src/nmutil/singlepipe.py

   1 """ Pipeline API.  For multi-input and multi-output variants, see multipipe.
   2
   3     Associated development bugs:
   4     * http://bugs.libre-riscv.org/show_bug.cgi?id=64
   5     * http://bugs.libre-riscv.org/show_bug.cgi?id=57
   6
   7     Important: see Stage API (stageapi.py) in combination with below
   8
   9     RecordBasedStage:
  10     ----------------
  11
  12     A convenience class that takes an input shape, output shape, a
  13     "processing" function and an optional "setup" function.  Honestly
  14     though, there's not much more effort to just... create a class
  15     that returns a couple of Records (see ExampleAddRecordStage in
  16     examples).
  17
  18     PassThroughStage:
  19     ----------------
  20
  21     A convenience class that takes a single function as a parameter,
  22     that is chain-called to create the exact same input and output spec.
  23     It has a process() function that simply returns its input.
  24
  25     Instances of this class are completely redundant if handed to
  26     StageChain, however when passed to UnbufferedPipeline they
  27     can be used to introduce a single clock delay.
  28
  29     ControlBase:
  30     -----------
  31
  32     The base class for pipelines.  Contains previous and next ready/valid/data.
  33     Also has an extremely useful "connect" function that can be used to
  34     connect a chain of pipelines and present the exact same prev/next
  35     ready/valid/data API.
  36
  37     Note: pipelines basically do not become pipelines as such until
  38     handed to a derivative of ControlBase.  ControlBase itself is *not*
  39     strictly considered a pipeline class.  Wishbone and AXI4 (master or
  40     slave) could be derived from ControlBase, for example.
  41     UnbufferedPipeline:
  42     ------------------
  43
  44     A simple stalling clock-synchronised pipeline that has no buffering
  45     (unlike BufferedHandshake).  Data flows on *every* clock cycle when
  46     the conditions are right (this is nominally when the input is valid
  47     and the output is ready).
  48
  49     A stall anywhere along the line will result in a stall back-propagating
  50     down the entire chain.  The BufferedHandshake by contrast will buffer
  51     incoming data, allowing previous stages one clock cycle's grace before
  52     also having to stall.
  53
  54     An advantage of the UnbufferedPipeline over the Buffered one is
  55     that the amount of logic needed (number of gates) is greatly
  56     reduced (no second set of buffers basically)
  57
  58     The disadvantage of the UnbufferedPipeline is that the valid/ready
  59     logic, if chained together, is *combinatorial*, resulting in
  60     progressively larger gate delay.
  61
  62     PassThroughHandshake:
  63     ------------------
  64
  65     A Control class that introduces a single clock delay, passing its
  66     data through unaltered.  Unlike RegisterPipeline (which relies
  67     on UnbufferedPipeline and PassThroughStage) it handles ready/valid
  68     itself.
  69
  70     RegisterPipeline:
  71     ----------------
  72
  73     A convenience class that, because UnbufferedPipeline introduces a single
  74     clock delay, when its stage is a PassThroughStage, it results in a Pipeline
  75     stage that, duh, delays its (unmodified) input by one clock cycle.
  76
  77     BufferedHandshake:
  78     ----------------
  79
  80     nmigen implementation of buffered pipeline stage, based on zipcpu:
  81     https://zipcpu.com/blog/2017/08/14/strategies-for-pipelining.html
  82
  83     this module requires quite a bit of thought to understand how it works
  84     (and why it is needed in the first place).  reading the above is
  85     *strongly* recommended.
  86
  87     unlike john dawson's IEEE754 FPU STB/ACK signalling, which requires
  88     the STB / ACK signals to raise and lower (on separate clocks) before
  89     data may proceeed (thus only allowing one piece of data to proceed
  90     on *ALTERNATE* cycles), the signalling here is a true pipeline
  91     where data will flow on *every* clock when the conditions are right.
  92
  93     input acceptance conditions are when:
  94         * incoming previous-stage strobe (p.valid_i) is HIGH
  95         * outgoing previous-stage ready   (p.ready_o) is LOW
  96
  97     output transmission conditions are when:
  98         * outgoing next-stage strobe (n.valid_o) is HIGH
  99         * outgoing next-stage ready   (n.ready_i) is LOW
 100
 101     the tricky bit is when the input has valid data and the output is not
 102     ready to accept it.  if it wasn't for the clock synchronisation, it
 103     would be possible to tell the input "hey don't send that data, we're
 104     not ready".  unfortunately, it's not possible to "change the past":
 105     the previous stage *has no choice* but to pass on its data.
 106
 107     therefore, the incoming data *must* be accepted - and stored: that
 108     is the responsibility / contract that this stage *must* accept.
 109     on the same clock, it's possible to tell the input that it must
 110     not send any more data.  this is the "stall" condition.
 111
 112     we now effectively have *two* possible pieces of data to "choose" from:
 113     the buffered data, and the incoming data.  the decision as to which
 114     to process and output is based on whether we are in "stall" or not.
