This patch changes the IEW drain check to include the FU pool as there
can be instructions that are "stored" in FU completion events and thus
not covered by the existing checks. With this patch, we simply include
a check to see if all the FUs are considered non-busy in the next
tick.
Without this patch, the pc-switcheroo-full regression fails after
minor changes to the cache timing (aligning to clock edge).
This patch fixes an outstanding issue in the cache timing calculations
where an atomic access returned a time in Cycles, but the port
forwarded it on as if it was in Ticks.
A separate patch will update the regression stats.
This patch fixes a bug in the granularity calculation. For example, if
the high bit is 6 (counting from 0) and we have one interleaving bit,
then the granularity is now 2 ** (6 - 1 + 1) = 64.
This patch changes the updards snoop packet to avoid allocating and
later deleting it. As the code executes in 0 time and the lifetime of
the packet does not extend beyond the block there is no reason to heap
allocate it.
This patch removes the printing of the SparseHist total in the
stats.txt output file. This has been removed as a sparse histogram has
no total, and therefore this was printing out the value of a
non-local, unrelated variable.
This patch adds separate actions for requests that missed in the local cache
and messages were sent out to get the requested line. These separate actions
are required for differentiating between the hit and miss latencies in the
statistics collected.
This patch adds separate actions for requests that missed in the local cache
and messages were sent out to get the requested line. These separate actions
are required for differentiating between the hit and miss latencies in the
statistics collected.
The patch started of with removing the global variables from the profiler for
profiling the miss latency of requests made to the cache. The corrresponding
histograms have been moved to the Sequencer. These are combined together when
the histograms are printed. Separate histograms are now maintained for
tracking latency of all requests together, of hits only and of misses only.
A particular set of histograms used to use the type GenericMachineType defined
in one of the protocol files. This patch removes this type. Now, everything
that relied on this type would use MachineType instead. To do this, SLICC has
been changed so that multiple machine types can be declared by a controller
in its preamble.
This patch removes the following three files: RubySlicc_Profiler.sm,
RubySlicc_Profiler_interface.cc and RubySlicc_Profiler_interface.hh.
Only one function prototyped in the file RubySlicc_Profiler.sm. Rest of the
code appearing in any of these files is not in use. Therefore, these files
are being removed.
That one single function, profileMsgDelay(), is being moved to the protocol
files where it is in use. If we need any of these deleted functions, I think
the right way to make them visible is to have the AbstractController class in
a .sm and let the controller state machine inherit from this class. The
AbstractController class can then have the prototypes of these profiling
functions in its definition.
2013-06-24 08:59:08 -05:00
Joel Hestness ext:(%2C%20Nilay%20Vaish%20%3Cnilay%40cs.wisc.edu%3E)
The m_size variable attempted to track m_prio_heap.size(), but it did so
incorrectly due to the functions reanalyzeMessages and reanalyzeAllMessages().
Since this variable is intended to track m_prio_heap.size(), we can simply
replace instances where m_size is referenced with m_prio_heap.size(), which
has the added bonus of removing the need for m_size.
Note: This patch also removes an extraneous DPRINTF format string designator
from reanalyzeAllMessages()
Committed by: Nilay Vaish <nilay@cs.wisc.edu>
Previously, .sm files were allowed to use the same name for a type and a
variable. This is unnecessarily confusing and has some bad side effects, like
not being able to declare later variables in the same scope with the same type.
This causes the compiler to complain and die on things like Address Address.
Committed by: Nilay Vaish <nilay@cs.wisc.edu>
Change all occurrances of Address as a variable name to instead use Addr.
Address is an allowed name in slicc even when Address is also being used as a
type, leading to declarations of "Address Address". While this works, it
prevents adding another field of type Address because the compiler then thinks
Address is a variable name, not type.
Committed by: Nilay Vaish <nilay@cs.wisc.edu>
The current implementation of the x87 never updates the x87 tag
word. This is currently not a big issue since the simulated x87 never
checks for stack overflows, however this becomes an issue when
switching between a virtualized CPU and a simulated CPU. This
changeset adds support, which is enabled by default, for updating the
tag register to every floating point microop that updates the stack
top using the spm mechanism.
