Introduces COFFVCRuntimeBootstrapper that loads/initialize vc runtime libraries. In COFF, we *must* jit-link vc runtime libraries as COFF relocation types have no proper way to deal with out-of-reach data symbols ragardless of linking mode. (even dynamic version msvcrt.lib have tons of static data symbols that must be jit-linked) This class tries to load vc runtime library files from msvc installations with an option to override the path.
There are some complications when dealing with static version of vc runtimes. First, they need static initializers to be ran that requires COFFPlatform support but orc runtime will not be usable before vc runtimes are fully initialized. (as orc runtime will use msvc stl libraries) COFFPlatform that will be introduced in a following up patch will collect static initializers and run them manually in host before boostrapping itself. So, the user will have to do the following.
1. Create COFFPlatform that addes static initializer collecting passes.
2. LoadVCRuntime
3. InitializeVCRuntime
4. COFFPlatform.bootstrap()
Second, the internal crt initialization function had to be reimplemented in orc side. There are other ways of doing this, but this is the simplest implementation that makes platform fully responsible for static initializer. The complication comes from the fact that crt initialization functions (such as acrt_initialize or dllmain_crt_process_attach) actually run all static initializers by traversing from `__xi_a` symbol to `__xi_z`. This requires symbols to be contiguously allocated in sections alphabetically sorted in memory, which is not possible right now and not practical in jit setting. We might ignore emission of `__xi_a` and `__xi_z` symbol and allocate them ourselves, but we have to take extra care after orc runtime boostrap has been done -- as that point orc runtime should be the one running the static initializers.
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D130456
...with more fixes.
The original patch was reverted in 3e9cc543f2 due to bot failures caused by
a missing dependence on librt. That issue was fixed in 32d8d23cd0, but that
commit also broke sanitizer bots due to a bug in SimplePackedSerialization:
empty ArrayRef<char>s triggered a zero-byte memcpy from a null source. The
ArrayRef<char> serialization issue was fixed in 67220c2ad7, and this patch has
also been updated with a new custom SharedMemorySegFinalizeRequest message that
should avoid serializing empty ArrayRefs in the first place.
https://reviews.llvm.org/D128544
The original commit was reverted in 3e9cc543f2 due to buildbot failures, which
should be fixed by the addition of dependencies on librt.
Differential Revision: https://reviews.llvm.org/D128544
This is an implementation of orc::MemoryMapper that maps shared memory
pages in both executor and controller process and writes directly to
them avoiding transferring content over EPC. All allocations are properly
deinitialized automatically on the executor side at shutdown by the
ExecutorSharedMemoryMapperService.
Reviewed By: lhames
Differential Revision: https://reviews.llvm.org/D128544
(Reapply after revert in e9ce1a5880 due to
Fuchsia test failures. Removed changes in lib/ExecutionEngine/ other
than error categories, to be checked in more detail and reapplied
separately.)
Bulk remove many of the more trivial uses of ManagedStatic in the llvm
directory, either by defining a new getter function or, in many cases,
moving the static variable directly into the only function that uses it.
Differential Revision: https://reviews.llvm.org/D129120
Bulk remove many of the more trivial uses of ManagedStatic in the llvm
directory, either by defining a new getter function or, in many cases,
moving the static variable directly into the only function that uses it.
Differential Revision: https://reviews.llvm.org/D129120
runFinalizeActions takes an AllocActions vector and attempts to run its finalize
actions. If any finalize action fails then all paired dealloc actions up to the
failing pair are run, and the error(s) returned. If all finalize actions succeed
then a vector containing the dealloc actions is returned.
runDeallocActions takes a vector<WrapperFunctionCall> containing dealloc action
calls and runs them all, returning any error(s).
These helpers are intended to simplify the implementation of
JITLinkMemoryManager::InFlightAlloc::finalize and
JITLinkMemoryManager::deallocate overrides by taking care of execution (and
potential roll-back) of allocation actions.
In SimpleRemoteEPC, calls to from callWrapperAsync to sendMessage may fail.
The handlers may or may not be sent failure messages by handleDisconnect,
depending on when that method is run. This patch adds a check for an un-failed
handler, and if it finds one sends it a failure message.
This reverts commit dfd74db981.
SimpleRemoteEPC should share dispatch with the ExecutionSession, rather than
having two different dispatch systems on the controller side.
SimpleRemoteEPCServer::Dispatch doesn't need to be shared.
Renames SimpleRemoteEPCServer::Dispatcher to SimpleRemoteEPCDispatcher and
moves it into OrcShared. SimpleRemoteEPCServer::ThreadDispatcher is similarly
moved and renamed to DynamicThreadPoolSimpleRemoteEPCDispatcher.
This will allow these classes to be reused by SimpleRemoteEPC on the controller
side of the connection.
With the removal of OrcRPCExecutorProcessControl and OrcRPCTPCServer in
6aeed7b19c the ORC RPC library no longer has any in-tree users.
Clients needing serialization for ORC should move to Simple Packed
Serialization (usually by adopting SimpleRemoteEPC for remote JITing).
