Pete Cooper [Thu, 30 Apr 2015 22:15:59 +0000 (22:15 +0000)]
Don't always apply kill flag in thumb2 ABS pseudo expansion.
The expansion for t2ABS was always setting the kill flag on the rsb instruction.
It should instead only be set on rsb if it was set on the original ABS instruction.
[X86] Use 4 byte preferred aggregate alignment on Win32
This helps reduce the frequency of stack realignment prologues in 32-bit
X86 Windows code. Before this change and the corresponding clang change,
we would take the max of the type preferred alignment and the explicit
alignment on the alloca.
If you don't override aggregate alignment in datalayout, you get a
default of 8. This dates back to 2007 / r34356, and changing it seems
prohibitively difficult at this point.
InstCombineSimplifyDemanded: Remove nsw/nuw flags when optimizing demanded bits
When optimizing demanded bits of the operands of an Add we have to
remove the nsw/nuw flags as we have no guarantee anymore that we don't
wrap. This is legal here because the top bit is not demanded. In fact
this operaion was already performed but missed in the case of an Add
with a constant on the right side. To fix this this patch refactors the
code to unify the code paths in SimplifyDemandedUseBits() handling of
Add/Sub:
- The transformation of Add->Or is removed from the simplify demand
code because the equivalent transformation exists in
InstCombiner::visitAdd()
- KnownOnes/KnownZero are not adjusted for Add x, C anymore as
computeKnownBits() already performs these computations.
- The simplification of the operands is unified. In this new version
constant on the right side of a Sub are shrunk now as I could not find
a reason why not to do so.
- The special case for clearing nsw/nuw in ShrinkDemandedConstant() is
not necessary anymore as the caller does that already.
InstCombine: Move Sub->Xor rule from SimplifyDemanded to InstCombine
The rule that turns a sub to xor if the LHS is 2^n-1 and the remaining bits
are known zero, does not use the demanded bits at all: Move it to the
normal InstCombine code path.
Andrea Di Biagio [Thu, 30 Apr 2015 21:03:29 +0000 (21:03 +0000)]
Fix for PR23103. Correctly propagate the 'IsUndef' flag to the register operands of a commuted instruction.
Revision 220239 exposed a latent bug in method
'TargetInstrInfo::commuteInstruction'. When commuting the operands of a machine
instruction, method 'commuteInstruction' didn't correctly propagate the
'IsUndef' flag to the register operands of the new (commuted) instruction.
Before this patch, the following instruction:
%vreg4<def> = VADDSDrr %vreg14, %vreg5<undef>; FR64:%vreg4,%vreg14,%vreg5
was wrongly converted by method 'commuteInstruction' into:
%vreg4<def> = VADDSDrr %vreg5, %vreg14<undef>; FR64:%vreg4,%vreg5,%vreg14
The correct instruction should have been:
%vreg4<def> = VADDSDrr %vreg5<undef>, %vreg14; FR64:%vreg4,%vreg5,%vreg14
This patch fixes the problem in method 'TargetInstrInfo::commuteInstruction'.
When swapping the operands of a machine instruction, we now make sure that
'IsUndef' flags are correctly set.
Added test case 'pr23103.ll'.
Kevin Enderby [Thu, 30 Apr 2015 20:30:42 +0000 (20:30 +0000)]
For llvm-objdump, with the -archive-headers and -macho options, use the -non-verbose
option to print the archive headers using raw numeric values. Also add the -archive-member-offsets
for use with these to also trigger printing of the offset of the archive member from the start
of the archive.
Matt Arsenault [Thu, 30 Apr 2015 19:35:41 +0000 (19:35 +0000)]
MachineVerifier: Don't crash if MachineOperand has no parent
If you somehow added a MachineOperand to an instruction
that did not have the parent set, the verifier would
crash since it attempts to use the operand's parent.
[ARM] Do not generate invalid encoding for stack adjust, even if this is just
temporary.
Because of that:
1. The machine verifier was complaining on such code.
2. The generate code worked just because the thumb reduction size pass fixed the
opcode.
Jan Vesely [Thu, 30 Apr 2015 17:15:56 +0000 (17:15 +0000)]
Reinstate revisions r234755, r234759, r234760
changes:
Don't apply on hexagon and NVPTX since they no longer claim to support UADDO/USUBO
Add location to getConstant
Drop comment about the ops being turned into expand
Rafael Espindola [Thu, 30 Apr 2015 14:21:49 +0000 (14:21 +0000)]
Write sections mostly in one pass.
During ELF writing, there is no need to further relax the sections, so we
should not be creating fragments. This patch avoids doing so in all cases
but debug section compression (that is next).
Also, the ELF format is fairly simple to write. We can do a single pass over
the sections to write them out and compute the section header table.
Diego Novillo [Thu, 30 Apr 2015 13:22:48 +0000 (13:22 +0000)]
Fix private constructor for ScaledNumber.
Summary:
The private constructor for ScaledNumber was using uint64_t instead of
DigitsT. This was preventing instantiations of ScaledNumber with
anything other than uint64_t types.
In implementing the tests, I ran into another issue. Operators >>= and
<<= did not have variants for accepting other ScaledNumber as the shift
argument. This is expected by the SCALED_NUMBER_BOP.
It makes no sense to allow shifting a ScaledNumber by another
ScaledNumber, so the patch includes two new templates for shifting
ScaledNumbers.
