Commit Graph

342 Commits

Author SHA1 Message Date
Dehao Chen 396f62444d Use emplace_back to replace size() and resize().
Summary: This speeds-up thin-link for ~29% for large programs.

Reviewers: tejohnson

Reviewed By: tejohnson

Subscribers: grandinj, sanjoy, llvm-commits

Differential Revision: https://reviews.llvm.org/D35145

llvm-svn: 307543
2017-07-10 15:31:53 +00:00
Evgeniy Stepanov b9f1b014e1 Infer relocation model from module flags in relocatable LTO link.
Fix for PR33096.

llvm-svn: 303578
2017-05-22 21:11:35 +00:00
Peter Collingbourne 7b30f16c9f Re-apply r299168 and r299169 now that the libdeps are fixed.
llvm-svn: 299184
2017-03-31 04:47:07 +00:00
Peter Collingbourne f10698b940 Revert r299168 and r299169 due to library dependency issues.
http://bb.pgr.jp/builders/i686-mingw32-RA-on-linux/builds/25073/steps/build_llvmclang/logs/stdio

llvm-svn: 299171
2017-03-31 02:44:50 +00:00
Peter Collingbourne d9717aa0e4 LTO: Reduce memory consumption by creating an in-memory symbol table for InputFiles. NFCI.
Introduce symbol table data structures that can be potentially written to
disk, have the LTO library build those data structures using temporarily
constructed modules and redirect the LTO library implementation to go through
those data structures. This allows us to remove the LLVMContext and Modules
owned by InputFile.

With this change I measured a peak memory consumption decrease from 5.4GB to
2.8GB in a no-op incremental ThinLTO link of Chromium on Linux. The impact on
memory consumption is larger in COFF linkers where we are currently forced
to materialize all metadata in order to read linker options. Peak memory
consumption linking a large piece of Chromium for Windows with full LTO and
debug info decreases from >64GB (OOM) to 15GB.

Part of PR27551.

Differential Revision: https://reviews.llvm.org/D31364

llvm-svn: 299168
2017-03-31 02:28:30 +00:00
Peter Collingbourne 192d8520de More accurate header inclusions. NFC.
llvm-svn: 298960
2017-03-28 23:35:34 +00:00
Peter Collingbourne 0d56b959ad LTO: Replace InputFile::Symbol::getFlags() with predicate accessors. NFC.
This makes the predicates independent of the flag representation
and makes the code a little easier to read.

llvm-svn: 298951
2017-03-28 22:31:35 +00:00
Teresa Johnson 0c6a4ff8dc [ThinLTO] Add support for emitting minimized bitcode for thin link
Summary:
The cumulative size of the bitcode files for a very large application
can be huge, particularly with -g. In a distributed build environment,
all of these files must be sent to the remote build node that performs
the thin link step, and this can exceed size limits.

The thin link actually only needs the summary along with a bitcode
symbol table. Until we have a proper bitcode symbol table, simply
stripping the debug metadata results in significant size reduction.

Add support for an option to additionally emit minimized bitcode
modules, just for use in the thin link step, which for now just strips
all debug metadata. I plan to add a cc1 option so this can be invoked
easily during the compile step.

However, care must be taken to ensure that these minimized thin link
bitcode files produce the same index as with the original bitcode files,
as these original bitcode files will be used in the backends.

Specifically:
1) The module hash used for caching is typically produced by hashing the
written bitcode, and we want to include the hash that would correspond
to the original bitcode file. This is because we want to ensure that
changes in the stripped portions affect caching. Added plumbing to emit
the same module hash in the minimized thin link bitcode file.
2) The module paths in the index are constructed from the module ID of
each thin linked bitcode, and typically is automatically generated from
the input file path. This is the path used for finding the modules to
import from, and obviously we need this to point to the original bitcode
files. Added gold-plugin support to take a suffix replacement during the
thin link that is used to override the identifier on the MemoryBufferRef
constructed from the loaded thin link bitcode file. The assumption is
that the build system can specify that the minimized bitcode file has a
name that is similar but uses a different suffix (e.g. out.thinlink.bc
instead of out.o).

Added various tests to ensure that we get identical index files out of
the thin link step.

Reviewers: mehdi_amini, pcc

Subscribers: Prazek, llvm-commits

Differential Revision: https://reviews.llvm.org/D31027

llvm-svn: 298638
2017-03-23 19:47:39 +00:00
Peter Collingbourne 128423f99b LTO: Fix a potential race condition in the caching API.
After the call to sys::fs::exists succeeds, indicating a cache hit, we call
AddFile and the client will open the file using the supplied path. If the
client is using cache pruning, there is a potential race between the pruner
and the client. To avoid this, change the caching API so that it provides
a MemoryBuffer to the client, and have clients use that MemoryBuffer where
possible.

This scheme won't work with the gold plugin because the plugin API expects a
file path. So we have the gold plugin use the buffer identifier as a path and
live with the race for now. (Note that the gold plugin isn't actually affected
by the problem at the moment because it doesn't support cache pruning.)

This effectively reverts r279883 modulo the change to use the existing path
in the gold plugin.

Differential Revision: https://reviews.llvm.org/D31063

llvm-svn: 298020
2017-03-17 00:34:07 +00:00
Peter Collingbourne ab76a19afb LTO: When creating a local cache, create the cache directory if it does not already exist.
Differential Revision: https://reviews.llvm.org/D30519

llvm-svn: 296726
2017-03-02 02:02:38 +00:00
Peter Collingbourne 8dabd7e7de gold-plugin: Remove unused variable.
llvm-svn: 296533
2017-02-28 22:40:30 +00:00
Teresa Johnson 95597ae2c9 [ThinLTO] PrintStatistics when we exit early for thinlto-index-only
Summary:
This is necessary to get stats from the ThinLink printed before the
early exit when compiling in a distributed build.

