Commit Graph

62 Commits

Author SHA1 Message Date
Lang Hames 018452e6bc [MCJIT] Fix an iterator invalidation bug in MCJIT::finalizeObject.
The finalizeObject method calls generateCodeForModule on each of the currently
'added' objects, but generateCodeForModule moves objects out of the 'added'
set as it's called. To avoid iterator invalidation issues, the added set is
copied out before any calls to generateCodeForModule.

This should fix http://llvm.org/PR20851 .

llvm-svn: 217291
2014-09-05 23:38:35 +00:00
David Blaikie d1101574f7 unique_ptrify MCJIT::emitObject
llvm-svn: 217067
2014-09-03 19:57:35 +00:00
David Blaikie ed9709d928 unique_ptrify a bunch of stuff through RuntimeDyld::loadObject
llvm-svn: 217065
2014-09-03 19:48:09 +00:00
David Blaikie 196e323c01 unique_ptrify passing the TargetMachine to ExecutionEngine::MCJITCtor
llvm-svn: 216988
2014-09-02 22:41:07 +00:00
Eric Christopher 79cc1e3ae7 Reinstate "Nuke the old JIT."
Approved by Jim Grosbach, Lang Hames, Rafael Espindola.

This reinstates commits r215111, 215115, 215116, 215117, 215136.

llvm-svn: 216982
2014-09-02 22:28:02 +00:00
Rafael Espindola 7271c19420 Give ExecutionEngine of top level buffers.
Long term the idea if for the engine to not own the buffers, but for now
this is consistent with the rest of the API.

llvm-svn: 216484
2014-08-26 21:04:04 +00:00
Rafael Espindola 48af1c2a1a Don't own the buffer in object::Binary.
Owning the buffer is somewhat inflexible. Some Binaries have sub Binaries
(like Archive) and we had to create dummy buffers just to handle that. It is
also a bad fit for IRObjectFile where the Module wants to own the buffer too.

Keeping this ownership would make supporting IR inside native objects
particularly painful.

This patch focuses in lib/Object. If something elsewhere used to own an Binary,
now it also owns a MemoryBuffer.

This patch introduces a few new types.

* MemoryBufferRef. This is just a pair of StringRefs for the data and name.
  This is to MemoryBuffer as StringRef is to std::string.
* OwningBinary. A combination of Binary and a MemoryBuffer. This is needed
  for convenience functions that take a filename and return both the
  buffer and the Binary using that buffer.

The C api now uses OwningBinary to avoid any change in semantics. I will start
a new thread to see if we want to change it and how.

llvm-svn: 216002
2014-08-19 18:44:46 +00:00
Rafael Espindola 2a8a2795ee Make it explicit that ExecutionEngine takes ownership of the modules.
llvm-svn: 215967
2014-08-19 04:04:25 +00:00
Benjamin Kramer a7c40ef022 Canonicalize header guards into a common format.
Add header guards to files that were missing guards. Remove #endif comments
as they don't seem common in LLVM (we can easily add them back if we decide
they're useful)

Changes made by clang-tidy with minor tweaks.

llvm-svn: 215558
2014-08-13 16:26:38 +00:00
Eric Christopher b9fd9ed37e Temporarily Revert "Nuke the old JIT." as it's not quite ready to
be deleted. This will be reapplied as soon as possible and before
the 3.6 branch date at any rate.

Approved by Jim Grosbach, Lang Hames, Rafael Espindola.

This reverts commits r215111, 215115, 215116, 215117, 215136.

llvm-svn: 215154
2014-08-07 22:02:54 +00:00
Rafael Espindola f8b27c41e8 Nuke the old JIT.
I am sure we will be finding bits and pieces of dead code for years to
come, but this is a good start.

