removes cast by performing the lshr on smaller types. However, currently there
is no trunc(lshr (sext A), Cst) variant.
This patch add such optimization by transforming trunc(lshr (sext A), Cst)
to ashr A, Cst.
Differential Revision: http://reviews.llvm.org/D12520
llvm-svn: 247271
GlobalsAA must by definition be preserved in function passes, but the passmanager doesn't know that. Make each pass explicitly preserve GlobalsAA.
llvm-svn: 247263
The assertion was weaker than it should be and gave the impression we're growing the number of base defining values being considered during the fixed point interation. That's not true. The tighter form of the assert is useful documentation.
llvm-svn: 247221
This change is simply enhancing the existing inference algorithm to handle insertelement instructions by conservatively inserting a new instruction to propagate the vector of associated base pointers. In the process, I'm ripping out the peephole optimizations which mostly helped cover the fact this hadn't been done.
Note that most of the newly inserted nodes will be nearly immediately removed by the post insertion optimization pass introduced in 246718. Arguably, we should be trying harder to avoid the malloc traffic here, but I'd rather get the code correct, then worry about compile time.
Unlike previous extensions of the algorithm to handle more case, I discovered the existing code was causing miscompiles in some cases. In particular, we had an implicit assumption that the peephole covered *all* insert element instructions, so if we had a value directly based on a insert element the peephole didn't cover, we proceeded as if it were a base anyways. Not good. I believe we had the same issue with shufflevector which is why I adjusted the predicate for them as well.
Differential Revision: http://reviews.llvm.org/D12583
llvm-svn: 247210
Previously, the base pointer algorithm wasn't deterministic. The core fixed point was (of course), but we were inserting new nodes and optimizing them in an order which was unspecified and variable. We'd somewhat hacked around this for testing by sorting by value name, but that doesn't solve the general determinism problem.
Instead, we can use the order of traversal over the def/use graph to give us a single consistent ordering. Today, this is a DFS order, but the exact order doesn't mater provided it's deterministic for a given input.
(Q: It is safe to rely on a deterministic order of operands right?)
Note that this only fixes the determinism within a single inference step. The inference step is currently invoked many times in a non-deterministic order. That's a future change in the sequence. :)
Differential Revision: http://reviews.llvm.org/D12640
llvm-svn: 247208
Visit disjoint sets in a deterministic order based on the maximum BitSetNM
index, otherwise the order in which we visit them will depend on pointer
comparisons. This was being exposed by MSan.
llvm-svn: 247201
with the new pass manager, and no longer relying on analysis groups.
This builds essentially a ground-up new AA infrastructure stack for
LLVM. The core ideas are the same that are used throughout the new pass
manager: type erased polymorphism and direct composition. The design is
as follows:
- FunctionAAResults is a type-erasing alias analysis results aggregation
interface to walk a single query across a range of results from
different alias analyses. Currently this is function-specific as we
always assume that aliasing queries are *within* a function.
- AAResultBase is a CRTP utility providing stub implementations of
various parts of the alias analysis result concept, notably in several
cases in terms of other more general parts of the interface. This can
be used to implement only a narrow part of the interface rather than
the entire interface. This isn't really ideal, this logic should be
hoisted into FunctionAAResults as currently it will cause
a significant amount of redundant work, but it faithfully models the
behavior of the prior infrastructure.
- All the alias analysis passes are ported to be wrapper passes for the
legacy PM and new-style analysis passes for the new PM with a shared
result object. In some cases (most notably CFL), this is an extremely
naive approach that we should revisit when we can specialize for the
new pass manager.
- BasicAA has been restructured to reflect that it is much more
fundamentally a function analysis because it uses dominator trees and
loop info that need to be constructed for each function.
All of the references to getting alias analysis results have been
updated to use the new aggregation interface. All the preservation and
other pass management code has been updated accordingly.
The way the FunctionAAResultsWrapperPass works is to detect the
available alias analyses when run, and add them to the results object.
This means that we should be able to continue to respect when various
passes are added to the pipeline, for example adding CFL or adding TBAA
passes should just cause their results to be available and to get folded
into this. The exception to this rule is BasicAA which really needs to
be a function pass due to using dominator trees and loop info. As
a consequence, the FunctionAAResultsWrapperPass directly depends on
BasicAA and always includes it in the aggregation.
This has significant implications for preserving analyses. Generally,
most passes shouldn't bother preserving FunctionAAResultsWrapperPass
because rebuilding the results just updates the set of known AA passes.
The exception to this rule are LoopPass instances which need to preserve
all the function analyses that the loop pass manager will end up
needing. This means preserving both BasicAAWrapperPass and the
aggregating FunctionAAResultsWrapperPass.
Now, when preserving an alias analysis, you do so by directly preserving
that analysis. This is only necessary for non-immutable-pass-provided
alias analyses though, and there are only three of interest: BasicAA,
GlobalsAA (formerly GlobalsModRef), and SCEVAA. Usually BasicAA is
preserved when needed because it (like DominatorTree and LoopInfo) is
marked as a CFG-only pass. I've expanded GlobalsAA into the preserved
set everywhere we previously were preserving all of AliasAnalysis, and
I've added SCEVAA in the intersection of that with where we preserve
SCEV itself.
One significant challenge to all of this is that the CGSCC passes were
actually using the alias analysis implementations by taking advantage of
a pretty amazing set of loop holes in the old pass manager's analysis
management code which allowed analysis groups to slide through in many
cases. Moving away from analysis groups makes this problem much more
obvious. To fix it, I've leveraged the flexibility the design of the new
PM components provides to just directly construct the relevant alias
analyses for the relevant functions in the IPO passes that need them.
This is a bit hacky, but should go away with the new pass manager, and
is already in many ways cleaner than the prior state.
Another significant challenge is that various facilities of the old
alias analysis infrastructure just don't fit any more. The most
significant of these is the alias analysis 'counter' pass. That pass
relied on the ability to snoop on AA queries at different points in the
analysis group chain. Instead, I'm planning to build printing
functionality directly into the aggregation layer. I've not included
that in this patch merely to keep it smaller.
Note that all of this needs a nearly complete rewrite of the AA
documentation. I'm planning to do that, but I'd like to make sure the
new design settles, and to flesh out a bit more of what it looks like in
the new pass manager first.
Differential Revision: http://reviews.llvm.org/D12080
llvm-svn: 247167
Predicating stores requires creating extra blocks. It's much cleaner if we do this in one pass instead of mutating the CFG while writing vector instructions.
Besides which we can make use of helper functions to update domtree for us, reducing the work we need to do.
llvm-svn: 247139
This change extends the bitset lowering pass to support bitsets that may
contain either functions or global variables. A function bitset is lowered to
a jump table that is laid out before one of the functions in the bitset.
Also add support for non-string bitset identifier names. This allows for
distinct metadata nodes to stand in for names with internal linkage,
as done in D11857.
Differential Revision: http://reviews.llvm.org/D11856
llvm-svn: 247080
removes cast by performing the lshr on smaller types. However, currently there
is no trunc(lshr (sext A), Cst) variant.
This patch add such optimization by transforming trunc(lshr (sext A), Cst)
to ashr A, Cst.
Differential Revision: http://reviews.llvm.org/D12520
llvm-svn: 246997
Trivial multiplication by zero may survive the worklist. We tried to
reassociate the multiplication with a division instruction, causing us
to divide by zero; bail out instead.
This fixes PR24726.
llvm-svn: 246939
As a first step towards a new implementation of the base pointer inference algorithm, introduce an abstraction for BDVs, strengthen the assertions around them, and rewrite the BDV relation code in terms of the abstraction which includes an explicit notion of whether the BDV is also a base. The later is motivated by the fact we had a bug where insertelement was always assumed to be a base pointer even though the BDV code knew it wasn't. The strengthened assertions in this patch would have caught that bug.
The next step will be to separate the DefiningValueMap into a BDV use list cache (entirely within findBasePointers) and a base pointer cache. Having the former will allow me to use a deterministic visit order when visiting BDVs in the inference algorithm and remove a bunch of ordering related hacks. Before actually doing the last step, I'm likely going to extend the lattice with a 'BaseN' (seen only base inputs) state so that I can kill the post process optimization step.
Phabricator Revision: http://reviews.llvm.org/D12608
llvm-svn: 246809
The visit order being used in the base pointer inference algorithm is currently non-deterministic. When working on http://reviews.llvm.org/D12583, I discovered that we were relying on a peephole optimization to get deterministic ordering in one of the test cases.
This change is intented to let me test and land http://reviews.llvm.org/D12583. The current code will not be long lived. I'm starting to investigate a rewrite of the algorithm which will combine the post-process step into the initial algorithm and make the visit order determistic. Before doing that, I wanted to make sure the existing code was complete and the test were stable. Hopefully, patches should be up for review for the new algorithm this week or early next.
llvm-svn: 246801
Summary:
Add a `cleanupendpad` instruction, used to mark exceptional exits out of
cleanups (for languages/targets that can abort a cleanup with another
exception). The `cleanupendpad` instruction is similar to the `catchendpad`
instruction in that it is an EH pad which is the target of unwind edges in
the handler and which itself has an unwind edge to the next EH action.
The `cleanupendpad` instruction, similar to `cleanupret` has a `cleanuppad`
argument indicating which cleanup it exits. The unwind successors of a
`cleanuppad`'s `cleanupendpad`s must agree with each other and with its
`cleanupret`s.
Update WinEHPrepare (and docs/tests) to accomodate `cleanupendpad`.
Reviewers: rnk, andrew.w.kaylor, majnemer
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D12433
llvm-svn: 246751
Summary:
This patch introduces a side table in Merge Functions to
efficiently remove functions from the function set when functions
they refer to are merged. Previously these functions would need to
be compared lg(N) times to find the appropriate FunctionNode in the
tree to defer. With the recent determinism changes, this comparison
is more expensive. In addition, the removal function would not always
actually remove the function from the set (i.e. after remove(F),
there would sometimes still be a node in the tree which contains F).
With these changes, these functions are properly deferred, and so more
functions can be merged. In addition, when there are many merged
functions (and thus more deferred functions), there is a speedup:
chromium: 48678 merged -> 49380 merged; 6.58s -> 5.49s
libxul.so: 41004 merged -> 41030 merged; 8.02s -> 6.94s
mysqld: 1607 merged -> 1607 merged (same); 0.215s -> 0.212s (probably noise)
Author: jrkoenig
Reviewers: jfb, dschuff
Subscribers: llvm-commits, nlewycky
Differential revision: http://reviews.llvm.org/D12537
llvm-svn: 246735
Fix a bug in change 246133. I didn't handle the case where we had a cycle in the use graph and could add an instruction we were about to erase back on to the worklist. Oddly, I have not been able to write a small test case for this, even with the AssertingVH added. I have confirmed the basic theory for the fix on a large failing example, but all attempts to reduce that to something appropriate for a test case have failed.
Differential Revision: http://reviews.llvm.org/D12575
llvm-svn: 246718
There was infinite loop because it was trying to change assume(true) into
assume(true)
Also added handling when assume(false) appear
http://reviews.llvm.org/D12516
llvm-svn: 246697
After hitting @llvm.assume(X) we can:
- propagate equality that X == true
- if X is icmp/fcmp (with eq operation), and one of operand
is constant we can change all variables with constants in the same BasicBlock
http://reviews.llvm.org/D11918
llvm-svn: 246695
This makes RemoveDuplicatePHINodes more effective and fixes an assertion
failure. Triggering the assertions requires a DenseSet reallocation
so this change only contains a constructive test.
