gem5/tests/configs/base_config.py

290 lines
11 KiB
Python
Raw Normal View History

# Copyright (c) 2012-2013 ARM Limited
# All rights reserved.
#
# The license below extends only to copyright in the software and shall
# not be construed as granting a license to any other intellectual
# property including but not limited to intellectual property relating
# to a hardware implementation of the functionality of the software
# licensed hereunder. You may use the software subject to the license
# terms below provided that you ensure that this notice is replicated
# unmodified and in its entirety in all distributions of the software,
# modified or unmodified, in source code or in binary form.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are
# met: redistributions of source code must retain the above copyright
# notice, this list of conditions and the following disclaimer;
# redistributions in binary form must reproduce the above copyright
# notice, this list of conditions and the following disclaimer in the
# documentation and/or other materials provided with the distribution;
# neither the name of the copyright holders nor the names of its
# contributors may be used to endorse or promote products derived from
# this software without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
#
# Authors: Andreas Sandberg
# Andreas Hansson
from abc import ABCMeta, abstractmethod
import m5
from m5.objects import *
from m5.proxy import *
m5.util.addToPath('../configs/common')
import FSConfig
from Caches import *
_have_kvm_support = 'BaseKvmCPU' in globals()
class BaseSystem(object):
"""Base system builder.
This class provides some basic functionality for creating an ARM
system with the usual peripherals (caches, GIC, etc.). It allows
customization by defining separate methods for different parts of
the initialization process.
"""
__metaclass__ = ABCMeta
def __init__(self, mem_mode='timing', mem_class=SimpleMemory,
cpu_class=TimingSimpleCPU, num_cpus=1, num_threads=1,
checker=False,
mem_size=None):
"""Initialize a simple base system.
Keyword Arguments:
mem_mode -- String describing the memory mode (timing or atomic)
mem_class -- Memory controller class to use
cpu_class -- CPU class to use
num_cpus -- Number of CPUs to instantiate
checker -- Set to True to add checker CPUs
mem_size -- Override the default memory size
"""
self.mem_mode = mem_mode
self.mem_class = mem_class
self.cpu_class = cpu_class
self.num_cpus = num_cpus
self.num_threads = num_threads
self.checker = checker
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
def create_cpus(self, cpu_clk_domain):
"""Return a list of CPU objects to add to a system."""
cpus = [ self.cpu_class(clk_domain=cpu_clk_domain,
numThreads=self.num_threads,
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
cpu_id=i)
for i in range(self.num_cpus) ]
if self.checker:
for c in cpus:
c.addCheckerCpu()
return cpus
def create_caches_private(self, cpu):
"""Add private caches to a CPU.
Arguments:
cpu -- CPU instance to work on.
"""
cpu.addPrivateSplitL1Caches(L1_ICache(size='32kB', assoc=1),
L1_DCache(size='32kB', assoc=4))
def create_caches_shared(self, system):
"""Add shared caches to a system.
Arguments:
system -- System to work on.
Returns:
A bus that CPUs should use to connect to the shared cache.
"""
system.toL2Bus = L2XBar(clk_domain=system.cpu_clk_domain)
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
system.l2c = L2Cache(clk_domain=system.cpu_clk_domain,
size='4MB', assoc=8)
system.l2c.cpu_side = system.toL2Bus.master
system.l2c.mem_side = system.membus.slave
return system.toL2Bus
def init_cpu(self, system, cpu, sha_bus):
"""Initialize a CPU.
Arguments:
system -- System to work on.
cpu -- CPU to initialize.
"""
if not cpu.switched_out:
self.create_caches_private(cpu)
cpu.createInterruptController()
cpu.connectAllPorts(sha_bus if sha_bus != None else system.membus,
system.membus)
# System has caches before the membus -> add snoop filter
if sha_bus and system.membus.snoop_filter == NULL:
system.membus.snoop_filter = SnoopFilter()
def init_kvm(self, system):
"""Do KVM-specific system initialization.
Arguments:
system -- System to work on.
"""
system.vm = KvmVM()
def init_system(self, system):
"""Initialize a system.
Arguments:
system -- System to initialize.
"""
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
self.create_clk_src(system)
system.cpu = self.create_cpus(system.cpu_clk_domain)
if _have_kvm_support and \
any([isinstance(c, BaseKvmCPU) for c in system.cpu]):
self.init_kvm(system)
sha_bus = self.create_caches_shared(system)
# System has caches before the membus -> add snoop filter
if sha_bus and system.membus.snoop_filter == NULL:
system.membus.snoop_filter = SnoopFilter()
for cpu in system.cpu:
self.init_cpu(system, cpu, sha_bus)
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
def create_clk_src(self,system):
# Create system clock domain. This provides clock value to every
# clocked object that lies beneath it unless explicitly overwritten
# by a different clock domain.
system.voltage_domain = VoltageDomain()
system.clk_domain = SrcClockDomain(clock = '1GHz',
voltage_domain =
system.voltage_domain)
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
# Create a seperate clock domain for components that should
# run at CPUs frequency
system.cpu_clk_domain = SrcClockDomain(clock = '2GHz',
voltage_domain =
system.voltage_domain)
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
@abstractmethod
def create_system(self):
"""Create an return an initialized system."""
