Files
chipyard/docs/Customization/Custom-Core.rst
2020-07-05 21:05:21 -07:00

335 lines
20 KiB
ReStructuredText

.. _custom_core:
Adding a custom core
====================
You may want to integrate a custom RISC-V core into the Chipyard framework. This documentation page provides a step-to-step
instruction on how to achieve this.
.. note::
RoCC is currently not supported by cores other than Rocket and BOOM. Please use Rocket or BOOM as the RoCC base core if you need to use RoCC.
Wrap Verilog Module with Blackbox (Optional)
--------------------------------------------
Since Chipyard uses Scala and Chisel, if the top module of your core is not in Chisel, you will first need to create a Verilog
blackbox for it so that it can be processed by Chipyard. See :ref:`incorporating-verilog-blocks` for instructions.
Create Parameter Case Classes
-----------------------------
Chipyard will generate a core for every ``InstantiableTileParams`` object it discovered in the ``TilesLocated(InSubsystem)`` key.
This object is derived from``TileParams``, a trait containing the information needed to create a tile. All cores must have
their own implementation of ``InstantiableTileParams``, as well as ``CoreParams`` which is passed as a field in ``TileParams``.
``TileParams`` holds the parameters for the tile, which are the same for every generated core, while ``CoreParams``
contains the parameters for individual cores. They must be implemented as case classes with fields that can be overridden by
other config fragments as the constructor parameters. See the appendix at the bottom of the page for a list of
variable to be implemented. You can also add custom fields to them, but standard fields should always be preferred.
``InstantiableTileParams[TileType]`` holds the constructor of ``TileType`` on top of the fields of ``TileParams``.
All custom cores will also need to implement ``instantiate()`` in their tile parameter class to return a new instance
of the tile class ``TileType``.
``TileParams``, ``InstantiableTileParams[TileType]`` and ``CoreParams`` contains the following fields (you may ignore
any fields marked "Rocket specific" and use their default values, although it is recommended to use them if you
need a custom field with similar purposes):
.. code-block:: scala
trait TileParams {
val core: CoreParams // Core parameters (see below)
val icache: Option[ICacheParams] // Rocket specific: I1 cache option
val dcache: Option[DCacheParams] // Rocket specific: D1 cache option
val btb: Option[BTBParams] // Rocket specific: BTB / branch predictor option
val hartId: Int // Hart ID: Must be unique within a design config (This MUST be a case class parameter)
val beuAddr: Option[BigInt] // Rocket specific: Bus Error Unit for Rocket Core
val blockerCtrlAddr: Option[BigInt] // Rocket specific: Bus Blocker for Rocket Core
val name: Option[String] // Name of the core
}
abstract class InstantiableTileParams[TileType <: BaseTile] extends TileParams {
def instantiate(crossing: TileCrossingParamsLike, lookup: LookupByHartIdImpl)
(implicit p: Parameters): TileType
}
trait CoreParams {
val bootFreqHz: BigInt // Frequency
val useVM: Boolean // Support virtual memory
val useUser: Boolean // Support user mode
val useSupervisor: Boolean // Support supervisor mode
val useDebug: Boolean // Support RISC-V debug specs
val useAtomics: Boolean // Support A extension
val useAtomicsOnlyForIO: Boolean // Support A extension for memory-mapped IO (may be true even if useAtomics is false)
val useCompressed: Boolean // Support C extension
val useVector: Boolean = false // Support V extension
val useSCIE: Boolean // Support custom instructions (in custom-0 and custom-1)
val useRVE: Boolean // Use E base ISA
val mulDiv: Option[MulDivParams] // *Rocket specific: M extension related setting (Use Some(MulDivParams()) to indicate M extension supported)
val fpu: Option[FPUParams] // F and D extensions and related setting (see below)
val fetchWidth: Int // Max # of insts fetched every cycle
val decodeWidth: Int // Max # of insts decoded every cycle
val retireWidth: Int // Max # of insts retired every cycle
val instBits: Int // Instruction bits (if 32 bit and 64 bit are both supported, use 64)
val nLocalInterrupts: Int // # of local interrupts (see SiFive interrupt cookbook)
val nPMPs: Int // # of Physical Memory Protection units
val pmpGranularity: Int // Size of the smallest unit of region for PMP unit (must be power of 2)
val nBreakpoints: Int // # of hardware breakpoints supported (in RISC-V debug specs)
val useBPWatch: Boolean // Support hardware breakpoints
val nPerfCounters: Int // # of supported performance counters
val haveBasicCounters: Boolean // Support basic counters defined in the RISC-V counter extension
val haveFSDirty: Boolean // If true, the core will set FS field in mstatus CSR to dirty when appropriate
val misaWritable: Boolean // Support writable misa CSR (like variable instruction bits)
val haveCFlush: Boolean // Rocket specific: enables Rocket's custom instruction extension to flush the cache
val nL2TLBEntries: Int // # of L2 TLB entries
val mtvecInit: Option[BigInt] // mtvec CSR (of V extension) initial value
val mtvecWritable: Boolean // If mtvec CSR is writable
// Normally, you don't need to change these values (except lrscCycles)
def customCSRs(implicit p: Parameters): CustomCSRs = new CustomCSRs
def hasSupervisorMode: Boolean = useSupervisor || useVM
def instBytes: Int = instBits / 8
def fetchBytes: Int = fetchWidth * instBytes
// Rocket specific: Longest possible latency of Rocket core D1 cache. Simply set it to the default value 80 if you don't use it.