The new tag words is generated by the helper function
X86ISA::genX87Tags(). This function is currently limited to flagging a
stack position as valid or invalid and does not try to distinguish
between the valid, zero, and special states.
This changeset actually fixes two issues:
* The lfpimm instruction didn't work correctly when applied to a
floating point constant (it did work for integers containing the
bit string representation of a constant) since it used
reinterpret_cast to convert a double to a uint64_t. This caused a
compilation error, at least, in gcc 4.6.3.
* The instructions loading floating point constants in the x87
processor didn't work correctly since they just stored a truncated
integer instead of a double in the floating point register. This
changeset fixes the old microcode by using lfpimm instruction
instead of the limm instructions.
The current implementation of fprem simply does an fmod and doesn't
simulate any of the iterative behavior in a real fprem. This isn't
normally a problem, however, it can lead to problems when switching
between CPU models. If switching from a real CPU in the middle of an
fprem loop to a simulated CPU, the output of the fprem loop becomes
correupted. This changeset changes the fprem implementation to work
like the one on real hardware.
Reuse the address finalization code in the TLB instead of replicating
it when handling MMIO. This patch also adds support for injecting
memory mapped IPR requests into the memory system.
The rflags register is spread across several different registers. Most
of the flags are stored in MISCREG_RFLAGS, but some are stored in
microcode registers. When accessing RFLAGS, we need to reconstruct it
from these registers. This changeset adds two functions,
X86ISA::getRFlags() and X86ISA::setRFlags(), that take care of this
magic.
This changeset fixes two problems in the FABS and FCHS
implementation. First, the ISA parser expects the assignment in
flag_code to be a pure assignment and not an and-assignment, which
leads to the isa_parser omitting the misc reg update. Second, the FCHS
and FABS macro-ops don't set the SetStatus flag, which means that the
default micro-op version, which doesn't update FSW, is executed.
This changeset adds the following stats to KVM:
* numVMHalfEntries: Number of entries into KVM to finalize pending
IO operations without executing guest instructions. These typically
happen as a result of a drain where the guest must finalize some
operations before the guest state is consistent.
* numExitSignal: Number of VM exits that have been triggered by a
signal. These usually happen as a result of the timer that limits
the time spent in KVM.
We used to use the KVM CPU's clock to specify the host frequency. This
was not ideal for several reasons. One of them being that the clock
parameter of a CPU determines the frequency of some of the components
connected to the CPU. This changeset adds a separate hostFreq
parameter that should be used to specify the host frequency until we
add code to autodetect it. The hostFactor should still be used to
specify the conversion factor between the host performance and that of
the simulated system.
We currently execute instructions in the guest and then handle any IO
request right after we break out of the virtualized environment. This
has the effect of executing IO requests in the exact same tick as the
first instruction in the sequence that was just run. There seem to be
cases where this simplification upsets some timing-sensitive devices.
This changeset splits execute and IO (and other services) across
multiple ticks. This is implemented by adding a separate
RunningService state to the CPU state machine. When a VM requires
service, it enters into this state and pending IO is then serviced in
the future instead of immediately. The delay between getting the
request and servicing it depends on the number of cycles executed in
the guest, which allows other components to catch up with the CPU.
Update the system's totalNumInst counter when exiting from KVM and
maintain an internal absolute instruction count instead of relying on
the one from perf.
The TSC value stored in MISCREG_TSC is actually just an offset from
the current CPU cycle to the actual TSC value. Writes with
side-effects to the TSC subtract the current cycle count before
storing the new value, while reads add the current cycle count. When
switching CPUs, the current value is copied without side-effects. This
works as long as the source and the destination CPUs have the same
clock frequencies. The TSC will jump, sometimes backwards, if they
have different clock frequencies. Most OSes assume the TSC to be
monotonic and break when this happens.
This changeset makes sure that the TSC is copied with side-effects to
ensure that the offset is updated to match the new CPU.
HG changset 34e3295b0e39 introduced a check in the main simulation
loop that discards exit events that happen at the same tick as another
exit event. This was supposed to fix a problem where a simulation
script got confused by multiple exit events. This obviously breaks the
simulator since it can hide important simulation events, such as a
simulation failure, that happen at the same time as a non-fatal
simulation event.
