maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/Transforms/ConstProp/loads.ll

272 lines
6.0 KiB
LLVM
Raw Normal View History

opt: Rename -default-data-layout flag to -data-layout and make it always override the layout. There isn't much point in a flag that only works if the data layout is empty. Differential Revision: https://reviews.llvm.org/D30014 llvm-svn: 295468
2017-02-18 01:36:52 +08:00
; RUN: opt < %s -data-layout="e-p:64:64:64-p1:16:16:16-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64" -instcombine -S | FileCheck %s --check-prefix=LE
; RUN: opt < %s -data-layout="E-p:64:64:64-p1:16:16:16-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64" -instcombine -S | FileCheck %s --check-prefix=BE
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; {{ 0xDEADBEEF, 0xBA }, 0xCAFEBABE}
fix PR5287, a serious regression from my previous patches. Thanks to Duncan for the nice tiny testcase. llvm-svn: 84992
2009-10-24 13:22:15 +08:00
@g1 = constant {{i32,i8},i32} {{i32,i8} { i32 -559038737, i8 186 }, i32 -889275714 }
@g2 = constant double 1.0
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; { 0x7B, 0x06B1BFF8 }
fix PR5287, a serious regression from my previous patches. Thanks to Duncan for the nice tiny testcase. llvm-svn: 84992
2009-10-24 13:22:15 +08:00
@g3 = constant {i64, i64} { i64 123, i64 112312312 }
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
; Simple load
define i32 @test1() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%r = load i32, i32* getelementptr ({{i32,i8},i32}, {{i32,i8},i32}* @g1, i32 0, i32 0, i32 0)
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
ret i32 %r
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0xDEADBEEF
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test1(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i32 -559038737
; 0xDEADBEEF
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test1(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i32 -559038737
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
}
; PR3152
; Load of first 16 bits of 32-bit value.
define i16 @test2() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%r = load i16, i16* bitcast(i32* getelementptr ({{i32,i8},i32}, {{i32,i8},i32}* @g1, i32 0, i32 0, i32 0) to i16*)
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
ret i16 %r
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0xBEEF
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test2(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i16 -16657
; 0xDEAD
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test2(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i16 -8531
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
}
Look through addrspacecast in IsConstantOffsetFromGlobal llvm-svn: 213000
2014-07-15 06:39:26 +08:00
define i16 @test2_addrspacecast() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%r = load i16, i16 addrspace(1)* addrspacecast(i32* getelementptr ({{i32,i8},i32}, {{i32,i8},i32}* @g1, i32 0, i32 0, i32 0) to i16 addrspace(1)*)
Look through addrspacecast in IsConstantOffsetFromGlobal llvm-svn: 213000
2014-07-15 06:39:26 +08:00
ret i16 %r
Fix assert when inlining a constantexpr addrspacecast The pointer size of the addrspacecasted pointer might not have matched, so this would have hit an assert in accumulateConstantOffset. I think this was here to allow constant folding of a load of an addrspacecasted constant. Accumulating the offset through the addrspacecast doesn't make much sense, so something else is necessary to allow folding the load through this cast. llvm-svn: 243300
2015-07-28 02:31:03 +08:00
; FIXME: Should be able to load through a constant addrspacecast.
Look through addrspacecast in IsConstantOffsetFromGlobal llvm-svn: 213000
2014-07-15 06:39:26 +08:00
; 0xBEEF
; LE-LABEL: @test2_addrspacecast(
Fix assert when inlining a constantexpr addrspacecast The pointer size of the addrspacecasted pointer might not have matched, so this would have hit an assert in accumulateConstantOffset. I think this was here to allow constant folding of a load of an addrspacecasted constant. Accumulating the offset through the addrspacecast doesn't make much sense, so something else is necessary to allow folding the load through this cast. llvm-svn: 243300
2015-07-28 02:31:03 +08:00
; XLE: ret i16 -16657
; LE: load i16, i16 addrspace(1)* addrspacecast
Look through addrspacecast in IsConstantOffsetFromGlobal llvm-svn: 213000
2014-07-15 06:39:26 +08:00
; 0xDEAD
; BE-LABEL: @test2_addrspacecast(
Fix assert when inlining a constantexpr addrspacecast The pointer size of the addrspacecasted pointer might not have matched, so this would have hit an assert in accumulateConstantOffset. I think this was here to allow constant folding of a load of an addrspacecasted constant. Accumulating the offset through the addrspacecast doesn't make much sense, so something else is necessary to allow folding the load through this cast. llvm-svn: 243300
2015-07-28 02:31:03 +08:00
; XBE: ret i16 -8531
; BE: load i16, i16 addrspace(1)* addrspacecast
Look through addrspacecast in IsConstantOffsetFromGlobal llvm-svn: 213000
2014-07-15 06:39:26 +08:00
}
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
; Load of second 16 bits of 32-bit value.
define i16 @test3() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%r = load i16, i16* getelementptr(i16, i16* bitcast(i32* getelementptr ({{i32,i8},i32}, {{i32,i8},i32}* @g1, i32 0, i32 0, i32 0) to i16*), i32 1)
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
ret i16 %r
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0xDEAD
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test3(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i16 -8531
; 0xBEEF
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test3(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i16 -16657
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
}
; Load of 8 bit field + tail padding.
