Mwait works as follows:
1. A cpu monitors an address of interest (monitor instruction)
2. A cpu calls mwait - this loads the cache line into that cpu's cache.
3. The cpu goes to sleep.
4. When another processor requests write permission for the line, it is
evicted from the sleeping cpu's cache. This eviction is forwarded to the
sleeping cpu, which then wakes up.
Committed by: Nilay Vaish <nilay@cs.wisc.edu>
This changeset adds probe points that can be used to implement PMU
counters for CPU stats. The following probes are supported:
* BaseCPU::ppCycles / Cycles
* BaseCPU::ppRetiredInsts / RetiredInsts
* BaseCPU::ppRetiredLoads / RetiredLoads
* BaseCPU::ppRetiredStores / RetiredStores
* BaseCPU::ppRetiredBranches RetiredBranches
This patch optimises the passing of StaticInstPtr by avoiding copying
the reference-counting pointer. This avoids first incrementing and
then decrementing the reference-counting pointer.
The call paths for de-scheduling a thread are halt() and suspend(), from
the thread context. There is no call to deallocateContext() in general,
though some CPUs chose to define it. This patch removes the function
from BaseCPU and the cores which do not require it.
activate(), suspend(), and halt() used on thread contexts had an optional
delay parameter. However this parameter was often ignored. Also, when used,
the delay was seemily arbitrarily set to 0 or 1 cycle (no other delays were
ever specified). This patch removes the delay parameter and 'Events'
associated with them across all ISAs and cores. Unused activate logic
is also removed.
This patch changes how faults are passed between methods in an attempt
to copy as few reference-counting pointer instances as possible. This
should avoid unecessary copies being created, contributing to the
increment/decrement of the reference counters.
The o3 cpu could attempt to schedule inactive threads under round-robin SMT
mode.
This is because it maintained an independent priority list of threads from the
active thread list. This priority list could be come stale once threads were
inactive, leading to the cpu trying to fetch/commit from inactive threads.
Additionally the fetch queue is now forcibly flushed of instrctuctions
from the de-scheduled thread.
Relevant output:
24557000: system.cpu: [tid:1]: Calling deactivate thread.
24557000: system.cpu: [tid:1]: Removing from active threads list
24557500: system.cpu:
FullO3CPU: Ticking main, FullO3CPU.
24557500: system.cpu.fetch: Running stage.
24557500: system.cpu.fetch: Attempting to fetch from [tid:1]
Check for free entries in Load Queue and Store Queue separately to
avoid cases when load cannot be renamed due to full Store Queue and
vice versa.
This work was done while Binh was an intern at AMD Research.
This patch removes the stat totalCommittedInsts. This variable was used for
recording the total number of instructions committed across all the threads
of a core. The instructions committed by each thread are recorded invidually.
The total would now be generated by summing these individual counts.
A number of calls to isEmpty() and numFreeEntries()
should be thread-specific.
In cpu.cc, the fact that tid is /*commented*/ out is a bug. Say the rob
has instructions from thread 0 (isEmpty() returns false), and none from
thread 1. If we are trying to squash all of thread 1, then
readTailInst(thread 1) will be called because rob->isEmpty() returns
false. The result is end_it is not in the list and the while
statement loops indefinitely back over the cpu's instList.
In iew_impl.hh, all threads are told they have the entire remaining IQ, when
each thread actually has a certain allocation. The result is extra stalls at
the iew dispatch stage which the rename stage usually takes care of.
In commit_impl.hh, rob->readHeadInst(thread 1) can be called if the rob only
contains instructions from thread 0. This returns a dummyInst (which may work
since we are trying to squash all instructions, but hardly seems like the right
way to do it).
In rob_impl.hh this fix skips the rest of the function more frequently and is
more efficient.
Committed by: Nilay Vaish <nilay@cs.wisc.edu>
The probe patch is motivated by the desire to move analytical and trace code
away from functional code. This is achieved by the probe interface which is
essentially a glorified observer model.
What this means to users:
* add a probe point and a "notify" call at the source of an "event"
* add an isolated module, that is being used to carry out *your* analysis (e.g. generate a trace)
* register that module as a probe listener
Note: an example is given for reference in src/cpu/o3/simple_trace.[hh|cc] and src/cpu/SimpleTrace.py
What is happening under the hood:
* every SimObject maintains has a ProbeManager.
