isdir isn't a nice way to check if an FS.Base is a File or a Dir as was
initially assumed, it literally checks if a path can be stat-ed and is
reported as a directory by stat. This means that if a directory is
going to be created as part of the build, the result of that test will
change depending on whether that part of the build has happened
successfully before.
A better check which behaves as originally intended is to check whether
the Node is an instance of the SCons.Node.FS.Dir class.
Change-Id: Id041917d50b768a8205769c0a05320f92b09993c
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/17128
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
2c242d6 introduced implicit-fallthrough errors when building against
ARM.
Added "default: return new Unknown(machInst);" to offending switch
statements; please verify this is the corret behavior
Signed-off-by: Ryan Gambord
Change-Id: I5f5e3661ec562d4a3b2699e07d1195e6877ff959
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/17071
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Added additional information to the PrefetchInfo data structure
- Whether the event is triggered by a cache miss
- Whether the event is a write or a read
- Size of the data accessed
- Data accessed by the request
Change-Id: I070f3ffe837ea960a357388e7f2b8a61d7b2196c
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/16583
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Some accesses only need to search for a tag in the tag array, with
no need to touch the data array. This is the case for CleanEvicts,
evicts that don't find a corresponding block entry (since a write
cannot be done in parallel with tag lookup), and maintenance
operations.
Change-Id: I7365a915500b5d7ab636d49a9acc627072a7f58e
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/14878
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
When dealing with writebacks, as soon as the packet metadata arrives
there will be a tag lookup, done sequentially because a write can't
be done in parallel. While the tag lookup is being done, the payload
will arrive. When both the payload are present and the tag is correct
block entry is determined the fill happens.
Change-Id: If1a0085d742458b675bfc012b6d908d9d9a25e32
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/14877
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Previously the bus delay was being ignored for the access latency
calculation, and then applied on top of the access latency. This
patch fixes the order, as first the packet must arrive before the
access starts.
Change-Id: I6d55299a911d54625c147814dd423bfc63ef1b65
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/14876
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
The header and payload delay have already been accounted and
zeroed previous to calling this function. The probe is not
allowed to modify the packet, therefore no extra delays are
added, and it is safe to remove the todo note.
Change-Id: I8ddf7e189fbe609cdec34364f3c013427930daf7
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/14875
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
In the b_transport method of the SCMasterPort class, there is a check
which determines whether the packet being sent to gem5 should be
deleted once the call to sendAtomic returns. This was deleting the
packet if extension was *not* nullptr.
This check should delete the packet if the extension *is* nullptr. The
reasoning is that the extension will equal nullptr if there was no
gem5 packet in an extension and a new one needed to be allocated. If
there was an extension, ie if extension is not nullptr, then that's
where the packet came from which therefore doesn't belong to us. In
that case, we need to leave it alone and let its owner clean it up.
With the check reversed, this method will either leak allocated packets
it should delete, or delete packets it shouldn't that someone else will
likely try to use later.
Change-Id: I61578d910be6e5085b9fc0ddaa82468b1ac68578
Reviewed-on: https://gem5-review.googlesource.com/c/16949
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
These bring in some pieces that those headers use but were only
coincidentally included by something else when they were used.
Change-Id: I5f119260d8f25d914d8545a60834f23f65f82d0c
Reviewed-on: https://gem5-review.googlesource.com/c/16948
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
The problem occurs when a proxy is being resolved to another proxy
that hasn't been resolved yet. The problematic case that was
triggering this issues in the VGIC. It was caused by parameters
looking a bit like this:
gic = Param.GicV2(Parent.any)
some_param = Param.Int(Self.gic.some_param)
When 'some_param' was resolved, it found the 'gic' parameter in
Self. However, that parameter hadn't been resolved yet, so the
existing code was setting the proxy evaluation context to the
unresolved Parent.any proxy without first unproxying it.
It seems like this bug depends on the graph traversal order and I have
so far only seen it when compiling gem5 with Python 3.
Change-Id: Iea12cc138765e70bfd6bb776b1efa012364db066
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/16004
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
TLB:getMasterPort is used to obtain the PageWalkMasterPort if present and
hides the BaseTLB::getMasterPort().
The TLB::getMasterPort() is renamed according to the expected behavior.
Change-Id: If4f61189094a706d59805cd10f4f814e5830eda8
Reviewed-on: https://gem5-review.googlesource.com/c/16648
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
This will ensure that the value of USE_SYSTEMC is consistent throughout
the build. It also has the side effect that USE_SYSTEMC can be forced
to a particular value if you're confident you know what you're doing
and want to override these checks.
