This puts all the data formatting (byte rotation based on lowest three
bits of the address, byte reversal, sign extension, zero extension)
in loadstore1. Writeback now simply sends the data provided to the
register files.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
So that the dcache could in future be used by an MMU, this moves
logic to do with data formatting, rA updates for update-form
instructions, and handling of unaligned loads and stores out of
dcache and into loadstore1. For now, dcache connects only to
loadstore1, and loadstore1 now has the connection to writeback.
Dcache generates a stall signal to loadstore1 which indicates that
the request presented in the current cycle was not accepted and
should be presented again. However, loadstore1 doesn't currently
use it because we know that we can never hit the circumstances
where it might be set.
For unaligned transfers, loadstore1 generates two requests to
dcache back-to-back, and then waits to see two acks back from
dcache (cycles where d_in.valid is true).
Loadstore1 now has a FSM for tracking how many acks we are
expecting from dcache and for doing the rA update cycles when
necessary. Handling for reservations and conditional stores is
still in dcache.
Loadstore1 now generates its own stall signal back to decode2,
so we no longer need the logic in execute1 that generated the stall
for the first two cycles.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Since we removed one cycle from the load hit case, we actually no
longer need the extra cycle provided by having the LOAD_UPDATE
state. Therefore this makes the load hit case in the IDLE and
NEXT_DWORD states go to LOAD_UPDATE2 rather than LOAD_UPDATE.
Then we remove LOAD_UPDATE and then rename LOAD_UPDATE2 to
LOAD_UPDATE.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Currently we don't get the result from a load that hits in the dcache
until the fourth cycle after the instruction was presented to
loadstore1. This trims this back to 3 cycles by taking the low order
bits of the address generated in loadstore1 into dcache directly (not
via the output register of loadstore1) and using them to address the
read port of the dcache data RAM. We use the lower 12 address bits
here in the expectation that any reasonable data cache design will
have a set size of 4kB or less in order to avoid the aliasing problems
that can arise with a virtually-indexed physically-tagged cache if
the set size is greater than the smallest page size provided by the
MMU.
With this we can get rid of r2 and drive the signals going to
writeback from r1, since the load hit data is now available one
cycle earlier. We need a multiplexer on the read address of the
data cache RAM in order to handle the second doubleword of an
unaligned access.
One small complication is that we now need an extra cycle in the case
of an unaligned load which misses in the data cache and which reads
the 2nd-last and last doublewords of a cache line. This is the reason
for the PRE_NEXT_DWORD state; if we just go straight to NEXT_DWORD
then we end up having the write of the last doubleword of the cache
line and the read of that same doubleword occurring in the same
cycle, which means we read stale data rather than the just-fetched
data.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
From:
commit 8e0389b973
Author: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Date: Wed Oct 23 12:08:55 2019 +1100
ram: Rework main RAM interface
We need to change the name.
Signed-off-by: Michael Neuling <mikey@neuling.org>
This involves plumbing the (existing) 'reserve' and 'rc' bits in
the decode tables down to dcache, and 'rc' and 'store_done' bits
from dcache to writeback.
It turns out that we had 'RC' set in the 'rc' column for several
ordinary stores and for the attn instruction. This corrects them
to 'NONE', and sets the 'rc' column to 'ONE' for the conditional
stores.
In writeback we now have logic to set CR0 when the input from dcache
has rc = 1.
In dcache we have the reservation itself, which has a valid bit
and the address down to cache line granularity. We don't currently
store the reservation length. For a store conditional which fails,
we set a 'cancel_store' signal which inhibits the write to the
cache and prevents the state machine from starting a bus cycle or
going to the STORE_WAIT_ACK state. Instead we set r1.stcx_fail
which causes the instruction to complete in the next cycle with
rc=1 and store_done=0.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
For an unaligned load or store, we do the first doubleword (dword) of
the transfer as normal, but then go to a new NEXT_DWORD state of the
state machine to do the cache tag lookup for the second dword of the
transfer. From the NEXT_DWORD state we have much the same transitions
to other states as from the IDLE state (the transitions for OP_LOAD_HIT
are a bit different but almost identical for the other op values).