* during initialization (src/python/m5/simulate.py) first regProbePoints and
the regProbeListeners is called on each SimObject. this hooks up the probe
point notify calls with the listeners.
FAQs:
Why did you develop probe points:
* to remove trace, stats gathering, analytical code out of the functional code.
* the belief that probes could be generically useful.
What is a probe point:
* a probe point is used to notify upon a given event (e.g. cpu commits an instruction)
What is a probe listener:
* a class that handles whatever the user wishes to do when they are notified
about an event.
What can be passed on notify:
* probe points are templates, and so the user can generate probes that pass any
type of argument (by const reference) to a listener.
What relationships can be generated (1:1, 1:N, N:M etc):
* there isn't a restriction. You can hook probe points and listeners up in a
1:1, 1:N, N:M relationship. They become useful when a number of modules
listen to the same probe points. The idea being that you can add a small
number of probes into the source code and develop a larger number of useful
analysis modules that use information passed by the probes.
Can you give examples:
* adding a probe point to the cpu's commit method allows you to build a trace
module (outputting assembler), you could re-use this to gather instruction
distribution (arithmetic, load/store, conditional, control flow) stats.
Why is the probe interface currently restricted to passing a const reference:
* the desire, initially at least, is to allow an interface to observe
functionality, but not to change functionality.
* of course this can be subverted by const-casting.
What is the performance impact of adding probes:
* when nothing is actively listening to the probes they should have a
relatively minor impact. Profiling has suggested even with a large number of
probes (60) the impact of them (when not active) is very minimal (<1%).
Restructured rename map and free list to clean up some
extraneous code and separate out common code that can
be reused across different register classes (int and fp
at this point). Both components now consist of a set
of Simple* objects that are stand-alone rename map &
free list for each class, plus a Unified* object that
presents a unified interface across all register
classes and then redirects accesses to the appropriate
Simple* object as needed.
Moved free list initialization to PhysRegFile to better
isolate knowledge of physical register index mappings
to that class (and remove the need to pass a number
of parameters to the free list constructor).
Causes a small change to these stats:
cpu.rename.int_rename_lookups
cpu.rename.fp_rename_lookups
because they are now categorized on a per-operand basis
rather than a per-instruction basis.
That is, an instruction with mixed fp/int/misc operand
types will have each operand categorized independently,
where previously the lookup was categorized based on
the instruction type.
It had a bunch of fields (and associated constructor
parameters) thet it didn't really use, and the array
initialization was needlessly verbose.
Also just hardwired the getReg() method to aleays
return true for misc regs, rather than having an array
of bits that we always kept marked as ready.
No need for PhysRegFile to be a template class, or
have a pointer back to the CPU. Also made some methods
for checking the physical register type (int vs. float)
based on the phys reg index, which will come in handy later.
Previously, nextCycle() could return the *current* cycle if the current tick was
already aligned with the clock edge. This behavior is not only confusing (not
quite what the function name implies), but also caused problems in the
drainResume() function. When exiting/re-entering the sim loop (e.g., to take
checkpoints), the CPUs will drain and resume. Due to the previous behavior of
nextCycle(), the CPU tick events were being rescheduled in the same ticks that
were already processed before draining. This caused divergence from runs that
did not exit/re-entered the sim loop. (Initially a cycle difference, but a
significant impact later on.)
This patch separates out the two behaviors (nextCycle() and clockEdge()),
uses nextCycle() in drainResume, and uses clockEdge() everywhere else.
Nothing (other than name) should change except for the drainResume timing.
Virtualized CPUs and the fastmem mode of the atomic CPU require direct
access to physical memory. We currently require caches to be disabled
when using them to prevent chaos. This is not ideal when switching
between hardware virutalized CPUs and other CPU models as it would
require a configuration change on each switch. This changeset
introduces a new version of the atomic memory mode,
'atomic_noncaching', where memory accesses are inserted into the
memory system as atomic accesses, but bypass caches.
To make memory mode tests cleaner, the following methods are added to
the System class:
* isAtomicMode() -- True if the memory mode is 'atomic' or 'direct'.
* isTimingMode() -- True if the memory mode is 'timing'.
* bypassCaches() -- True if caches should be bypassed.
The old getMemoryMode() and setMemoryMode() methods should never be
used from the C++ world anymore.