Currently, the only way to get a CPU to stop after a fixed number of
instructions/loads is to set a property on the CPU that causes a
SimLoopExitEvent to be scheduled when the CPU is constructed. This is
clearly not ideal in cases where the simulation script wants the CPU
to stop at multiple instruction counts (e.g., SimPoint generation).
This changeset adds the methods scheduleInstStop() and
scheduleLoadStop() to the BaseCPU. These methods are exported to
Python and are designed to be used from the simulation script. By
using these methods instead of the old properties, a simulation script
can schedule a stop at any point during simulation or schedule
multiple stops. The number of instructions specified when scheduling a
stop is relative to the current point of execution.
This patch removes per processor cycle count, histogram for filter stats,
histogram for multicasts, histogram for prefetch wait, some function
prototypes that do not have definitions.
The Profiler class does not need an event for dumping statistics
periodically. This is because there is a method for dumping statistics
for all the sim objects periodically. Since Ruby is a sim object, its
statistics are also included.
This moves event and transition count statistics for cache controllers to
gem5's statistics. It does the same for the statistics associated with the
memory controller in ruby.
All the cache/directory/dma controllers individually collect the event and
transition counts. A callback function, collateStats(), has been added that
is invoked on the controller version 0 of each controller class. This
function adds all the individual controller statistics to a vector
variables. All the code for registering the statistical variables and
collating them is generated by SLICC. The patch removes the files
*_Profiler.{cc,hh} and *_ProfileDumper.{cc,hh} which were earlier used for
collecting and dumping statistics respectively.
This patch adds a new flag to specify if the data values for a given vector
should be printed in one line in the stats.txt file. The default behavior
will be to print the data in multiple lines. It makes changes to print
functions to enforce this behavior.
in the TLB
Some architectures (currently only x86) require some fixing-up of
physical addresses after a normal address translation. This is usually
to remap devices such as the APIC, but could be used for other memory
mapped devices as well. When running the CPU in a using hardware
virtualization, we still need to do these address fix-ups before
inserting the request into the memory system. This patch moves this
patch allows that code to be used by such CPUs without doing full
address translations.
The custom Python loader didn't comply with PEP302 for two reasons:
* Previously, we would overwrite old modules on name
conflicts. PEP302 explicitly states that: "If there is an existing
module object named 'fullname' in sys.modules, the loader must use
that existing module".
* The "__package__" attribute wasn't set. PEP302: "The __package__
attribute must be set."
This changeset addresses both of these issues.
Some architectures have special registers in the guest that can be
used to do cycle accounting. This is generally preferrable since the
prevents the guest from seeing a non-monotonic clock. This changeset
adds a virtual method, getHostCycles(), that the architecture-specific
code can override to implement this functionallity. The default
implementation uses the hwCycles counter.
timer_create can apparently return -1 and set errno to EAGAIN if the
kernel suffered a temporary failure when allocating a timer. This
happens from time to time, so we need to handle it.
It is now required to initialize the thread context by calling
startup() on it. Failing to do so currently causes decoder in
x86-based CPUs to get very confused when restoring from checkpoints.
Some Linux versions disable updates (regB.set = 1) to prevent the chip
from updating its internal state while the OS is updating it. Support
for this was already there, this patch merely disables the check in
writeReg that prevented it from being enabled. The patch also includes
support for disabling the divider, which is used to control when clock
updates should start after setting the internal RTC state.
These changes are required to boot most vanilla Linux distributions
that update the RTC settings at boot.
Rewrite reg A & B handling to use the bitunion stuff instead of bit
masking. Add better error messages when the kernel tries to enable
unsupported stuff.
This patch changes the class names of the variuos DRAM configurations
to better reflect what memory they are based on. The speed and
interface width is now part of the name, and also the alias that is
used to select them on the command line.
Some minor changes are done to the actual parameters, to better
reflect the named configurations. As a result of these changes the
regressions change slightly and the stats will be bumped in a separate
patch.
This patch adds a histogram to track how many bytes are accessed in an
open row before it is closed. This metric is useful in characterising
a workload and the efficiency of the DRAM scheduler. For example, a
DDR3-1600 device requires 44 cycles (tRC) before it can activate
another row in the same bank. For a x32 interface (8 bytes per cycle)
that means 8 x 44 = 352 bytes must be transferred to hide the
preparation time.