Adds a 'start' method to SimpleRemoteEPCTransport to defer transport startup
until the client has been configured. This avoids races on client members if the
first messages arrives while the client is being configured.
Also fixes races on the file descriptors in FDSimpleRemoteEPCTransport.
This reintroduces "[ORC] Introduce EPCGenericRTDyldMemoryManager."
(bef55a2b47) and "[lli] Add ChildTarget dependence
on OrcTargetProcess library." (7a219d801b) which were
reverted in 99951a5684 due to bot failures.
The root cause of the bot failures should be fixed by "[ORC] Fix uninitialized
variable." (0371049277) and "[ORC] Wait for
handleDisconnect to complete in SimpleRemoteEPC::disconnect."
(320832cc9b).
This reverts commit bef55a2b47 while I investigate
failures on some bots. Also reverts "[lli] Add ChildTarget dependence on
OrcTargetProcess library." (7a219d801b) which was
a fallow-up to bef55a2b47.
EPCGenericRTDyldMemoryMnaager is an EPC-based implementation of the
RuntimeDyld::MemoryManager interface. It enables remote-JITing via EPC (backed
by a SimpleExecutorMemoryManager instance on the executor side) for RuntimeDyld
clients.
The lli and lli-child-target tools are updated to use SimpleRemoteEPC and
SimpleRemoteEPCServer (rather than OrcRemoteTargetClient/Server), and
EPCGenericRTDyldMemoryManager for MCJIT tests.
By enabling remote-JITing for MCJIT and RuntimeDyld-based ORC clients,
EPCGenericRTDyldMemoryManager allows us to deprecate older remote-JITing
support, including OrcTargetClient/Server, OrcRPCExecutorProcessControl, and the
Orc RPC system itself. These will be removed in future patches.
Removing the 'ess' suffix improves the ergonomics without sacrificing clarity.
Since this class is likely to be used more frequently in the future it's worth
some short term pain to fix this now.
EPCGenericDylibManager provides an interface for loading dylibs and looking up
symbols in the executor, implemented using EPC-calls to functions in the
executor.
SimpleExecutorDylibManager is an executor-side service that provides the
functions used by EPCGenericDylibManager.
SimpleRemoteEPC is updated to use an EPCGenericDylibManager instance to
implement the ExecutorProcessControl loadDylib and lookup methods. In a future
commit these methods will be removed, and clients updated to use
EPCGenericDylibManagers directly.
Two typos, one unsused include and some leftovers from the TargetProcessControl -> ExecutorProcessControl renaming
Reviewed By: xgupta
Differential Revision: https://reviews.llvm.org/D110260
Finalization and deallocation actions are a key part of the upcoming
JITLinkMemoryManager redesign: They generalize the existing finalization and
deallocate concepts (basically "copy-and-mprotect", and "munmap") to include
support for arbitrary registration and deregistration of parts of JIT linked
code. This allows us to register and deregister eh-frames, TLV sections,
language metadata, etc. using regular memory management calls with no additional
IPC/RPC overhead, which should both improve JIT performance and simplify
interactions between ORC and the ORC runtime.
The SimpleExecutorMemoryManager class provides executor-side support for memory
management operations, including finalization and deallocation actions.
This support is being added in advance of the rest of the memory manager
redesign as it will simplify the introduction of an EPC based
RuntimeDyld::MemoryManager (since eh-frame registration/deregistration will be
expressible as actions). The new RuntimeDyld::MemoryManager will in turn allow
us to remove older remote allocators that are blocking the rest of the memory
manager changes.
This is a small first step towards reorganization of the ORC libraries:
Declarations for types and function names (as strings) to be found in the
"ORC runtime bootstrap" set are moved into OrcRTBridge.h / OrcRTBridge.cpp.
The current implementation of the "ORC runtime bootstrap" functions is moved
into OrcRTBootstrap.h and OrcRTBootstrap.cpp. It is likely that this code will
eventually be moved into ORT-RT proper (in compiler RT).
The immediate goal of this change is to make these bootstrap functions usable
for clients other than SimpleRemoteEPC/SimpleRemoteEPCServer. The first planned
client is a new RuntimeDyld::MemoryManager that will run over EPC, which will
allow us to remove the old OrcRemoteTarget code.
This reapplies bb27e45643 (SimpleRemoteEPC
support) and 2269a941a4 (#include <mutex>
fix) with further fixes to support building with LLVM_ENABLE_THREADS=Off.
This reverts commit 5629afea91 ("[ORC] Add missing
include."), and bb27e45643 ("[ORC] Add
SimpleRemoteEPC: ExecutorProcessControl over SPS + abstract transport.").
The SimpleRemoteEPC patch currently assumes availability of threads, and needs
to be rewritten with LLVM_ENABLE_THREADS guards.
SimpleRemoteEPC is an ExecutorProcessControl implementation (with corresponding
new server class) that uses ORC SimplePackedSerialization (SPS) to serialize and
deserialize EPC-messages to/from byte-buffers. The byte-buffers are sent and
received via a new SimpleRemoteEPCTransport interface that can be implemented to
run SimpleRemoteEPC over whatever underlying transport system (IPC, RPC, network
sockets, etc.) best suits your use case.
