProgress & ~({{`N-1{1'b0}}, 1'b1} << (Din[5:0]-1)); // lower "InProgress" to signify completion
24'h201000: intThreshold[1] <= #1 Din[2:0];
24'h201004: intInProgress <= #1 intInProgress & ~({{`N-1{1'b0}}, 1'b1} << (Din[5:0]-1)); // lower "InProgress" to signify completion
24'h200000: intThreshold[0] <= #1 Din[2:0];
24'h200004: intInProgress <= #1 intInProgress & ~({{`N-1{1'b0}}, 1'b1} << (Din[5:0]-1)); // lower "InProgress" to signify completion
24'h201000: intThreshold[1] <= #1 Din[2:0];
24'h201004: intInProgress <= #1 intInProgress & ~({{`N-1{1'b0}}, 1'b1} << (Din[5:0]-1)); // lower "InProgress" to signify completion
endcase
// Read synchronously because a read can have side effect of changing intInProgress
if (memread) begin
@ -245,4 +246,3 @@ module plic_apb import cvw::*; #(parameter cvw_t P) (
assign MExtInt = |(threshMask[0] & priorities_with_irqs[0]);
assign SExtInt = |(threshMask[1] & priorities_with_irqs[1]);
endmodule

40
src/uncore/ram_ahb.sv
View File

@ -30,33 +30,33 @@
module ram_ahb import cvw::*; #(parameter cvw_t P,
parameter BASE=0, RANGE = 65535) (
input logic HCLK, HRESETn,
input logic HSELRam,
input logic HCLK, HRESETn,
input logic HSELRam,
input logic [P.PA_BITS-1:0] HADDR,
input logic HWRITE,
input logic HREADY,
input logic [1:0] HTRANS,
input logic HWRITE,
input logic HREADY,
input logic [1:0] HTRANS,
input logic [P.XLEN-1:0] HWDATA,
input logic [P.XLEN/8-1:0] HWSTRB,
output logic [P.XLEN-1:0] HREADRam,
output logic HRESPRam, HREADYRam
output logic HRESPRam, HREADYRam
);
localparam ADDR_WIDTH = $clog2(RANGE/8);
localparam OFFSET = $clog2(P.XLEN/8);
localparam ADDR_WIDTH = $clog2(RANGE/8);
localparam OFFSET = $clog2(P.XLEN/8);
logic [P.XLEN/8-1:0] ByteMask;
logic [P.PA_BITS-1:0] HADDRD, RamAddr;
logic initTrans;
logic memwrite, memwriteD, memread;
logic nextHREADYRam;
logic DelayReady;
logic initTrans;
logic memwrite, memwriteD, memread;
logic nextHREADYRam;
logic DelayReady;
// a new AHB transactions starts when HTRANS requests a transaction,
// the peripheral is selected, and the previous transaction is completing
assign initTrans = HREADY & HSELRam & HTRANS[1] ;
assign memwrite = initTrans & HWRITE;
assign memread = initTrans & ~HWRITE;
assign memwrite = initTrans & HWRITE;
assign memread = initTrans & ~HWRITE;
flopenr #(1) memwritereg(HCLK, ~HRESETn, HREADY, memwrite, memwriteD);
flopenr #(P.PA_BITS) haddrreg(HCLK, ~HRESETn, HREADY, HADDR, HADDRD);
@ -74,7 +74,6 @@ module ram_ahb import cvw::*; #(parameter cvw_t P,
ram1p1rwbe #(.DEPTH(RANGE/8), .WIDTH(P.XLEN)) memory(.clk(HCLK), .ce(1'b1),
.addr(RamAddr[ADDR_WIDTH+OFFSET-1:OFFSET]), .we(memwriteD), .din(HWDATA), .bwe(HWSTRB), .dout(HREADRam));
// use this to add arbitrary latency to ram. Helps test AHB controller correctness
if(`RAM_LATENCY > 0) begin
logic [7:0] NextCycle, Cycle;
@ -89,15 +88,15 @@ module ram_ahb import cvw::*; #(parameter cvw_t P,
always_ff @(posedge HCLK)
if (~HRESETn) CurrState <= #1 READY;
else CurrState <= #1 NextState;
else CurrState <= #1 NextState;
always_comb begin
case(CurrState)
READY: if(initTrans & ~CycleFlag) NextState = DELAY;
else NextState = READY;
DELAY: if(CycleFlag) NextState = READY;
else NextState = DELAY;
default: NextState = READY;
else NextState = READY;
DELAY: if(CycleFlag) NextState = READY;
else NextState = DELAY;
default: NextState = READY;
endcase
end
@ -110,4 +109,3 @@ module ram_ahb import cvw::*; #(parameter cvw_t P,
end
endmodule

19
src/uncore/rom_ahb.sv
View File

@ -28,24 +28,23 @@
module rom_ahb import cvw::*; #(parameter cvw_t P,
parameter BASE=0, RANGE = 65535) (
input logic HCLK, HRESETn,
input logic HSELRom,
input logic [P.PA_BITS-1:0] HADDR,
input logic HREADY,
input logic [1:0] HTRANS,
output logic [P.XLEN-1:0] HREADRom,
output logic HRESPRom, HREADYRom
input logic HCLK, HRESETn,
input logic HSELRom,
input logic [P.PA_BITS-1:0] HADDR,
input logic HREADY,
input logic [1:0] HTRANS,
output logic [P.XLEN-1:0] HREADRom,
output logic HRESPRom, HREADYRom
);
localparam ADDR_WIDTH = $clog2(RANGE/8);
localparam OFFSET = $clog2(P.XLEN/8);
localparam OFFSET = $clog2(P.XLEN/8);
// Never stalls
assign HREADYRom = 1'b1;
assign HRESPRom = 0; // OK
assign HRESPRom = 0; // OK
// single-ported ROM
rom1p1r #(ADDR_WIDTH, P.XLEN, P.FPGA)
memory(.clk(HCLK), .ce(1'b1), .addr(HADDR[ADDR_WIDTH+OFFSET-1:OFFSET]), .dout(HREADRom));
endmodule

