This change touches several subsystems, tools and experiments
(sal, util, cmake, import-trace, generic-tracing, nanojpeg), and
changes details not worth separate commits.
Change-Id: Icd1d664d1be5cfc2212dbf77801c271183214d08
This tool can now import extended trace information with the
--extended-trace command-line parameter. The existing importers cease
using artificial access_info_t objects in favor of passing through the
original Trace_Event wherever possible. This allows us to import
extended trace information for all importers.
Change-Id: I3613e9d05d5e69ad49e96f4dc5ba0b1c4ef95a11
This change became necessary as we observed weird fail-client SIGSEGV
crashes with both Bochs and Gem5 backends and different experiments.
Some Fail* components are instantiated statically: the
SimulatorController instance "simulator", containing the
ListenerManager and the CoroutineManager, and the active
ExperimentFlow subclass(es)
(experiments/instantiate-experiment*.ah.in). The experiment(s) is
registered as an active flow in the CoroutineManager at startup.
As plugins (which are ExperimentFlows themselves) are often created on
an experiment's stack, ExperimentFlows deregister themselves on
destruction (e.g., when leaving the plugin variable's scope). The
core problem is, that the creation and destruction order of statically
instantiated objects depends on the link order; if the experiment is
destroyed after the CoroutineManager, its automatic self-deregistering
feature talks to the smoking ruins of the latter.
This change removes all static instantiations of ExperimentFlow and
replaces them with constructions on the heap. Additionally it makes
sure that the CoroutineManager recognizes that a shutdown is in
progress, and refrains from touching potentially already destroyed
data structures when a (mistakenly globally instantiated)
ExperimentFlow deregisters in this case.
Change-Id: I8a7d42fb141222cd2cce6040ab1a01f9de61be24
This change adapts the gem5 backend to the Register class interface change
in commit 52723a8. The necessary modifications suggested adding the "misc"
registers from gem5, too.
Change-Id: I32561c3fc905b9cd396e32ce80c791c01d5682fb
The extended trace register list is needed in multiple locations; the CPU
class is the logical module to contain this information.
Increased number of x86 registers to be traced; we can remove those
that prove unusable for fault-space pruning later on.
Change-Id: Ic46ecdbc55167a6d92872c190317fc0d1a3ad92d
This allows a register to be a member of multiple UniformRegisterSets.
Needed to architecture-specifically store, e.g., a list of registers to
record in traces.
Change-Id: Ia5d20768540efe252a84df967d43e569f107da3a
The SConscript in src/core/sal/gem5 is now generated via CMake
(SConscript.in). No more hardcoded relative paths -> Fail* build
folder can now be anywhere. Experiment and Plugin libraries are now
set automagically (using ${EXPERIMENTS_ACTIVATED} /
${PLUGINS_ACTIVATED})
Generated SConscript now resides in binary dir.
Change-Id: I1bf2e17c83c95ffdcf6801c02481064fcb63bfb0
This commit introduces a virtual dtor in the ArmCPUState class to
prevent the compiler from producing the following warning:
"deleting object of polymorphic class type 'fail::Gem5ArmCPU' which
has non-virtual destructor might cause undefined behaviour"
Change-Id: I5029662064e72ae87a9f0e36aeaa309e4cb2291b
The build system now allows incremental gem5 builds. Unfortunately,
the current solution always requires re-linking the executable.
Without the enforcement of re-linking, the fail code will be rebuilt
but not linked into gem5.
The number of cores for building gem5 is derived from /proc/cpuinfo.
As before, only the gem5.debug configuration is supported.
Change-Id: Ib13b15d1ecd62196eb251e0fd00953f4eb052feb
The correct flags accessor methods should be used to write the flags register,
since values (especially the arithmetic flags) might be cached.
Change-Id: I338927cfdd1f111966be539668914c9256b54c8f
It is no longer required (and recommended) to write a dependency
check (using conditional compilation) in the experiment code. Instead,
the Fail code performs these checks itself. If an experiment is
compiled without having the required configuration flags set to
enabled, the compiler emits an error. This is because the constructors
of unused listener classes are declared as protected. (Note that
this works only because the Fail code does not create any listeners.)
In a next step, one should remove all the dependency guards in the
existing experiments (not needed anymore).
Change-Id: I1331813a432a752322d510b4102da53ff53294f0
Doxygen skips undesired directories and files now. In addition, the
documentation of the "fail" namespace has been fixed. Note that there
are still several warnings (due to incomplete documentations) in the
Doxygen output.
Change-Id: Idad4f1ecff453765b307fa40a5c1cebc0c2ce2bb
The checkpoint which is produced by this save method is a little bit
different to the checkpoint which is produced by the --take-checkpoint
command. It differs in the save-parameters so_state, funcExeInst, intRegs,
_upc, _nupc, _when. Tests have shown that it probably does not affect the
course of the program execution.
Change-Id: Id776a10f2d40f71643e9edbb45d7368609309df4
The checkpoint which is produced by this save method is a little bit
different to the checkpoint which is produced by the --take-checkpoint
command. It differs in the save-parameters so_state, funcExeInst, intRegs,
_upc, _nupc, _when. Tests have shown that it probably does not affect the
course of the program execution.
Change-Id: I19b3fc809288224532e0ed6b7910a45115cb1c5d
The previous implementation wasn't in a working state because
the register content retrieval was buggy. (For example, RT_FP
does *not* denote a "floating point" register. Instead, it is
the frame pointer!)
Change-Id: I31fd80d374c945adaf35b47958d6437a8e2d48c3
Now, the gem5 implementation equals the Bochs variant. Note that it's
*not* necessary to enable CONFIG_EVENTS_BREAKPOINTS_RANGE in order to
use range breakpoints.
In addition, gem5 distinguishes between macro- and microops. With the
new implementation, onBreakpoint() is only called when a macroop
changes.
Change-Id: Ib86d1802fc70c20d22ca1a1ece0e8d1221b2e7db
This adds an interface for a backend-specific notion of time, e.g. CPU
cycles since simulator start, and a concrete implementation for the
Bochs backend. This is needed to record CPU idle times (e.g., HLT
instruction), and for target backends capable of more timing-accurate
execution.
This change also modifies the tracing plugin to add the time to all
trace events.
Change-Id: I93ac1d54c07f32b0b8f84f333417741d8e9c8288
Encapsulated gem5-specific code into wrapper functions to separate the
build process (Fail: CMake, gem5: scons). Added some gem5-related FIXMEs.
Another CMake related FIXME added. +some cosmetics.
Change-Id: Id84b480127b1f13aed6a0ee97f3583f410d531c5
This allows the commandline parameter parser to modify argc, as it finds
arguments for the Fail* client. Additionally argv is correctly null
terminated when removing arguments.
This fixes a bug introduced in eb17e9ef82.
Change-Id: Iabe84530790ecb7c587b0af139127015aad868d5
After each simulator break, T32Tracer retrieves the latest (16)
trace records from the T32. Memory address and value can now
be evaluated easily from the trace record.
TODO:Nevertheless we still have to traverse the trace to
find the instruction causing the access.
If CommandLine.hpp and (indirectly) optionparser.h is #included in
FailBochsInit.ah, bochs compilation fails (for, e.g., gui/x.cc, at least
on Debian 6).