def lrscCycles: Int
def dcacheReqTagBits: Int = 6
def minFLen: Int = 32
def vLen: Int = 0
def sLen: Int = 0
def eLen(xLen: Int, fLen: Int): Int = xLen max fLen
def vMemDataBits: Int = 0
}
case class FPUParams(
minFLen: Int = 32, // Minimum floating point length (no need to change)
fLen: Int = 64, // Maximum floating point length, use 32 if only single precision is supported
divSqrt: Boolean = true, // Div/Sqrt operation supported
sfmaLatency: Int = 3, // Rocket specific: Fused multiply-add pipeline latency (single precision)
dfmaLatency: Int = 4 // Rocket specific: Fused multiply-add pipeline latency (double precision)
)
Most of the fields here are originally designed for the Rocket core and thus contain some implementation-specific details, but
many of them are general enough to be useful for other cores. It is strongly recommended to use these fields instead
of creating your own custom fields when applicable.
You will also need a ``CanAttachTile`` class to add the tile config into the config system, with the following format:
.. literalinclude:: ../../generators/chipyard/src/main/scala/example/TutorialTile.scala
:language: scala
:lines: 61-67
.. note::
Implementations may choose to ignore some fields here or use them in a non-standard way, but using an inaccurate
value may break Chipyard components that rely on them (e.g. an inaccurate indication of supported ISA extension will
result in an incorrect test suite being generated) as well as any custom modules that use them. ALWAYS document any
fields you ignore or with altered usage in your core implementation, and if you are implementing other devices that
would look up these config values, also document them. "Rocket specific" values are generally safe to ignore, but
you should document them if you use them.
Create Tile Class
-----------------
In Chipyard, all Tiles are diplomatically instantiated. In the first phase, diplomatic nodes which specify Tile-to-System
interconnects are evaluated, while in the second "Module Implementation" phase, hardware is elaborated.
See :ref:`tilelink_and_diplomacy` for more details. In this step, you will need to implement a tile class for your core.
All tile classes implement ``BaseTile`` and will normally implement ``SinksExternalInterrupts`` and ``SourcesExternalNotifications``,
which allow the tile to accept external interrupt. A typical tile has the following form:
.. literalinclude:: ../../generators/chipyard/src/main/scala/example/TutorialTile.scala
:language: scala
:lines: 87-125, 143
Connect TileLink Buses
----------------------
Chipyard use TileLink as its onboard bus protocol. If your core doesn't use TileLink, you will need to insert converters
between the core's memory protocol and TileLink in the Tile module.
in the tile class. Below is an example of how to connect a core using AXI4 to the TileLink bus with converters provided by
Rocket chip:
.. literalinclude:: ../../generators/chipyard/src/main/scala/example/TutorialTile.scala
:language: scala
:lines: 133-142
Remember, you may not need all of these intermediate widgets. See :ref:`diplomatic_widgets` for the meaning of each intermediate
widget. If you are using TileLink, then you only need the tap node and the TileLink node used by your components. Chipyard also
provides converters for AHB, APB and AXIS, and most of the AXI4 widgets has equivalent widget for these bus protocol; see the
source files in ``generators/rocket-chip/src/main/scala/amba`` for more info.
