| Age | Commit message (Collapse) | Author |
|---|
|
Suspending the current thread context while draining due to a quiesce
pseudo instruction (for example a wfi instruction) could deadlock the
cpu and prevent it from successfully draining. This change ensures
that the cpu is not draining before suspending the thread context.
Change-Id: I7c019847f5a870d4bc9ce2b19936bc3dc45e5fd7
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5881
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
Add the power_gating_on_idle option to control whether a core
automatically enters the power gated state. The default behaviour is
to transition to clock gated when idle, but not to power gated. When
this option is set to true, the core automatically transitions to the
power gated state after a configurable latency.
Change-Id: Ida98c7fc532de4140d0e511c25613769b47b3702
Reviewed-on: https://gem5-review.googlesource.com/5741
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
If the CPU has been clock gated for a sufficient amount of time
(configurable via pwrGatingLatency), the CPU will go into the OFF
power state. This does not model hardware, just behaviour.
Change-Id: Ib3681d1ffa6ad25eba60f47b4020325f63472d43
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/3969
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
Change-Id: I017852eac183fac3f914fdb96d7e72a56ea9d682
Reviewed-by: Nathanael Premillieu <nathanael.premillieu@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5121
Reviewed-by: Matthias Jung <jungma@eit.uni-kl.de>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
Fixes compile error for gem5.fast on CLANG due to unused variable.
Change-Id: Iabe777a27d75ee8bfa7b214fff577aed3c7582c7
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-on: https://gem5-review.googlesource.com/4980
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
|
|
The size of the store entry in the LSQ is used to indicate a fault in
the execution of the store. At the same time, a store that is
predicated false will also have 0 size in the corresponding store
queue entry. This changeset ensures that we check if the store was
predicated false before checking the size field. This way we avoid
printing stores as faulting when they are only predicated false.
Change-Id: Ie07982197bd73d7b44d26a3257d54ecb103a952a
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/4821
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
The O3CPU allows stores to commit before they are completed and as
soon as they enter the store queue. This is the reason why stores are
verified by the the checker CPU, separately, once they complete
and after they are sent to the memory.
Store conditionals, on the other hand, have an additional writeback
stage in the pipeline as they return their result to a register,
similarly to loads. This is the reason why they do not commit
before they receive a response from the memory. This allows store
conditionals to be verified by the checker CPU as soon as they
commit in the same way as all other non-store insturctions.
At the same time, the presense of a checker CPU should not require
changes to way we handle instructions. This change removes explicit
calls to:
* incorrectly set the extra data of the request to 0 (a subsequent
call to completeAcc already does this without making any ISA
assumptions about the return value of the failed store conditional)
* complete failing store conditionals
Change-Id: If21d70b21caa55b35e9fdcc50f254c590465d3c3
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/4820
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
The kernel stat mechanism should really be refactored and moved somewhere
else, but in the mean time there's some old cruft that can be cleared away.
Change-Id: I21e725de590dda0d20bf3bc675bbe976c7b1bd86
Reviewed-on: https://gem5-review.googlesource.com/4600
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
When different sizes were set for the choice and global saturation
counter (e.g. ex5_big), the threshold calculation used the wrong
size. Thus the branch predictor always predicted "not taken" for
choice > global.
Change-Id: I076549ff1482e2280cef24a0d16b7bb2122d4110
Reviewed-on: https://gem5-review.googlesource.com/4560
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
|
|
Coverage was wrongly set to PartialAddrRangeCoverage in the case of
disjoint adjacent ranges
Change-Id: I29aaf5145e6cdcf5f0b8f4e009d57ee57bd4c944
Signed-off-by: Pau Cabre <pau.cabre@metempsy.com>
Reviewed-on: https://gem5-review.googlesource.com/4640
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
When a split load hits a memory region where IPRs are mapped, the
Writebackevent which is scheduled for that was carrying a data packet
that was not correctly initialized which caused an assertion to fire
when the Writeback event is processed.
Change-Id: I71a4e291f0086f7468d7e8124a0a8f098088972f
Signed-off-by: Matthias Hille <matthiashille8@gmail.com>
Reported-by: Matthias Hille <matthiashille8@gmail.com>
Reviewed-on: https://gem5-review.googlesource.com/4620
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
The VM's event queue is normally used for devices in multi-core KVM
mode. Add a helper method, BaseKvmCPU::deviceEventQueue(), to access
this queue. This makes the intention of code migrating to device event
queues clearer.
