Suspending the current thread context while draining due to a quiesce
pseudo instruction (for example a wfi instruction) could deadlock the
cpu and prevent it from successfully draining. This change ensures
that the cpu is not draining before suspending the thread context.
Change-Id: I7c019847f5a870d4bc9ce2b19936bc3dc45e5fd7
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5881
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Events were not being attached to counters after a checkpoint resume.
By not storing the enable private variable from the stored state the
recreation of the event to counter association is automatically carried.
The enable state is stored in the reg_pmcnten.
Change-Id: I46344df0882a9050c900efb2e8996d64dbfbf297
Reviewed-by: Sascha Bischoff <sascha.bischoff@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5761
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Add the power_gating_on_idle option to control whether a core
automatically enters the power gated state. The default behaviour is
to transition to clock gated when idle, but not to power gated. When
this option is set to true, the core automatically transitions to the
power gated state after a configurable latency.
Change-Id: Ida98c7fc532de4140d0e511c25613769b47b3702
Reviewed-on: https://gem5-review.googlesource.com/5741
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
If the CPU has been clock gated for a sufficient amount of time
(configurable via pwrGatingLatency), the CPU will go into the OFF
power state. This does not model hardware, just behaviour.
Change-Id: Ib3681d1ffa6ad25eba60f47b4020325f63472d43
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/3969
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
A recent serial device refactoring changed the name of the parameter
that the terminal device gets attached to on the UART. The x86 and
SPARC platform devices didn't get updated though, and were still using
the old name. This change updates those objects.
Reported-by: Kanad Basu <kanad.kut@gmail.com>
Change-Id: I0824a9df8639062d8561420ea9ffea26b8b7e2e9
Reviewed-on: https://gem5-review.googlesource.com/5781
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
This test has been broken since 70176fecd1 in 2014. The problem was
that the array size in the test was technically not constant because it
was based on an int variable that wasn't declared as const. That
prevented g++ from resolving it as a template parameter. Before the
change mentioned above, the implementation wasn't based on templates.
Change-Id: I6819cf522f9ba4636ac661da368b9bcbae0a813f
Reviewed-on: https://gem5-review.googlesource.com/5821
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Recent Linux kernels for AArch64 have changed their start addresses
but we still want to relocate the kernel to 0x80080000 which
required hacking the load_addr_mask in Realview.py to be 0x7ffffff
from 0xfffffff to mask off the proper number of MSBs to load the
kernel in the desired location. To avoid having to make this change
in the future again, we auto-calculate the load_addr_mask if it is
specified as 0x0 in the System sim-object to find the most restrictive
address mask instead of having the configuration specify it. If the
configuration does specify the address mask, we use it instead of
auto-calculating.
Change-Id: I18aabb5d09945c6e3e3819c9c8036ea24b6c35cf
Signed-off-by: Geoffrey Blake <Geoffrey.Blake@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/2323
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
The aarch64 boot loader was distributed using a BSD license that was
using non-standard formatting. Updated the license to match gem5's
canonical license format and removed the separete LICENSE.txt file.
Change-Id: I660b73ca5ddd922763a2b72051c73d539248ebcf
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5728
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Added a check to avoid unlocking/locking the same event queue. Also,
added an optional parameter to enable the migration to be skipped. This
can be useful to disable the synchronization for certain runtime
conditions.
Change-Id: I4b03b3ffff4f9503153cd41dd8aa78705bf16cc4
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5730
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
This change adds a script to generate graphs from the stats file
output by the configuration script low_power_sweep.py.
The graphs show stacked bars for time spent and energy consumed
wherein each component of the stacked bar represents a DRAM power
state (Idle, Refresh, Active, Active Power-down, Precharge Power-down
and Self-refresh). The script generates one plot per delay value. It
also generates a pdf (--pdf option) in which the graphs are laid out
such that you can easily compare how the increasing delay and other
swept params affect the resulting energy.
