This has been causing the prologue_end to be incorrectly positioned.
Patch by Vladimir Radosavljevic.
Differential Revision: http://reviews.llvm.org/D11293
llvm-svn: 246309
When combiner AA is enabled, look at stores on the same chain.
Non-aliasing stores are moved to the same chain so the existing
code fails because it expects to find an adajcent store on a consecutive
chain.
Because of how DAGCombiner tries these store combines,
MergeConsecutiveStores doesn't see the correct set of stores on the chain
when it visits the other stores. Each store individually has its chain
fixed before trying to merge consecutive stores, and then tries to merge
stores from that point before the other stores have been processed to
have their chains fixed. To fix this, attempt to use FindBetterChain
on any possibly neighboring stores in visitSTORE.
Suppose you have 4 32-bit stores that should be merged into 1 vector
store. One store would be visited first, fixing the chain. What happens is
because not all of the store chains have yet been fixed, 2 of the stores
are merged. The other 2 stores later have their chains fixed,
but because the other stores were already merged, they have different
memory types and merging the two different sized stores is not
supported and would be more difficult to handle.
llvm-svn: 246307
Summary:
This patch removes two remaining places where pointer value comparisons
are used to order functions: comparing range annotation metadata, and comparing
block address constants. (These are both rare cases, and so no actual
non-determinism was observed from either case).
The fix for range metadata is simple: the annotation always consists of a pair
of integers, so we just order by those integers.
The fix for block addresses is more subtle. Two constants are the same if they
are the same basic block in the same function, or if they refer to corresponding
basic blocks in each respective function. Note that in the first case, merging
is trivially correct. In the second, the correctness of merging relies on the
fact that the the values of block addresses cannot be compared. This change is
actually an enhancement, as these functions could not previously be merged (see
merge-block-address.ll).
There is still a problem with cross function block addresses, in that constants
pointing to a basic block in a merged function is not updated.
This also more robustly compares floating point constants by all fields of their
semantics, and fixes a dyn_cast/cast mixup.
Author: jrkoenig
Reviewers: dschuff, nlewycky, jfb
Subscribers llvm-commits
Differential revision: http://reviews.llvm.org/D12376
llvm-svn: 246305
A readnone tailcall may still have a chain of computation which follows
it that would invalidate a tailcall lowering. Don't skip the analysis
in such cases.
This fixes PR24613.
llvm-svn: 246304
lib.exe has a feature to create import library files (which contain
short import files) from module-definition files. Previously, we were
using that feature, but it turned out that the feature is not complete
for us.
There seems no way to specify "Import Types" in module-definition file.
lib.exe always adds "_" to given symbols and specify IMPORT_NAME_UNDECORATE.
We need more fine-grainded control on that value.
This patch teaches LLD to create short import files itself.
We are still using lib.exe, but the use of the tool is limited to create
empty import library files. We then create short import files and add them
to the empty files as new members.
This patch does not intend to change the functionality. LLD produces
the same import libraries as before. I'll make another change to create
different import libraries in a follow-up patch.
llvm-svn: 246292
handle more allocas with loads past the end of the alloca.
I suspect there are some related crashers with slightly different
patterns, but I'll fix those and add test cases as I find them.
Thanks to David Majnemer for the excellent test case reduction here.
Made this super simple to debug and fix.
llvm-svn: 246289
For external users, the memory locations into which we generate scalar values
may be of interest. This change introduces two functions that allow to obtain
(or create) the AllocInsts for a given BasePointer.
We use this change to simplify the code in BlockGenerators.
llvm-svn: 246285
This patch includes a fix for a llvm-readobj test. With this patch,
the tool does no longer print out COFF headers for the short import
file, but that's probably desirable because the header for the short
import file is dummy.
llvm-svn: 246283
This allows users to extend the IslNodeBuilder to create their own optimization
passes. This feature is not used in Polly's codebase itself, but as these
funtions are not performance critical, the cost of making experiments of
external users easier seems low enough to do so.
llvm-svn: 246281
Added codegen for array section in 'depend' clause of 'task' directive. It emits to pointers, one for the begin of array section and another for the end of array section. Size of the section is calculated as (end + 1 - start) * sizeof(basic_element_type).
llvm-svn: 246278
COFF short import files are special kind of files that contains only
DLL-exported symbol names. That's different from object files because
it has no data except symbol names.
This change implements a SymbolicFile interface for the short import
files so that symbol names can be accessed through that interface.
llvm-ar is now able to read the file and create symbol table entries
for short import files.
llvm-svn: 246276
After months of work there are only 4 tests still failing in C++03.
This patch fixes those tests.
All of the libc++ builders should be green.
llvm-svn: 246275
Summary:
This patch properly constrains the converting assignment operator in C++03. It also fixes a bug where std::forward was given the wrong type.
The following two tests begin passing in C++03:
* `unique_ptr.single.asgn/move_convert.pass.cpp`
* `unique_ptr.single.asgn/move_convert13.fail.cpp`
Reviewers: mclow.lists
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D12173
llvm-svn: 246272
Summary:
This patch marks *most* tests for `std::promise`, `std::future` and `std::shared_future` as unsupported in C++03. These tests fail in C++03 mode because they attempt to copy a `std::future` even though it is a `MoveOnly` type. AFAIK the missing move-semantics in `std::future` is the only reason these tests fail but without move semantics these classes are useless. For example even though `std::promise::set_value` and `std::promise::set_exception(...)` work in C++03 `std::promise` is still useless because we cannot call `std::promise::get_future(...)`.
It might be possible to hack `std::move(...)` like we do for `std::unique_ptr` to make the move semantics work but I don't think it is worth the effort. Instead I think we should leave the `<future>` header as-is and mark the failing tests as `UNSUPPORTED`. I don't believe there are any users of `std::future` or `std::promise` in C++03 because they are so unusable. Therefore I am not concerned about losing test coverage and possibly breaking users. However because there are still parts of `<future>` that work in C++03 it would be wrong to `#ifdef` out the entire header.
@mclow.lists Should we take further steps to prevent the use of `std::promise`, `std::future` and `std::shared_future` in C++03?
Note: This patch also cleans up the tests and converts them to use `support/test_allocator.h` instead of a duplicate class in `test/std/futures/test_allocator.h`.
Reviewers: mclow.lists
Subscribers: vsk, mclow.lists, cfe-commits
Differential Revision: http://reviews.llvm.org/D12135
llvm-svn: 246271
Summary:
This patch optimizes basic_string::compare to use strcmp when the default char_traits has been given.
See PR19900 for more information. https://llvm.org/bugs/show_bug.cgi?id=19900
Reviewers: mclow.lists
Subscribers: bkramer, cfe-commits
Differential Revision: http://reviews.llvm.org/D12355
llvm-svn: 246266
InputFiles.h:98:53: error: invalid use of incomplete type ‘class lld::elf2::SymbolBody’
return SymbolBodies[SymbolIndex - FirstNonLocal]->getReplacement();
llvm-svn: 246262
For targets that didn't support this, this will let us respect the
langref instead of failing to select.
Note that we don't need to change the 32-bit x86/PPC lowerings (to
account for the result type/# difference) because they're both
custom and bypass type legalization.
llvm-svn: 246258
Petar Jovanovic