92
src/uncore/uartPC16550D.sv
View File

@ -45,7 +45,7 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
output logic INTR, TXRDYb, RXRDYb, // interrupt and ready lines
// Clocks
output logic BAUDOUTb, // active low baud clock
input logic RCLK, // usually BAUDOUTb tied to RCLK externally
input logic RCLK, // usually BAUDOUTb tied to RCLK externally
// E1A Driver
input logic SIN, DSRb, DCDb, CTSb, RIb, // UART external serial and flow-control inputs
output logic SOUT, RTSb, DTRb, OUT1b, OUT2b // UART external serial and flow-control outputs
@ -71,9 +71,9 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
logic DLAB; // Divisor Latch Access Bit (LCR bit 7)
// Baud and rx/tx timing
logic baudpulse, txbaudpulse, rxbaudpulse; // high one system clk cycle each baud/16 period
logic [16+UART_PRESCALE-1:0] baudcount;
logic [3:0] rxoversampledcnt, txoversampledcnt; // count oversampled-by-16
logic baudpulse, txbaudpulse, rxbaudpulse; // high one system clk cycle each baud/16 period
logic [16+UART_PRESCALE-1:0] baudcount;
logic [3:0] rxoversampledcnt, txoversampledcnt; // count oversampled-by-16
logic [3:0] rxbitsreceived, txbitssent;
statetype rxstate, txstate;
@ -122,6 +122,7 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
///////////////////////////////////////////
// Input synchronization: 2-stage synchronizer
///////////////////////////////////////////
always_ff @(posedge PCLK) begin
{SINd, DSRbd, DCDbd, CTSbd, RIbd} <= #1 {SIN, DSRb, DCDb, CTSb, RIb};
{SINsync, DSRbsync, DCDbsync, CTSbsync, RIbsync} <= #1 loop ? {SOUTbit, ~MCR[0], ~MCR[3], ~MCR[1], ~MCR[2]} :
@ -132,6 +133,7 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
///////////////////////////////////////////
// Register interface (Table 1, note some are read only and some write only)
///////////////////////////////////////////
always_ff @(posedge PCLK, negedge PRESETn)
if (~PRESETn) begin // Table 3 Reset Configuration
IER <= #1 4'b0;
@ -162,13 +164,13 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
if (~MEMWb & (A == 3'b101))
LSR[6:1] <= #1 Din[6:1]; // recommended only for test, see 8.6.3
else begin
LSR[0] <= #1 rxdataready; // Data ready
LSR[1] <= #1 (LSR[1] | RXBR[10]) & ~squashRXerrIP;; // overrun error
LSR[2] <= #1 (LSR[2] | RXBR[9]) & ~squashRXerrIP; // parity error
LSR[3] <= #1 (LSR[3] | RXBR[8]) & ~squashRXerrIP; // framing error
LSR[4] <= #1 (LSR[4] | rxbreak) & ~squashRXerrIP; // break indicator
LSR[5] <= #1 THRE; // THRE
LSR[6] <= #1 ~txsrfull & THRE; // TEMT
LSR[0] <= #1 rxdataready; // Data ready
LSR[1] <= #1 (LSR[1] | RXBR[10]) & ~squashRXerrIP;; // overrun error
LSR[2] <= #1 (LSR[2] | RXBR[9]) & ~squashRXerrIP; // parity error
LSR[3] <= #1 (LSR[3] | RXBR[8]) & ~squashRXerrIP; // framing error
LSR[4] <= #1 (LSR[4] | rxbreak) & ~squashRXerrIP; // break indicator
LSR[5] <= #1 THRE; // THRE
LSR[6] <= #1 ~txsrfull & THRE; // TEMT
if (rxfifohaserr) LSR[7] <= #1 1; // any bits in FIFO have error
end
@ -208,6 +210,7 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
///////////////////////////////////////////
// Ross Thompson: Found a bug. If the baud rate dividers DLM, and DLL are reloaded
// the baudcount is not reset to {DLM, DLL, UART_PRESCALE}
always_ff @(posedge PCLK, negedge PRESETn)
if (~PRESETn) begin
baudcount <= #1 1;
@ -233,17 +236,18 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
///////////////////////////////////////////
// receive timing and control
///////////////////////////////////////////
always_ff @(posedge PCLK, negedge PRESETn)
if (~PRESETn) begin
rxoversampledcnt <= #1 0;
rxstate <= #1 UART_IDLE;
rxbitsreceived <= #1 0;
rxtimeoutcnt <= #1 0;
rxoversampledcnt <= #1 0;
rxstate <= #1 UART_IDLE;
rxbitsreceived <= #1 0;
rxtimeoutcnt <= #1 0;
end else begin
if (rxstate == UART_IDLE & ~SINsync) begin // got start bit
rxstate <= #1 UART_ACTIVE;
rxstate <= #1 UART_ACTIVE;
rxoversampledcnt <= #1 0;
rxbitsreceived <= #1 0;
rxbitsreceived <= #1 0;
if (~rxfifotimeout) rxtimeoutcnt <= #1 0; // reset timeout when new character is arriving. Jacob Pease: Only if the timeout was not already reached. p.16 PC16550D.pdf
end else if (rxbaudpulse & (rxstate == UART_ACTIVE)) begin
rxoversampledcnt <= #1 rxoversampledcnt + 1; // 16x oversampled counter
@ -260,13 +264,14 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
// ***explain why
if(QEMU) assign rxcentered = rxbaudpulse & (rxoversampledcnt[1:0] == 2'b10); // implies rxstate = UART_ACTIVE
else assign rxcentered = rxbaudpulse & (rxoversampledcnt == 4'b1000); // implies rxstate = UART_ACTIVE
else assign rxcentered = rxbaudpulse & (rxoversampledcnt == 4'b1000); // implies rxstate = UART_ACTIVE
assign rxbitsexpected = 4'd1 + (4'd5 + {2'b00, LCR[1:0]}) + {3'b000, LCR[3]} + 4'd1; // start bit + data bits + (parity bit) + stop bit
///////////////////////////////////////////
// receive shift register, buffer register, FIFO
///////////////////////////////////////////
always_ff @(posedge PCLK, negedge PRESETn)
if (~PRESETn) rxshiftreg <= #1 10'b0000000001; // initialize so that there is a valid stop bit
else if (rxcentered) rxshiftreg <= #1 {rxshiftreg[8:0], SINsync}; // capture bit
@ -282,11 +287,11 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
assign rxdata = LCR[3] ? rxdata9[7:0] : rxdata9[8:1]; // discard parity bit
// ERROR CONDITIONS
assign rxparity = ^rxdata;
assign rxparityerr = (rxparity ^ rxparitybit ^ ~evenparitysel) & LCR[3]; // Check even/odd parity (*** check if LCR needs to be inverted)
assign rxparity = ^rxdata;
assign rxparityerr = (rxparity ^ rxparitybit ^ ~evenparitysel) & LCR[3]; // Check even/odd parity (*** check if LCR needs to be inverted)
assign rxoverrunerr = fifoenabled ? (rxfifoentries == 15) : rxdataready; // overrun if FIFO or receive buffer register full
assign rxframingerr = ~rxstopbit; // framing error if no stop bit
assign rxbreak = rxframingerr & (rxdata9 == 9'b0); // break when 0 for start + data + parity + stop time
assign rxbreak = rxframingerr & (rxdata9 == 9'b0); // break when 0 for start + data + parity + stop time
// receive FIFO and register
always_ff @(posedge PCLK)
@ -298,7 +303,7 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
end else if (rxstate == UART_DONE) begin
RXBR <= #1 {rxoverrunerr, rxparityerr, rxframingerr, rxdata}; // load recevive buffer register
if (rxoverrunerr) $warning("UART RX Overrun Err\n");
if (rxparityerr) $warning("UART RX Parity Err\n");
if (rxparityerr) $warning("UART RX Parity Err\n");
if (rxframingerr) $warning("UART RX Framing Err\n");
if (fifoenabled) begin
rxfifo[rxfifohead] <= #1 {rxoverrunerr, rxparityerr, rxframingerr, rxdata};
@ -339,14 +344,14 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
else assign rxfullbitunwrapped[i] = ({1'b0,rxfifohead}==i | rxfullbitunwrapped[i-1]) & (rxfifotailunwrapped != i);
end
for (i=0; i<16; i++) begin:rx
assign RXerrbit[i] = |rxfifo[i][10:8]; // are any of the error conditions set?
assign RXerrbit[i] = |rxfifo[i][10:8]; // are any of the error conditions set?
assign rxfullbit[i] = rxfullbitunwrapped[i] | rxfullbitunwrapped[i+16];
/* if (i > 0)
assign rxfullbit[i] = ((rxfifohead==i) | rxfullbit[i-1]) & (rxfifotail != i);
else
assign rxfullbit[0] = ((rxfifohead==i) | rxfullbit[15]) & (rxfifotail != i);*/
end
assign rxfifohaserr = |(RXerrbit & rxfullbit);
assign rxfifohaserr = |(RXerrbit & rxfullbit);
// receive buffer register and ready bit
always_ff @(posedge PCLK, negedge PRESETn) // track rxrdy for DMA mode (FCR3 = FCR0 = 1)
@ -361,22 +366,23 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
if (fifodmamodesel) RXRDYb = ~rxfifodmaready;
else RXRDYb = rxfifoempty;
end else begin
RBR = RXBR;
RBR = RXBR;
RXRDYb = ~rxdataready;
end
///////////////////////////////////////////
// transmit timing and control
///////////////////////////////////////////
always_ff @(posedge PCLK, negedge PRESETn)
if (~PRESETn) begin
txoversampledcnt <= #1 0;
txstate <= #1 UART_IDLE;
txbitssent <= #1 0;
txstate <= #1 UART_IDLE;
txbitssent <= #1 0;
end else if ((txstate == UART_IDLE) & txsrfull) begin // start transmitting
txstate <= #1 UART_ACTIVE;
txstate <= #1 UART_ACTIVE;
txoversampledcnt <= #1 1;
txbitssent <= #1 0;
txbitssent <= #1 0;
end else if (txbaudpulse & (txstate == UART_ACTIVE)) begin
txoversampledcnt <= #1 txoversampledcnt + 1;
if (txnextbit) begin // transmit at end of phase
@ -390,11 +396,12 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
assign txbitsexpected = 4'd1 + (4'd5 + {2'b00, LCR[1:0]}) + {3'b000, LCR[3]} + 4'd1 + {3'b000, LCR[2]} - 4'd1; // start bit + data bits + (parity bit) + stop bit(s) - 1
// *** explain; is this necessary?
if (QEMU) assign txnextbit = txbaudpulse & (txoversampledcnt[1:0] == 2'b00); // implies txstate = UART_ACTIVE
else assign txnextbit = txbaudpulse & (txoversampledcnt == 4'b0000); // implies txstate = UART_ACTIVE
else assign txnextbit = txbaudpulse & (txoversampledcnt == 4'b0000); // implies txstate = UART_ACTIVE
///////////////////////////////////////////
// transmit holding register, shift register, FIFO
///////////////////////////////////////////
always_comb begin // compute value for parity and tx holding register
nexttxdata = fifoenabled ? txfifo[txfifotail] : TXHR; // pick from FIFO or holding register
case (LCR[1:0]) // compute parity from appropriate number of bits
@ -426,9 +433,9 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
if (~MEMWb & A == 3'b000 & ~DLAB) begin // writing transmit holding register or fifo
if (fifoenabled) begin
txfifo[txfifohead] <= #1 Din;
txfifohead <= #1 txfifohead + 1;
txfifohead <= #1 txfifohead + 1;
end else begin
TXHR <= #1 Din;
TXHR <= #1 Din;
txhrfull <= #1 1;
end
$write("%c",Din); // for testbench
@ -436,12 +443,12 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
if (txstate == UART_IDLE) begin // move data into tx shift register if available
if (fifoenabled) begin
if (~txfifoempty & ~txsrfull) begin
txsr <= #1 txdata;
txsr <= #1 txdata;
txfifotail <= #1 txfifotail+1;
txsrfull <= #1 1;
txsrfull <= #1 1;
end
end else if (txhrfull) begin
txsr <= #1 txdata;
txsr <= #1 txdata;
txhrfull <= #1 0;
txsrfull <= #1 1;
end
@ -470,13 +477,13 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
HeadPointerLastMove <= 1'b0;
end
assign txfifoempty = (txfifohead == txfifotail) & ~HeadPointerLastMove;
assign txfifoempty = (txfifohead == txfifotail) & ~HeadPointerLastMove;
// verilator lint_off WIDTH
assign txfifoentries = (txfifohead >= txfifotail) ? (txfifohead-txfifotail) :
(txfifohead + 16 - txfifotail);
// verilator lint_on WIDTH
//assign txfifofull = (txfifoentries == 4'b1111);
assign txfifofull = (txfifohead == txfifotail) & HeadPointerLastMove;
//assign txfifofull = (txfifoentries == 4'b1111);
assign txfifofull = (txfifohead == txfifotail) & HeadPointerLastMove;
// transmit buffer ready bit
always_ff @(posedge PCLK, negedge PRESETn) // track txrdy for DMA mode (FCR3 = FCR0 = 1)
@ -490,11 +497,12 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
// Transmitter pin
assign SOUTbit = txsr[11]; // transmit most significant bit
assign SOUT = loop ? 1 : (LCR[6] ? 0 : SOUTbit); // tied to 1 during loopback or 0 during break
assign SOUT = loop ? 1 : (LCR[6] ? 0 : SOUTbit); // tied to 1 during loopback or 0 during break
///////////////////////////////////////////
// interrupts
///////////////////////////////////////////
assign RXerr = |LSR[4:1]; // LS interrupt if any of the flags are true
assign RXerrIP = RXerr & ~squashRXerrIP; // intr squashed upon reading LSR
assign rxdataavailintr = fifoenabled ? rxfifotriggered : rxdataready;
@ -533,15 +541,15 @@ module uartPC16550D #(parameter UART_PRESCALE, QEMU) (
// modem control logic
///////////////////////////////////////////
assign loop = MCR[4];
assign loop = MCR[4];
assign DTRb = ~MCR[0] | loop; // disable modem signals in loopback mode
assign RTSb = ~MCR[1] | loop;
assign OUT1b = ~MCR[2] | loop;
assign OUT2b = ~MCR[3] | loop;
assign DLAB = LCR[7];
assign evenparitysel = LCR[4];
assign fifoenabled = FCR[0];
assign evenparitysel = LCR[4];
assign fifoenabled = FCR[0];
assign fifodmamodesel = FCR[3];
always_comb
case (FCR[7:6])

