Simon Atanasyan [Tue, 9 Jul 2019 12:55:55 +0000 (12:55 +0000)]
[mips] Implement sge/sgeu pseudo instructions
The `sge/sgeu Dst, Src1, Src2/Imm` pseudo instructions set register
`Dst` to 1 if register `Src1` is greater than or equal `Src2/Imm` and
to 0 otherwise.
Tim Northover [Tue, 9 Jul 2019 11:35:35 +0000 (11:35 +0000)]
OpaquePtr: add Type parameter to Loads analysis API.
This makes the functions in Loads.h require a type to be specified
independently of the pointer Value so that when pointers have no structure
other than address-space, it can still do its job.
Most callers had an obvious memory operation handy to provide this type, but a
SROA and ArgumentPromotion were doing more complicated analysis. They get
updated to merge the properties of the various instructions they were
considering.
Dump the DWARF information about call sites and call site parameters into
debug info sections.
The patch also provides an interface for the interpretation of instructions
that could load values of a call site parameters in order to generate DWARF
about the call site parameters.
([13/13] Introduce the debug entry values.)
Co-authored-by: Ananth Sowda <asowda@cisco.com> Co-authored-by: Nikola Prica <nikola.prica@rt-rk.com> Co-authored-by: Ivan Baev <ibaev@cisco.com>
Differential Revision: https://reviews.llvm.org/D60716
Alex Bradbury [Tue, 9 Jul 2019 10:56:18 +0000 (10:56 +0000)]
[RISCV] Fix RISCVTTIImpl::getIntImmCost for immediates where getMinSignedBits() > 64
APInt::getSExtValue will assert if getMinSignedBits() > 64. This can happen,
for instance, if examining an i128. Avoid this assertion by checking
Imm.getMinSignedBits() <= 64 before doing
getTLI()->isLegalAddImmediate(Imm.getSExtValue()). We could directly check
getMinSignedBits() <= 12 but it seems better to reuse the isLegalAddImmediate
helper for this.
[LegalizeTypes] Fix saturation bug for smul.fix.sat
Summary:
Make sure we use SETGE instead of SETGT when checking
if the sign bit is zero at SMULFIXSAT expansion.
The faulty expansion occured when doing "expand" of
SMULFIXSAT and the scale was exactly matching the
size of the smaller type. For example doing
i64 Z = SMULFIXSAT X, Y, 32
and expanding X/Y/Z into using two i32 values.
The problem was that we sometimes did not saturate
to min when overflowing.
Here is an example using Q3.4 numbers:
Consider that we are multiplying X and Y.
X = 0x80 (-8.0 as Q3.4)
Y = 0x20 (2.0 as Q3.4)
To avoid loss of precision we do a widening
multiplication, getting a 16 bit result
Z = 0xF000 (-16.0 as Q7.8)
To detect negative overflow we should check if
the five most significant bits in Z are less than -1.
Assume that we name the 4 most significant bits
as HH and the next 4 bits as HL. Then we can do the
check by examining if
(HH < -1) or (HH == -1 && "sign bit in HL is zero").
The fault was that we have been doing the check as
(HH < -1) or (HH == -1 && HL > 0)
instead of
(HH < -1) or (HH == -1 && HL >= 0).
In our example HH is -1 and HL is 0, so the old
code did not trigger saturation and simply truncated
the result to 0x00 (0.0). With the bugfix we instead
detect that we should saturate to min, and the result
will be set to 0x80 (-8.0).
Reland "[LiveDebugValues] Emit the debug entry values"
Emit replacements for clobbered parameters location if the parameter
has unmodified value throughout the funciton. This is basic scenario
where we can use the debug entry values.
([12/13] Introduce the debug entry values.)
Co-authored-by: Ananth Sowda <asowda@cisco.com> Co-authored-by: Nikola Prica <nikola.prica@rt-rk.com> Co-authored-by: Ivan Baev <ibaev@cisco.com>
Differential Revision: https://reviews.llvm.org/D58042
[Loop Peeling] Add support for peeling of loops with multiple exits
This patch modifies the loop peeling transformation so that
it does not expect that there is only one loop exit from latch.