Change-Id: Id80b0947bfde27e11e5505b23a3adb30f793a43f
Reviewed-by: Wendy Elsasser <wendy.elsasser@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5727
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
This patch adds two regression tests that execute the script in the
configs dir for triggering low power mode transitions. A separate
test is required for each page policy because for close-adaptive
page policy the DRAM goes into the Precharge Power-down mode while
for open-adaptive page policy it goes into the Activate Power-down
mode.
Change-Id: Iad61af23f132db046f2857cc3ef64b2bf42cf5e4
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5726
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
This change adds a Python script to sweep a few parameters with a
goal to trigger transitions to the low power states of the DRAM
controller.
This script is largely based on the sweep.py but is helpful to study
the impact of inter-transaction delay on the behaviour of the DRAM
in addition to typical sweep params like stride size, bank
utilization and read percent. An idle period is added as the last
traffic generator state to target hitting self-refresh.
Change-Id: I34380afffbf2de9f4e997dfe9fff5e615e077524
Reviewed-by: Wendy Elsasser <wendy.elsasser@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5725
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
This patch syncs the DRAMPower library of gem5 to the
external github (https://github.com/ravenrd/DRAMPower).
The version pulled in is the commit:
90d6290f802c29b3de9e10233ceee22290907ce6
from 30th Oct. 2016.
This change also modifies the DRAM Ctrl interaction with the
DRAMPower, due to changes in the lib API in the above version.
Previously multiple functions were called to prepare the power
lib before calling the function that would calculate the enery. With
the new API, these functions are encompassed inside the function to
calculate the energy and therefore should now be removed from the
DRAM controller.
The other key difference is the introduction of a new function called
calcWindowEnergy which can be useful for any system that wants
to do measurements over intervals. For gem5 DRAM ctrl that means we
now need to accumulate the window energy measurements into the total
stat.
Change-Id: I3570fff2805962e166ff2a1a3217ebf2d5a197fb
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5724
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
This patch allows the user to specify more than one
power model for any given Clocked Object. This is
useful and some times necessary to properly model
components that have multiple power sources (or
subcomponents) but gem5 doesn't model them.
The ideal solution would be to have a DictParam
to replace the VectorParam so each model can have
a name and can be identified in the stats file.
Change-Id: I4080a7054a16b56069c44750a7a9ce4e674cdf9d
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5733
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
When ignoring writes to the Dummy ISA device (DummyISADevice),
additionally print the value being ignored in the diagnostic.
Sometimes it is useful to know exactly what we are dropping ...
Change-Id: I9a01623611f0da0aa12b065fbb2031aa27e2c036
Signed-off-by: Sean McGoogan <Sean.McGoogan@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5731
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
There are cases where it is desirable to load a kernel and a set of
additional objects. This can, for example, be useful for testing where
the bootstrap code can be loaded from one object (the kernel) and the
test proper from another.
This changeset adds this functionality by adding a kernel_extras
vector parameter to the System class. Object files in this vector are
loaded in order after the kernel when running in full system mode.
Change-Id: I06f57c6a65a17b02eb4267bed0aa829f21bcfa3b
Reviewed-on: https://gem5-review.googlesource.com/5703
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Jason Lowe-Power <jason@lowepower.com>
Writes to DCCMVAC (Data Cache line Clean by VA to PoC) system register
shouldn't flush the pipeline as a result of the operation. This addition
was wrongly introduced for supporting self-modifying code. Software
barriers should be used instead.
Change-Id: Idf0c27d2e49ca01be19888ae5523b8f8eaefa7b3
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5362
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
This Patch is removing the FlushPipe ArmFault, which was used for
flushing the pipeline in favour of the general IsSquashAfter StaticInstr
flag. Using a fault was preventing tracers from tracing barriers like
ISB and from adding them to the instruction count
Change-Id: I176e9254eca904694f2f611eb486c55e50ec61ff
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Giacomo Gabrielli <giacomo.gabrielli@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5361
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
This patch introduces the ARM A32/T32/A64 CRC Instructions, which are
mandatory since ARMv8.1. The UNPREDICTABLE behaviours are implemented as
follows:
1) CRC32(C)X (64 bit) instructions are decoded as Undefined in Aarch32
2) The instructions support predication in Aarch32
3) Using R15(PC) as source/dest operand is permitted in Aarch32
Change-Id: Iaf29b05874e1370c7615da79a07f111ded17b6cc
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Giacomo Gabrielli <giacomo.gabrielli@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5521
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Simplify the patch creation helper script by removing support for
automatic conversion to Mercurial's patch format. We no longer use
Mercurial as a primary repository, so this support is not likely to be
needed again.