 115     i.e. when the next stage is no longer ready, the output comes from
 116     the buffer if a stall had previously occurred, otherwise it comes
 117     direct from processing the input.
 118
 119     this allows us to respect a synchronous "travelling STB" with what
 120     dan calls a "buffered handshake".
 121
 122     it's quite a complex state machine!
 123
 124     SimpleHandshake
 125     ---------------
 126
 127     Synchronised pipeline, Based on:
 128     https://github.com/ZipCPU/dbgbus/blob/master/hexbus/rtl/hbdeword.v
 129 """
 130
 131 from nmigen import Signal, Mux, Module, Elaboratable, Const
 132 from nmigen.cli import verilog, rtlil
 133 from nmigen.hdl.rec import Record
 134
 135 from nmutil.queue import Queue
 136 import inspect
 137
 138 from nmutil.iocontrol import (PrevControl, NextControl, Object, RecordObject)
 139 from nmutil.stageapi import (_spec, StageCls, Stage, StageChain, StageHelper)
 140 from nmutil import nmoperator
 141
 142
 143 class RecordBasedStage(Stage):
 144     """ convenience class which provides a Records-based layout.
 145         honestly it's a lot easier just to create a direct Records-based
 146         class (see ExampleAddRecordStage)
 147     """
 148     def __init__(self, in_shape, out_shape, processfn, setupfn=None):
 149         self.in_shape = in_shape
 150         self.out_shape = out_shape
 151         self.__process = processfn
 152         self.__setup = setupfn
 153     def ispec(self): return Record(self.in_shape)
 154     def ospec(self): return Record(self.out_shape)
 155     def process(seif, i): return self.__process(i)
 156     def setup(seif, m, i): return self.__setup(m, i)
 157
 158
 159 class PassThroughStage(StageCls):
 160     """ a pass-through stage with its input data spec identical to its output,
 161         and "passes through" its data from input to output (does nothing).
 162
 163         use this basically to explicitly make any data spec Stage-compliant.
 164         (many APIs would potentially use a static "wrap" method in e.g.
 165          StageCls to achieve a similar effect)
 166     """
 167     def __init__(self, iospecfn): self.iospecfn = iospecfn
 168     def ispec(self): return self.iospecfn()
 169     def ospec(self): return self.iospecfn()
 170
 171
 172 class ControlBase(StageHelper, Elaboratable):
 173     """ Common functions for Pipeline API.  Note: a "pipeline stage" only
 174         exists (conceptually) when a ControlBase derivative is handed
 175         a Stage (combinatorial block)
 176
 177         NOTE: ControlBase derives from StageHelper, making it accidentally
 178         compliant with the Stage API.  Using those functions directly
 179         *BYPASSES* a ControlBase instance ready/valid signalling, which
 180         clearly should not be done without a really, really good reason.
 181     """
 182     def __init__(self, stage=None, in_multi=None, stage_ctl=False, maskwid=0):
 183         """ Base class containing ready/valid/data to previous and next stages
 184
 185             * p: contains ready/valid to the previous stage
 186             * n: contains ready/valid to the next stage
 187
 188             Except when calling Controlbase.connect(), user must also:
 189             * add data_i member to PrevControl (p) and
 190             * add data_o member to NextControl (n)
 191             Calling ControlBase._new_data is a good way to do that.
 192         """
 193         print ("ControlBase", self, stage, in_multi, stage_ctl)
 194         StageHelper.__init__(self, stage)
 195
 196         # set up input and output IO ACK (prev/next ready/valid)
 197         self.p = PrevControl(in_multi, stage_ctl, maskwid=maskwid)
 198         self.n = NextControl(stage_ctl, maskwid=maskwid)
 199
 200         # set up the input and output data
 201         if stage is not None:
 202             self._new_data("data")
 203
 204     def _new_data(self, name):
 205         """ allocates new data_i and data_o
 206         """
 207         self.p.data_i, self.n.data_o = self.new_specs(name)
 208
 209     @property
 210     def data_r(self):
 211         return self.process(self.p.data_i)
 212
 213     def connect_to_next(self, nxt):
 214         """ helper function to connect to the next stage data/valid/ready.
 215         """
 216         return self.n.connect_to_next(nxt.p)
 217
 218     def _connect_in(self, prev):
 219         """ internal helper function to connect stage to an input source.
 220             do not use to connect stage-to-stage!
 221         """
 222         return self.p._connect_in(prev.p)
 223
 224     def _connect_out(self, nxt):
 225         """ internal helper function to connect stage to an output source.
 226             do not use to connect stage-to-stage!