Thanks to Lang Hames for making MCJIT a good replacement!

llvm-svn: 215111
2014-08-07 14:21:18 +00:00
Rafael Espindola ce47a05c7c Replace comment about ownership with std::unique_ptr.
llvm-svn: 214533
2014-08-01 18:09:32 +00:00
Rafael Espindola b9a23cdcdb Delete dead code.
llvm-svn: 214370
2014-07-31 01:14:09 +00:00
David Blaikie 7a1e775a7e PR19553: Memory leak in RuntimeDyldELF::createObjectImageFromFile
This starts in MCJIT::getSymbolAddress where the
unique_ptr<object::Binary> is release()d and (after a cast) passed to a
single caller, MCJIT::addObjectFile.

addObjectFile calls RuntimeDyld::loadObject.
RuntimeDld::loadObject calls RuntimeDyldELF::createObjectFromFile

And the pointer is never owned at this point. I say this point, because
the alternative codepath, RuntimeDyldMachO::createObjectFile certainly
does take ownership, so this seemed like a good hint that this was a/the
right place to take ownership.

llvm-svn: 207580
2014-04-29 21:52:46 +00:00
Craig Topper e73658ddbb [C++] Use 'nullptr'.
llvm-svn: 207394
2014-04-28 04:05:08 +00:00
Lang Hames 868d4b3122 Add an option to MCJIT to have it forward all sections to the
RTDyldMemoryManager, regardless of whether it thinks they're "required for
execution".

Currently, RuntimeDyld only passes sections that are "required for execution"
to the RTDyldMemoryManager, and takes "required for execution" to mean exactly
"contains symbols or relocations". There are two problems with this:
(1) It can drop sections with anonymous data that is referenced by code.
(2) It leaves the JIT client no way to inspect interesting sections that aren't
    actually required to run the program (e.g dwarf sections).

A test case is still in the works.

Future work: We may want to replace this with a generic section filtering
mechanism, but that will require more consideration. For now, this flag at least
allows clients to volunteer to do the filtering themselves.

Fixes <rdar://problem/15177691>.

llvm-svn: 204398
2014-03-20 21:06:46 +00:00
Craig Topper b51ff603ea [C++11] Add 'override' keyword to virtual methods that override their base class.
llvm-svn: 203344
2014-03-08 07:51:20 +00:00
Ahmed Charles 56440fd820 Replace OwningPtr<T> with std::unique_ptr<T>.
This compiles with no changes to clang/lld/lldb with MSVC and includes
overloads to various functions which are used by those projects and llvm
which have OwningPtr's as parameters. This should allow out of tree
projects some time to move. There are also no changes to libs/Target,
which should help out of tree targets have time to move, if necessary.

llvm-svn: 203083
2014-03-06 05:51:42 +00:00
Lang Hames 937ec54951 Extend RTDyld API to enable optionally precomputing the total amount of memory
required for all sections in a module. This can be useful when targets or
code-models place strict requirements on how sections must be laid out
in memory.

If RTDyldMemoryManger::needsToReserveAllocationSpace() is overridden to return
true then the JIT will call the following method on the memory manager, which
can be used to preallocate the necessary memory.

void RTDyldMemoryManager::reserveAllocationSpace(uintptr_t CodeSize,
                                                 uintptr_t DataSizeRO,
                                                 uintptr_t DataSizeRW)

Patch by Vaidas Gasiunas. Thanks very much Viadas!

llvm-svn: 201259
2014-02-12 21:30:07 +00:00
Juergen Ributzka 5fe955cb75 Add target analysis passes to the codegen pipeline for MCJIT.
This patch adds the target analysis passes (usually TargetTransformInfo) to the
codgen pipeline. We also expose now the AddAnalysisPasses method through the C
API, because the optimizer passes would also benefit from better target-specific
cost models.

Reviewed by Andrew Kaylor

llvm-svn: 199926
2014-01-23 19:23:28 +00:00
Lang Hames 173c69f226 Re-apply r196639: Add support for archives and object file caching under MCJIT.
I believe the bot failures on linux systems were due to overestimating the
alignment of object-files within archives, which are only guaranteed to be
two-byte aligned. I have reduced the alignment in
RuntimeDyldELF::createObjectImageFromFile accordingly.

llvm-svn: 198737
2014-01-08 04:09:09 +00:00
NAKAMURA Takumi 87e0880606 Whitespace cleanups.
llvm-svn: 196654
2013-12-07 11:21:42 +00:00
Lang Hames 567befd88f Revert r196639 while I investigate a bot failure.
llvm-svn: 196641
2013-12-07 04:25:19 +00:00
Lang Hames a691358078 Add support for archives and object file caching under MCJIT.
Patch by Andy Kaylor, with minor edits to resolve merge conflicts.

