The style checker complains about line length and ordering for these
files. This fix should make these two files kosher.
Change-Id: I822a0518a98d9e379a543d2017e90c4e9666a58d
Reviewed-on: https://gem5-review.googlesource.com/3380
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Brandon Potter <Brandon.Potter@amd.com>
If a (regular) store is followed closely enough by a locked load that
overlaps, the LSQ will forward the store's data to the locked load and
never tell the cache about the locked load. As a result, the cache will
not lock the address and all future store-conditional requests on that
address will fail. This patch fixes that by preventing forwarding if
the memory request is a locked load and adding another case to the LSQ
forwarding logic that delays the locked load request if a store in the
LSQ contains all or part of the data that is requested.
[Merge second and last if blocks because their bodies are the same.]
Change-Id: I895cc2b9570035267bdf6ae3fdc8a09049969841
Reviewed-on: https://gem5-review.googlesource.com/2400
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Tony Gutierrez <anthony.gutierrez@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
We don't implement the GICD_IGROUPRn registers, which is allowed, but
to be correct, they should be RAZ/WI (read as zero, writes ignored).
Change-Id: I8039baf72f45c0095f41e165b8e327c79b1ac082
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2620
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Currently, if the remote gdb stub fails to read a byte from an incoming
packet because the connection has been dropped, the read call will return
anyway and the calling code will have no way to know something bad
happened. It might reattempt the read over and over again waiting for some
particular byte, doomed to never make forward progress.
This change modifies the remote GDB code so that if a read or write call
fails, it will instead detach from the debugger and continue. Before this
change, When simulating a port scan, ie connecting to the debugger port
and then immediately dropping the connection using this command:
nc -v -n -z -w 1 127.0.0.1 7000
gem5 would enter the previously described death spiral. After it, gem5
detaches from the bad connection and resumes execution. Subsequently
attaching with gdb was successful.
This code is written in a C centric style, and would benefit from some
refactoring.
Change-Id: Ie3c0bb35b9cfe3671d0f731e3907548bae0d292f
Reviewed-on: https://gem5-review.googlesource.com/3180
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Clang's UBSAN implementation complains about macros that expand to
expressions that contain 'defined'. Explicitly set the various feature
macros to 1 or 0 to avoid this issue.
Change-Id: Iba239dacfe526c43ab9c5da5183a350fc4fdc57d
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Andreas Hansson <andreas.hansson@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/3241
Factor out the ISA ness of the switching header generating function. Also
turn it into a SCons builder which builds a single header, and a wrapping
method which uses the builder on a group of header files which all target
the same subdirectory.
Change-Id: I87705f97b6ebd9baebd4ebcfea19cc1218a64ad0
Reviewed-on: https://gem5-review.googlesource.com/2983
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
This option invokes the Listener::loopbackOnly() static function which
will make the port listeners bind to the loopback device exclusively and
ignore connections on other devices. That prevents external agents like
port scanners from disrupting simulations with spurious connections.
Change-Id: I46b22165046792a6f970826c109bdbce7db25c84
Reviewed-on: https://gem5-review.googlesource.com/3082
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
The loopback device will allow access to various services like remote GDB
debugging, connecting to the terminal, etc., without letting external
agents like port scanners connect and disrupting the simulation.
Change-Id: I76dccbf152fa278ae9f342b25f7e345a1329fbe4
Reviewed-on: https://gem5-review.googlesource.com/3080
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
The PyBind11 changes slightly modified gem5's internal debug
interfaces. The corresponding change to the public API went missing
before the new bindings were merged. This change updates the Python
glue to use the new interface.
Change-Id: I3ecca5a3f6c35b99d55126d697371124f81a12dd
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Matteo Andreozzi <matteo.andreozzi@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/3140
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Add support for a memory mapped m5op interface. When enabled, the TLB
intercepts accesses in the 64KiB region designated by the
ArmTLB.m5ops_base parameter. An access to this range maps to a
specific m5op call. The upper 8 bits of the offset into the range
denote the m5op function to call and the lower 8 bits denote the
subfunction.
Change-Id: I55fd8ac1afef4c3cc423b973870c9fe600a843a2
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2964
The state transfer code wasn't reading back PSTATE correctly from the
CPU prior to updating the thread context and was incorreclty writing
the register as a 32-bit value when updating KVM. Correctly read back
the state before updating gem5's view of PSTATE and cast the value to
a uint64_t.