pass
@abstractmethod
def create_root(self):
"""Create and return a simulation root using the system
defined by this class."""
pass
class BaseSESystem(BaseSystem):
"""Basic syscall-emulation builder."""
def __init__(self, **kwargs):
BaseSystem.__init__(self, **kwargs)
def init_system(self, system):
BaseSystem.init_system(self, system)
def create_system(self):
system = System(physmem = self.mem_class(),
membus = SystemXBar(),
mem_mode = self.mem_mode,
multi_thread = (self.num_threads > 1))
system.system_port = system.membus.slave
system.physmem.port = system.membus.master
self.init_system(system)
return system
def create_root(self):
system = self.create_system()
m5.ticks.setGlobalFrequency('1THz')
return Root(full_system=False, system=system)
class BaseSESystemUniprocessor(BaseSESystem):
"""Basic syscall-emulation builder for uniprocessor systems.
Note: This class is only really needed to provide backwards
compatibility in existing test cases.
"""
def __init__(self, **kwargs):
BaseSESystem.__init__(self, **kwargs)
def create_caches_private(self, cpu):
# The atomic SE configurations do not use caches
if self.mem_mode == "timing":
# @todo We might want to revisit these rather enthusiastic L1 sizes
cpu.addTwoLevelCacheHierarchy(L1_ICache(size='128kB'),
L1_DCache(size='256kB'),
L2Cache(size='2MB'))
def create_caches_shared(self, system):
return None
class BaseFSSystem(BaseSystem):
"""Basic full system builder."""
def __init__(self, **kwargs):
BaseSystem.__init__(self, **kwargs)
def init_system(self, system):
BaseSystem.init_system(self, system)
# create the memory controllers and connect them, stick with
# the physmem name to avoid bumping all the reference stats
system.physmem = [self.mem_class(range = r)
for r in system.mem_ranges]
for i in xrange(len(system.physmem)):
system.physmem[i].port = system.membus.master
# create the iocache, which by default runs at the system clock
system.iocache = IOCache(addr_ranges=system.mem_ranges)
system.iocache.cpu_side = system.iobus.master
system.iocache.mem_side = system.membus.slave
def create_root(self):
system = self.create_system()
m5.ticks.setGlobalFrequency('1THz')
return Root(full_system=True, system=system)
class BaseFSSystemUniprocessor(BaseFSSystem):
"""Basic full system builder for uniprocessor systems.
Note: This class is only really needed to provide backwards
compatibility in existing test cases.
"""
def __init__(self, **kwargs):
BaseFSSystem.__init__(self, **kwargs)
def create_caches_private(self, cpu):
cpu.addTwoLevelCacheHierarchy(L1_ICache(size='32kB', assoc=1),
L1_DCache(size='32kB', assoc=4),
L2Cache(size='4MB', assoc=8))
def create_caches_shared(self, system):
return None
class BaseFSSwitcheroo(BaseFSSystem):
"""Uniprocessor system prepared for CPU switching"""
def __init__(self, cpu_classes, **kwargs):
BaseFSSystem.__init__(self, **kwargs)
self.cpu_classes = tuple(cpu_classes)
sim: Add the notion of clock domains to all ClockedObjects This patch adds the notion of source- and derived-clock domains to the ClockedObjects. As such, all clock information is moved to the clock domain, and the ClockedObjects are grouped into domains. The clock domains are either source domains, with a specific clock period, or derived domains that have a parent domain and a divider (potentially chained). For piece of logic that runs at a derived clock (a ratio of the clock its parent is running at) the necessary derived clock domain is created from its corresponding parent clock domain. For now, the derived clock domain only supports a divider, thus ensuring a lower speed compared to its parent. Multiplier functionality implies a PLL logic that has not been modelled yet (create a separate clock instead). The clock domains should be used as a mechanism to provide a controllable clock source that affects clock for every clocked object lying beneath it. The clock of the domain can (in a future patch) be controlled by a handler responsible for dynamic frequency scaling of the respective clock domains. All the config scripts have been retro-fitted with clock domains. For the System a default SrcClockDomain is created. For CPUs that run at a different speed than the system, there is a seperate clock domain created. This domain incorporates the CPU and the associated caches. As before, Ruby runs under its own clock domain. The clock period of all domains are pre-computed, such that no virtual functions or multiplications are needed when calling clockPeriod. Instead, the clock period is pre-computed when any changes occur. For this to be possible, each clock domain tracks its children.
2013-06-27 11:49:49 +02:00
def create_cpus(self, cpu_clk_domain):
cpus = [ cclass(clk_domain = cpu_clk_domain,
cpu_id=0,
switched_out=True)
for cclass in self.cpu_classes ]
cpus[0].switched_out = False
return cpus