I'll explain the issue with a small example. In the following function
there's a duplicate PHI, %4 and %5 are identical. When this is found
the DenseSet in RemoveDuplicatePHINodes contains %2, %3 and %4.
define void @F() {
br label %1
; <label>:1 ; preds = %1, %0
%2 = phi i32 [ 42, %0 ], [ %4, %1 ]
%3 = phi i32 [ 42, %0 ], [ %5, %1 ]
%4 = phi i32 [ 42, %0 ], [ 23, %1 ]
%5 = phi i32 [ 42, %0 ], [ 23, %1 ]
br label %1
}
after RemoveDuplicatePHINodes runs the function looks like this. %3 has
changed and is now identical to %2, but RemoveDuplicatePHINodes never
saw this.
define void @F() {
br label %1
; <label>:1 ; preds = %1, %0
%2 = phi i32 [ 42, %0 ], [ %4, %1 ]
%3 = phi i32 [ 42, %0 ], [ %4, %1 ]
%4 = phi i32 [ 42, %0 ], [ 23, %1 ]
br label %1
}
If the DenseSet does a reallocation now it will reinsert all
keys and stumble over %3 now having a different hash value than it had
when inserted into the map for the first time. This change clears the
set whenever a PHI is deleted and starts the progress from the
beginning, allowing %3 to be deleted and avoiding inconsistent DenseSet
state. This potentially has a negative performance impact because
it rescans all PHIs, but I don't think that this ever makes a difference
in practice.
llvm-svn: 246694
We were bailing to two places if our runtime checks failed. If the initial overflow check failed, we'd go to ScalarPH. If any other check failed, we'd go to MiddleBlock. This caused us to have to have an extra PHI per induction and reduction as the vector loop's exit block was not dominated by its latch.
There's no need to have this behavior - if we just always go to ScalarPH we can get rid of a bunch of complexity.
llvm-svn: 246637
This reduces the complexity of createEmptyBlock() and will open the door to further refactoring.
The test change is simply because we're now constant folding a trivial test.
llvm-svn: 246634
It makes things easier to understand if this is in a helper method. This is part of my ongoing spaghetti-removal operation on createEmptyLoop.
llvm-svn: 246632
There's no need to widen canonical induction variables. It's just as efficient to create a *new*, wide, induction variable.
Consider, if we widen an indvar, then we'll have to truncate it before its uses anyway (1 trunc). If we create a new indvar instead, we'll have to truncate that instead (1 trunc) [besides which IndVars should go and clean up our mess after us anyway on principle].
This lets us remove a ton of special-casing code.
llvm-svn: 246631
Vectorized loops only ever have one induction variable. All induction PHIs from the scalar loop are rewritten to be in terms of this single indvar.
We were trying very hard to pick an indvar that already existed, even if that indvar wasn't canonical (didn't start at zero). But trying so hard is really fruitless - creating a new, canonical, indvar only results in one extra add in the worst case and that add is trivially easy to push through the PHI out of the loop by instcombine.
If we try and be less clever here and instead let instcombine clean up our mess (as we do in many other places in LV), we can remove unneeded complexity.
llvm-svn: 246630
Teach FunctionAttr to infer the nonnull attribute on return values of functions which never return a potentially null value. This is done both via a conservative local analysis for the function itself and a optimistic per-SCC analysis. If no function in the SCC returns anything which could be null (other than values from other functions in the SCC), we can conclude no function returned a null pointer. Even if some function within the SCC returns a null pointer, we may be able to locally conclude that some don't.
Differential Revision: http://reviews.llvm.org/D9688
llvm-svn: 246476
Summary:
JumpThreading shouldn't duplicate a convergent call, because that would move a convergent call into a control-inequivalent location. For example,
if (cond) {
...
} else {
...
}
convergent_call();
if (cond) {
...
} else {
...
}
should not be optimized to
if (cond) {
...
convergent_call();
...
} else {
...
convergent_call();
...
}
Test Plan: test/Transforms/JumpThreading/basic.ll
Patch by Xuetian Weng.
Reviewers: resistor, arsenm, jingyue
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D12484
llvm-svn: 246415
PR24605 is caused due to an incorrect insert point in instcombine's IR
builder. When simplifying
%t = add X Y
...
%m = icmp ... %t
the replacement for %t should be placed before %t, not before %m, as
there could be a use of %t between %t and %m.
llvm-svn: 246315
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
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
After hitting @llvm.assume(X) we can:
- propagate equality that X == true
- if X is icmp/fcmp (with eq operation), and one of operand
is constant we can change all variables with constants in the same BasicBlock
http://reviews.llvm.org/D11918
llvm-svn: 246243
This patch changes the analysis diagnostics produced when loops with
floating-point recurrences or memory operations are identified. The new messages
say "cannot prove it is safe to reorder * operations; allow reordering by
specifying #pragma clang loop vectorize(enable)". Depending on the type of
diagnostic the message will include additional options such as ffast-math or
__restrict__.
This patch also allows the vectorize(enable) pragma to override the low pointer
memory check threshold. When the hint is given a higher threshold is used.
See the clang patch for the options produced for each diagnostic.
llvm-svn: 246187
Unlike scalar operations, we can perform vector operations on element types that
are smaller than the native integer types. We type-promote scalar operations if
they are smaller than a native type (e.g., i8 arithmetic is promoted to i32
arithmetic on Arm targets). This patch detects and removes type-promotions
within the reduction detection framework, enabling the vectorization of small
size reductions.
In the legality phase, we look through the ANDs and extensions that InstCombine
creates during promotion, keeping track of the smaller type. In the
profitability phase, we use the smaller type and ignore the ANDs and extensions
in the cost model. Finally, in the code generation phase, we truncate the result
of the reduction to allow InstCombine to rewrite the entire expression in the
smaller type.
This fixes PR21369.
http://reviews.llvm.org/D12202
Patch by Matt Simpson <mssimpso@codeaurora.org>!
llvm-svn: 246149
... and move it into LoopUtils where it can be used by other passes, just like ReductionDescriptor. The API is very similar to ReductionDescriptor - that is, not very nice at all. Sorting these both out will come in a followup.
NFC
llvm-svn: 246145
A release fence acts as a publication barrier for stores within the current thread to become visible to other threads which might observe the release fence. It does not require the current thread to observe stores performed on other threads. As a result, we can allow store-load and load-store forwarding across a release fence.
We do need to make sure that stores before the fence can't be eliminated even if there's another store to the same location after the fence. In theory, we could reorder the second store above the fence and *then* eliminate the former, but we can't do this if the stores are on opposite sides of the fence.
Note: While more aggressive then what's there, this patch is still implementing a really conservative ordering. In particular, I'm not trying to exploit undefined behavior via races, or the fact that the LangRef says only 'atomic' accesses are ordered w.r.t. fences.
Differential Revision: http://reviews.llvm.org/D11434
llvm-svn: 246134
When computing base pointers, we introduce new instructions to propagate the base of existing instructions which might not be bases. However, the algorithm doesn't make any effort to recognize when the new instruction to be inserted is the same as an existing one already in the IR. Since this is happening immediately before rewriting, we don't really have a chance to fix it after the pass runs without teaching loop passes about statepoints.
I'm really not thrilled with this patch. I've rewritten it 4 different ways now, but this is the best I've come up with. The case where the new instruction is just the original base defining value could be merged into the existing algorithm with some complexity. The problem is that we might have something like an extractelement from a phi of two vectors. It may be trivially obvious that the base of the 0th element is an existing instruction, but I can't see how to make the algorithm itself figure that out. Thus, I resort to the call to SimplifyInstruction instead.
Note that we can only adjust the instructions we've inserted ourselves. The live sets are still being tracked in side structures at this point in the code. We can't easily muck with instructions which might be in them. Long term, I'm really thinking we need to materialize the live pointer sets explicitly in the IR somehow rather than using side structures to track them.
Differential Revision: http://reviews.llvm.org/D12004
llvm-svn: 246133
This patch ensures that every analysis diagnostic produced by the vectorizer
will be printed if the loop has a vectorization hint on it. The condition has
also been improved to prevent printing when a disabling hint is specified.
llvm-svn: 246132
As Sanjoy pointed out over in http://reviews.llvm.org/D11819, a switch on an icmp should always be able to become a branch instruction. This patch generalizes that notion slightly to prove that the default case of a switch is unreachable if the cases completely cover all possible bit patterns in the condition. Once that's done, the switch to branch conversion kicks in just fine.
Note: Duplicate case values are disallowed by the LangRef and verifier.
Differential Revision: http://reviews.llvm.org/D11995
llvm-svn: 246125
The problem here were the function analyses invoked by the function pass
manager from the new IPO pass. I looked at other IPO passes needing
dominance information and the only one that requires it (partial
inliner) does not use the standard dependency mechanism.
This patch mimics what the partial inliner does to compute dominance,
post-dominance and loop info. One thing I like about this approach is
that I can delay the computation of all this until I actually need it.
This should bring the ASAN buildbot back to green. If there's a better
way to fix this, I'll do it in a follow-up patch.
llvm-svn: 246066
This was only added to preserve the old ScalarRepl's use of SSAUpdater
which was originally to avoid use of dominance frontiers. Now, we only
need a domtree, and we'll need a domtree right after this pass as well
and so it makes perfect sense to always and only use the dom-tree
powered mem2reg. This was flag-flipper earlier and has stuck reasonably
so I wanted to gut the now-dead code out of SROA before we waste more
time with it. Among other things, this will make passmanager porting
easier.
llvm-svn: 246028
Summary: When comparing basic blocks, there is an additional check that two Value*'s should have the same ID, which interferes with merging equivalent constants of different kinds (such as a ConstantInt and a ConstantPointerNull in the included testcase). The cmpValues function already ensures that the two values in each function are the same, so removing this check should not cause incorrect merging.
Also, the type comparison is redundant, based on reviewing the code and testing on the test suite and several large LTO bitcodes.
Author: jrkoenig
Reviewers: nlewycky, jfb, dschuff
Subscribers: llvm-commits
Differential revision: http://reviews.llvm.org/D12302
llvm-svn: 246001
Summary:
This change makes the variable argument intrinsics, `llvm.va_start` and
`llvm.va_copy`, and the `va_arg` instruction behave as they do on Windows
inside a `CallingConv::X86_64_Win64` function. It's needed for a Clang patch
I have to add support for GCC's `__builtin_ms_va_list` constructs.
Reviewers: nadav, asl, eugenis
CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D1622
llvm-svn: 245990
Extend signed relational comparison instrumentation with a special
case for comparisons with -1. This fixes an MSan false positive when
such comparison is used as a sign bit test.
https://llvm.org/bugs/show_bug.cgi?id=24561
llvm-svn: 245980
The loop minimum iterations check below ensures the loop has enough trip count so the generated
vector loop will likely be executed, and it covers the overflow check.
Differential Revision: http://reviews.llvm.org/D12107.
llvm-svn: 245952
Eventually, we will need sample profiles to be incorporated into the
inliner's cost models. To do this, we need the sample profile pass to
be a module pass.
This patch makes no functional changes beyond the mechanical adjustments
needed to run SampleProfile as a module pass.
llvm-svn: 245940
It doesn't solve the problem, when for example we load something, and
then assume that it is the same as some constant value, because
globalopt will fail on unknown load instruction. The proposed solution
would be to skip some instructions that we can't evaluate and they are
safe to skip (f.e. load, assume and many others) and see if they are
required to perform optimization (f.e. we don't care about ephemeral
instructions that may appear using @llvm.assume())
http://reviews.llvm.org/D12266
llvm-svn: 245919
This patch adds support for dfsan on aarch64-linux with 42-bit VMA
(current default config for 64K pagesize kernels). The support is
enabled by defining the SANITIZER_AARCH64_VMA to 42 at build time
for both clang/llvm and compiler-rt. The default VMA is 39 bits.
llvm-svn: 245840
TL-DR: SROA is followed by EarlyCSE which requires the DominatorTree.