The SimpleRemoteEPCServer class provides executor-side support. It uses a
customizable SimpleRemoteEPCServer::Dispatcher object to dispatch wrapper
function calls to prevent the RPC thread from being blocked (a problem in some
earlier remote-JIT server implementations). Almost all functionality (beyond the
bare basics needed to bootstrap) is implemented as wrapper functions to keep the
implementation simple and uniform.
Compared to previous remote JIT utilities (OrcRemoteTarget*,
OrcRPCExecutorProcessControl), more consideration has been given to
disconnection and error handling behavior: Graceful disconnection is now always
initiated by the ORC side of the connection, and failure at either end (or in
the transport) will result in Errors being delivered to both ends to enable
controlled tear-down of the JIT and Executor (in the Executor's case this means
"as controlled as the JIT'd code allows").
The introduction of SimpleRemoteEPC will allow us to remove other remote-JIT
support from ORC (including the legacy OrcRemoteTarget* code used by lli, and
the OrcRPCExecutorProcessControl and OrcRPCEPCServer classes), and then remove
ORC RPC itself.
The llvm-jitlink and llvm-jitlink-executor tools have been updated to use
SimpleRemoteEPC over file descriptors. Future commits will move lli and other
tools and example code to this system, and remove ORC RPC.
Replace the existing WrapperFunctionResult type in
llvm/include/ExecutionEngine/Orc/Shared/TargetProcessControlTypes.h with a
version adapted from the ORC runtime's implementation.
Also introduce the SimplePackedSerialization scheme (also adapted from the ORC
runtime's implementation) for wrapper functions to avoid manual serialization
and deserialization for calls to runtime functions involving common types.
Moves all headers from Orc/RPC to Orc/Shared, and from the llvm::orc::rpc
namespace into llvm::orc::shared. Also renames RPCTypeName to
SerializationTypeName and Function to RPCFunction.
In addition to being a more reasonable home for this code, this will make it
easier for the upcoming Orc runtime to re-use the Serialization system for
creating and parsing wrapper-function binary blobs.
The LLVM_ENABLE_MODULES builds currently randomly fail due depending on the
headers generated by the intrinsics_gen target, but the current dependency only model
the non-modules dependencies:
```
While building module 'LLVM_ExecutionEngine' imported from llvm-project/llvm/lib/ExecutionEngine/Orc/Shared/TargetProcessControlTypes.cpp:13:
While building module 'LLVM_intrinsic_gen' imported from llvm-project/llvm/include/llvm/ExecutionEngine/Orc/ThreadSafeModule.h:17:
In file included from <module-includes>:1:
In file included from llvm-project/llvm/include/llvm/IR/Argument.h:18:
llvm/include/llvm/IR/Attributes.h:75:14: fatal error: 'llvm/IR/Attributes.inc' file not found
#include "llvm/IR/Attributes.inc"
^~~~~~~~~~~~~~~~~~~~~~~~
```
Depending on whether intrinsics_gen runs before compiling Orc/Shared files we either fail or include an outdated Attributes.inc
in module builds. The Clang modules require these additional dependencies as including/importing one module requires all
includes headers by that module to be parsable.
Differential Revision: https://reviews.llvm.org/D92873
No longer rely on an external tool to build the llvm component layout.
Instead, leverage the existing `add_llvm_componentlibrary` cmake function and
introduce `add_llvm_component_group` to accurately describe component behavior.
These function store extra properties in the created targets. These properties
are processed once all components are defined to resolve library dependencies
and produce the header expected by llvm-config.
Differential Revision: https://reviews.llvm.org/D90848
implementation.
This patch aims to improve support for out-of-process JITing using OrcV2. It
introduces two new class templates, OrcRPCTargetProcessControlBase and
OrcRPCTPCServer, which together implement the TargetProcessControl API by
forwarding operations to an execution process via an Orc-RPC Endpoint. These
utilities are used to implement out-of-process JITing from llvm-jitlink to
a new llvm-jitlink-executor tool.
This patch also breaks the OrcJIT library into three parts:
-- OrcTargetProcess: Contains code needed by the JIT execution process.
-- OrcShared: Contains code needed by the JIT execution and compiler
processes
-- OrcJIT: Everything else.
This break-up allows JIT executor processes to link against OrcTargetProcess
and OrcShared only, without having to link in all of OrcJIT. Clients executing
JIT'd code in-process should start linking against OrcTargetProcess as well as
OrcJIT.
In the near future these changes will enable:
-- Removal of the OrcRemoteTargetClient/OrcRemoteTargetServer class templates
which provided similar functionality in OrcV1.
-- Restoration of Chapter 5 of the Building-A-JIT tutorial series, which will
serve as a simple usage example for these APIs.
-- Implementation of lazy, cross-target compilation in lli's -jit-kind=orc-lazy
mode.
Sunho Kim