Summary:
This doesn't make much difference to MIPS32, but it will simplify a
MIPS64r6 bugfix which will follow shortly by removing unnecessary
sign-extension of parameters.
Add convenient overloads for CreateInsertElement and CreateExtractElement methods in IRBuilder
Summary:
This pathc add convenient overloads for CreateInsertElement and CreateExtractElement methods in IRBuilder
where vector index can be uint64_t instead of Value*.
[TableGen] Merge a variable assignment and a return to drop curly braces. Fold an assignment into an if. Use auto on the result of a couple dyn_casts. NFC
Sanjoy Das [Thu, 30 Apr 2015 04:56:04 +0000 (04:56 +0000)]
[InstCombine] Add new rule for MIN(MAX(~A, ~B), ~C) et. al.
Summary:
Optimizing these well are especially interesting for IRCE since it
"clamps" values by generating this sort of pattern through SCEV
expressions.
Owen Anderson [Thu, 30 Apr 2015 04:06:32 +0000 (04:06 +0000)]
Semantically revert r236031, which is not a good idea for in-order targets.
At the least it should be guarded by some kind of target hook.
It also introduced catastrophic compile time and code quality
regressions on some out of tree targets (test case still being
reduced/sanitized).
Sanjay agreed with reverting this patch until these issues can be
resolved.
Hans Wennborg [Thu, 30 Apr 2015 00:57:37 +0000 (00:57 +0000)]
Switch lowering: use profile info to build weight-balanced binary search trees
This will cause hot nodes to appear closer to the root.
The literature says building the tree like this makes it a near-optimal (in
terms of search time given key frequencies) binary search tree. In LLVM's case,
we can do up to 3 comparisons in each leaf node, so it might be better to opt
for lower tree height in some cases; that's something to look into in the
future.
[WinEH] Start EH preparation for 32-bit x86, it uses no arguments
32-bit x86 MSVC-style exceptions are functionaly similar to 64-bit, but
they take no arguments. Instead, they implicitly use the value of EBP
passed in by the caller as a pointer to the parent's frame. In LLVM, we
can represent this as llvm.frameaddress(1), and feed that into all of
our calls to llvm.framerecover.
The next steps are:
- Add an alloca to the fs:00 linked list of handlers
- Add something like llvm.sjlj.lsda or generalize it to store in the
alloca
- Move state number calculation to WinEHPrepare, arrange for
FunctionLoweringInfo to call it
- Use the state numbers to insert explicit loads and stores in the IR
David Blaikie [Wed, 29 Apr 2015 21:22:39 +0000 (21:22 +0000)]
[opaque pointer type] Pass GlobalAlias the actual pointer type rather than decomposing it into pointee type + address space
Many of the callers already have the pointer type anyway, and for the
couple of callers that don't it's pretty easy to call PointerType::get
on the pointee type and address space.
This avoids LLParser from using PointerType::getElementType when parsing
GlobalAliases from IR.
Tim Northover [Wed, 29 Apr 2015 19:16:38 +0000 (19:16 +0000)]
ARM: mark branch-like instructions with correct flags.
There's probably no way to test BXJ, but if the compiler ever did emit it
during CodeGen it would have to be a block terminator so "isBranch" is
appropriate.
BLX is more tricky. Clearly a call, but it affects surprisingly little.
Adrian Prantl [Wed, 29 Apr 2015 16:52:17 +0000 (16:52 +0000)]
Temporarily relax a check in the debug info verifier.
The clang frontend helps out GDB by emitting the members of local anonymous
unions as artificial local variables with shared storage. When SROA splits
the storage for artificial local variables that are smaller than the entire
union, the overhang piece will be outside of the allotted space for the
variable and this check fails.
x86 Windows uses the '_' prefix for all global symbols, and this was
mistakenly being applied to frameescape labels, which are not externally
visible global symbols. They use the private global prefix 'L'.
The *right* way to fix this is probably to stop masquerading this label
as an ExternalSymbol and create a new SDNode type. These labels are not
"external", and we know they will be resolved by assembly time. Having a
custom SDNode type would allow us to do better X86 address mode
matching, so it's probably worth doing eventually.
Finish off PR23080 by renaming the debug info IR constructs from `MD*`
to `DI*`. The last of the `DIDescriptor` classes were deleted in
r235356, and the last of the related typedefs removed in r235413, so
this has all baked for about a week.
Note: If you have out-of-tree code (like a frontend), I recommend that
you get everything compiling and tests passing with the *previous*
commit before updating to this one. It'll be easier to keep track of
what code is using the `DIDescriptor` hierarchy and what you've already
updated, and I think you're extremely unlikely to insert bugs. YMMV of
course.
Back to *this* commit: I did this using the rename-md-di-nodes.sh
upgrade script I've attached to PR23080 (both code and testcases) and
filtered through clang-format-diff.py. I edited the tests for
test/Assembler/invalid-generic-debug-node-*.ll by hand since the columns
were off-by-three. It should work on your out-of-tree testcases (and
code, if you've followed the advice in the previous paragraph).
Some of the tests are in badly named files now (e.g.,
test/Assembler/invalid-mdcompositetype-missing-tag.ll should be
'dicompositetype'); I'll come back and move the files in a follow-up
commit.
[mips] [IAS] Inline assemble-time shifting out of createLShiftOri. NFC.
Summary:
Do the assemble-time shifts from createLShiftOri at the source, which groups all the shifting together, closer to the main logic path, and
store the results in concisely-named variables to improve code clarity.
projects / llvm / log