James Henderson [Thu, 31 Jan 2019 13:58:48 +0000 (13:58 +0000)]
[CommandLine] Improve help text for cl::values style options
In order to make an option value truly optional, both the ValueOptional
and an empty-named value are required. This empty-named value appears in
the command-line help text, which is not ideal.
This change improves the help text for these sort of options in a number
of ways:
1) ValueOptional options with an empty-named value now print their help
text twice: both without and then with '=<value>' after the name. The
latter version then lists the allowed values after it.
2) Empty-named values with no help text in ValueOptional options are not
listed in the permitted values.
3) Otherwise empty-named options are printed as =<empty> rather than
simply '='.
4) Option values without help text do not have the '-' separator
printed.
It also tweaks the llvm-symbolizer -functions help text to not print a
trailing ':' as that looks bad combined with 1) above.
Sjoerd Meijer [Thu, 31 Jan 2019 08:38:06 +0000 (08:38 +0000)]
[ARM] Thumb2: ConstantMaterializationCost
Constants can also be materialised using the negated value and a MVN, and this
case seem to have been missed for Thumb2. To check the constant materialisation
costs, we now call getT2SOImmVal twice, once for the original constant and then
also for its negated value, and this function checks if the constant can both
be splatted or rotated.
This was revealed by a test that optimises for minsize: instead of a LDR
literal pool load and having a literal pool entry, just a MVN with an immediate
is smaller (and also faster).
Sjoerd Meijer [Thu, 31 Jan 2019 08:07:30 +0000 (08:07 +0000)]
[SelectionDAG] Codesize: don't expand SHIFT to SHIFT_PARTS
And instead just generate a libcall. My motivating example on ARM was a simple:
shl i64 %A, %B
for which the code bloat is quite significant. For other targets that also
accept __int128/i128 such as AArch64 and X86, it is also beneficial for these
cases to generate a libcall when optimising for minsize. On these 64-bit targets,
the 64-bits shifts are of course unaffected because the SHIFT/SHIFT_PARTS
lowering operation action is not set to custom/expand.
Douglas Yung [Thu, 31 Jan 2019 07:58:34 +0000 (07:58 +0000)]
Fixup test after r352704 since it changes how paths may be emitted.
On Unix/Mac OS X, normpath() returns the path unchanged (FileCheck), but
on case-insensitive filesystems (like NTFS on Windows), it converts the
path to lowercase (filecheck) which was causing the test to fail.
Petr Hosek [Thu, 31 Jan 2019 06:21:01 +0000 (06:21 +0000)]
[CMake] Unify scripts for generating VCS headers
Previously, there were two different scripts for generating VCS headers:
one used by LLVM and one used by Clang. They were both similar, but
different. They were both broken in their own ways, for example the one
used by Clang didn't properly handle monorepo resulting in an incorrect
version information reported by Clang.
This change unifies two the scripts by introducing a new script that's
used from both LLVM and Clang, ensures that the new script supports both
monorepo and standalone SVN and Git setups, and removes the old scripts.
Max Kazantsev [Thu, 31 Jan 2019 06:19:25 +0000 (06:19 +0000)]
[SCEV] Prohibit SCEV transformations for huge SCEVs
Currently SCEV attempts to limit transformations so that they do not work with
big SCEVs (that may take almost infinite compile time). But for this, it uses heuristics
such as recursion depth and number of operands, which do not give us a guarantee
that we don't actually have big SCEVs. This situation is still possible, though it is not
likely to happen. However, the bug PR33494 showed a bunch of simple corner case
tests where we still produce huge SCEVs, even not reaching big recursion depth etc.
This patch introduces a concept of 'huge' SCEVs. A SCEV is huge if its expression
size (intoduced in D35989) exceeds some threshold value. We prohibit optimizing
transformations if any of SCEVs we are dealing with is huge. This gives us a reliable
check that we don't spend too much time working with them.
As the next step, we can possibly get rid of old limiting mechanisms, such as recursion
depth thresholds.
