mem-ruby: Make bitSelect use bits<Addr>

There is no need to replicate bits<Addr>' functionality.

As a side effect, ADDRESS_WIDTH is no longer used and was removed

Change-Id: Ia5679f3976c81f779665d82cb758850092f2a293
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21085
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-ruby: Fix maskLowOrderBits

The function was wrong when number = 63. Also, use the more reliable
src/base/bitfield.hh's mbits when posible.

maskLowOrderBits has only been kept because SLICC does not accept
a templated function.

Change-Id: I8dd680da02ceb9e614e2f9cbf8f1ac52cead8d45
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21084
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-ruby: Remove shiftLowOrderBits

There is no need to encapsulate a shift operation.

Change-Id: Ie711d8d4975d1d9dde656cc2284a048410cfdadb
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21083
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-ruby: Remove maskHighOrderBits

Function was not being used. If needed, src/base/bitfield.hh's
mbits can be used instead:

maskHighOrderBits(addr, pos) == mbits<Addr>(addr, 64-pos, 0)

Change-Id: I3abd041f8d256ec157ba7502182d8588721c2a05
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21082
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-ruby: Remove bitRemove

bitRemove is not being used anywhere. If needed, can be used
as src/base/bitfield.hh's bits:

bitRemove(addr, small, big) ==
    ((bits<Addr>(addr, 63, big + 1) << small) |
    bits<Addr>(addr, small, 0))

Change-Id: I45fd3bc0271ccb659d6a94e3dd00ca095dfd6aa7
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21081
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Make use of DRAMCtrl::AddrMap in the traffic generators

Use the enum defined in the memory controller rather than custom
strings and int that are later converted to the DRAMCtrl::AddrMap
enum.

Change-Id: Ie5b19f915f9990fd2b7505d4d1b17b6fc2100f9e
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21080
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

misc: Added line wrapping functionality for Sim-Object desc

Descriptions were previously printed on one line, unless explicitly broken
when writing the description of the Sim-Object. In this commit, line
wrapping is enabled when printing these descriptions. Developers, when
writing the Sim-Object descriptions, may now over multiple lines with
triple double-quotes and still have the description output correctly when
viewing the Sim-Objects within the CLI.

E.g.: X86System previously had the following load_addr_mask component which
was output as:

load_addr_mask
            default: 18446744073709551615
               desc: Address to mask loading binaries with, if 0, system \
auto-calculates the mask to be the most restrictive, otherwise it obeys a \
custom mask.

This was defined by the developer via:

load_addr_mask = Param.UInt64(0xffffffffffffffff,
            "Address to mask loading binaries with, if 0, system "
            "auto-calculates the mask to be the most restrictive, "
            "otherwise it obeys a custom mask.")

This is now displayed as:

load_addr_mask
            default: 18446744073709551615
               desc: Address to mask loading binaries with, if 0,
                     system auto-calculates the mask to be the most
                     restrictive, otherwise it obeys a custom mask.

JiraID: Gem5-57
Built: Linux (GCC)
Tested: Ran quick tests for X86, ARM, and RISC-V
Change-Id: If012304e50af60f6ba10c1fa2b44da8bac1c09cf
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21179
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

mem: Use new-style stats in the XBar models

Migrate to new-world stats with an explicit hierarchy in all of the
XBar models.

Change-Id: I18b6746a1303ca415638e6d382fb4757607f1123
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21141
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>

mem-cache: Switch to new-style stats

This change puts cache and tag stats into a Stats::Group struct. This
makes it easier to identify stat updates (they are prefixed with
stat.) and adds hierarchy information for output formats that need it.

Change-Id: I2b8e9138f1cb977abb445ec864d80a79b588481d
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21140
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem: Convert DRAM controller to new-style stats

Note that this changes the stat format used by the DRAM
controller. Previously, it would have a structure looking a bit like
this:

  - system
    - dram: Main DRAM controller
    - dram_0: Rank 0
    - dram_1: Rank 1

This structure can't be replicated with new-world stats since stats
are confined to the SimObject name space. This means that the new
structure looks like this:

  - system
    - dram: Main DRAM controller
      - rank0: Rank 0
      - rank1: Rank 1

Change-Id: I7435cfaf137c94b0c18de619d816362dd0da8125
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21142
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Wendy Elsasser <wendy.elsasser@arm.com>

ruby: 2x protocols has typo/syntax error that fails building

MOESI_hammer and MOESI_CMP_token contain incorrect lines.

