arch-arm: Provide SVE support to the TarmacParser

This patch is providing SVE support to the tarmac parser, so that
it is recognizing Vector & Predicate entries.

Change-Id: I268e621cffa05644d3f1d80170b067aacaa2d5ea
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Giacomo Gabrielli <giacomo.gabrielli@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21560
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arm: Set the number of FloatRegs to zero.

ARM no longer uses the floating point register file and uses the
vector registers instead. This avoids checkpointing a bunch of unused
registers, making it hard to tell where floating point instructions
are keeping their values, etc.

Change-Id: I23145ba750f1dd9ff5b815395e073c410120840d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22524
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

tests,base: Added GTests for base/match.cc

In order to aid testing the method "match.getExpressions()" has been added.

Change-Id: I11acf9bed286ee2809dfa3d05ef573dea85eb786
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22503
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Fix a bug in getCurrentInstCount in the checker CPU.

An earlier change accidentally left out the actualTC-> prefix in the
getCurrentInstCount method which was supposed to delegate the call to
another thread context. Without that, it just called itself and would
infinitely recurse.

This bug was pointed out in email by Robert Henry.

Change-Id: Ibf1fee6b48ff87790309c6d435bd76fa95c6cab9
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22623
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

power: Replace gtoh and htog with betoh and htobe.

We already know what endianness to use when within power.

Change-Id: Id4ced279d21c56855307a5a8da51654101a13786
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22371
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

x86: Replace htog and gtoh with htole and letoh.

We already know what endianness to use from within x86.

Change-Id: Ie92568efe8b23fbb7d9edad55fef09c6302cbe62
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22370
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mips: Replace gtoh and htog with letoh and htole.

We already know what endianness to use from within MIPS.

Change-Id: Ic4cd295a7a66c4c8ef55ebcf976fe6637567391f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22369
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

sparc: Replace htog and gtoh with htobe and betoh.

We know what endianness to use when we're implicitly working with
SPARC.

Change-Id: I85eaac1da087a8086b9450b762a52323f2498e2e
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22368
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

systemc: Remove boost dependency caused by tlm

This commit replaces the tlm header file, which caused the boost
dependency.

Change-Id: Ie4b1af71202522d8139e9a861144863097188072
Signed-off-by: Hoa Nguyen <hoanguyen@ucdavis.edu>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22624
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

fastmodel: Plumb the ITB and DTB through the IRIS thread context.

These might be necessary to, for instance, translate virtual addresses.
A custom TLB which uses the IRIS API will be written which can be
substituted in for the normal ARM TLB.

Change-Id: Ic44822db6692ca3a4ca13875b2260b08547a24da
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22116
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>

fastmodel: Implement inst count events in the IRIS thread contexts.

These use the IRIS stepping API.

Change-Id: Ib45744cb0928fece664187e4df6b25b064b19f0e
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22115
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: Simplify AMO code generation templates

This change simplifies the isa template for the atomic memory
operation (AMO). Previously the flow had unecessary if statements that
ended up breaking build using clang, due to variables that could
seemingly be used before they were unitialized.

Change-Id: I1b46dfd5f1e90377245c4f649c08b6532b507b9c
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22603
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Use std::array for registers in SimpleThread.

If the number of one of the register types is zero (useful on ARM in
the near future), memset will complain that it's given the length of
the array without multiplying by the size of the array elements. This
is a false positive since the length of the array and the number of
elements are both zero.

To avoid that warning/error and to simplify and update the SimpleThread
class slightly, this change replaces the C style arrays with
std::array.

Change-Id: Ifedd081a1940a578765c4d585e623236008ace67
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22523
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch-arm: Annotate original address in CMOs

This is needed when a CMO triggers an exception (e.g. DataAbort) In that
case the faulting address should be the one encoded in the instruction
rather than the cacheline address:

According to armarm:
If a memory fault that sets FAR_EL1 is generated from a data cache
maintenance or other DC instruction, FAR_EL1[63:0] holds the address
specified in the register argument of the instruction.