We now do the preparation of the data to be written in loadstore1,
that is, byte reversal if necessary and rotation by a number of
bytes based on the low 3 bits of the address. We do rotation not
shifting so we have the bytes that need to go into the second
doubleword in the right place in the low bytes of the data sent to
dcache. The rotation and byte reversal are done in a single step
with one multiplexer per byte by setting the select inputs for each
byte appropriately.
This also fixes writeback to not write the register value until it
has received both pieces of an unaligned load value.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
The obscure bug is that a non-cacheable load with update would never
do the update and would never complete the instruction. This is fixed
by making state NC_LOAD_WAIT_ACK go to LOAD_UPDATE2 if r1.req.update
is set.
The slow load forms with update can go to LOAD_UPDATE2 at the end
rather than LOAD_UPDATE, thus saving a cycle. Loads with a cache
hit need the LOAD_UPDATE state in the third cycle since they are
not writing back until the 4th cycle, when the state is LOAD_UPDATE2.
Slow loads (cacheable loads that miss and non-cacheable loads)
currently go to LOAD_UPDATE in the cycle after they see
r1.wb.ack = 1 for the last time, but that cycle is the cycle where
they write back, and the following cycle does nothing. Going to
LOAD_UPDATE2 in those cases saves a cycle and makes them consistent
with the load hit case.
The logic in the RELOAD_WAIT_ACK case doesn't need to check
r1.req.load = '1' since we only ever use RELOAD_WAIT_ACK for loads.
There are also some whitespace fixes and a typo fix.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This removes the constraint that loads and stores are single-issue,
at the expense of a stall of at least 2 cycles for every load and
store.
To do this, we plumb the existing stall signal that was generated
in dcache to core, where it gets ORed with the stall signal from
execute1. Execute1 generates a stall signal for the first two
cycles of each load and store, and dcache generates the stall
signal in the 3rd and subsequent cycles if it needs to.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
It turns out that CR logical instructions have the truth table of
the operation embedded in the instruction word. This means that we
can collect the two input operand bits into a 2-bit value and use
that as the index to select the appropriate bit from the instruction
word.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Diamond doesn't like the "" & method of converting std_logic to a single bit
std_logic_vector. Thanks to Olof Kindgren for this patch.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
No need to open code a bunch of random termios flags, use cfmakeraw().
Most of the time we want ctrl-c to exit the current simulation, so
make that the default.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
A number of people have tripped up on GHDL backend issues. We
require either the LLVM or gcc backend, because the mcode backend
can't link against libraries which we use for simulated memory
and UART. Make that clearer, and point people at the Docker images
if they are having issues building ghdl.
Also point at the prebuilt micropython image.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
Some distros don't have a version of ghdl with the LLVM or GCC backend,
so add a Docker image as an alternative.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
A first pass at ghdl synthesis using yosys and nextpnr. It runs hello world
or micropython if the FPGA has enough block RAM (eg ECP5 85F). The hello
world testcase also loops UART rx to tx in software (ie not a hardware
loopback).
It uses Docker images, so no software needs to be installed. If you prefer
podman you can use that too. Edit Makefile.synth to configure your FPGA,
JTAG device etc.
To build:
make -f Makefile.synth
and to program:
make -f Makefile.synth prog
A few issues:
We need to add PLL support. Right now Microwatt runs at whatever the
external clock frequency is and the baud rate gets scaled by how far off
50MHz it is. This means on the ecp5-evn with a 12 MHz clock rate the baud
rate is a quite strange 27650 (115200 * 50 / 12). On my OrangeCrab with a
50MHz clock the UART is 115200.
It uses a large amount of resources, way more than it should. There are
still some ghdl/yosys issues to be sorted out.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
GHDL doesn't seem to have a way to specify the location of the object
file it writes, so right now they are all ending up in the root
directory. The Makefile rules did not reflect that, so make would
continually the files in fpga/*
Fix the rules to match what GHDL is doing.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
dependencies.py was pretty terrible at actually determining
dependencies. This improves it and also adds a --synth option.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
Micropython has been able to fit into 384kB for ages, so lets reduce our
simulated RAM. This is useful for testing if micropython will run on an
ECP5 85k, which has enough BRAM for 384kB but not enough for 512kB.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
We might want a non power of 2 amount of RAM in order to fit into an
FPGA, so create log2ceil and use it when calculating the number of
memory bits.
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
The current version of micropython in tests/micropython.bin is ancient.
