Moving FlowAfterFree and UseAfterFree.qll as a general purpose lib.

This commit is contained in:
Benjamin Rodes
2024-01-22 11:18:03 -05:00
parent 2181fcf284
commit da10e6ca5b
4 changed files with 12 additions and 4 deletions

View File

@@ -13,7 +13,7 @@
import cpp
import semmle.code.cpp.dataflow.new.DataFlow
import FlowAfterFree
import semmle.code.cpp.security.flowafterfree.FlowAfterFree
import DoubleFree::PathGraph
/**

View File

@@ -1,164 +0,0 @@
import cpp
import semmle.code.cpp.dataflow.new.DataFlow
private import semmle.code.cpp.ir.IR
/**
* Holds if `(b1, i1)` strictly post-dominates `(b2, i2)`
*/
bindingset[i1, i2]
predicate strictlyPostDominates(IRBlock b1, int i1, IRBlock b2, int i2) {
b1 = b2 and
i1 > i2
or
b1.strictlyPostDominates(b2)
}
/**
* Holds if `(b1, i1)` strictly dominates `(b2, i2)`
*/
bindingset[i1, i2]
predicate strictlyDominates(IRBlock b1, int i1, IRBlock b2, int i2) {
b1 = b2 and
i1 < i2
or
b1.strictlyDominates(b2)
}
signature module FlowFromFreeParamSig {
/**
* Signature for a predicate that holds if `n.asExpr() = e` and `n` is a sink in
* the `FlowFromFreeConfig` module.
*/
predicate isSink(DataFlow::Node n, Expr e);
/**
* Holds if `dealloc` is a deallocation expression and `e` is an expression such
* that `isFree(_, e)` holds for some `isFree` predicate satisfying `isSinkSig`,
* and this source-sink pair should be excluded from the analysis.
*/
bindingset[dealloc, e]
predicate isExcluded(DeallocationExpr dealloc, Expr e);
/**
* Holds if `sink` should be considered a `sink` when the source of flow is `source`.
*/
bindingset[source, sink]
default predicate sourceSinkIsRelated(DataFlow::Node source, DataFlow::Node sink) { any() }
}
/**
* Constructs a `FlowFromFreeConfig` module that can be used to find flow between
* a pointer being freed by some deallocation function, and a user-specified sink.
*
* In order to reduce false positives, the set of sinks is restricted to only those
* that satisfy at least one of the following two criteria:
* 1. The source dominates the sink, or
* 2. The sink post-dominates the source.
*/
module FlowFromFree<FlowFromFreeParamSig P> {
module FlowFromFreeConfig implements DataFlow::StateConfigSig {
class FlowState instanceof Expr {
FlowState() { isFree(_, _, this, _) }
string toString() { result = super.toString() }
}
predicate isSource(DataFlow::Node node, FlowState state) { isFree(node, _, state, _) }
pragma[inline]
predicate isSink(DataFlow::Node sink, FlowState state) {
exists(Expr e, DataFlow::Node source, DeallocationExpr dealloc |
P::isSink(sink, e) and
isFree(source, _, state, dealloc) and
e != state and
not P::isExcluded(dealloc, e) and
P::sourceSinkIsRelated(source, sink)
)
}
predicate isBarrierIn(DataFlow::Node n) {