 227         """
 228         return self.n._connect_out(nxt.n)
 229
 230     def connect(self, pipechain):
 231         """ connects a chain (list) of Pipeline instances together and
 232             links them to this ControlBase instance:
 233
 234                       in <----> self <---> out
 235                        |                   ^
 236                        v                   |
 237                     [pipe1, pipe2, pipe3, pipe4]
 238                        |    ^  |    ^  |     ^
 239                        v    |  v    |  v     |
 240                      out---in out--in out---in
 241
 242             Also takes care of allocating data_i/data_o, by looking up
 243             the data spec for each end of the pipechain.  i.e It is NOT
 244             necessary to allocate self.p.data_i or self.n.data_o manually:
 245             this is handled AUTOMATICALLY, here.
 246
 247             Basically this function is the direct equivalent of StageChain,
 248             except that unlike StageChain, the Pipeline logic is followed.
 249
 250             Just as StageChain presents an object that conforms to the
 251             Stage API from a list of objects that also conform to the
 252             Stage API, an object that calls this Pipeline connect function
 253             has the exact same pipeline API as the list of pipline objects
 254             it is called with.
 255
 256             Thus it becomes possible to build up larger chains recursively.
 257             More complex chains (multi-input, multi-output) will have to be
 258             done manually.
 259
 260             Argument:
 261
 262             * :pipechain: - a sequence of ControlBase-derived classes
 263                             (must be one or more in length)
 264
 265             Returns:
 266
 267             * a list of eq assignments that will need to be added in
 268               an elaborate() to m.d.comb
 269         """
 270         assert len(pipechain) > 0, "pipechain must be non-zero length"
 271         assert self.stage is None, "do not use connect with a stage"
 272         eqs = [] # collated list of assignment statements
 273
 274         # connect inter-chain
 275         for i in range(len(pipechain)-1):
 276             pipe1 = pipechain[i]                # earlier
 277             pipe2 = pipechain[i+1]              # later (by 1)
 278             eqs += pipe1.connect_to_next(pipe2) # earlier n to later p
 279
 280         # connect front and back of chain to ourselves
 281         front = pipechain[0]                # first in chain
 282         end = pipechain[-1]                 # last in chain
 283         self.set_specs(front, end) # sets up ispec/ospec functions
 284         self._new_data("chain") # NOTE: REPLACES existing data
 285         eqs += front._connect_in(self)      # front p to our p
 286         eqs += end._connect_out(self)       # end n   to our n
 287
 288         return eqs
 289
 290     def set_input(self, i):
 291         """ helper function to set the input data (used in unit tests)
 292         """
 293         return nmoperator.eq(self.p.data_i, i)
 294
 295     def __iter__(self):
 296         yield from self.p # yields ready/valid/data (data also gets yielded)
 297         yield from self.n # ditto
 298
 299     def ports(self):
 300         return list(self)
 301
 302     def elaborate(self, platform):
 303         """ handles case where stage has dynamic ready/valid functions
 304         """
 305         m = Module()
 306         m.submodules.p = self.p
 307         m.submodules.n = self.n
 308
 309         self.setup(m, self.p.data_i)
 310
 311         if not self.p.stage_ctl:
 312             return m
 313
 314         # intercept the previous (outgoing) "ready", combine with stage ready
 315         m.d.comb += self.p.s_ready_o.eq(self.p._ready_o & self.stage.d_ready)
 316
 317         # intercept the next (incoming) "ready" and combine it with data valid
 318         sdv = self.stage.d_valid(self.n.ready_i)
 319         m.d.comb += self.n.d_valid.eq(self.n.ready_i & sdv)
 320
 321         return m
 322
 323
 324 class BufferedHandshake(ControlBase):
 325     """ buffered pipeline stage.  data and strobe signals travel in sync.
 326         if ever the input is ready and the output is not, processed data
 327         is shunted in a temporary register.
 328
 329         Argument: stage.  see Stage API above
 330
 331         stage-1   p.valid_i >>in   stage   n.valid_o out>>   stage+1
 332         stage-1   p.ready_o <<out  stage   n.ready_i <<in    stage+1
 333         stage-1   p.data_i  >>in   stage   n.data_o  out>>   stage+1
 334                               |             |
 335                             process --->----^
 336                               |             |
 337                               +-- r_data ->-+
 338
 339         input data p.data_i is read (only), is processed and goes into an
 340         intermediate result store [process()].  this is updated combinatorially.
 341
 342         in a non-stall condition, the intermediate result will go into the
 343         output (update_output).  however if ever there is a stall, it goes
 344         into r_data instead [update_buffer()].
 345
 346         when the non-stall condition is released, r_data is the first
 347         to be transferred to the output [flush_buffer()], and the stall
 348         condition cleared.
 349
 350         on the next cycle (as long as stall is not raised again) the
 351         input may begin to be processed and transferred directly to output.