Change-Id: I83f6efb9ccaee9c548b33257a4d6128153750c76
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5162
Reviewed-by: Gabe Black <gabeblack@google.com>
The helper script to list changes currently only looks at the changes
between a branch point and two heads. This helps performance, but
sometimes misclassifies changes that exist both in the upstream
branch before the branch point and in a feature branch. Such changes
should normally not exist, but can be the product of an incorrect
rebase.
This change adds an option to the helper script to search the entire
upstream repo history when determining the set of upstream
changes. When this option is provided, the script lists the changes
that exist upstream before the branch point and in the feature branch
as "Incorrectly rebased changes".
Change-Id: I4cb72cea2152c49d0317dc43613be94a0a2de1e5
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5161
These had a lot of code which duplicated what was already in the
Float param value class. Also, printing into the ini file with "%f"
forces python to truncate values which require more precision than the
fixed float format supports.
Change-Id: Iad9623b71a31d17b69c184082585dcbb881eaa20
Reviewed-on: https://gem5-review.googlesource.com/5622
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
ARMv8 Tracers might want to be able to read the intWidth field of the
ArmStaticInst object. The field is specifying the bit width of the
integer registers used by the current instruction.
Change-Id: Iaee3123823a2c7380917001c453377c1c12e54a7
Signed-off-by: Giacomo Travaglini <giacomo.travaglini@arm.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5661
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
The python conversion module was really repetitive and fragmented,
where some types of conversions use common code, and some use hand
written case statements which did something very similar. Also, some
types like Voltage could only handle V and mV but no other scaling
prefix.
This change restructures the module to centralize a lot of the unit
handling code into toFloat, and makes the various other functions use
it.
Change-Id: Ic8529203cc226c9b551b8535a444e3f2f25ad1eb
Reviewed-on: https://gem5-review.googlesource.com/5621
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Make site_init.py manage sys.path, and the "default" tool set
PYTHONPATH on any environment that's created. The paths to add are
tracked in a common gem5_python_paths.py.
Change-Id: I3387d4394d47a2f9c83322644cfd05909c6890fa
Reviewed-on: https://gem5-review.googlesource.com/5564
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
This imports various environment variables into the scons environment,
and sets some general properties on it. These are basically just copied
directly from the SConstruct and have the same behavior here.
gem5_tool_list will be used later on to add scons "tools" which should
be automatically added to new Environment objects.
Change-Id: Ib255955090c7b1e1cb80c703c18a9c867fcf1c9e
Reviewed-on: https://gem5-review.googlesource.com/5562
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
In an effort to shrink and modularize the main scons files, this
change pulls the scons and python version checking code out of the
main file and into site_init.py which runs before the SConstruct
starts. This will be a place to put really generic code which has to
do with the very fundemental aspects of getting scons to work. Other
checks, like checks for particular tools or particular versions of
tools, will happen in other more specialized files.
Change-Id: Icd00ecadbe1141aef4dbadcf42d6ddef1f3a701f
Reviewed-on: https://gem5-review.googlesource.com/5561
Reviewed-by: Gabe Black <gabeblack@google.com>
Maintainer: Gabe Black <gabeblack@google.com>
When the XBar receives a Writeback/WriteClean packet, it doesn't need
to snoop the upstream caches. It only queries the snoop filter and
sets the blockCached flag accordingly. This is in line with the
recvTimingReq.
Change-Id: I0ae22f21491d75a111019124bb95bac7b16d3cd3
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Anouk Van Laer <anouk.vanlaer@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5043
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Nikos Nikoleris <nikos.nikoleris@arm.com>
A program running in EL0 is allowed to execute CMOs when the UCI bit
in SCTLR is set. The execution of dc ivac, however, would fault
uncoditionally when executed from EL0. This change aligns the
permission checks for dc ivac with the rest of the CMOs.