If you are using some other bus protocol, you may implement your own converters, using the files in ``generators/rocket-chip/src/main/scala/amba``
as the template, but it is not recommended unless you are familiar with TileLink.
``memAXI4Node`` is an AXI4 master node and is defined as following in our example:
.. literalinclude:: ../../generators/chipyard/src/main/scala/example/TutorialTile.scala
:language: scala
:lines: 126-132
where ``portName`` and ``idBits`` (number of bits to represent a port ID) are the parameter provides by the tile.
Make sure to read :ref:`node_types` to check out what type of nodes Chipyard supports and their parameters!
Also, by default, there are boundary buffers for both master and slave connections to the bus when they are leaving the tile, and you
can override the following two functions to control how to buffer the bus requests/responses:
.. code-block:: scala
protected def makeMasterBoundaryBuffers(implicit p: Parameters): TLBuffer
protected def makeSlaveBoundaryBuffers(implicit p: Parameters): TLBuffer
You can find more information on ``TLBuffer`` in :ref:`diplomatic_widgets`.
Connect Interrupt
-----------------
Chipyard allows a tile to either receive interrupts from other devices or initiate interrupts to notify other cores/devices.
In the tile that inherited ``SinksExternalInterrupts``, one can create a ``TileInterrupts`` object (a Chisel bundle) and
call ``decodeCoreInterrupts`` with the object as the argument. You can then read the interrupt bits from the object.
The definition of ``TileInterrupts`` is
.. code-block:: scala
class TileInterrupts(implicit p: Parameters) extends CoreBundle()(p) {
val debug = Bool() // debug interrupt
val mtip = Bool() // Machine level timer interrupt
val msip = Bool() // Machine level software interrupt
val meip = Bool() // Machine level external interrupt
val seip = usingSupervisor.option(Bool()) // Valid only if supervisor mode is supported
val lip = Vec(coreParams.nLocalInterrupts, Bool()) // Local interrupts
}
This function should be in the implementation class since it involves hardware generation.
Also, the tile can also notify other cores or devices for some events by calling following functions in ``SourcesExternalNotifications``
from the implementation class:
.. code-block:: scala
def reportHalt(could_halt: Option[Bool]) // Triggered when there is an unrecoverable hardware error (halt the machine)
def reportHalt(errors: Seq[CanHaveErrors]) // Varient for standard error bundle (used only by cache when there's an ECC error)
reportCease(could_cease: Option[Bool], quiescenceCycles: Int = 8) // Triggered when the core stop retiring instructions (like clock gating)
reportWFI(could_wfi: Option[Bool]) // Triggered when a WFI instruction is executed
Create Implementation Class
---------------------------
The implementation class for your core is of the following form:
.. literalinclude:: ../../generators/chipyard/src/main/scala/example/TutorialTile.scala
:language: scala
:lines: 145-149, 160
In the body of this class, you can look up any parameters by calling ``p({key})``, where ``{key}`` is the config key of
the value you want to look up. For a list of frequently used keys, see the appendix below.
If you create an AXI4 node (or equivalents), you will need to connect them to your core. You can connect a port like this:
.. literalinclude:: ../../generators/chipyard/src/main/scala/example/TutorialTile.scala
:language: scala
:lines: 151-159
Create Config Fragments to Integrate the Core
---------------------------------------------
To use your core in a Chipyard config, you would need a config fragment that would create a ``TileParams`` object of your core in
the current config. An example of such config will be like this:
.. literalinclude:: ../../generators/chipyard/src/main/scala/example/TutorialTile.scala
:language: scala
:lines: 162-179
Chipyard looks up the tile parameters in the field ``TilesLocated(InSubsystem)``, whose type is a list of ``InstantiableTileParams``.
This config fragment simply appends new tile parameters to the end of this list.
Now you have finished all the steps to prepare your cores for Chipyard! To generate the custom core, simply follow the instructions
in :ref:`custom_chisel` to add your project to the build system, then create a config by following the steps in :ref:`hetero_socs_`.
You can now run any desired workflow for the new config just as you do for the built-in cores.
If you would like to see how an actual core are integrated into Chipyard, ``generators/ariane/src/main/scala/ArianeTile.scala``
provides a concrete example of integrating a third party Verilog core Ariane.
Appendix: Common Config Keys
----------------------------
Chipyard provide a set of keys to store standard parameters. Below are some of the most common key used in core integration.