There is no reason not to require it up-front for SROA.
Some history is necessary to understand why we ended-up here.
r123437 switched the second (Legacy)SROA in the optimizer pipeline to
use SSAUpdater in order to avoid recomputing the costly
DominanceFrontier. The purpose was to speed-up the compile-time.
Later r123609 removed the need for the DominanceFrontier in
(Legacy)SROA.
Right after, some cleanup was made in r123724 to remove any reference
to the DominanceFrontier. SROA existed in two flavors: SROA_SSAUp and
SROA_DT (the latter replacing SROA_DF).
The second argument of `createScalarReplAggregatesPass` was renamed
from `UseDomFrontier` to `UseDomTree`.
I believe this is were a mistake was made. The pipeline was not
updated and the call site was still:
PM->add(createScalarReplAggregatesPass(-1, false));
At that time, SROA was immediately followed in the pipeline by
EarlyCSE which required alread the DominatorTree. Not requiring
the DominatorTree in SROA didn't save anything, but unfortunately
it was lost at this point.
When the new SROA Pass was introduced in r163965, I believe the goal
was to have an exact replacement of the existing SROA, this bug
slipped through.
You can see currently:
$ echo "" | clang -x c++ -O3 -c - -mllvm -debug-pass=Structure
...
...
FunctionPass Manager
SROA
Dominator Tree Construction
Early CSE
After this patch:
$ echo "" | clang -x c++ -O3 -c - -mllvm -debug-pass=Structure
...
...
FunctionPass Manager
Dominator Tree Construction
SROA
Early CSE
This improves the compile time from 88s to 23s for PR17855.
https://llvm.org/bugs/show_bug.cgi?id=17855
And from 113s to 12s for PR16756
https://llvm.org/bugs/show_bug.cgi?id=16756
Reviewers: chandlerc
Differential Revision: http://reviews.llvm.org/D12267
From: Mehdi Amini <mehdi.amini@apple.com>
llvm-svn: 245820
Summary:
WinEHPrepare is going to require that cleanuppad and catchpad produce values
of token type which are consumed by any cleanupret or catchret exiting the
pad. This change updates the signatures of those operators to require/enforce
that the type produced by the pads is token type and that the rets have an
appropriate argument.
The catchpad argument of a `CatchReturnInst` must be a `CatchPadInst` (and
similarly for `CleanupReturnInst`/`CleanupPadInst`). To accommodate that
restriction, this change adds a notion of an operator constraint to both
LLParser and BitcodeReader, allowing appropriate sentinels to be constructed
for forward references and appropriate error messages to be emitted for
illegal inputs.
Also add a verifier rule (noted in LangRef) that a catchpad with a catchpad
predecessor must have no other predecessors; this ensures that WinEHPrepare
will see the expected linear relationship between sibling catches on the
same try.
Lastly, remove some superfluous/vestigial casts from instruction operand
setters operating on BasicBlocks.
Reviewers: rnk, majnemer
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D12108
llvm-svn: 245797
Summary:
Merge functions previously relied on unsigned comparisons of pointer values to
order functions. This caused observable non-determinism in the compiler for
large bitcode programs. Basically, opt -mergefuncs program.bc | md5sum produces
different hashes when run repeatedly on the same machine. Differing output was
observed on three large bitcodes, but it was less frequent on the smallest file.
It is possible that this only manifests on the large inputs, hence remaining
undetected until now.
This patch fixes this by removing (almost, see below) all places where
comparisons between pointers are used to order functions. Most of these changes
are local, but the comparison of global values requires assigning an identifier
to each local in the order it is visited. This is very similar to the way the
comparison function identifies Value*'s defined within a function. Because the
order of visiting the functions and their subparts is deterministic, the
identifiers assigned to the globals will be as well, and the order of functions
will be deterministic.
With these changes, there is no more observed non-determinism. There is also
only minor slowdowns (negligible to 4%) compared to the baseline, which is
likely a result of the fact that global comparisons involve hash lookups and not
just pointer comparisons.
The one caveat so far is that programs containing BlockAddress constants can
still be non-deterministic. It is not clear what the right solution is here. In
particular, even if the global numbers are used to order by function, we still
need a way to order the BasicBlock*'s. Unfortunately, we cannot just bail out
and fail to order the functions or consider them equal, because we require a
total order over functions. Note that programs with BlockAddress constants are
relatively rare, so the impact of leaving this in is minor as long as this pass
is opt-in.
Author: jrkoenig
Reviewers: nlewycky, jfb, dschuff
Subscribers: jevinskie, llvm-commits, chapuni
Differential revision: http://reviews.llvm.org/D12168
llvm-svn: 245762
The original checkin was buggy, this change has a fix.
Original commit message:
[InstCombine] Transform A & (L - 1) u< L --> L != 0
Summary:
This transform is never a pessimization at the IR level (since it
replaces an `icmp` with another), and has potentiall payoffs:
1. It may make the `icmp` fold away or become loop invariant.
2. It may make the `A & (L - 1)` computation dead.
This shows up in Java, in range checks generated by array accesses of
the form `a[i & (a.length - 1)]`.
Reviewers: reames, majnemer
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D12210
llvm-svn: 245753
Gets a bit tricky in the ValueMapper, of course - not sure if we should
just expose a list of explicit types for each Value so that the
ValueMapper can be neutral to these special cases (it's OK for things
like load, where the explicit type is the result type - but when that's
not the case, it means plumbing through another "special" type... )
llvm-svn: 245728
The module splitter splits a module into linkable partitions. It will
be used to implement parallel LTO code generation.
This initial version of the splitter does not attempt to deal with the
somewhat subtle symbol visibility issues around module splitting. These
will be dealt with in a future change.
Differential Revision: http://reviews.llvm.org/D12132
llvm-svn: 245662
Summary:
This transform is never a pessimization at the IR level (since it
replaces an `icmp` with another), and has potentiall payoffs:
1. It may make the `icmp` fold away or become loop invariant.
2. It may make the `A & (L - 1)` computation dead.
This shows up in Java, in range checks generated by array accesses of
the form `a[i & (a.length - 1)]`.
Reviewers: reames, majnemer
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D12210
llvm-svn: 245635
and make it always preserve debug locations, since all callers wanted this
behavior anyway.
This is addressing a post-commit review feedback for r245589.
NFC (inside the LLVM tree).
llvm-svn: 245622
This patch adds support for asan on aarch64-linux with 42-bit VMA
(current default config for 64K pagesize kernels). The support is
enabled by defining the SANITIZER_AARCH64_VMA to 42 at build time
for both clang/llvm and compiler-rt. The default VMA is 39 bits.
llvm-svn: 245594
Summary:
Refactor, NFC
Extracts computeOverflowForSignedAdd and isKnownNonNegative from NaryReassociate to ValueTracking in case
others need it.
Reviewers: reames
Subscribers: majnemer, llvm-commits
Differential Revision: http://reviews.llvm.org/D11313
llvm-svn: 245591
Instruction::dropUnknownMetadata(KnownSet) is supposed to preserve all
metadata in KnownSet, but the condition for DebugLocs was inverted.
Most users of dropUnknownMetadata() actually worked around this by not
adding LLVMContext::MD_dbg to their list of KnowIDs.
This is now made explicit.
llvm-svn: 245589
Caught by the famous "DebugLoc describes the currect SubProgram" assertion.
When GVN is removing a nonlocal load it updates the debug location of the
SSA value it replaced the load with with the one of the load. In the
testcase this actually overwrites a valid debug location with an empty one.
In reality GVN has to make an arbitrary choice between two equally valid
debug locations. This patch changes to behavior to only update the
location if the value doesn't already have a debug location.
llvm-svn: 245588
Since Ashutosh made findDefsUsedOutsideOfLoop public, we can clean this
up.
Now clients that don't compute DefsUsedOutsideOfLoop can just call
versionLoop() and computing DefsUsedOutsideOfLoop will happen
implicitly. With that there is no reason to expose addPHINodes anymore.
Ashutosh, you can now drop the calls to findDefsUsedOutsideOfLoop and
addPHINodes in LVerLICM and things should just work.
llvm-svn: 245579
Summary: We know that -x & 1 is equivalent to x & 1, avoid using negation for testing if a negative integer is even or odd.
Reviewers: majnemer
Subscribers: junbuml, mssimpso, gberry, mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D12156
llvm-svn: 245569
Usually DSE is not supposed to remove lifetime intrinsics, but it's
actually ok to remove them for dead objects in terminating blocks,
because they convey no extra information there. Until we hit a lifetime
start that cannot be removed, that is. Because from that point on the
lifetime intrinsics become interesting again, e.g. for stack coloring.
Reviewers: reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11710
llvm-svn: 245542
analyses into LLVM's Analysis library rather than having them in
a Transforms library.
This is motivated by the need to have the core AliasAnalysis
infrastructure be aware of the ObjCARCAliasAnalysis. However, it also
seems like a nice and clean separation. Everything was very easy to move
and this doesn't create much clutter in the analysis library IMO.
Differential Revision: http://reviews.llvm.org/D12133
llvm-svn: 245541
folding the code into the main Analysis library.
There already wasn't much of a distinction between Analysis and IPA.
A number of the passes in Analysis are actually IPA passes, and there
doesn't seem to be any advantage to separating them.
Moreover, it makes it hard to have interactions between analyses that
are both local and interprocedural. In trying to make the Alias Analysis
infrastructure work with the new pass manager, it becomes particularly
awkward to navigate this split.
I've tried to find all the places where we referenced this, but I may
have missed some. I have also adjusted the C API to continue to be
equivalently functional after this change.
Differential Revision: http://reviews.llvm.org/D12075
llvm-svn: 245318
After hitting @llvm.assume(X) we can:
- propagate equality that X == true
- if X is icmp/fcmp (with eq operation), and one of operand
is constant we can change all variables with constants in the same BasicBlock
http://reviews.llvm.org/D11918
llvm-svn: 245265
PR24469 resulted because DeleteDeadInstruction in handleNonLocalStoreDeletion was
deleting the next basic block iterator. Fixed the same by resetting the basic block iterator
post call to DeleteDeadInstruction.
llvm-svn: 245195
This change makes ScalarEvolution a stand-alone object and just produces
one from a pass as needed. Making this work well requires making the
object movable, using references instead of overwritten pointers in
a number of places, and other refactorings.
I've also wired it up to the new pass manager and added a RUN line to
a test to exercise it under the new pass manager. This includes basic
printing support much like with other analyses.
But there is a big and somewhat scary change here. Prior to this patch
ScalarEvolution was never *actually* invalidated!!! Re-running the pass
just re-wired up the various other analyses and didn't remove any of the
existing entries in the SCEV caches or clear out anything at all. This
might seem OK as everything in SCEV that can uses ValueHandles to track
updates to the values that serve as SCEV keys. However, this still means
that as we ran SCEV over each function in the module, we kept
accumulating more and more SCEVs into the cache. At the end, we would
have a SCEV cache with every value that we ever needed a SCEV for in the
entire module!!! Yowzers. The releaseMemory routine would dump all of
this, but that isn't realy called during normal runs of the pipeline as
far as I can see.
To make matters worse, there *is* actually a key that we don't update
with value handles -- there is a map keyed off of Loop*s. Because
LoopInfo *does* release its memory from run to run, it is entirely
possible to run SCEV over one function, then over another function, and
then lookup a Loop* from the second function but find an entry inserted
for the first function! Ouch.
To make matters still worse, there are plenty of updates that *don't*
trip a value handle. It seems incredibly unlikely that today GVN or
another pass that invalidates SCEV can update values in *just* such
a way that a subsequent run of SCEV will incorrectly find lookups in
a cache, but it is theoretically possible and would be a nightmare to
debug.