Change-Id: Ifb10f553a6d7445c8d562f658cf9d0b1f4c577ff
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/4287
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
|
|
The KVM CPU sometimes needs to access devices when drain() is
called. This typically happens on ARM when synchronizing devices that
use the system register interface. When called from drain(), the event
queue isn't locked since drain is called from the outside when the
simulator isn't servicing any events. In such cases, performing a
migration to the device's queue will unlock a mutex that isn't
locked. This typically results in a deadlock when resuming the system
since the lock will be in an undefined state.
Change-Id: Ibdcc2e034e916a929124f297e72aae306cf66728
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/4286
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
|
|
The introduction of a new vector register class broke rename in the O3
CPU due to an unhandled register class in
DefaultRename<Impl>::renameSrcRegs(). This patch fixes adds the
necessary handling to avoid a panic when the vector register file is
used.
Change-Id: Ie380ab35ec4a151db15402f25b25b58931ee0581
Reviewed-by: Giacomo Gabrielli <giacomo.gabrielli@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/4140
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
Checkpointing a system with out-of-order CPUs might get stuck if
one of the CPUs has been put to sleep. The quiesce instruction
cannot get drained hence checkpointing never finishes.
This commit resolves that by activating all suspended thread
contexts when draining the system.
Change-Id: I817ab1672b4ead777bd8e12a0445829481c46fdc
Reviewed-by: Sascha Bischoff <sascha.bischoff@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/3970
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
|
|
Change-Id: I897b6162a827216b7bad74d955c0e50e06a5a3ec
Signed-off-by: Sean Wilson <spwilson2@wisc.edu>
Reviewed-on: https://gem5-review.googlesource.com/3926
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
|
|
Change-Id: Ifafdcf4692d58a17f90e66ff8de8fa3e146c34bb
Signed-off-by: Sean Wilson <spwilson2@wisc.edu>
Reviewed-on: https://gem5-review.googlesource.com/3924
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
Change-Id: If765c6100d67556f157e4e61aa33c2b7eeb8d2f0
Signed-off-by: Sean Wilson <spwilson2@wisc.edu>
Reviewed-on: https://gem5-review.googlesource.com/3923
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
|
|
By setting the BaseCPU parameter wait_for_dbg_connection, the GDB
server blocks during initialisation waiting for the remote debugger to
connect before starting the simulated CPU.
Change-Id: I4d62c68ce9adf69344bccbb44f66e30b33715a1c
[ Update info message to include remote GDB port, rename param. ]
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/3963
Reviewed-by: Gabe Black <gabeblack@google.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
|
|
The BaseArmKvmCPU is responsible for forwarding the IRQ and FIQ
signals from gem5's simulated GIC to KVM. However, these signals
shouldn't be used when the in-kernel GIC emulator is used.
Instead of delivering the interrupts to the guest, we should just
ignore them since any such pending interrupts are likely to be an
artifact of CPU switching or incorrect draining.
Change-Id: I083b72639384272157f92f44a6606bdf0be7413c
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Sudhanshu Jha <sudhanshu.jha@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/3660
|
|
Reiley's update :) of the isa parser definitions. My addition of the
vector element operand concept for the ISA parser. Nathanael's modification
creating a hierarchy between vector registers and its constituencies to the
isa parser.
Some fixes/updates on top to consider instructions as vectors instead of
floating when they use the VectorRF. Some counters added to all the
models to keep faithful counts.
Change-Id: Id8f162a525240dfd7ba884c5a4d9fa69f4050101
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2706
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
This patch adds some more functionality to the cpu model and the arch to
interface with the vector register file.
This change consists mainly of augmenting ThreadContexts and ExecContexts
with calls to get/set full vectors, underlying microarchitectural elements
or lanes. Those are meant to interface with the vector register file. All
classes that implement this interface also get an appropriate implementation.
This requires implementing the vector register file for the different
models using the VecRegContainer class.
This change set also updates the Result abstraction to contemplate the
possibility of having a vector as result.
The changes also affect how the remote_gdb connection works.
There are some (nasty) side effects, such as the need to define dummy
numPhysVecRegs parameter values for architectures that do not implement
vector extensions.
Nathanael Premillieu's work with an increasing number of fixes and
improvements of mine.
Change-Id: Iee65f4e8b03abfe1e94e6940a51b68d0977fd5bb
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
[ Fix RISCV build issues and CC reg free list initialisation ]
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2705
|
|
The Result union used to collect the result of an instruction is now a
class of its own, with its constructor, and explicit casting methods for
cleanliness.
This is also a stepping stone to have vector registers, and instructions
that produce a vector register as output.
Change-Id: I6f40c11cb5e835d8b11f7804a4e967aff18025b9
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2703
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
With the hierarchical RegId there are a lot of functions that are
redundant now.
