dependencies. The basic situation was this: consider if we had:
store1
...
store2
...
store3
Where memdep thinks that store3 depends on store2 and store2 depends
on store1. The problem happens when we delete store2: The code in
question was updating dep info for store3 to be store1. This is a
spiffy optimization, but is not safe at all, because aliasing isn't
transitive. This bug isn't exposed today with DSE because DSE will only
zap store2 if it is identifical to store 3, and in this case, it is
safe to update it to depend on store1. However, memcpyopt is not so
fortunate, which is presumably why the "dropInstruction" code used to
exist.
Since this doesn't actually provide a speedup in practice, just rip the
code out.
llvm-svn: 60263
an entry in the nonlocal deps map, don't reset entries
referencing that instruction to [dirty, null], instead, set
them to [dirty,next] where next is the instruction after the
deleted one. Use this information in the non-local deps
code to avoid rescanning entire blocks.
This speeds up GVN slightly by avoiding pointless work. On
403.gcc this makes GVN 1.5% faster.
llvm-svn: 60256
Put a some code back to handle buggy behavior that GVN expects: it wants
loads to depend on each other, and accesses to depend on their allocations.
llvm-svn: 60240
Document the Dirty value more precisely, use it for the uninitialized
DepResultTy value. Change reverse mappings to be from an instruction*
instead of DepResultTy, and stop tracking other forms. This makes it more
clear that we only care about the instruction cases.
Eliminate a DepResultTy,bool pair by using Dirty in the local case as well,
shrinking the map and simplifying the code.
This speeds up GVN by ~3% on 403.gcc.
llvm-svn: 60232
query. This makes it crystal clear what cases can escape from MemDep that
the clients have to handle. This also gives the clients a nice simplified
interface to it that is easy to poke at.
This patch also makes DepResultTy and MemoryDependenceAnalysis::DepType
private, yay.
llvm-svn: 60231
of a pointer/int pair instead of a manually bitmangled pointer.
This forces clients to think a little more about checking the
appropriate pieces and will be useful for internal
implementation improvements later.
I'm not particularly happy with this. After going through this
I don't think that the clients of memdep should be exposed to
the internal type at all. I'll fix this in a subsequent commit.
This has no functionality change.
llvm-svn: 60230
properly updates the reverse dependency map when it installs updated
dependencies for instructions that depend on the removed instruction.
llvm-svn: 60222
value. It must now be as if the pointer were allocated and has not escaped to
the caller. Thanks to Dan Gohman for pointing out the error in the original
and helping devise this definition.
llvm-svn: 59940
indicate functions that allocate, such as operator new, or list::insert. The
actual definition is slightly less strict (for now).
No changes to the bitcode reader/writer, asm printer or verifier were needed.
llvm-svn: 59934
g++ -m32 -c -g -DIN_GCC -W -Wall -Wwrite-strings -Wmissing-format-attribute -fno-common -mdynamic-no-pic -DHAVE_CONFIG_H -Wno-unused -DTARGET_NAME=\"i386-apple-darwin9.5.0\" -I. -I. -I../../llvm-gcc.src/gcc -I../../llvm-gcc.src/gcc/. -I../../llvm-gcc.src/gcc/../include -I./../intl -I../../llvm-gcc.src/gcc/../libcpp/include -I../../llvm-gcc.src/gcc/../libdecnumber -I../libdecnumber -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.obj/include -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.src/include -DENABLE_LLVM -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.obj/../llvm.src/include -D_DEBUG -D_GNU_SOURCE -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -I. -I. -I../../llvm-gcc.src/gcc -I../../llvm-gcc.src/gcc/. -I../../llvm-gcc.src/gcc/../include -I./../intl -I../../llvm-gcc.src/gcc/../libcpp/include -I../../llvm-gcc.src/gcc/../libdecnumber -I../libdecnumber -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.obj/include -I/Volumes/Sandbox/Buildbot/llvm/full-llvm/build/llvm.src/include ../../llvm-gcc.src/gcc/llvm-types.cpp -o llvm-types.o
../../llvm-gcc.src/gcc/llvm-convert.cpp: In member function 'void TreeToLLVM::EmitMemCpy(llvm::Value*, llvm::Value*, llvm::Value*, unsigned int)':
../../llvm-gcc.src/gcc/llvm-convert.cpp:1496: error: 'memcpy_i32' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp:1496: error: 'memcpy_i64' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp: In member function 'void TreeToLLVM::EmitMemMove(llvm::Value*, llvm::Value*, llvm::Value*, unsigned int)':
../../llvm-gcc.src/gcc/llvm-convert.cpp:1512: error: 'memmove_i32' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp:1512: error: 'memmove_i64' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp: In member function 'void TreeToLLVM::EmitMemSet(llvm::Value*, llvm::Value*, llvm::Value*, unsigned int)':
../../llvm-gcc.src/gcc/llvm-convert.cpp:1528: error: 'memset_i32' is not a member of 'llvm::Intrinsic'
../../llvm-gcc.src/gcc/llvm-convert.cpp:1528: error: 'memset_i64' is not a member of 'llvm::Intrinsic'
make[3]: *** [llvm-convert.o] Error 1
make[3]: *** Waiting for unfinished jobs....