With this refactoring, I've fixed all this by actually destroying and
recreating the ScalarEvolution object from run to run. Technically, this
could increase the amount of malloc traffic we see, but then again it is
also technically correct. ;] I don't actually think we're suffering from
tons of malloc traffic from SCEV because if we were, the fact that we
never clear the memory would seem more likely to have come up as an
actual problem before now. So, I've made the simple fix here. If in fact
there are serious issues with too much allocation and deallocation,
I can work on a clever fix that preserves the allocations (while
clearing the data) between each run, but I'd prefer to do that kind of
optimization with a test case / benchmark that shows why we need such
cleverness (and that can test that we actually make it faster). It's
possible that this will make some things faster by making the SCEV
caches have higher locality (due to being significantly smaller) so
until there is a clear benchmark, I think the simple change is best.
Differential Revision: http://reviews.llvm.org/D12063
llvm-svn: 245193
If we can ignore NaNs, fmin/fmax libcalls can become compare and select
(this is what we turn std::min / std::max into).
This IR should then be optimized in the backend to whatever is best for
any given target. Eg, x86 can use minss/maxss instructions.
This should solve PR24314:
https://llvm.org/bugs/show_bug.cgi?id=24314
Differential Revision: http://reviews.llvm.org/D11866
llvm-svn: 245187
Bitwise arithmetic can obscure a simple sign-test. If replacing the
mask with a truncate is preferable if the type is legal because it
permits us to rephrase the comparison more explicitly.
llvm-svn: 245171
ByteSize and BitSize should not be size_t but unsigned, considering
1) They are at most 2^16 and 2^19, respectively.
2) BitSize is an argument to Type::getIntNTy which takes unsigned.
Also, use the correct utostr instead itostr and cache the string result.
Thanks to James Touton for reporting this!
llvm-svn: 245167
Some personality routines require funclet exit points to be clearly
marked, this is done by producing a token at the funclet pad and
consuming it at the corresponding ret instruction. CleanupReturnInst
already had a spot for this operand but CatchReturnInst did not.
Other personality routines don't need to use this which is why it has
been made optional.
llvm-svn: 245149
This patch makes the Merge Functions pass faster by calculating and comparing
a hash value which captures the essential structure of a function before
performing a full function comparison.
The hash is calculated by hashing the function signature, then walking the basic
blocks of the function in the same order as the main comparison function. The
opcode of each instruction is hashed in sequence, which means that different
functions according to the existing total order cannot have the same hash, as
the comparison requires the opcodes of the two functions to be the same order.
The hash function is a static member of the FunctionComparator class because it
is tightly coupled to the exact comparison function used. For example, functions
which are equivalent modulo a single variant callsite might be merged by a more
aggressive MergeFunctions, and the hash function would need to be insensitive to
these differences in order to exploit this.
The hashing function uses a utility class which accumulates the values into an
internal state using a standard bit-mixing function. Note that this is a different interface
than a regular hashing routine, because the values to be hashed are scattered
amongst the properties of a llvm::Function, not linear in memory. This scheme is
fast because only one word of state needs to be kept, and the mixing function is
a few instructions.
The main runOnModule function first computes the hash of each function, and only
further processes functions which do not have a unique function hash. The hash
is also used to order the sorted function set. If the hashes differ, their
values are used to order the functions, otherwise the full comparison is done.
Both of these are helpful in speeding up MergeFunctions. Together they result in
speedups of 9% for mysqld (a mostly C application with little redundancy), 46%
for libxul in Firefox, and 117% for Chromium. (These are all LTO builds.) In all
three cases, the new speed of MergeFunctions is about half that of the module
verifier, making it relatively inexpensive even for large LTO builds with
hundreds of thousands of functions. The same functions are merged, so this
change is free performance.
Author: jrkoenig
Reviewers: nlewycky, dschuff, jfb
Subscribers: llvm-commits, aemerson
Differential revision: http://reviews.llvm.org/D11923
llvm-svn: 245140
This seems to only work some of the time. In some situations,
this seems to use a nonsensical type and isn't actually aware of the
memory being accessed. e.g. if branch condition is an icmp of a pointer,
it checks the addressing mode of i1.
llvm-svn: 245137
Summary: Similar to the change we applied to ASan. The same test case works.
Reviewers: samsonov
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11961
llvm-svn: 245067
This introduces the basic functionality to support "token types".
The motivation stems from the need to perform operations on a Value
whose provenance cannot be obscured.
There are several applications for such a type but my immediate
motivation stems from WinEH. Our personality routine enforces a
single-entry - single-exit regime for cleanups. After several rounds of
optimizations, we may be left with a terminator whose "cleanup-entry
block" is not entirely clear because control flow has merged two
cleanups together. We have experimented with using labels as operands
inside of instructions which are not terminators to indicate where we
came from but found that LLVM does not expect such exotic uses of
BasicBlocks.
Instead, we can use this new type to clearly associate the "entry point"
and "exit point" of our cleanup. This is done by having the cleanuppad
yield a Token and consuming it at the cleanupret.
The token type makes it impossible to obscure or otherwise hide the
Value, making it trivial to track the relationship between the two
points.
What is the burden to the optimizer? Well, it turns out we have already
paid down this cost by accepting that there are certain calls that we
are not permitted to duplicate, optimizations have to watch out for
such instructions anyway. There are additional places in the optimizer
that we will probably have to update but early examination has given me
the impression that this will not be heroic.
Differential Revision: http://reviews.llvm.org/D11861
llvm-svn: 245029
its creation function.
This required shifting a bunch of method definitions to be out-of-line
so that we could leave most of the implementation guts in the .cpp file.
llvm-svn: 245021
I've used forward declarations and reorderd the source code some to make
this reasonably clean and keep as much of the code as possible in the
source file, including all the stratified set details. Just the basic AA
interface and the create function are in the header file, and the header
file is now included into the relevant locations.
llvm-svn: 245009
Summary:
This patch implements my promised optimization to reunites certain sexts from
operands after we extract the constant offset. See the header comment of
reuniteExts for its motivation.
One key building block that enables this optimization is Bjarke's poison value
analysis (D11212). That helps to prove "a +nsw b" can't overflow.
Reviewers: broune
Subscribers: jholewinski, sanjoy, llvm-commits
Differential Revision: http://reviews.llvm.org/D12016
llvm-svn: 245003
AliasAnalysis in LoopIdiomRecognize.
The previous commit to LIR, r244879, exposed some scary bug in the loop
pass pipeline with an assert failure that showed up on several bots.
This patch got reverted as part of getting that revision reverted, but
they're actually independent and unrelated. This patch has no functional
change and should be completely safe. It is also useful for my current
work on the AA infrastructure.
llvm-svn: 244993
If <src> is non-zero we can safely set the flag to true, and this
results in less code generated for, e.g. ffs(x) + 1 on FreeBSD.
Thanks to majnemer for suggesting the fix and reviewing.
Code generated before the patch was applied:
0: 0f bc c7 bsf %edi,%eax
3: b9 20 00 00 00 mov $0x20,%ecx
8: 0f 45 c8 cmovne %eax,%ecx
b: 83 c1 02 add $0x2,%ecx
e: b8 01 00 00 00 mov $0x1,%eax
13: 85 ff test %edi,%edi
15: 0f 45 c1 cmovne %ecx,%eax
18: c3 retq
Code generated after the patch was applied:
0: 0f bc cf bsf %edi,%ecx
3: 83 c1 02 add $0x2,%ecx
6: 85 ff test %edi,%edi
8: b8 01 00 00 00 mov $0x1,%eax
d: 0f 45 c1 cmovne %ecx,%eax
10: c3 retq
It seems we can still use cmove and save another 'test' instruction, but
that can be tackled separately.
Differential Revision: http://reviews.llvm.org/D11989
llvm-svn: 244947
We used to be over-conservative about preserving inbounds. Actually, the second
GEP (which applies the constant offset) can inherit the inbounds attribute of
the original GEP, because the resultant pointer is equivalent to that of the
original GEP. For example,
x = GEP inbounds a, i+5
=>
y = GEP a, i // inbounds removed
x = GEP inbounds y, 5 // inbounds preserved
llvm-svn: 244937
DeadStoreElimination does eliminate a store if it stores a value which was loaded from the same memory location.
So far this worked only if the store is in the same block as the load.
Now we can also handle stores which are in a different block than the load.
Example:
define i32 @test(i1, i32*) {
entry:
%l2 = load i32, i32* %1, align 4
br i1 %0, label %bb1, label %bb2
bb1:
br label %bb3
bb2:
; This store is redundant
store i32 %l2, i32* %1, align 4
br label %bb3
bb3:
ret i32 0
}
Differential Revision: http://reviews.llvm.org/D11854
llvm-svn: 244901
Consider this code:
BB:
%i = phi i32 [ 0, %if.then ], [ %c, %if.else ]
%add = add nsw i32 %i, %b
...
In this common case the add can be moved to the %if.else basic block, because
adding zero is an identity operation. If we go though %if.then branch it's
always a win, because add is not executed; if not, the number of instructions
stays the same.
This pattern applies also to other instructions like sub, shl, shr, ashr | 0,
mul, sdiv, div | 1.
Patch by Jakub Kuderski!
llvm-svn: 244887
simplified form to remove redundant checks and simplify the code for
popcount recognition. We don't actually need to handle all of these
cases.
I've left a FIXME for one in particular until I finish inspecting to
make sure we don't actually *rely* on the predicate in any way.
llvm-svn: 244879
Most SSE/AVX (non-constant) vector shift instructions only use the lower 64-bits of the 128-bit shift amount vector operand, this patch calls SimplifyDemandedVectorElts to optimize for this.
I had to refactor some of my recent InstCombiner work on the vector shifts to avoid quite a bit of duplicate code, it means that SimplifyX86immshift now (re)decodes the type of shift.
Differential Revision: http://reviews.llvm.org/D11938
llvm-svn: 244872
Summary: This patch moves the check of OptimizeForSize before traversing over all basic blocks in current loop. If OptimizeForSize is set to true, no non-trivial unswitch is ever allowed. Therefore, the early exit will help reduce compilation time. This patch should be NFC.
Reviewers: reames, weimingz, broune
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11997
llvm-svn: 244868
code into methods on LoopIdiomRecognize.
This simplifies the code somewhat and also makes it much easier to move
the analyses around. Ultimately, the separate class wasn't providing
significant value over methods -- it contained the precondition basic
block and the current loop. The current loop is already available and
the precondition block wasn't needed everywhere and is easy to pass
around.
In several cases I just moved things to be static functions because they
already accepted most of their inputs as arguments.
This doesn't fix the way we manage analyses yet, that will be the next
patch, but it already makes the code over 50 lines shorter.
No functionality changed.
llvm-svn: 244851
complexity.
There is only one function that was called from multiple locations, and
that was 'getBranch' which has a reasonable one-line spelling already:
dyn_cast<BranchInst>(BB->getTerminator). We could make this shorter, but
it doesn't seem to add much value. Instead, we should avoid calling it
so many times on the same basic blocks, but that will be in a subsequent
patch.
The other functions are only called in one location, so inline them
there, and take advantage of this to use direct early exit and reduce
indentation. This makes it much more clear what is being tested for, and
in fact makes it clear now to me that there are simpler ways to do this
work. However, this patch just does the mechanical inlining. I'll clean
up the functionality of the code to leverage loop simplified form more
effectively in a follow-up.
Despite lots of early line breaks due to early-exit, this is still
shorter than it was before.
llvm-svn: 244841
a significant code cleanup here.
