Consolidate and/andc/nand, or/orc/nor and xor/eqv, using a common
invert on the input and output. This saves us about 200 LUTs.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
The current scheme has UART0 repeating throughout the UART address range.
This patch tightens the address checking so that it only occurs once.
Signed-off-by: Alastair D'Silva <alastair@d-silva.org>
This adds support for set associativity to the icache. It can still
be direct mapped by setting NUM_WAYS to 1.
The replacement policy uses a simple tree-PLRU for each set.
This is only lightly tested, tests pass but I have to double check
that we are using the ways effectively and not creating duplicates.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We only ever access the cache memory for at most the wishbone bus
width at a time. So having the BRAMs organized as a cache-line-wide
port is a waste of resources.
Instead, use a wishbone-wide memory and store a line as consecutive
rows in the BRAM.
This significantly improves BRAM usage in the FPGA as we can now use
more rows in the BRAM blocks. It also saves a few LUTs and muxes.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The goal is to have the icache fit in BRAM by latching the output
into a register. In order to avoid timing issues , we need to give
the BRAM a full cycle on reads, and thus we souce the BRAM address
directly from fetch1 latched NIA.
(Note: This will be problematic if/when we want to hash the address,
we'll probably be better off having fetch1 latch a fully hashed address
along with the normal one, so the icache can use the former to address
the BRAM and pass the latter along)
One difficulty is that we cannot really stall the icache without adding
more combo logic that would break the "one full cycle" BRAM model. This
means that on stalls from decode, by the time we stall fetch1, it has
already gone to the next address, which the icache is already latching.
We work around this by having a "stash" buffer in fetch2 that will stash
away the icache output on a stall, and override the output of the icache
with the content of the stash buffer when unstalling.
This requires a rewrite of the stop/step debug logic as well. We now
do most of the hard work in fetch1 which makes more sense.
Note: Vivado is still not inferring an built-in output register for the
BRAMs. I don't want to add another cycle... I don't fully understand why
it wouldn't be able to treat current_row as such but clearly it won't. At
least the timing seems good enough now for 100Mhz, possibly more.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This adds a new entity 'rotator' which contains combinatorial logic
for rotating and masking 64-bit values. It implements the operations
of the rlwinm, rlwnm, rlwimi, rldicl, rldicr, rldic, rldimi, rldcl,
rldcr, sld, slw, srd, srw, srad, sradi, sraw and srawi instructions.
It consists of a 3-stage 64-bit rotator using 4:1 multiplexors at
each stage, two mask generators, output logic and control logic.
The insn_type_t values used for these instructions have been reduced
to just 5: OP_RLC, OP_RLCL and OP_RLCR for the rotate and mask
instructions (clear both left and right, clear left, clear right
variants), OP_SHL for left shifts, and OP_SHR for right shifts.
The control signals for the rotator are derived from the opcode
and from the is_32bit and is_signed fields of the decode_rom_t.
The rotator is instantiated as an entity in execute1 so that we can
be sure we only have one of it.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This changes the names of the mul_32bit and mul_signed fields of
decode_rom_t to is_32bit and is_signed, so they can be used with
other types of operations besides multiplies.
This plumbs the is_32bit and is_signed flags down into execute1,
though they are not used at this point.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This aims to simplify the logic between the instruction image and
the register file read address ports and reduce the size of the decode
tables. With this patch, the input_reg_a column of the decode tables
can only select RA or zeroes, the input_reg_b column can only select
RB or a constant (0, -1, or an immediate value from the instruction),
and the input_reg_c columns can only select RS or zeroes.
That means that the rotate/shift/logical ops now have their first
input coming in via the input_reg_c column. That means we need to
add a read_data3 field to the Decode2ToExecuteType record, but that
will go away again when we split out the rotate/mask/logical ops to
their own unit.
As a related but not tightly connected change, this patch also sets
the read1_enable signal to the register file be 0 when RA=0 and the
input_reg_a for the instruction is RA_OR_ZERO (previously it was 1).
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
All of the PPC add and subtract instructions, including carrying
and extended versions, do much the same arithmetic operation:
result = (I xor A) + B + C
where A is the value from RA, I provides a logical inversion of A
(i.e. I is 0 or -1), B is either from RB or is a constant 0 or -1,
and C is 0, 1 or the carry bit from XER (CA).
To consolidate all the add/subtract instructions into a single
OP_ADD, we add a column to decode_rom_t to indicate when A should
be inverted, and change the input_carry field to a 3-state selector
to select C in the equation above.
This also adds a new "CONST_M1" value for input_reg_b_t to indicate
that B is a constant -1. This allows us to implement addme and
subfme.
The addex instruction appears not to exist, so the comments referring
to it are removed.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Experimentation on POWER9 indicates that the invalid form of lbzux
with RA=0 uses just RB as the address, not R0 + RB. Extrapolating
this to all update-form loads and stores with RA=0, change all the
update-form loads and stores to use RA_OR_ZERO rather than RA.