rm fsf-funding.pod gcov.pod gfdl.pod cpp.pod gpl.pod gcc.pod
make[2]: *** [all-stage1-gcc] Error 2
make[1]: *** [stage1-bubble] Error 2
make: *** [all] Error 2
llvm-svn: 59809
Use it to safely handle less-than-or-equals-to exit conditions in loops. These
also occur when the loop exit branch is exit on true because SCEV inverses the
icmp predicate.
Use it again to handle non-zero strides, but only with an unsigned comparison
in the exit condition.
llvm-svn: 59528
If this patch causes a performance regression for anyone, please let me know,
and it can be fixed in a different way with much more effort.
llvm-svn: 59384
information. This logically replaces the "Desc" classes in
MachineModuleInfo. Nice features of these classes are that they:
1. Are much more efficient than MMI because they don't create a
temporary parallel data structure for debug info that has to be
'serialized' and 'deserialized' into/out of the module.
2. These provide a much cleaner abstraction for debug info than
MMI, which will make it easier to change the implementation in
the future (to be MDNode-based).
3. These are much easier to use than the MMI interfaces, requiring
a lot less code in the front-ends.
4. These can be used to both create (for frontends) and read (for
codegen) debug information. DebugInfoBuilder can only be used
to create the nodes.
So far, this is implemented just enough to support the debug info
generation needs of clang. This can and should be extended to
support the full set of debug info constructs, and we should switch
llvm-gcc and llc over to using this in the near future.
This code also has a ton of FIXMEs in it, because the way we
currently represent debug info in LLVM IR is basically insane in a
variety of details. This sort of issue should be fixed when we
eventually reimplement debug info on top of MDNodes.
llvm-svn: 58954
pointer bitcasts and GEP's", and centralize the
logic in Value::getUnderlyingObject. The
difference with stripPointerCasts is that
stripPointerCasts only strips GEPs if all
indices are zero, while getUnderlyingObject
strips GEPs no matter what the indices are.
llvm-svn: 56922
- Recognize expressions like "x > -1 ? x : 0" as min/max and turn them
into expressions like "x < 0 ? 0 : x", which is easily recognizable
as a min/max operation.
- Refrain from folding expression like "y/2 < 1" to "y < 2" when the
comparison is being used as part of a min or max idiom, like
"y/2 < 1 ? 1 : y/2". In that case, the division has another use, so
folding doesn't eliminate it, and obfuscates the min/max, making it
harder to recognize as a min/max operation.
These benefit ScalarEvolution, CodeGen, and anything else that wants to
recognize integer min and max.
llvm-svn: 56246
its callers to emit a space character before calling it when a
space is needed.
This fixes several spurious whitespace issues in
ScalarEvolution's debug dumps. See the test changes for
examples.
This also fixes odd space-after-tab indentation in the output
for switch statements, and changes calls from being printed like
this:
call void @foo( i32 %x )
to this:
call void @foo(i32 %x)
llvm-svn: 56196
when a readonly declaration is called, set a
flag. This is faster and uses less memory.
In theory it is less accurate, because before
only those internal globals that were read
by someone were being marked "Ref", but now
all are. But in practice, thanks to other
passes, all internal globals of the kind
considered here will be both read and stored
to: those only read will have been turned
into constants, and those only stored to will
have been deleted.
llvm-svn: 56143
(1) code left over from the days of ConstantPointerRef:
if a use of a function is a GlobalValue then that is
not considered a reason to add an edge from the external
node, even though the use may be as an initializer for
an externally visible global! There might be some point
to this behaviour when the use is by an alias (though the
code predated aliases by some centuries), but I think
PR2782 is a better way of handling that. (2) If function
F calls function G, and also G is a parameter to the
call, then an F->G edge is not added to the callgraph.
While this doesn't seem to matter much, adding such an
edge makes the callgraph more regular.
