mirror of
https://github.com/RPCS3/llvm-mirror.git
synced 2024-10-26 14:33:02 +02:00
0d99339102
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
llvm-svn: 230786
109 lines
4.8 KiB
LLVM
109 lines
4.8 KiB
LLVM
; RUN: llc < %s -mtriple=arm-apple-ios -mattr=+v6,+vfp2 | FileCheck %s
|
|
|
|
@quant_coef = external global [6 x [4 x [4 x i32]]] ; <[6 x [4 x [4 x i32]]]*> [#uses=1]
|
|
@dequant_coef = external global [6 x [4 x [4 x i32]]] ; <[6 x [4 x [4 x i32]]]*> [#uses=1]
|
|
@A = external global [4 x [4 x i32]] ; <[4 x [4 x i32]]*> [#uses=1]
|
|
|
|
; CHECK-LABEL: dct_luma_sp:
|
|
define fastcc i32 @dct_luma_sp(i32 %block_x, i32 %block_y, i32* %coeff_cost) {
|
|
entry:
|
|
; Make sure to use base-updating stores for saving callee-saved registers.
|
|
; CHECK: push
|
|
; CHECK-NOT: sub sp
|
|
; CHECK: push
|
|
%predicted_block = alloca [4 x [4 x i32]], align 4 ; <[4 x [4 x i32]]*> [#uses=1]
|
|
br label %cond_next489
|
|
|
|
cond_next489: ; preds = %cond_false, %bb471
|
|
%j.7.in = load i8* null ; <i8> [#uses=1]
|
|
%i.8.in = load i8* null ; <i8> [#uses=1]
|
|
%i.8 = zext i8 %i.8.in to i32 ; <i32> [#uses=4]
|
|
%j.7 = zext i8 %j.7.in to i32 ; <i32> [#uses=4]
|
|
%tmp495 = getelementptr [4 x [4 x i32]], [4 x [4 x i32]]* %predicted_block, i32 0, i32 %i.8, i32 %j.7 ; <i32*> [#uses=2]
|
|
%tmp496 = load i32* %tmp495 ; <i32> [#uses=2]
|
|
%tmp502 = load i32* null ; <i32> [#uses=1]
|
|
%tmp542 = getelementptr [6 x [4 x [4 x i32]]], [6 x [4 x [4 x i32]]]* @quant_coef, i32 0, i32 0, i32 %i.8, i32 %j.7 ; <i32*> [#uses=1]
|
|
%tmp543 = load i32* %tmp542 ; <i32> [#uses=1]
|
|
%tmp548 = ashr i32 0, 0 ; <i32> [#uses=3]
|
|
%tmp561 = sub i32 0, %tmp496 ; <i32> [#uses=3]
|
|
%abscond563 = icmp sgt i32 %tmp561, -1 ; <i1> [#uses=1]
|
|
%abs564 = select i1 %abscond563, i32 %tmp561, i32 0 ; <i32> [#uses=1]
|
|
%tmp572 = mul i32 %abs564, %tmp543 ; <i32> [#uses=1]
|
|
%tmp574 = add i32 %tmp572, 0 ; <i32> [#uses=1]
|
|
%tmp576 = ashr i32 %tmp574, 0 ; <i32> [#uses=7]
|
|
%tmp579 = icmp eq i32 %tmp548, %tmp576 ; <i1> [#uses=1]
|
|
br i1 %tmp579, label %bb712, label %cond_next589
|
|
|
|
cond_next589: ; preds = %cond_next489
|
|
%tmp605 = getelementptr [6 x [4 x [4 x i32]]], [6 x [4 x [4 x i32]]]* @dequant_coef, i32 0, i32 0, i32 %i.8, i32 %j.7 ; <i32*> [#uses=1]
|
|
%tmp606 = load i32* %tmp605 ; <i32> [#uses=1]
|
|
%tmp612 = load i32* null ; <i32> [#uses=1]
|
|
%tmp629 = load i32* null ; <i32> [#uses=1]
|
|
%tmp629a = sitofp i32 %tmp629 to double ; <double> [#uses=1]
|
|
%tmp631 = fmul double %tmp629a, 0.000000e+00 ; <double> [#uses=1]
|
|
%tmp632 = fadd double 0.000000e+00, %tmp631 ; <double> [#uses=1]
|
|
%tmp642 = call fastcc i32 @sign( i32 %tmp576, i32 %tmp561 ) ; <i32> [#uses=1]
|