This then means that all decode ROM entries with insn_type = LDST
have input_reg_a = RA_OR_ZERO.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
The register file is currently implemented as a whole pile of individual
1-bit registers instead of LUT memory which is a huge waste of FPGA
space.
This is caused by the output signal exposing the register file to the
outside world for simulation debug.
This removes that output, and moves the dumping of the register file
to the register file module itself. This saves about 8% of fpga on
the little Arty A7-35T.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
The const* fields of decode_rom_t drove multiplexers in decode2 that
picked out various instruction fields and put them into the const*
fields of the Decode2ToExecute1Type record, from where they were
used in execute1. However, the code in execute1 can just as easily
use the appropriate fields of the original instruction word, since
that is now available in execute1. This therefore changes the
code to do that, resulting in smaller decode tables.
Suggested-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
The l?arx and st?cx. instructions are defined to use the normal indexed
mode address calculations, i.e. (RA|0) + RB. Fix their entries in the
decode table to say RA_OR_ZERO rather than RA.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This changes decode_op_31_array from being indexed by a ppc_insn_t
(which is derived from the instruction word by a whole series of
if/elsif statements) to being indexed directly by bits 10...1 of
the instruction word. With this we no longer need ppc_insn.
This then means that the decode1 stage doesn't distinguish between
mfcr and mfocrf, or between mtcrf and mtocrf, since those are
distinguished by the value in bit 20 of the instruction. To
accommodate that, execute1 changes so that the one op value (OP_MFCR)
does either the mfcr or the mfocrf behaviour depending on bit 20
of the instruction word; and similarly for mtcrf/mtocrf.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This comprises the 64-bit rotate and mask instructions. In order to
reduce the table index to 3 bits, we combine rldcl and rdlcr into a
single op (OP_RLDCX), and choose the right mask at execute time based
on bit 1 of the instruction word.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This changes the decoding of major opcode 19 from using the ppc_insn_t
index to using bits of the instruction word directly. Opcode 19 has
a 10-bit minor opcode field (bits 10..1) but the space is sparsely
filled. Therefore we index a table of single-bit entries with the
10-bit minor opcode to filter out the illegal minor opcodes, and
index a table using just 3 bits -- 5, 3 and 2 -- of the instruction
to get the decode entry. This groups together all the instructions
in 4 columns of the opcode map as a single entry. That means that
mcrf and all the CR logical ops get grouped together, and bcctr, bclr
and bctar get grouped together. At present the CR logical ops are not
implemented, so their grouping has no impact.
The code for bclr and bcctr in execute1 is now common, using a single
op, and it now determines the branch address by looking at bit 10 of
the instruction word at execute time.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
With this, we have a table for most major opcodes and separate
tables for each major opcode that has further decoding required.
These tables are still mostly indexed by the ppc_insn_t values,
however.
A few things are still decoded completely at the top level: nop,
attn and sim_config.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Instead of doing mfctr, mflr, mftb, mtctr, mtlr as separate ops,
just pass down mfspr and mtspr ops with the spr number and let
execute1 decode which SPR we're addressing. This will help reduce
the number of instruction bits decode1 needs to look at.
In fact we now pass down the whole instruction from decode2 to
execute1. We will need more bits of the instruction in future,
and the tools should just optimize away any that we don't end
up using. Since the 'aa' bit was just a copy of an instruction
bit, we can now remove it from the record.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Hopefully it's not too timing catastrophic. The variable newcrf will
be handy for the other CR ops when we implement them I suspect.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We can now pass both the input clock and target clock frequency
via generics. Add support for both 50Mhz and 100Mhz target freqs
for both cases.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This seems dependent on the FPGA type/size, so we should probably
make it a toplevel generic, but for now this helps on the
Arty A7-35
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
This moves the data formatting for read data to after a register,
instead of before, in order to improve timing. The data formatting
is now effectively combinational logic on the input side of the
writeback stage.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
It's always set when f_out.redirect is set, so may as well set it once
at the end. It's all combo from the register.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Do the +4 in a single place. This shouldn't cause any difference
in behaviour as these are sequential variable assignments.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
These are a copy of the A7-35 definitions with 35 changed to 100.
The A7-100 uses the same .xdc file (arty_a7-35.xdc) as the A7-35
since the only difference between the two is the FPGA part; the
hardware and connections on the two boards are identical.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This moves the negation of negative operands for signed divide and
modulus operations out of the decode2 stage and into the divider.
If either of the operands for a signed divide or modulus operation
is negative, the divider now takes an extra cycle to negate the
operands that are negative.
The interface to the divider now has an 'is_signed' signal rather
than a 'neg_result' signal, and the dividend and divisor can be
negative, so divider_tb had to be updated for the new interface.
The reason for doing this is that one of the worst timing violations
on the Arty A7-100 at 100MHz involved the carry chain in the adders
that did the negation of the dividend and divisor in the decode stage.
Moving the negations to a separate cycle fixes that and also seems to
reduce the total number of slice LUTs used.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
These are intended to be combinatorial. The previous code was giving
warnings in vivado about registers/latches with no clock defined.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Anton Blanchard
Paul Mackerras