Change-Id: I6d0dadbef6e70db57438b01a76c5def3bdd2d974
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/+/22443
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-ruby: Reset Ruby Sequencer Outstanding Requests stats

Change-Id: I14b106e0eb7abd9c14badeedf35d6d1c9f198f98
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22446
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: optional instantiation of GICv3 ITS

GICv3 ITS is an optional component of GICv3. The previous behaviour
was for a stub ITS to be created by default, which resulted in a crash
for use cases where a GICv3 with no ITS is required.
This patch removes the instantiation of the ITS by default and adds
checks for its presence both in initialization and device tree
generation code.

Change-Id: Id424924c8c1152d512aaa2837de4aa60329ec234
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22423
Reviewed-by: Ciro Santilli <ciro.santilli@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Modify compressor to appease newer compilers

The type of the local unique_ptr variable was different from the return type.

In C++11 because of such difference, a copy-ellision would not be possible,
and that required the use of a std::move.

In C++14 the restriction of same types being required was removed, so
std::move would not be needed anymore.

With the addition of the -Wredundant-move warning in newer compilers, having
the std::move on the return became an issue, breaking compilation.

Change-Id: I45d18dfc500bb5db5fe360814feb91853c735a19
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22403
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>

mem-cache: Implement a perfect compressor

Implement a perfect compressor that always manages to compresses data
exactly to its maximum allowed compression ratio. This allows tracking
a compression upper bound.

Change-Id: Ibc68bf2dc84b75207795d5ba6304b9ed6dbeae8f
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21160
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>

mem-cache: Make BDI a multi compressor

BDI is a compressor containing multiple sub-compressors.

Change-Id: I98411e2ef9dcc2182801a172dfc59ed7a8ee7dd4
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21159
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>

mem-cache: Implement a multi compressor

Implement a compressor that contains multiple sub-compressors and
choses the one that provides the best compression results for each
compression.

Change-Id: I758cf67c84bd85edbea16b2a07b2068b00454461
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21158
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Implement BDI sub-compressors

Implement sub-compressors of BDI as public compressors so that
they can be used separately.

Change-Id: I710e35f39f4abb82fd02fd33b1b86a3f214c12cb
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21157
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Bobby R. Bruce <bbruce@ucdavis.edu>

mem-cache: Implement a repeated values compressor

The repeated values compressor can only compress data composed solely
repeated instances of the same value.

Change-Id: If2c4f47ad4af492d202ec2017e30ba52ee67e307
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21156
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Bobby R. Bruce <bbruce@ucdavis.edu>

mem-cache: Implement a zero compressor

The zero compressor can only compress data composed solely of zero
bits.

Change-Id: I8b359c03776a8748abd144a178bda944b5a1b766
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21155
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Bobby R. Bruce <bbruce@ucdavis.edu>

mem-cache: Implement FPC-D cache compression

Implementation of Frequent Pattern Compression with limited Dictionary
support (FPC-D) cache compressor, as described in "Opportunistic
Compression for Direct-Mapped DRAM Caches", by Alameldeen et al.

Change-Id: I26cc1646f95400b6a006f89754f6b2952f5b4aeb
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21154
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Bobby R. Bruce <bbruce@ucdavis.edu>

arch,cpu: Move endianness conversion of inst bytes into the ISA.

It doesn't matter if the bytes are converted before or after they're
fed into the decoder. The ISA already knows what endianness to use
implicitly, and this frees the CPU which doesn't from having to worry
about it.

Change-Id: Id6574ee81bbf4f032c1d7b2901a664f2bd014fbc
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22343
Maintainer: Gabe Black <gabeblack@google.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem: Delete the packet accessors which use guest endianness.

These accessors create an extra dependency on the guest OS, and can be
avoided. Now that all their uses have been removed, they aren't needed
any more.

Change-Id: I466c07fef99bce2d7964c07a7ac3dd398691378b
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/13465
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

arch-x86: Fix FLDCW_P and FNSTCW_P to use rip.

FLDCW_P and FNSTCW_P should use rip to compute address.

Change-Id: Ide7327e243d42bdd8791e43773385b2a79d45418
Signed-off-by: Zhengrong Wang <seanzw@ucla.edu>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22483
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

tests: add squashfs make to m5-exit

An ARM squashfs rootfs that runs m5 exit can be generated for example
with:

make ARCH=arm_A64 CROSS_COMPILE=aarch64-linux-gnu- squashfs

The existing Makefile.x86 was not used as a basis because we would
like to provide a setup that allows users to use their own compilers
if they wish, without requiring dockcross.