Bug #135 points out that more recent versions are much smaller and they
also handle restart when ctrl+D is pressed.
Save all three versions of the file (elf, bin and hex) in micropython/
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
Shrink hello_world a bit (from 12kB to 8kB).
Include the built images
Add 0x10 and 0x100 entry points
Signed-off-by: Anton Blanchard <anton@linux.ibm.com>
This seems just to have been missed in commit f291efa266 ("decode1:
Mark ALU ops using carry as pipelined").
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This adds a register in the middle of the countzero computation,
so that we now have two cycles to count leading or trailing zeroes
instead of just one. Execute1 now outputs a one-cycle stall signal
when it encounters a cntlz* or cnttz* instruction. With this,
the countzero path no longer fails timing on the Artix-7 at 100MHz.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This allows us to use the bypass at the input of execute1 for the
address and data operands for loadstore1.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This enables back-to-back execution of integer instructions where
the first instruction writes a GPR and the second reads the same
GPR. This is done with a set of multiplexers at the start of
execute1 which enable any of the three input operands to be taken
from the output of execute1 (i.e. r.e.write_data) rather than the
input from decode2 (i.e. e_in.read_data[123]).
This also requires changes to the hazard detection and handling.
Decode2 generates a signal indicating that the GPR being written
is available for bypass, which is true for instructions that are
executed in execute1 (rather than loadstore1/dcache). The
gpr_hazard module stores this "bypassable" bit, and if the same
GPR needs to be read by a subsequent instruction, it outputs a
"use_bypass" signal rather than generating a stall. The
use_bypass signal is then latched at the output of decode2 and
passed down to execute1 to control the input multiplexer.
At the moment there is no bypass on the inputs to loadstore1, but that
is OK because all load and store instructions are marked as
single-issue.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This implements logic in the logical entity to calculate the results
of the popcnt* and prty* instructions. We now have one insn_type_t
value for the 3 popcnt variants and one for the two prty variants,
using the length field of the decode_rom_t to distinguish between
them. The implementations in logical.vhdl using recursive
algorithms rather than the simple functions in ppc_fx_insns.vhdl.
This gives a saving of about 140 slice LUTs on the A7-100 and
improves timing slightly.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This handles OP_CMP like a subtraction; the main adder computes
~RA + RB + 1, and the condition codes are computed from the results.
A direct comparison of the two input operands is used to calculate the
EQ bit of the condition result. The LT and GT bits are computed from
the MSB of the subtraction result, the carry out from the subtraction,
and the MSBs of the operands. For a 32-bit comparison, the 32-bit
carry and bit 31 of the result and input operands are used; for a
64-bit comparison, the 64-bit carry and bit 63 of the operands and
result are used.
It turns out to be more convenient to use the 'signed' field of
the decode table to distinguish signed from unsigned comparisons,
rather than the insn_type. Therefore this uses OP_CMP for both
cmp and cmpl, which also has the benefit of reducing the number
of values in insn_type_t.
Doing this saves over 200 slice LUTs on the Arty A7-100 and improves
timing slightly as well.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This moves the sign extension done by the extsb, extsh and extsw
instructions back into execute1. This means that we no longer need
any data formatting in writeback for results coming from execute1,
so this modifies writeback so the data formatter inputs come
directly from the loadstore unit output. The condition code
updates for RC=1 form instructions are now done on the value from
execute1 rather than the output of the data formatter, which should
help timing.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
For multiply and divide operations, execute1 now records the
destination GPR number, RC and OE from the instruction, and the
XER value. This means that the multiply and divide units don't
need to record those values and then send them back to execute1.
This makes the interface to those units a bit simpler. They
simply report an overflow signal along with the result value, and
execute1 takes care of updating XER if necessary.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
With this, the divider is a unit that execute1 sends operands to and
which sends its results back to execute1, which then send them to
writeback. Execute1 now sends a stall signal when it gets a divide
or modulus instruction until it gets a valid signal back from the
divider. Divide and modulus instructions are no longer marked as
single-issue.
The data formatting step that used to be done in decode2 for div
and mod instructions is now done in execute1. We also do the
absolute value operation in that same cycle instead of taking an
extra cycle inside the divider for signed operations with a
negative operand.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Paul Mackerras
Anton Blanchard
Michael Neuling