The idea behind the simplification is that instead of having the regId,
telling which kind of register read/write/rename/lookup/etc. and then
the function panic_if'ing if the regId is not of the appropriate type,
we provide an interface that decides what kind of register to read
depending on the register type of the given regId.
Change-Id: I7d52e9e21fc01205ae365d86921a4ceb67a57178
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
[ Fix RISCV build issues ]
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2702
|
|
Mimic the changes done on the architectural register indexes on the
physical register indexes. This is specific to the O3 model. The
structure, called PhysRegId, contains a register class, a register
index and a flat register index. The flat register index is kept
because it is useful in some cases where the type of register is not
important (dependency graph and scoreboard for example). Instead
of directly using the structure, most of the code is working with
a const PhysRegId* (typedef to PhysRegIdPtr). The actual PhysRegId
objects are stored in the regFile.
Change-Id: Ic879a3cc608aa2f34e2168280faac1846de77667
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2701
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
Replace the unified register mapping with a structure associating
a class and an index. It is now much easier to know which class of
register the index is referring to. Also, when adding a new class
there is no need to modify existing ones.
Change-Id: I55b3ac80763702aa2cd3ed2cbff0a75ef7620373
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
[ Fix RISCV build issues ]
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2700
|
|
Change-Id: Idd5992463bcf9154f823b82461070d1f1842cea3
Signed-off-by: Sean Wilson <spwilson2@wisc.edu>
Reviewed-on: https://gem5-review.googlesource.com/3746
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
|
|
If a (regular) store is followed closely enough by a locked load that
overlaps, the LSQ will forward the store's data to the locked load and
never tell the cache about the locked load. As a result, the cache will
not lock the address and all future store-conditional requests on that
address will fail. This patch fixes that by preventing forwarding if
the memory request is a locked load and adding another case to the LSQ
forwarding logic that delays the locked load request if a store in the
LSQ contains all or part of the data that is requested.
[Merge second and last if blocks because their bodies are the same.]
Change-Id: I895cc2b9570035267bdf6ae3fdc8a09049969841
Reviewed-on: https://gem5-review.googlesource.com/2400
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Tony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
|
|
Use the PyBind11 wrapping infrastructure instead of SWIG to generate
wrappers for functionality that needs to be exported to Python. This
has several benefits:
* PyBind11 can be redistributed with gem5, which means that we have
full control of the version used. This avoid a large number of
hard-to-debug SWIG issues we have seen in the past.
* PyBind11 doesn't rely on a custom C++ parser, instead it relies on
wrappers being explicitly declared in C++. The leads to slightly
more boiler-plate code in manually created wrappers, but doesn't
doesn't increase the overall code size. A big benefit is that this
avoids strange compilation errors when SWIG doesn't understand
modern language features.
* Unlike SWIG, there is no risk that the wrapper code incorporates
incorrect type casts (this has happened on numerous occasions in
the past) since these will result in compile-time errors.
As a part of this change, the mechanism to define exported methods has
been redesigned slightly. New methods can be exported either by
declaring them in the SimObject declaration and decorating them with
the cxxMethod decorator or by adding an instance of
PyBindMethod/PyBindProperty to the cxx_exports class variable. The
decorator has the added benefit of making it possible to add a
docstring and naming the method's parameters.
The new wrappers have the following known issues:
* Global events can't be memory managed correctly. This was the
case in SWIG as well.
Change-Id: I88c5a95b6cf6c32fa9e1ad31dfc08b2e8199a763
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Andreas Hansson <andreas.hansson@arm.com>
Reviewed-by: Andrew Bardsley <andrew.bardsley@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2231
Reviewed-by: Tony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Pierre-Yves Péneau <pierre-yves.peneau@lirmm.fr>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
|
|
This also allows checkpointing of a Kvm GIC via the Pl390 model.
Change-Id: Ic85d81cfefad630617491b732398f5e6a5f34c0b
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2444
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Weiping Liao <weipingliao@google.com>
|
|
This change adds the ability to print a message at intervals
of committed instruction count to indicate progress in the
trace replay.
Change-Id: I8363502354c42bfc52936d2627986598b63a5797
Reviewed-by: Rekai Gonzalez Alberquilla <rekai.gonzalezalberquilla@arm.com>
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2321
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
|
|
simulations
Modifies the clone system call and adds execve system call. Requires allowing
processes to steal thread contexts from other processes in the same system
object and the ability to detach pieces of process state (such as MemState)
to allow dynamic sharing.
|
|
This changeset adds functionality that allows system calls to retry without
affecting thread context state such as the program counter or register values
for the associated thread context (when system calls return with a retry
fault).