Change-Id: I1f9ac429d0f4dcb0241f21c8c9b831bee7aa37a4
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21259
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

fastmodel: Add glue code which adapts fastmodels to run in gem5.

Fast Models are models written by ARM which emulate different
components of a computer system. They can be combined into small
subsystems and then exported as systemc modules.

To enable this code, you'll need to set USE_ARM_FASTMODEL variable to
true. This CL does not include the fast models themselves, or a license
to use them or the associated tools. To build these fast models, you'll
need to set some scons variables. These variables should be set as
described in the fast model distribution.

* PVLIB_HOME
* MAXCORE_HOME
* ARMLMD_LICENSE_FILE

Some minor patches to source filesdistributed with the fast model code
may be necessary since their use of systemc is not necessarily 100%
standards compliant.

Change-Id: Id53814b95d8aa320da4d4f2159be0736fc12eb73
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20799
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

sim: Convert power modelling framework to new-style stats

Change-Id: I1dd3ea3d37bb4464637222aa5bc5d88cc7d9b66a
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21143
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>

stats: Add a preDumpStats() callback to Stats::Group

Some objects need to know that we are about to dump stats to perform
prepare statistics. This is currently done by registering a callback
with the stat system. Expose this callback as a virtual method
in Stats::Group to make this pattern more convenient.

Change-Id: I5aa475b7d04c288e45f5f413ab7a1907b971dae5
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21139
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>

stats: Don't output index in vectors of length 1

New-world stats didn't implement the old gem5 quirk where the index of
an element in a vector of size 1 is ignored. This affects the object
path when printed in a stat file. Implement the same quirk as the
original code to make sure names of old- and new-world stats are
compatible.

Change-Id: I12696d40baaed2d0bb1ef061e87c3b1d0f4bb4b1
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21163
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>

stats: Correctly print new-style dist stat names

The text stat printer didn't output fully qualified names for
new-style dist stats. Fix that.

Change-Id: I61bc7403fbd760bbbb0641e8e593781e6d03a4ed
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21162
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>

mem-ruby: prevent cacheProbe being called multiple times

The cacheProbe() function will return the victim entry, and it gets
called for multiple times in trigger function in a single miss. This
will cause a problem when we try to add a new replacement policy to
the Ruby system. Certain policy, like RRIP, will modify the block
information every time the getVictim() function gets called. To
prevent future problems, we need to store the victim entry, so that
we only call it once in one miss.

Change-Id: Ic5ca05f789d9bbfb963b8e993ef707020f243702
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21099
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Pouya Fotouhi <pfotouhi@ucdavis.edu>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

ext: Updated Pybind11 to version 2.4.1.

This updates Pybind11 from version 2.2.1 to version 2.4.1. This fixes
warning/error received when "<experiment/optional>" is used when
compiling using c++14 with clang. It should be noted that
"ext/pybind11/include/pybind11/std.h" has been changed to include a fix
added by commit ba42457254cc362eddc099f22b60d469cc6369e0. This is
necessary to avoid build errors.

Built: Linux (gcc, c++11) and MacOS (clang, c++14).
Tested: Ran quick tests for X86, ARM, and RISC-V.
Deprecates: https://gem5-review.googlesource.com/c/public/gem5/+/21019
Change-Id: Ie9783511cb6be50136076a55330e645f4f36d075
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21119
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Fix checker cpu instantiation

This change uses the params as instantiated from the default
constructor to create the checker cpu. If any of these parameters are
invalid for the checker cpu, the simulation will exit with a warning.

Change-Id: I0e58ed096c9ea5f413f2e9b64d8d184d9b0fc84e
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21079
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu, mem: Changing AtomicOpFunctor* for unique_ptr<AtomicOpFunctor>

This change is based on modify the way we move the AtomicOpFunctor*
through gem5 in order to mantain proper ownership of the object and
ensuring its destruction when it is no longer used.

Doing that we fix at the same time a memory leak in Request.hh
where we were assigning a new AtomicOpFunctor* without destroying the
previous one.

This change creates a new type AtomicOpFunctor_ptr as a
std::unique_ptr<AtomicOpFunctor> and move its ownership as needed. Except
for its only usage when AtomicOpFunc() is called.

