If no --variant / --benchmark is specified, it's more reasonable to
prune or run *all* variants/benchmarks (using the wildcard "%")
instead of defaulting to "none"/"none". The trivial case with only
one single variant/benchmark (which may still be "none"/"none" if
import-trace's default is used) is still covered by this new default
behavior.
Change-Id: I0e9001137d5e052183dd74211e2edbcfab749528
The FESamplingPruner implements the fault-expansion variance reduction
technique (FE-VRT) as described in: Smith, D. Todd and Johnson, Barry W. and
Andrianos, Nikos and Profeta, III, Joseph A., "A variance-reduction technique
via fault-expansion for fault-coverage estimation" (1997), 366--374.
Change-Id: I04a0c9bb2622974278bd8c73793e51451119e650
This change moves prune-trace's --variants-exclude / --benchmarks-exclude
capabilities to Database::get_variants() to make it available to all users.
Change-Id: Icbc6bb1a3ae7c846d2de40b881f47a9cc1ed7bbf
Otherwise it's not possible to keep a "basic" and a "sampling" pruned
version of one variant in the same database.
Change-Id: Ic71eb27ea16df23e2289cbf9f96ae10209745791
This change makes the RegisterImporter continue importing if it
encounters an instruction pointer that is not part of the disassembled
ELF binary (and, thus, cannot be disassembled). This is OK if we
don't want to inject into registers used by these instructions.
Change-Id: Ia9b5e7f789367f8386d63f235451dae5d399610d
The fix for tools/convert-trace/Gem5Converter.cc repairs a real issue
which caused a runtime warning to never be displayed.
Change-Id: I5d94acdbffc99d5f2a78f047062c011c6f8ab4e3
This change extends import-trace's help functionality for importers,
adds the ?-operator to option "-i" and refactors main.cc a little.
The extensions are "doing something useful with the AliasedRegistry"
and are merely listings of all importers' prime aliases.
Change-Id: I7bb184fc45dd9f90664e37455edfccc704d99ef1
This change implements a generic registry in order to clean up import-trace's
code - it's possible (and reasonable) to use the registry for pruners as well.
Importer now extends AliasedRegisterable; all importers have been adapted
to suit the interface/abstract methods.
Each AliasedRegisterable should have at least one alias (the class' name
is a sensible choice) but can have several. The first specified alias is
the class' prime alias which can be used e.g. to list all registered objects.
Change-Id: If6daa34edce35a3b0194e4ba67ed3b44b74a49b0
This change limits fault injection to general-purpose registers, instead of
relying on the LLVM/Fail* bridge to only recognize the status register
(EFLAGS on x86) and general-purpose registers. Since this bridge just
learned to translate x86's control and segment registers, and these
registers need special handling for fault injection (def/use pruning does
not work here), only import register accesses from the RT_GP subset.
Status register and instruction pointer injection remain functional, and
import-trace now should work architecture independently.
Change-Id: Id8ad2f0a9dab1861bf16ea9443c3bdfe7213d3fa
Before this change, running prune-trace with, e.g.
"prune-trace -d fsp_mibench -v bitmap% --benchmark-exclude clockcnv"
resulted in an implied "--benchmark none", rendering --benchmark-exclude
ineffective and resulting in nothing being pruned. Now, the "none" default
only applies when neither --benchmark nor --benchmark-exclude (analogously
for --variant / --variant-exclude) is provided.
Change-Id: Ic7c88919d7cfde1261749a745dc6a679472ff348
Using Database::insert_multiple() instead of prepared statements
speeds up trace import by a factor of 3-4. While being there, we now
properly deal with nonexistent extended trace values (i.e., put NULLs
into the DB).
Side note: The ElfImporter should switch to insert_multiple(), too.
Change-Id: I96785e9775e3ef4f242fd50720d5c34adb4e88a1
Based on the database layout given by the pruner.
Run ./run.py -c <path to mysql.cnf>
(Default config ~/.my.cnf)
- Checks if objdump table exists
- Added view for results per instruction
- Added config file support for table details
- Overview data loaded at server startup
- Result type mapping configurable via config file
Based on Flask and MySQLdb
Change-Id: Ib49eac8f5c1e0ab23921aedb5bc53c34d0cde14d
Memory accesses that don't belong to the preceding IP event in the
trace *do* have a use case: a hardware interrupt causes the CPU to
push its state onto the (kernel) stack. At the moment we cannot
distinguish this case from a malformed trace (as we don't record the
occurrence of interrupts), hence this warning needs to be disabled for
now.
This reverts commit 84edd02b6f.
cluster runs.
If this output file is enabled, all running processes try to write to the
same file on the shared filesystem. They block each other which leads to
massive I/O wait time and CPU idle time.
This change reduces the runtime e.g. from several hours (12+) to few minutes
(20).
Change-Id: I028628af31c845fc517e5daca5b4f981eade3cf4
If the --debug option is set, the line number table of the elf binary will
be imported into the database. The information will be stored in the
"dbg_mapping" table.
If the --sources option is set, the source files will be imported
into the database. Only the files that were actually used in the
elf binary will be imported.
Change-Id: I0e9de6b456bc42b329c1700c25e5839d9552cdbb
As these tools work closely together with fail components, its
easiest, to build them in this context. As these tools don't
really matter for fail use, they might never be pushed to the
master branch.
Change-Id: I8c8bd80376d0475f08a531a995d829e85032371b
As non-gzipped trace-files cause the tool to always import zero events, the
input file is now openend as in the dump-trace tool, where opening non-gzipped
files obviously works fine.