(Note that internal fields are hidden)
.. code-block:: scala
// keys
// Parameters exposed to the top-level design, set based on external requirements, etc. See RISC-V debug specs for more info.
case object DebugModuleKey extends Field[Option[DebugModuleParams]](Some(DebugModuleParams()))
case object BootROMParams extends Field[BootROMParams] // See chipyard boot process tutorial
case object CLINTKey extends Field[Option[CLINTParams]](None) // Core Local Interrupter setting (See SiFive Interrupt Cookbook)
case object PLICKey extends Field[Option[PLICParams]](None) // Platform Level Interrupt Controller setting (See SiFive Interrupt Cookbook)
case object CacheBlockBytes extends Field[Int](64) // # of bytes in a cache block
case object BroadcastKey extends Field(BroadcastParams()) // L2 Cache broadcast setting
case object BankedL2Key extends Field(BankedL2Params()) // L2 Cache memory setting
case object PgLevels extends Field[Int](2) // Page Level of virtual memory
case object ASIdBits extends Field[Int](0) // Max # of bits for Address Space Identifer (See specs)
case object ExtMem extends Field[Option[MemoryPortParams]](None) // External DRAM setting
case object ExtBus extends Field[Option[MasterPortParams]](None) // External (off-chip) output bus setting
case object ExtIn extends Field[Option[SlavePortParams]](None) // External (off-chip) input bus setting
case object MaxHartIdBits extends Field[Int] // Max # of bits used to represent a Hart ID
case object XLen extends Field[Int] // Instruction bits (32 or 64)
case object BuildRoCC extends Field[Seq[Parameters => LazyRoCC]](Nil) // See custom ROCC tutorial
// Values
case class DebugModuleParams (
nDMIAddrSize : Int = 7, // Size of the Debug Bus Address
nProgramBufferWords: Int = 16, // Number of 32-bit words for Program Buffer
nAbstractDataWords : Int = 4, // Number of 32-bit words for Abstract Commands
nScratch : Int = 1, // Number of scratch memories used
hasBusMaster : Boolean = false, // Whether or not a bus master should be included
clockGate : Boolean = true, // Use clock gating
maxSupportedSBAccess : Int = 32, // Maximum transaction size supported by System Bus Access logic.
supportQuickAccess : Boolean = false, // Whether or not to support the quick access command.
supportHartArray : Boolean = true, // Whether or not to implement the hart array register (if >1 hart).
nHaltGroups : Int = 1, // Number of halt groups (group of harts that are halted together)
nExtTriggers : Int = 0, // Number of extra triggers
hasHartResets : Boolean = false, // Whether harts can be reseted with debugging system
hasImplicitEbreak : Boolean = false, // There is an additional RO program buffer word containing an ebreak
hasAuthentication : Boolean = false, // Has authentication (to prevent unauthorized users to use debugging system)
crossingHasSafeReset : Boolean = true // Include "safe" logic in Async Crossings so that only one side needs to be reset.
)
case class CLINTParams(
baseAddress: BigInt = 0x02000000, // Default interrupt handler base address for CLINT
intStages: Int = 0 // # of cycles (stages) interrupts are delayed
)
case class PLICParams(
baseAddress: BigInt = 0xC000000, // Default interrupt handler base address for PLIC
maxPriorities: Int = 7, // Maximum allowed interrupt priority (cannot be over 7)
intStages: Int = 0, // # of cycles (stages) interrupts are delayed
maxHarts: Int = PLICConsts.maxMaxHarts // Maximum number or hart / core connected to it
)
case class BroadcastParams(
nTrackers: Int = 4, // # of broadcast tracker
bufferless: Boolean = false // Bufferless broadcast
)
case class BankedL2Params(
nBanks: Int = 1 // Number of banks in L2 cache
)
case class MasterPortParams(
base: BigInt, // Base memory address for this port
size: BigInt, // Size of this external memory
beatBytes: Int, // Interface width in bytes
idBits: Int, // # of bits in the port ID
maxXferBytes: Int = 256, // Maximum bytes in one transfer transaction
executable: Boolean = true // If the data from this port can be executed as instruciton
)
/** Specifies the width of external slave ports */
case class SlavePortParams(
beatBytes: Int, // Interface width in bytes
idBits: Int, // # of bits in the port ID
sourceBits: Int // # of bits in the source address
)
case class MemoryPortParams(
master: MasterPortParams, // The memory port setting
nMemoryChannels: Int // Number of memory channel
)