n.asIndirectExpr() = any(AddressOfExpr aoe)
or
n.asIndirectExpr() = any(Call call).getAnArgument()
or
exists(Expr e |
n.asIndirectExpr() = e.(PointerDereferenceExpr).getOperand() or
n.asIndirectExpr() = e.(ArrayExpr).getArrayBase()
|
e = any(StoreInstruction store).getDestinationAddress().getUnconvertedResultExpression()
)
or
n.asExpr() instanceof ArrayExpr
}
}
import DataFlow::GlobalWithState<FlowFromFreeConfig>
}
/**
* Holds if `outgoing` is a dataflow node that represents the pointer passed to
* `dealloc` after the call returns (i.e., the post-update node associated with
* the argument to `dealloc`), and `incoming` is the corresponding argument
* node going into `dealloc` (i.e., the pre-update node of `outgoing`).
*/
predicate isFree(DataFlow::Node outgoing, DataFlow::Node incoming, Expr e, DeallocationExpr dealloc) {
exists(Expr conv |
e = conv.getUnconverted() and
conv = dealloc.getFreedExpr().getFullyConverted() and
incoming = outgoing.(DataFlow::PostUpdateNode).getPreUpdateNode() and
conv = incoming.asConvertedExpr()
) and
// Ignore realloc functions
not exists(dealloc.(FunctionCall).getTarget().(AllocationFunction).getReallocPtrArg())
}
/**
* Holds if `fc` is a function call that is the result of expanding
* the `ExFreePool` macro.
*/
predicate isExFreePoolCall(FunctionCall fc, Expr e) {
e = fc.getArgument(0) and
(
exists(MacroInvocation mi |
mi.getMacroName() = "ExFreePool" and
mi.getExpr() = fc
)
or
fc.getTarget().hasGlobalName("ExFreePool")
)
}
/**
* Holds if either `source` strictly dominates `sink`, or `sink` strictly
* post-dominates `source`.
*/
bindingset[source, sink]
predicate defaultSourceSinkIsRelated(DataFlow::Node source, DataFlow::Node sink) {
exists(IRBlock b1, int i1, IRBlock b2, int i2 |
source.hasIndexInBlock(b1, i1) and
sink.hasIndexInBlock(b2, i2)
|
strictlyDominates(b1, i1, b2, i2)
or
strictlyPostDominates(b2, i2, b1, i1)
)
}
/**
* `dealloc1` is a deallocation expression, `e` is an expression that dereferences a
* pointer, and the `(dealloc1, e)` pair should be excluded by the `FlowFromFree` library.
*
* Note that `e` is not necessarily the expression deallocated by `dealloc1`. It will
* be bound to the second deallocation as identified by the `FlowFromFree` library.
*
* From https://learn.microsoft.com/en-us/windows-hardware/drivers/ddi/wdm/nf-wdm-mmfreepagesfrommdl:
* "After calling MmFreePagesFromMdl, the caller must also call ExFreePool
* to release the memory that was allocated for the MDL structure."
*/
bindingset[dealloc1, e]
predicate isExcludedMmFreePageFromMdl(DeallocationExpr dealloc1, Expr e) {
exists(DeallocationExpr dealloc2 | isFree(_, _, e, dealloc2) |
dealloc1.(FunctionCall).getTarget().hasGlobalName("MmFreePagesFromMdl") and
isExFreePoolCall(dealloc2, _)
)
}