31
src/uncore/uart_apb.sv
View File

@ -29,18 +29,18 @@
////////////////////////////////////////////////////////////////////////////////////////////////
module uart_apb import cvw::*; #(parameter cvw_t P) (
input logic PCLK, PRESETn,
input logic PSEL,
input logic [2:0] PADDR,
input logic [P.XLEN-1:0] PWDATA,
input logic PCLK, PRESETn,
input logic PSEL,
input logic [2:0] PADDR,
input logic [P.XLEN-1:0] PWDATA,
input logic [P.XLEN/8-1:0] PSTRB,
input logic PWRITE,
input logic PENABLE,
output logic [P.XLEN-1:0] PRDATA,
output logic PREADY,
input logic SIN, DSRb, DCDb, CTSb, RIb, // from E1A driver from RS232 interface
output logic SOUT, RTSb, DTRb, // to E1A driver to RS232 interface
output logic OUT1b, OUT2b, INTR, TXRDYb, RXRDYb); // to CPU
input logic PWRITE,
input logic PENABLE,
output logic [P.XLEN-1:0] PRDATA,
output logic PREADY,
input logic SIN, DSRb, DCDb, CTSb, RIb, // from E1A driver from RS232 interface
output logic SOUT, RTSb, DTRb, // to E1A driver to RS232 interface
output logic OUT1b, OUT2b, INTR, TXRDYb, RXRDYb); // to CPU
// UART interface signals
logic [2:0] entry;
@ -49,10 +49,10 @@ module uart_apb import cvw::*; #(parameter cvw_t P) (
assign memwrite = PWRITE & PENABLE & PSEL; // only write in access phase
assign memread = ~PWRITE & PENABLE & PSEL;
assign PREADY = 1'b1; // CLINT never takes >1 cycle to respond
assign entry = PADDR[2:0];
assign MEMRb = ~memread;
assign MEMWb = ~memwrite;
assign PREADY = 1'b1; // CLINT never takes >1 cycle to respond
assign entry = PADDR[2:0];
assign MEMRb = ~memread;
assign MEMWb = ~memwrite;
if (P.XLEN == 64) begin:uart
always_comb begin
@ -97,4 +97,3 @@ module uart_apb import cvw::*; #(parameter cvw_t P) (
);
endmodule

77
src/uncore/uncore.sv
View File

@ -29,58 +29,58 @@
module uncore import cvw::*; #(parameter cvw_t P)(
// AHB Bus Interface
input logic HCLK, HRESETn,
input logic TIMECLK,
input logic HCLK, HRESETn,
input logic TIMECLK,
input logic [P.PA_BITS-1:0] HADDR,
input logic [P.AHBW-1:0] HWDATA,
input logic [P.XLEN/8-1:0] HWSTRB,
input logic HWRITE,
input logic [2:0] HSIZE,
input logic [2:0] HBURST,
input logic [3:0] HPROT,
input logic [1:0] HTRANS,
input logic HMASTLOCK,
input logic HWRITE,
input logic [2:0] HSIZE,
input logic [2:0] HBURST,
input logic [3:0] HPROT,
input logic [1:0] HTRANS,
input logic HMASTLOCK,
input logic [P.AHBW-1:0] HRDATAEXT,
input logic HREADYEXT, HRESPEXT,
input logic HREADYEXT, HRESPEXT,
output logic [P.AHBW-1:0] HRDATA,
output logic HREADY, HRESP,
output logic HSELEXT,
output logic HREADY, HRESP,
output logic HSELEXT,
// peripheral pins
output logic MTimerInt, MSwInt, // Timer and software interrupts from CLINT
output logic MExtInt, SExtInt, // External interrupts from PLIC
output logic [63:0] MTIME_CLINT, // MTIME, from CLINT
input logic [31:0] GPIOIN, // GPIO pin input value
output logic [31:0] GPIOOUT, GPIOEN, // GPIO pin output value and enable
input logic UARTSin, // UART serial input
output logic UARTSout, // UART serial output
output logic SDCCmdOut, // SD Card command output
output logic SDCCmdOE, // SD Card command output enable
input logic SDCCmdIn, // SD Card command input
input logic [3:0] SDCDatIn, // SD Card data input
output logic SDCCLK // SD Card clock
output logic MTimerInt, MSwInt, // Timer and software interrupts from CLINT
output logic MExtInt, SExtInt, // External interrupts from PLIC
output logic [63:0] MTIME_CLINT, // MTIME, from CLINT
input logic [31:0] GPIOIN, // GPIO pin input value
output logic [31:0] GPIOOUT, GPIOEN, // GPIO pin output value and enable
input logic UARTSin, // UART serial input
output logic UARTSout, // UART serial output
output logic SDCCmdOut, // SD Card command output
output logic SDCCmdOE, // SD Card command output enable
input logic SDCCmdIn, // SD Card command input
input logic [3:0] SDCDatIn, // SD Card data input
output logic SDCCLK // SD Card clock
);
logic [P.XLEN-1:0] HREADRam, HREADSDC;
logic [10:0] HSELRegions;
logic HSELDTIM, HSELIROM, HSELRam, HSELCLINT, HSELPLIC, HSELGPIO, HSELUART, HSELSDC;
logic HSELDTIMD, HSELIROMD, HSELEXTD, HSELRamD, HSELCLINTD, HSELPLICD, HSELGPIOD, HSELUARTD, HSELSDCD;
logic HRESPRam, HRESPSDC;
logic HREADYRam, HRESPSDCD;
logic [10:0] HSELRegions;
logic HSELDTIM, HSELIROM, HSELRam, HSELCLINT, HSELPLIC, HSELGPIO, HSELUART, HSELSDC;
logic HSELDTIMD, HSELIROMD, HSELEXTD, HSELRamD, HSELCLINTD, HSELPLICD, HSELGPIOD, HSELUARTD, HSELSDCD;
logic HRESPRam, HRESPSDC;
logic HREADYRam, HRESPSDCD;
logic [P.XLEN-1:0] HREADBootRom;
logic HSELBootRom, HSELBootRomD, HRESPBootRom, HREADYBootRom, HREADYSDC;
logic HSELNoneD;
logic UARTIntr,GPIOIntr;
logic SDCIntM;
logic HSELBootRom, HSELBootRomD, HRESPBootRom, HREADYBootRom, HREADYSDC;
logic HSELNoneD;
logic UARTIntr,GPIOIntr;
logic SDCIntM;
logic PCLK, PRESETn, PWRITE, PENABLE;
logic [3:0] PSEL, PREADY;
logic [31:0] PADDR;
logic PCLK, PRESETn, PWRITE, PENABLE;
logic [3:0] PSEL, PREADY;
logic [31:0] PADDR;
logic [P.XLEN-1:0] PWDATA;
logic [P.XLEN/8-1:0] PSTRB;
logic [3:0][P.XLEN-1:0] PRDATA;
logic [P.XLEN-1:0] HREADBRIDGE;
logic HRESPBRIDGE, HREADYBRIDGE, HSELBRIDGE, HSELBRIDGED;
logic HRESPBRIDGE, HREADYBRIDGE, HSELBRIDGE, HSELBRIDGED;
// Determine which region of physical memory (if any) is being accessed
// Use a trimmed down portion of the PMA checker - only the address decoders
@ -154,9 +154,9 @@ module uncore import cvw::*; #(parameter cvw_t P)(
.SDCIntM
);
end else begin : sdc
assign SDCCLK = 0;
assign SDCCLK = 0;
assign SDCCmdOut = 0;
assign SDCCmdOE = 0;
assign SDCCmdOE = 0;
end
// AHB Read Multiplexer
@ -189,4 +189,3 @@ module uncore import cvw::*; #(parameter cvw_t P)(
HSELCLINTD, HSELGPIOD, HSELUARTD, HSELPLICD, HSELSDCD, HSELNoneD});
flopenr #(1) hselbridgedelayreg(HCLK, ~HRESETn, HREADY, HSELBRIDGE, HSELBRIDGED);
endmodule