Reviewers: mehdi_amini

Subscribers: Prazek, llvm-commits

Differential Revision: https://reviews.llvm.org/D29461

llvm-svn: 293908
2017-02-02 17:33:53 +00:00
Peter Collingbourne 6201d78653 gold-plugin: Simplify naming of object files created with save-temps or obj-path.
Now we never append a number to the file name for task ID 0.

Differential Revision: https://reviews.llvm.org/D29160

llvm-svn: 293132
2017-01-26 02:07:05 +00:00
Peter Collingbourne 5b8a1bd4a2 gold-plugin: Add the file path to the file open error diagnostic.
llvm-svn: 293013
2017-01-25 03:35:28 +00:00
Dehao Chen 2797800595 Pass sample pgo flags to thinlto.
Summary: ThinLTO needs to invoke SampleProfileLoader pass during link time in order to annotate profile correctly after module importing.

Reviewers: davidxl, mehdi_amini, tejohnson

Subscribers: pcc, davide, llvm-commits, mehdi_amini

Differential Revision: https://reviews.llvm.org/D27790

llvm-svn: 289957
2016-12-16 16:48:46 +00:00
Benjamin Kramer e6ba5efa80 Apply clang-tidy's 'performance-faster-string-find' check to LLVM.
No functionality change intended.

llvm-svn: 288235
2016-11-30 10:01:11 +00:00
Mehdi Amini 48f296059d Fix gold plugin after Error API changes
llvm-svn: 286571
2016-11-11 06:04:30 +00:00
Teresa Johnson ad17679abd Split Bitcode/ReaderWriter.h into separate reader and writer headers
Summary:
Split ReaderWriter.h which contains the APIs into both the BitReader and
BitWriter libraries into BitcodeReader.h and BitcodeWriter.h.

This is to address Chandler's concern about sharing the same API header
between multiple libraries (BitReader and BitWriter). That concern is
why we create a single bitcode library in our downstream build of clang,
which led to r286297 being reverted as it added a dependency that
created a cycle only when there is a single bitcode library (not two as
in upstream).

Reviewers: mehdi_amini

Subscribers: dlj, mehdi_amini, llvm-commits

Differential Revision: https://reviews.llvm.org/D26502

llvm-svn: 286566
2016-11-11 05:34:58 +00:00
Peter Collingbourne 7576cb0fa7 Bitcode: Remove the remnants of the BitcodeDiagnosticInfo class.
The BitcodeReader no longer produces BitcodeDiagnosticInfo diagnostics.
The only remaining reference was in the gold plugin; the code there has been
dead since we stopped producing InvalidBitcodeSignature error codes in r225562.
While at it remove the InvalidBitcodeSignature error code.

llvm-svn: 286326
2016-11-09 01:09:11 +00:00
Rafael Espindola 7912110ddc Make the LTO comdat api more symbol table friendly.
In an IR symbol table I would expect the comdats to be represented as:

- A table of strings, one for each comdat name.
- Each symbol has an optional index into that table.

The natural api for accessing that would be

InputFile:
ArrayRef<StringRef> getComdatTable() const;

Symbol:
int getComdatIndex() const;

This patch implements an API as close to that as possible.  The
implementation on top of the current IRObjectFile is a bit hackish,
but should map just fine over a symbol table and is very convenient to
use.

llvm-svn: 285061
2016-10-25 12:02:03 +00:00
Teresa Johnson ec544c552e [ThinLTO] Default backend threads to heavyweight_hardware_concurrency
Summary:
Changes default backend parallelism from thread::hardware_concurrency to
the new llvm::heavyweight_hardware_concurrency, which for X86 Linux
defaults to the number of physical cores (and will fall back to
thread::hardware_concurrency otherwise). This avoid oversubscribing
the physical cores using hyperthreading.

Reviewers: mehdi_amini, pcc

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D25775

llvm-svn: 284618
2016-10-19 17:35:01 +00:00
Rafael Espindola 62382c930d Return a StringRef instead of a Comdat*.
This is a small step in making this interface compatible with an
bitcode symbol table.

llvm-svn: 284408
2016-10-17 18:51:02 +00:00
Peter Collingbourne 80186a57d6 LTO: Simplify caching interface.
The NativeObjectOutput class has a design problem: it mixes up the caching
policy with the interface for output streams, which makes the client-side
code hard to follow and would for example make it harder to replace the
cache implementation in an arbitrary client.

This change separates the two aspects by moving the caching policy
to a separate field in Config, replacing NativeObjectOutput with a
NativeObjectStream class which only deals with streams and does not need to
be overridden by most clients and introducing an AddFile callback for adding
files (e.g. from the cache) to the link.

Differential Revision: https://reviews.llvm.org/D24622

llvm-svn: 282299
2016-09-23 21:33:43 +00:00
Teresa Johnson 896fee2846 [gold] Split plugin options controlling ThinLTO and codegen parallelism.
Summary:
As suggested in D24826, use different options for ThinLTO backend
parallelism from the option controlling regular LTO code gen
parallelism. They are already split in the LTO API, and this enables
controlling them with different clang options.

Reviewers: pcc, mehdi_amini

Subscribers: dexonsmith, llvm-commits, mehdi_amini

Differential Revision: https://reviews.llvm.org/D24873

llvm-svn: 282290
2016-09-23 20:35:19 +00:00
Teresa Johnson 3f212b8908 [ThinLTO] Emit files for distributed builds for all modules
With the new LTO API in r278338, we stopped emitting the individual
index files and imports files for some modules in the distributed backend
case (thinlto-index-only plugin option).

Specifically, this is when the linker decides not to include a module in the
link, because it was in an archive library and did not have a strong
reference to it. Not creating the expected output files makes the
distributed build system implementation more difficult, in terms of
checking for the expected outputs of the thin link, and scheduling the
backend jobs. To address this, the gold-plugin will write dummy empty
.thinlto.bc and .imports files for modules not included in the link
(which LTO never sees).