Change-Id: Ic516f9d8217cb1ae1f0a19500e5da0336da9fd4f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20919
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem: Delete the now unused Message*Port classes.

This port type is no longer used.

Change-Id: If4abbb774819644bea58fd82e00dfdec8f79b5a6
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20822
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Gabe Black <gabeblack@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

x86: Templatize the IntMasterPort.

This makes the IntMasterPort usable with any class, making it possible
to avoid inheriting from IntDevice.

It also makes IntMasterPort inherit directly from QueuedMasterPort,
skipping over MessageMasterPort.

Change-Id: I9d218556c838ea567ced5f6fa4d57a3ec9d28d31
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20821
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>

x86: Templatize IntSlavePort.

This makes the device IntSlavePort calls back into based on a template
parameter so that IntDevice doesn't have to be in the inheritance
hierarchy to use it.

It also makes IntSlavePort inherit from SimpleTimingPort directly,
skipping over MessageSlavePort.

Change-Id: Ic3213edc9c3ed5e506ee1e9f5e082cd47d7c7998
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20820
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>

x86: Turn the local APIC into a PioDevice instead of a BasicPioDevice.

This will let us accept several address ranges through our pio port
instead of just one, and that will in turn let us accept interrupt
requests and pio requests through the same port.

Change-Id: I70b78c8cd0edca7fe58b3d4cd241e41d9e0f2c20
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20819
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

gpu-compute: Fix overriden errors

When building Gem5 with GPU protocols, overriden errors were thrown
from files in gpu-compute. After adding override to the files, the
errors were resolved and Gem5 builds successfully.

Change-Id: Iab3a0768caf82c226e8bbee5690a834bf92d1e03
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20939
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev, x86: Delete the now unused X86 specific interrupt pins/lines.

Change-Id: I3915f0ad673119b551dcc4c5cedec180a9b88735
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20702
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev, x86: Convert x86 devices to the generic int pins.

Change-Id: I4551ad00cf205c31555c90b53e87bc206a8d8729
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20701
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-x86: ignore non-temporal hint for movntps/movntpd SSE insts

Making the implementation of movntps/movntpd consistent with other
non-temporal instructions. We are ignoring the hint here, and
implementing those instructions as cacheable instructions.

This change adds a warning to let user know about this workaround.
Also, this change add the address check for second part of move.

Change-Id: I811652b24cf39ca2f5c5d4c9e9e417f69190b55c
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20408
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev: Terminal output's dump name conflicts

The recently Terminal dump options name introduced
in patch 1c72e90 conflict with MacOS stdio.h header.

From stdio.h:

#define stdin __stdinp
#define stdout __stdoutp
#define stderr __stderrp

To remove this conflict, the TerminalDump options are renamed.

Change-Id: I1b60fa1031328fc32d6a58bf7889b6e479d95219
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20959
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-x86: Change warn to warn_once for NT instructions

Change-Id: I50353716f2a913b9b106b140644d95991879f662
Signed-off-by: Hoa Nguyen <hoanguyen@ucdavis.edu>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21039
Reviewed-by: Gabe Black <gabeblack@google.com>
Reviewed-by: Pouya Fotouhi <pfotouhi@ucdavis.edu>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

python: Don't try to bind a stat group to the NULL simobject.

That is a SimObject like object which is used when a SimObject
parameter is purposefully left empty, vs. being accidentally left
empty through a typo or accidental ommission.

It doesn't have a getCCObject method, and attempting to use it anyway
causes gem5 to crash.

Change-Id: Ie86321fbdbcc41cf88b7009184423acd7b64484b
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21059
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

misc: Update MAINTAINERS

Change-Id: I13ca12d4dd170ce3db03d851829df9bc62d1a74c
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20999
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Anthony Gutierrez <anthony.gutierrez@amd.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Conditionally enable HDLcd when doing DTB autogen

This is a preparation change for a real DTB autogen implementation

Change-Id: Ia0c1c5e65ea96036e55455eb4222cec12944d33a
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20331
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Add HDLcd DTB autogeneration

A Display has been defined. Its sole purpose is to generate the device
tree node to be referenced by the HDLcd device.  The encoder parameters
are based on the existing node defined in:

system/arm/dt/armv8.dts

Change-Id: I6cdeb0437dce207dbd0f2c65c16b224245eb74e1
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20330
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: PSTATE.PAN changes should inval cached regs in TLB