114
src/wally/cvw.sv
View File

@ -47,8 +47,9 @@ typedef struct packed {
// RISC-V Features
logic ZICSR_SUPPORTED;
logic ZIFENCEI_SUPPORTED;
logic [11:0] COUNTERS;
logic ZICOUNTERS_SUPPORTED;
logic [11:0] COUNTERS;
logic ZICNTR_SUPPORTED;
logic ZIHPM_SUPPORTED;
logic ZFH_SUPPORTED;
logic SSTC_SUPPORTED;
logic VIRTMEM_SUPPORTED;
@ -56,6 +57,10 @@ typedef struct packed {
logic BIGENDIAN_SUPPORTED;
logic SVADU_SUPPORTED;
logic ZMMUL_SUPPORTED;
logic ZICBOM_SUPPORTED;
logic ZICBOZ_SUPPORTED;
logic ZICBOP_SUPPORTED;
logic SVPBMT_SUPPORTED;
// Microarchitectural Features
logic BUS_SUPPORTED;
@ -63,31 +68,31 @@ typedef struct packed {
logic ICACHE_SUPPORTED;
// TLB configuration. Entries should be a power of 2
int ITLB_ENTRIES;
int DTLB_ENTRIES;
int ITLB_ENTRIES;
int DTLB_ENTRIES;
// Cache configuration. Sizes should be a power of two
// typical configuration 4 ways, 4096 ints per way, 256 bit or more lines
int DCACHE_NUMWAYS;
int DCACHE_WAYSIZEINBYTES;
int DCACHE_LINELENINBITS;
int ICACHE_NUMWAYS;
int ICACHE_WAYSIZEINBYTES;
int ICACHE_LINELENINBITS;
int DCACHE_NUMWAYS;
int DCACHE_WAYSIZEINBYTES;
int DCACHE_LINELENINBITS;
int ICACHE_NUMWAYS;
int ICACHE_WAYSIZEINBYTES;
int ICACHE_LINELENINBITS;
// Integer Divider Configuration
// IDIV_BITSPERCYCLE must be 1, 2, or 4
int IDIV_BITSPERCYCLE;
int IDIV_BITSPERCYCLE;
logic IDIV_ON_FPU;
// Legal number of PMP entries are 0, 16, or 64
int PMP_ENTRIES;
int PMP_ENTRIES;
// Address space
logic [63:0] RESET_VECTOR;
// WFI Timeout Wait
int WFI_TIMEOUT_BIT;
int WFI_TIMEOUT_BIT;
// Peripheral Addresses
// Peripheral memory space extends from BASE to BASE+RANGE
@ -129,24 +134,23 @@ typedef struct packed {
logic GPIO_LOOPBACK_TEST;
// Hardware configuration
int UART_PRESCALE ;
int UART_PRESCALE ;
// Interrupt configuration
int PLIC_NUM_SRC;
int PLIC_NUM_SRC;
logic PLIC_NUM_SRC_LT_32;
int PLIC_GPIO_ID;
int PLIC_UART_ID;
logic BPRED_SUPPORTED;
BranchPredictorType BPRED_TYPE;
int BPRED_NUM_LHR;
int BPRED_SIZE;
int BTB_SIZE;
int PLIC_GPIO_ID;
int PLIC_UART_ID;
logic BPRED_SUPPORTED;
BranchPredictorType BPRED_TYPE;
int BPRED_NUM_LHR;
int BPRED_SIZE;
int BTB_SIZE;
// FPU division architecture
int RADIX;
int DIVCOPIES;
int RADIX;
int DIVCOPIES;
// bit manipulation
logic ZBA_SUPPORTED;
@ -199,47 +203,47 @@ typedef struct packed {
int PMPCFG_ENTRIES;
// Floating point constants for Quad, Double, Single, and Half precisions
int Q_LEN;
int Q_NE;
int Q_NF;
int Q_BIAS;
int Q_LEN;
int Q_NE;
int Q_NF;
int Q_BIAS;
logic [1:0] Q_FMT;
int D_LEN;
int D_NE;
int D_NF;
int D_BIAS;
int D_LEN;
int D_NE;
int D_NF;
int D_BIAS;
logic [1:0] D_FMT;
int S_LEN;
int S_NE;
int S_NF;
int S_BIAS;
int S_LEN;
int S_NE;
int S_NF;
int S_BIAS;
logic [1:0] S_FMT;
int H_LEN;
int H_NE;
int H_NF;
int H_BIAS;
int H_LEN;
int H_NE;
int H_NF;
int H_BIAS;
logic [1:0] H_FMT;
// Floating point length FLEN and number of exponent (NE) and fraction (NF) bits
int FLEN;
int NE ;
int NF ;
int NE ;
int NF ;
logic [1:0] FMT ;
int BIAS;
int BIAS;
// Floating point constants needed for FPU paramerterization
int FPSIZES;
int FMTBITS;
int LEN1 ;
int NE1 ;
int NF1 ;
int FPSIZES;
int FMTBITS;
int LEN1 ;
int NE1 ;
int NF1 ;
logic [1:0] FMT1 ;
int BIAS1;
int LEN2 ;
int NE2 ;
int NF2 ;
int BIAS1;
int LEN2 ;
int NE2 ;
int NF2 ;
logic [1:0] FMT2 ;
int BIAS2;
int BIAS2;
// largest length in IEU/FPU
int CVTLEN;
@ -251,7 +255,7 @@ typedef struct packed {
// division constants
int DIVN ;
int LOGR;
int LOGR ;
int RK ;
int LOGRK ;
int FPDUR ;

