Files
archived-llvm/include/llvm/CodeGen/TargetPassConfig.h
Chandler Carruth fd5a8723ce Introduce the "retpoline" x86 mitigation technique for variant #2 of the speculative execution vulnerabilities disclosed today, specifically identified by CVE-2017-5715, "Branch Target Injection", and is one of the two halves to Spectre..
Summary:
First, we need to explain the core of the vulnerability. Note that this
is a very incomplete description, please see the Project Zero blog post
for details:
https://googleprojectzero.blogspot.com/2018/01/reading-privileged-memory-with-side.html

The basis for branch target injection is to direct speculative execution
of the processor to some "gadget" of executable code by poisoning the
prediction of indirect branches with the address of that gadget. The
gadget in turn contains an operation that provides a side channel for
reading data. Most commonly, this will look like a load of secret data
followed by a branch on the loaded value and then a load of some
predictable cache line. The attacker then uses timing of the processors
cache to determine which direction the branch took *in the speculative
execution*, and in turn what one bit of the loaded value was. Due to the
nature of these timing side channels and the branch predictor on Intel
processors, this allows an attacker to leak data only accessible to
a privileged domain (like the kernel) back into an unprivileged domain.

The goal is simple: avoid generating code which contains an indirect
branch that could have its prediction poisoned by an attacker. In many
cases, the compiler can simply use directed conditional branches and
a small search tree. LLVM already has support for lowering switches in
this way and the first step of this patch is to disable jump-table
lowering of switches and introduce a pass to rewrite explicit indirectbr
sequences into a switch over integers.

However, there is no fully general alternative to indirect calls. We
introduce a new construct we call a "retpoline" to implement indirect
calls in a non-speculatable way. It can be thought of loosely as
a trampoline for indirect calls which uses the RET instruction on x86.
Further, we arrange for a specific call->ret sequence which ensures the
processor predicts the return to go to a controlled, known location. The
retpoline then "smashes" the return address pushed onto the stack by the
call with the desired target of the original indirect call. The result
is a predicted return to the next instruction after a call (which can be
used to trap speculative execution within an infinite loop) and an
actual indirect branch to an arbitrary address.

On 64-bit x86 ABIs, this is especially easily done in the compiler by
using a guaranteed scratch register to pass the target into this device.
For 32-bit ABIs there isn't a guaranteed scratch register and so several
different retpoline variants are introduced to use a scratch register if
one is available in the calling convention and to otherwise use direct
stack push/pop sequences to pass the target address.

This "retpoline" mitigation is fully described in the following blog
post: https://support.google.com/faqs/answer/7625886

We also support a target feature that disables emission of the retpoline
thunk by the compiler to allow for custom thunks if users want them.
These are particularly useful in environments like kernels that
routinely do hot-patching on boot and want to hot-patch their thunk to
different code sequences. They can write this custom thunk and use
`-mretpoline-external-thunk` *in addition* to `-mretpoline`. In this
case, on x86-64 thu thunk names must be:
```
  __llvm_external_retpoline_r11
```
or on 32-bit:
```
  __llvm_external_retpoline_eax
  __llvm_external_retpoline_ecx
  __llvm_external_retpoline_edx
  __llvm_external_retpoline_push
```
And the target of the retpoline is passed in the named register, or in
the case of the `push` suffix on the top of the stack via a `pushl`
instruction.

There is one other important source of indirect branches in x86 ELF
binaries: the PLT. These patches also include support for LLD to
generate PLT entries that perform a retpoline-style indirection.

The only other indirect branches remaining that we are aware of are from
precompiled runtimes (such as crt0.o and similar). The ones we have
found are not really attackable, and so we have not focused on them
here, but eventually these runtimes should also be replicated for
retpoline-ed configurations for completeness.

For kernels or other freestanding or fully static executables, the
compiler switch `-mretpoline` is sufficient to fully mitigate this
particular attack. For dynamic executables, you must compile *all*
libraries with `-mretpoline` and additionally link the dynamic
executable and all shared libraries with LLD and pass `-z retpolineplt`
(or use similar functionality from some other linker). We strongly
recommend also using `-z now` as non-lazy binding allows the
retpoline-mitigated PLT to be substantially smaller.