Change-Id: Id94e355fec345d2e952539a7dce7fbd21ed220c6
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20983
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

system-arm: Add ITS node in platforms/vexpress_gem5_v2_base.dtsi

This is aligning sources with DTB autogeneration

Change-Id: Icf369eb85719c91da770398b45645d8b03d8abf3
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20982
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: Fix Data Abort ISS when caused by Atomic operation

Data Aborts caused by an atomic instruction have a special rule for
their syndrome:
From a ISS point of view they count as read if a read to that address
would generate a fault; they count as writes otherwise (ISS.WnR bit)
This patch is implementing this in the TLB. For permission faults we
need to explicitly check if a read would trigger a fault
(e.g. checking for the AP bits) since permissions can allow read-only
accesses.
For other MMU exceptions (like translation faults) we are confident the
nature of the access doesn't affect the genration of a fault.
This means that if the access is atomic, we treat it as a read from an
ISS.WnR point of view.

Change-Id: Ia524aa6ae07f81513cdc26c516b5fd9b01a931c3
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20981
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: ISV bit in DataAbort should check for translation stage

According to the ESR spec, the ISV bit is set to 1 only for stage 2
aborts.

Change-Id: Id524ef36e82184f741e968ddba04ca8ccdd4ad58
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20980
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: PSTATE.PAN affecting EL2 only when HCR_EL2.E2H=1

Change-Id: I6df0cdcbadca17f30d3de3bed887f75c739b00f0
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20979
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch, x86: Rework the debug faults and microops.

This makes the non-fatal microops advance the PC, and adds missing
functions. The *_once Faults now also can be run once per *something*.
They would previously be run once per Fault invoke function which is
common to all M5WarnOnceFaults. The warn_once microop will now warn
once per message.

Change-Id: I05974b93f3b2700077a411b243679c2ff0e8c2cb
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20739
Reviewed-by: Gabe Black <gabeblack@google.com>
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

systemc: Make TLM port wrappers for tlm_base_(target|initiator)_socket.

These are useful if using TLM sockets without using the standard TLM
protocol. For instance, when used with ARM's fast models, this can wrap
sockets which carry the opaque GICv3Comms protocol.

Change-Id: I329a919068f958abbde2cb83683d3a3ae2e05a20
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20860
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

systemc: Expose the utility functions for payload->packet conversion.

These can be used anywhere someone needs to turn a packet into a
payload or vice-versa.

Change-Id: I26a1f134e6fbec81b230a2853cf79264e80bf2dc
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20859
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Allow IOMMU binding to HDLcd

Change-Id: I894080e7bd76e7efedef141c937e1561c0c0527c
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Ciro Santilli <ciro.santilli@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20841
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Store the IOMMU reference from within the SMMU::connect

Change-Id: I35718a71dc040ee4acad9eee2a07076ebb571304
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Ciro Santilli <ciro.santilli@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20840
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev: Enable DTB IOMMU binding with a DMA object

This happens by storing a iommu reference in the dma device, and by
calling the addIommuProperty method once doing dtb autogeneration for
the device

Change-Id: Ibd585addac686a9eeaba6959f459d81901087549
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Ciro Santilli <ciro.santilli@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20839
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

sparc: Fix a warning/error in tlb.cc.

gcc has started to not like memseting an ojbect to zero in some cases.
Cast the TlbEntry pointer to a void * before memsetting it to placate
gcc.

Change-Id: Iccb3c326fdb82f1f111329ff1a80bb6719cace47
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20830
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>

dev: Generalize the x86 int source/sink pins.

Sink pins are now templated based on the underlying device types, and
the pins themselves are based on the new, generalized Port mechanism.

Change-Id: I09c678c56f6eb6b4a167c12f221ae0241fe99c2d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20700
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Fix BDI size calculation

The bitmask field indicates to which base a delta refers, and in
the original paper it is fixed and proportional to the highest
number of bases allowed in the compressed data.

Change-Id: I271bf2e19e0765de52b933eaf6d4fcc2ce25d185
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/19748
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

base: Add a perfect bloom filter

Add a bloom filter that keeps track of all observed entries, and
thus has no false negatives nor false positives.

Change-Id: Iba784e617a99c77554c688470d9b9e12c260f23b
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/18879
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>

stats: Add M5_VAR_USED to variable

Fixes compile error for gem5.fast due to unused variable.