Augmented a gold v1.12+ test, since that version of gold has the handling
for notifying on modules not being included in the link.

llvm-svn: 282100
2016-09-21 19:12:05 +00:00
Peter Collingbourne 0758644461 gold: Simplify. Do not unnecessarily enumerate Obj's symbols.
llvm-svn: 281437
2016-09-14 02:55:16 +00:00
Teresa Johnson 26a462877b [ThinLTO] Move loading of cache entry to client
Summary:
Have the cache pass back the path to the cache entry when it
is ready to be loaded, instead of a buffer.

For gold-plugin we can simply pass this file back to gold directly,
which avoids expensive writing of a separate tmp file. Ensure
the cache entry is not deleted on cleanup by adjusting the setting
of the IsTemporary flags.

Moved the loading of the buffer into llvm-lto2 to maintain current
behavior.

Reviewers: mehdi_amini

Subscribers: llvm-commits, mehdi_amini

Differential Revision: https://reviews.llvm.org/D23946

llvm-svn: 279883
2016-08-26 23:29:14 +00:00
Teresa Johnson 57891a50a8 [ThinLTO/gold] Add caching support to gold-plugin
Summary:
With support now in the new LTO API for caching (r279576), add
optional ThinLTO caching in the gold-plugin.

Reviewers: mehdi_amini

Subscribers: mehdi_amini, llvm-commits

Differential Revision: https://reviews.llvm.org/D23836

llvm-svn: 279631
2016-08-24 15:11:47 +00:00
Mehdi Amini 6ec233313a Fix Gold Plugin after API change in the LTO API (constify callback type)
llvm-svn: 279440
2016-08-22 16:41:58 +00:00
Mehdi Amini dc4c8cf9ac [LTO] Handles commons in monolithic LTO
The gold-plugin was doing this internally, now the API is handling
commons correctly based on the given resolution.

Differential Revision: https://reviews.llvm.org/D23739

llvm-svn: 279417
2016-08-22 06:25:46 +00:00
Teresa Johnson 765941a841 [gold/ThinLTO] Restore ThinLTO file management in gold plugin
Summary:
The gold-plugin changes added along with the new LTO API in r278338 had
the effect of removing the management of the PluginInputFile that
ensured the files weren't released back to gold until the backend
threads were complete. Add back the old file handling.

Fixes PR29020.

Reviewers: mehdi_amini

Subscribers: mehdi_amini, llvm-commits, hjl.tools

Differential Revision: https://reviews.llvm.org/D23721

llvm-svn: 279356
2016-08-20 01:24:07 +00:00
Teresa Johnson f2b5ec6ef4 [ThinLTO] Keep common symbols in ThinLTO modules
Summary:
Skip the merging of common symbols for ThinLTO modules, they will be
merged by the final native object link. Trying to merge the symbols and
add to a combined module will incorrectly enable the common symbol to be
internalized in the ThinLTO module. Additionally, we will not want to
create a combined module for ThinLTO distributed builds.

This fixes failures in 7 cpu2006 benchmarks from the new LTO API in
ThinLTO mode.

Reviewers: mehdi_amini

Subscribers: pcc, llvm-commits, mehdi_amini

Differential Revision: https://reviews.llvm.org/D23637

llvm-svn: 279023
2016-08-18 01:08:50 +00:00
Mehdi Amini eccffada33 [LTO] Change addSaveTemps API: do not add dot to the supplied prefix path
Summary:
It does not play well with directories (end up with a bunch of hidden
files).
Also, do not strip the 0 suffix for the first task, especially since
0 can be used by ThinLTO as well now.

Reviewers: tejohnson

Subscribers: mehdi_amini, pcc, llvm-commits

Differential Revision: https://reviews.llvm.org/D23612

llvm-svn: 279014
2016-08-18 00:12:33 +00:00
Mehdi Amini 970800e0c8 [LTO] Introduce an Output class to wrap the output stream creation (NFC)
Summary:
While NFC for now, this will allow more flexibility on the client side
to hold state necessary to back up the stream.
Also when adding caching, this class will grow in complexity.

Note I blindly modified the gold-plugin as I can't compile it.

Reviewers: tejohnson

Subscribers: mehdi_amini, llvm-commits

Differential Revision: https://reviews.llvm.org/D23542

llvm-svn: 278907
2016-08-17 06:23:09 +00:00
Saleem Abdulrasool 98541b09f4 Revert "gold: add a cast to appease std::max NFC"
This was fixed differently by Teresa and this should no longer be needed.

llvm-svn: 278622
2016-08-14 05:07:20 +00:00
Saleem Abdulrasool 0bc85613f7 gold: add a cast to appease std::max NFC
llvm-svn: 278567
2016-08-12 21:56:12 +00:00
Teresa Johnson 9ba95f99f3 Restore "Resolution-based LTO API."
This restores commit r278330, with fixes for a few bot failures:
- Fix a late change I had made to the save temps output file that I
  missed due to existing files sitting on my disk
- Fix a bunch of Windows bot failures with "ambiguous call to overloaded
  function" due to confusion between llvm::make_unique vs
  std::make_unique (preface the new make_unique calls with "llvm::")
- Attempt to fix a modules bot failure by adding a missing include
  to LTO/Config.h.

Original change:

Resolution-based LTO API.

Summary:
This introduces a resolution-based LTO API. The main advantage of this API over
existing APIs is that it allows the linker to supply a resolution for each
symbol in each object, rather than the combined object as a whole. This will
become increasingly important for use cases such as ThinLTO which require us
to process symbol resolutions in a more complicated way than just adjusting
linkage.

Patch by Peter Collingbourne.

Reviewers: rafael, tejohnson, mehdi_amini

Subscribers: lhames, tejohnson, mehdi_amini, llvm-commits

Differential Revision: https://reviews.llvm.org/D20268

llvm-svn: 278338
2016-08-11 14:58:12 +00:00
Teresa Johnson cbf684e6c6 Revert "Resolution-based LTO API."
This reverts commit r278330.