Change-Id: I19c54cf0575b405f191f45aaf1e4a05c3f2e69ae
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22223
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
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>

arch-arm: generic method for getting an ArmSystem

This patch provides a generic method for casting a System object
into an ArmSystem object. This is specially useful in dev-arm,
since devices by default obtain a generic System reference which
needs to be casted to use ArmSystem-specific functionality.

Change-Id: Ib100002413cb48cd93772dcf38f13be65badd1d3
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22426
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev-arm: Add SMMUv3 to VExpress_GEM5_V*

The VExpress_GEM5_V* Platforms will now optionally make use of the
SMMUv3.
In order to attach a devices to it, a user must simply use the
attachSmmu method, making sure the device it is not part of the
_on_chip_devices().

Change-Id: Ib819eb50d43dba1f5e5d1a1f7159ac4fbaccff6e
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/+/21559
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

config: Add --kernel-cmd option in fs_bigLITTLE.py

Allow users to specify their custom Linux kernel command.

Change-Id: I1e88523c918369ea5dd86c088eca82471663e76a
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20900
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

config: Add --mem-size option in fs_bigLITTLE.py

Allow users to specify system memory size.

Change-Id: I1e0c099dc08d04f71b406d8cc86850f68b6048cb
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/20899
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>

fastmodel: Add CortexA76x[234] models.

These use the parameterization added in earlier commits.

Change-Id: Id7b99b97894f8fc1f1e5cc34e3e5d32146fed1c7
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21505
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

fastmodel: Enable auto bridging and use it to simplify CortexA76x1.

This lets us avoid having to set up bridges for all the different
interrupt signals coming out of the CPU. When we have more cores, like
in the x2, x3, and x4 versions of the CPU, we won't have to have a
set of bridges for each set of signals, and can connect them all to
external ports using array notation, keeping everything simple,
concise, and maintainable.

Change-Id: I1a5f707073868516e93c106dc17d105409de668a
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21504
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Chun-Chen TK Hsu <chunchenhsu@google.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

fastmodel: Templatize the xn versions of the CortexA76.

This will make it a lot easier and more succinct to define the x2-x4
versions of that CPU.

Change-Id: I951cd3af4419c62892c57968e729fd11c0e4a59e
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21503
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu-o3,tests: add page boundary access test

The O3 model supports unaligned accesses across page boundaries.
This changeset provides a stress test for the feature.

Other benchmarks exercise it, but their coverage is spotty. This
test does nothing else except poke bytes which straddle page
boundaries.

Change-Id: Ic20c2862bcb7af7be091030c12916ba499db10c7
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22303
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

mem-cache: Fix missing header in associative set

Add missing intmath header to AssociativeSet, so that isPowerOf2 can
be used.

error: there are no arguments to 'isPowerOf2' that depend on a template
parameter, so a declaration of 'isPowerOf2' must be available

Change-Id: Ib2b194f9e71284ee439786bdb76d99858e57e2f5
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22444
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-ruby: Fixed pipeline squashes caused by aliased requests

This patch was created by Bihn Pham during his internship at AMD.

This patch fixes a very significant performance bug when using the O3
CPU model and Ruby. The issue was Ruby returned false when it received
a request to the same address that already has an outstanding request or
when the memory is blocked. As a result, O3 unnecessary squashed the
pipeline and re-executed instructions. This fix merges readRequestTable
and writeRequestTable in Sequencer into a single request table that
keeps track of all requests and allows multiple outstanding requests to
the same address. This prevents O3 from squashing the pipeline.

Change-Id: If934d57b4736861e342de0ab18be4feec464273d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21219
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Anthony Gutierrez <anthony.gutierrez@amd.com>
Tested-by: kokoro <noreply+kokoro@google.com>

configs: Add baremetal.py example script

Change-Id: I7a4818836decbf743f6c9b1f2e5361a43fca2d10
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21799
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-o3: bugfix for partial faults in x86

The c58cb8c9 changeset broke some code related to checking
consistency model guarantees (found in X86 benchmarks).