Nico Weber [Thu, 31 Jan 2019 00:40:43 +0000 (00:40 +0000)]
lit: Let lit.util.which() return a normcase()ed path
LLVMConfig.with_environment() uses os.path.normcase(os.path.normpath(x)) to
normalize temporary env vars. LLVMConfig.use_clang() uses with_environment() to
temporarily set PATH and then look for clang there. This means that on Windows,
clang will be run with a path like c:\foo\bin\clang.EXE (with a lower-case
"C:").
lit.util.which() used to not do this, which means the executables added in
clang/test/lit.cfg.py (e.g. c-index-test) were run with a path like
C:\foo\bin\c-index-test.EXE (because both CMake and GN happen to write
clang_tools_dir with an upper-case C to lit.site.cfg.py).
clang/test/Index/pch-from-libclang.c requires that both c-index-test and clang
use _exactly_ the same resource dir path (same case and everything), because a
hash of the resource directory is used as module cache path.
This patch is necessary but not sufficient to make pch-from-libclang.c pass on
Windows.
Thomas Lively [Thu, 31 Jan 2019 00:35:37 +0000 (00:35 +0000)]
[LegalizeVectorTypes] Allow illegal indices when splitting extract_vector_elt
Summary:
Fixes PR40267, in which the removed assertion was triggering on
perfectly valid IR. As far as I can tell, constant out of bounds
indices should be allowed when splitting extract_vector_elt, since
they will simply be propagated as out of bounds indices in the
resulting split vector and handled appropriately elsewhere.
Craig Topper [Thu, 31 Jan 2019 00:04:48 +0000 (00:04 +0000)]
[LegalizeTypes] Use report_fatal_error instead of llvm_unreachable in the default case of some type legalization handlers that can be reached with intrinsics with result or operands that aren't legal types.
These can be triggered by mistakenly using a 64-bit mode only intrinsics with a -mtriple=i686. Using report_fatal_error gives a better experience for this mistake in release builds instead of probably crashing.
We already do this for some of the vector type legalization handles.
Craig Topper [Thu, 31 Jan 2019 00:04:46 +0000 (00:04 +0000)]
[X86] Remove handling of ISD::INTRINSIC_WO_CHAIN in ReplaceNodeResults.
I believe this was there to handle avx512bw intrinsics that returned i64 type in 32-bit mode. But all those intrinsics have since been changed to v64i1 results or replaced with generic IR.
Heejin Ahn [Wed, 30 Jan 2019 23:53:36 +0000 (23:53 +0000)]
[WebAssembly] Remove TODO on wasm.extract.exception intrinsic (NFC)
Summary:
We planned to delete this intrinsic and do custom lowering from
`wasm.get.exception`, which has a token argument, to
`EXTRACT_EXCEPTION`, a wasm pseudo instruction that simulates popping a
value from the wasm stack.
To do that, we need to introduce a new `WebAssemblyISD` node for this,
which itself is not a problem, but also have to introduce the
`WebAssemblyISD` namespace in SelectionDAGBuilder.cpp. I don't think any
other targets are doing that in the file. And also putting a
target-specific intrinsic in the common file is a little weird too. (All
other intrinsic functions in this `visitIntrinsicCall` functions are not
target-specific ones. Other target-specific intrinsics are usually
handled in `lib/Target/[TargetName]/[TargetName]ISelLowering.cpp`. The
reason we can't do this is it has a token argument.
Anyway, so I think I prefer the current code with one redundant
intrinsic more than adding one more `WebAssemblyISD` node and
also introducing the `WebAssemblyISD` namespace into
SelectionDAGBuilder.cpp. What do you think?
Zachary Turner [Wed, 30 Jan 2019 23:52:32 +0000 (23:52 +0000)]
[RuntimeDyld] Don't try to allocate sections with align 0.
ELF sections allow 0 for the alignment, which is specified to
be the same as 1. However many clients do not expect this and
will behave poorly in the presence of a 0-aligned section (for
example by trying to modulo something by the section alignment).
We can be more polite by making sure that we always pass a
non-zero value to clients.
Heejin Ahn [Wed, 30 Jan 2019 22:44:45 +0000 (22:44 +0000)]
[WebAssembly] Restore stack pointer right after catch instruction
Summary:
After the staack is unwound due to a thrown exxception,
`__stack_pointer` global can point to an invalid address. So
a `global.set` to restore `__stack_pointer` should be inserted right
after `catch` instruction.
But after r352598 the `global.set` instruction is inserted not right
after `catch` but after `block` - `br-on-exn` - `end_block` -
`extract_exception` sequence. This CL fixes it.