I made a change to the save temps output that is causing issues with the
bots. Didn't realize this because I had older output files sitting on
disk in my test output directory.

llvm-svn: 278331
2016-08-11 13:03:56 +00:00
Teresa Johnson f99573b3ee Resolution-based LTO API.
Summary:
This introduces a resolution-based LTO API. The main advantage of this API over
existing APIs is that it allows the linker to supply a resolution for each
symbol in each object, rather than the combined object as a whole. This will
become increasingly important for use cases such as ThinLTO which require us
to process symbol resolutions in a more complicated way than just adjusting
linkage.

Patch by Peter Collingbourne.

Reviewers: rafael, tejohnson, mehdi_amini

Subscribers: lhames, tejohnson, mehdi_amini, llvm-commits

Differential Revision: https://reviews.llvm.org/D20268

Address review comments

llvm-svn: 278330
2016-08-11 12:56:40 +00:00
Teresa Johnson 1e2708c9e0 [ThinLTO/gold] Support for getting list of included objects from gold
Summary:
In the distributed backend case, the ThinLink step and the final native
object link are separate processes. This can be problematic when archive
libraries are involved in the link (e.g. via --start-lib/--end-lib
pairs). The linker only includes objects from libraries when
there is a strong reference to them, and depending on the intervening
ThinLTO backend processes' importing/inlining, the strong references
may appear different in the two link steps. See D22356 and D22467
for two scenarios where this causes issues.

To ensure that the final link includes the same objects, this patch
adds support for an "=filename" form of the thinlto-index-only plugin
option, in which case objects gold included in the link are emitted to
the given filename. This should be used as input to the final link (e.g.
via the @filename option to gold), instead of listing all the objects
within --start-lib/--end-lib pairs again.

Note that the support for the gold callback that identifies included
objects was added in gold version 1.12.

Reviewers: davidxl, mehdi_amini

Subscribers: llvm-commits, mehdi_amini

Differential Revision: https://reviews.llvm.org/D22677

llvm-svn: 276450
2016-07-22 18:20:22 +00:00
Teresa Johnson 35e0204eec [ThinLTO/gold] Perform index-based weak/linkonce resolution
Summary:
Invoke the weak/linkonce symbol resolution support (already used by
libLTO) that operates via the summary index.

This ensures prevailing linkonce are kept, by making them weak, and
marks preempted copies as available_externally when possible.

With this change, the older support for keeping the prevailing linkonce
(by changing their symbol resolution) is removed.

Reviewers: mehdi_amini

Subscribers: llvm-commits, mehdi_amini

Differential Revision: http://reviews.llvm.org/D22302

llvm-svn: 275474
2016-07-14 21:13:24 +00:00
Teresa Johnson b907d06151 [ThinLTO/gold] Enable symbol resolution in distributed backend case
While testing a follow-on change to enable index-based symbol resolution
and internalization in the distributed backends, I realized that a test
case change I made in r275247 was only required because we were not
analyzing symbols in the claimed files in thinlto-index-only mode.

In the fixed test case there should be no internalization because we are
linking in -shared mode, so f() is in fact exported, which is detected
properly when we analyze symbols in thinlto-index-only mode. Note that
this is not (yet) a correctness issue (because we are not yet performing
the index-based linkage optimizations in the distributed backends -
that's coming in a follow-on patch).

llvm-svn: 275277
2016-07-13 16:35:56 +00:00
Teresa Johnson 27694571b1 [ThinLTO/gold] ThinLTO internalization fixes
Internalization was missing cases where we originally had a local symbol
that was promoted eagerly but not actually exported. This is because we
were only internalizing the set of global (non-local) symbols that were
PREVAILAING_DEF_IRONLY. Instead, collect the set of global symbols that
are referenced outside of a single IR file, and skip internalization for
those.

llvm-svn: 275247
2016-07-13 03:42:41 +00:00
Rafael Espindola aee1a0e407 Don't use the new x86 relax relocations on the gold plugin.
Should bring back the bots with old versions.

llvm-svn: 273022
2016-06-17 17:53:57 +00:00
Peter Collingbourne 96efdd6107 IR: Introduce local_unnamed_addr attribute.
If a local_unnamed_addr attribute is attached to a global, the address
is known to be insignificant within the module. It is distinct from the
existing unnamed_addr attribute in that it only describes a local property
of the module rather than a global property of the symbol.

This attribute is intended to be used by the code generator and LTO to allow
the linker to decide whether the global needs to be in the symbol table. It is
possible to exclude a global from the symbol table if three things are true:
- This attribute is present on every instance of the global (which means that
  the normal rule that the global must have a unique address can be broken without
  being observable by the program by performing comparisons against the global's
  address)
- The global has linkonce_odr linkage (which means that each linkage unit must have
  its own copy of the global if it requires one, and the copy in each linkage unit
  must be the same)
- It is a constant or a function (which means that the program cannot observe that
  the unique-address rule has been broken by writing to the global)

Although this attribute could in principle be computed from the module
contents, LTO clients (i.e. linkers) will normally need to be able to compute
this property as part of symbol resolution, and it would be inefficient to
materialize every module just to compute it.

See:
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20160509/356401.html
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-20160516/356738.html
for earlier discussion.

Part of the fix for PR27553.

Differential Revision: http://reviews.llvm.org/D20348

llvm-svn: 272709
2016-06-14 21:01:22 +00:00
Teresa Johnson 7ab1f69272 [ThinLTO/gold] Enable summary-based internalization
Summary: Enable existing summary-based importing support in the gold-plugin.

Reviewers: mehdi_amini

Subscribers: llvm-commits, mehdi_amini

Differential Revision: http://reviews.llvm.org/D21080

llvm-svn: 272239
2016-06-09 01:14:13 +00:00
Teresa Johnson 4cafbdd4bd [ThinLTO/gold] Pass import lists by reference (NFC)
In my earlier patch r271690 I missed passing a map by reference in one
place.

llvm-svn: 271732
2016-06-03 20:46:32 +00:00
Teresa Johnson 56c68353fa [ThinLTO/gold] Pass down the imports lists from the thin link (NFC)
No longer need to compute the imports in each backend thread.

llvm-svn: 271690
2016-06-03 17:15:54 +00:00
Benjamin Kramer 82de7d323d Apply clang-tidy's misc-move-constructor-init throughout LLVM.
No functionality change intended, maybe a tiny performance improvement.

llvm-svn: 270997
2016-05-27 14:27:24 +00:00
Peter Collingbourne 1eaa97f439 Linker: teach the IR mover to return llvm::Error.
This will be needed in order to consistently return an Error
to clients of the API being developed in D20268.