When manually apply similar transformations to `-mretpoline` to the
Linux kernel we observed very small performance hits to applications
running typical workloads, and relatively minor hits (approximately 2%)
even for extremely syscall-heavy applications. This is largely due to
the small number of indirect branches that occur in performance
sensitive paths of the kernel.

When using these patches on statically linked applications, especially
C++ applications, you should expect to see a much more dramatic
performance hit. For microbenchmarks that are switch, indirect-, or
virtual-call heavy we have seen overheads ranging from 10% to 50%.

However, real-world workloads exhibit substantially lower performance
impact. Notably, techniques such as PGO and ThinLTO dramatically reduce
the impact of hot indirect calls (by speculatively promoting them to
direct calls) and allow optimized search trees to be used to lower
switches. If you need to deploy these techniques in C++ applications, we
*strongly* recommend that you ensure all hot call targets are statically
linked (avoiding PLT indirection) and use both PGO and ThinLTO. Well
tuned servers using all of these techniques saw 5% - 10% overhead from
the use of retpoline.

We will add detailed documentation covering these components in
subsequent patches, but wanted to make the core functionality available
as soon as possible. Happy for more code review, but we'd really like to
get these patches landed and backported ASAP for obvious reasons. We're
planning to backport this to both 6.0 and 5.0 release streams and get
a 5.0 release with just this cherry picked ASAP for distros and vendors.

This patch is the work of a number of people over the past month: Eric, Reid,
Rui, and myself. I'm mailing it out as a single commit due to the time
sensitive nature of landing this and the need to backport it. Huge thanks to
everyone who helped out here, and everyone at Intel who helped out in
discussions about how to craft this. Also, credit goes to Paul Turner (at
Google, but not an LLVM contributor) for much of the underlying retpoline
design.

Reviewers: echristo, rnk, ruiu, craig.topper, DavidKreitzer

Subscribers: sanjoy, emaste, mcrosier, mgorny, mehdi_amini, hiraditya, llvm-commits

Differential Revision: https://reviews.llvm.org/D41723

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@323155 91177308-0d34-0410-b5e6-96231b3b80d8
2018-01-22 22:05:25 +00:00