This changeset adds some documentation to the code and obviates
the problem.

Change-Id: Ied9c6b0b1d237538efe4beb2f97ef76248ce2746
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22283
Maintainer: Brandon Potter <Brandon.Potter@amd.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>

ext: Fix undefined-behavior bug in bitshift

If a small number or zero is passed in, fp64_exp could be very
negative (-1000 for example). The intent of the line is to evaluate to
zero in these cases, but what it actually did was bitshift right by
1000, which is undefined behavior (according to ubsan) that so happens
to result in 0 on GCC/most architectures. This commit changes the code
to check for cases where the bitshift is larger than the width of the
integer.

Change-Id: I8de4bd8ad170f0321d54689460de449b7f8fb60a
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21859
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

alpha: Convert htog and gtoh to htole and letoh.

When within the Alpha arch, we know the guest endianness already.

Change-Id: Iee22a33451a06c6e6a5c7d1e8c71f0b00e98ae3f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22367
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>

arch,sim: Make copyStringArray take an explicit endianness.

Change-Id: I5cf4291b19dd2d2bdbbf145ad8e00994fabf5547
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22366
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

kern: When dumping dmesg, detect the byte order dynamically.

The dmesg dumper has access to the system object and so has access to
the getGuestByteOrder accessor. Use that instead of TheISA to determine
the byte order.

Change-Id: I4df7b1bcd807aaced1d7dc8d2030123e2d4d1d2b
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22365
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

sim: Use the system and OS to get endianness.

This converts the syscall implementations to either use the
OS::byteOrder constant or, if that's not available, the system's
getGuestByteOrder() accessor, to determine the byte order, instead of
relying on TheISA to provide the correct accessor.

Change-Id: Idf7b02ee8d73990224ceac9a5efaec91a5ebf79f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22364
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arch: Make endianness a property of the OS class syscalls can consume.

That way the syscall implementations won't have to find the right
endianness to use on their own, typically by referring to TheISA.

Change-Id: I186b2f419d5dbee72cc9b5abce7356f3143f0c83
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22363
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev: Make the virtio devices track endianness explicitly.

These classes now track what endianness they're supposed to use
explicitly, initially set by the getGuestByteOrder accessor on the
system object. In the future, if the endianness depends on the
version of the VirtIO spec as the comment suggest, it will be easier
to dynamically set the endianness in the various structures based on
the version being used,

Since there isn't anything special about the virt IO versions of these
converters other than their types, and since the endianness conversion
infrastructure can be taught how to convert new types, the code was
switched over to using the standard htog and gtoh but with the
explicit byte order provided.

This also gets rid of the final use of TheISA in the dev directory.

Change-Id: I9345e3295eb27fc5eb87e8ce0d8d424ad1e75d2d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22273
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

sim: Add a getGuestByteOrder accessor to the system class.

This goes along with the existing getPageBytes, etc., accessors, and
paves the way for this to be a parameter of the System class.

Change-Id: Ibfe2d591185d23beccdd5bbff1092dc07b1278ac
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22272
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

misc: Update MAINTAINERS file

Add systemc tag and add Bobby as a tests maintainer.

Change-Id: Idc890365387f7e99543067d2e02ce0caadf785df
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22304
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Anthony Gutierrez <anthony.gutierrez@amd.com>
Tested-by: kokoro <noreply+kokoro@google.com>

fastmodel: Refactor the CortexA76x1 model for MP support.

This change inverts the relationship between the fast model and gem5
CPUs, and factors out the parts of the CortexA76x1 which are per core
vs. per cluster.

Change-Id: I33eacc2461f08c7fd1784936b230e96c768c0e79
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21501
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev: Adding support for large BAR

During PCI setup, this patch checks if a Base Address Register (BAR) is
used as a large BAR (64 bits rather than 32), and return proper address
range. The order which updates are done is decided by kernel, so this
patch implements both cases (writing lower or upper bits first).

Bit 2 in a BAR indicates a 64-bit decoder (10X to be more exact, 11X is
reserved).

The addresses in BARAddrs are full addresses and are set to zero for BAR
providing upper 32 bits to avoid conflicts in addr ranges reported.