This patch adds a frontend and backend static latency to the DRAM
controller by delaying the responses. Two parameters expressing the
frontend and backend contributions in absolute time are added to the
controller, and the appropriate latency is added to the responses when
adding them to the (infinite) queued port for sending.
For writes and reads that hit in the write buffer, only the frontend
latency is added. For reads that are serviced by the DRAM, the static
latency is the sum of the pipeline latencies of the entire frontend,
backend and PHY. The default values are chosen based on having roughly
10 pipeline stages in total at 500 MHz.
In the future, it would be sensible to make the controller use its
clock and convert these latencies (and a few of the DRAM timings) to
cycles.
This patch does some minor tidying up of the MSHR and MSHRQueue. The
clean up started as part of some ad-hoc tracing and debugging, but
seems worthwhile enough to go in as a separate patch.
The highlights of the changes are reduced scoping (private) members
where possible, avoiding redundant new/delete, and constructor
initialisation to please static code analyzers.
This patch prunes the TraceCPU as the code is stale and the
functionality that it provided can now be achieved with the TrafficGen
using its trace playback mode.
The TraceCPU was able to play back pre-recorded memory traces of a few
different formats, and to achieve this level of flexibility with the
TrafficGen, use the util/encode_packet_trace (with suitable
modifications) to create a protobuf trace off-line.
Add a check which ensures that the minumum period for the LINEAR and
RANDOM traffic generator states is less than or equal to the maximum
period. If the minimum period is greater than the maximum period a
fatal is triggered.
This patch fixes a bug with the traffic generator which occured when
reading in the state transitions from the configuration
file. Previously, the size of the vector which stored the transitions
was used to get the size of the transitions matrix, rather than using
the number of states. Therefore, if there were more transitions than
states, i.e. some transitions has a probability of less than 1, then
the traffic generator would fatal when trying to check the
transitions.
This issue has been addressed by using the number of input states,
rather then the number of transitions.
This patch adds an optional request elasticity to the traffic
generator, effectievly compensating for it in the case of the linear
and random generators, and adding it in the case of the trace
generator. The accounting is left with the top-level traffic
generator, and the individual generators do the necessary math as part
of determining the next packet tick.
Note that in the linear and random generators we have to compensate
for the blocked time to not be elastic, i.e. without this patch the
aforementioned generators will slow down in the case of back-pressure.
This patch changes the queued port for a conventional master port and
stalls the traffic generator when requests are not immediately
accepted. This is a first step to allowing elasticity in the injection
of requests.
The patch also adds stats for the sent packets and retries, and
slightly changes how the nextPacketTick and getNextPacket
interact. The advancing of the trace is now moved to getNextPacket and
nextPacketTick is only responsible for answering the question when the
next packet should be sent.
This patch moves the responsibility for sending packets out of the
generator states and leaves it with the top-level traffic
generator. The main aim of this patch is to enable a transition to
non-queued ports, i.e. with send/retry flow control, and to do so it
is much more convenient to not wrap the port interactions and instead
leave it all local to the traffic generator.
The generator states now only govern when they are ready to send
something new, and the generation of the packets to send. They thus
have no knowledge of the port that is used.
This patch simplifies the object hierarchy of the traffic generator by
getting rid of the StateGraph class and folding this functionality
into the traffic generator itself.
The main goal of this patch is to facilitate upcoming changes by
reducing the number of affected layers.
This patch introduces a mirrored internal snoop port to facilitate
easy addition of flow control for the snoop responses that are turned
into normal responses on their return. To perform this, the slave
ports of the coherent bus are wrapped in internal master ports that
are passed as the source ports to the response layer in question.
As a result of this patch, there is more contention for the response
resources, and as such system performance will decrease slightly.
A consequence of the mirrored internal port is that the port the bus
tells to retry (the internal one) and the port actually retrying (the
mirrored) one are not the same. Thus, the existing check in tryTiming
is not longer correct. In fact, the test is redundant as the layer is
only in the retry state while calling sendRetry on the waiting port,
and if the latter does not immediately call the bus then the retry
state is left. Consequently the check is removed.
This patch makes the buses multi layered, and effectively creates a
crossbar structure with distributed contention ports at the
destination ports. Before this patch, a bus could have a single
request, response and snoop response in flight at any time, and with
these changes there can be as many requests as connected slaves (bus
master ports), and as many responses as connected masters (bus slave
ports).