Change-Id: I0a6ff5b77b897c756b20a20f65c420f42386360f
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2963
Reviewed-by: Rahul Thakur <rjthakur@google.com>
For Value() nodes, the get_contents() method and the get_text_contents()
method are just aliases to the same thing, both of which return a value
which has already been converted using str(). The str() which was included
explicitly in the SConscript was redundant, except that it showed that the
value from get_contents was being treated as a string. To avoid the
redundancy but to still leave a hint to the value's type, this change
converts those bits of code to use get_text_contents() and removes the
str().
Change-Id: I8f7da9b652f749e741b553c9a0e2248ae15ec3ca
Reviewed-on: https://gem5-review.googlesource.com/3084
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Use the PyBind11 wrapping infrastructure instead of SWIG to generate
wrappers for functionality that needs to be exported to Python. This
has several benefits:
* PyBind11 can be redistributed with gem5, which means that we have
full control of the version used. This avoid a large number of
hard-to-debug SWIG issues we have seen in the past.
* PyBind11 doesn't rely on a custom C++ parser, instead it relies on
wrappers being explicitly declared in C++. The leads to slightly
more boiler-plate code in manually created wrappers, but doesn't
doesn't increase the overall code size. A big benefit is that this
avoids strange compilation errors when SWIG doesn't understand
modern language features.
* Unlike SWIG, there is no risk that the wrapper code incorporates
incorrect type casts (this has happened on numerous occasions in
the past) since these will result in compile-time errors.
As a part of this change, the mechanism to define exported methods has
been redesigned slightly. New methods can be exported either by
declaring them in the SimObject declaration and decorating them with
the cxxMethod decorator or by adding an instance of
PyBindMethod/PyBindProperty to the cxx_exports class variable. The
decorator has the added benefit of making it possible to add a
docstring and naming the method's parameters.
The new wrappers have the following known issues:
* Global events can't be memory managed correctly. This was the
case in SWIG as well.
Change-Id: I88c5a95b6cf6c32fa9e1ad31dfc08b2e8199a763
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Andreas Hansson <andreas.hansson@arm.com>
Reviewed-by: Andrew Bardsley <andrew.bardsley@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2231
Reviewed-by: Tony Gutierrez <anthony.gutierrez@amd.com>
Reviewed-by: Pierre-Yves Péneau <pierre-yves.peneau@lirmm.fr>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
A previous change forced scons to spawn child processes by exec-ing it
directly rather than going through the shell because the command line
length would be too long for the shell to handle. Now that incremental
linking should keep the command line lengths more under control, that
change should no longer be necessary.
Change-Id: I9e82a62083afd1414324a7fd697bd6d4b76367ae
Reviewed-on: https://gem5-review.googlesource.com/2947
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
The groups won't be perfectly balanced or optimally planned, but this
requires no thought and breaks the object files down into a reasonable
number of reasonably sized groups.
Change-Id: I6542fc807aaf356a9be751093f68e2e29f0b1586
Reviewed-on: https://gem5-review.googlesource.com/2946
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
The groups will be linked together into intermediate partially linked
object files. Right now the hierarchy is assumed to be flat, but with some
effort it could be extended to allow truly hierarchical linking.
Change-Id: I77b77710554e5f05e8b00720a0170afaf4afac2d
Reviewed-on: https://gem5-review.googlesource.com/2945
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Add a case for the ATA command ATAPI_IDENTIFY_DEVICE.
This avoids the panic: Unsupported ATA command when booting a recent Linux
kernel. This was tested on 4.8.13.
Change-Id: Ib297a2c02da0730d8698c59801254dd0f5ee9f7f
Signed-off-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-on: https://gem5-review.googlesource.com/2863
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Gabe Black <gabeblack@google.com>
When a branch micro-op belongs to a flow and the micro-op does not change
the nPC and just updates the nuPC (like a 'rep movs' flow), branching()
function always returns not-taken no matter actual micro-branch outcome.
Provided fix adds to the equation nuPC attribute checking since these kind
of branch micro-op only updates that pointer.
This issue has been found while debugging the performance of a copy-loop
implemented with memcopy function. Without the fix, 'rep movss' internal
micro-branch was always predicted as not-taken causing an squash event
after every branch micro-branch execution.