This functionality is needed to solve problems with blocking system calls
in multi-process or multi-threaded simulations where information is passed
between processes/threads. Blocking system calls can cause deadlock because
the simulator itself is single threaded. There is only a single thread
servicing the event queue which can cause deadlock if the thread hits a
blocking system call instruction.
To illustrate the problem, consider two processes using the producer/consumer
sharing model. The processes can use file descriptors and the read and write
calls to pass information to one another. If the consumer calls the blocking
read system call before the producer has produced anything, the call will
block the event queue (while executing the system call instruction) and
deadlock the simulation.
The solution implemented in this changeset is to recognize that the system
calls will block and then generate a special retry fault. The fault will
be sent back up through the function call chain until it is exposed to the
cpu model's pipeline where the fault becomes visible. The fault will trigger
the cpu model to replay the instruction at a future tick where the call has
a chance to succeed without actually going into a blocking state.
In subsequent patches, we recognize that a syscall will block by calling a
non-blocking poll (from inside the system call implementation) and checking
for events. When events show up during the poll, it signifies that the call
would not have blocked and the syscall is allowed to proceed (calling an
underlying host system call if necessary). If no events are returned from the
poll, we generate the fault and try the instruction for the thread context
at a distant tick. Note that retrying every tick is not efficient.
As an aside, the simulator has some multi-threading support for the event
queue, but it is not used by default and needs work. Even if the event queue
was completely multi-threaded, meaning that there is a hardware thread on
the host servicing a single simulator thread contexts with a 1:1 mapping
between them, it's still possible to run into deadlock due to the event queue
barriers on quantum boundaries. The solution of replaying at a later tick
is the simplest solution and solves the problem generally.
|
|
A KVM VM is typically a child of the System object already, but for
solving future issues with configuration graph resolution, the most
logical way to keep track of this object is for it to be an actual
parameter of the System object.
Change-Id: I965ded22203ff8667db9ca02de0042ff1c772220
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
Used cppclean to help identify useless includes and removed them. This
involved erroneously included headers, but also cases where forward
declarations could have been used rather than a full include.
|
|
|
|
The headers declared in export_method_cxx_predecls are redundant since a
SimObject's main header is automatically included.
Change-Id: Ied9e84630b36960e54efe91d16f8c66fba7e0da0
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-by: Joe Gross <joseph.gross@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
|
|
This patch implements an L-TAGE predictor, based on André Seznec's code
available from CBP-2
(http://hpca23.cse.tamu.edu/taco/camino/cbp2/cbp-src/realistic-seznec.h).
Signed-off-by Jason Lowe-Power <jason@lowepower.com>
|
|
The Minor and o3 cpu models share the branch prediction
code. Minor relies on the BPredUnit::squash() function
to update the branch predictor tables on a branch mispre-
diction. This is fine because Minor executes in-order, so
the update is on the correct path. However, this causes the
branch predictor to be updated on out-of-order branch
mispredictions when using the o3 model, which should not
be the case.
This patch guards against speculative update of the branch
prediction tables. On a branch misprediction, BPredUnit::squash()
calls BpredUnit::update(..., squashed = true). The underlying
branch predictor tests against the value of squashed. If it is
true, it restores any speculatively updated internal state
it might have (e.g., global/local branch history), then returns.
If false, it updates its prediction tables. Previously, exist-
ing predictors did not test against the "squashed" parameter.
To accomodate for this change, the Minor model must now call
BPredUnit::squash() then BPredUnit::update(..., squashed = false)
on branch mispredictions. Before, calling BpredUnit::squash()
performed the prediction tables update.
The effect is a slight MPKI improvement when using the o3
model. A further patch should perform the same modifications
for the indirect target predictor and BTB (less critical).
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
|
|
The tournament predictor is presented as doing speculative
update of the global history and non-speculative update
of the local history used to generate the branch prediction.
However, the code does speculative update of both histories.
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
|
|
The target of taken conditional direct branches does not
need to be resolved in IEW: the target can be computed at
decode, usually using the decoded instruction word and the PC.
The higher-than-necessary penalty is taken only on conditional
branches that are predicted taken but miss in the BTB. Thus,
this is mostly inconsequential on IPC if the BTB is big/associative
enough (fewer capacity/conflict misses). Nonetheless, what gem5
simulates is not representative of how conditional branch targets
can be handled.