Change-Id: Ib1664b100e95a741174610f8cdbdebaff9635f66
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20759
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

stats: Fix incorrect name conflict panic with grouped stats

Info::setName() performs a sanity check to ensure that the same stat
name isn't used twice. This doesn't work for new-style stats with a
parent group since the name is only unique within the group. Disable
the check for new-style stats since these usually use names generated
from member variable names.

Change-Id: I590abe6040407c6a4fe582c0782a418165ff5588
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20760
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

stats: Ignore non-Group objects in stat hierarchy

Some objects, such as SystemC modules, are not a subclass of
Stat::Group. Calling the addStatGroup function on them causes errors.
This changes ignores those objects that are not Stat::Group in the stat
hierarchy.

Signed-off-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
Change-Id: I9b62419417b7af7331461fbfaf15e45a4ee2b35f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20680
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

base: Make Bloom Filter counting by default

Since a boolean bool filter is a saturating bloom filter with a
single bit per entry, generalize them by using SatCounter instead
of int for the filter entries.

Change-Id: I7f54e28d54de5671e0770b02ed9161735e6bd339
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/18877
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>

base: Make Bulk inherit from MultiBitSel Bloom Filter

Cleanup bulk's code and make it inherit from MultiBitSel.

Change-Id: I83154feb30bb3dac3d02743bcafbdcb57489c2fd
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/18876
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>

mem-ruby: Move Bloom Filters to base

All Bloom Filters are completely independent of Ruby, and
therefore can be used everywhere.

As a side effect, Ruby was not using the filters, so
their dependency was removed.

Change-Id: Ic5f430610c33c0791fb81c79101ebe737189497e
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/18875
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

mem: Mark MemObject as deprecated.

It's constructor will now warn that it's deprecated and suggest using
ClockedObject directly. This change also gets rid of the params()
method and the Params typedef since they are functionally equivalent to
the ClockedObject versions.

It also removes the include of mem/port.hh which is not used in
mem_object.hh. This may break code which purposefully or (more likely)
accidentally depended on that transitive include from mem_object.hh.

Change-Id: I6dab3ba626e3f3ab6a6bd86edcf4f5cb4d6d2c45
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20720
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Reset HPPI when clearing an LPI

Change-Id: I2a69e6cef69aa48d7c265d59915b859e5eac2bcc
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20638
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Add resetHppi method in the GICv3 cpu interface

The method is used for resetting the highest priority pending interrupt
interrupt from the cpu interface if it matches the intid passed as an
argument.

Change-Id: I9fbc4cb3e05a1cc32f853b6afab5c2bc99369435
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20637
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Cleanup GICv3 initialization

This patch is removing the unnecessary initState() / reset() methods
from GICv3 classes, since we can initialize everything at
construction/init time

Change-Id: Ia70edcc4ca4f11878fac0024342e4f2cd81883a0
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20636
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Initialize GICD_TYPER once at construction time

Change-Id: Ib4dfdf7005709c22b4ba95099b1192f6edd6ff06
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20635
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Writes to IGRPEN1_EL3 triggering update

Change-Id: I56804eb1bfc8913bd0d3cab05865a382bf270bc1
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20634
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Fix GICv3 ITS cmdq wrapping

Change-Id: I979e8d1378d5b5d2647158798479cf4238f2c349
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20633
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Fix mapping between IGRPEN1_EL3 and IGRPEN1_EL1

Previous mapping was wrong because it was checking which security bits
it was accessing by using the inSecureState() function, whereas it
should have used the isSecureBelowEL3().  This patch is not making the
sostitution since it is optimizing the mapping furthermore by avoiding
updating both IGRPEN1_EL1 and  IGRPEN1_EL3 on writes.  The IGRPEN1_EL1
register is used as a storage, and any reads/writes to IGRPEN1_EL3 is
routed to that register.

Change-Id: Id318ec44e19d4f844e4e3410d74d0c4f89810811
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20632
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Implement message-based SPIs

Change-Id: I35e79dfd572c3e0d9cadc8e0aab01befd6004ece
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20631
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev: Scrub out some lingering uses of MemObject.

MemObject doesn't do anything any more, and is basically just an alias
for ClockedObject.

Change-Id: Ic0e1658609e4e1d7f4b829fbc421f222e4869dee
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20719
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Add GICD_SGIR register

The Distributor Software Generated Interrupt Register is implemented
only if affinity routing is disabled. Since this configuration is
currently not supported in gem5, it has to be treated as RES0.