It modifies only transformation. Update of branch weights remains
only for exit from latch.
The motivation is that in follow-up patch I plan to enable loop peeling for
loops with multiple exits but only if other exits then from latch one goes to
block with call to deopt.
Prepare for making SwitchInstProfUpdateWrapper strict
This patch removes the test part that relates to the non-strict
behavior of SwitchInstProfUpdateWrapper and changes
the assertion to llvm_unreachable() to allow the check in
release builds.
This patch prepares SwitchInstProfUpdateWrapper to become
strict with one line change. That is need to revert it easily
if any failure will arise.
[LoopInfo] Update getExitEdges to accept vector of pairs for non const BasicBlock
D63921 requires getExitEdges fills a vector of Edge pairs where
BasicBlocks are not constant.
The rest Loop API mostly returns non-const BasicBlocks, so to be more consistent with
other Loop API getExitEdges is modified to return non-const BasicBlocks as well.
Kai Luo [Tue, 9 Jul 2019 02:55:08 +0000 (02:55 +0000)]
[PowerPC][Peephole] Combine extsw and sldi after instruction selection
Summary:
`extsw` and `sldi` are supposed to be combined if they are in the same
BB in instruction selection phase. This patch handles the case where
extsw and sldi are not in the same BB.
Jinsong Ji [Tue, 9 Jul 2019 02:27:35 +0000 (02:27 +0000)]
[MachinePipeliner] Fix Phi refers to Phi in same stage in 1st epilogue
Summary:
This is exposed by functional testing on PowerPC.
In some pipelined loops, Phi refer to phi did not get value defined by
the Phi, hence getting wrong value later.
As the comment mentioned, we should "use the value defined by the Phi,
unless we're generating the firstepilog and the Phi refers to a Phi
in a different stage.", so Phi refering to same stage Phi should use
the value defined by the Phi here.
[WebAssembly] Make sret parameter work with AddMissingPrototypes
Summary:
Even with functions with `no-prototype` attribute, there can be an
argument `sret` (structure return) attribute, which is an optimization
when a function return type is a struct. Fixes PR42420.
Yonghong Song [Tue, 9 Jul 2019 01:51:36 +0000 (01:51 +0000)]
[BPF] add new intrinsics preserve_{array,union,struct}_access_index
For background of BPF CO-RE project, please refer to
http://vger.kernel.org/bpfconf2019.html
In summary, BPF CO-RE intends to compile bpf programs
adjustable on struct/union layout change so the same
program can run on multiple kernels with adjustment
before loading based on native kernel structures.
In order to do this, we need keep track of GEP(getelementptr)
instruction base and result debuginfo types, so we
can adjust on the host based on kernel BTF info.
Capturing such information as an IR optimization is hard
as various optimization may have tweaked GEP and also
union is replaced by structure it is impossible to track
fieldindex for union member accesses.
Three intrinsic functions, preserve_{array,union,struct}_access_index,
are introducted.
addr = preserve_array_access_index(base, index, dimension)
addr = preserve_union_access_index(base, di_index)
addr = preserve_struct_access_index(base, gep_index, di_index)
here,
base: the base pointer for the array/union/struct access.
index: the last access index for array, the same for IR/DebugInfo layout.
dimension: the array dimension.
gep_index: the access index based on IR layout.
di_index: the access index based on user/debuginfo types.
For example, for the following example,
$ cat test.c
struct sk_buff {
int i;
int b1:1;
int b2:2;
union {
struct {
int o1;
int o2;
} o;
struct {
char flags;
char dev_id;
} dev;
int netid;
} u[10];
};
static int (*bpf_probe_read)(void *dst, int size, const void *unsafe_ptr)
= (void *) 4;
Note that @llvm.preserve.{struct,union}.access.index calls have metadata llvm.preserve.access.index
attached to instructions to provide struct/union debuginfo type information.
For &ctx->u[5].dev.dev_id,
. The "%6 = ..." represents struct member "u" with index 2 for IR layout and index 3 for DI layout.
. The "%7 = ..." represents array subscript "5".
. The "%8 = ..." represents union member "dev" with index 1 for DI layout.