Together with address interleaving, this patch enables us to create
high-throughput memory interconnects, e.g. 50+ GByte/s.
This patch makes the flow control and state updates of the coherent
bus more clear by separating the two cases, i.e. forward as a snoop
response, or turn it into a normal response.
With this change it is also more clear what resources are being
occupied, and that we effectively bypass the busy check for the second
case. As a result of the change in resource usage some stats change.
This patch does some minor housekeeping on the bus code, removing
redundant code, and moving the extraction of the destination id to the
top of the functions using it.
This patch adds a basic set of stats which are hard to impossible to
implement using only communication monitors, and are needed for
insight such as bus utilization, transactions through the bus etc.
Stats added include throughput and transaction distribution, and also
a two-dimensional vector capturing how many packets and how much data
is exchanged between the masters and slaves connected to the bus.
This patch changes the set used to track outstanding requests to an
unordered set (part of C++11 STL). There is no need to maintain the
order, and hopefully there might even be a small performance benefit.
This patch adds a typical (leaning towards fast) LPDDR3 configuration
based on publically available data. As expected, it looks very similar
to the LPDDR2-S4 configuration, only with a slightly lower burst time.
This patch adapts the existing LPDDR2 configuration to make use of the
multi-channel functionality. Thus, to get a x64 interface two
controllers should be instantiated using the makeMultiChannel method.
The page size and ranks are also adapted to better suit with a typical
LPDDR2 part.
This patch removes the explicit memset as it is redundant and causes
the simulator to touch the entire space, forcing the host system to
allocate the pages.
Anonymous pages are mapped on the first access, and the page-fault
handler is responsible for zeroing them. Thus, the pages are still
zeroed, but we avoid touching the entire allocated space which enables
us to use much larger memory sizes as long as not all the memory is
actually used.
This patch changes how the streams are created to avoid the size
limitation on the coded streams. As we only read/write a single
message at a time, there is never any message larger than a few
bytes. However, the coded stream eventually complains that its
internal counter reaches 64+ MByte if the total file size exceeds this
value.
Based on suggestions in the protobuf discussion forums, the coded
stream is now created for every message that is read/written. The
result is that the internal byte count never goes about tens of bytes,
and we can read/write any size file that the underlying file I/O can
handle.
This patch changes the type of the hash function for BasicBlockRanges
to match the original definition of the templatized type. Without
this, clang raises a warning and combined with the "-Werror" flag this
causes compilation to fail.
This is the x86 version of the ARM changeset baa17ba80e06. In case an
instruction has been squashed by the o3 cpu, this patch allows page
table walker to avoid carrying out a pending translation that the
instruction requested for.
Currently call and return instructions are marked as IsCall and IsReturn. Thus, the
branch predictor does not use RAS for these instructions. Similarly, the number of
function calls that took place is recorded as 0. This patch marks these instructions
as they should be.
Currently all the integer microops are marked as IntAluOp and the floating
point microops are marked as FloatAddOp. This patch adds support for marking
different microops differently. Now IntMultOp, IntDivOp, FloatDivOp,
FloatMultOp, FloatCvtOp, FloatSqrtOp classes will be used as well. This will
help in providing different latencies for different op class.
This patch changes the way cache statistics are collected in ruby.
As of now, there is separate entity called CacheProfiler which holds
statistical variables for caches. The CacheMemory class defines different
functions for accessing the CacheProfiler. These functions are then invoked
in the .sm files. I find this approach opaque and prone to error. Secondly,
we probably should not be paying the cost of a function call for recording
statistics.
Instead, this patch allows for accessing statistical variables in the
.sm files. The collection would become transparent. Secondly, it would happen
in place, so no function calls. The patch also removes the CacheProfiler class.
--HG--
rename : src/mem/slicc/ast/InfixOperatorExprAST.py => src/mem/slicc/ast/OperatorExprAST.py
having separate params for the local/globalHistoryBits and the
local/globalPredictorSize can lead to inconsistencies if they
are not carefully set. this patch dervies the number of bits
necessary to index into the local/global predictors based on
their size.
the value of the localHistoryTableSize for the ARM O3 CPU has been
increased to 1024 from 64, which is more accurate for an A15 based
on some correlation against A15 hardware.