The handling of analyses in this pass is overly complex and can be
simplified significantly, but the right way to do that is to simplify
all of the code not just the analyses, and that'll require pretty
extensive edits that would be noisy with formatting changes mixed into
them.
llvm-svn: 244828
To be clear: this is an *optimization* not a correctness change.
CodeGenPrep likes to duplicate icmps feeding branch instructions to take advantage of x86's ability to fuze many comparison/branch patterns into a single micro-op and to reduce the need for materializing i1s into general registers. PlaceSafepoints likes to place safepoint polls right at the end of basic blocks (immediately before terminators) when inserting entry and backedge safepoints. These two heuristics interact in a somewhat unfortunate way where the branch terminating the original block will be controlled by a condition driven by unrelocated pointers. This forces the register allocator to keep both the relocated and unrelocated values of the pointers feeding the icmp alive over the safepoint poll.
One simple fix would have been to just adjust PlaceSafepoints to move one back in the basic block, but you can reach similar cases as a result of LICM or other hoisting passes. As a result, doing a post insertion fixup seems to be more robust.
I considered doing this in CodeGenPrep itself, but having to update the live sets of already rewritten safepoints gets complicated fast. In particular, you can't just use def/use information because by moving the icmp, we're extending the live range of it's inputs potentially.
Instead, this patch teaches RewriteStatepointsForGC to make the required adjustments before making the relocations explicit in the IR. This change really highlights the fact that RSForGC is a CodeGenPrep-like pass which is performing target specific lowering. In the long run, we may even want to combine the two though this would require a lot more smarts to be integrated into RSForGC first. We currently rely on being able to run a set of cleanup passes post rewriting because the IR RSForGC generates is pretty damn ugly.
Differential Revision: http://reviews.llvm.org/D11819
llvm-svn: 244821
When rewriting the IR such that base pointers are available for every live pointer, we potentially need to duplicate instructions to propagate the base. The original code had only handled PHI and Select under the belief those were the only instructions which would need duplicated. When I added support for vector instructions, I'd added a collection of hacks for ExtractElement which caught most of the common cases. Of course, I then found the one test case my hacks couldn't cover. :)
This change removes all of the early hacks for extract element. By defining extractelement as a BDV (rather than trying to look through it), we can extend the rewriting algorithm to duplicate the extract as needed. Note that a couple of peephole optimizations were left in for the moment, because while we now handle extractelement as a first class citizen, we're not yet handling insertelement. That change will follow in the near future.
llvm-svn: 244808
AliasAnalysis.
Same as the other commits, the TLI access from an alias analysis is
going away and isn't very clean -- it is better to explicitly mark the
dependencies.
llvm-svn: 244785
just depend on it directly.
This was particularly frustrating because there was a really wide
mixture of using a member variable and re-extracting it from the AA that
happened to be around. I think the result is much more clear.
I've also deleted all of the pointless null checks and used references
across the APIs where I could to make it explicit that this cannot be
null in a useful fashion.
llvm-svn: 244780
r243382 changed the behavior to always require a set of memchecks to be
passed to LoopVer. This change restores the prior behavior as an
alternative to the new behavior. This allows the checks to be
implicitly taken from the LAA object.
Patch by Ashutosh Nema!
llvm-svn: 244763
As discussed in D11886, this patch moves the SSE/AVX vector blend folding to instcombiner from PerformINTRINSIC_WO_CHAINCombine (which allows us to remove this completely).
InstCombiner already had partial support for this, I just had to add support for zero (ConstantAggregateZero) masks and also the case where both selection inputs were the same (allowing us to ignore the mask).
I also moved all the relevant combine tests into InstCombine/blend_x86.ll
Differential Revision: http://reviews.llvm.org/D11934
llvm-svn: 244723
`InstCombiner::OptimizeOverflowCheck` was asserting an
invariant (operands to binary operations are ordered by decreasing
complexity) that wasn't really an invariant. Fix this by instead having
`InstCombiner::OptimizeOverflowCheck` establish the invariant if it does
not hold.
llvm-svn: 244676
Summary: This patch adds check for dead blocks and skip them for processSwitchInst(). This will help reduce compilation time.
Reviewers: reames, hans
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11953
llvm-svn: 244656
Summary: LowerSwitch crashed with the attached test case after deleting the default block. This happened because the current implementation of deleting dead blocks is wrong. After the default block being deleted, it contains no instruction or terminator, and it should no be traversed anymore. However, since the iterator is advanced before processSwitchInst() function is executed, the block advanced to could be deleted inside processSwitchInst(). The deleted block would then be visited next and crash dyn_cast<SwitchInst>(Cur->getTerminator()) because Cur->getTerminator() returns a nullptr. This patch fixes this problem by recording dead default blocks into a list, and delete them after all processSwitchInst() has been done. It still possible to visit dead default blocks and waste time process them. But it is a compile time issue, and I plan to have another patch to add support to skip dead blocks.
Reviewers: kariddi, resistor, hans, reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11852
llvm-svn: 244642
Summary:
For LTO we need to enable this pass in the LTO pipeline,
as it is skipped during the "-flto -c" compile step (when PrepareForLTO is
set).
Reviewers: rnk
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11919
llvm-svn: 244622
The select pattern recognition in ValueTracking (as used by InstCombine
and SelectionDAGBuilder) only knew about integer patterns. This teaches
it about minimum and maximum operations.
matchSelectPattern() has been extended to return a struct containing the
existing Flavor and a new enum defining the pattern's behavior when
given one NaN operand.
C minnum() is defined to return the non-NaN operand in this case, but
the idiomatic C "a < b ? a : b" would return the NaN operand.
ARM and AArch64 at least have different instructions for these different cases.
llvm-svn: 244580
This adds somewhat basic preparation functionality including:
- Formation of funclets via coloring basic blocks.
- Cloning of polychromatic blocks to ensure that funclets have unique
program counters.
- Demotion of values used between different funclets.
- Some amount of cleanup once we have removed predecessors from basic
blocks.
- Verification that we are left with a CFG that makes some amount of
sense.
N.B. Arguments and numbering still need to be done.
Differential Revision: http://reviews.llvm.org/D11750
llvm-svn: 244558
This patch and a relatec clang patch solve the problem of having to explicitly enable analysis when specifying a loop hint pragma to get the diagnostics. Passing AlwasyPrint as the pass name (see below) causes the front-end to print the diagnostic if the user has specified '-Rpass-analysis' without an '=<target-pass>’. Users of loop hints can pass that compiler option without having to specify the pass and they will get diagnostics for only those loops with loop hints.
llvm-svn: 244555
This patch moves checking the threshold of runtime pointer checks to the vectorization requirements (late diagnostics) and emits a diagnostic that infroms the user the loop would be vectorized if not for exceeding the pointer-check threshold. Clang will also append the options that can be used to allow vectorization.
llvm-svn: 244523
As discussed in D11760, this patch moves the (V)PSRA(WD) arithmetic shift-by-constant folding to InstCombine to match the logical shift implementations.
Differential Revision: http://reviews.llvm.org/D11886
llvm-svn: 244495
This patch moves the verification of fast-math to just before vectorization is done. This way we can tell clang to append the command line options would that allow floating-point commutativity. Specifically those are enableing fast-math or specifying a loop hint.
llvm-svn: 244489
Sometimes interleaving is not beneficial, as determined by the cost-model and sometimes it is disabled by a loop hint (by the user). This patch modifies the diagnostic messages to make it clear why interleaving wasn't done.
llvm-svn: 244485
This change adds the unroll metadata "llvm.loop.unroll.enable" which directs
the optimizer to unroll a loop fully if the trip count is known at compile time, and
unroll partially if the trip count is not known at compile time. This differs from
"llvm.loop.unroll.full" which explicitly does not unroll a loop if the trip count is not
known at compile time.
The "llvm.loop.unroll.enable" is intended to be added for loops annotated with
"#pragma unroll".
llvm-svn: 244466
Summary:
This adds a hook to TTI which enables us to selectively turn on by default
interleaved access vectorization for targets on which we have have performed
the required benchmarking.
Reviewers: rengolin
Subscribers: rengolin, llvm-commits
Differential Revision: http://reviews.llvm.org/D11901
llvm-svn: 244449
The scalarizer can cache incorrect entries when walking up a chain of
insertelement instructions. This occurs when it encounters more than one
instruction that it is not actively searching for, as it unconditionally caches
every element it finds. The fix is to only cache the first element that it
isn't searching for so we don't overwrite correct entries.
Reviewers: hfinkel
Differential Revision: http://reviews.llvm.org/D11559
llvm-svn: 244448
This is the full set of checks that clients can further filter. IOW,
it's client-agnostic. This makes LAA complete in the sense that it now
provides the two main results of its analysis precomputed:
1. memory dependences via getDepChecker().getInsterestingDependences()
2. run-time checks via getRuntimePointerCheck().getChecks()
However, as a consequence we now compute this information pro-actively.
Thus if the client decides to skip the loop based on the dependences
we've computed the checks unnecessarily. In order to see whether this
was a significant overhead I checked compile time on SPEC2k6 LTO bitcode
files. The change was in the noise.
The checks are generated in canCheckPtrAtRT, at the same place where we
used to call groupChecks to merge checks.
llvm-svn: 244368
Summary: llvm::ConstantFoldTerminator function can convert SwitchInst with single case (and default) to a conditional BranchInst. This patch adds support to preserve make.implicit metadata on this conversion.
Reviewers: sanjoy, weimingz, chenli
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D11841
llvm-svn: 244348
This patch fixes the sse2/avx2 vector shift by constant instcombine call to correctly deal with the fact that the shift amount is formed from the entire lower 64-bit and not just the lowest element as it currently assumes.
e.g.
%1 = tail call <4 x i32> @llvm.x86.sse2.psrl.d(<4 x i32> %v, <4 x i32> <i32 15, i32 15, i32 15, i32 15>)
In this case, (V)PSRLD doesn't perform a lshr by 15 but in fact attempts to shift by 64424509455 ((15 << 32) | 15) - giving a zero result.
In addition, this review also recognizes shift-by-zero from a ConstantAggregateZero type (PR23821).
Differential Revision: http://reviews.llvm.org/D11760
llvm-svn: 244341
As a follow-up to r244181, resolve uniquing cycles underneath distinct
nodes on the fly. This prevents uniquing cycles in early operands from
affecting later operands. It also removes an iteration through distinct
nodes' operands.
No real functional change here, just more prompt resolution of temporary
nodes.
llvm-svn: 244302
After r244074, we now have a successors() method to iterate over
all the successors of a TerminatorInst. This commit changes a bunch
of eligible loops to use it.
llvm-svn: 244260
iisUnmovableInstruction() had a list of instructions hardcoded which are
considered unmovable. The list lacked (at least) an entry for the va_arg
and cmpxchg instructions.
Fix this by introducing a new Instruction::mayBeMemoryDependent()
instead of maintaining another instruction list.
Patch by Matthias Braun <matze@braunis.de>.
Differential Revision: http://reviews.llvm.org/D11577
rdar://problem/22118647
llvm-svn: 244244
This is the first mechanical step in preparation for making this and all
the other alias analysis passes available to the new pass manager. I'm
factoring out all the totally boring changes I can so I'm moving code
around here with no other changes. I've even minimized the formatting
churn.
I'll reformat and freshen comments on the interface now that its located
in the right place so that the substantive changes don't triger this.
llvm-svn: 244197
around a DataLayout interface in favor of directly querying DataLayout.
This wrapper specifically helped handle the case where this no
DataLayout, but LLVM now requires it simplifynig all of this. I've
updated callers to directly query DataLayout. This in turn exposed
a bunch of places where we should have DataLayout readily available but
don't which I've fixed. This then in turn exposed that we were passing
DataLayout around in a bunch of arguments rather than making it readily
available so I've also fixed that.
