| Age | Commit message (Collapse) | Author |
|---|
|
There is a test that thinks this should hold:
((&gInt >> 8) << 8) != ((&gInt << 8) >> 8)
For example, if the address is 0x00123400, this means:
0x00123400 != 0x00123400
which is obviously not true. Kill the bogus assumption as it causes
occasional failures in the tests. This is done by ORing the address with
1 so that we can have:
0x00123400 != 0x00123401
Convert also the respective truncated 32bit pointers to 64bit.
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
Core: explicitly create CallSite from Instruction
|
|
Added new option --warnings-only-to-file which causes warnings to be written to warnings.txt only.
|
|
Core: MCJIT functions need unique names
|
|
Module: simplify is_object checks
|
|
Moved printFileLine() to be part of KInstruction
|
|
|
|
Fix internal fork without new pathOS.id
|
|
Using klee_message instead of llvm:errs
|
|
|
|
|
|
Newer LLVMs do not allow implicit conversion from Instruction to
CallSite. We see this error:
Internal/Support/ModuleUtil.h:36:19: note: candidate function not viable: no known conversion from 'llvm::Instruction *' to 'llvm::CallSite' for 1st argument
llvm::Function *getDirectCallTarget(llvm::CallSite);
^
So explicitly create a CallSite from Instruction.
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
|
|
to warnings.txt only. Disabled by default.
|
|
convert iterators using static_cast
|
|
We will use newer MCJIT with newer LLVM versions. But it needs unique
names of functions or a wrong function can be called. So prepend
"dispatcher_" to function names (even for older LLVMs).
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
Newer versions of LLVM do not allow to implicitly cast iterators to
pointers where they point. So convert all such uses to explicit
static_cast, the same as LLVM code does.
Otherwise we see errors like:
lib/Core/Executor.cpp:548:15: error: no viable conversion from 'Module::iterator' (aka 'ilist_iterator<llvm::Function>') to 'llvm::Function *'
Function *f = i;
^ ~
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
object::Binary has isObject method, which can be used to check whether
it is an object::ObjectFile. Use that, since dyn_casting of
object::Binary is not allowed in newer LLVMs:
lib/Module/ModuleUtil.cpp:304:78: error: cannot convert ‘llvm::object::ObjectFile’ to ‘llvm::object::ObjectFile*’ in initialization
else if (object::ObjectFile *o = dyn_cast<object::ObjectFile>(child.get()))
^
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
Teach KLEE to respect the requested memory alignment of allocated memory
|
|
llvm: stop using global context
|
|
It was marked as deprecated long time ago and finally removed in LLVM
3.9. Remove all uses of getGlobalContext and create our own context.
Propagate it all over the code then.
[v2] use ctx, not C as name
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
variables when possible.
Previously an alignment 8 was always used which did not faithfully
emulate what was either explicitly requested in the LLVM IR or what
the default alignment was for the target.
|
|
CommandLine: do not copy list in optionIsSet
|
|
Pass the list as reference. Otherwise we can get errors with newer LLVM
like:
lib/Basic/ConstructSolverChain.cpp:26:19: error: call to deleted constructor of 'llvm::cl::list<QueryLoggingSolverType>'
if (optionIsSet(queryLoggingOptions, SOLVER_KQUERY)) {
^~~~~~~~~~~~~~~~~~~
llvm/Support/CommandLine.h:1466:3: note: 'list' has been explicitly marked deleted here
list(const list &) = delete;
^
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
klee: remove use of deprecated 'register'
|
|
As per:
http://llvm.org/releases/3.6.0/docs/ReleaseNotes.html#the-opt-option-std-compile-opts-was-removed
the -std-compile-opts was effectively an alias of -O3. And since it was
removed in later LLVM versions, stop relying on that.
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
New compilers warn about using 'register' as follows:
ConstantDivision.cpp:66:3: warning: 'register' storage class specifier is deprecated and incompatible with C++1z [-Wdeprecated-register]
Remove the register specifier -- the compilers are clever enough to know
what to do.
Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
|
|
delcypher/fix_klee_get_direct_call_bitcast_weak_alias
Fix assertion failure when klee::getDirectCallTarget() is used on call to bitcasted weak alias
|
|
|
|
controlled by a new parameter `moduleIsFullyLinked`. When
true the linkage type of a weak alias is ignored. It is legal to do
this when the module is fully linked because there won't be another
function that could override the weak alias.
This fixes a previous assertion failure in `klee::getDirectCallTarget()`
triggered by the `test/regression/2016-11-24-bitcast-weak-alias.c` test case.
|
|
`klee::getDirectCallTarget(llvm::CallSite)`.
The problem is that it doesn't handle bitcasted functions that call a
weak alias.
|
|
Silenced warning: comparison of integers of different signs ('const i…
|
|
'const unsigned long long')
|
|
too strict limitations (LLVM >= 3.0)
|
|
|
|
ReadExpr::create() was missing an opportunity to constant fold
|
|
|
|
width greater than 64
|
|
|
|
correctly applied
|
|
constant arrays.
|
|
Added error message when STP fails to fork.
|
|
|
|
Removing unused lib/SMT directory
|
|
|
|
Modify scripts and a test to allow ASan/UBSan builds.
|
|
compiler warnings, one by adding an assert, and the other by refactoring the choose() function.
|
|
Revert "Increased the type size for the stop-after-n-instructions option to a…"
|
|
a…"
|
|
Increased the type size for the stop-after-n-instructions option to a…
|