While doing that, we can actually move ReplacePhysRegs pass after
LateEHPrepare and merge EHRestoreStackPointer pass into LateEHPrepare,
and now placing `global.set` to `__stack_pointer` right after `catch` is
much easier. Otherwise it is hard to guarantee that `global.set` is
still right after `catch` and not touched with other transformations, in
which case we have to do something to hoist it.
Jessica Paquette [Wed, 30 Jan 2019 21:16:04 +0000 (21:16 +0000)]
[GlobalISel][AArch64] Add instruction selection support for @llvm.log2
This teaches GlobalISel to emit a RTLib call for @llvm.log2 when it encounters
it.
It updates the existing floating point tests to show that we don't fall back on
the intrinsic, and select the correct instructions. It also adds a legalizer
test for G_FLOG2.
Jessica Paquette [Wed, 30 Jan 2019 21:03:52 +0000 (21:03 +0000)]
[GlobalISel][AArch64] Add instruction selection support for @llvm.sqrt
This teaches the legalizer about G_FSQRT in AArch64. Also adds a legalizer
test for G_FSQRT, a selection test for it, and updates existing floating point
tests.
Erik Pilkington [Wed, 30 Jan 2019 20:34:35 +0000 (20:34 +0000)]
Add a 'dynamic' parameter to the objectsize intrinsic
This is meant to be used with clang's __builtin_dynamic_object_size.
When 'true' is passed to this parameter, the intrinsic has the
potential to be folded into instructions that will be evaluated
at run time. When 'false', the objectsize intrinsic behaviour is
unchanged.
Philip Reames [Wed, 30 Jan 2019 19:21:11 +0000 (19:21 +0000)]
SimplifyDemandedVectorElts for all intrinsics
The point is that this simplifies integration of new intrinsics into SimplifiedDemandedVectorElts, and ensures we don't miss any existing ones.
This is intended to be NFC-ish, but as seen from the diffs, can produce slightly different output. This is due to order of transforms w/in instcombine resulting in two slightly different fixed points. That's something we should fix, but isn't a problem w/this patch per se.
Wolfgang Pieb [Wed, 30 Jan 2019 18:34:07 +0000 (18:34 +0000)]
[DEBUGINFO] Handle restore instructions in LiveDebugValues
The LiveDebugValues pass recognizes spills but not restores, which can
cause large gaps in location information for some variables, depending
on control flow. This patch make LiveDebugValues recognize restores and
generate appropriate DBG_VALUE instructions.
Matt Arsenault [Wed, 30 Jan 2019 17:52:23 +0000 (17:52 +0000)]
GlobalISel: Add assert that legalize mutation makes sense
I've repeatedly encountered bugs resulting from custom legalize
mutations returning nonsense legalize results, such as increasing the
number of elements for FewerElements. Add an assert function to make
sure the type to mutate to is consistent with the legalize action.
Simon Pilgrim [Wed, 30 Jan 2019 16:15:59 +0000 (16:15 +0000)]
[utils] Fix update scripts output when run on python3.
This fixes a "bytes-like object is required, not 'str'" python3 error I hit on update_llc_test_checks.py (but present on the other scripts as well) by matching what update_mca_test_checks.py already does, plus I've added an explicit 'utf-8' encoding.
Stefan Granitz [Wed, 30 Jan 2019 15:10:08 +0000 (15:10 +0000)]
[CMake] Accept ENTITLEMENTS in llvm_add_library()
Summary: We added support for code signing entitlements in add_llvm_executable() with D54443. In the future it would be useful to have this functionality available also for libraries.
Jordan Rupprecht [Wed, 30 Jan 2019 14:58:13 +0000 (14:58 +0000)]
[llvm-objcopy] Support -X|--discard-locals.
Summary:
This adds support for the --discard-locals flag, which acts similarly to --discard-all, except it only applies to compiler-generated symbols (i.e. symbols starting with `.L` in ELF).
I am not sure about COFF local symbols: those appear to also use `.L` in most cases, but also use just `L` in other cases, so for now I am just leaving it unimplemented there.
George Rimar [Wed, 30 Jan 2019 14:08:55 +0000 (14:08 +0000)]
[llvm-readobj] - Simplify the code.
We have a Field struct which has a StringRef member Str.
The code needs to create and keep alive the temporarily
std::string variables because of that.
That is not convenient and makes the code be more complicated
than it could be.
There seems to be no reason to keep Str be StringRef.
The patch changes it to be std::string and
rearranges the code around slightly.