No functionality changed.
llvm-svn: 244189
Rotate the algorithm for remapping distinct nodes in order to simplify
how uniquing cycles get resolved. This removes some of the recursion,
and, most importantly, exposes all uniquing cycles at the top-level.
Besides being a little more efficient -- temporary MDNodes won't live as
long -- the clearer logic should help protect against bugs like those
fixed in r243961 and r243976.
What are uniquing cycles? Why do they present challenges when remapping
metadata?
!0 = !{!1}
!1 = !{!0}
!0 and !1 form a simple uniquing cycle. When remapping from one
metadata graph to another, every uniquing cycle gets "duplicated"
through a dance:
!0-temp = !{!1?} ; map(!0): clone !0, VM[!0] = !0-temp
!1-temp = !{!0?} ; ..map(!1): clone !1, VM[!1] = !1-temp
!1-temp = !{!0-temp} ; ..map(!1): remap !1's operands
!2 = !{!0-temp} ; ..map(!1): uniquify: !1-temp => !2
!0-temp = !{!2} ; map(!0): remap !0's operands
!3 = !{!2} ; map(!0): uniquify: !0-temp => !3
; Result
!2 = !{!3}
!3 = !{!2}
(In the two "uniquify" steps above, the operands of !X-temp are compared
to the operands of !X. If they're the same, then !X-temp gets RAUW'ed
to !X; if they're different, then !X-temp is promoted to a new unique
node. The latter case always hits in for uniquing cycles, so we
duplicate all the nodes involved.)
Why is this a problem? Uniquable Metadata nodes that have temporary
node as transitive operands keep RAUW support until the temporary nodes
get finalized. With non-cycles, this happens automatically: when a
uniquable node's count of unresolved operands drops to zero, it
immediately sheds its own RAUW support (possibly triggering the same in
any node that references it). However, uniquing cycles create a
reference cycle, and uniqued nodes that transitively reference a
uniquing cycle are "stuck" in an unresolved state until someone calls
`MDNode::resolveCycles()` on a node in the unresolved subgraph.
Distinct nodes should help here (and mostly do): since they aren't
uniqued anywhere, they are guaranteed not to be RAUW'ed. They
effectively form a barrier between uniqued nodes, breaking some uniquing
cycles, and shielding uniqued nodes from uniquing cycles.
Unfortunately, with this barrier in place, the unresolved subgraph(s)
can be disjoint from the top-level node. The mapping algorithm needs to
find at least one representative from each disjoint subgraph. But which
nodes are *stuck*, and which will get resolved automatically? And which
nodes are in the unresolved subgraph? The old logic was conservative.
This commit rotates the logic for distinct nodes, so that we have access
to unresolved nodes at the top-level call to `llvm::MapMetadata()`.
Each time we return to the top-level, we know that all temporaries have
been RAUW'ed away. Here, it's safe (and necessary) to call
`resolveCycles()` immediately on unresolved operands.
This should also perform better than the old algorithm. The recursion
stack is shorter, temporary nodes don't live as long, and there are
fewer tracking references to unresolved nodes. As the debug info graph
introduces more 'distinct' nodes, remapping should incrementally get
cheaper and cheaper.
Aside from possible performance improvements (and reduced cruft in the
`LLVMContext`), there should be no functionality change here.
llvm-svn: 244181
Rename `remap()` to `remapOperands()`, and restrict its contract to
remapping operands. Previously, it also called `mapToMetadata()`, but
this logic is hard to reason about externally. In particular, this
refactors `mapUniquedNode()` to avoid redundant mapping calls, taking
advantage of the RAUWs that are already in place.
llvm-svn: 244168
The only place that tries to return a CallGraph by value
(CallGraphAnalysis::run) doesn't seem to be used right now, but it's a
reasonable bit of cleanup anyway.
llvm-svn: 244122
Create wrapper methods in the Function class for the OptimizeForSize and MinSize
attributes. We want to hide the logic of "or'ing" them together when optimizing
just for size (-Os).
Currently, we are not consistent about this and rely on a front-end to always set
OptimizeForSize (-Os) if MinSize (-Oz) is on. Thus, there are 18 FIXME changes here
that should be added as follow-on patches with regression tests.
This patch is NFC-intended: it just replaces existing direct accesses of the attributes
by the equivalent wrapper call.
Differential Revision: http://reviews.llvm.org/D11734
llvm-svn: 243994
r243883 and r243961 made a use-after-free far more likely:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux-fast/builds/6041/steps/check-llvm%20asan/logs/stdio
Unresolved nodes get inserted into the `Cycles` array. If they later
get resolved through RAUW, we need to update the reference. It's
interesting that this never hit before (maybe an asan-ified clang
bootstrap with `-flto -g` would have hit it, but I admit I haven't tried
anything quite that crazy).
llvm-svn: 243976
This change was done as an audit and is by inspection. The new EH
system is still very much a work in progress. NFC for the landingpad
case.
llvm-svn: 243965
r243883 started moving 'distinct' nodes instead of duplicated them in
lib/Linker. This had the side-effect of sometimes not cloning uniqued
nodes that reference them. I missed a corner case:
!named = !{!0}
!0 = !{!1}
!1 = distinct !{!0}
!0 is the entry point for "remapping", and a temporary clone (say,
!0-temp) is created and mapped in case we need to model a uniquing
cycle.
Recursive descent into !1. !1 is distinct, so we leave it alone,
but update its operand to !0-temp.
Pop back out to !0. Its only operand, !1, hasn't changed, so we don't
need to use !0-temp. !0-temp goes out of scope, and we're finished
remapping, but we're left with:
!named = !{!0}
!0 = !{!1}
!1 = distinct !{null} ; uh oh...
Previously, if !0 and !0-temp ended up with identical operands, then
!0-temp couldn't have been referenced at all. Now that distinct nodes
don't get duplicated, that assumption is invalid. We need to
!0-temp->replaceAllUsesWith(!0) before freeing !0-temp.
I found this while running an internal `-flto -g` bootstrap. Strangely,
there was no case of this in the open source bootstrap I'd done before
commit...
llvm-svn: 243961
through PHI nodes across iterations.
This patch teaches the new advanced loop unrolling heuristics to propagate
constants into the loop from the preheader and around the backedge after
simulating each iteration. This lets us brute force solve simple recurrances
that aren't modeled effectively by SCEV. It also makes it more clear why we
need to process the loop in-order rather than bottom-up which might otherwise
make much more sense (for example, for DCE).
This came out of an attempt I'm making to develop a principled way to account
for dead code in the unroll estimation. When I implemented
a forward-propagating version of that it produced incorrect results due to
failing to propagate *cost* between loop iterations through the PHI nodes, and
it occured to me we really should at least propagate simplifications across
those edges, and it is quite easy thanks to the loop being in canonical and
LCSSA form.
Differential Revision: http://reviews.llvm.org/D11706
llvm-svn: 243900
Instead of cloning distinct `MDNode`s when linking in a module, just
move them over. The module linker destroys the source module, so the
old node would otherwise just be leaked on the context. Create the new
node in place. This also reduces the number of cloned uniqued nodes
(since it's less likely their operands have changed).
This mapping strategy is only correct when we're discarding the source,
so the linker turns it on via a ValueMapper flag, `RF_MoveDistinctMDs`.
There's nothing observable in terms of `llvm-link` output here: the
linked module should be semantically identical.
I'll be adding more 'distinct' nodes to the debug info metadata graph in
order to break uniquing cycles, so the benefits of this will partly come
in future commits. However, we should get some gains immediately, since
we have a fair number of 'distinct' `DILocation`s being linked in.
llvm-svn: 243883
This is a minor optimization to only check for unresolved operands
inside `mapDistinctNode()` if the operands have actually changed. This
shouldn't really cause any change in behaviour. I didn't actually see a
slowdown in a profile, I was just poking around nearby and saw the
opportunity.
llvm-svn: 243866
This introduces new instructions neccessary to implement MSVC-compatible
exception handling support. Most of the middle-end and none of the
back-end haven't been audited or updated to take them into account.
Differential Revision: http://reviews.llvm.org/D11097
llvm-svn: 243766
The patch changes the SLPVectorizer::vectorizeStores to choose the immediate
succeeding or preceding candidate for a store instruction when it has multiple
consecutive candidates. In this way it has better chance to find more slp
vectorization opportunities.
Differential Revision: http://reviews.llvm.org/D10445
llvm-svn: 243666
The reason I was passing this vector by value in the constructor so that
I wouldn't have to copy when initializing the corresponding member but
then I forgot the std::move.
The use-case is LoopDistribution which filters the checks then
std::moves it to LoopVersioning's constructor. With this interface we
can avoid any copies.
llvm-svn: 243616
Before, we were passing the pointer partitions to LAA. Now, we get all
the checks from LAA and filter out the checks within partitions in
LoopDistribution.
This effectively concludes the steps to move filtering memchecks from
LAA into its clients. There is still some cleanup left to remove the
unused interfaces in LAA that still take PtrPartition.
(Moving this functionality to LoopDistribution also requires
needsChecking on pointers to be made public.)
llvm-svn: 243613
Bonus change to remove emacs major mode marker from SystemZMachineFunctionInfo.cpp because emacs already knows it's C++ from the extension. Also fix typo "appeary" in AMDGPUMCAsmInfo.h.
llvm-svn: 243585
Summary:
returns_twice (most importantly, setjmp) functions are
optimization-hostile: if local variable is promoted to register, and is
changed between setjmp() and longjmp() calls, this update will be
undone. This is the reason why "man setjmp" advises to mark all these
locals as "volatile".
This can not be enough for ASan, though: when it replaces static alloca
with dynamic one, optionally called if UAR mode is enabled, it adds a
whole lot of SSA values, and computations of local variable addresses,
that can involve virtual registers, and cause unexpected behavior, when
these registers are restored from buffer saved in setjmp.
To fix this, just disable dynamic alloca and UAR tricks whenever we see
a returns_twice call in the function.
Reviewers: rnk
Subscribers: llvm-commits, kcc
Differential Revision: http://reviews.llvm.org/D11495
llvm-svn: 243561
ASan shadow on Android starts at address 0 for both historic and
performance reasons. This is possible because the platform mandates
-pie, which makes lower memory region always available.
This is not such a good idea on 64-bit platforms because of MAP_32BIT
incompatibility.
This patch changes Android/AArch64 mapping to be the same as that of
Linux/AAarch64.
llvm-svn: 243548
Summary:
As added initially, statepoints required their call targets to be a
constant pointer null if ``numPatchBytes`` was non-zero. This turns out
to be a problem ergonomically, since there is no way to mark patchable
statepoints as calling a (readable) symbolic value.
This change remove the restriction of requiring ``null`` call targets
for patchable statepoints, and changes PlaceSafepoints to maintain the
symbolic call target through its transformation.
Reviewers: reames, swaroop.sridhar
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11550
llvm-svn: 243502
Before the patch, the checks were generated internally in
addRuntimeCheck. Now, we use the new overloaded version of
addRuntimeCheck that takes the ready-made set of checks as a parameter.
The checks are now generated by the client (LoopDistribution) with the
new RuntimePointerChecking::generateChecks API.
Also the new printChecks API is used to print out the checks for
debugging.
This is to continue the transition over to the new model whereby clients
will get the full set of checks from LAA, filter it and then pass it to
LoopVersioning and in turn to addRuntimeCheck.
llvm-svn: 243382
Summary:
If a scale or a base register can be rewritten as "Zext({A,+,1})" then
LSR will now consider a formula of that form in its normal cost
computation.
