mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-10-03 08:55:51 +00:00
198d8baafb
One of several parallel first steps to remove the target type of pointers,
replacing them with a single opaque pointer type.
This adds an explicit type parameter to the gep instruction so that when the
first parameter becomes an opaque pointer type, the type to gep through is
still available to the instructions.
* This doesn't modify gep operators, only instructions (operators will be
handled separately)
* Textual IR changes only. Bitcode (including upgrade) and changing the
in-memory representation will be in separate changes.
* geps of vectors are transformed as:
getelementptr <4 x float*> %x, ...
->getelementptr float, <4 x float*> %x, ...
Then, once the opaque pointer type is introduced, this will ultimately look
like:
getelementptr float, <4 x ptr> %x
with the unambiguous interpretation that it is a vector of pointers to float.
* address spaces remain on the pointer, not the type:
getelementptr float addrspace(1)* %x
->getelementptr float, float addrspace(1)* %x
Then, eventually:
getelementptr float, ptr addrspace(1) %x
Importantly, the massive amount of test case churn has been automated by
same crappy python code. I had to manually update a few test cases that
wouldn't fit the script's model (r228970,r229196,r229197,r229198). The
python script just massages stdin and writes the result to stdout, I
then wrapped that in a shell script to handle replacing files, then
using the usual find+xargs to migrate all the files.
update.py:
import fileinput
import sys
import re
ibrep = re.compile(r"(^.*?[^%\w]getelementptr inbounds )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")
normrep = re.compile( r"(^.*?[^%\w]getelementptr )(((?:<\d* x )?)(.*?)(| addrspace\(\d\)) *\*(|>)(?:$| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$))")
def conv(match, line):
if not match:
return line
line = match.groups()[0]
if len(match.groups()[5]) == 0:
line += match.groups()[2]
line += match.groups()[3]
line += ", "
line += match.groups()[1]
line += "\n"
return line
for line in sys.stdin:
if line.find("getelementptr ") == line.find("getelementptr inbounds"):
if line.find("getelementptr inbounds") != line.find("getelementptr inbounds ("):
line = conv(re.match(ibrep, line), line)
elif line.find("getelementptr ") != line.find("getelementptr ("):
line = conv(re.match(normrep, line), line)
sys.stdout.write(line)
apply.sh:
for name in "$@"
do
python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name"
rm -f "$name.tmp"
done
The actual commands:
From llvm/src:
find test/ -name *.ll | xargs ./apply.sh
From llvm/src/tools/clang:
find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}"
From llvm/src/tools/polly:
find test/ -name *.ll | xargs ./apply.sh
After that, check-all (with llvm, clang, clang-tools-extra, lld,
compiler-rt, and polly all checked out).
The extra 'rm' in the apply.sh script is due to a few files in clang's test
suite using interesting unicode stuff that my python script was throwing
exceptions on. None of those files needed to be migrated, so it seemed
sufficient to ignore those cases.
Reviewers: rafael, dexonsmith, grosser
Differential Revision: http://reviews.llvm.org/D7636
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@230786 91177308-0d34-0410-b5e6-96231b3b80d8
107 lines
3.7 KiB
LLVM
107 lines
3.7 KiB
LLVM
; RUN: opt < %s -sroa -S | FileCheck %s
|
|
target datalayout = "e-p:64:64:64-p1:16:16:16-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:64:64-f80:128-v64:64:64-v128:128:128-a0:0:64-n8:16:32:64"
|
|
|
|
declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind
|
|
declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1) nounwind
|
|
|
|
define void @no_split_on_non_byte_width(i32) {
|
|
; This tests that allocas are not split into slices that are not byte width multiple
|
|
%arg = alloca i32 , align 8
|
|
store i32 %0, i32* %arg
|
|
br label %load_i32
|
|
|
|
load_i32:
|
|
; CHECK-LABEL: load_i32:
|
|
; CHECK-NOT: bitcast {{.*}} to i1
|
|
; CHECK-NOT: zext i1
|
|
%r0 = load i32* %arg
|
|
br label %load_i1
|
|
|
|
load_i1:
|
|
; CHECK-LABEL: load_i1:
|
|
; CHECK: bitcast {{.*}} to i1
|
|
%p1 = bitcast i32* %arg to i1*
|
|
%t1 = load i1* %p1
|
|
ret void
|
|
}
|
|
|
|
; PR18726: Check that we use memcpy and memset to fill out padding when we have
|
|
; a slice with a simple single type whose store size is smaller than the slice
|
|
; size.