llvm-svn: 196639
2013-12-07 03:05:51 +00:00
Yaron Keren 1ec9df3322 Replaced non-ASCII character.
llvm-svn: 193324
2013-10-24 10:04:47 +00:00
Andrew Kaylor c89fc826b2 Optimizing MCJIT module state tracking
Patch co-developed with Yaron Keren.

llvm-svn: 193291
2013-10-24 00:19:14 +00:00
Andrew Kaylor 2fb40ce1dc FIXME comment shouldn't have been doxygen style
llvm-svn: 193131
2013-10-21 23:27:02 +00:00
Andrew Kaylor 4fba04942d Improving MCJIT/RuntimeDyld thread safety
llvm-svn: 193094
2013-10-21 17:42:06 +00:00
Yaron Keren fb95582bf1 Added comments from Andrew Kaylor.
llvm-svn: 193033
2013-10-19 09:03:20 +00:00
Andrew Kaylor c442a76c60 Adding support for deregistering EH frames with MCJIT.
Patch by Yaron Keren

llvm-svn: 192753
2013-10-16 00:14:21 +00:00
Andrew Kaylor 7bb1344c67 Adding multiple object support to MCJIT EH frame handling
llvm-svn: 192504
2013-10-11 21:25:48 +00:00
Andrew Kaylor 1b2cfb6495 Adding support and tests for multiple module handling in lli
llvm-svn: 191938
2013-10-04 00:49:38 +00:00
Filip Pizlo 7aa695e026 This threads SectionName through the allocateCodeSection/allocateDataSection APIs, both in C++ and C land.
It's useful for the memory managers that are allocating a section to know what the name of the section is.  
At a minimum, this is useful for low-level debugging - it's customary for JITs to be able to tell you what 
memory they allocated, and as part of any such dump, they should be able to tell you some meta-data about 
what each allocation is for.  This allows clients that supply their own memory managers to do this.  
Additionally, we also envision the SectionName being useful for passing meta-data from within LLVM to an LLVM 
client.

This changes both the C and C++ APIs, and all of the clients of those APIs within LLVM.  I'm assuming that 
it's safe to change the C++ API because that API is allowed to change.  I'm assuming that it's safe to change 
the C API because we haven't shipped the API in a release yet (LLVM 3.3 doesn't include the MCJIT memory 
management C API).

llvm-svn: 191804
2013-10-02 00:59:25 +00:00
Andrew Kaylor ea395924d2 Adding multiple module support for MCJIT.
Tests to follow.

PIC with small code model and  EH frame handling will not work with multiple modules.  There are also some rough edges to be smoothed out for remote target support.

llvm-svn: 191722
2013-10-01 01:47:35 +00:00
David Tweed 2e7efedd39 Minor changes to the MCJITTest unittests to use the correct API for finalizing
the JIT object (including XFAIL an ARM test that now needs fixing). Also renames
internal function for consistency.

llvm-svn: 182085
2013-05-17 10:01:46 +00:00
Filip Pizlo 9bc53e8467 SectionMemoryManager shouldn't be a JITMemoryManager. Previously, the
EngineBuilder interface required a JITMemoryManager even if it was being used 
to construct an MCJIT. But the MCJIT actually wants a RTDyldMemoryManager. 
Consequently, the SectionMemoryManager, which is meant for MCJIT, derived 
from the JITMemoryManager and then stubbed out a bunch of JITMemoryManager 
methods that weren't relevant to the MCJIT.

This patch fixes the situation: it teaches the EngineBuilder that 
RTDyldMemoryManager is a supertype of JITMemoryManager, and that it's 
appropriate to pass a RTDyldMemoryManager instead of a JITMemoryManager if 
we're using the MCJIT. This allows us to remove the stub methods from 
SectionMemoryManager, and make SectionMemoryManager a direct subtype of 
RTDyldMemoryManager.

llvm-svn: 181820
2013-05-14 19:29:00 +00:00
Andrew Kaylor ced4e8ff6e Re-enabling MCJIT object caching with memory leak fixed
llvm-svn: 180575
2013-04-25 21:02:36 +00:00
Rafael Espindola 72780ed996 Revert "Adding object caching support to MCJIT"
This reverts commit 07f03923137a91e3cca5d7fc075a22f8c9baf33a.