 352     """
 353
 354     def elaborate(self, platform):
 355         self.m = ControlBase.elaborate(self, platform)
 356
 357         result = _spec(self.stage.ospec, "r_tmp")
 358         r_data = _spec(self.stage.ospec, "r_data")
 359
 360         # establish some combinatorial temporaries
 361         o_n_validn = Signal(reset_less=True)
 362         n_ready_i = Signal(reset_less=True, name="n_i_rdy_data")
 363         nir_por = Signal(reset_less=True)
 364         nir_por_n = Signal(reset_less=True)
 365         p_valid_i = Signal(reset_less=True)
 366         nir_novn = Signal(reset_less=True)
 367         nirn_novn = Signal(reset_less=True)
 368         por_pivn = Signal(reset_less=True)
 369         npnn = Signal(reset_less=True)
 370         self.m.d.comb += [p_valid_i.eq(self.p.valid_i_test),
 371                      o_n_validn.eq(~self.n.valid_o),
 372                      n_ready_i.eq(self.n.ready_i_test),
 373                      nir_por.eq(n_ready_i & self.p._ready_o),
 374                      nir_por_n.eq(n_ready_i & ~self.p._ready_o),
 375                      nir_novn.eq(n_ready_i | o_n_validn),
 376                      nirn_novn.eq(~n_ready_i & o_n_validn),
 377                      npnn.eq(nir_por | nirn_novn),
 378                      por_pivn.eq(self.p._ready_o & ~p_valid_i)
 379         ]
 380
 381         # store result of processing in combinatorial temporary
 382         self.m.d.comb += nmoperator.eq(result, self.data_r)
 383
 384         # if not in stall condition, update the temporary register
 385         with self.m.If(self.p.ready_o): # not stalled
 386             self.m.d.sync += nmoperator.eq(r_data, result) # update buffer
 387
 388         # data pass-through conditions
 389         with self.m.If(npnn):
 390             data_o = self._postprocess(result) # XXX TBD, does nothing right now
 391             self.m.d.sync += [self.n.valid_o.eq(p_valid_i), # valid if p_valid
 392                               nmoperator.eq(self.n.data_o, data_o), # update out
 393                              ]
 394         # buffer flush conditions (NOTE: can override data passthru conditions)
 395         with self.m.If(nir_por_n): # not stalled
 396             # Flush the [already processed] buffer to the output port.
 397             data_o = self._postprocess(r_data) # XXX TBD, does nothing right now
 398             self.m.d.sync += [self.n.valid_o.eq(1),  # reg empty
 399                               nmoperator.eq(self.n.data_o, data_o), # flush
 400                              ]
 401         # output ready conditions
 402         self.m.d.sync += self.p._ready_o.eq(nir_novn | por_pivn)
 403
 404         return self.m
 405
 406
 407 class MaskCancellable(ControlBase):
 408     """ Mask-activated Cancellable pipeline
 409
 410         Argument: stage.  see Stage API above
 411
 412         stage-1   p.valid_i >>in   stage   n.valid_o out>>   stage+1
 413         stage-1   p.ready_o <<out  stage   n.ready_i <<in    stage+1
 414         stage-1   p.data_i  >>in   stage   n.data_o  out>>   stage+1
 415                               |             |
 416                               +--process->--^
 417     """
 418     def __init__(self, stage, maskwid, in_multi=None, stage_ctl=False):
 419         ControlBase.__init__(self, stage, in_multi, stage_ctl, maskwid)
 420
 421
 422     def elaborate(self, platform):
 423         self.m = m = ControlBase.elaborate(self, platform)
 424
 425         # store result of processing in combinatorial temporary
 426         result = _spec(self.stage.ospec, "r_tmp")
 427         m.d.comb += nmoperator.eq(result, self.data_r)
 428
 429         # establish if the data should be passed on.  cancellation is
 430         # a global signal.
 431         # XXX EXCEPTIONAL CIRCUMSTANCES: inspection of the data payload
 432         # is NOT "normal" for the Stage API.
 433         p_valid_i = Signal(reset_less=True)
 434         print ("self.p.data_i", self.p.data_i)
 435         m.d.comb += p_valid_i.eq((self.p.mask_i.bool()))
 436         #m.d.comb += p_valid_i.eq((self.p.data_i.ctx.idmask & \
 437         #                         ~self.cancelmask)) # nonzero
 438
 439         # if idmask nonzero, data gets passed on.
 440         m.d.sync += self.n.valid_o.eq(p_valid_i)
 441         with m.If(p_valid_i):
 442             m.d.sync += self.n.mask_o.eq(self.p.mask_i)
 443             data_o = self._postprocess(result) # XXX TBD, does nothing right now
 444             m.d.sync += nmoperator.eq(self.n.data_o, data_o) # update output
 445         with m.Else():
 446             m.d.sync += self.n.mask_o.eq(0)
 447
 448         # output valid if
 449         # input always "ready"
 450         m.d.comb += self.p._ready_o.eq(Const(1))
 451
 452         return self.m
 453
 454
 455 class SimpleHandshake(ControlBase):
 456     """ simple handshake control.  data and strobe signals travel in sync.
 457         implements the protocol used by Wishbone and AXI4.