CPUs need to test that the memory system is in the right mode in two
places, when the CPU is initialized (unless it's switched out) and on
a drainResume(). This led to some code duplication in the CPU
models. This changeset introduces the verifyMemoryMode() method which
is called by BaseCPU::init() if the CPU isn't switched out. The
individual CPU models are responsible for calling this method when
resuming from a drain as this code is CPU model specific.
The changes made by the changeset 9376 were not quite correct. The patch made
changes to the code which resulted in decoder not getting initialized correctly
when the state was restored from a checkpoint.
This patch adds a startup function to each ISA object. For x86, this function
sets the required state in the decoder. For other ISAs, the function is empty
right now.
Cleanup the serialization code for the simple CPUs and the O3 CPU. The
CPU-specific code has been replaced with a (un)serializeThread that
serializes the thread state / context of a specific thread. Assuming
that the thread state class uses the CPU-specific thread state uses
the base thread state serialization code, this allows us to restore a
checkpoint with any of the CPU models.
Previously, the O3 CPU could stop in the middle of a microcode
sequence. This patch makes sure that the pipeline stops when it has
committed a normal instruction or exited from a microcode
sequence. Additionally, it makes sure that the pipeline has no
instructions in flight when it is drained, which should make draining
more robust.
Draining is controlled in the commit stage, which checks if the next
PC after a committed instruction is in microcode. If this isn't the
case, it requests a squash of all instructions after that the
instruction that just committed and immediately signals a drain stall
to the fetch stage. The CPU then continues to execute until the
pipeline and all associated buffers are empty.
The defer_registration parameter is used to prevent a CPU from
initializing at startup, leaving it in the "switched out" mode. The
name of this parameter (and the help string) is confusing. This patch
renames it to switched_out, which should be more descriptive.
This patch cleans up the CPU switching functionality by making sure
that CPU models consistently call the parent on switchOut() and
takeOverFrom(). This has the following implications that might alter
current functionality:
* The call to BaseCPU::switchout() in the O3 CPU is moved from
signalDrained() (!) to switchOut().
* A call to BaseSimpleCPU::switchOut() is introduced in the simple
CPUs.
The O3 CPU used to copy its thread context to a SimpleThread in order
to do serialization. This was a bit of a hack involving two static
SimpleThread instances and a magic constructor that was only used by
the O3 CPU.
This patch moves the ThreadContext serialization code into two global
procedures that, in addition to the normal serialization parameters,
take a ThreadContext reference as a parameter. This allows us to reuse
the serialization code in all ThreadContext implementations.
The entire O3 pipeline used to be initialized from init(), which is
called before initState() or unserialize(). This causes the pipeline
to be initialized from an incorrect thread context. This doesn't
currently lead to correctness problems as instructions fetched from
the incorrect start PC will be squashed a few cycles after
initialization.
This patch will affect the regressions since the O3 CPU now issues its
first instruction fetch to the correct PC instead of 0x0.
This patch adds checks to all CPU models to make sure that the memory
system is in the correct mode at startup and when resuming after a
drain. Previously, we only checked that the memory system was in the
right mode when resuming. This is inadequate since this is a
configuration error that should be detected at startup as well as when
resuming. Additionally, since the check was done using an assert, it
wasn't performed when NDEBUG was set (e.g., the fast target).
The ISA class on stores the contents of ID registers on many
architectures. In order to make reset values of such registers
configurable, we make the class inherit from SimObject, which allows
us to use the normal generated parameter headers.
This patch introduces a Python helper method, BaseCPU.createThreads(),
which creates a set of ISAs for each of the threads in an SMT
system. Although it is currently only needed when creating
multi-threaded CPUs, it should always be called before instantiating
the system as this is an obvious place to configure ID registers
identifying a thread/CPU.
isSyscall was originally created because during handling of a syscall in SE
mode the threadcontext had to be updated. However, in many places this is used
in FS mode (e.g. fault handlers) and the name doesn't make much sense. The
boolean actually stops gem5 from squashing speculative and non-committed state
when a write to a threadcontext happens, so re-name the variable to something
more appropriate
This patch moves the draining interface from SimObject to a separate
class that can be used by any object needing draining. However,
objects not visible to the Python code (i.e., objects not deriving
from SimObject) still depend on their parents informing them when to
drain. This patch also gets rid of the CountedDrainEvent (which isn't
really an event) and replaces it with a DrainManager.
This patch addresses the comments and feedback on the preceding patch
that reworks the clocks and now more clearly shows where cycles
(relative cycle counts) are used to express time.