. The "%10 = ..." represents struct member "dev_id" with index 1 for both IR and DI layout.
Basically, traversing the use-def chain recursively for the 3rd argument of bpf_probe_read() and
examining all preserve_*_access_index calls, the debuginfo struct/union/array access index
can be achieved.
The intrinsics also contain enough information to regenerate codes for IR layout.
For array and structure intrinsics, the proper GEP can be constructed.
For union intrinsics, replacing all uses of "addr" with "base" should be enough.
The test case ThinLTO/X86/lazyload_metadata.ll is adjusted to reflect the
new addition of the metadata.
Signed-off-by: Yonghong Song <yhs@fb.com>
Differential Revision: https://reviews.llvm.org/D61810
Philip Reames [Tue, 9 Jul 2019 01:27:45 +0000 (01:27 +0000)]
[LoopPred] Extend LFTR normalization to the inverse EQ case
A while back, I added support for NE latches formed by LFTR. I didn't think that quite through, as LFTR will also produce the inverse EQ form for some loops and I hadn't handled that. This change just adds handling for that case as well.
Let unaliased Args track which Alias they were created from, and use that in Arg::getAsString() for diagnostics
With this, `clang-cl /source-charset:utf-16 test.cc` now prints `invalid
value 'utf-16' in '/source-charset:utf-16'` instead of `invalid value
'utf-16' in '-finput-charset=utf-16'` before, and several other clang-cl
flags produce much less confusing output as well.
Fixes PR29106.
Since an arg and its alias can have different arg types (joined vs not)
and different values (because of AliasArgs<>), I chose to give the Alias
its own Arg object. For convenience, I just store the alias directly in
the unaliased arg – there aren't many arg objects at runtime, so that
seems ok.
Finally, I changed Arg::getAsString() to use the alias's representation
if it's present – that function was already documented as being the
suitable function for diagnostics, and most callers already used it for
diagnostics.
Implementation-wise, Arg::accept() previously used to parse things as
the unaliased option. The core of that switch is now extracted into a
new function acceptInternal() which parses as the _aliased_ option, and
the previously-intermingled unaliasing is now done as an explicit step
afterwards.
(This also changes one place in lld that didn't use getAsString() for
diagnostics, so that that one place now also prints the flag as the user
wrote it, not as it looks after it went through unaliasing.)
[Attributor] Deduce the "returned" argument attribute
Deduce the "returned" argument attribute by collecting all potentially
returned values.
Not only the unique return value, if any, can be used by subsequent
attributes but also the set of all potentially returned values as well
as the mapping from returned values to return instructions that they
originate from (see AAReturnedValues::checkForallReturnedValues).
Change in statistics (-stats) for LLVM-TS + Spec2006, totaling ~19% more "returned" arguments.
Rong Xu [Mon, 8 Jul 2019 21:03:12 +0000 (21:03 +0000)]
llvm-profdata] Handle the cases of overlapping input file and output file
Currently llvm-profdata does not expect the same file name for the input profile
and the output profile.
>llvm-profdata merge A.profraw B.profraw -o B.profraw
The above command runs successfully but the resulted B.profraw is not correct.
This patch fixes the issue by moving the initialization of writer after loading
the profile.
For the show command, the following will report a confusing error of
"Empty raw profile file":
>llvm-profdata show B.profraw -o B.profraw
It's harder to fix as we need to output something before loading the input profile.
I don't think that a fix for this is worth the effort. I just make the error explicit for
the show command.
[InstCombine] fold insertelement into splat of same scalar
Forming the canonical splat shuffle improves analysis and
may allow follow-on transforms (although some possibilities
are missing as shown in the test diffs).
The backend generically turns these patterns into build_vector,
so there should be no codegen regressions. All targets are
expected to be able to lower splats efficiently.
Teach the symbolizer lib symbolize objects directly.
Currently, the symbolizer lib can only symbolize a file on disk.
This patch teaches the symbolizer lib to symbolize objects.
llvm-objdump needs this to support archive disassembly with source info.
Matt Arsenault [Mon, 8 Jul 2019 19:03:38 +0000 (19:03 +0000)]
AMDGPU: Make s34 the FP register
Make the FP register callee saved.