Change-Id: I93303d36ac83dab9ed6837c81e77c9dfb778f409
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22082
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

base: Name segments after their index

Current loader is performing a linear scan of the section table for
every segment in the elf since it is naming every segment after the
sections it contains. With this patch we are just naming segments
after their index.
This is in any case how they are referenced when a readelf --segments
command is issued on the elf file.

Change-Id: I599400fcdfc0b80ac64632aba36781bd876777f0
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21999
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu-o3: Fix handling of some mem. order violations

This patch fixes the handling of memory order violations due to snoops
targeting out-of-order loads: the re-execution triggered in these cases
is achieved by raising a ReExec fault, but such a fault was not handled
correctly after the code changes introduced in changeset 46da8fb.

Change-Id: I2abe161a90468412f56cb28dcc92729326cba1cd
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21819
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Timothy Hayes <timothy.hayes@arm.com>
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

dev: Remove TheISA from ns_gige.cc.

It was not being used there.

Change-Id: Ib39cfb52553e0556f7a6ad616e1029c74fa90e1f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22271
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

dev: Stop including config/the_isa.hh unnecessarily.

This file was included in a few files which didn't use TheISA.

Change-Id: Ib296b88dc6cfe9d487ee31cf385bb872d2cffaf2
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22270
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

dev: Get PageBytes from the system in the ARM generic timer.

These will ultimately by ArmISA::PageBytes, but this is more consistent
with other devices which don't know what ISA they're part of.

Change-Id: Iac13d5010564512207ed009377a771ee5949eff3
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22269
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

dev: Get PageBytes from the system in the ide_disk model.

This avoids having to use TheISA::.

Change-Id: I020860ab343f9b6fafbcb0e23479d0b64f094512
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22268
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

dev: Remove using namespace std and TheISA from ide_disk.cc.

Neither was necessary since either that namespace wasn't used, or was
used explicitly when referenced.

Change-Id: I96d2791c09250255e7e65d47cbb6f65eb4fcb3ba
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22267
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

fastmodel: Helper function to setup FastModels for simulation

This function sets up ARM license, simulation name, and minimum
synchronize latency in FastModels. This function should be called once
per simulation.

Change-Id: Ic3408955aaff9f8b4e2b72d2f2b0da97b41bfc3f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22183
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

arm: Get rid of using namespace std and TheISA in realview.cc.

Neither was actually used by the nearly empty file.

Change-Id: Ief1b77b18c8c616511bf9870e1667439a42dfcbd
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22266
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

dev: Remove SINIC_VTOPHYS and related code.

The code in this #ifdef isn't turned on by anything, and either has or
likely will bitrot, especially since there are no tests to even
determine manually if the code they guard works. They are also
preceeded by panics which say that the code they guard is known not to
work now anyway.

This change also gets rid of TheISA in that file since the only reason
it was around was for vtophys in the guarded code.

Change-Id: I59fd8974d0dd3d7ab0d5a8ccfa6a446d2da41eb0
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22265
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

alpha: Remove TheISA from the Alpha devices.

These are the from the various bits of the tsunami platform. They
primarily consisted of "using TheISA" which could be replaced with
using AlphaISA or removed altogether (I went with the later), and use
of TheISA:: which I replaced with AlphaISA::.

Change-Id: Ic52577c65241a92a3f1ae318a19431f8faa50a66
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22264
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

x86: Remove TheISA from x86 devices.

This was really only in the PC platform class.

Change-Id: I5365d965ea335a7c45be9f80706a875b19ed0417
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22263
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

sparc: Create a helper functions to install firmware images.

The first function handles the repetitive process of creating an
ObjectFile for a particular purpose and checking if that was
successful.
The second conditionally offsets the images in case they were, for
instance, loaded from an ELF file which already had them in the right
place. It offsets them so that their entry point (which will be zero
for raw images) lines up with the appropriate entry address (which will
be at the start of raw images).

This is more correct in more cases, and also removes a lot of
redundancy. There's still a lot of redundancy in the code which sets
up the symbol tables, but there are some irregularities which make that
harder to wrap in a helper function.

Change-Id: I2fee8b2175faff284ff9e007307f7769043497a1
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21469
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Add a repeated value pattern to compressors

The repeated value pattern checks if values are composed of multiple
instances of the same value. If successful, the bits of the repeated
value are included only once in the compressed data.