Change-Id: I9ffcb31b26fc17547f74a4f1d43ce72c59786fa8
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20630
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

python: Make the dot writer handle unconnected Port vector elements.

Change-Id: I5234013716cdcce5fc39e239dc3d92cd1f2b7799
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20699
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev: Enable Terminal output's dump to stdout

While the default option is to dump the Terminal content in a file (e.g.
m5out/system.terminal), with this patch it will be possible to choose to
dump it to standard output.

Change-Id: If51c2fd671fa3eb0867a855b5f7d3b0df9cad025
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20639
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: State update when setting MISCREG_ICC_IGRPENx register

This is because by enabling ainterrupt group at the cpu interface, we
need to check if a previously pending interrupt needs to be forwarded to
the PE.
We are doing the same when globally enabling irqs in the distributor
(GICD_CTLR).

Change-Id: I80aeb87b2a58a108de899006d5a2f12eadbe6c2e
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20629
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm, dev-arm: MISCREG_ICC_IGRPEN1_EL1 using AA64 banking

Change-Id: Ic08ac1e7f3ebef408a83aa068ce15e9dfe2aa3cd
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20628
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm, dev-arm: MISCREG_ICC_AP1R0_EL1 using AA64 banking

Change-Id: Ide93464f62288fbe8f409f718487a15512c01295
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20627
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm, dev-arm: MISCREG_ICC_CTLR_EL1 using AA64 banking

Change-Id: Ib1691f1cba08251a36ceb959849b61c33cc3e93b
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20626
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: MISCREG_ICC_BPR1_EL1 using AA64 banking

Change-Id: Ib30c7a49490f05f88ddfd7572dd360cb92647f81
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20625
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Add read/writeBanked helpers to GICv3

These will be used by AA64 security banked registers in GICv3.

Change-Id: Ia980c4f5c14187ab9c18da1d1d596562644111ae
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20624
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: Add explicit AArch64 MiscReg banking

Change-Id: I89836d14491a51b1573f45c8012e3ad12b107d24
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20623
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: Use same template across all MSR inst

Change-Id: Ifb9f1db288e401761b71ccf426e370c475e5663f
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20622
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: SySDC64 Instructions (CMO) using MiscRegIndex

Change-Id: Ia66d6abf965b1d33579e8fa048608d99c93ff2ce
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20621
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev: Fix segmentation fault in VirtIOBlock

GEM5 got a segmentation fault when the size is large in
VirtIOBlock::write. This change uses a vector to avoid this segmentation
fault.

Signed-off-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
Change-Id: I26272686a6e7e39cdf2389657ecd38ce90261144
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20679
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: fix GDB stub after SVE

The SVE patches made registers longer by increasing NumVecElemPerVecReg,
but the GDB XML was not updated to account for that, and as a result GDB
connections were failing with:

Remote 'g' packet reply is too long

This commit introduces NumVecElemPerSimdVecReg which counts only the SIMD
register sizes to get it back working. SVE GDB support is not added here.

Change-Id: I4191b9f1999ae02b0308863db4cc9b5b16a27d6d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20468
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Rewrite ICC_BPR0/ICC_BPR1 handling

The patch is fixing BPR reads in AA32, by removing the line

Gicv3::GroupId group =
    misc_reg == MISCREG_ICC_BPR0_EL1 ? Gicv3::G0S : Gicv3::G1S;

Where a read to ICC_BPR0 will return a G1S group.
The patch is also fixing Security banking accesses.

Change-Id: I28f1d1244c44d4b8b202d3141f8380943c7c1c86
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20620
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Add GICv3 unimplemented Hyp Active Priorities Group regs

ICH_APxR1, ICH_APxR2, ICH_APxR3 are implemented only if supporting more
than 6 bits of priority. Since this is not the case, they are currently
unimplemented.
According to spec, unimplemented registers are RAZ/WI.