Looks like it broke the valgrind bot:

http://lab.llvm.org:8011/builders/llvm-x86_64-linux-vg_leak/builds/649

llvm-svn: 180249
2013-04-25 03:47:41 +00:00
Andrew Kaylor 1d2d8e0e84 Adding object caching support to MCJIT
llvm-svn: 180146
2013-04-23 21:26:38 +00:00
Chandler Carruth 802d755533 Sort includes for all of the .h files under the 'lib' tree. These were
missed in the first pass because the script didn't yet handle include
guards.

Note that the script is now able to handle all of these headers without
manual edits. =]

llvm-svn: 169224
2012-12-04 07:12:27 +00:00
Andrew Kaylor d8ffd9c7e7 Add interface for object-based JIT events.
This patch adds the interface to expose events from MCJIT when an object is emitted or freed and implements the MCJIT functionality to send those events.  The IntelJITEventListener implementation is left empty for now.  It will be fleshed out in a future patch.

llvm-svn: 167475
2012-11-06 18:51:59 +00:00
Andrew Kaylor a714efc1bd Add a method to indicate section address re-assignment is finished.
Prior to this patch RuntimeDyld attempted to re-apply relocations every time reassignSectionAddress was called (via MCJIT::mapSectionAddress).  In addition to being inefficient and redundant, this led to a problem when a section was temporarily moved too far away from another section with a relative relocation referencing the section being moved.  To fix this, I'm adding a new method (finalizeObject) which the client can call to indicate that it is finished rearranging section addresses so the relocations can safely be applied.

llvm-svn: 167400
2012-11-05 20:57:16 +00:00
Andrew Kaylor adc70568f9 Clean-up of memory buffer and object ownership model in MCJIT
llvm-svn: 165053
2012-10-02 21:18:39 +00:00
Jim Grosbach 6d61397c73 Better const handling for RuntimeDyld and MCJIT.
mapSectionAddress() wasn't consistent.

llvm-svn: 163843
2012-09-13 21:50:06 +00:00
Jim Grosbach bea6753f4f MCJIT: Tidy up the constructor.
The MCJIT doesn't need or want a TargetJITInfo. That's vestigal from the old
JIT, so just remove it.

rdar://12119347

llvm-svn: 162280
2012-08-21 15:42:49 +00:00
Andrew Kaylor 1a568c3a4f Enable lazy compilation in MCJIT
llvm-svn: 161438
2012-08-07 18:33:00 +00:00
Danil Malyshev bfee542cce Move getPointerToNamedFunction() from JIT/MCJIT to JITMemoryManager.
llvm-svn: 153607
2012-03-28 21:46:36 +00:00
Chandler Carruth e26dafeb79 Revert a series of commits to MCJIT to get the build working in CMake
(and hopefully on Windows). The bots have been down most of the day
because of this, and it's not clear to me what all will be required to
fix it.

The commits started with r153205, then r153207, r153208, and r153221.
The first commit seems to be the real culprit, but I couldn't revert
a smaller number of patches.

When resubmitting, r153207 and r153208 should be folded into r153205,
they were simple build fixes.

llvm-svn: 153241
2012-03-22 05:44:06 +00:00
Danil Malyshev 02fc6e86a6 Based on this discussion: http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20120305/138477.html
1. Declare a virtual function getPointerToNamedFunction() in JITMemoryManager
2. Move the implementation of getPointerToNamedFunction() form JIT/MCJIT to DefaultJITMemoryManager.

llvm-svn: 153205
2012-03-21 18:26:47 +00:00
Jim Grosbach 0ddb3a4963 ExecutionEngine interface to re-map addresses for engines that support it.
llvm-svn: 148264
2012-01-16 23:50:55 +00:00