Change-Id: Ia7045b4e33a91fd8d712fe1ca689f7f8cb4e5feb
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21153
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Inform unused bits instead of bytes in compressor pattern

Increase pattern precision by giving the number of unmatched bits
instead of bytes.

Change-Id: I5efbe9c31672cc973b4c89c741cdc8cc28d26285
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21152
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Add a masked const value pattern to compressors

The masked pattern compares values to masked const non-dictionary values
to determine whether these match. If successful, the bits that do not
match must be added to the compressed data.

Change-Id: I4c53568694dab916136fe384cb2ee10e554f7136
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21151
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Add a masked pattern to compressors

The masked pattern compares masked values to masked dictionary entries
to determine whether these values match. If successful, the bits that
do not match must be added to the compressed data.

Change-Id: I4b1c8feb0faa99576382b54a73a20c353f965d2a
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21150
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Add an uncompressed pattern to compressors

The uncompressed pattern always stores the original data, and therefore
it is always successful. All of the derived classes of the dictionary
compressor must have this pattern as the last pattern of the pattern
factory.

Change-Id: I2a38fd56630d88ef8b918220dc4c2824a196a8a2
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21149
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Templatize DictionaryCompressor

Templatize DictionaryCompressor so that the dictionary entries' sizes
can be changed.

Change-Id: I3d89e3c692a721cefcd7e3c55d2ccdefa425f614
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21148
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>

mem-cache: Factor out CPack's dictionary functionality

Factor out dictionary functionality of CPack, so that it can be
used easily for other compressors.

As a side effect, create an addToDictionary function to allow
subclasses to chose how to handle insertion.

Change-Id: I02fae4e98b02db5a40467ec470b71020d5e867cb
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21147
Tested-by: kokoro <noreply+kokoro@google.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>

mem-cache: Use shouldAllocate() instead of CPack's decompress()

Split decompression functionality using the proper function to
determine if a dictionary entry should be allocated after
decompression or not.

Change-Id: I4995304f4c4508c03c9fc1685f04511622969556
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21146
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>

mem-cache: Limit compression size

Add a threshold so that if the compressed size is greater than it,
the compression is abandoned, and the data is considered uncompressible.

Change-Id: Ic416195b06ec440a40263b75bd0f0383cde2ea6a
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21144
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>

mem-cache: Do not try to compress dataless packets

Fix filling blocks so that packets that do not contain data do not
generate a compression attempt. This can happen, for example, with
invalidation responses, which will trigger a packet data access
assertion.

Change-Id: I2a1e7983657f6e5e770b148ab62c9de9ac3986ac
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22164
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

ext: Remove non-source files from fputils

Remove the autoconf-based build system and GNU-style information files
from fputils. After this change, we only keep the files we will need
to integrate into gem5's main source tree.

Change-Id: I2ddf1d07d9cb51bcd91fc63f1ae43c7f46129933
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22163
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

mem: Fix DRAM controller to operate on its own address space

Typically, a memory controller is assigned an address range of the
form [start, end). This address range might be interleaved and
therefore only a non-continuous subset of the addresses in the address
range is handed by this controller.

Prior to this patch, the DRAM controller was unaware of the
interleaving and as a result the address range could affect the
mapping of addresses to DRAM ranks, rows and columns. This patch
changes the DRAM controller, to transform the input address to a
continuous range of the form [0, size). As a result the DRAM
controller always operates on a dense and continuous address range
regardlesss of the system configuration.

Change-Id: I7d273a630928421d1854658c9bb0ab34e9360851
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/19328
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Wendy Elsasser <wendy.elsasser@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mem-cache: Avoid promotion of incompatible deferred targets

Often a request that hits on an MSHR has to be deferred as it can't be
serviced by the current response.

For example, a request that requires writable has to be deferred when
the response is expected to bring in a read-only copy of the
block. However, there are cases where the response, although not
expected to do so, brings a writable copy and as a result we also
service deferred targets. In such cases, we promote deferred targets
up until the first that can't be serviced by the current response
(e.g., cache maintainance operation). If the first deferred target is
incompatible we don't promote any targets at all.