166
src/wally/wallypipelinedcore.sv
View File

@ -32,12 +32,12 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
input logic MTimerInt, MExtInt, SExtInt, MSwInt,
input logic [63:0] MTIME_CLINT,
// Bus Interface
input logic [P.AHBW-1:0] HRDATA,
input logic [P.AHBW-1:0] HRDATA,
input logic HREADY, HRESP,
output logic HCLK, HRESETn,
output logic [P.PA_BITS-1:0] HADDR,
output logic [P.AHBW-1:0] HWDATA,
output logic [P.XLEN/8-1:0] HWSTRB,
output logic [P.PA_BITS-1:0] HADDR,
output logic [P.AHBW-1:0] HWDATA,
output logic [P.XLEN/8-1:0] HWSTRB,
output logic HWRITE,
output logic [2:0] HSIZE,
output logic [2:0] HBURST,
@ -55,15 +55,15 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
logic IntDivE, W64E;
logic CSRReadM, CSRWriteM, PrivilegedM;
logic [1:0] AtomicM;
logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE;
logic [P.XLEN-1:0] SrcAM;
logic [P.XLEN-1:0] ForwardedSrcAE, ForwardedSrcBE;
logic [P.XLEN-1:0] SrcAM;
logic [2:0] Funct3E;
logic [31:0] InstrD;
logic [31:0] InstrM, InstrOrigM;
logic [P.XLEN-1:0] PCSpillF, PCE, PCLinkE;
logic [P.XLEN-1:0] PCM;
logic [P.XLEN-1:0] CSRReadValW, MDUResultW;
logic [P.XLEN-1:0] UnalignedPCNextF, PC2NextF;
logic [P.XLEN-1:0] PCSpillF, PCE, PCLinkE;
logic [P.XLEN-1:0] PCM;
logic [P.XLEN-1:0] CSRReadValW, MDUResultW;
logic [P.XLEN-1:0] UnalignedPCNextF, PC2NextF;
logic [1:0] MemRWM;
logic InstrValidD, InstrValidE, InstrValidM;
logic InstrMisalignedFaultM;
@ -83,31 +83,31 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
logic [4:0] RdE, RdM, RdW;
logic FPUStallD;
logic FWriteIntE;
logic [P.FLEN-1:0] FWriteDataM;
logic [P.XLEN-1:0] FIntResM;
logic [P.XLEN-1:0] FCvtIntResW;
logic [P.FLEN-1:0] FWriteDataM;
logic [P.XLEN-1:0] FIntResM;
logic [P.XLEN-1:0] FCvtIntResW;
logic FCvtIntW;
logic FDivBusyE;
logic FRegWriteM;
logic FCvtIntStallD;
logic FpLoadStoreM;
logic [4:0] SetFflagsM;
logic [P.XLEN-1:0] FIntDivResultW;
logic [P.XLEN-1:0] FIntDivResultW;
// memory management unit signals
logic ITLBWriteF;
logic ITLBMissF;
logic [P.XLEN-1:0] SATP_REGW;
logic [P.XLEN-1:0] SATP_REGW;
logic STATUS_MXR, STATUS_SUM, STATUS_MPRV;
logic [1:0] STATUS_MPP, STATUS_FS;
logic [1:0] PrivilegeModeW;
logic [P.XLEN-1:0] PTE;
logic [P.XLEN-1:0] PTE;
logic [1:0] PageType;
logic sfencevmaM;
logic SelHPTW;
// PMA checker signals
var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0];
var logic [P.PA_BITS-3:0] PMPADDR_ARRAY_REGW[P.PMP_ENTRIES-1:0];
var logic [7:0] PMPCFG_ARRAY_REGW[P.PMP_ENTRIES-1:0];
// IMem stalls
@ -116,14 +116,14 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
// cpu lsu interface
logic [2:0] Funct3M;
logic [P.XLEN-1:0] IEUAdrE;
logic [P.XLEN-1:0] WriteDataM;
logic [P.XLEN-1:0] IEUAdrM;
logic [P.LLEN-1:0] ReadDataW;
logic [P.XLEN-1:0] IEUAdrE;
logic [P.XLEN-1:0] WriteDataM;
logic [P.XLEN-1:0] IEUAdrM;
logic [P.LLEN-1:0] ReadDataW;
logic CommittedM;
// AHB ifu interface
logic [P.PA_BITS-1:0] IFUHADDR;
logic [P.PA_BITS-1:0] IFUHADDR;
logic [2:0] IFUHBURST;
logic [1:0] IFUHTRANS;
logic [2:0] IFUHSIZE;
@ -131,9 +131,9 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
logic IFUHREADY;
// AHB LSU interface
logic [P.PA_BITS-1:0] LSUHADDR;
logic [P.XLEN-1:0] LSUHWDATA;
logic [P.XLEN/8-1:0] LSUHWSTRB;
logic [P.PA_BITS-1:0] LSUHADDR;
logic [P.XLEN-1:0] LSUHWDATA;
logic [P.XLEN/8-1:0] LSUHWSTRB;
logic LSUHWRITE;
logic LSUHREADY;
@ -192,12 +192,12 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
.PCE, .PCLinkE, .FWriteIntE, .FCvtIntE, .IEUAdrE, .IntDivE, .W64E,
.Funct3E, .ForwardedSrcAE, .ForwardedSrcBE,
// Memory stage interface
.SquashSCW, // from LSU
.MemRWM, // read/write control goes to LSU
.AtomicM, // atomic control goes to LSU
.SquashSCW, // from LSU
.MemRWM, // read/write control goes to LSU
.AtomicM, // atomic control goes to LSU
.WriteDataM, // Write data to LSU
.Funct3M, // size and signedness to LSU
.SrcAM, // to privilege and fpu
.Funct3M, // size and signedness to LSU
.SrcAM, // to privilege and fpu
.RdE, .RdM, .FIntResM, .FlushDCacheM,
.BranchD, .BranchE, .JumpD, .JumpE,
// Writeback stage
@ -219,24 +219,24 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
.LSUHADDR, .HRDATA, .LSUHWDATA, .LSUHWSTRB, .LSUHSIZE,
.LSUHBURST, .LSUHTRANS, .LSUHWRITE, .LSUHREADY,
// connect to csr or privilege and stay the same.
.PrivilegeModeW, .BigEndianM, // connects to csr
.PMPCFG_ARRAY_REGW, // connects to csr
.PMPADDR_ARRAY_REGW, // connects to csr
.PrivilegeModeW, .BigEndianM, // connects to csr
.PMPCFG_ARRAY_REGW, // connects to csr
.PMPADDR_ARRAY_REGW, // connects to csr
// hptw keep i/o
.SATP_REGW, // from csr
.STATUS_MXR, // from csr
.STATUS_SUM, // from csr
.STATUS_MPRV, // from csr
.STATUS_MPP, // from csr
.sfencevmaM, // connects to privilege
.DCacheStallM, // connects to privilege
.LoadPageFaultM, // connects to privilege
.StoreAmoPageFaultM, // connects to privilege
.LoadMisalignedFaultM, // connects to privilege
.LoadAccessFaultM, // connects to privilege
.HPTWInstrAccessFaultF, // connects to privilege
.StoreAmoMisalignedFaultM, // connects to privilege
.StoreAmoAccessFaultM, // connects to privilege
.SATP_REGW, // from csr
.STATUS_MXR, // from csr
.STATUS_SUM, // from csr
.STATUS_MPRV, // from csr
.STATUS_MPP, // from csr
.sfencevmaM, // connects to privilege
.DCacheStallM, // connects to privilege
.LoadPageFaultM, // connects to privilege
.StoreAmoPageFaultM, // connects to privilege
.LoadMisalignedFaultM, // connects to privilege
.LoadAccessFaultM, // connects to privilege
.HPTWInstrAccessFaultF, // connects to privilege
.StoreAmoMisalignedFaultM, // connects to privilege
.StoreAmoAccessFaultM, // connects to privilege
.InstrUpdateDAF,
.PCSpillF, .ITLBMissF, .PTE, .PageType, .ITLBWriteF, .SelHPTW,
.LSUStallM);
@ -292,14 +292,14 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
.PMPCFG_ARRAY_REGW, .PMPADDR_ARRAY_REGW,
.FRM_REGW,.BreakpointFaultM, .EcallFaultM, .wfiM, .IntPendingM, .BigEndianM);
end else begin
assign CSRReadValW = 0;
assign CSRReadValW = 0;
assign UnalignedPCNextF = PC2NextF;
assign RetM = 0;
assign TrapM = 0;
assign wfiM = 0;
assign IntPendingM = 0;
assign sfencevmaM = 0;
assign BigEndianM = 0;
assign RetM = 0;
assign TrapM = 0;
assign wfiM = 0;
assign IntPendingM = 0;
assign sfencevmaM = 0;
assign BigEndianM = 0;
end
// multiply/divide unit
@ -310,45 +310,45 @@ module wallypipelinedcore import cvw::*; #(parameter cvw_t P) (
.MDUResultW, .DivBusyE);
end else begin // no M instructions supported
assign MDUResultW = 0;
assign DivBusyE = 0;
assign DivBusyE = 0;
end
// floating point unit
if (P.F_SUPPORTED) begin:fpu
fpu #(P) fpu(
.clk, .reset,
.FRM_REGW, // Rounding mode from CSR
.InstrD, // instruction from IFU
.ReadDataW(ReadDataW[P.FLEN-1:0]),// Read data from memory
.ForwardedSrcAE, // Integer input being processed (from IEU)
.StallE, .StallM, .StallW, // stall signals from HZU
.FlushE, .FlushM, .FlushW, // flush signals from HZU
.RdE, .RdM, .RdW, // which FP register to write to (from IEU)
.STATUS_FS, // is floating-point enabled?
.FRegWriteM, // FP register write enable
.FRM_REGW, // Rounding mode from CSR
.InstrD, // instruction from IFU
.ReadDataW(ReadDataW[P.FLEN-1:0]), // Read data from memory
.ForwardedSrcAE, // Integer input being processed (from IEU)
.StallE, .StallM, .StallW, // stall signals from HZU
.FlushE, .FlushM, .FlushW, // flush signals from HZU
.RdE, .RdM, .RdW, // which FP register to write to (from IEU)
.STATUS_FS, // is floating-point enabled?
.FRegWriteM, // FP register write enable
.FpLoadStoreM,
.ForwardedSrcBE, // Integer input for intdiv
.ForwardedSrcBE, // Integer input for intdiv
.Funct3E, .Funct3M, .IntDivE, .W64E, // Integer flags and functions
.FPUStallD, // Stall the decode stage
.FWriteIntE, .FCvtIntE, // integer register write enable, conversion operation
.FWriteDataM, // Data to be written to memory
.FIntResM, // data to be written to integer register
.FCvtIntResW, // fp -> int conversion result to be stored in int register
.FCvtIntW, // fpu result selection
.FDivBusyE, // Is the divide/sqrt unit busy (stall execute stage)
.IllegalFPUInstrD, // Is the instruction an illegal fpu instruction
.SetFflagsM, // FPU flags (to privileged unit)
.FPUStallD, // Stall the decode stage
.FWriteIntE, .FCvtIntE, // integer register write enable, conversion operation
.FWriteDataM, // Data to be written to memory
.FIntResM, // data to be written to integer register
.FCvtIntResW, // fp -> int conversion result to be stored in int register
.FCvtIntW, // fpu result selection
.FDivBusyE, // Is the divide/sqrt unit busy (stall execute stage)
.IllegalFPUInstrD, // Is the instruction an illegal fpu instruction
.SetFflagsM, // FPU flags (to privileged unit)
.FIntDivResultW);
end else begin // no F_SUPPORTED or D_SUPPORTED; tie outputs low
assign FPUStallD = 0;
assign FWriteIntE = 0;
assign FCvtIntE = 0;
assign FIntResM = 0;
assign FCvtIntW = 0;
assign FDivBusyE = 0;
end else begin // no F_SUPPORTED or D_SUPPORTED; tie outputs low
assign FPUStallD = 0;
assign FWriteIntE = 0;
assign FCvtIntE = 0;
assign FIntResM = 0;
assign FCvtIntW = 0;
assign FDivBusyE = 0;
assign IllegalFPUInstrD = 1;
assign SetFflagsM = 0;
assign FpLoadStoreM = 0;
assign SetFflagsM = 0;
assign FpLoadStoreM = 0;
end
endmodule