View File

@@ -14,8 +14,8 @@
import cpp
import semmle.code.cpp.dataflow.new.DataFlow
import semmle.code.cpp.ir.IR
import FlowAfterFree
import UseAfterFree
import semmle.code.cpp.security.flowafterfree.FlowAfterFree
import semmle.code.cpp.security.flowafterfree.UseAfterFree
import UseAfterFreeTrace::PathGraph
module UseAfterFreeParam implements FlowFromFreeParamSig {

View File

@@ -1,146 +0,0 @@
import cpp
private import FlowAfterFree
private import semmle.code.cpp.ir.IR
/**
* Holds if `call` is a call to a function that obviously
* doesn't dereference its `i`'th argument.
*/
private predicate externalCallNeverDereferences(FormattingFunctionCall call, int arg) {
exists(int formatArg |
pragma[only_bind_out](call.getFormatArgument(formatArg)) =
pragma[only_bind_out](call.getArgument(arg)) and
call.getFormat().(FormatLiteral).getConvSpec(formatArg) != "%s"
)
}
predicate isUse0(Expr e) {
not isFree(_, _, e, _) and
(
e = any(PointerDereferenceExpr pde).getOperand()
or
e = any(PointerFieldAccess pfa).getQualifier()
or
e = any(ArrayExpr ae).getArrayBase()
or
e = any(Call call).getQualifier()
or
// Assume any function without a body will dereference the pointer
exists(int i, Call call, Function f |
e = call.getArgument(i) and
f = call.getTarget() and
not f.hasEntryPoint() and
// Exclude known functions we know won't dereference the pointer.
// For example, a call such as `printf("%p", myPointer)`.
not externalCallNeverDereferences(call, i)
)
)
}
module ParameterSinks {
import semmle.code.cpp.ir.ValueNumbering
predicate flowsToUse(DataFlow::Node n) {
isUse0(n.asExpr())
or
exists(DataFlow::Node succ |
flowsToUse(succ) and
DataFlow::localFlowStep(n, succ)
)
}
private predicate flowsFromParam(DataFlow::Node n) {
flowsToUse(n) and
(
n.asParameter().getUnspecifiedType() instanceof PointerType
or
exists(DataFlow::Node prev |
flowsFromParam(prev) and
DataFlow::localFlowStep(prev, n)
)
)
}
private predicate step(DataFlow::Node n1, DataFlow::Node n2) {
flowsFromParam(n1) and
flowsFromParam(n2) and
DataFlow::localFlowStep(n1, n2)
}
private predicate paramToUse(DataFlow::Node n1, DataFlow::Node n2) = fastTC(step/2)(n1, n2)
private predicate hasFlow(
DataFlow::Node source, InitializeParameterInstruction init, DataFlow::Node sink
) {
pragma[only_bind_out](source.asParameter()) = pragma[only_bind_out](init.getParameter()) and
paramToUse(source, sink) and
isUse0(sink.asExpr())
}
private InitializeParameterInstruction getAnAlwaysDereferencedParameter0() {
exists(DataFlow::Node source, DataFlow::Node sink, IRBlock b1, int i1, IRBlock b2, int i2 |
hasFlow(pragma[only_bind_into](source), result, pragma[only_bind_into](sink)) and
source.hasIndexInBlock(b1, pragma[only_bind_into](i1)) and
sink.hasIndexInBlock(b2, pragma[only_bind_into](i2)) and
strictlyPostDominates(b2, i2, b1, i1)
)
}
private CallInstruction getAnAlwaysReachedCallInstruction() {
exists(IRFunction f | result.getBlock().postDominates(f.getEntryBlock()))
}
pragma[nomagic]
private predicate callHasTargetAndArgument(Function f, int i, Instruction argument) {
exists(CallInstruction call |
call.getStaticCallTarget() = f and
call.getArgument(i) = argument and
call = getAnAlwaysReachedCallInstruction()
)
}
pragma[nomagic]
private predicate initializeParameterInFunction(Function f, int i) {
exists(InitializeParameterInstruction init |
pragma[only_bind_out](init.getEnclosingFunction()) = f and
init.hasIndex(i) and
init = getAnAlwaysDereferencedParameter()
)
}
pragma[nomagic]
private predicate alwaysDereferencedArgumentHasValueNumber(ValueNumber vn) {
exists(int i, Function f, Instruction argument |
callHasTargetAndArgument(f, i, argument) and
initializeParameterInFunction(pragma[only_bind_into](f), pragma[only_bind_into](i)) and
vn.getAnInstruction() = argument
)
}
InitializeParameterInstruction getAnAlwaysDereferencedParameter() {
result = getAnAlwaysDereferencedParameter0()
or
exists(ValueNumber vn |
alwaysDereferencedArgumentHasValueNumber(vn) and
vn.getAnInstruction() = result
)
}
}
module IsUse {
private import semmle.code.cpp.ir.dataflow.internal.DataFlowImplCommon
predicate isUse(DataFlow::Node n, Expr e) {
isUse0(e) and n.asExpr() = e
or
exists(CallInstruction call, InitializeParameterInstruction init |
n.asOperand().getDef().getUnconvertedResultExpression() = e and
pragma[only_bind_into](init) = ParameterSinks::getAnAlwaysDereferencedParameter() and
viableParamArg(call, DataFlow::instructionNode(init), n) and
pragma[only_bind_out](init.getEnclosingFunction()) =
pragma[only_bind_out](call.getStaticCallTarget())
)
}
}
import IsUse