Differential Revision: http://reviews.llvm.org/D20550

llvm-svn: 270967
2016-05-27 05:21:35 +00:00
Teresa Johnson 683abe79b2 [ThinLTO/gold] Handle bitcode archives
Summary:
Several changes were required for ThinLTO links involving bitcode
archive static libraries. With this patch clang/llvm bootstraps with
ThinLTO and gold.

The first is that the gold callbacks get_input_file and
release_input_file can normally be used to get file information for
each constituent bitcode file within an archive. However, these
interfaces lock the underlying file and can't be for each archive
constituent for ThinLTO backends where we get all the input files up
front and don't release any until after the backend threads complete.
However, it is sufficient to only get and release once per file, and
then each consituent bitcode file can be accessed via get_view. This
required saving some information to identify which file handle is the
"leader" for each claimed file sharing the same file descriptor, and
other information so that get_input_file isn't necessary later when
processing the backends.

Second, the module paths in the index need to distinguish between
different constituent bitcode files within the same archive file,
otherwise they will all end up with the same archive file path.
Do this by appending the offset within the archive for the start of the
bitcode file, returned by get_input_file when we claim each bitcode file,
and saving that along with the file handle.

Third, rather than have the function importer try to load a file based
on the module path identifier (which now contains a suffix to
distinguish different bitcode files within an archive), use a custom
module loader. This is the same approach taken in libLTO, and I am using
the support refactored into the new LTO.h header in r270509. The module
loader parses the bitcode files out of the memory buffers returned from
gold via the get_view callback and saved in a map. This also means that
we call the function importer directly, rather than add it to the pass
pipeline (which was in the plan to do already for other reasons).

Reviewers: pcc, joker.eph

Subscribers: llvm-commits, joker.eph

Differential Revision: http://reviews.llvm.org/D20559

llvm-svn: 270814
2016-05-26 01:46:41 +00:00
Rafael Espindola 8c34dd8257 Delete Reloc::Default.
Having an enum member named Default is quite confusing: Is it distinct
from the others?

This patch removes that member and instead uses Optional<Reloc> in
places where we have a user input that still hasn't been maped to the
default value, which is now clear has no be one of the remaining 3
options.

llvm-svn: 269988
2016-05-18 22:04:49 +00:00
Reid Kleckner 8e96c3e9dd [ThinLTO] Use semicolon to separate path prefix replacement
Summary:
Colons can appear in Windows paths after drive letters. Both colon and
semicolon are valid characters in filenames, but neither are very
common. Semicolon seems just as good, and makes the test pass on
Windows.

Reviewers: tejohnson

Subscribers: joker.eph, llvm-commits

Differential Revision: http://reviews.llvm.org/D20332

llvm-svn: 269798
2016-05-17 18:43:22 +00:00
Teresa Johnson bbd10b4579 [ThinLTO] Option to control path of distributed backend files
Summary:
Add support to control where files for a distributed backend (the
individual index files and optional imports files) are created.

This is invoked with a new thinlto-prefix-replace option in the gold
plugin and llvm-lto. If specified, expects a string of the form
"oldprefix:newprefix", and instead of generating these files in the
same directory path as the corresponding bitcode file, will use a path
formed by replacing the bitcode file's path prefix matching oldprefix
with newprefix.

Also add a new replace_path_prefix helper to Path.h in libSupport.

Depends on D19636.

Reviewers: joker.eph

Subscribers: llvm-commits, joker.eph

Differential Revision: http://reviews.llvm.org/D19644

llvm-svn: 269771
2016-05-17 14:45:30 +00:00
Teresa Johnson 532e33be42 [ThinLTO] Use correct pipeline for ThinLTO in gold-plugin.
This change is the gold side of the change made in D17115 and clang
patch r261045 to add a ThinLTO specific pipeline that moves more of
the optimization to the backends.

llvm-svn: 269386
2016-05-13 01:25:31 +00:00
Teresa Johnson 8570fe47ef [ThinLTO] Add option to emit imports files for distributed backends
Summary:
Add support for emission of plaintext lists of the imported files for
each distributed backend compilation. Used for distributed build file
staging.

Invoked with new gold-plugin thinlto-emit-imports-files option, which is
only valid with thinlto-index-only (i.e. for distributed builds), or
from llvm-lto with new -thinlto-action=emitimports value.

Depends on D19556.

Reviewers: joker.eph

Subscribers: llvm-commits, joker.eph

Differential Revision: http://reviews.llvm.org/D19636

llvm-svn: 269067
2016-05-10 15:54:09 +00:00
Teresa Johnson 84174c3771 Restore "[ThinLTO] Emit individual index files for distributed backends"
This restores commit r268627:
    Summary:
    When launching ThinLTO backends in a distributed build (currently
    supported in gold via the thinlto-index-only plugin option), emit
    an individual index file for each backend process as described here:
    http://lists.llvm.org/pipermail/llvm-dev/2016-April/098272.html

    ...