In addition, the new code should be faster as well as
simpler.
llvm-svn: 55987
call (thus changing the call site) it didn't
inform the callgraph about this. But the
call site does matter - as shown by the testcase,
the callgraph become invalid after the inliner
ran (with an edge between two functions simply
missing), resulting in wrong deductions by
GlobalsModRef.
llvm-svn: 55872
because it does not maintain a correct list
of callsites. I discovered (see following
commit) that the inliner will create a wrong
callgraph if it is fed a callgraph with
correct edges but incorrect callsites. These
were created by Prune-EH, and while it wasn't
done via removeCallEdgeTo, it could have been
done via removeCallEdgeTo, which is an accident
waiting to happen. Use removeCallEdgeFor
instead.
llvm-svn: 55859
analysis would bail out without removing function
records for other members of the SCC (which may exist
if those functions read or wrote global variables).
Since these are initialized to "readnone", this
resulted in incorrect alias analysis results.
llvm-svn: 55714
callgraph, when one member of a SCC calls another
then the analysis would drop to mod-ref because
there is (usually) no function info for the callee
yet; fix this. Teach the analysis about function
attributes, in particular the readonly attribute
(which requires being careful about globals).
llvm-svn: 55696
use raw_ostream instead of std::ostream. Among other goodness,
this speeds up llvm-dis of kc++ with a release build from 0.85s
to 0.49s (88% faster).
Other interesting changes:
1) This makes Value::print be non-virtual.
2) AP[S]Int and ConstantRange can no longer print to ostream directly,
use raw_ostream instead.
3) This fixes a bug in raw_os_ostream where it didn't flush itself
when destroyed.
4) This adds a new SDNode::print method, instead of only allowing "dump".
A lot of APIs have both std::ostream and raw_ostream versions, it would
be useful to go through and systematically anihilate the std::ostream
versions.
This passes dejagnu, but there may be minor fallout, plz let me know if
so and I'll fix it.
llvm-svn: 55263
continue past the first conditional branch when looking for a
relevant test. This helps it avoid using MAX expressions in
loop trip counts in more cases.
llvm-svn: 54697
type lattice value for an Argument*, giving clients the opportunity to
use something other than Top for it if they choose to."
Patch by John McCall!
llvm-svn: 54589
version uses a new algorithm for evaluating the binomial coefficients
which is significantly more efficient for AddRecs of more than 2 terms
(see the comments in the code for details on how the algorithm works).
It also fixes some bugs: it removes the arbitrary length restriction for
AddRecs, it fixes the silent generation of incorrect code for AddRecs
which require a wide calculation width, and it fixes an issue where we
were incorrectly truncating the iteration count too far when evaluating
an AddRec expression narrower than the induction variable.
There are still a few related issues I know of: I think there's
still an issue with the SCEVExpander expansion of AddRec in terms of
the width of the induction variable used. The hack to avoid generating
too-wide integers shouldn't be necessary; instead, the callers should be
considering the cost of the expansion before expanding it (in addition
to not expanding too-wide integers, we might not want to expand
expressions that are really expensive, especially when optimizing for
size; calculating an length-17 32-bit AddRec currently generates about 250
instructions of straight-line code on X86). Also, for long 32-bit
AddRecs on X86, CodeGen really sucks at scheduling the code. I'm planning on
filing follow-up PRs for these issues.
llvm-svn: 54332
time applying to the implicit comparison in smin expressions. The
correct way to transform an inequality into the opposite
inequality, either signed or unsigned, is with a not expression.
I looked through the SCEV code, and I don't think there are any more
occurrences of this issue.
llvm-svn: 54194
SGT exit condition. Essentially, the correct way to flip an inequality
in 2's complement is the not operator, not the negation operator.
That said, the difference only affects cases involving INT_MIN.
Also, enhance the pre-test search logic to be a bit smarter about
inequalities flipped with a not operator, so it can eliminate the smax
from the iteration count for simple loops.
llvm-svn: 54184
circumstances we could end up remapping a dependee to the same instruction
that we're trying to remove. Handle this properly by just falling back to
a conservative solution.
llvm-svn: 54132
bail after 256-bits to avoid producing code that the backends can't handle.
Previously, we capped it at 64-bits, preferring to miscompile in those cases.
This change also reverts much of r52248 because the invariants the code was
expecting are now being met.
llvm-svn: 53812
Move GetConstantStringInfo to lib/Analysis. Remove
string output routine from Constant. Update all
callers. Change debug intrinsic api slightly to
accomodate move of routine, these now return values
instead of strings.
