The transformation happening here is that we want to turn a
"mul(ext(X), ext(X))" into a "vmull(X, X)", stripping off the extension. We have
to make sure that X still has a valid vector type - possibly recreate an
extension to a smaller type. In case of a extload of a memory type smaller than
64 bit we used create a ext(load()). The problem with doing this - instead of
recreating an extload - is that an illegal type is exposed.
This patch fixes this by creating extloads instead of ext(load()) sequences.
Fixes PR15970.
radar://13871383
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181842 91177308-0d34-0410-b5e6-96231b3b80d8
ARM FastISel is currently only enabled for iOS non-Thumb1, and I'm working on
enabling it for other targets. As a first step I've fixed some of the tests.
Changes to ARM FastISel tests:
- Different triples don't generate the same relocations (especially
movw/movt versus constant pool loads). Use a regex to allow either.
- Mangling is different. Use a regex to allow either.
- The reserved registers are sometimes different, so registers get
allocated in a different order. Capture the names only where this
occurs.
- Add -verify-machineinstrs to some tests where it works. It doesn't
work everywhere it should yet.
- Add -fast-isel-abort to many tests that didn't have it before.
- Split out the VarArg test from fast-isel-call.ll into its own
test. This simplifies test setup because of --check-prefix.
Patch by JF Bastien
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181801 91177308-0d34-0410-b5e6-96231b3b80d8
return values are bitcasts.
The chain had previously been being clobbered with the entry node to
the dag, which sometimes caused other code in the function to be
erroneously deleted when tailcall optimization kicked in.
<rdar://problem/13827621>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181696 91177308-0d34-0410-b5e6-96231b3b80d8
Now even the small structures could be passed within byval (small enough
to be stored in GPRs).
In regression tests next function prototypes are checked:
PR15293:
%artz = type { i32 }
define void @foo(%artz* byval %s)
define void @foo2(%artz* byval %s, i32 %p, %artz* byval %s2)
foo: "s" stored in R0
foo2: "s" stored in R0, "s2" stored in R2.
Next AAPCS rules are checked:
5.5 Parameters Passing, C.4 and C.5,
"ParamSize" is parameter size in 32bit words:
-- NSAA != 0, NCRN < R4 and NCRN+ParamSize > R4.
Parameter should be sent to the stack; NCRN := R4.
-- NSAA != 0, and NCRN < R4, NCRN+ParamSize < R4.
Parameter stored in GPRs; NCRN += ParamSize.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@181148 91177308-0d34-0410-b5e6-96231b3b80d8
Optimize CONCAT_VECTOR nodes that merge EXTRACT_SUBVECTOR values that extract from the same vector.
rdar://13402653
PR15866
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180871 91177308-0d34-0410-b5e6-96231b3b80d8
register-indirect address with an offset of 0.
It used to be that a DBG_VALUE is a register-indirect value if the offset
(operand 1) is nonzero. The new convention is that a DBG_VALUE is
register-indirect if the first operand is a register and the second
operand is an immediate. For plain registers use the combination reg, reg.
rdar://problem/13658587
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180816 91177308-0d34-0410-b5e6-96231b3b80d8
For now, we just reschedule instructions that use the copied vregs and
let regalloc elliminate it. I would really like to eliminate the
copies on-the-fly during scheduling, but we need a complete
implementation of repairIntervalsInRange() first.
The general strategy is for the register coalescer to eliminate as
many global copies as possible and shrink live ranges to be
extended-basic-block local. The coalescer should not have to worry
about resolving local copies (e.g. it shouldn't attemp to reorder
instructions). The scheduler is a much better place to deal with local
interference. The coalescer side of this equation needs work.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180193 91177308-0d34-0410-b5e6-96231b3b80d8
-- C.4 and C.5 statements, when NSAA is not equal to SP.
-- C.1.cp statement for VA functions. Note: There are no VFP CPRCs in a
variadic procedure.
Before this patch "NSAA != 0" means "don't use GPRs anymore ". But there are
some exceptions in AAPCS.
1. For non VA function: allocate all VFP regs for CPRC. When all VFPs are allocated
CPRCs would be sent to stack, while non CPRCs may be still allocated in GRPs.
2. Check that for VA functions all params uses GPRs and then stack.
No exceptions, no CPRCs here.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@180011 91177308-0d34-0410-b5e6-96231b3b80d8
This reverts commit r179840 with a fix to test/DebugInfo/two-cus-from-same-file.ll
I'm not sure why that test only failed on ARM & MIPS and not X86 Linux, even
though the debug info was clearly invalid on all of them, but this ought to fix
it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179996 91177308-0d34-0410-b5e6-96231b3b80d8
Rather than just splitting the input type and hoping for the best, apply
a bit more cleverness. Just splitting the types until the source is
legal often leads to an illegal result time, which is then widened and a
scalarization step is introduced which leads to truly horrible code
generation. With the loop vectorizer, these sorts of operations are much
more common, and so it's worth extra effort to do them well.