Instead of bumping the existing patch I chose to make this a separate
patch, merely to try and focus the discussion around a smaller set of
changes. The two patches will be pushed together though.
This changes done as part of this patch are mostly following directly
from the introduction of the wrapper class, and change enough code to
make things compile and run again. There are definitely more places
where int/uint/Tick is still used to represent cycles, and it will
take some time to chase them all down. Similarly, a lot of parameters
should be changed from Param.Tick and Param.Unsigned to
Param.Cycles.
In addition, the use of curTick is questionable as there should not be
an absolute cycle. Potential solutions can be built on top of this
patch. There is a similar situation in the o3 CPU where
lastRunningCycle is currently counting in Cycles, and is still an
absolute time. More discussion to be had in other words.
An additional change that would be appropriate in the future is to
perform a similar wrapping of Tick and probably also introduce a
Ticks class along with suitable operators for all these classes.
This patch introduces the notion of a clock update function that aims
to avoid costly divisions when turning the current tick into a
cycle. Each clocked object advances a private (hidden) cycle member
and a tick member and uses these to implement functions for getting
the tick of the next cycle, or the tick of a cycle some time in the
future.
In the different modules using the clocks, changes are made to avoid
counting in ticks only to later translate to cycles. There are a few
oddities in how the O3 and inorder CPU count idle cycles, as seen by a
few locations where a cycle is subtracted in the calculation. This is
done such that the regression does not change any stats, but should be
revisited in a future patch.
Another, much needed, change that is not done as part of this patch is
to introduce a new typedef uint64_t Cycle to be able to at least hint
at the unit of the variables counting Ticks vs Cycles. This will be
done as a follow-up patch.
As an additional follow up, the thread context still uses ticks for
the book keeping of last activate and last suspend and this should
probably also be changed into cycles as well.
This patch makes the Tick unsigned and removes the UTick typedef. The
ticks should never be negative, and there was only one major issue
with removing it, caused by the o3 CPU using a -1 as an initial value.
The patch has no impact on any regressions.
This patch fixes some problems with the drain/switchout functionality
for the O3 cpu and for the ARM ISA and adds some useful debug print
statements.
This is an incremental fix as there are still a few bugs/mem leaks with the
switchout code. Particularly when switching from an O3CPU to a
TimingSimpleCPU. However, when switching from O3 to O3 cores with the ARM ISA
I haven't encountered any more assertion failures; now the kernel will
typically panic inside of simulation.
This patch moves send/recvTiming and send/recvTimingSnoop from the
Port base class to the MasterPort and SlavePort, and also splits them
into separate member functions for requests and responses:
send/recvTimingReq, send/recvTimingResp, and send/recvTimingSnoopReq,
send/recvTimingSnoopResp. A master port sends requests and receives
responses, and also receives snoop requests and sends snoop
responses. A slave port has the reciprocal behaviour as it receives
requests and sends responses, and sends snoop requests and receives
snoop responses.
For all MemObjects that have only master ports or slave ports (but not
both), e.g. a CPU, or a PIO device, this patch merely adds more
clarity to what kind of access is taking place. For example, a CPU
port used to call sendTiming, and will now call
sendTimingReq. Similarly, a response previously came back through
recvTiming, which is now recvTimingResp. For the modules that have
both master and slave ports, e.g. the bus, the behaviour was
previously relying on branches based on pkt->isRequest(), and this is
now replaced with a direct call to the apprioriate member function
depending on the type of access. Please note that send/recvRetry is
still shared by all the timing accessors and remains in the Port base
class for now (to maintain the current bus functionality and avoid
changing the statistics of all regressions).
The packet queue is split into a MasterPort and SlavePort version to
facilitate the use of the new timing accessors. All uses of the
PacketQueue are updated accordingly.
With this patch, the type of packet (request or response) is now well
defined for each type of access, and asserts on pkt->isRequest() and
pkt->isResponse() are now moved to the appropriate send member
functions. It is also worth noting that sendTimingSnoopReq no longer
returns a boolean, as the semantics do not alow snoop requests to be
rejected or stalled. All these assumptions are now excplicitly part of
the port interface itself.
This patch introduces port access methods that separates snoop
request/responses from normal memory request/responses. The
differentiation is made for functional, atomic and timing accesses and
builds on the introduction of master and slave ports.