 458
 459         Argument: stage.  see Stage API above
 460
 461         stage-1   p.valid_i >>in   stage   n.valid_o out>>   stage+1
 462         stage-1   p.ready_o <<out  stage   n.ready_i <<in    stage+1
 463         stage-1   p.data_i  >>in   stage   n.data_o  out>>   stage+1
 464                               |             |
 465                               +--process->--^
 466         Truth Table
 467
 468         Inputs   Temporary  Output Data
 469         -------  ---------- -----  ----
 470         P P N N  PiV& ~NiR&  N P
 471         i o i o  PoR  NoV    o o
 472         V R R V              V R
 473
 474         -------   -    -     - -
 475         0 0 0 0   0    0    >0 0    reg
 476         0 0 0 1   0    1    >1 0    reg
 477         0 0 1 0   0    0     0 1    process(data_i)
 478         0 0 1 1   0    0     0 1    process(data_i)
 479         -------   -    -     - -
 480         0 1 0 0   0    0    >0 0    reg
 481         0 1 0 1   0    1    >1 0    reg
 482         0 1 1 0   0    0     0 1    process(data_i)
 483         0 1 1 1   0    0     0 1    process(data_i)
 484         -------   -    -     - -
 485         1 0 0 0   0    0    >0 0    reg
 486         1 0 0 1   0    1    >1 0    reg
 487         1 0 1 0   0    0     0 1    process(data_i)
 488         1 0 1 1   0    0     0 1    process(data_i)
 489         -------   -    -     - -
 490         1 1 0 0   1    0     1 0    process(data_i)
 491         1 1 0 1   1    1     1 0    process(data_i)
 492         1 1 1 0   1    0     1 1    process(data_i)
 493         1 1 1 1   1    0     1 1    process(data_i)
 494         -------   -    -     - -
 495     """
 496
 497     def elaborate(self, platform):
 498         self.m = m = ControlBase.elaborate(self, platform)
 499
 500         r_busy = Signal()
 501         result = _spec(self.stage.ospec, "r_tmp")
 502
 503         # establish some combinatorial temporaries
 504         n_ready_i = Signal(reset_less=True, name="n_i_rdy_data")
 505         p_valid_i_p_ready_o = Signal(reset_less=True)
 506         p_valid_i = Signal(reset_less=True)
 507         m.d.comb += [p_valid_i.eq(self.p.valid_i_test),
 508                      n_ready_i.eq(self.n.ready_i_test),
 509                      p_valid_i_p_ready_o.eq(p_valid_i & self.p.ready_o),
 510         ]
 511
 512         # store result of processing in combinatorial temporary
 513         m.d.comb += nmoperator.eq(result, self.data_r)
 514
 515         # previous valid and ready
 516         with m.If(p_valid_i_p_ready_o):
 517             data_o = self._postprocess(result) # XXX TBD, does nothing right now
 518             m.d.sync += [r_busy.eq(1),      # output valid
 519                          nmoperator.eq(self.n.data_o, data_o), # update output
 520                         ]
 521         # previous invalid or not ready, however next is accepting
 522         with m.Elif(n_ready_i):
 523             data_o = self._postprocess(result) # XXX TBD, does nothing right now
 524             m.d.sync += [nmoperator.eq(self.n.data_o, data_o)]
 525             # TODO: could still send data here (if there was any)
 526             #m.d.sync += self.n.valid_o.eq(0) # ...so set output invalid
 527             m.d.sync += r_busy.eq(0) # ...so set output invalid
 528
 529         m.d.comb += self.n.valid_o.eq(r_busy)
 530         # if next is ready, so is previous
 531         m.d.comb += self.p._ready_o.eq(n_ready_i)
 532
 533         return self.m
 534
 535
 536 class UnbufferedPipeline(ControlBase):
 537     """ A simple pipeline stage with single-clock synchronisation
 538         and two-way valid/ready synchronised signalling.
 539
 540         Note that a stall in one stage will result in the entire pipeline
 541         chain stalling.
 542
 543         Also that unlike BufferedHandshake, the valid/ready signalling does NOT
 544         travel synchronously with the data: the valid/ready signalling
 545         combines in a *combinatorial* fashion.  Therefore, a long pipeline
 546         chain will lengthen propagation delays.