Add a legalization hook for the operands of a TRUNCATE node, which will
be encountered after the result type has been legalized, but if the
operand type is still illegal. If simple splitting of both types
ends up with the result type of each half still being legal, just
do that (v16i16 -> v16i8 on ARM, for example). If, however, that would
result in an illegal result type (v8i32 -> v8i8 on ARM, for example),
we can get more clever with power-two vectors. Specifically,
split the input type, but also widen the result element size, then
concatenate the halves and truncate again. For example on ARM,
To perform a "%res = v8i8 trunc v8i32 %in" we transform to:
%inlo = v4i32 extract_subvector %in, 0
%inhi = v4i32 extract_subvector %in, 4
%lo16 = v4i16 trunc v4i32 %inlo
%hi16 = v4i16 trunc v4i32 %inhi
%in16 = v8i16 concat_vectors v4i16 %lo16, v4i16 %hi16
%res = v8i8 trunc v8i16 %in16
This allows instruction selection to generate three VMOVN instructions
instead of a sequences of moves, stores and loads.
Update the ARMTargetTransformInfo to take this improved legalization
into account.
Consider the simplified IR:
define <16 x i8> @test1(<16 x i32>* %ap) {
%a = load <16 x i32>* %ap
%tmp = trunc <16 x i32> %a to <16 x i8>
ret <16 x i8> %tmp
}
define <8 x i8> @test2(<8 x i32>* %ap) {
%a = load <8 x i32>* %ap
%tmp = trunc <8 x i32> %a to <8 x i8>
ret <8 x i8> %tmp
}
Previously, we would generate the truly hideous:
.syntax unified
.section __TEXT,__text,regular,pure_instructions
.globl _test1
.align 2
_test1: @ @test1
@ BB#0:
push {r7}
mov r7, sp
sub sp, sp, #20
bic sp, sp, #7
add r1, r0, #48
add r2, r0, #32
vld1.64 {d24, d25}, [r0:128]
vld1.64 {d16, d17}, [r1:128]
vld1.64 {d18, d19}, [r2:128]
add r1, r0, #16
vmovn.i32 d22, q8
vld1.64 {d16, d17}, [r1:128]
vmovn.i32 d20, q9
vmovn.i32 d18, q12
vmov.u16 r0, d22[3]
strb r0, [sp, #15]
vmov.u16 r0, d22[2]
strb r0, [sp, #14]
vmov.u16 r0, d22[1]
strb r0, [sp, #13]
vmov.u16 r0, d22[0]
vmovn.i32 d16, q8
strb r0, [sp, #12]
vmov.u16 r0, d20[3]
strb r0, [sp, #11]
vmov.u16 r0, d20[2]
strb r0, [sp, #10]
vmov.u16 r0, d20[1]
strb r0, [sp, #9]
vmov.u16 r0, d20[0]
strb r0, [sp, #8]
vmov.u16 r0, d18[3]
strb r0, [sp, #3]
vmov.u16 r0, d18[2]
strb r0, [sp, #2]
vmov.u16 r0, d18[1]
strb r0, [sp, #1]
vmov.u16 r0, d18[0]
strb r0, [sp]
vmov.u16 r0, d16[3]
strb r0, [sp, #7]
vmov.u16 r0, d16[2]
strb r0, [sp, #6]
vmov.u16 r0, d16[1]
strb r0, [sp, #5]
vmov.u16 r0, d16[0]
strb r0, [sp, #4]
vldmia sp, {d16, d17}
vmov r0, r1, d16
vmov r2, r3, d17
mov sp, r7
pop {r7}
bx lr
.globl _test2
.align 2
_test2: @ @test2
@ BB#0:
push {r7}
mov r7, sp
sub sp, sp, #12
bic sp, sp, #7
vld1.64 {d16, d17}, [r0:128]
add r0, r0, #16
vld1.64 {d20, d21}, [r0:128]
vmovn.i32 d18, q8
vmov.u16 r0, d18[3]
vmovn.i32 d16, q10
strb r0, [sp, #3]
vmov.u16 r0, d18[2]
strb r0, [sp, #2]
vmov.u16 r0, d18[1]
strb r0, [sp, #1]
vmov.u16 r0, d18[0]
strb r0, [sp]
vmov.u16 r0, d16[3]
strb r0, [sp, #7]
vmov.u16 r0, d16[2]
strb r0, [sp, #6]
vmov.u16 r0, d16[1]
strb r0, [sp, #5]
vmov.u16 r0, d16[0]
strb r0, [sp, #4]
ldm sp, {r0, r1}
mov sp, r7
pop {r7}
bx lr
Now, however, we generate the much more straightforward:
.syntax unified
.section __TEXT,__text,regular,pure_instructions
.globl _test1
.align 2
_test1: @ @test1
@ BB#0:
add r1, r0, #48
add r2, r0, #32
vld1.64 {d20, d21}, [r0:128]
vld1.64 {d16, d17}, [r1:128]
add r1, r0, #16
vld1.64 {d18, d19}, [r2:128]
vld1.64 {d22, d23}, [r1:128]
vmovn.i32 d17, q8
vmovn.i32 d16, q9
vmovn.i32 d18, q10
vmovn.i32 d19, q11
vmovn.i16 d17, q8
vmovn.i16 d16, q9
vmov r0, r1, d16
vmov r2, r3, d17
bx lr
.globl _test2
.align 2
_test2: @ @test2
@ BB#0:
vld1.64 {d16, d17}, [r0:128]
add r0, r0, #16
vld1.64 {d18, d19}, [r0:128]