|
|
|
|
%union.Foo = type { x86_fp80, i64, i64 }
|
|
|
|
@foo_copy_source = external constant %union.Foo
|
|
@i64_sink = global i64 0
|
|
|
|
define void @memcpy_fp80_padding() {
|
|
%x = alloca %union.Foo
|
|
|
|
; Copy from a global.
|
|
%x_i8 = bitcast %union.Foo* %x to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* %x_i8, i8* bitcast (%union.Foo* @foo_copy_source to i8*), i32 32, i32 16, i1 false)
|
|
|
|
; Access a slice of the alloca to trigger SROA.
|
|
%mid_p = getelementptr %union.Foo, %union.Foo* %x, i32 0, i32 1
|
|
%elt = load i64* %mid_p
|
|
store i64 %elt, i64* @i64_sink
|
|
ret void
|
|
}
|
|
; CHECK-LABEL: define void @memcpy_fp80_padding
|
|
; CHECK: alloca x86_fp80
|
|
; CHECK: call void @llvm.memcpy.p0i8.p0i8.i32
|
|
; CHECK: load i64* getelementptr inbounds (%union.Foo* @foo_copy_source, i64 0, i32 1)
|
|
; CHECK: load i64* getelementptr inbounds (%union.Foo* @foo_copy_source, i64 0, i32 2)
|
|
|
|
define void @memset_fp80_padding() {
|
|
%x = alloca %union.Foo
|
|
|
|
; Set to all ones.
|
|
%x_i8 = bitcast %union.Foo* %x to i8*
|
|
call void @llvm.memset.p0i8.i32(i8* %x_i8, i8 -1, i32 32, i32 16, i1 false)
|
|
|
|
; Access a slice of the alloca to trigger SROA.
|
|
%mid_p = getelementptr %union.Foo, %union.Foo* %x, i32 0, i32 1
|
|
%elt = load i64* %mid_p
|
|
store i64 %elt, i64* @i64_sink
|
|
ret void
|
|
}
|
|
; CHECK-LABEL: define void @memset_fp80_padding
|
|
; CHECK: alloca x86_fp80
|
|
; CHECK: call void @llvm.memset.p0i8.i32(i8* %{{.*}}, i8 -1, i32 16, i32 16, i1 false)
|
|
; CHECK: store i64 -1, i64* @i64_sink
|
|
|
|
%S.vec3float = type { float, float, float }
|
|
%U.vec3float = type { <4 x float> }
|
|
|
|
declare i32 @memcpy_vec3float_helper(%S.vec3float*)
|
|
|
|
define i32 @memcpy_vec3float_widening(%S.vec3float* %x) {
|
|
; CHECK-LABEL: @memcpy_vec3float_widening(
|
|
; PR18726: Check that SROA does not rewrite a 12-byte memcpy into a 16-byte
|
|
; vector store, hence accidentally putting gibberish onto the stack.
|
|
entry:
|
|
; Create a temporary variable %tmp1 and copy %x[0] into it
|
|
%tmp1 = alloca %S.vec3float, align 4
|
|
%0 = bitcast %S.vec3float* %tmp1 to i8*
|
|
%1 = bitcast %S.vec3float* %x to i8*
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* %0, i8* %1, i32 12, i32 4, i1 false)
|
|
|
|
; The following block does nothing; but appears to confuse SROA
|
|
%unused1 = bitcast %S.vec3float* %tmp1 to %U.vec3float*
|
|
%unused2 = getelementptr inbounds %U.vec3float, %U.vec3float* %unused1, i32 0, i32 0
|
|
%unused3 = load <4 x float>* %unused2, align 1
|
|
|
|
; Create a second temporary and copy %tmp1 into it
|
|
%tmp2 = alloca %S.vec3float, align 4
|
|
%2 = bitcast %S.vec3float* %tmp2 to i8*
|
|
%3 = bitcast %S.vec3float* %tmp1 to i8*
|
|
; CHECK: alloca
|
|
; CHECK-NOT: store <4 x float>
|
|
call void @llvm.memcpy.p0i8.p0i8.i32(i8* %2, i8* %3, i32 12, i32 4, i1 false)
|
|
|
|
%result = call i32 @memcpy_vec3float_helper(%S.vec3float* %tmp2)
|
|
ret i32 %result
|
|
; CHECK: ret i32 %result
|
|
}
|