434 lines
16 KiB
C++

//===- TargetPassConfig.h - Code Generation pass options --------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
/// Target-Independent Code Generator Pass Configuration Options pass.
//
//===----------------------------------------------------------------------===//
#ifndef LLVM_CODEGEN_TARGETPASSCONFIG_H
#define LLVM_CODEGEN_TARGETPASSCONFIG_H
#include "llvm/Pass.h"
#include "llvm/Support/CodeGen.h"
#include <cassert>
#include <string>
namespace llvm {
class LLVMTargetMachine;
struct MachineSchedContext;
class PassConfigImpl;
class ScheduleDAGInstrs;
// The old pass manager infrastructure is hidden in a legacy namespace now.
namespace legacy {
class PassManagerBase;
} // end namespace legacy
using legacy::PassManagerBase;
/// Discriminated union of Pass ID types.
///
/// The PassConfig API prefers dealing with IDs because they are safer and more
/// efficient. IDs decouple configuration from instantiation. This way, when a
/// pass is overriden, it isn't unnecessarily instantiated. It is also unsafe to
/// refer to a Pass pointer after adding it to a pass manager, which deletes
/// redundant pass instances.
///
/// However, it is convient to directly instantiate target passes with
/// non-default ctors. These often don't have a registered PassInfo. Rather than
/// force all target passes to implement the pass registry boilerplate, allow
/// the PassConfig API to handle either type.
///
/// AnalysisID is sadly char*, so PointerIntPair won't work.
class IdentifyingPassPtr {
union {
AnalysisID ID;
Pass *P;
};
bool IsInstance = false;
public:
IdentifyingPassPtr() : P(nullptr) {}
IdentifyingPassPtr(AnalysisID IDPtr) : ID(IDPtr) {}
IdentifyingPassPtr(Pass *InstancePtr) : P(InstancePtr), IsInstance(true) {}
bool isValid() const { return P; }
bool isInstance() const { return IsInstance; }
AnalysisID getID() const {
assert(!IsInstance && "Not a Pass ID");
return ID;
}
Pass *getInstance() const {
assert(IsInstance && "Not a Pass Instance");
return P;
}
};
template <> struct isPodLike<IdentifyingPassPtr> {
static const bool value = true;
};
/// Target-Independent Code Generator Pass Configuration Options.
///
/// This is an ImmutablePass solely for the purpose of exposing CodeGen options
/// to the internals of other CodeGen passes.
class TargetPassConfig : public ImmutablePass {
private:
PassManagerBase *PM = nullptr;
AnalysisID StartBefore = nullptr;
AnalysisID StartAfter = nullptr;
AnalysisID StopBefore = nullptr;
AnalysisID StopAfter = nullptr;
bool Started = true;
bool Stopped = false;
bool AddingMachinePasses = false;
/// Set the StartAfter, StartBefore and StopAfter passes to allow running only
/// a portion of the normal code-gen pass sequence.
///
/// If the StartAfter and StartBefore pass ID is zero, then compilation will
/// begin at the normal point; otherwise, clear the Started flag to indicate
/// that passes should not be added until the starting pass is seen. If the
/// Stop pass ID is zero, then compilation will continue to the end.
///
/// This function expects that at least one of the StartAfter or the
/// StartBefore pass IDs is null.
void setStartStopPasses();
protected:
LLVMTargetMachine *TM;
PassConfigImpl *Impl = nullptr; // Internal data structures
bool Initialized = false; // Flagged after all passes are configured.
// Target Pass Options
// Targets provide a default setting, user flags override.
bool DisableVerify = false;
/// Default setting for -enable-tail-merge on this target.
bool EnableTailMerge = true;
/// Require processing of functions such that callees are generated before
/// callers.
bool RequireCodeGenSCCOrder = false;
/// Add the actual instruction selection passes. This does not include
/// preparation passes on IR.
bool addCoreISelPasses();
public:
TargetPassConfig(LLVMTargetMachine &TM, PassManagerBase &pm);
// Dummy constructor.
TargetPassConfig();
~TargetPassConfig() override;
static char ID;
/// Get the right type of TargetMachine for this target.
template<typename TMC> TMC &getTM() const {
return *static_cast<TMC*>(TM);
}
//
void setInitialized() { Initialized = true; }
CodeGenOpt::Level getOptLevel() const;
/// Describe the status of the codegen
/// pipeline set by this target pass config.
/// Having a limited codegen pipeline means that options
/// have been used to restrict what codegen is doing.
/// In particular, that means that codegen won't emit
/// assembly code.
bool hasLimitedCodeGenPipeline() const;
/// If hasLimitedCodeGenPipeline is true, this method
/// returns a string with the name of the options, separated
/// by \p Separator that caused this pipeline to be limited.
std::string
getLimitedCodeGenPipelineReason(const char *Separator = "/") const;
/// Check if the codegen pipeline is limited in such a way that it
/// won't be complete. When the codegen pipeline is not complete,
/// this means it may not be possible to generate assembly from it.
bool willCompleteCodeGenPipeline() const {
return !hasLimitedCodeGenPipeline() || (!StopAfter && !StopBefore);
}
void setDisableVerify(bool Disable) { setOpt(DisableVerify, Disable); }
bool getEnableTailMerge() const { return EnableTailMerge; }
void setEnableTailMerge(bool Enable) { setOpt(EnableTailMerge, Enable); }
bool requiresCodeGenSCCOrder() const { return RequireCodeGenSCCOrder; }