Change-Id: Ib3e13be51120b7c0f0053b83b76bde03e1b7dd4e
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22127
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

mem-cache: Fix MSHR whole line write tracking

The MSHR keeps track of outstanding writes and services them as a
whole line write whenever possible. To do this the outstanding writes
have to be compatible (e.g., not strictly ordered). Prior to this
change, due to this tracking mechanism, the MSHR would not service a
WriteLineReq with flags that do not allow merging as a full line write
even if it was the first target triggering an assertion. This
changeset fixes this bug.

Change-Id: I2cbf5ece0c108c1fcfe6855e8f194408d5ab8ce2
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22126
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Tested-by: kokoro <noreply+kokoro@google.com>

configs: Update HPI to use ArmDTB and ArmITB.

Change-Id: I67693e9d79e89f151e30c585ad565deac53c77c1
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22203
Reviewed-by: Ciro Santilli <ciro.santilli@arm.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

configs: Fix undefined BaseCPU

When using NULL ISA BaseCPU is undefined, and therefore the
isinstance call generates a NameError.

Change-Id: Ia4582606b775cdb20829966f8e312a333a55b6f3
Signed-off-by: Daniel R. Carvalho <odanrc@yahoo.com.br>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21959
Reviewed-by: Bobby R. Bruce <bbruce@ucdavis.edu>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

tests, base: Fixed incorrect implementation of StrTest.ToLower test.

Change-Id: I96672052f8c9da9d4f61ff0e8eed324032b1afac
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22123
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

ext: Update fputils to rev 0721843

This patch updates fputils to the latest revision (0721843) from the
upstream repository (github.com/andysan/fputils). This effectively
only updates the license of the library to match gem5's 3-clause BSD
license.

Change-Id: I46d8a5dfac2f4a95e66ee82a15288ac424d7df90
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22125
Reviewed-by: Anthony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Anthony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>

mips,riscv: Get rid of some Alpha cruft in these System classes.

The consolePanicEvent pointer and addConsoleFuncEvent template were
inherited from Alpha and were not used (and probably make no sense) for
MIPS or RISCV which (to my knowledge) don't have the idea of a
"console" binary.

Change-Id: I109b866a65f69c7334062f7304c7b18acc51d99d
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/21782
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>

cpu: Get rid of the nextInstEventCount method.

This was only used by the KVM CPU, and it has access to all it needs to
figure out that value locally without requiring all the ThreadContexts
to implement an equivalent function.

Change-Id: I17a14ce669db2519edf129db761ebd8dc3bd4129
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22114
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Get rid of the serviceInstCountEvents method.

This was useful when transitioning away from the CPU based
comInstEventQueue, but now that objects backing the ThreadContexts have
access to the underlying comInstEventQueue and can manipulate it
directly, they don't need to do so through a generic interface.

Getting rid of this function narrows and simplifies the interface.

Change-Id: I202d466d266551675ef6792d38c658d8a8f1cb8b
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22113
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

fastmodel: Use getCurrentInstCount for totalInsts().

This had been using a custom totalInsts method on the iris
ThreadContext, but since that's equivalent to what the totalInsts
method does only through a different mechanism, we can
drop that and use getCurrentInstCount instead.

Change-Id: I058fec13e81f28285281e136635d53a2e849cb47
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22112
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

fastmodel: Implement getCurrentInstCount.

This uses the step counter the iris API provides.

Change-Id: Ic916888fa256d0aa65042d3e6695d9bf4ba32c86
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22111
Reviewed-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Maintainer: Giacomo Travaglini <giacomo.travaglini@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Switch off of the CPU's comInstEventQueue.

This switches to letting the ThreadContexts use a thread based/local
comInstEventQueue instead of falling back to the CPU's array. Because
the implementation is no longer shared and it's not given where the
comInstEventQueue (or other implementation) should be accessed, the
default implementation has been removed.

Also, because nobody is using the CPU's array of event queues, those
have been removed.

Change-Id: I515e6e00a2174067a928c33ef832bc5c840bdf7f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22110
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Access inst events through ThreadContext instead of the CPU.

Also delete the CPU interface.

Change-Id: I62a6b0a9a303d672f4083bdedf393f9f6d07331f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22109
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Delegate comInstEventQueue methods to the ThreadContexts.