Change-Id: I0f2d1153245ff17ce4a828c6b7496cb9ded6bd5b
Reviewed-on: https://gem5-review.googlesource.com/c/16810
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
Garnet utilizes round robin policy to select a VC for
transmission ar Network Interface and Routers. The current logic
for round robin is only fair if all the virtual networks are active
at a given router. If the router or network interface is not
receiving traffic in from any vnet then the priority is always taken
up by the next vnet in numerically (or loops back to 0).
This fix changes the way we perform round robin arbitration. When
a VC is selected in a cycle, the round robin pointer is set to the VC
next to it and is iterated from there on. If any VC does not have a
flit in a given cycle, it will lose its turn until the next round.
At maximum traffic this will model round robin correctly even if
a certain VNET is not active at that unit.
Change-Id: I9bf805221054f9f25bee14b57ff521f4ce4ca980
Reviewed-on: https://gem5-review.googlesource.com/c/16688
Reviewed-by: Jieming Yin <Jieming.Yin@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
During the O3PipeView execution, a potential invalid iterator is used to
Update the instruction storeTick field.
If the store_idx iterator is the first() of the StoreQueue, the
corresponding instruction is removed from the queue, leaving the iterator
invalid and not usable in the TRACING_ON block.
This patch uses the store_inst variable to access (and update) the
instruction tick, instead of the (potential) invalid one.
Change-Id: I671052ef282b9048e5239da8629b89e8afa86bf0
Reviewed-on: https://gem5-review.googlesource.com/c/16322
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
A cache that forwards a request to the memory below does not fill and
forwards the response with the data to cache above. This change
ensures that the flags of the original response are also preserved.
Change-Id: I244b20b073c31b976358816c5b14bba413b8271f
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/16182
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Python 2.7 used to return lists for operations such as map and range,
this has changed in Python 3. To make the configs Python 3 compliant,
add explicit conversions from iterators to lists where needed, replace
xrange with range, and fix changes to exec syntax.
This change doesn't fix import paths since that might require us to
restructure the configs slightly.
Change-Id: Idcea8482b286779fc98b4e144ca8f54069c08024
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/16002
Reviewed-by: Gabe Black <gabeblack@google.com>
We currently use the Python version used by scons to marshal Python
code. This doesn't work when building gem5 with Python 3 support since
scons typically runs in Python 2.7. Add a custom marshal helper that
links with the same library as gem5 to generate byte code that is
guaranteed to work in gem5's Python interpreter.
Change-Id: I665b0f2078726d4c055d74a3e668a580fc613b59
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/16422
Reviewed-by: Gabe Black <gabeblack@google.com>
Start using sc_main and sc_main_result from the systemc module, and
stop using the versions of those functions which are attached to the
SystemC_Kernel SimObject.
Change-Id: I802898038c80ed36e6a9176211cffb7e0fde2d7e
Reviewed-on: https://gem5-review.googlesource.com/c/16564
Maintainer: Gabe Black <gabeblack@google.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
These will be how systemc and tlm APIs which are not attached to
SimObjects will be exposed. This avoids having to artificially attach
them to wrapping SimObjects for instance, which is a bit awkward
and non-obvious.
The python code which attaches the systemc and tlm modules to the
m5 modules lives in src/python/m5/__init__.py, but the modules
themselves live in src/systemc/python to keep all the systemc code
grouped together. It might be a little confusing to have a small part
of the glue that adds those modules in a separate place (__init__.py),
but that is, as far as I can tell, unavoidable, and it's better in my
opinion to keep the systemc code grouped together than to put it
alongside the other python code and __init__.py.
Change-Id: Iecb218daec5e15772152b5ad22b51f43b86c3d4b
Reviewed-on: https://gem5-review.googlesource.com/c/16563
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
There is a circular dependency between params and proxy at import
time. This causes issues for Python 3. Add the imports to the specific
methods with the dependencies to make the import happen when the
method is executed instead.
Change-Id: I770112fd3c07c395459e204976942bda3dc7236f
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/15993
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Juha Jäykkä <juha.jaykka@arm.com>
Add a sticky variable (PYTHON_CONFIG) to select which python-config
version to use. This can, for example, be used to build with Python 3
or with Python 2.7 in a custom location.
Change-Id: I1f4c00d66f85a9c99f50fe4d746b69dd82b60b4b
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/16003
Reviewed-by: Gabe Black <gabeblack@google.com>
Python 3 has restructured some packages. Specifically, __builtin__ has
been renamed to builtins and urlparse has been included in urllib.
Change-Id: I81f8f3942471db1043006a36abbad6e5a49e0a43
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/15994
Reviewed-by: Juha Jäykkä <juha.jaykka@arm.com>