void setRequiresCodeGenSCCOrder(bool Enable = true) {
setOpt(RequireCodeGenSCCOrder, Enable);
}
/// Allow the target to override a specific pass without overriding the pass
/// pipeline. When passes are added to the standard pipeline at the
/// point where StandardID is expected, add TargetID in its place.
void substitutePass(AnalysisID StandardID, IdentifyingPassPtr TargetID);
/// Insert InsertedPassID pass after TargetPassID pass.
void insertPass(AnalysisID TargetPassID, IdentifyingPassPtr InsertedPassID,
bool VerifyAfter = true, bool PrintAfter = true);
/// Allow the target to enable a specific standard pass by default.
void enablePass(AnalysisID PassID) { substitutePass(PassID, PassID); }
/// Allow the target to disable a specific standard pass by default.
void disablePass(AnalysisID PassID) {
substitutePass(PassID, IdentifyingPassPtr());
}
/// Return the pass substituted for StandardID by the target.
/// If no substitution exists, return StandardID.
IdentifyingPassPtr getPassSubstitution(AnalysisID StandardID) const;
/// Return true if the pass has been substituted by the target or
/// overridden on the command line.
bool isPassSubstitutedOrOverridden(AnalysisID ID) const;
/// Return true if the optimized regalloc pipeline is enabled.
bool getOptimizeRegAlloc() const;
/// Return true if the default global register allocator is in use and
/// has not be overriden on the command line with '-regalloc=...'
bool usingDefaultRegAlloc() const;
/// High level function that adds all passes necessary to go from llvm IR
/// representation to the MI representation.
/// Adds IR based lowering and target specific optimization passes and finally
/// the core instruction selection passes.
/// \returns true if an error occured, false otherwise.
bool addISelPasses();
/// Add common target configurable passes that perform LLVM IR to IR
/// transforms following machine independent optimization.
virtual void addIRPasses();
/// Add passes to lower exception handling for the code generator.
void addPassesToHandleExceptions();
/// Add pass to prepare the LLVM IR for code generation. This should be done
/// before exception handling preparation passes.
virtual void addCodeGenPrepare();
/// Add common passes that perform LLVM IR to IR transforms in preparation for
/// instruction selection.
virtual void addISelPrepare();
/// addInstSelector - This method should install an instruction selector pass,
/// which converts from LLVM code to machine instructions.
virtual bool addInstSelector() {
return true;
}
/// This method should install an IR translator pass, which converts from
/// LLVM code to machine instructions with possibly generic opcodes.
virtual bool addIRTranslator() { return true; }
/// This method may be implemented by targets that want to run passes
/// immediately before legalization.
virtual void addPreLegalizeMachineIR() {}
/// This method should install a legalize pass, which converts the instruction
/// sequence into one that can be selected by the target.
virtual bool addLegalizeMachineIR() { return true; }
/// This method may be implemented by targets that want to run passes
/// immediately before the register bank selection.
virtual void addPreRegBankSelect() {}
/// This method should install a register bank selector pass, which
/// assigns register banks to virtual registers without a register
/// class or register banks.
virtual bool addRegBankSelect() { return true; }
/// This method may be implemented by targets that want to run passes
/// immediately before the (global) instruction selection.
virtual void addPreGlobalInstructionSelect() {}
/// This method should install a (global) instruction selector pass, which
/// converts possibly generic instructions to fully target-specific
/// instructions, thereby constraining all generic virtual registers to
/// register classes.
virtual bool addGlobalInstructionSelect() { return true; }
/// Add the complete, standard set of LLVM CodeGen passes.
/// Fully developed targets will not generally override this.
virtual void addMachinePasses();
/// Create an instance of ScheduleDAGInstrs to be run within the standard
/// MachineScheduler pass for this function and target at the current
/// optimization level.
///
/// This can also be used to plug a new MachineSchedStrategy into an instance
/// of the standard ScheduleDAGMI:
/// return new ScheduleDAGMI(C, make_unique<MyStrategy>(C), /*RemoveKillFlags=*/false)
///
/// Return NULL to select the default (generic) machine scheduler.
virtual ScheduleDAGInstrs *
createMachineScheduler(MachineSchedContext *C) const {
return nullptr;
}
/// Similar to createMachineScheduler but used when postRA machine scheduling
/// is enabled.
virtual ScheduleDAGInstrs *
createPostMachineScheduler(MachineSchedContext *C) const {
return nullptr;
}
/// printAndVerify - Add a pass to dump then verify the machine function, if
/// those steps are enabled.
void printAndVerify(const std::string &Banner);
/// Add a pass to print the machine function if printing is enabled.
void addPrintPass(const std::string &Banner);
/// Add a pass to perform basic verification of the machine function if
/// verification is enabled.
void addVerifyPass(const std::string &Banner);
/// Check whether or not GlobalISel should abort on error.