 547
 548         Argument: stage.  see Stage API, above
 549
 550         stage-1   p.valid_i >>in   stage   n.valid_o out>>   stage+1
 551         stage-1   p.ready_o <<out  stage   n.ready_i <<in    stage+1
 552         stage-1   p.data_i  >>in   stage   n.data_o  out>>   stage+1
 553                               |             |
 554                             r_data        result
 555                               |             |
 556                               +--process ->-+
 557
 558         Attributes:
 559         -----------
 560         p.data_i : StageInput, shaped according to ispec
 561             The pipeline input
 562         p.data_o : StageOutput, shaped according to ospec
 563             The pipeline output
 564         r_data : input_shape according to ispec
 565             A temporary (buffered) copy of a prior (valid) input.
 566             This is HELD if the output is not ready.  It is updated
 567             SYNCHRONOUSLY.
 568         result: output_shape according to ospec
 569             The output of the combinatorial logic.  it is updated
 570             COMBINATORIALLY (no clock dependence).
 571
 572         Truth Table
 573
 574         Inputs  Temp  Output  Data
 575         -------   -   -----   ----
 576         P P N N ~NiR&  N P
 577         i o i o  NoV   o o
 578         V R R V        V R
 579
 580         -------   -    - -
 581         0 0 0 0   0    0 1    reg
 582         0 0 0 1   1    1 0    reg
 583         0 0 1 0   0    0 1    reg
 584         0 0 1 1   0    0 1    reg
 585         -------   -    - -
 586         0 1 0 0   0    0 1    reg
 587         0 1 0 1   1    1 0    reg
 588         0 1 1 0   0    0 1    reg
 589         0 1 1 1   0    0 1    reg
 590         -------   -    - -
 591         1 0 0 0   0    1 1    reg
 592         1 0 0 1   1    1 0    reg
 593         1 0 1 0   0    1 1    reg
 594         1 0 1 1   0    1 1    reg
 595         -------   -    - -
 596         1 1 0 0   0    1 1    process(data_i)
 597         1 1 0 1   1    1 0    process(data_i)
 598         1 1 1 0   0    1 1    process(data_i)
 599         1 1 1 1   0    1 1    process(data_i)
 600         -------   -    - -
 601
 602         Note: PoR is *NOT* involved in the above decision-making.
 603     """
 604
 605     def elaborate(self, platform):
 606         self.m = m = ControlBase.elaborate(self, platform)
 607
 608         data_valid = Signal() # is data valid or not
 609         r_data = _spec(self.stage.ospec, "r_tmp") # output type
 610
 611         # some temporaries
 612         p_valid_i = Signal(reset_less=True)
 613         pv = Signal(reset_less=True)
 614         buf_full = Signal(reset_less=True)
 615         m.d.comb += p_valid_i.eq(self.p.valid_i_test)
 616         m.d.comb += pv.eq(self.p.valid_i & self.p.ready_o)
 617         m.d.comb += buf_full.eq(~self.n.ready_i_test & data_valid)
 618
 619         m.d.comb += self.n.valid_o.eq(data_valid)
 620         m.d.comb += self.p._ready_o.eq(~data_valid | self.n.ready_i_test)
 621         m.d.sync += data_valid.eq(p_valid_i | buf_full)
 622
 623         with m.If(pv):
 624             m.d.sync += nmoperator.eq(r_data, self.data_r)
 625         data_o = self._postprocess(r_data) # XXX TBD, does nothing right now
 626         m.d.comb += nmoperator.eq(self.n.data_o, data_o)
 627
 628         return self.m
 629
 630
 631 class UnbufferedPipeline2(ControlBase):
 632     """ A simple pipeline stage with single-clock synchronisation
 633         and two-way valid/ready synchronised signalling.
 634
 635         Note that a stall in one stage will result in the entire pipeline
 636         chain stalling.
 637
 638         Also that unlike BufferedHandshake, the valid/ready signalling does NOT
 639         travel synchronously with the data: the valid/ready signalling
 640         combines in a *combinatorial* fashion.  Therefore, a long pipeline
 641         chain will lengthen propagation delays.
 642
 643         Argument: stage.  see Stage API, above
 644
 645         stage-1   p.valid_i >>in   stage   n.valid_o out>>   stage+1
 646         stage-1   p.ready_o <<out  stage   n.ready_i <<in    stage+1
 647         stage-1   p.data_i  >>in   stage   n.data_o  out>>   stage+1
 648                               |             |    |
 649                               +- process-> buf <-+
 650         Attributes:
 651         -----------
 652         p.data_i : StageInput, shaped according to ispec
 653             The pipeline input
 654         p.data_o : StageOutput, shaped according to ospec
 655             The pipeline output
 656         buf : output_shape according to ospec
 657             A temporary (buffered) copy of a valid output
 658             This is HELD if the output is not ready.  It is updated
 659             SYNCHRONOUSLY.
 660
 661         Inputs  Temp  Output Data
 662         -------   -   -----
 663         P P N N ~NiR&  N P   (buf_full)
 664         i o i o  NoV   o o
 665         V R R V        V R
 666
 667         -------   -    - -
 668         0 0 0 0   0    0 1   process(data_i)
 669         0 0 0 1   1    1 0   reg (odata, unchanged)
 670         0 0 1 0   0    0 1   process(data_i)
 671         0 0 1 1   0    0 1   process(data_i)
 672         -------   -    - -
 673         0 1 0 0   0    0 1   process(data_i)
 674         0 1 0 1   1    1 0   reg (odata, unchanged)
 675         0 1 1 0   0    0 1   process(data_i)
 676         0 1 1 1   0    0 1   process(data_i)
 677         -------   -    - -
 678         1 0 0 0   0    1 1   process(data_i)
 679         1 0 0 1   1    1 0   reg (odata, unchanged)
 680         1 0 1 0   0    1 1   process(data_i)
 681         1 0 1 1   0    1 1   process(data_i)
 682         -------   -    - -
 683         1 1 0 0   0    1 1   process(data_i)
 684         1 1 0 1   1    1 0   reg (odata, unchanged)
 685         1 1 1 0   0    1 1   process(data_i)
 686         1 1 1 1   0    1 1   process(data_i)