Change-Id: If2575dbeb93ed657c7b8ddd9f14f41b5cc7bf7c6
In some cases the write-pilot is located at the upper boundary of the
experiment and thus is in a race situation with the experiment's end.
If the experiment's end occurs first, the campaign ends and complains
about missing data, otherwise everything is fine.
This patch circumvents this via using "the first" writing pilot; iff the
only write is located at the experiment's end, the race will still occur,
but cleverly written experiment code can, according to hsc, circumvent it.
Change-Id: I6a27a8c4770c04ea8dcaef8aa7bd85d18f43f0b5
The Fail* tools expect trace events to be ordered in a specific way:
memory-access events are supposed to come *after* the instruction
event for the instruction that caused them. Using a different order
may cause subtle problems with both fault-space pruning and fast
forwarding. This change introduces a warning message when such a
malformed trace is detected (i.e., when the instruction pointer of a
memory-access event does not match the preceding instruction event).
Change-Id: I8ae7420fd8ff26e2574590748bdcc5a63db76490
As non-gzipped trace files cause import-trace to always import zero
events, the input file is now openend as in the dump-trace tool, where
opening non-gzipped files obviously works fine.
In the medium term we should find a centralized solution for this,
instead of re-implementing it all over the place.
Change-Id: I75845c03c0bbdc2b6b578b83d492b7dbbb40f051
The ElfImporter is not a real trace importer, but we locate it
into the import-trace utility, since here the infrastructure is
already in place to import things related to an elf binary into
the database.
The ElfImporter calls objdump and dissassembles an elf binary
and imports the results into the database.
Change-Id: I6e35673c8dbee3b7e8dfc7549d10e5dca9b55935
Richard noticed that instr2 values are off by one when done with the
MemoryImporter vs. with his own importer. The core problem is that
the dynamic instruction counter in the Importer base class
(Importer::copy_to_database, instruction_count_t instr) gets increased
*after* reporting an IP event to the importer implementation; this has
the side-effect that memory access events have a +1 dynamic
instruction count offset with regard to the IP event of the
instruction they belong to.
Bottom line: IP events and all memory events belonging to that
instruction should have the same dynamic instruction number.
Christian argued for the numbers starting with 0, which, as a side
effect, relativizes the repercussions of the change introduced in the
previous commit, as the new "first" event gets the sequence number 0
now.
- All experiments and importers only dealing with memory accesses
(MemoryImporter) are affected by this change: The dynamic
instruction count now starts with 0 instead of 1. Together with
the previous commit, the only change is one additional dynamic
instruction at position 0. Note that existing trace files do not
have this additional instruction, which shifts all trace positions
by 1.
- All importers that process *only* IP events (InstructionImporter,
RandomJumpImporter, RegisterImporter) won't see any difference.
Commit 036e340, though, introduced a +1 offset.
- Experiments that use these instruction counts for navigating to
the target instruction must be checked to properly deal with the
dynamic instruction #0 (no forwarding necessary). All dynamic
instruction offsetting should now work uniformly for both memory
accesses and all other fault models. To be sure everything works
in order, sanity-check the current absolute instruction pointer
right before fault injection.
Change-Id: I3f509f1b47836fa78fd029a7bb7c36c878912d97
Without this change, import-trace won't recognize, e.g., the -e and -t
parameters if they come after a parameter that was added by the Importer:
import-trace -i objdump --objdump arm-none-eabi-objdump -e B.elf -t C.tc
[...]
[import-trace 14:37:32] couldn't open trace.pb
Change-Id: I9532b01e432055479c79d801b1ca2736a8fd21cc
The RegisterImporter splits each register into 1 byte chunks. The
--do-not-split flag prohibits this splitting. Be aware, that def/use
pruning won't work correctly in mixed-width cases (EAX/AX/AH/AL).
Change-Id: Ifa1930bdd9f317a6fd3ae50c4ff3cffc97504640
During the prune step the data_width of the injected location was not
propagated before. It is now stored in fsppilot (database layout change!) and
sent in the fsppilot protobuf message.
Change-Id: I0562f6fc8957adea0f8a9fb63469ca5e3f4b7b2d
There's one fspgroup entry for every trace entry, the pilot_id is
therefore *not* part of the (unique) primary key. If this had been
right in the first place, it would have revealed an equivalence-based
fault-space pruning bug early ... :-/
Change-Id: I449d4985645c6631c0a8db0c64510364677b1354
If you only need raw failure counts, no-effect write equivalence classes
are redundant and only slow down access to the trace and fspgroup tables.
This switch prevents any accesstype='W' entries from finding their way into
the trace table.
Change-Id: Ifb415994063a2107769bc80ebd2fd780de5a4dda
Initially this was implemented by directly passing through trace
events to the MemoryImporter, keeping a record of conditional jumps
and opcodes, and UPDATEing all inserted rows in a second pass when the
MemoryImporter is finished.
Unfortunately, UPDATE is very slow, and keeping all information in
memory till the end doesn't scale indefinitely. Therefore the
implementation now delays passing memory access events upwards to the
MemoryImporter only until enough branch history is aggregated, and
taps into Importer's database operations with a set of new virtual
functions that are called downwards.
Change-Id: I159b2533932087087fb3049f4ff07a5f17a25a00
A MemoryImporter that additionally imports Relyzer-style conditional
branch history, instruction opcodes, and a virtual
duration=time2-time1+1 column (MariaDB 5.2+ only) for fault-space
pruning purposes.
Change-Id: I6764a26fa8aae21655be44134b88fdee85e67ff6