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
|
|
cachePorts currently constrains the number of store packets written to the
D-Cache each cycle), but loads currently affect this variable. This leads
to unexpected congestion (e.g., setting cachePorts to a realistic 1 will
in fact allow a store to WB only if no loads have accessed the D-Cache
this cycle). In the absence of arbitration, this patch decouples how many
loads can be done per cycle from how many stores can be done per cycle.
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
|
|
Previously all traffic generators would use the same value for write
requests. With this change traffic generators use their master id as
the payload of write requests making them more useful for the
memchecker.
Change-Id: Id1a6b8f02853789b108ef6003f4c32ab929bb123
Reviewed-by: Andreas Hansson <andreas.hansson@arm.com>
Reviewed-by: Stephan Diestelhorst <stephan.diestelhorst@arm.com>
|
|
First of five patches adding RISC-V to GEM5. This patch introduces the
base 64-bit ISA (RV64I) in src/arch/riscv for use with syscall emulation.
The multiply, floating point, and atomic memory instructions will be added
in additional patches, as well as support for more detailed CPU models.
The loader is also modified to be able to parse RISC-V ELF files, and a
"Hello world\!" example for RISC-V is added to test-progs.
Patch 2 will implement the multiply extension, RV64M; patch 3 will implement
the floating point (single- and double-precision) extensions, RV64FD;
patch 4 will implement the atomic memory instructions, RV64A, and patch 5
will add support for timing, minor, and detailed CPU models that is missing
from the first four patches (such as handling locked memory).
[Removed several unused parameters and imports from RiscvInterrupts.py,
RiscvISA.py, and RiscvSystem.py.]
[Fixed copyright information in RISC-V files copied from elsewhere that had
ARM licenses attached.]
[Reorganized instruction definitions in decoder.isa so that they are sorted
by opcode in preparation for the addition of ISA extensions M, A, F, D.]
[Fixed formatting of several files, removed some variables and
instructions that were missed when moving them to other patches, fixed
RISC-V Foundation copyright attribution, and fixed history of files
copied from other architectures using hg copy.]
[Fixed indentation of switch cases in isa.cc.]
[Reorganized syscall descriptions in linux/process.cc to remove large
number of repeated unimplemented system calls and added implmementations
to functions that have received them since it process.cc was first
created.]
[Fixed spacing for some copyright attributions.]
[Replaced the rest of the file copies using hg copy.]
[Fixed style check errors and corrected unaligned memory accesses.]
[Fix some minor formatting mistakes.]
Signed-off by: Alec Roelke
Signed-off by: Jason Lowe-Power <jason@lowepower.com>
|
|
This function was used by the now-defunct InOrderCPU model. Since this
model is no longer in gem5, this function was not called from anywhere in
the code.
|
|
Modify the opClass assigned to AArch64 FP instructions from SimdFloat* to
Float*. Also create the FloatMemRead and FloatMemWrite opClasses, which
distinguishes writes to the INT and FP register banks.
Change the latency of (Simd)FloatMultAcc to 5, based on the Cortex-A72,
where the "latency" of FMADD is 3 if the next instruction is a FMADD and
has only the augend to destination dependency, otherwise it's 7 cycles.
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
|
|
networktest is essentially a collection of synthetic traffic patterns
for the network. The protocol name and the tester having the same name
led to multiple python configuration files with the same name, adding
confusion. This patch renames networktest to garnet_synthetic_traffic,
and also adds more synthetic traffic patterns.
|
|
The drain did not wait until stages were ready again. Therefore, as a
result of messages in the TimeBuffer being drain, the state after the
drain was not consistent and asserts fired in some places when the
draining happened after a stage got blocked, but before the notification
arrived to the previous stages.
Change-Id: Ib50b3b40b7f745b62c1eba2931dec76860824c71
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
|
|
This change adds a Trace CPU param to exit simulation early,
i.e. when the first (any one) trace execution is complete. With
this change the user gets a choice to configure exit as either
when the last CPU finishes (default) or first CPU finishes
replay. Configuring an early exit enables simulating and
measuring stats strictly when memory-system resources are being
stressed by all Trace CPUs.
Change-Id: I3998045fdcc5cd343e1ca92d18dd7f7ecdba8f1d
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
|
|
This change subtracts the time offset present in the trace from
all the event times when nodes and request are sent so that the
replay starts immediately when the simulation starts. This makes
the stats accurate when the time offset in traces is large, for
example when traces are generated in the middle of a workload
execution. It also solves the problem of unnecessary DRAM
refresh events that would keep occuring during the large time
offset before even a single request is replayed into the system.
Change-Id: Ie0898842615def867ffd5c219948386d952af7f7
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
|