6
src/wally/wallypipelinedsoc.sv
View File

@ -48,9 +48,9 @@ module wallypipelinedsoc import cvw::*; #(parameter cvw_t P) (
output logic HREADY,
// I/O Interface
input logic TIMECLK, // optional for CLINT MTIME counter
input logic [31:0] GPIOIN, // inputs from GPIO
output logic [31:0] GPIOOUT, // output values for GPIO
output logic [31:0] GPIOEN, // output enables for GPIO
input logic [31:0] GPIOIN, // inputs from GPIO
output logic [31:0] GPIOOUT, // output values for GPIO
output logic [31:0] GPIOEN, // output enables for GPIO
input logic UARTSin, // UART serial data input
output logic UARTSout, // UART serial data output
input logic SDCCmdIn, // SDC Command input

12
testbench/common/wallyTracer.sv
View File

@ -139,6 +139,7 @@ module wallyTracer(rvviTrace rvvi);
CSRArray[12'h344] = testbench.dut.core.priv.priv.csr.csrm.MIP_REGW;
CSRArray[12'h304] = testbench.dut.core.priv.priv.csr.csrm.MIE_REGW;
CSRArray[12'h301] = testbench.dut.core.priv.priv.csr.csrm.MISA_REGW;
CSRArray[12'h30A] = testbench.dut.core.priv.priv.csr.csrm.MENVCFG_REGW;
CSRArray[12'hF14] = testbench.dut.core.priv.priv.csr.csrm.MHARTID_REGW;
CSRArray[12'h340] = testbench.dut.core.priv.priv.csr.csrm.MSCRATCH_REGW;
CSRArray[12'h342] = testbench.dut.core.priv.priv.csr.csrm.MCAUSE_REGW;
@ -157,6 +158,7 @@ module wallyTracer(rvviTrace rvvi);
CSRArray[12'h105] = testbench.dut.core.priv.priv.csr.csrs.csrs.STVEC_REGW;
CSRArray[12'h141] = testbench.dut.core.priv.priv.csr.csrs.csrs.SEPC_REGW;
CSRArray[12'h106] = testbench.dut.core.priv.priv.csr.csrs.csrs.SCOUNTEREN_REGW;
CSRArray[12'h10A] = testbench.dut.core.priv.priv.csr.csrs.csrs.SENVCFG_REGW;
CSRArray[12'h180] = testbench.dut.core.priv.priv.csr.csrs.csrs.SATP_REGW;
CSRArray[12'h140] = testbench.dut.core.priv.priv.csr.csrs.csrs.SSCRATCH_REGW;
CSRArray[12'h143] = testbench.dut.core.priv.priv.csr.csrs.csrs.STVAL_REGW;
@ -189,6 +191,7 @@ module wallyTracer(rvviTrace rvvi);
CSRArray[12'h344] = CSRArrayOld[12'h344];
CSRArray[12'h304] = CSRArrayOld[12'h304];
CSRArray[12'h301] = CSRArrayOld[12'h301];
CSRArray[12'h30A] = CSRArrayOld[12'h30A];
CSRArray[12'hF14] = CSRArrayOld[12'hF14];
CSRArray[12'h340] = CSRArrayOld[12'h340];
CSRArray[12'h342] = CSRArrayOld[12'h342];
@ -207,6 +210,7 @@ module wallyTracer(rvviTrace rvvi);
CSRArray[12'h105] = CSRArrayOld[12'h105];
CSRArray[12'h141] = CSRArrayOld[12'h141];
CSRArray[12'h106] = CSRArrayOld[12'h106];
CSRArray[12'h10A] = CSRArrayOld[12'h10A];
CSRArray[12'h180] = CSRArrayOld[12'h180];
CSRArray[12'h140] = CSRArrayOld[12'h140];
CSRArray[12'h143] = CSRArrayOld[12'h143];
@ -308,6 +312,7 @@ module wallyTracer(rvviTrace rvvi);
CSRArrayOld[12'h344] = CSRArray[12'h344];
CSRArrayOld[12'h304] = CSRArray[12'h304];
CSRArrayOld[12'h301] = CSRArray[12'h301];
CSRArrayOld[12'h30A] = CSRArray[12'h30A];
CSRArrayOld[12'hF14] = CSRArray[12'hF14];
CSRArrayOld[12'h340] = CSRArray[12'h340];
CSRArrayOld[12'h342] = CSRArray[12'h342];
@ -326,6 +331,7 @@ module wallyTracer(rvviTrace rvvi);
CSRArrayOld[12'h105] = CSRArray[12'h105];
CSRArrayOld[12'h141] = CSRArray[12'h141];
CSRArrayOld[12'h106] = CSRArray[12'h106];
CSRArrayOld[12'h10A] = CSRArray[12'h10A];
CSRArrayOld[12'h180] = CSRArray[12'h180];
CSRArrayOld[12'h140] = CSRArray[12'h140];
CSRArrayOld[12'h143] = CSRArray[12'h143];
@ -352,6 +358,7 @@ module wallyTracer(rvviTrace rvvi);
assign #2 CSR_W[12'h305] = (CSRArrayOld[12'h305] != CSRArray[12'h305]) ? 1 : 0;
assign #2 CSR_W[12'h341] = (CSRArrayOld[12'h341] != CSRArray[12'h341]) ? 1 : 0;
assign #2 CSR_W[12'h306] = (CSRArrayOld[12'h306] != CSRArray[12'h306]) ? 1 : 0;
assign #2 CSR_W[12'h30A] = (CSRArrayOld[12'h30A] != CSRArray[12'h30A]) ? 1 : 0;
assign #2 CSR_W[12'h320] = (CSRArrayOld[12'h320] != CSRArray[12'h320]) ? 1 : 0;
assign #2 CSR_W[12'h302] = (CSRArrayOld[12'h302] != CSRArray[12'h302]) ? 1 : 0;
assign #2 CSR_W[12'h303] = (CSRArrayOld[12'h303] != CSRArray[12'h303]) ? 1 : 0;
@ -374,6 +381,7 @@ module wallyTracer(rvviTrace rvvi);
assign #2 CSR_W[12'h105] = (CSRArrayOld[12'h105] != CSRArray[12'h105]) ? 1 : 0;
assign #2 CSR_W[12'h141] = (CSRArrayOld[12'h141] != CSRArray[12'h141]) ? 1 : 0;
assign #2 CSR_W[12'h106] = (CSRArrayOld[12'h106] != CSRArray[12'h106]) ? 1 : 0;
assign #2 CSR_W[12'h10A] = (CSRArrayOld[12'h10A] != CSRArray[12'h10A]) ? 1 : 0;
assign #2 CSR_W[12'h180] = (CSRArrayOld[12'h180] != CSRArray[12'h180]) ? 1 : 0;
assign #2 CSR_W[12'h140] = (CSRArrayOld[12'h140] != CSRArray[12'h140]) ? 1 : 0;
assign #2 CSR_W[12'h143] = (CSRArrayOld[12'h143] != CSRArray[12'h143]) ? 1 : 0;
@ -394,6 +402,7 @@ module wallyTracer(rvviTrace rvvi);
assign rvvi.csr_wb[0][0][12'h303] = CSR_W[12'h303];
assign rvvi.csr_wb[0][0][12'h344] = CSR_W[12'h344];
assign rvvi.csr_wb[0][0][12'h304] = CSR_W[12'h304];
assign rvvi.csr_wb[0][0][12'h30A] = CSR_W[12'h30A];
assign rvvi.csr_wb[0][0][12'h301] = CSR_W[12'h301];
assign rvvi.csr_wb[0][0][12'hF14] = CSR_W[12'hF14];
assign rvvi.csr_wb[0][0][12'h340] = CSR_W[12'h340];
@ -411,6 +420,7 @@ module wallyTracer(rvviTrace rvvi);
assign rvvi.csr_wb[0][0][12'h105] = CSR_W[12'h105];
assign rvvi.csr_wb[0][0][12'h141] = CSR_W[12'h141];
assign rvvi.csr_wb[0][0][12'h106] = CSR_W[12'h106];
assign rvvi.csr_wb[0][0][12'h10A] = CSR_W[12'h10A];
assign rvvi.csr_wb[0][0][12'h180] = CSR_W[12'h180];
assign rvvi.csr_wb[0][0][12'h140] = CSR_W[12'h140];
assign rvvi.csr_wb[0][0][12'h143] = CSR_W[12'h143];
@ -431,6 +441,7 @@ module wallyTracer(rvviTrace rvvi);
assign rvvi.csr[0][0][12'h303] = CSRArray[12'h303];
assign rvvi.csr[0][0][12'h344] = CSRArray[12'h344];
assign rvvi.csr[0][0][12'h304] = CSRArray[12'h304];
assign rvvi.csr[0][0][12'h30A] = CSRArray[12'h30A];
assign rvvi.csr[0][0][12'h301] = CSRArray[12'h301];
assign rvvi.csr[0][0][12'hF14] = CSRArray[12'hF14];
assign rvvi.csr[0][0][12'h340] = CSRArray[12'h340];
@ -448,6 +459,7 @@ module wallyTracer(rvviTrace rvvi);
assign rvvi.csr[0][0][12'h105] = CSRArray[12'h105];
assign rvvi.csr[0][0][12'h141] = CSRArray[12'h141];
assign rvvi.csr[0][0][12'h106] = CSRArray[12'h106];
assign rvvi.csr[0][0][12'h10A] = CSRArray[12'h10A];
assign rvvi.csr[0][0][12'h180] = CSRArray[12'h180];
assign rvvi.csr[0][0][12'h140] = CSRArray[12'h140];
assign rvvi.csr[0][0][12'h143] = CSRArray[12'h143];