The CpuPort class was removed before the KVM patches were committed,
which means that the KVM interface currently doesn't compile. This
changeset adds the BaseKvmCPU::KVMCpuPort class which derives from
MasterPort. This class is used on the data and instruction ports
instead of the old CpuPort.
This changeset adds a 'numInsts' stat to the KVM-based CPU. It also
cleans up the variable names in kvmRun to make the distinction between
host cycles and estimated simulated cycles clearer. As a bonus
feature, it also fixes a warning (unreferenced variable) when
compiling in fast mode.
Add a debug print (when the Checkpoint debug flag is set) on serialize
and unserialize. Additionally, dump the KVM state before
serializing. The KVM state isn't dumped after unserializing since the
state is loaded lazily on the next KVM entry.
Device accesses are normally uncacheable. This change probably doesn't
make any difference since we normally disable caching when KVM is
active. However, there might be devices that check this, so we'd
better enable this flag to be safe.
The vsyscall address for gettimeofday is 0xffffffffff600000ul. The offset
therefore should be 0x0 instead of 0x410. This can be cross checked with
the file sysdeps/unix/sysv/linux/x86_64/gettimeofday.c in source of glibc.
Committed by: Nilay Vaish <nilay@cs.wisc.edu>
The existing implementation can read uninitialized data or stale information
from the cached PageTable entries.
1) Add a valid bit for the cache entries. Simply using zero for the virtual
address to signify invalid entries is not sufficient. Speculative, wrong-path
accesses frequently access page zero. The current implementation would return
a uninitialized TLB entry when address zero was accessed and the PageTable
cache entry was invalid.
2) When unmapping/mapping/remaping a page, invalidate the corresponding
PageTable cache entry if one already exists.
The 'lret' instruction reloads instruction pointer and code segment from the
stack and then pops them. But the popping part is missing from the current
implementation. This caused incorrect behavior in some code related to the
Fiasco OS. Microops are being added to rectify the behavior of the instruction.
Committed by: Nilay Vaish <nilay@cs.wisc.edu>
Due to recent changes to clocking system in Ruby and the way Ruby restores
state from a checkpoint, garnet was failing to run from a checkpointed state.
The problem is that Ruby resets the time to zero while warming up the caches.
If any component records a local copy of the time (read calls curCycle())
before the simulation has started, then that component will not operate until
that time is reached. In the context of this particular patch, the Garnet
Network class calls curCycle() at multiple places. Any non-operational
component can block in requests in the memory system, which the system
interprets as a deadlock. This patch makes changes so that Garnet can
successfully run from checkpointed state.
It adds a globally visible time at which the actual execution started. This
time is initialized in RubySystem::startup() function. This variable is only
meant for components with in Ruby. This replaces the private variable that
was maintained within Garnet since it is not possible to figure out the
correct time when the value of this variable can be set.
The patch also does away with all cases where curCycle() is called with in
some Ruby component before the system has actually started executing. This
is required due to the quirky manner in which ruby restores from a checkpoint.
This patch adds an address mapping scheme where the channel
interleaving takes place on a cache line granularity. It is similar to
the existing RaBaChCo that interleaves on a DRAM page, but should give
higher performance when there is less locality in the address
stream.
This patch changes the slightly ambigious names used for the address
mapping scheme to be more descriptive, and actually spell out what
they do. With this patch we also open up for adding more flavours of
open- and close-type mappings, i.e. interleaving across channels with
the open map.
This patch enables the use of the generator behaviours outside the
TrafficGen module. This is useful e.g. to allow packet replay modes
for other devices in the system without having to replace them with a
TrafficGen in the configuration files.
This change also enables more specific behaviours to be composed as
specific modules, e.g. BaseBandModem can use a number of generators
and have application-specific parameters based around a specific set
of generators.
This patch adds a WideIO 200 MHz configuration that can be used as a
baseline to compare with DDRx and LPDDRx. Note that it is a single
channel and that it should be replicated 4 times. It is based on
publically available information and attempts to capture an envisioned
8 Gbit single-die part (i.e. without TSVs).
This patch provides useful printouts throughut the memory system. This
includes pretty-printed cache tags and function call messages
(call-stack like).