This unbreaks llvm-gcc bootstrap.
llvm-svn: 52884
string output routine from Constant. Update all
callers. Change debug intrinsic api slightly to
accomodate move of routine, these now return values
instead of strings.
llvm-svn: 52748
inserting extractvalues. In particular, this prevents the insertion of
extractvalues that can't be folded away later. Also add an example of when this
stuff is needed.
llvm-svn: 52328
I'm at it, rename it to FindInsertedValue.
The only functional change is that newly created instructions are no longer
added to instcombine's worklist, but that is not really necessary anyway (and
I'll commit some improvements next that will completely remove the need).
llvm-svn: 52315
This fixes several minor bugs (such as returning noalias
for comparisons between external weak functions an null) but
is mostly a cleanup.
llvm-svn: 52299
take into account the instrucion pointed by InsertPt. Thanks to it,
returning the new value of InsertPt to the InsertBinop() caller can be
avoided. The bug was, actually, in visitAddRecExpr() method which wasn't
correctly handling changes of InsertPt. There shouldn't be any
performance regression, as -gvn pass (run after -indvars) removes any
redundant binops.
llvm-svn: 52291
Add a safety measure. It isn't safe to assume in ScalarEvolutionExpander that
all loops are in canonical form (but it should be safe for loops that have
AddRecs).
xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
llvm-svn: 52275
with code that was expecting different bit widths for different values.
Make getTruncateOrZeroExtend a method on ScalarEvolution, and use it.
llvm-svn: 52248
is longer than the second one) should stop after finding one. Added break
instruction guarantees it. It also changes difference between offsets to
absolute value of this difference in the condition.
llvm-svn: 51875
out of instcombine into a new file in libanalysis. This also teaches
ComputeNumSignBits about the number of sign bits in a constantint.
llvm-svn: 51863
Analysis/ConstantFolding to fold ConstantExpr's, then make instcombine use it
to try to use targetdata to fold constant expressions on void instructions.
Also extend the icmp(inttoptr, inttoptr) folding to handle the case where
int size != ptr size.
llvm-svn: 51559
SCCP like sparse lattice analysis with relative ease. Just pick your
lattice function and implement the transfer function and you're good.
Just make sure you don't break monotonicity ;-)
llvm-svn: 50961
by an instance of LibCallInfo to provide mod/ref info of
standard library functions. This is powerful enough to
say that 'sqrt' is readonly except that it modifies errno,
or that "printf doesn't store to memory unless the %n
constraint is present" etc.
llvm-svn: 50827
Currently is sufficient to describe mod/ref behavior but will hopefully
eventually be extended for other purposes.
This isn't used by anything yet.
llvm-svn: 50820
manually performing the comparison. This allows the special
case to work correctly even in the case where someone is
experimenting with a different comparison function :-).
llvm-svn: 49670
Parse reversed smax and umax as smin and umin and express them with negative
or binary-not SCEVs (which are really just subtract under the hood).
Parse 'xor %x, -1' as (-1 - %x).
Remove dead code (ConstantInt::get always returns a ConstantInt).
Don't use getIntegerSCEV(-1, Ty). The first value is an int, then it gets
passed into a uint64_t. Instead, create the -1 directly from
ConstantInt::getAllOnesValue().
llvm-svn: 47360
Also, noalias arguments are be considered "like" stack allocated ones for this purpose, because
the only way they can be modref'ed is if they escape somewhere in the current function.
llvm-svn: 47247
variable (with step 1) and m is its final value. Then, the correct trip
count is SMAX(m,n)-n. Previously, we used SMAX(0,m-n), but m-n may
overflow and can't in general be interpreted as signed.
Patch by Nick Lewycky.
llvm-svn: 47007
to the RHS. This simple change allows to compute loop iteration count
for loops with condition similar to the one in the testcase (which seems
to be quite common).
llvm-svn: 46959
arbitrary iteration.
The patch:
1) changes SCEVSDivExpr into SCEVUDivExpr,
2) replaces PartialFact() function with BinomialCoefficient(); the
computations (essentially, the division) in BinomialCoefficient() are
performed with the apprioprate bitwidth necessary to avoid overflow;
unsigned division is used instead of the signed one.
Computations in BinomialCoefficient() require support from the code
generator for APInts. Currently, we use a hack rounding up the
neccessary bitwidth to the nearest power of 2. The hack is easy to turn
off in future.
One remaining issue: we assume the divisor of the binomial coefficient
formula can be computed accurately using 16 bits. It means we can handle
AddRecs of length up to 9. In future, we should use APInts to evaluate
the divisor.
Thanks to Nicholas for cooperation!
llvm-svn: 46955
Chris Lattner