Depends on D9180
Reviewers: qcolombet, atrick
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D9181
llvm-svn: 243348
Summary:
Was D9784: "Remove loop variant range check when induction variable is
strictly increasing"
This change re-implements D9784 with the two differences:
1. It does not use SCEVExpander and does not generate new
instructions. Instead, it does a quick local search for existing
`llvm::Value`s that it needs when modifying the `icmp`
instruction.
2. It is more general -- it deals with both increasing and decreasing
induction variables.
I've added all of the tests included with D9784, and two more.
As an example on what this change does (copied from D9784):
Given C code:
```
for (int i = M; i < N; i++) // i is known not to overflow
if (i < 0) break;
a[i] = 0;
}
```
This transformation produces:
```
for (int i = M; i < N; i++)
if (M < 0) break;
a[i] = 0;
}
```
Which can be unswitched into:
```
if (!(M < 0))
for (int i = M; i < N; i++)
a[i] = 0;
}
```
I went back and forth on whether the top level logic should live in
`SimplifyIndvar::eliminateIVComparison` or be put into its own
routine. Right now I've put it under `eliminateIVComparison` because
even though the `icmp` is not *eliminated*, it no longer is an IV
comparison. I'm open to putting it in its own helper routine if you
think that is better.
Reviewers: reames, nicholas, atrick
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11278
llvm-svn: 243331
Now that we are generating sane codegen for vector sext/zext nodes on SSE targets, this patch uses instcombine to replace the SSE41/AVX2 pmovsx and pmovzx intrinsics with the equivalent native IR code.
Differential Revision: http://reviews.llvm.org/D11503
llvm-svn: 243303
This reverts commit r243167.
Duncan pointed out that dyn_cast can return null in these cases, so this
was an unsafe commit to make. Sorry for the noise.
Worryingly there were no tests which fail...
llvm-svn: 243302
r243250 appeared to break clang/test/Analysis/dead-store.c on one of the build
slaves, but I couldn't reproduce this failure locally. Probably a false
positive as I saw this test was broken by r243246 or r243247 too but passed
later without people fixing anything.
llvm-svn: 243253
Summary:
This patch updates TargetTransformInfoImplCRTPBase::getGEPCost to consider
addressing modes. It now returns TCC_Free when the GEP can be completely folded
to an addresing mode.
I started this patch as I refactored SLSR. Function isGEPFoldable looks common
and is indeed used by some WIP of mine. So I extracted that logic to getGEPCost.
Furthermore, I noticed getGEPCost wasn't directly tested anywhere. The best
testing bed seems CostModel, but its getInstructionCost method invokes
getAddressComputationCost for GEPs which provides very coarse estimation. So
this patch also makes getInstructionCost call the updated getGEPCost for GEPs.
This change inevitably breaks some tests because the cost model changes, but
nothing looks seriously wrong -- if we believe the new cost model is the right
way to go, these tests should be updated.
This patch is not perfect yet -- the comments in some tests need to be updated.
I want to know whether this is a right approach before fixing those details.
Reviewers: chandlerc, hfinkel
Subscribers: aschwaighofer, llvm-commits, aemerson
Differential Revision: http://reviews.llvm.org/D9819
llvm-svn: 243250
Summary:
This patch improves trivial loop unswitch.
The current trivial loop unswitch only checks if loop header's terminator contains a trivial unswitch condition. But if the loop header only has one reachable successor (due to intentionally or unintentionally missed code simplification), we should consider the successor as part of the loop header. Therefore, instead of stopping at loop header's terminator, we should keep traversing its successors within loop until reach a *real* conditional branch or switch (whose condition can not be constant folded). This change will enable a single -loop-unswitch pass to unswitch multiple trivial conditions (unswitch one trivial condition could open opportunity to unswitch another one in the same loop), while the old implementation can unswitch only one per pass.
Reviewers: reames, broune
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11481
llvm-svn: 243203
This patch extend LoopReroll pass to hand the loops which
is similar to the following:
while (len > 1) {
sum4 += buf[len];
sum4 += buf[len-1];
len -= 2;
}
llvm-svn: 243171
Since both places which set this variable do so with dyn_cast, and not
dyn_cast_or_null, its impossible to get a nullptr here, so we can remove
the check.
llvm-svn: 243167
Instead of the pattern
for (auto I = x.rbegin(), E = x.end(); I != E; ++I)
we can use make_range to construct the reverse range and iterate using
that instead.
llvm-svn: 243163
Summary:
This threshold limited FunctionAttrs ability to prove arguments to be read-only.
In NVPTX, a specialized instruction ld.global.nc can be used to load memory
with non-coherent texture cache. We notice that in SHOC [1] benchmark, some
function arguments are not marked with readonly because FunctionAttrs reaches
a hardcoded threshold when analysis uses.
Removing this threshold won't cause significant regression in compilation time, because the worst-case time complexity of the algorithm is still O(# of instructions) for each parameter.
Patched by Xuetian Weng.
[1] https://github.com/vetter/shoc
Reviewers: nlewycky, jingyue, nicholas
Subscribers: nicholas, test, llvm-commits
Differential Revision: http://reviews.llvm.org/D11311
llvm-svn: 243141
The names for instructions inserted were previous dependent on iteration order. By deriving the names from the original instructions, we can avoid instability in tests without resorting to ordered traversals. It also makes the IR mildly easier to read at large scale.
llvm-svn: 243140
We had a few places where we did
for (unsigned i = 0, e = STy->getNumElements(); i != e; ++i) {
but those could instead do
for (auto *EltTy : STy->elements()) {
llvm-svn: 243136
Summary:
Resolving a branch allows us to ignore blocks that won't be executed, and thus make our estimate more accurate.
This patch is intended to be applied after D10205 (though it could be applied independently).
Reviewers: chandlerc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D10206
llvm-svn: 243084
The new code should hopefully be equivalent to the old code; it just uses a worklist to track instructions which need to visited rather than iterating over all instructions visited each time. This should be faster, but the primary benefit is that the purpose should be more clear and the diff of adding another instruction type (forthcoming) much more obvious.
Differential Revision: http://reviews.llvm.org/D11480
llvm-svn: 243071
Deleting much of the code using trace-rewrite-statepoints and use idiomatic DEBUG statements instead. This includes adding operator<< to a helper class.
llvm-svn: 243054
We don't need to pass in the map from BDV to PhiStates; we can instead handle that externally and let the MeetPhiStates helper class just meet PhiStates.
llvm-svn: 243045
Summary:
Scalarizer has two data structures that hold information about changes
to the function, Gathered and Scattered. These are cleared in finish()
at the end of runOnFunction() if finish() detects any changes to the
function.
However, finish() was checking for changes by only checking if
Gathered was non-empty. The function visitStore() only modifies
Scattered without touching Gathered. As a result, Scattered could have
ended up having stale data if Scalarizer only scalarized store
instructions. Since the data in Scattered is used during the execution
of the pass, this introduced dangling pointer errors.
The fix is to check whether both Scattered and Gathered are empty
before deciding what to do in finish(). This also fixes a problem
where the Function can be modified although the pass returns false.
Reviewers: rnk
Subscribers: rnk, srhines, llvm-commits
Differential Revision: http://reviews.llvm.org/D10459
llvm-svn: 243040
We currently version `__asan_init` and when the ABI version doesn't match, the linker gives a `undefined reference to '__asan_init_v5'` message. From this, it might not be obvious that it's actually a version mismatch error. This patch makes the error message much clearer by changing the name of the undefined symbol to be `__asan_version_mismatch_check_xxx` (followed by the version string). We obviously don't want the initializer to be named like that, so it's a separate symbol that is used only for the purpose of version checking.
Reviewed at http://reviews.llvm.org/D11004
llvm-svn: 243003
the general GMR-in-non-LTO flag.
Without this, we have the global information during the CGSCC pipeline
for GVN and such, but don't have it available during the late loop
optimizations such as the vectorizer. Moreover, after the CGSCC pipeline
has finished we have substantially more accurate and refined call graph
information, function annotations, etc, which will make GMR even more
powerful than it is early in the pipelien.
Note that we have to play silly games with preserving AliasAnalysis
(which is now trivially preserved) in order to let a module analysis
magically be preserved into the entire function pass pipeline.
Simultaneously we have to not make GMR an immutable pass in order to be
able to re-run it and collect fresh data on the final call graph.
llvm-svn: 242999
preparation for de-coupling the AA implementations.
In order to do this, they had to become fake-scoped using the
traditional LLVM pattern of a leading initialism. These can't be actual
scoped enumerations because they're bitfields and thus inherently we use
them as integers.
I've also renamed the behavior enums that are specific to reasoning
about the mod/ref behavior of functions when called. This makes it more
clear that they have a very narrow domain of applicability.
I think there is a significantly cleaner API for all of this, but
I don't want to try to do really substantive changes for now, I just
want to refactor the things away from analysis groups so I'm preserving
the exact original design and just cleaning up the names, style, and
lifting out of the class.
Differential Revision: http://reviews.llvm.org/D10564
llvm-svn: 242963
Summary:
While working on a project I wound up generating a fairly large lookup table (10k entries) of callbacks inside of a static constructor. Clang was taking upwards of ~10 minutes to compile the lookup table. I generated a smaller test case (http://www.inolen.com/static_initializer_test.ll) that, after running with -ftime-report, pointed fingers at GlobalOpt and MemCpyOptimizer.
Running globalopt took around ~9 minutes. The slowdown came from how GlobalOpt merged stores from static constructors individually into the global initializer in EvaluateStaticConstructor. For each store it discovered and wanted to commit, it would copy the existing global initializer and then merge in the individual store. I changed this so that stores are now grouped by global, and sorted from most significant to least significant by their GEP indexes (e.g. a store to GEP 0, 0 comes before GEP 0, 0, 1). With this representation, the existing initializer can be copied and all new stores merged into it in a single pass.
With this patch and http://reviews.llvm.org/D11198, the lookup table that was taking ~10 minutes to compile now compiles in around 5 seconds. I've ran 'make check' and the test-suite, which all passed.
I'm not really sure who to tag as a reviewer, Lang mentioned that Chandler may be appropriate.
Reviewers: chandlerc, nlewycky
Subscribers: nlewycky, llvm-commits
Differential Revision: http://reviews.llvm.org/D11200
llvm-svn: 242935
pipeline.
Even before I started improving its runtime, it was already crazy fast
once the call graph exists, and if we can get it to be conservatively
correct, will still likely catch a lot of interesting and useful cases.
So it may well be useful to enable by default.
But more importantly for me, this should make it easier for me to test
that changes aren't breaking it in fundamental ways by enabling it for
normal builds.
llvm-svn: 242895
Currently, a load from an alloca that is used in as single block and is not preceded
by a store is replaced by undef. This is not always correct if the single block is
inside a loop.
Fix the logic so that:
1) If there are no stores in the block, replace the load with an undef, as before.
2) If there is a store (regardless of where it is in the block w.r.t the load), bail
out, and let the rest of mem2reg handle this alloca.
Patch by: gil.rapaport@intel.com
Differential Revision: http://reviews.llvm.org/D11355
llvm-svn: 242884
In r242510, non-instrumented allocas are now moved into the first basic block. This patch limits that to only move allocas that are present *after* the first instrumented one (i.e. only move allocas up). A testcase was updated to show behavior in these two cases. Without the patch, an alloca could be moved down, and could cause an invalid IR.
Differential Revision: http://reviews.llvm.org/D11339
llvm-svn: 242883
through APIs that are no longer necessary now that the update API has
been removed.
This will make changes to the AA interfaces significantly less
disruptive (I hope). Either way, it seems like a really nice cleanup.
llvm-svn: 242882
part of simplifying its interface and usage in preparation for porting
to work with the new pass manager.
Note that this will likely expose that we have dead arguments, members,
and maybe even pass requirements for AA. I'll be cleaning those up in
seperate patches. This just zaps the actual update API.