Using the provided test, branch mispredition went from 1922 without the fix
to 7.
Change-Id: I1bcbefae26aef47e3135817ef99b53d0ea0a98fa
The command line can be too long, causing bash to choke. This means we can't
use any shell syntax like shell variables or redirection when linking, but
that should be easy to avoid.
Change-Id: Ie6c8ecab337cef6bd3c7e403346ced06f46f0993
Reviewed-on: https://gem5-review.googlesource.com/2780
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Steve Reinhardt <stever@gmail.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Allow global stats such as sim_seconds in power equations to make it
possible to convert from event stats to rate stats.
Change-Id: I429abe0ffadc0dbd162eb39e0897143be472ef65
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2720
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
This patch fixes a bug where increasing the mmap region too much causes
it to run into already-allocated memory, which causes gem5 to fail an
assertion. Previously, the stack was incorrectly set up such that the
end of the mmap region and the top of the stack were the same address
and both would grow downward. With this patch, the top of the stack has
been separated from the end of mmap and moved up, and the mmap region
now grows upward instead of downward.
[Rebase to master branch and remove dependencies.]
Change-Id: I7271ff478fff2994f918bc5003a6139b9ba6a520
Reviewed-on: https://gem5-review.googlesource.com/2680
Maintainer: Alec Roelke <ar4jc@virginia.edu>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
There is currently no good way of extracting the current operating
voltage in MathExprPowerModels. This change adds a magic variable,
'voltage', that can be referenced from such expressions to get the
current operating voltage.
Change-Id: Ice3c9a4a221921a542de5da52f83f3f88862d246
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Sascha Bischoff <sascha.bischoff@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2662
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
MathExprPower model currently doesn't print any useful error messages
if an expression fails to evaluate. To add insult to injury, the model
only detects a failure when dumping stats and not at
initialization. This change adds a verification step in startup() that
ensures that all of the referenced stats actually exist.
Change-Id: I8f71c73341578d5882c8d93e482f5383fbda5f1d
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Sascha Bischoff <sascha.bischoff@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2661
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Multiple outstanding DMA requests introduced new DMA states that didn't
be considered into slicc code. This patch implements the missed DMA state
changes on MOESI_CMP_directory protocol.
Change-Id: I700d441d76556b7e77e0d507904af6ec6ba59cc2
Signed-off-by: Michael LeBeane <michael.lebeane@amd.com>
Reviewed-on: https://gem5-review.googlesource.com/2380
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Michael LeBeane <Michael.Lebeane@amd.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Some of the functions in the Linux toolchain that allocate memory make
use of paired LR and SC instructions, which didn't work properly for
that toolchain. This patch fixes that so attempting to use those
functions doesn't cause an endless loop of failed SC instructions.
Change-Id: If27696323dd6229a0277818e3744fbdf7180fca7
Reviewed-on: https://gem5-review.googlesource.com/2340
Maintainer: Alec Roelke <ar4jc@virginia.edu>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Previously, RISC-V in gem5 only supported RISC-V's Newlib toolchain
(riscv64-unknown-elf-*) due to incorrect assumptions made in the initial
setup of the user stack in SE mode. This patch fixes that by referring
to the RISC-V proxy kernel code (https://github.com/riscv/riscv-pk) and
setting up the stack according to how it does it. Now binaries compiled
using the Linux toolchain (riscv64-unknown-linux-gnu-*) will run as
well.
[Update for recent changes to MemState to add accessors and mutators to
get its members.]
Change-Id: I6d2c486df7688efe3df54273e9aa0fd686851285
Reviewed-on: https://gem5-review.googlesource.com/2305
Maintainer: Alec Roelke <ar4jc@virginia.edu>
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
This patch adds support for debugging with remote GDB to RISC-V. Using
GDB compiled with the RISC-V GNU toolchain, it is possible to pause
and continue execution, view debugging information, etc. As with the
rest of RISC-V, this does not support full-system mode.
Change-Id: I2d3a8be614725e1be4b4c283f9fb678a0a30578d
Reviewed-on: https://gem5-review.googlesource.com/2304
Maintainer: Alec Roelke <ar4jc@virginia.edu>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Previously, if a memory operation referenced an address that caused the
data to wrap around to the beginning of the memory (such as -1 or
0xFFFFFFFFFFFFFFFF), an assert would fail during address translation and
gem5 would crash. This patch fixes that by checking for such a case in
RISC-V's TLB code and returning a fault from translateData if that would
happen. Because RISC-V does support unaligned memory accesses, no
checking is performed to make sure that an access doesn't cross a cache
line.