    Differential Revision: http://reviews.llvm.org/D19556

Address msan failures by avoiding std::prev on map.end(), the
theory is that this is causing issues due to some known UB problems
in __tree.

llvm-svn: 269059
2016-05-10 13:48:23 +00:00
Vitaly Buka 1df2338bb6 Revert "[ThinLTO] Emit individual index files for distributed backends"
MemorySanitizer: use-of-uninitialized-value in lib/Bitcode/Writer/BitcodeWriter.cpp:364:70
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/12544/steps/check-llvm%20msan/logs/stdio

This reverts commit 0c4a898ea550699d1b2f4fe3767251c8f9a48d52.

llvm-svn: 268660
2016-05-05 18:31:00 +00:00
Teresa Johnson 9254ebe3c0 [ThinLTO] Emit individual index files for distributed backends
Summary:
When launching ThinLTO backends in a distributed build (currently
supported in gold via the thinlto-index-only plugin option), emit
an individual index file for each backend process as described here:
http://lists.llvm.org/pipermail/llvm-dev/2016-April/098272.html

The individual index file encodes the summary and module information
required for implementing the importing/exporting decisions made
for a given module in the thin link step.
This is in place of the current mechanism that uses the combined index
to make importing decisions in each back end independently. It is an
enabler for doing global summary based optimizations in the thin link
step (which will be recorded in the individual index files), and reduces
the size of the index that must be sent to each backend process, and
the amount of work to scan it in the backends.

Rather than create entirely new ModuleSummaryIndex structures (and all
the included unique_ptrs) for each backend index file, a map is created
to record all of the GUID and summary pointers needed for a particular
index file. The IndexBitcodeWriter walks this map instead of the full
index (hiding the details of managing the appropriate summary iteration
in a new iterator subclass). This is more efficient than walking the
entire combined index and filtering out just the needed summaries during
each backend bitcode index write.

Depends on D19481.

Reviewers: joker.eph

Subscribers: llvm-commits, joker.eph

Differential Revision: http://reviews.llvm.org/D19556

llvm-svn: 268627
2016-05-05 13:44:56 +00:00
Evgeniy Stepanov d194e51726 [gold] Fix linkInModule and extend common.ll test.
Fix early exit from linkInModule. IRMover::move returns false on
success and true on error.

Add a few more cases of merged common linkage variables with
different sizes and alignments.

llvm-svn: 267437
2016-04-25 18:23:29 +00:00
Davide Italiano 8d728fccb8 [gold-plugin] Remove dead assignment. NFC.
llvm-svn: 267429
2016-04-25 17:18:45 +00:00
Teresa Johnson 723cd05f83 [gold] Gate value name discarding under save-temps
Summary:
This removes a couple of flags added to control this behavior, and
simply keeps all value names when save-temps is specified.

Reviewers: rafael

Subscribers: llvm-commits, pcc, davide

Differential Revision: http://reviews.llvm.org/D19384

llvm-svn: 267279
2016-04-23 05:15:59 +00:00
Teresa Johnson 143d15bc29 Re-enable "[gold-plugin] Disable name for values other than GlobalValue"
This restores r266871 with a fix for gold tests relying on the value
names, when using a release compiler, by adding a way to disable the
default discarding. Update affected tests to use the new mechanism so
that value names are preserved as expected, regardless of how the
compiler was built.

llvm-svn: 266881
2016-04-20 15:16:57 +00:00
Teresa Johnson b35cc691ea [ThinLTO] Prevent importing of "llvm.used" values
Summary:
This patch prevents importing from (and therefore exporting from) any
module with a "llvm.used" local value. Local values need to be promoted
and renamed when importing, and their presense on the llvm.used variable
indicates that there are opaque uses that won't see the rename. One such
example is a use in inline assembly.

See also the discussion at:
http://lists.llvm.org/pipermail/llvm-dev/2016-April/098047.html

As part of this, move collectUsedGlobalVariables out of Transforms/Utils
and into IR/Module so that it can be used more widely. There are several
other places in LLVM that used copies of this code that can be cleaned
up as a follow on NFC patch.

Reviewers: joker.eph

Subscribers: pcc, llvm-commits, joker.eph

Differential Revision: http://reviews.llvm.org/D18986

llvm-svn: 266877
2016-04-20 14:39:45 +00:00
Teresa Johnson f0bedf5343 Revert "[gold-plugin] Disable name for values other than GlobalValue"
This reverts commit r266871. Setting the default based on the NDEBUG
flag is causing test failures. Need to figure out whether to change this
approach or update tests.

llvm-svn: 266872
2016-04-20 13:18:47 +00:00
Teresa Johnson d3ded4a441 [gold-plugin] Disable name for values other than GlobalValue
Summary:
Applies Mehdi's optimization (r263086) to disable value names other than
for GlobalValues to LTO/ThinLTO performed via the gold-plugin, in the
same manner as it is applied in libLTO.

Reviewers: rafael, joker-eph

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D19269

llvm-svn: 266871
2016-04-20 13:01:37 +00:00
Duncan P. N. Exon Smith ed8fdb2a0e IR: Rename API for enabling ODR uniquing of DITypes, NFC
As per David's review, rename everything in the new API for ODR type
uniquing of debug info.

    ensureDITypeMap  => enableDebugTypeODRUniquing
    destroyDITypeMap => disableDebugTypeODRUniquing
    hasDITypeMap     => isODRUniquingDebugTypes

llvm-svn: 266713
2016-04-19 04:55:25 +00:00
Mehdi Amini b550cb1750 [NFC] Header cleanup
Removed some unused headers, replaced some headers with forward class declarations.

Found using simple scripts like this one:
clear && ack --cpp -l '#include "llvm/ADT/IndexedMap.h"' | xargs grep -L 'IndexedMap[<]' | xargs grep -n --color=auto 'IndexedMap'

Patch by Eugene Kosov <claprix@yandex.ru>

Differential Revision: http://reviews.llvm.org/D19219

From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 266595
2016-04-18 09:17:29 +00:00
Rafael Espindola 197871d532 Fix a use after move.
llvm-svn: 266566
2016-04-17 18:56:49 +00:00
Rafael Espindola 3c1c9875b9 Keep only the splitCodegen version that takes a factory.
This makes it much easier to see that all created TargetMachines are
equivalent.

llvm-svn: 266564
2016-04-17 18:42:27 +00:00
Duncan P. N. Exon Smith 5ab2be094e IR: Use an explicit map for debug info type uniquing
Rather than relying on the structural equivalence of DICompositeType to
merge type definitions, use an explicit map on the LLVMContext that
LLParser and BitcodeReader consult when constructing new nodes.
Each non-forward-declaration DICompositeType with a non-empty
'identifier:' field is stored/loaded from the type map, and the first
definiton will "win".