Change-Id: Ifd7f7a3d42b4575c2f7aff3b95d5a47ac1e61842
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20619
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Allow 32-bit access to GITS_TYPER

Change-Id: I9d19174b38ba70f82050102f955ccc162965d1fb
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20618
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Cpu interface groupEnabled check for global enable

Gicv3CPUInterface::groupEnabled should check for global enable flags at
distributor level:
- Gicv3Distributor.EnableGrp0
- Gicv3Distributor.EnableGrp1S
- Gicv3Distributor.EnableGrp1NS

Change-Id: I1c855b0e4c2bc8f1cd0a8f086b9450f516177b08
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20617
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Check if INTID group is enabled when reading HPPIRx

If it is not enabled, it should return INTID_SPOURIOUS

Change-Id: I4dfa8b9fcea874b4d281cd154dd38752b05e1d59
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20616
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Writing GICD_CTLR should trigger an update

This is the case where an interrupt is pending, but the distributor is
masking it. As soon as the group gets enabled, the interrupt should be
forwarded to the PE.

Change-Id: Ie428780bde7e4726688adf78dfcc4d43d1b45261
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20615
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Rewrite GICv3 update

The GICv3 update methods are method which are invoked anytime the model
needs to evaluate a change in its state, which most of the time means
managing the state of an interrupt (forwarding it to a PE, deasserting
it, etc).
The way it is currently done is a little bit obscure and doesn't
handle correctly IRQ prioritization.
Example:
An IRQ which is handled by the redistributor (PPI or LPI) was not
competing with any pending interrupts coming from the distributor (SPIs)
once raised by a peripheral.

Also the way the pending state of an interrupt was removed at the
cpu interface level wasn't happening in place where this was actually
happening (E.g. when activating it), but happened with a weird
fullUpdate semantic, where if there was a pending interrupt in a
cpu interface, all cpu interfaces had their pending interrupt (if any)
been disabled.

With this patch, state update always starts at the distributor, and
it goes down until the cpu interface where a Gicv3CPUInterface::update
method selects the winning interrupt coming from distributor/redistributor
to be forwarded to the PE.

Change-Id: I1c517cbc4bf107cc2d7ae7beb2692e3cf5187a40
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20614
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Fix GICv3 IGRPMOD writes

Writes to IGRPMOD were not right shifting the value, which resulted in
interrupts having a IGRPMOD value > 1, whereas the only allowed values
are 0 and 1.

Change-Id: Id491bd1b184d6e5abeeea25ea272eeb91792ccf7
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20613
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: SGI registers undecoded in AArch32

Change-Id: I64d3e639e1beaa507263637d59499aafeb5a19f8
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20612
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: Fix read/setMiscReg for AArch32 GICv3 ICC regs

The readMiscReg/setMiscReg methods were not forwarding register
reads/writes to the cpu interface when in AArch32.

Change-Id: Ide983e793b8033a88d31fe6ea87eaeffe9b093f5
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20611
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>

dev-arm: Fix SGI generation

The patch is fixing the following aspects of SGIs

* The conditons over which an SGI can be forwarded to a PE
* SGIs in AArch32 (see below)

It is in fact refactoring SGI generation under a common method in the
cpu interface. It is abandoning the implicit fallthrough mechanism not
only for cosmetic reasons, but also because checking "misc_reg ==" was
only working if the register was an AArch64 one (e.g.
MISCREG_ICC_SGI0R_EL1) and not the AArch32 counterpart (MISCREG_SGI0R).

Change-Id: I6fedfb80388666f4f1d20f6abef378a9f093aa83
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20610
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Gicv3 ITS device tree autogen

This patch adds device tree automatic generation for Gicv3 ITS.

Change-Id: Ic01500ffa691b331f527c5c2c785ff715660b0c2
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20609
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: modify GICv3 ITS default addr

The current default base address for GICv3 ITS stated in RealView is
0x2c120000. The redistributors base address is 0x2c010000; each
instantiated core has an associated redistributor with memory region
size 0x40000 (with GICv4 extension, enabled by default). With 8 cores,
the redistributor range spans to 0x2c210000, creating a conflict with
the ITS address space.

This patch changes the ITS base address to 0x2e010000 which guarantees
no overlapping with the redistributor.

Change-Id: I7dc1af9e69ac037f85ae96f0985554f1fb8372a0
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20608
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-x86: Adding warning for movnti

We are ignoring the non-temporal hint here, and implementing this
instruction as a cacheable instruction.

This change adds a warning to let user know about this workaround.

Change-Id: I2e40437a44282fe9cf7772a25a8870bd8729a6ed
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20428
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

dev-arm: Improper translation slot release in SMMUv3

The SMMUv3SlaveInterface is using the xlateSlotsRemaining to model a
limit on the number of translation requests it can receive from the
master device.