vmovn.i32 d16, q8
vmovn.i32 d17, q9
vmovn.i16 d16, q8
vmov r0, r1, d16
bx lr
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179989 91177308-0d34-0410-b5e6-96231b3b80d8
Previously, when spilling 64-bit paired registers, an LDMIA with both
a FrameIndex and an offset was produced. This kind of instruction
shouldn't exist, and the extra operand was being confused with the
predicate, causing aborts later on.
This removes the invalid 0-offset from the instruction being
produced.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179956 91177308-0d34-0410-b5e6-96231b3b80d8
Adding another CU-wide list, in this case of imported_modules (since they
should be relatively rare, it seemed better to add a list where each element
had a "context" value, rather than add a (usually empty) list to every scope).
This takes care of DW_TAG_imported_module, but to fully address PR14606 we'll
need to expand this to cover DW_TAG_imported_declaration too.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179836 91177308-0d34-0410-b5e6-96231b3b80d8
This pattern occurs in SROA output due to the way vector arguments are lowered
on ARM.
The testcase from PR15525 now compiles into this, which is better than the code
we got with the old scalarrepl:
_Store:
ldr.w r9, [sp]
vmov d17, r3, r9
vmov d16, r1, r2
vst1.8 {d16, d17}, [r0]
bx lr
Differential Revision: http://llvm-reviews.chandlerc.com/D647
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@179106 91177308-0d34-0410-b5e6-96231b3b80d8
The Thumb2SizeReduction pass avoids false CPSR dependencies, except it
still aggressively creates tMOVi8 instructions because they are so
common.
Avoid creating false CPSR dependencies even for tMOVi8 instructions when
the the CPSR flags are known to have high latency. This allows integer
computation to overlap floating point computations.
Also process blocks in a reverse post-order and propagate high-latency
flags to successors.
<rdar://problem/13468102>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178773 91177308-0d34-0410-b5e6-96231b3b80d8
This is helps on architectures where i8,i16 are not legal but we have byte, and
short loads/stores. Allowing us to merge copies like the one below on ARM.
copy(char *a, char *b, int n) {
do {
int t0 = a[0];
int t1 = a[1];
b[0] = t0;
b[1] = t1;
radar://13536387
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178546 91177308-0d34-0410-b5e6-96231b3b80d8
This reverts commit 342d92c7a0.
Turns out we're going with a different schema design to represent
DW_TAG_imported_modules so we won't need this extra field.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178215 91177308-0d34-0410-b5e6-96231b3b80d8
This is just the basic groundwork for supporting DW_TAG_imported_module but I
wanted to commit this before pushing support further into Clang or LLVM so that
this rather churny change is isolated from the rest of the work. The major
churn here is obviously adding another field (within the common DIScope prefix)
to all DIScopes (files, classes, namespaces, lexical scopes, etc). This should
be the last big churny change needed for DW_TAG_imported_module/using directive
support/PR14606.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@178099 91177308-0d34-0410-b5e6-96231b3b80d8
NEON is not IEEE 754 compliant, so we should avoid lowering single-precision
floating point operations with NEON unless unsafe-math is turned on. The
equivalent VFP instructions are IEEE 754 compliant, but in some cores they're
much slower, so some archs/OSs might still request it to be on by default,
such as Swift and Darwin.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@177651 91177308-0d34-0410-b5e6-96231b3b80d8