This map is opt-in: clients that expect ODR types from different modules
to be merged must call LLVMContext::ensureDITypeMap.

  - Clients that just happen to load more than one Module in the same
    LLVMContext won't magically merge types.

  - Clients (like LTO) that want to continue to merge types based on ODR
    identifiers should opt-in immediately.

I have updated LTOCodeGenerator.cpp, the two "linking" spots in
gold-plugin.cpp, and llvm-link (unless -disable-debug-info-type-map) to
set this.

With this in place, it will be straightforward to remove the DITypeRef
concept (i.e., referencing types by their 'identifier:' string rather
than pointing at them directly).

llvm-svn: 266549
2016-04-17 03:58:21 +00:00
Evgeniy Stepanov 268826a287 [gold] Save bitcode for module partitions (save-temps + split codegen).
llvm-svn: 265583
2016-04-06 18:32:13 +00:00
Marianne Mailhot-Sarrasin a5a750eaf1 gold-plugin: Fixed typo in an error message.
llvm-svn: 264860
2016-03-30 12:20:53 +00:00
Davide Italiano 289a43ed0a [gold] Emit a diagnostic in case we fail to remove a file.
llvm-svn: 263914
2016-03-20 20:12:33 +00:00
Davide Italiano 2788a825e7 [gold] Use early return to simplify.
llvm-svn: 263872
2016-03-19 02:34:33 +00:00
Teresa Johnson 26ab5772b0 [ThinLTO] Renaming of function index to module summary index (NFC)
(Resubmitting after fixing missing file issue)

With the changes in r263275, there are now more than just functions in
the summary. Completed the renaming of data structures (started in
r263275) to reflect the wider scope. In particular, changed the
FunctionIndex* data structures to ModuleIndex*, and renamed related
variables and comments. Also renamed the files to reflect the changes.

A companion clang patch will immediately succeed this patch to reflect
this renaming.

llvm-svn: 263513
2016-03-15 00:04:37 +00:00
Teresa Johnson cec0cae313 Revert "[ThinLTO] Renaming of function index to module summary index (NFC)"
This reverts commit r263490. Missed a file.

llvm-svn: 263493
2016-03-14 21:18:10 +00:00
Teresa Johnson 892920b358 [ThinLTO] Renaming of function index to module summary index (NFC)
With the changes in r263275, there are now more than just functions in
the summary. Completed the renaming of data structures (started in
r263275) to reflect the wider scope. In particular, changed the
FunctionIndex* data structures to ModuleIndex*, and renamed related
variables and comments. Also renamed the files to reflect the changes.

A companion clang patch will immediately succeed this patch to reflect
this renaming.

llvm-svn: 263490
2016-03-14 21:05:56 +00:00
Teresa Johnson 76a1c1d0ba [ThinLTO] Support for reference graph in per-module and combined summary.
Summary:
This patch adds support for including a full reference graph including
call graph edges and other GV references in the summary.

The reference graph edges can be used to make importing decisions
without materializing any source modules, can be used in the plugin
to make file staging decisions for distributed build systems, and is
expected to have other uses.

The call graph edges are recorded in each function summary in the
bitcode via a list of <CalleeValueIds, StaticCount> tuples when no PGO
data exists, or <CalleeValueId, StaticCount, ProfileCount> pairs when
there is PGO, where the ValueId can be mapped to the function GUID via
the ValueSymbolTable. In the function index in memory, the call graph
edges reference the target via the CalleeGUID instead of the
CalleeValueId.

The reference graph edges are recorded in each summary record with a
list of referenced value IDs, which can be mapped to value GUID via the
ValueSymbolTable.

Addtionally, a new summary record type is added to record references
from global variable initializers. A number of bitcode records and data
structures have been renamed to reflect the newly expanded scope of the
summary beyond functions. More cleanup will follow.

Reviewers: joker.eph, davidxl

Subscribers: joker.eph, llvm-commits

Differential Revision: http://reviews.llvm.org/D17212

llvm-svn: 263275
2016-03-11 18:52:24 +00:00
Evgeniy Stepanov 4dc3c8df45 [gold] Fix common symbols handling.
LLVM Gold plugin decides which instance of a common symbol it wants
based on the symbol size in claim_file_hook. If the file that
contains the chosen instance is later dropped from the link, we end
up with an undefined reference.

This change delays this decision until the set of the included files
is known.

llvm-svn: 263180
2016-03-11 00:51:57 +00:00
Mehdi Amini d2d989609f Fix GOLD plugin build after r262976
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 262981
2016-03-09 01:55:15 +00:00
Benjamin Kramer 39988a03a5 [gold] Avoid assertion failures when taking a pointer to an empty vector.
llvm-svn: 262926
2016-03-08 14:02:46 +00:00
Teresa Johnson 3b8f6126ac Fix bot failure from r262721: unintented change in gold-plugin save-temps
The split code gen task ID should not be appended to save-temps output
file when the parallelism factor is 1 (not actually splitting).

llvm-svn: 262731
2016-03-04 18:16:00 +00:00
Teresa Johnson a17f2cd1a3 [ThinLTO] Ensure prevailing linkonce emitted as weak in ThinLTO backends
Summary:
Since IR files are all compiled into separate independent object files
in ThinLTO mode, the prevailing linkonce symbols must be emitted in its
object file even if it is no longer referenced there, e.g. if no
references remain in the module after inlining, since it may be
referenced by another ThinLTO compiled object file. This is done by
changing LDPR_PREVAILING_DEF_IRONLY* symbols to LDPR_PREVAILING_DEF,
which converts the prevailing linkonce to weak. We also don't need the
other prevailing IRONLY handling for internalization, which is not
currently performed for ThinLTO.

Test case included.