[Update creation of page table fault to use make_shared.]
[Add comment explaining the change and assertion that the memory request
isn't zero size.]
Change-Id: I7b8ef9a5838f30184dbdbd0c7c1655e1c04a9410
Reviewed-on: https://gem5-review.googlesource.com/2345
Maintainer: Alec Roelke <ar4jc@virginia.edu>
Reviewed-by: Brandon Potter <Brandon.Potter@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Sometimes an ld instruction will be split across a
cache boundary. Previously RISC-V was set to not
allow this. This patch fixes that.
Change-Id: I8bc8ea6d67f65a9b3662e14c4037f4224799d20f
Reviewed-on: https://gem5-review.googlesource.com/2341
Maintainer: Alec Roelke <ar4jc@virginia.edu>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
The Binary Point Register (BPR) specifies which bits belong to the
group priority field (which are used for preemption) and which to the
subpriority field (which are ignored for preemption).
Change-Id: If51e669d23b49047b69b82ab363dd01a936cc93b
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2443
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Weiping Liao <weipingliao@google.com>
The aforementioned registers (Interrupt Processor Targets Registers) are
banked per-CPU, but are read-only. This patch eliminates the per-CPU
storage of these values that are simply computed.
Change-Id: I52cafc2f58e87dd54239a71326c01f4923544689
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2442
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Weiping Liao <weipingliao@google.com>
Instructions that use the coprocessor interface check the current
program status to determine whether the current context has the
priviledges to read from/write to the coprocessor. Some modes allow
the execution of coprocessor instructions, some others do not allow it,
while some other modes are unexpected (e.g., executing an AArch32
instruction while being in an AArch64 mode).
Previously we would unconditionally trigger a panic if we were in an
unexpected mode. This change removes the panic and replaces it
with an Undefined Instruction fault that triggers if and when a
coprocessor instruction commits in an unexpected mode. This allows
speculative coprocessor instructions from unexpected modes to execute
but prevents them from gettting committed.
Change-Id: If2776d5bae2471cdbaf76d0e1ae655f501bfbf01
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Rekai Gonzalez Alberquilla <rekai.gonzalezalberquilla@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2281
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Weiping Liao <weipingliao@google.com>
A completed write to a memory location that is Write-Through Cacheable
has to be visible to an external observer without the need of explicit
cache maintenance. This change adds support for Write-Through
Cacheable Normal memory and treats it as Non-cacheable. This incurs a
small penalty as accesses to the memory do not fill in the cache but
does not violate the properties of the memory type.
Change-Id: Iee17ef9d952a550be9ad660b1e60e9f6c4ef2c2d
Reviewed-by: Giacomo Gabrielli <giacomo.gabrielli@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2280
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
There are cases where Drainable objects need to create new objects in
Drainable::resume(). In such cases, the local drain state will be
inherited from the DrainManager. We currently set the state to Running
as soon as we start resuming the simulator. This means that new
objects are created in the Running state rather than the Drained
state, which the resume code assumes. Depending on the traversal order
in DrainManager::resume(), this sometimes triggers a panic because the
object being resumed is in the wrong state.
This change introduces a new drain state, Resuming, that the
DrainManager enters as soon as it starts resuming the
simulator. Objects that are created while resuming are created in this
state. Such objects are then resumed in a subsequent pass over the
list of Drainable objects that need to be resumed. Once all objects
have been resumed, the simulator enters the Running state.
Change-Id: Ieee8645351ffbdec477e9cd2ff86fc795e459617
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2600
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Weiping Liao <weipingliao@google.com>
The kernel and gem5 derive MPIDR values from CPU IDs in slightly
different ways. This means that guests running in a multi-CPU setup
sometimes fail to bring up secondary CPUs. Fix this by overriding the
MPIDR value in virtual CPUs just after they have been instantiated.
Change-Id: I916d44978a9c855ab89c80a083af45b0cea6edac
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-by: Sascha Bischoff <sascha.bischoff@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2461
Reviewed-by: Weiping Liao <weipingliao@google.com>