|
%tmp650 = mul i32 %tmp606, %tmp642 ; <i32> [#uses=1]
|
|
%tmp656 = mul i32 %tmp650, %tmp612 ; <i32> [#uses=1]
|
|
%tmp658 = shl i32 %tmp656, 0 ; <i32> [#uses=1]
|
|
%tmp659 = ashr i32 %tmp658, 6 ; <i32> [#uses=1]
|
|
%tmp660 = sub i32 0, %tmp659 ; <i32> [#uses=1]
|
|
%tmp666 = sub i32 %tmp660, %tmp496 ; <i32> [#uses=1]
|
|
%tmp667 = sitofp i32 %tmp666 to double ; <double> [#uses=2]
|
|
call void @levrun_linfo_inter( i32 %tmp576, i32 0, i32* null, i32* null )
|
|
%tmp671 = fmul double %tmp667, %tmp667 ; <double> [#uses=1]
|
|
%tmp675 = fadd double %tmp671, 0.000000e+00 ; <double> [#uses=1]
|
|
%tmp678 = fcmp oeq double %tmp632, %tmp675 ; <i1> [#uses=1]
|
|
br i1 %tmp678, label %cond_true679, label %cond_false693
|
|
|
|
cond_true679: ; preds = %cond_next589
|
|
%abscond681 = icmp sgt i32 %tmp548, -1 ; <i1> [#uses=1]
|
|
%abs682 = select i1 %abscond681, i32 %tmp548, i32 0 ; <i32> [#uses=1]
|
|
%abscond684 = icmp sgt i32 %tmp576, -1 ; <i1> [#uses=1]
|
|
%abs685 = select i1 %abscond684, i32 %tmp576, i32 0 ; <i32> [#uses=1]
|
|
%tmp686 = icmp slt i32 %abs682, %abs685 ; <i1> [#uses=1]
|
|
br i1 %tmp686, label %cond_next702, label %cond_false689
|
|
|
|
cond_false689: ; preds = %cond_true679
|
|
%tmp739 = icmp eq i32 %tmp576, 0 ; <i1> [#uses=1]
|
|
br i1 %tmp579, label %bb737, label %cond_false708
|
|
|
|
cond_false693: ; preds = %cond_next589
|
|
ret i32 0
|
|
|
|
cond_next702: ; preds = %cond_true679
|
|
ret i32 0
|
|
|
|
cond_false708: ; preds = %cond_false689
|
|
ret i32 0
|
|
|
|
bb712: ; preds = %cond_next489
|
|
ret i32 0
|
|
|
|
bb737: ; preds = %cond_false689
|
|
br i1 %tmp739, label %cond_next791, label %cond_true740
|
|
|
|
cond_true740: ; preds = %bb737
|
|
%tmp761 = call fastcc i32 @sign( i32 %tmp576, i32 0 ) ; <i32> [#uses=1]
|
|
%tmp780 = load i32* null ; <i32> [#uses=1]
|
|
%tmp785 = getelementptr [4 x [4 x i32]], [4 x [4 x i32]]* @A, i32 0, i32 %i.8, i32 %j.7 ; <i32*> [#uses=1]
|
|
%tmp786 = load i32* %tmp785 ; <i32> [#uses=1]
|
|
%tmp781 = mul i32 %tmp780, %tmp761 ; <i32> [#uses=1]
|
|
%tmp787 = mul i32 %tmp781, %tmp786 ; <i32> [#uses=1]
|
|
%tmp789 = shl i32 %tmp787, 0 ; <i32> [#uses=1]
|
|
%tmp790 = ashr i32 %tmp789, 6 ; <i32> [#uses=1]
|
|
br label %cond_next791
|
|
|
|
cond_next791: ; preds = %cond_true740, %bb737
|
|
%ilev.1 = phi i32 [ %tmp790, %cond_true740 ], [ 0, %bb737 ] ; <i32> [#uses=1]
|
|
%tmp796 = load i32* %tmp495 ; <i32> [#uses=1]
|
|
%tmp798 = add i32 %tmp796, %ilev.1 ; <i32> [#uses=1]
|
|
%tmp812 = mul i32 0, %tmp502 ; <i32> [#uses=0]
|
|
%tmp818 = call fastcc i32 @sign( i32 0, i32 %tmp798 ) ; <i32> [#uses=0]
|
|
unreachable
|
|
}
|
|
|
|
declare i32 @sign(i32, i32)
|
|
|
|
declare void @levrun_linfo_inter(i32, i32, i32*, i32*)
|