Differential Revision: http://reviews.llvm.org/D11325
llvm-svn: 242881
change because the diff is *useless*. I assure you, I just switched to
early-return in this function.
Cleanup in preparation for my next commit, as requested in code review!
llvm-svn: 242880
Summary: The current code in LoopUnswtich::processCurrentLoop() mixes trivial loop unswitch and non-trivial loop unswitch together. It goes over all basic blocks in the loop and checks if a condition is trivial or non-trivial unswitch condition. However, trivial unswitch condition can only occur in the loop header basic block (where it controls whether or not the loop does something at all). This refactoring separate trivial loop unswitch and non-trivial loop unswitch. Before going over all basic blocks in the loop, it checks if the loop header contains a trivial unswitch condition. If so, unswitch it. Otherwise, go over all blocks like before but don't check trivial condition any more since they are not possible to be in the other blocks. This code has no functionality change.
Reviewers: meheff, reames, broune
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11276
llvm-svn: 242873
types and loads, loads or stores widened past the size of an alloca,
etc.
This started off with a bug report about big-endian behavior with
bitfields and loads and stores to a { i32, i24 } struct. An initial
attempt to fix this was sent for review in D10357, but that didn't
really get to the root of the problem.
The core issue was that canConvertValue and convertValue in SROA were
handling different bitwidth integers by doing a zext of the integer. It
wouldn't do a trunc though, only a zext! This would in turn lead SROA to
form an i24 load from an i24 alloca, zext it to i32, and then use it.
This would at least produce the wrong value for big-endian systems.
One of my many false starts here was to correct the computation for
big-endian systems by shifting. But this doesn't actually work because
the original code has a 64-bit store to the entire 8 bytes, and a 32-bit
load of the last 4 bytes, and because the alloc size is 8 bytes, we
can't lose that last (least significant if bigendian) byte! The real
problem here is that we're forming an i24 load in SROA which is actually
not sufficiently wide to load all of the necessary bits here. The source
has an i32 load, and SROA needs to form that as well.
The straightforward way to do this is to disable the zext logic in
canConvertValue and convertValue, forcing us to actually load all
32-bits. This seems like a really good change, but it in turn breaks
several other parts of SROA.
First in the chain of knock-on failures, we had places where we were
doing integer-widening promotion even though some of the integer loads
or stores extended *past the end* of the alloca's memory! There was even
a comment about preventing this, but it only prevented the case where
the type had a different bit size from its store size. So I added checks
to handle the cases where we actually have a widened load or store and
to avoid trying to special integer widening promotion in those cases.
Second, we actually rely on the ability to promote in the face of loads
past the end of an alloca! This is important so that we can (for
example) speculate loads around PHI nodes to do more promotion. The bits
loaded are garbage, but as long as they aren't used and the alignment is
suitable high (which it wasn't in the test case!) this is "fine". And we
can't stop promoting here, lots of things stop working well if we do. So
we need to add specific logic to handle the extension (and truncation)
case, but *only* where that extension or truncation are over bytes that
*are outside the alloca's allocated storage* and thus totally bogus to
load or store.
And of course, once we add back this correct handling of extension or
truncation, we need to correctly handle bigendian systems to avoid
re-introducing the exact bug that started us off on this chain of misery
in the first place, but this time even more subtle as it only happens
along speculated loads atop a PHI node.
I've ported an existing test for PHI speculation to the big-endian test
file and checked that we get that part correct, and I've added several
more interesting big-endian test cases that should help check that we're
getting this correct.
Fun times.
llvm-svn: 242869
We insert a bitcast which obfuscates the getCalledFunction for the utility
function which looks up attributes from the called function. Loosing ABI
changing parameter attributes is a bad thing.
rdar://21516488
llvm-svn: 242807
A bit more code cleanup: delete some a trivial true assertion and supporting code, remove a redundant cast, and use count in assertions where feasible.
llvm-svn: 242805
We can use builders to simplify part of the code and we only check for the existance of the metadata value; this enables us to delete some redundant code.
llvm-svn: 242751
Summary:
Arguments to llvm.localescape must be static allocas. They must be at
some statically known offset from the frame or stack pointer so that
other functions can access them with localrecover.
If we ever want to instrument these, we can use more indirection to
recover the addresses of these local variables. We can do it during
clang irgen or with the asan module pass.
Reviewers: eugenis
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11307
llvm-svn: 242726
Not sure if the optimizer will save the call as getCalledFunction()
is not a trivial access function but the code is clearer this way.
llvm-svn: 242641
directly model in the new PM.
This also was an incredibly brittle and expensive update API that was
never fully utilized by all the passes that claimed to preserve AA, nor
could it reasonably have been extended to all of them. Any number of
places add uses of values. If we ever wanted to reliably instrument
this, we would want a callback hook much like we have with ValueHandles,
but doing this for every use addition seems *extremely* expensive in
terms of compile time.
The only user of this update mechanism is GlobalsModRef. The idea of
using this to keep it up to date doesn't really work anyways as its
analysis requires a symmetric analysis of two different memory
locations. It would be very hard to make updates be sufficiently
rigorous to *guarantee* symmetric analysis in this way, and it pretty
certainly isn't true today.
However, folks have been using GMR with this update for a long time and
seem to not be hitting the issues. The reported issue that the update
hook fixes isn't even a problem any more as other changes to
GetUnderlyingObject worked around it, and that issue stemmed from *many*
years ago. As a consequence, a prior patch provided a flag to control
the unsafe behavior of GMR, and this patch removes the update mechanism
that has questionable compile-time tradeoffs and is causing problems
with moving to the new pass manager. Note the lack of test updates --
not one test in tree actually requires this update, even for a contrived
case.
All of this was extensively discussed on the dev list, this patch will
just enact what that discussion decides on. I'm sending it for review in
part to show what I'm planning, and in part to show the *amazing* amount
of work this avoids. Every call to the AA here is something like three
to six indirect function calls, which in the non-LTO pipeline never do
any work! =[
Differential Revision: http://reviews.llvm.org/D11214
llvm-svn: 242605
Instrumentation and the runtime library were in disagreement about
ASan shadow offset on Android/AArch64.
This fixes a large number of existing tests on Android/AArch64.
llvm-svn: 242595
Since r230724 ("Skip promotable allocas to improve performance at -O0"), there is a regression in the generated debug info for those non-instrumented variables. When inspecting such a variable's value in LLDB, you often get garbage instead of the actual value. ASan instrumentation is inserted before the creation of the non-instrumented alloca. The only allocas that are considered standard stack variables are the ones declared in the first basic-block, but the initial instrumentation setup in the function breaks that invariant.
This patch makes sure uninstrumented allocas stay in the first BB.
Differential Revision: http://reviews.llvm.org/D11179
llvm-svn: 242510
Internalizing an individual comdat group member without also internalizing
the other members of the comdat can break comdat semantics. For example,
if a module contains a reference to an internalized comdat member, and the
linker chooses a comdat group from a different object file, this will break
the reference to the internalized member.
This change causes the internalizer to only internalize comdat members if all
other members of the comdat are not externally visible. Once a comdat group
has been fully internalized, there is no need to apply comdat rules to its
members; later optimization passes (e.g. globaldce) can legally drop individual
members of the comdat. So we drop the comdat attribute from all comdat members.
Differential Revision: http://reviews.llvm.org/D10679
llvm-svn: 242423
This new wrapper pass is useful when we want to do branch probability analysis conditionally (e.g. only in PGO mode) but don't want to add one more pass dependence.
http://reviews.llvm.org/D11241
llvm-svn: 242349
Summary:
This is a trivial code change with no functionality effect.
When LoopUnswitch determines trivial unswitch condition, it checks whether the loop header's terminator instruction is a branch instruction or switch instruction since trivial unswitch condition can only apply to these two instruction types. The current code does not fail the check directly on other instruction types, but check the nullness of LoopExitBB variable instead. The added else clause makes the check fail immediately on other instruction types and makes the code more obvious.
Reviewers: reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D11239
llvm-svn: 242345
Self-referential constants containing references to a merged function
no longer cause the MergeFunctions pass to infinite loop. Also adds a
reproduction IR which would otherwise fail, which was isolated from a similar
issue in Chromium.
Author: jrkoenig
Reviewers: nlewycky, jfb
Subscribers: llvm-commits, nlewycky, jfb
Differential Revision: http://reviews.llvm.org/D11208
llvm-svn: 242337
During estimation of unrolling effect we should be able to propagate
constants through casts.
Differential Revision: http://reviews.llvm.org/D10207
llvm-svn: 242257
This is useful when we want to do block frequency analysis
conditionally (e.g. only in PGO mode) but don't want to add
one more pass dependence.
Patch by congh.
Approved by dexonsmith.
Differential Revision: http://reviews.llvm.org/D11196
llvm-svn: 242248
I am planning to add more nested classes inside RuntimePointerCheck so
all these triple-nesting would be hard to follow.
Also rename it to RuntimePointerChecking (i.e. append 'ing').
llvm-svn: 242218
Sometimes an incidentally created instruction can duplicate a Value used
elsewhere. It then often doesn't end up in the leader table. If it's later
removed, we attempt to remove it from the leader table and segfault.
Instead we should just ignore the removal request, which won't cause any
problems. The reverse situation, where the original instruction is replaced by
the new one (which you might think could leave the leader table empty) cannot
occur, because the incidental instruction will never be found in the first
place.
llvm-svn: 242199
Volatile loads and stores are made visible in global state regardless of
what memory is involved. It is not correct to disregard the ordering
and synchronization scope because it is possible to synchronize with
memory operations performed by hardware.
This partially addresses PR23737.
llvm-svn: 242126
Previously we would refrain from attempting to increase the linkage of
available_externally globals because they were considered weak for the
linker. Now they are treated more like a declaration instead of a weak
definition.
This was causing SSE alignment faults in Chromuim, when some code
assumed it could increase the alignment of a dllimported global that it
didn't control. http://crbug.com/509256
llvm-svn: 242091
When spotting that a loop can use ctpop, we were incorrectly replacing all uses of a value with a value derived from ctpop.
The bug here was exposed because we were replacing a use prior to the ctpop with the ctpop value and so we have a use before def, i.e., we changed
%tobool.5 = icmp ne i32 %num, 0
store i1 %tobool.5, i1* %ptr
br i1 %tobool.5, label %for.body.lr.ph, label %for.end
to
store i1 %1, i1* %ptr
%0 = call i32 @llvm.ctpop.i32(i32 %num)
%1 = icmp ne i32 %0, 0
br i1 %1, label %for.body.lr.ph, label %for.end
Even if we inserted the ctpop so that it dominates the store here, that would still be incorrect. The store doesn’t want the result of ctpop.
The fix is very simple, and involves replacing only the branch condition with the ctpop instead of all uses.
Reviewed by Hal Finkel.
llvm-svn: 242068
Enable runtime unrolling for loops with unroll count metadata ("#pragma unroll N")
and a runtime trip count. Also, do not unroll loops with unroll full metadata if the
loop has a runtime loop count. Previously, such loops would be unrolled with a
very large threshold (pragma-unroll-threshold) if runtime unrolled happened to be
enabled resulting in a very large (and likely unwise) unroll factor.
llvm-svn: 242047
Passes should never modify it, just use the const version. While there
reduce copying in LoopInterchange. No functional change intended.
llvm-svn: 242041
There is no suitable basic block to sink instructions in loops without
exits. The only way an instruction in a loop without exits can be used
is as an incoming value to a PHI. In such cases, the incoming block for
the corresponding value is unreachable.
This fixes PR24013.
Differential Revision: http://reviews.llvm.org/D10903
llvm-svn: 241987
Sanjay Patel