Shiva Chen [Wed, 30 Jan 2019 11:16:59 +0000 (11:16 +0000)]
[RISCV] Insert R_RISCV_ALIGN relocation type and Nops for code alignment when linker relaxation enabled
Linker relaxation may change code size. We need to fix up the alignment
of alignment directive in text section by inserting Nops and R_RISCV_ALIGN
relocation type. So then linker could satisfy the alignment by removing Nops.
To do this:
1. Add shouldInsertExtraNopBytesForCodeAlign target hook to calculate
the Nops we need to insert.
2. Add shouldInsertFixupForCodeAlign target hook to insert
R_RISCV_ALIGN fixup type.
Craig Topper [Wed, 30 Jan 2019 08:04:06 +0000 (08:04 +0000)]
[X86] Remove unnecessary code from the top of handleCompareFP in X86FloatingPoint.cpp.
There were checks to ensure some tables were sorted, but those tables aren't used by this function. The same tables are checked in the function that does use them. Maybe this was copy/pasted?
Craig Topper [Wed, 30 Jan 2019 07:33:24 +0000 (07:33 +0000)]
[X86] Remove a couple places where we unnecessarily pass 0 to the EmitPriority of some FP instruction aliases. NFC
As far as I can tell we already won't emit these aliases due to an operand count check in the tablegen code. Removing these because I couldn't make sense of the inconsistency between fadd and fmul from reading the code.
I checked the AsmMatcher and AsmWriter files before and after this change and there were no differences.
Craig Topper [Wed, 30 Jan 2019 03:43:41 +0000 (03:43 +0000)]
[IR] Use CallBase to simplify some code
Summary:
This patch does the following to simplify the asm-goto patch
-Move isInlineAsm from CallInst to CallBase to share with CallBrInst in the asm-goto patch.
-Forward CallSite's data_operands_begin()/data_operands_end() to CallBase's implementation.
-Forward CallSite's getOperandBundlesAsDefs to CallBase.
Heejin Ahn [Wed, 30 Jan 2019 03:21:57 +0000 (03:21 +0000)]
[WebAssembly] Exception handling: Switch to the new proposal
Summary:
This switches the EH implementation to the new proposal:
https://github.com/WebAssembly/exception-handling/blob/master/proposals/Exceptions.md
(The previous proposal was
https://github.com/WebAssembly/exception-handling/blob/master/proposals/old/Exceptions.md)
- Instruction changes
- Now we have one single `catch` instruction that returns a except_ref
value
- `throw` now can take variable number of operations
- `rethrow` does not have 'depth' argument anymore
- `br_on_exn` queries an except_ref to see if it matches the tag and
branches to the given label if true.
- `extract_exception` is a pseudo instruction that simulates popping
values from wasm stack. This is to make `br_on_exn`, a very special
instruction, work: `br_on_exn` puts values onto the stack only if it
is taken, and the # of values can vay depending on the tag.
- Now there's only one `catch` per `try`, this patch removes all special
handling for terminate pad with a call to `__clang_call_terminate`.
Before it was the only case there are two catch clauses (a normal
`catch` and `catch_all` per `try`).
- Make `rethrow` act as a terminator like `throw`. This splits BB after
`rethrow` in WasmEHPrepare, and deletes an unnecessary `unreachable`
after `rethrow` in LateEHPrepare.
- Now we stop at all catchpads (because we add wasm `catch` instruction
that catches all exceptions), this creates new
`findWasmUnwindDestinations` function in SelectionDAGBuilder.
- Now we use `br_on_exn` instrution to figure out if an except_ref
matches the current tag or not, LateEHPrepare generates this sequence
for catch pads:
```
catch
block i32
br_on_exn $__cpp_exception
end_block
extract_exception
```
- Branch analysis for `br_on_exn` in WebAssemblyInstrInfo
- Other various misc. changes to switch to the new proposal.
Thomas Lively [Wed, 30 Jan 2019 02:23:29 +0000 (02:23 +0000)]
[WebAssembly] Optimize BUILD_VECTOR lowering for size
Summary:
Implements custom lowering logic that finds the optimal value for the
initial splat of the vector and either uses it or uses v128.const if
it is available and if it would produce smaller code. This logic
replaces large TableGen ISEL patterns that would lower all non-splat
BUILD_VECTORs into a splat followed by a fixed number of replace_lane
instructions. This CL fixes PR39685.