This is tricky because now the FP needs to be spilled in the prolog
relative to the incoming SP register, rather than the frame register
used throughout the rest of the function. I don't like how this
bypassess the standard mechanism for CSR spills just to get the
correct insert point. I may look for a better solution, since all CSR
VGPRs may also need to have all lanes activated. Another option might
be to make getFrameIndexReference change the base register if the
frame index is a CSR, and then try to figure out the right insertion
point in emitProlog.
If there is a free VGPR lane available for SGPR spilling, try to use
it for the FP. If that would require intrtoducing a new VGPR spill,
try to use a free call clobbered SGPR. Only fallback to introducing a
new VGPR spill as a last resort.
This also doesn't attempt to handle SGPR spilling with scalar stores.
Keep the order of the basic blocks in the cloned loop as the original
loop
Summary:
Do the cloning in two steps, first allocate all the new loops, then
clone the basic blocks in the same order as the original loop.
Reviewer: Meinersbur, fhahn, kbarton, hfinkel
Reviewed By: hfinkel
Subscribers: hfinkel, hiraditya, llvm-commits
Tag: https://reviews.llvm.org/D64224
Differential Revision:
Chris Bieneman [Mon, 8 Jul 2019 18:29:29 +0000 (18:29 +0000)]
Fix issues building libraries as more than one type with Xcode
Summary:
CMake+Xcode doesn't seem to handle targets that only have object
sources. This patch works around that limitation by adding a dummy
soruce file to any library target that is generated by llvm_add_library
when object libraries are generated.
Object libraries are generated whenever llvm_add_library is passed more
than one library type, which is now the default case for clang static
libraries (which generate STATIC and OBJECT libraries).
Yonghong Song [Mon, 8 Jul 2019 17:08:28 +0000 (17:08 +0000)]
[BPF] add new intrinsics preserve_{array,union,struct}_access_index
For background of BPF CO-RE project, please refer to
http://vger.kernel.org/bpfconf2019.html
In summary, BPF CO-RE intends to compile bpf programs
adjustable on struct/union layout change so the same
program can run on multiple kernels with adjustment
before loading based on native kernel structures.
In order to do this, we need keep track of GEP(getelementptr)
instruction base and result debuginfo types, so we
can adjust on the host based on kernel BTF info.
Capturing such information as an IR optimization is hard
as various optimization may have tweaked GEP and also
union is replaced by structure it is impossible to track
fieldindex for union member accesses.
Three intrinsic functions, preserve_{array,union,struct}_access_index,
are introducted.
addr = preserve_array_access_index(base, index, dimension)
addr = preserve_union_access_index(base, di_index)
addr = preserve_struct_access_index(base, gep_index, di_index)
here,
base: the base pointer for the array/union/struct access.
index: the last access index for array, the same for IR/DebugInfo layout.
dimension: the array dimension.
gep_index: the access index based on IR layout.
di_index: the access index based on user/debuginfo types.
For example, for the following example,
$ cat test.c
struct sk_buff {
int i;
int b1:1;
int b2:2;
union {
struct {
int o1;
int o2;
} o;
struct {
char flags;
char dev_id;
} dev;
int netid;
} u[10];
};
static int (*bpf_probe_read)(void *dst, int size, const void *unsafe_ptr)
= (void *) 4;
Note that @llvm.preserve.{struct,union}.access.index calls have metadata llvm.preserve.access.index
attached to instructions to provide struct/union debuginfo type information.
For &ctx->u[5].dev.dev_id,
. The "%6 = ..." represents struct member "u" with index 2 for IR layout and index 3 for DI layout.
. The "%7 = ..." represents array subscript "5".
. The "%8 = ..." represents union member "dev" with index 1 for DI layout.
. The "%10 = ..." represents struct member "dev_id" with index 1 for both IR and DI layout.
Basically, traversing the use-def chain recursively for the 3rd argument of bpf_probe_read() and
examining all preserve_*_access_index calls, the debuginfo struct/union/array access index
can be achieved.
The intrinsics also contain enough information to regenerate codes for IR layout.
For array and structure intrinsics, the proper GEP can be constructed.
For union intrinsics, replacing all uses of "addr" with "base" should be enough.