34
testbench/testbench-fp.sv
View File

@ -113,8 +113,6 @@ module testbenchfp;
`include "parameter-defs.vh"
///////////////////////////////////////////////////////////////////////////////////////////////
// ||||||||| |||||||| ||||||| ||||||||| ||||||| |||||||| |||
@ -680,11 +678,11 @@ module testbenchfp;
// instantiate devices under test
if (TEST === "fma"| TEST === "mul" | TEST === "add" | TEST === "sub" | TEST === "all") begin : fma
fma #(P) fma(.Xs(Xs), .Ys(Ys), .Zs(Zs),
.Xe(Xe), .Ye(Ye), .Ze(Ze),
.Xm(Xm), .Ym(Ym), .Zm(Zm),
.XZero, .YZero, .ZZero, .Ss, .Se,
.OpCtrl(OpCtrlVal), .Sm, .InvA, .SCnt, .As, .Ps,
.ASticky);
.Xe(Xe), .Ye(Ye), .Ze(Ze),
.Xm(Xm), .Ym(Ym), .Zm(Zm),
.XZero, .YZero, .ZZero, .Ss, .Se,
.OpCtrl(OpCtrlVal), .Sm, .InvA, .SCnt, .As, .Ps,
.ASticky);
end
postprocess #(P) postprocess(.Xs(Xs), .Ys(Ys), .PostProcSel(UnitVal[1:0]),
@ -699,13 +697,13 @@ module testbenchfp;
.PostProcFlg(Flg), .PostProcRes(FpRes), .FCvtIntRes(IntRes));
if (TEST === "cvtfp" | TEST === "cvtint" | TEST === "all") begin : fcvt
fcvt #(P) fcvt (.Xs(Xs), .Xe(Xe), .Xm(Xm), .Int(SrcA), .ToInt(WriteIntVal),
fcvt fcvt (.Xs(Xs), .Xe(Xe), .Xm(Xm), .Int(SrcA), .ToInt(WriteIntVal),
.XZero(XZero), .OpCtrl(OpCtrlVal), .IntZero,
.Fmt(ModFmt), .Ce(CvtCalcExpE), .ShiftAmt(CvtShiftAmtE), .ResSubnormUf(CvtResSubnormUfE), .Cs(CvtResSgnE), .LzcIn(CvtLzcInE));
end
if (TEST === "cmp" | TEST === "all") begin: fcmp
fcmp #(P) fcmp (.Fmt(ModFmt), .OpCtrl(OpCtrlVal), .Xs, .Ys, .Xe, .Ye,
fcmp fcmp (.Fmt(ModFmt), .OpCtrl(OpCtrlVal), .Xs, .Ys, .Xe, .Ye,
.Xm, .Ym, .XZero, .YZero, .CmpIntRes(CmpRes),
.XNaN, .YNaN, .XSNaN, .YSNaN, .X, .Y, .CmpNV(CmpFlg[4]), .CmpFpRes(FpCmpRes));
end
@ -892,13 +890,13 @@ always @(negedge clk) begin
// check if result is correct
// - wait till the division result is done or one extra cylcle for early termination (to simulate the EM pipline stage)
// if(~((Res === Ans | NaNGood | NaNGood === 1'bx) & (ResFlg === AnsFlg | AnsFlg === 5'bx))&~((FDivBusyE===1'b1)|DivStart)&(UnitVal !== `CVTINTUNIT)&(UnitVal !== `CMPUNIT)) begin
ResMatch = (Res === Ans | NaNGood | NaNGood === 1'bx);
FlagMatch = (ResFlg === AnsFlg | AnsFlg === 5'bx);
divsqrtop = OpCtrlVal == `SQRT_OPCTRL | OpCtrlVal == `DIV_OPCTRL;
assign ResMatch = (Res === Ans | NaNGood | NaNGood === 1'bx);
assign FlagMatch = (ResFlg === AnsFlg | AnsFlg === 5'bx);
assign divsqrtop = OpCtrlVal == `SQRT_OPCTRL | OpCtrlVal == `DIV_OPCTRL;
assign DivDone = OldFDivBusyE & ~FDivBusyE;
//assign divsqrtop = OpCtrl[TestNum] == `SQRT_OPCTRL | OpCtrl[TestNum] == `DIV_OPCTRL;
CheckNow = (DivDone | ~divsqrtop) & (UnitVal !== `CVTINTUNIT)&(UnitVal !== `CMPUNIT);
assign CheckNow = (DivDone | ~divsqrtop) & (UnitVal !== `CVTINTUNIT)&(UnitVal !== `CMPUNIT);
if(~(ResMatch & FlagMatch) & CheckNow) begin
// if(~((Res === Ans | NaNGood | NaNGood === 1'bx) & (ResFlg === AnsFlg | AnsFlg === 5'bx))&(DivDone | (TEST != "sqrt" & TEST != "div"))&(UnitVal !== `CVTINTUNIT)&(UnitVal !== `CMPUNIT)) begin
errors += 1;
@ -974,10 +972,10 @@ module readvectors (
output logic [`FLEN-1:0] Ans,
output logic [`XLEN-1:0] SrcA,
output logic [4:0] AnsFlg,
output logic Xs, Ys, Zs, // sign bits of XYZ
output logic [`NE-1:0] Xe, Ye, Ze, // exponents of XYZ (converted to largest supported precision)
output logic [`NF:0] Xm, Ym, Zm, // mantissas of XYZ (converted to largest supported precision)
output logic XNaN, YNaN, ZNaN, // is XYZ a NaN
output logic Xs, Ys, Zs, // sign bits of XYZ
output logic [`NE-1:0] Xe, Ye, Ze, // exponents of XYZ (converted to largest supported precision)
output logic [`NF:0] Xm, Ym, Zm, // mantissas of XYZ (converted to largest supported precision)
output logic XNaN, YNaN, ZNaN, // is XYZ a NaN
output logic XSNaN, YSNaN, ZSNaN, // is XYZ a signaling NaN
output logic XSubnorm, ZSubnorm, // is XYZ denormalized
output logic XZero, YZero, ZZero, // is XYZ zero
@ -986,6 +984,7 @@ module readvectors (
output logic DivStart,
output logic [`FLEN-1:0] X, Y, Z, XPostBox
);
logic XEn, YEn, ZEn;
`include "parameter-defs.vh"
@ -1346,7 +1345,6 @@ module readvectors (
assign YEn = ~((Unit == `CVTINTUNIT)|(Unit == `CVTFPUNIT)|((Unit == `DIVUNIT)&OpCtrl[0]));
assign ZEn = (Unit == `FMAUNIT);
unpack #(P) unpack(.X, .Y, .Z, .Fmt(ModFmt), .Xs, .Ys, .Zs, .Xe, .Ye, .Ze,
.Xm, .Ym, .Zm, .XNaN, .YNaN, .ZNaN, .XSNaN, .YSNaN, .ZSNaN,
.XSubnorm, .XZero, .YZero, .ZZero, .XInf, .YInf, .ZInf,