/// When this is disabled, GlobalISel will fall back on SDISel instead of
/// erroring out.
bool isGlobalISelAbortEnabled() const;
/// Check whether or not a diagnostic should be emitted when GlobalISel
/// uses the fallback path. In other words, it will emit a diagnostic
/// when GlobalISel failed and isGlobalISelAbortEnabled is false.
virtual bool reportDiagnosticWhenGlobalISelFallback() const;
protected:
// Helper to verify the analysis is really immutable.
void setOpt(bool &Opt, bool Val);
/// Methods with trivial inline returns are convenient points in the common
/// codegen pass pipeline where targets may insert passes. Methods with
/// out-of-line standard implementations are major CodeGen stages called by
/// addMachinePasses. Some targets may override major stages when inserting
/// passes is insufficient, but maintaining overriden stages is more work.
///
/// addPreISelPasses - This method should add any "last minute" LLVM->LLVM
/// passes (which are run just before instruction selector).
virtual bool addPreISel() {
return true;
}
/// addMachineSSAOptimization - Add standard passes that optimize machine
/// instructions in SSA form.
virtual void addMachineSSAOptimization();
/// Add passes that optimize instruction level parallelism for out-of-order
/// targets. These passes are run while the machine code is still in SSA
/// form, so they can use MachineTraceMetrics to control their heuristics.
///
/// All passes added here should preserve the MachineDominatorTree,
/// MachineLoopInfo, and MachineTraceMetrics analyses.
virtual bool addILPOpts() {
return false;
}
/// This method may be implemented by targets that want to run passes
/// immediately before register allocation.
virtual void addPreRegAlloc() { }
/// createTargetRegisterAllocator - Create the register allocator pass for
/// this target at the current optimization level.
virtual FunctionPass *createTargetRegisterAllocator(bool Optimized);
/// addFastRegAlloc - Add the minimum set of target-independent passes that
/// are required for fast register allocation.
virtual void addFastRegAlloc(FunctionPass *RegAllocPass);
/// addOptimizedRegAlloc - Add passes related to register allocation.
/// LLVMTargetMachine provides standard regalloc passes for most targets.
virtual void addOptimizedRegAlloc(FunctionPass *RegAllocPass);
/// addPreRewrite - Add passes to the optimized register allocation pipeline
/// after register allocation is complete, but before virtual registers are
/// rewritten to physical registers.
///
/// These passes must preserve VirtRegMap and LiveIntervals, and when running
/// after RABasic or RAGreedy, they should take advantage of LiveRegMatrix.
/// When these passes run, VirtRegMap contains legal physreg assignments for
/// all virtual registers.
virtual bool addPreRewrite() {
return false;
}
/// This method may be implemented by targets that want to run passes after
/// register allocation pass pipeline but before prolog-epilog insertion.
virtual void addPostRegAlloc() { }
/// Add passes that optimize machine instructions after register allocation.
virtual void addMachineLateOptimization();
/// This method may be implemented by targets that want to run passes after
/// prolog-epilog insertion and before the second instruction scheduling pass.
virtual void addPreSched2() { }
/// addGCPasses - Add late codegen passes that analyze code for garbage
/// collection. This should return true if GC info should be printed after
/// these passes.
virtual bool addGCPasses();
/// Add standard basic block placement passes.
virtual void addBlockPlacement();
/// This pass may be implemented by targets that want to run passes
/// immediately before machine code is emitted.
virtual void addPreEmitPass() { }
/// Targets may add passes immediately before machine code is emitted in this
/// callback. This is called even later than `addPreEmitPass`.
// FIXME: Rename `addPreEmitPass` to something more sensible given its actual
// position and remove the `2` suffix here as this callback is what
// `addPreEmitPass` *should* be but in reality isn't.
virtual void addPreEmitPass2() {}
/// Utilities for targets to add passes to the pass manager.
///
/// Add a CodeGen pass at this point in the pipeline after checking overrides.
/// Return the pass that was added, or zero if no pass was added.
/// @p printAfter if true and adding a machine function pass add an extra
/// machine printer pass afterwards
/// @p verifyAfter if true and adding a machine function pass add an extra
/// machine verification pass afterwards.
AnalysisID addPass(AnalysisID PassID, bool verifyAfter = true,
bool printAfter = true);
/// Add a pass to the PassManager if that pass is supposed to be run, as
/// determined by the StartAfter and StopAfter options. Takes ownership of the
/// pass.
/// @p printAfter if true and adding a machine function pass add an extra
/// machine printer pass afterwards
/// @p verifyAfter if true and adding a machine function pass add an extra
/// machine verification pass afterwards.
void addPass(Pass *P, bool verifyAfter = true, bool printAfter = true);
/// addMachinePasses helper to create the target-selected or overriden
/// regalloc pass.
FunctionPass *createRegAllocPass(bool Optimized);
};
} // end namespace llvm
#endif // LLVM_CODEGEN_TARGETPASSCONFIG_H