For namespaces, this is consistent with mangling and GCC's debug info
behavior. For structs, GCC uses <anonymous struct> but we prefer
consistency between all anonymous entities but don't want to confuse
them with template arguments, etc, so we'll just go with parens in all
cases.
llvm-svn: 205398
While investigating some debug info issues, Eric and I came across a
particular template case where the location of a decl was quite
different from the range of the same decl. It might've been rather
helpful if the dumper had actually showed us this.
llvm-svn: 205396
LinkingContext and InputGraph are unnecessarily entangled. Most linker
input file data, e.g. the vector containing input files, the next index
of the input file, etc. are managed by InputGraph, but only the current
input file is for no obvious reason managed by LinkingContext.
This patch is to move code from LinkingContext to InputGraph to fix it.
It's now clear who's reponsible for managing input file state, which is
InputGraph, and LinkingContext is now free from that responsibility.
It improves the readability as we now have fewer dependencies between
classes. No functionality change.
Differential Revision: http://llvm-reviews.chandlerc.com/D3259
llvm-svn: 205394
and isTargetCygwin() to isTargetWindowsCygwin() to be consistent with the
four Windows environments in Triple.h.
Suggestion by Saleem Abdulrasool!
llvm-svn: 205393
This is a mechanical change addressing the various sign comparison warnings that
are identified by both clang and gcc. This helps cleanup some of the warning
spew that occurs during builds.
llvm-svn: 205390
For the purpose of calculating the cost of the loop at various vectorization
factors, we need to count dependencies of consecutive pointers as uniforms
(which means that the VF = 1 cost is used for all overall VF values).
For example, the TSVC benchmark function s173 has:
...
%3 = add nsw i64 %indvars.iv, 16000
%arrayidx8 = getelementptr inbounds %struct.GlobalData* @global_data, i64 0, i32 0, i64 %3
...
and we must realize that the add will be a scalar in order to correctly deduce
it to be profitable to vectorize this on PowerPC with VSX enabled. In fact, all
dependencies of a consecutive pointer must be a scalar (uniform), and so we
simply need to add all consecutive pointers to the worklist that currently
detects collects uniforms.
Fixes PR19296.
llvm-svn: 205387
I'm not sure the comment in the implementation really adds a lot of
value (it's clear that we emit zero when no symbol is provided, but it
doesn't explain why we would do that). Happy to iterate.
llvm-svn: 205386
This removes the magic-number-esque code creating/retrieving the same
label for a debug_loc entry from two places and removes the last small
piece of reusable logic from emitDebugLoc so that there will be less
duplication when refactoring it into two functions (one for debug_loc,
the other for debug_loc.dwo).
llvm-svn: 205382
You can either provide the function name, or function body text.
Also propagate the compilation error up from where it is checked so we can report compilation errors.
<rdar://problem/9898371>
llvm-svn: 205380
insertElementAt()'s third parameter is not only unused but also ignored
if you pass Position::END. The actual meaning of the parameter was obscure.
Differential Revision: http://llvm-reviews.chandlerc.com/D3256
llvm-svn: 205376
Group class is designed for GNU LD's --start-group and --end-group. There's
no obvious need to have two classes for it -- one as an abstract base class
and the other as a concrete class.
llvm-svn: 205375
Seems we didn't have any test coverage for merging... awesome. So I
added some - but hit an llvm-objdump bug while I was there. I'm choosing
not to shave that yak right now.
Code review feedback/bug catch by Adrian Prantl in r205360.
llvm-svn: 205373
We don't want to encourage the code to emit a lexical block for
a function that needs one in order for the line table to change,
we need to grab the line information from the body of the pattern
that we were instantiated from, this code should do that.
Modify the test case to ensure that we're still looking in the
right place for all of the scopes and also that we haven't
created a lexical block where we didn't need one.
llvm-svn: 205368
No test case (this would invoke UB by examining uninitialized members,
etc, at best - and this code is apparently untested anyway - I'm about
to fix that)
Code review feedback from Adrian Prantl on r205360.
llvm-svn: 205367
InputGraph has too many knobs and controls that are not being used. This
patch is to remove dead code, unused features and a class. There are two
things that worth noting, besides simple dead code removal:
1. ControlNode class is removed. We had it as the base class of Group
class, but it provides no functionality particularly meaningful. We now
have shallower class hierarchy that is easier to understand.
2. InputGraph provides a feature to replace a node with its internal data.
It is being used to "expand" some type of node, such as a Linker Script
node, with its actual files. We used to have two options when replacing
it -- ExpandOnly or ExpandAndReplace. ExpandOnly was to expand it but not
remove the node from the tree. There is no use of that option in the code,
so it was a dead feature.
Differential Revision: http://llvm-reviews.chandlerc.com/D3252
llvm-svn: 205363
Asserting with cast<T> did not actually make much sense because there was no
need to use dynamic casting in the first place. We could make the compiler to
statically type check these objects.
llvm-svn: 205350
This provides an initial implementation of getUnrollingPreferences for x86.
getUnrollingPreferences is used by the generic (concatenation) unroller, which
is distinct from the unrolling done by the loop vectorizer. Many modern x86
cores have some kind of uop cache and loop-stream detector (LSD) used to
efficiently dispatch small loops, and taking full advantage of this requires
unrolling small loops (small here means 10s of uops).
These caches also have limits on the number of taken branches in the loop, and
so we also cap the loop unrolling factor based on the maximum "depth" of the
loop. This is currently calculated with a partial DFS traversal (partial
because it will stop early if the path length grows too much). This is still an
approximation, and one that is both conservative (because it does not account
for branches eliminated via block placement) and optimistic (because it is only
recording the maximum depth over minimum paths). Nevertheless, because the
loops that fit in these uop caches are so small, it is not clear how much the
details matter.
The original set of patches posted for review produced the following test-suite
performance results (from the TSVC benchmark) at that time:
ControlLoops-dbl - 13% speedup
ControlLoops-flt - 15% speedup
Reductions-dbl - 7.5% speedup
llvm-svn: 205348
In preparation for an upcoming commit implementing unrolling preferences for
x86, this adds additional fields to the UnrollingPreferences structure:
- PartialThreshold and PartialOptSizeThreshold - Like Threshold and
OptSizeThreshold, but used when not fully unrolling. These are necessary
because we need different thresholds for full unrolling from those used when
partially unrolling (the full unrolling thresholds are generally going to be
larger).
- MaxCount - A cap on the unrolling factor when partially unrolling. This can
be used by a target to prevent the unrolled loop from exceeding some
resource limit independent of the loop size (such as number of branches).
There should be no functionality change for any in-tree targets.
llvm-svn: 205347