Signed-off-by: Yonghong Song <yhs@fb.com>
Differential Revision: https://reviews.llvm.org/D61810
Summary:
Before, they were one category of operands which could cause
crashes in non-sensical combinations, e.g. "f32.const symbol".
Now these are forced to be an error.
George Rimar [Mon, 8 Jul 2019 16:53:39 +0000 (16:53 +0000)]
[llvm\test\Object] - An initial step to cleanup the test cases.
This patch removes trivial-object-test.elf-i386,
trivial-object-test.elf-x86-64 and trivial-object-test2.elf-x86-64
precompiled objects from test/Object/Inputs folder.
I adjusted the existent test cases to use YAML instead.
[InstCombine] canonicalize insert+splat to/from element 0 of vector
We recognize a splat from element 0 in (VectorUtils) llvm::getSplatValue()
and also in ShuffleVectorInst::isZeroEltSplatMask(), so this converts
to that form for better matching.
The backend generically turns these patterns into build_vector,
so there should be no codegen difference.
Kevin P. Neal [Mon, 8 Jul 2019 16:18:18 +0000 (16:18 +0000)]
Teach the IRBuilder about fadd and friends.
The IRBuilder has calls to create floating point instructions like fadd.
It does not have calls to create constrained versions of them. This patch
adds support for constrained creation of fadd, fsub, fmul, fdiv, and frem.
Reviewed by: John McCall, Sanjay Patel
Approved by: John McCall
Differential Revision: https://reviews.llvm.org/D53157
Brian Homerding [Mon, 8 Jul 2019 15:57:56 +0000 (15:57 +0000)]
Add, and infer, a nofree function attribute
This patch adds a function attribute, nofree, to indicate that a function does
not, directly or indirectly, call a memory-deallocation function (e.g., free,
C++'s operator delete).
Alex Bradbury [Mon, 8 Jul 2019 14:52:36 +0000 (14:52 +0000)]
[Triple] Add isRISCV function
This matches isARM, isThumb, isAArch64 and similar helpers. Future commits
which clean-up code that currently checks for Triple::riscv32 ||
Triple::riscv64.
Differential Revision: https://reviews.llvm.org/D54215
Patch by Simon Cook.
Test case added by Alex Bradbury.
Petar Avramovic [Mon, 8 Jul 2019 14:45:52 +0000 (14:45 +0000)]
[MIPS GlobalISel] Register bank select for G_LOAD. Select i64 load
Select gprb or fprb when loaded value is used by either:
copy to physical register or
instruction with only one mapping available for that use operand.
Load of integer s64 is handled with narrowScalar when mapping is applied,
produced artifacts are combined away. Manually set gprb to all register
operands of instructions created during narrowScalar.
Petar Avramovic [Mon, 8 Jul 2019 14:36:36 +0000 (14:36 +0000)]
[MIPS GlobalISel] Register bank select for G_STORE. Select i64 store
Select gprb or fprb when stored value is defined by either:
copy from physical register or
instruction with only one mapping available for that def operand.
Store of integer s64 is handled with narrowScalar when mapping is applied,
produced artifacts are combined away. Manually set gprb to all register
operands of instructions created during narrowScalar.
[AMDGPU][MC] Corrected parsing of FLAT offset modifier
Summary of changes:
- simplified handling of FLAT offset: offset_s13 and offset_u12 have been replaced with flat_offset;
- provided information about error position for pre-gfx9 targets;
- improved errors handling.
[ARM] Relax constraints on operands of VQxDMLxDH instructions
Summary:
According to a recently updated Armv8-M spec
(https://static.docs.arm.com/ddi0553/bh/DDI0553B_h_armv8m_arm.pdf) the
32-bit width versions of the following instructions:
* VQDMLADH
* VQDMLADHX
* VQRDMLADH
* VQRDMLADHX
* VQDMLSDH
* VQDMLSDHX
* VQRDMLSDH
* VQRDMLSDHX
are no longer unpredictable when their output register is the same as
one of the input registers.
This patch updates the assembler parser and the corresponding tests
and also removes @earlyclobber from the instruction constraints.
projects / llvm / log