Before the introduction of this patch, the packets belonging to the
different phases of the protocol (request -> [forwarded snoop request
-> snoop response]* -> response) all use the same port access
functions, even though the snoop packets flow in the opposite
direction to the normal packet. That is, a coherent master sends
normal request and receives responses, but receives snoop requests and
sends snoop responses (vice versa for the slave). These two distinct
phases now use different access functions, as described below.
Starting with the functional access, a master sends a request to a
slave through sendFunctional, and the request packet is turned into a
response before the call returns. In a system without cache coherence,
this is all that is needed from the functional interface. For the
cache-coherent scenario, a slave also sends snoop requests to coherent
masters through sendFunctionalSnoop, with responses returned within
the same packet pointer. This is currently used by the bus and caches,
and the LSQ of the O3 CPU. The send/recvFunctional and
send/recvFunctionalSnoop are moved from the Port super class to the
appropriate subclass.
Atomic accesses follow the same flow as functional accesses, with
request being sent from master to slave through sendAtomic. In the
case of cache-coherent ports, a slave can send snoop requests to a
master through sendAtomicSnoop. Just as for the functional access
methods, the atomic send and receive member functions are moved to the
appropriate subclasses.
The timing access methods are different from the functional and atomic
in that requests and responses are separated in time and
send/recvTiming are used for both directions. Hence, a master uses
sendTiming to send a request to a slave, and a slave uses sendTiming
to send a response back to a master, at a later point in time. Snoop
requests and responses travel in the opposite direction, similar to
what happens in functional and atomic accesses. With the introduction
of this patch, it is possible to determine the direction of packets in
the bus, and no longer necessary to look for both a master and a slave
port with the requested port id.
In contrast to the normal recvFunctional, recvAtomic and recvTiming
that are pure virtual functions, the recvFunctionalSnoop,
recvAtomicSnoop and recvTimingSnoop have a default implementation that
calls panic. This is to allow non-coherent master and slave ports to
not implement these functions.
This patch unifies where initMemProxies is called, in the init()
method of each BaseCPU subclass, before TheISA::initCPU is
called. Moreover, it also ensures that initMemProxies is called in
both full-system and syscall-emulation mode, thus unifying also across
the modes. An additional check is added in the ThreadState to ensure
that initMemProxies is only called once.
Enables the CheckerCPU to be selected at runtime with the --checker option
from the configs/example/fs.py and configs/example/se.py configuration
files. Also merges with the SE/FS changes.
This patch adds a creation-time check to the CPU to ensure that the
interrupt controller is created for the cases where it is needed,
i.e. if the CPU is not being switched in later and not a checker CPU.
The patch also adds the "createInterruptController" call to a number
of the regression scripts.
This patch continues the unification of how the different CPU models
create and share their instruction and data ports. Most importantly,
it forces every CPU to have an instruction and a data port, and gives
these ports explicit getters in the BaseCPU (getDataPort and
getInstPort). The patch helps in simplifying the code, make
assumptions more explicit, andfurther ease future patches related to
the CPU ports.
The biggest changes are in the in-order model (that was not modified
in the previous unification patch), which now moves the ports from the
CacheUnit to the CPU. It also distinguishes the instruction fetch and
load-store unit from the rest of the resources, and avoids the use of
indices and casting in favour of keeping track of these two units
explicitly (since they are always there anyways). The atomic, timing
and O3 model simply return references to their already existing ports.
This patch adds the necessary flags to the SConstruct and SConscript
files for compiling using clang 2.9 and later (on Ubuntu et al and OSX
XCode 4.2), and also cleans up a bunch of compiler warnings found by
clang. Most of the warnings are related to hidden virtual functions,
comparisons with unsigneds >= 0, and if-statements with empty
bodies. A number of mismatches between struct and class are also
fixed. clang 2.8 is not working as it has problems with class names
that occur in multiple namespaces (e.g. Statistics in
kernel_stats.hh).
clang has a bug (http://llvm.org/bugs/show_bug.cgi?id=7247) which
causes confusion between the container std::set and the function
Packet::set, and this is currently addressed by not including the
entire namespace std, but rather selecting e.g. "using std::vector" in
the appropriate places.
Brings the CheckerCPU back to life to allow FS and SE checking of the
O3CPU. These changes have only been tested with the ARM ISA. Other
ISAs potentially require modification.
This patch performs minimal changes to move the instruction and data
ports from specialised subclasses to the base CPU (to the largest
degree possible). Ultimately it servers to make the CPU(s) have a
well-defined interface to the memory sub-system.