Patch

https://gem5-review.googlesource.com/c/public/gem5/+/19308/2

moved the resource acquire/release inside the SMMUTranslationProcess
constructor/destructor, for the sake of having a unique place for
calling the signalDrainDone.
While this is convenient, it breaks the original implementation,
which was freeing resources AFTER a translation has completed, but
BEFORE the final memory access (with the translated PA) is performed.
In other words the xlateSlotsRemaining is only modelling translation
slots and should be release once the PA gets produced.

The patch fixes this mismatch by restoring the resource release in
the right place (while keeping the acquire in the constructor)
and by adding a pendingMemAccess counter, which is keeping track
of a complete device memory request (translation + final access)
and will be used by the draining logic

Change-Id: I708fe2d0b6c96ed46f3f4f9a0512f8c1cc43a56c
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Adrian Herrera <adrian.herrera@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20260
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Implement invalidateASID in SMMUv3 WalkCache

This patch fixes a bug where issuing a invalidate-by-ASID command
(CMD_TLBI_NH_ASID) to the SMMU would cause Gem5 to crash.

Change-Id: I5b8343a17e43762fe3917560ae401a20be1e05b8
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20259
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Implement invalidateVA/VAA in SMMUv3 WalkCache

This patch implements VA/VAA invalidations in the SMMUv3 model.

As per SMMUv3.0 spec, if leaf bit is specified in the invalidation
command, only leaf entries within the walk cache need to be invalidated,
otherwise entries with intermediate translations are also invalidated.

Change-Id: I0eb1e1f1d8c00671a3c23d2a8fb756f2020d8d46
Reviewed-by: Michiel van Tol <michiel.vantol@arm.com>
Reviewed-by: Marc Mari Barcelo <marc.maribarcelo@arm.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20258
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-x86: implement movntq/movntdq instructions

Non-temporal quadword/double-quadword move instructions.
This change ignores the non-temporal hint and instructions are
implemented to send cacheable request to memory.
This would have some "performance" impact (i.e. having some cache
pollution) to get better "correctness" in behavior.

Change-Id: I2052ac0970f61a54bafb7332762debcb7103202d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20288
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

cpu: reset byte_enable across writeMem calls

data_write_req byteEnable which is used in ARM SVE partial writes was not
being zeroed between writes.

As a result, non-SVE memory write instructions such as STP that followed
SVE memory write instructions could still have the write mask active.

This could lead to wrong simulation behaviour, and to an assertion failure:

src/mem/packet.hh:1211: void Packet::writeData(uint8_t*) const: Assertion
`req->getByteEnable().size() == getSize()' failed. '`

Change-Id: I74b5a82675e9923b0ffdf2c1dd9afb00c91cb204
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20448
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev: Templatize PioPort.

When creating a base class which needs to be a SimObject, it's
necessary to decide ahead of time whether to use PioDevice or
BasicPioDevice in the hierarchy because they inherit from SimObject. If
they were added into the hierarchy later, then the original class would
inherit from SimObject, as would PioDevice. That would create a diamond
inheritance structure which would require virtual inheritance, and
that's a can of worms we'd rather not get into.

A big part of the PioPort mechanism is the PioPort itself which holds
a pointer to its parent device and delegates reads/writes to it. It
does that with a PioDevice pointer, and PioDevice declares virtual
functions for all the callbacks the port can call into.

Instead of that, this change templatizes PioPort based on the class of
the device that holds it. That will let you use a PioPort on *any*
class, as long as it has the methods PioPort depends on. That removes
the need to create an inheritance diamond to add a PioPort down the
line since PioDevice is no longer strictly required.

The PioDevice and BasicPioDevice classes are still around since they
still provide some additional functionality and there are existing
classes which depend on them.

Change-Id: I753afc1e0fa54b91217d54c1f8743c150537e960
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20568
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

ruby: Fix the way stall map size is checked for availability

To ensure that enqueuer observes the practical availability. We
check the message buffer queue size at the start of the cycle.
We also add the size of the stall queue to consider the total
queue size. However, messages can be moved from regular queue
to stall map. This leads to messages being considered twice leading
to false flow control. This patch fixes it by storing the stall map
size at the beginning of the cycle and considering it for checking
availability.

Change-Id: I6ea94f34fe5279b91f74e106d43263e55ec4bf06
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20389
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>