2
testbench/testbench-linux-imperas.sv
View File

@ -866,10 +866,12 @@ module testbench;
"medeleg": `checkCSR(`CSR_BASE.csrm.MEDELEG_REGW)
"mepc": `checkCSR(`CSR_BASE.csrm.MEPC_REGW)
"mtval": `checkCSR(`CSR_BASE.csrm.MTVAL_REGW)
"menvcfg": `checkCSR(`CSR_BASE.csrm.MENVCFG_REGW)
"sepc": `checkCSR(`CSR_BASE.csrs.csrs.SEPC_REGW)
"scause": `checkCSR(`CSR_BASE.csrs.csrs.SCAUSE_REGW)
"stvec": `checkCSR(`CSR_BASE.csrs.csrs.STVEC_REGW)
"stval": `checkCSR(`CSR_BASE.csrs.csrs.STVAL_REGW)
"senvcfg": `checkCSR(`CSR_BASE.csrs.SENVCFG_REGW)
"mip": begin
`checkCSR(`CSR_BASE.csrm.MIP_REGW)
if(!NO_SPOOFING) begin

3
tests/coverage/csrwrites.S
View File

@ -33,5 +33,8 @@ main:
csrrw t0, satp, zero
csrrw t0, stvec, zero
csrrw t0, sscratch, zero
li t0, -2
csrrw t1, menvcfg, t0
csrrw t2, senvcfg, t0
j done

2
tests/wally-riscv-arch-test/riscv-test-suite/rv32i_m/privilege/references/WALLY-trap-01.reference_output
View File

@ -1,3 +1,5 @@
FFFFFFFF # stimecmp readback
80000000 # menvcfg readback
00000aaa # Test 5.3.1.4: readback value from writing mie to enable interrupts # skipping instruction address fault since they're impossible with compressed instrs enabled
00000001 # mcause from an instruction access fault
00000000 # mtval of faulting instruction address (0x0)

2
tests/wally-riscv-arch-test/riscv-test-suite/rv32i_m/privilege/src/WALLY-TEST-LIB-32.h
View File

@ -280,7 +280,7 @@ end_trap_triggers:
la t4, 0x02004000 // MTIMECMP register in CLINT
li t5, 0xFFFFFFFF
sw t5, 0(t4) // set mtimecmp to 0xFFFFFFFF to really make sure time interrupts don't go off immediately after being enabled
j trap_handler_end_\MODE\() // skip the trap handler when it is being defined.
// ---------------------------------------------------------------------------------------------

7
tests/wally-riscv-arch-test/riscv-test-suite/rv32i_m/privilege/src/WALLY-trap-01.S
View File

@ -23,7 +23,7 @@
#include "WALLY-TEST-LIB-32.h"
RVTEST_ISA("RV32I_Zicsr")
RVTEST_ISA("RV32I_Sstc_Zicsr")
RVTEST_CASE(0,"//check ISA:=regex(.*32.*);check ISA:=regex(.*I.*); def Drvtest_mtrap_routine=True;def TEST_CASE_1=True;def NO_SAIL=True;",trap)
INIT_TESTS
@ -34,8 +34,9 @@ TRAP_HANDLER m, EXT_SIGNATURE=1 // turn on recording mtval and status bits on tr
li x28, 0x8
csrs mstatus, x28 // set mstatus.MIE bit to 1
WRITE_READ_CSR mie, 0xFFF // Enable interrupts from all sources // *** commented out until I can get the trap handler (and spike for time interrupts) to work correctly with interrupts
WRITE_READ_CSR stimecmp, 0xFFFFFFFF // set timer to high value so it doesn't go off immediately
WRITE_READ_CSR menvcfgh, 0x80000000 // Enable menvcfg.STCE
WRITE_READ_CSR mie, 0xFFF // Enable interrupts from all sources
// test 5.3.1.4 Basic trap tests
// instr address misaligned instructions are excluded from this test since they are impossible to cause when compressed instructions are enabled.

4
tests/wally-riscv-arch-test/riscv-test-suite/rv64i_m/privilege/references/WALLY-trap-01.reference_output
View File

@ -1,3 +1,7 @@
FFFFFFFF # stimecmp low bits
00000000 # stimecmp high bits
00000000 # menvcfg low bits
80000000 # menvcfg high bits
00000aaa # Test 5.3.1.4: readback value from writing mie to enable interrupts
00000000 # skipping instruction address fault since they're impossible with compressed instrs enabled
00000001 # mcause from an instruction access fault

2
tests/wally-riscv-arch-test/riscv-test-suite/rv64i_m/privilege/src/WALLY-TEST-LIB-64.h
View File

@ -274,7 +274,7 @@ end_trap_triggers:
la t4, 0x02004000 // MTIMECMP register in CLINT
li t5, 0xFFFFFFFF
sd t5, 0(t4) // set mtimecmp to 0xFFFFFFFF to really make sure time interrupts don't go off immediately after being enabled
j trap_handler_end_\MODE\() // skip the trap handler when it is being defined.
// ---------------------------------------------------------------------------------------------

6
tests/wally-riscv-arch-test/riscv-test-suite/rv64i_m/privilege/src/WALLY-trap-01.S
View File

@ -22,7 +22,7 @@
///////////////////////////////////////////
#include "WALLY-TEST-LIB-64.h"
RVTEST_ISA("RV64I_Zicsr")
RVTEST_ISA("RV64I_Sstc_Zicsr")
RVTEST_CASE(0,"//check ISA:=regex(.*64.*);check ISA:=regex(.*I.*);def TEST_CASE_1=True;def NO_SAIL=True;",trap)
INIT_TESTS
@ -33,7 +33,9 @@ TRAP_HANDLER m, EXT_SIGNATURE=1 // turn on recording mtval and status bits on tr
li x28, 0x8
csrs mstatus, x28 // set mstatus.MIE bit to 1
WRITE_READ_CSR mie, 0xFFF // Enable interrupts from all sources
WRITE_READ_CSR stimecmp, 0xFFFFFFFF // set timer to high value so it doesn't go off immediately
WRITE_READ_CSR menvcfg, 0x8000000000000000 // Enable menvcfg.STCE
WRITE_READ_CSR mie, 0xFFF // Enable interrupts from all sources
// test 5.3.1.4 Basic trap tests