Reviewers: davidxl, rafael

Subscribers: rafael, joker.eph, llvm-commits

Differential Revision: http://reviews.llvm.org/D16173

llvm-svn: 262727
2016-03-04 17:48:35 +00:00
Teresa Johnson 7cffaf3ad0 [ThinLTO] Launch importing backends in parallel threads from gold plugin
Summary:
Launch ThinLTO backends (LTO and codegen pipelines with importing) in
parallel using a ThreadPool, after creating the combined index.
The number of threads is controlled by the existing -jobs gold plugin
option, or the hardware concurrency if not specified.

The old behavior of exiting after creating the combined index can be
invoked via a new thinlto-index-only plugin option.

This commit involves just the ThinLTO-specific pieces of D15390, the NFC
and other restructuring pieces were committed independently:
  r262677: Add hardware_concurrency interface to llvm::thread (NFC)
  r262719: Change split code gen to use ThreadPool
  r262721: Refactor gold-plugin codegen to prepare for ThinLTO threads (NFC)

Reviewers: pcc, joker.eph, rafael

Subscribers: rafael, davidxl, llvm-commits, joker.eph

Differential Revision: http://reviews.llvm.org/D15390

llvm-svn: 262724
2016-03-04 17:06:02 +00:00
Teresa Johnson a9f65554b0 Refactor gold-plugin codegen to prepare for ThinLTO threads (NFC)
This is the NFC part remaining from D15390, which refactors the
current codegen() into a CodeGen class with various modular methods and
other helper functions that will be used by the follow-on ThinLTO piece.

llvm-svn: 262721
2016-03-04 16:36:06 +00:00
Evgeniy Stepanov 330c5a60c7 [gold] Handle modules that are not included in the link.
Gold has a newly added LDPT_GET_SYMBOLS_V3 callback that can
distinguish between a module that is not included in the link, and
one that is included but has its entire interface preempted by others.

Fixes PR26674.

llvm-svn: 262676
2016-03-04 00:23:29 +00:00
Rafael Espindola 40358fb86f Pass a std::unique_ptr to IRMover::move.
It was already the one "destroying" the source module, now the API
reflects that.

llvm-svn: 260989
2016-02-16 18:50:12 +00:00
Davide Italiano 72d0d89b59 [gold] Remove inconsistent llvm_unreachable().
Differential Revision:	 http://reviews.llvm.org/D16429

llvm-svn: 258561
2016-01-22 21:36:49 +00:00
Rafael Espindola 240529892b Disable use list order on the gold plugin.
It turns out that his is *really* slow. With this change the link of
clang with plugin-opt=emit-llvm goes from 41 to 26 seconds.

We can add an option to enable it again if needed.

llvm-svn: 256302
2015-12-22 23:45:49 +00:00
Teresa Johnson 031bed291e Use LLVM style variable name (NFC)
Fixes variable name from r255779.

llvm-svn: 255824
2015-12-16 21:37:48 +00:00
Teresa Johnson cb15b739ef Add RAII wrapper for gold plugin file management
Suggested in review of r255256.

llvm-svn: 255779
2015-12-16 16:34:06 +00:00
Rafael Espindola 9d2bfc4874 Use diagnostic handler in the LLVMContext
This patch converts code that has access to a LLVMContext to not take a
diagnostic handler.

This has a few advantages

* It is easier to use a consistent diagnostic handler in a single program.
* Less clutter since we are not passing a handler around.

It does make it a bit awkward to implement some C APIs that return a
diagnostic string. I will propose new versions of these APIs and
deprecate the current ones.

llvm-svn: 255571
2015-12-14 23:17:03 +00:00
Teresa Johnson 9d5b71b3d2 [ThinLTO] Release files in gold plugin during combined index (take 2)
Ensure we release the files even when they don't hold a function index
summary section, by restructuring the control flow a little bit.

llvm-svn: 255256
2015-12-10 16:11:23 +00:00
Rafael Espindola caabe22832 Slit lib/Linker in two.
A linker normally has two stages: symbol resolution and "moving stuff".

In lib/Linker there is the complication of lazy linking some globals,
but it was still far more mixed than it needed to.

This splits the linker into a lower level IRMover and the linker proper.
The IRMover just takes a list of globals to move and a callback that
lets the user control what is lazy linked.

The main motivation is that now tools/gold (and soon lld) can use their
own symbol resolution to instruct IRMover what to do.

llvm-svn: 255254
2015-12-10 14:19:35 +00:00
Teresa Johnson db51357c11 [ThinLTO] Release files read when creating combined index in gold plugin
This wasn't causing an issue since at HEAD we exit the linker completely
after creating the combined index.

llvm-svn: 255156
2015-12-09 21:11:42 +00:00
Teresa Johnson af9e93183d Delay context construction to when/if it is needed in gold plugin (NFC)
llvm-svn: 255146
2015-12-09 19:49:40 +00:00
Teresa Johnson b13dbd633a clang-format order of gold-plugin includes (NFC)
llvm-svn: 255144
2015-12-09 19:45:55 +00:00
Rafael Espindola f49a38fc08 Always pass a diagnostic handler to the linker.
Before this patch the diagnostic handler was optional. If it was not
passed, the one in the LLVMContext was used.

That is probably not a pattern we want to follow. If each area has an
optional callback, there is a sea of callbacks and it is hard to follow
which one is called.

Doing this also found cases where the callback is a nice addition, like
testing that no errors or warnings are reported.

The other option is to always use the diagnostic handler in the
LLVMContext. That has a few problems

* To implement the C API we would have to set the diag handler and then
  set it back to the original value.
* Code that creates the context might be far away from code that wants
  the diagnostics.

I do have a patch that implements the second option and will send that as
an RFC.

llvm-svn: 254777
2015-12-04 22:08:53 +00:00
Rafael Espindola 0e309fe860 Use references now that it is natural to do so.
The linker never takes ownership of a module or changes which module it
is refering to, making it natural to use references.

llvm-svn: 254449
2015-12-01 19:50:54 +00:00