These then just use the comInstEventQueue array from the CPU, but soon
they will actually be self contained and allow the thread context to
use whatever mechanism it wants.

Also, now that the thread contexts need to exist before instruction
count based events can be scheduled, setting up max instruction based
events needs to happen in init after the CPU subclasses have had a
chance to set up the threadContexts vector.

Change-Id: I34bb401633d277a60be74e30d5a478a149b972ea
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22108
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Make accesses to comInstEventQueue indirect through methods.

This lets us move the event queue itself around, or change how those
services are provided.

Change-Id: Ie36665b353cf9788968f253cf281a854a6eff4f4
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22107
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu,sim: Delegate PCEvent scheduling from Systems to ThreadContexts.

The System keeps track of what events are live so new ThreadContexts
can have the same set of events as the other ThreadContexts.

Change-Id: Id22bfa0af7592a43d97be1564ca067b08ac1de7c
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22106
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Make the ThreadContext a PCEventScope.

Both the thread and system's PCEventQueue are checked when appropriate.

Change-Id: I16c371339c91a37b5641860d974e546a30e23e13
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22105
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu,sim: Get rid of a bunch of conditional compilation for PCEvents.

These can now be built without referring to anything in ThreadContext
and so can be built even with the NULL ISA. This means the pcEventQueue
can be unconditionally built into the System class. Even though the
pcEventQueue is going away, this still makes it possible for System to
be a PCEventScope unconditionally.

Change-Id: Ia342bb7972b1b5ce95033176d72af4bfa343560f
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22104
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Don't print the CPU name when a (Break|Panic)PCEvent happens.

This requires reaching into the threadcontext to access the CPU
pointer, and also isn't all that useful since it's more important what
event happened, not what CPU happened to be running the code at that
time.

Change-Id: I368707c804dff9bd349f3261bdcd08be55c5d04a
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22103
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Pass the address to check into the PCEventQueue service method.

This prevents having to access it from within the ThreadContext.

Change-Id: I34f5815a11201b8fc41871c18bdbbcd0f40305cf
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22102
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

sim: Make the System object a PCEventScope.

This abstracts away the raw PCEventQueue managed by the System.

Change-Id: I04d773e6be90a891884a76841f15c3eecd5796ed
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22101
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Gabe Black <gabeblack@google.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Stop checking for PC changes when servicing a PCEventQueue.

First of all, this would arbitrarily skip events based on when they
were encountered in the queue. Second, this is one of the three places
where the ThreadContext is actually accessed in pc_event.cc. By
removing this and the other uses, this file can be included even when
using the NULL ISA, and a lot of #ifdefs can be removed.

Change-Id: If81f5e9ff9d3f9833145fec0b6062b4bda8d2e47
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22100
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

cpu: Create a PCEventScope class to abstract the scope of PCEvents.

This abstraction will allow scheduling PCEvents for a particular
ThreadContext, all contexts on a CPU, all contexts in a system, etc.,
and delegates scheduling and removing events to each particular scope.

Right now the PCEventQueue is the only implementor of the PCEventSCope
interface.

Change-Id: I8fb62931511136229915c2e19d36aae7ffdec9df
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22099
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Tested-by: kokoro <noreply+kokoro@google.com>

tests: Added GTests for base/str.cc

Adding these tests supercedes the unittest/strnumtest.cc
and unittest/tokentest.cc tests. They have thereby been removed.

Function "to_number" in base/str.hh previously failed to cast negative
float/double numbers. This was due to the use of
std::numeric_limits<T>::min() instead of std::numeric_limits<T>::lowest()
to determine whether a string-to-float/double conversion was
"Out of range". Tests "StrTest.ToNumberFloatNegative" and
"StrTest.ToNumberDoubleNegative" exposed this bug. It has been fixed.

Methods "split_first" and "split_last" in base/str.hh have had their
documentation updated to remove abiguity in their functionality.

Change-Id: I16e0fe40d884e22dd010db4045857eb6e7f33d4a
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/22084
Reviewed-by: Daniel Carvalho <odanrc@yahoo.com.br>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Tested-by: kokoro <noreply+kokoro@google.com>