 687         -------   -    - -
 688
 689         Note: PoR is *NOT* involved in the above decision-making.
 690     """
 691
 692     def elaborate(self, platform):
 693         self.m = m = ControlBase.elaborate(self, platform)
 694
 695         buf_full = Signal() # is data valid or not
 696         buf = _spec(self.stage.ospec, "r_tmp") # output type
 697
 698         # some temporaries
 699         p_valid_i = Signal(reset_less=True)
 700         m.d.comb += p_valid_i.eq(self.p.valid_i_test)
 701
 702         m.d.comb += self.n.valid_o.eq(buf_full | p_valid_i)
 703         m.d.comb += self.p._ready_o.eq(~buf_full)
 704         m.d.sync += buf_full.eq(~self.n.ready_i_test & self.n.valid_o)
 705
 706         data_o = Mux(buf_full, buf, self.data_r)
 707         data_o = self._postprocess(data_o) # XXX TBD, does nothing right now
 708         m.d.comb += nmoperator.eq(self.n.data_o, data_o)
 709         m.d.sync += nmoperator.eq(buf, self.n.data_o)
 710
 711         return self.m
 712
 713
 714 class PassThroughHandshake(ControlBase):
 715     """ A control block that delays by one clock cycle.
 716
 717         Inputs   Temporary          Output Data
 718         -------  ------------------  ----- ----
 719         P P N N  PiV& PiV| NiR| pvr   N P  (pvr)
 720         i o i o  PoR  ~PoR ~NoV       o o
 721         V R R V                       V R
 722
 723         -------   -    -    -   -     - -
 724         0 0 0 0   0    1    1   0     1 1   odata (unchanged)
 725         0 0 0 1   0    1    0   0     1 0   odata (unchanged)
 726         0 0 1 0   0    1    1   0     1 1   odata (unchanged)
 727         0 0 1 1   0    1    1   0     1 1   odata (unchanged)
 728         -------   -    -    -   -     - -
 729         0 1 0 0   0    0    1   0     0 1   odata (unchanged)
 730         0 1 0 1   0    0    0   0     0 0   odata (unchanged)
 731         0 1 1 0   0    0    1   0     0 1   odata (unchanged)
 732         0 1 1 1   0    0    1   0     0 1   odata (unchanged)
 733         -------   -    -    -   -     - -
 734         1 0 0 0   0    1    1   1     1 1   process(in)
 735         1 0 0 1   0    1    0   0     1 0   odata (unchanged)
 736         1 0 1 0   0    1    1   1     1 1   process(in)
 737         1 0 1 1   0    1    1   1     1 1   process(in)
 738         -------   -    -    -   -     - -
 739         1 1 0 0   1    1    1   1     1 1   process(in)
 740         1 1 0 1   1    1    0   0     1 0   odata (unchanged)
 741         1 1 1 0   1    1    1   1     1 1   process(in)
 742         1 1 1 1   1    1    1   1     1 1   process(in)
 743         -------   -    -    -   -     - -
 744
 745     """
 746
 747     def elaborate(self, platform):
 748         self.m = m = ControlBase.elaborate(self, platform)
 749
 750         r_data = _spec(self.stage.ospec, "r_tmp") # output type
 751
 752         # temporaries
 753         p_valid_i = Signal(reset_less=True)
 754         pvr = Signal(reset_less=True)
 755         m.d.comb += p_valid_i.eq(self.p.valid_i_test)
 756         m.d.comb += pvr.eq(p_valid_i & self.p.ready_o)
 757
 758         m.d.comb += self.p.ready_o.eq(~self.n.valid_o |  self.n.ready_i_test)
 759         m.d.sync += self.n.valid_o.eq(p_valid_i       | ~self.p.ready_o)
 760
 761         odata = Mux(pvr, self.data_r, r_data)
 762         m.d.sync += nmoperator.eq(r_data, odata)
 763         r_data = self._postprocess(r_data) # XXX TBD, does nothing right now
 764         m.d.comb += nmoperator.eq(self.n.data_o, r_data)
 765
 766         return m
 767
 768
 769 class RegisterPipeline(UnbufferedPipeline):
 770     """ A pipeline stage that delays by one clock cycle, creating a
 771         sync'd latch out of data_o and valid_o as an indirect byproduct
 772         of using PassThroughStage
 773     """
 774     def __init__(self, iospecfn):
 775         UnbufferedPipeline.__init__(self, PassThroughStage(iospecfn))
 776
 777
 778 class FIFOControl(ControlBase):
 779     """ FIFO Control.  Uses Queue to store data, coincidentally
 780         happens to have same valid/ready signalling as Stage API.