Change-Id: I1a532f37707c7dc0748b4375252c6ec0bbf95419
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5058
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Add a dummy serial device that discards any output and doesn't provide
any input. This device can be used to terminate UARTs that don't have
a default device (e.g., a terminal) attached.
Change-Id: I4a6b0b5037ce360f59bfb5c566e1698d113a1d26
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/4290
Reviewed-by: Gabe Black <gabeblack@google.com>
The UART models currently assume that they are always wired to a
terminal. While true at the moment, this isn't necessarily a valid
assumption. This change introduces the SerialDevice class that defines
the interface for serial devices. Currently, Terminal is the only
class that implements this interface.
Change-Id: I74fefafbbaf5ac1ec0d4ec0b5a0f4b246fdad305
Signed-off-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Curtis Dunham <curtis.dunham@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/4289
Reviewed-by: Gabe Black <gabeblack@google.com>
In the ISA instruction definitions, some classes were declared with
execute, etc., functions outside of the main template because they
had CPU specific signatures and would need to be duplicated with
each CPU plugged into them. Now that the instructions always just
use an ExecContext, there's no reason for those templates to be
separate. This change folds those templates together.
Change-Id: I13bda247d3d1cc07c0ea06968e48aa5b4aace7fa
Reviewed-on: https://gem5-review.googlesource.com/5401
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-by: Alec Roelke <ar4jc@virginia.edu>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
Some ISAs (namely RISC-V) implement getrlimit and setrlimit using
prlimit. This patch adds an implementation for the prlimit system call.
Normally prlimit is supposed to provide the functionality of both
getrlimit and setrlimit, but because gem5 does not support setrlimit
this implementation of prlimit will simply display a warning and return
a failure code if the setrlimit mode is used. The same thing will happen
if a pid other than 0 is passed to it.
Change-Id: I653af2d5a60e716f4d6286196be7600409efcef8
Reviewed-on: https://gem5-review.googlesource.com/5321
Maintainer: Brandon Potter <Brandon.Potter@amd.com>
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
The ISA parser used to generate different copies of exec functions
for each exec context class a particular CPU wanted to use. That's
since been changed so that those functions take a pointer to the base
ExecContext, so the code which would generate those extra functions
can be removed, and some functions which used to be templated on an
ExecContext subclass can be untemplated, or minimally less templated.
Now that some functions aren't going to be instantiated multiple times
with different signatures, there are also opportunities to collapse
templates and make many instruction definitions simpler within the
parser. Since those changes will be less mechanical, they're left for
later changes and will probably be done in smaller increments.
Change-Id: I0015307bb02dfb9c60380b56d2a820f12169ebea
Reviewed-on: https://gem5-review.googlesource.com/5381
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
These functions were already being treated as psuedo objects and had
properties assigned to them setting what their paths were. That's a bit
unusual and made it less obvious what the code was doing, but also
forced the "system" function to know what all the possible path
searching functions were so that they'd have their "path" property
initialized properly in a central location.
This change introduces a PathSearcFunc class which encapsulates the
mechanisms of the old code and makes it implicitly extensible so that
other path searching functions which might look in other directories
can be added in other places.
Change-Id: I7be28e51481a06ec83997677af99927709b18003
Reviewed-on: https://gem5-review.googlesource.com/5341
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
The fixed image writer (which was dumping .bmp images only) has been
replaced by the configurable one in HDLcd device. Default format is
Auto, which gives gem5 the freedom to choose the format it prefers.
Change-Id: I0643266556bb10b43cdebd628f6daa2cd5e105dd
Reviewed-by: Andreas Sandberg <andreas.sandberg@arm.com>
Reviewed-on: https://gem5-review.googlesource.com/5183
Reviewed-by: Jason Lowe-Power <jason@lowepower.com>
Maintainer: Andreas Sandberg <andreas.sandberg@arm.com>