 781
 782         data_i -> fifo.din -> FIFO -> fifo.dout -> data_o
 783     """
 784     def __init__(self, depth, stage, in_multi=None, stage_ctl=False,
 785                                      fwft=True, pipe=False):
 786         """ FIFO Control
 787
 788             * :depth: number of entries in the FIFO
 789             * :stage: data processing block
 790             * :fwft:  first word fall-thru mode (non-fwft introduces delay)
 791             * :pipe:  specifies pipe mode.
 792
 793             when fwft = True it indicates that transfers may occur
 794             combinatorially through stage processing in the same clock cycle.
 795             This requires that the Stage be a Moore FSM:
 796             https://en.wikipedia.org/wiki/Moore_machine
 797
 798             when fwft = False it indicates that all output signals are
 799             produced only from internal registers or memory, i.e. that the
 800             Stage is a Mealy FSM:
 801             https://en.wikipedia.org/wiki/Mealy_machine
 802
 803             data is processed (and located) as follows:
 804
 805             self.p  self.stage temp    fn temp  fn  temp  fp   self.n
 806             data_i->process()->result->cat->din.FIFO.dout->cat(data_o)
 807
 808             yes, really: cat produces a Cat() which can be assigned to.
 809             this is how the FIFO gets de-catted without needing a de-cat
 810             function
 811         """
 812         self.fwft = fwft
 813         self.pipe = pipe
 814         self.fdepth = depth
 815         ControlBase.__init__(self, stage, in_multi, stage_ctl)
 816
 817     def elaborate(self, platform):
 818         self.m = m = ControlBase.elaborate(self, platform)
 819
 820         # make a FIFO with a signal of equal width to the data_o.
 821         (fwidth, _) = nmoperator.shape(self.n.data_o)
 822         fifo = Queue(fwidth, self.fdepth, fwft=self.fwft, pipe=self.pipe)
 823         m.submodules.fifo = fifo
 824
 825         def processfn(data_i):
 826             # store result of processing in combinatorial temporary
 827             result = _spec(self.stage.ospec, "r_temp")
 828             m.d.comb += nmoperator.eq(result, self.process(data_i))
 829             return nmoperator.cat(result)
 830
 831         ## prev: make the FIFO (Queue object) "look" like a PrevControl...
 832         m.submodules.fp = fp = PrevControl()
 833         fp.valid_i, fp._ready_o, fp.data_i = fifo.we, fifo.writable, fifo.din
 834         m.d.comb += fp._connect_in(self.p, fn=processfn)
 835
 836         # next: make the FIFO (Queue object) "look" like a NextControl...
 837         m.submodules.fn = fn = NextControl()
 838         fn.valid_o, fn.ready_i, fn.data_o  = fifo.readable, fifo.re, fifo.dout
 839         connections = fn._connect_out(self.n, fn=nmoperator.cat)
 840
 841         # ok ok so we can't just do the ready/valid eqs straight:
 842         # first 2 from connections are the ready/valid, 3rd is data.
 843         if self.fwft:
 844             m.d.comb += connections[:2] # combinatorial on next ready/valid
 845         else:
 846             m.d.sync += connections[:2]  # non-fwft mode needs sync
 847         data_o = connections[2] # get the data
 848         data_o = self._postprocess(data_o) # XXX TBD, does nothing right now
 849         m.d.comb += data_o
 850
 851         return m
 852
 853
 854 # aka "RegStage".
 855 class UnbufferedPipeline(FIFOControl):
 856     def __init__(self, stage, in_multi=None, stage_ctl=False):
 857         FIFOControl.__init__(self, 1, stage, in_multi, stage_ctl,
 858                                    fwft=True, pipe=False)
 859
 860 # aka "BreakReadyStage" XXX had to set fwft=True to get it to work
 861 class PassThroughHandshake(FIFOControl):
 862     def __init__(self, stage, in_multi=None, stage_ctl=False):
 863         FIFOControl.__init__(self, 1, stage, in_multi, stage_ctl,
 864                                    fwft=True, pipe=True)
 865
 866 # this is *probably* BufferedHandshake, although test #997 now succeeds.
 867 class BufferedHandshake(FIFOControl):
 868     def __init__(self, stage, in_multi=None, stage_ctl=False):
 869         FIFOControl.__init__(self, 2, stage, in_multi, stage_ctl,
 870                                    fwft=True, pipe=False)
 871
 872
 873 """
 874 # this is *probably* SimpleHandshake (note: memory cell size=0)
 875 class SimpleHandshake(FIFOControl):
 876     def __init__(self, stage, in_multi=None, stage_ctl=False):
 877         FIFOControl.__init__(self, 0, stage, in_multi, stage_ctl,
 878                                    fwft=True, pipe=False)
 879 """