define i16 @test4() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%r = load i16, i16* getelementptr(i16, i16* bitcast(i32* getelementptr ({{i32,i8},i32}, {{i32,i8},i32}* @g1, i32 0, i32 0, i32 0) to i16*), i32 2)
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
ret i16 %r
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0x00BA
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test4(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i16 186
; 0xBA00
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test4(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i16 -17920
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
}
; Load of double bits.
define i64 @test6() {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%r = load i64, i64* bitcast(double* @g2 to i64*)
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
ret i64 %r
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0x3FF_0000000000000
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test6(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i64 4607182418800017408
; 0x3FF_0000000000000
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test6(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i64 4607182418800017408
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
}
; Load of double bits.
define i16 @test7() {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%r = load i16, i16* bitcast(double* @g2 to i16*)
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
ret i16 %r
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0x0000
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test7(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i16 0
; 0x3FF0
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test7(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i16 16368
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
}
; Double load.
define double @test8() {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%r = load double, double* bitcast({{i32,i8},i32}* @g1 to double*)
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
ret double %r
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test8(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret double 0xBADEADBEEF
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test8(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret double 0xDEADBEEFBA000000
teach libanalysis to fold int and fp loads from almost arbitrary non-type-safe constant initializers. This sort of thing happens quite a bit for 4-byte loads out of string constants, unions, bitfields, and an interesting endianness check from sqlite, which is something like this: const int sqlite3one = 1; # define SQLITE_BIGENDIAN (*(char *)(&sqlite3one)==0) # define SQLITE_LITTLEENDIAN (*(char *)(&sqlite3one)==1) # define SQLITE_UTF16NATIVE (SQLITE_BIGENDIAN?SQLITE_UTF16BE:SQLITE_UTF16LE) all of these macros now constant fold away. This implements PR3152 and is based on a patch started by Eli, but heavily modified and extended. llvm-svn: 84936
2009-10-23 14:23:49 +08:00
}
enhance FoldReinterpretLoadFromConstPtr to handle loads of up to 32 bytes (i256). llvm-svn: 84941
2009-10-23 14:50:36 +08:00
; i128 load.
define i128 @test9() {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%r = load i128, i128* bitcast({i64, i64}* @g3 to i128*)
enhance FoldReinterpretLoadFromConstPtr to handle loads of up to 32 bytes (i256). llvm-svn: 84941
2009-10-23 14:50:36 +08:00
ret i128 %r
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0x00000000_06B1BFF8_00000000_0000007B
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test9(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i128 2071796475790618158476296315
; 0x00000000_0000007B_00000000_06B1BFF8
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test9(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i128 2268949521066387161080
enhance FoldReinterpretLoadFromConstPtr to handle loads of up to 32 bytes (i256). llvm-svn: 84941
2009-10-23 14:50:36 +08:00
}
teach libanalysis to simplify vector loads with bitcast sources. This implements something out of Target/README.txt producing: _foo: ## @foo movl 4(%esp), %eax movapd LCPI1_0, %xmm0 movapd %xmm0, (%eax) ret $4 instead of: _foo: ## @foo movl 4(%esp), %eax movapd _b, %xmm0 mulpd LCPI1_0, %xmm0 addpd _a, %xmm0 movapd %xmm0, (%eax) ret $4 llvm-svn: 84942
2009-10-23 14:57:37 +08:00
; vector load.
define <2 x i64> @test10() {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%r = load <2 x i64>, <2 x i64>* bitcast({i64, i64}* @g3 to <2 x i64>*)
teach libanalysis to simplify vector loads with bitcast sources. This implements something out of Target/README.txt producing: _foo: ## @foo movl 4(%esp), %eax movapd LCPI1_0, %xmm0 movapd %xmm0, (%eax) ret $4 instead of: _foo: ## @foo movl 4(%esp), %eax movapd _b, %xmm0 mulpd LCPI1_0, %xmm0 addpd _a, %xmm0 movapd %xmm0, (%eax) ret $4 llvm-svn: 84942
2009-10-23 14:57:37 +08:00
ret <2 x i64> %r
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test10(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret <2 x i64> <i64 123, i64 112312312>
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test10(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret <2 x i64> <i64 123, i64 112312312>
teach libanalysis to simplify vector loads with bitcast sources. This implements something out of Target/README.txt producing: _foo: ## @foo movl 4(%esp), %eax movapd LCPI1_0, %xmm0 movapd %xmm0, (%eax) ret $4 instead of: _foo: ## @foo movl 4(%esp), %eax movapd _b, %xmm0 mulpd LCPI1_0, %xmm0 addpd _a, %xmm0 movapd %xmm0, (%eax) ret $4 llvm-svn: 84942
2009-10-23 14:57:37 +08:00
}
fix PR5287, a serious regression from my previous patches. Thanks to Duncan for the nice tiny testcase. llvm-svn: 84992
2009-10-24 13:22:15 +08:00
; PR5287
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; { 0xA1, 0x08 }
fix PR5287, a serious regression from my previous patches. Thanks to Duncan for the nice tiny testcase. llvm-svn: 84992
2009-10-24 13:22:15 +08:00
@g4 = internal constant { i8, i8 } { i8 -95, i8 8 }
define i16 @test11() nounwind {
entry:
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%a = load i16, i16* bitcast ({ i8, i8 }* @g4 to i16*)
fix PR5287, a serious regression from my previous patches. Thanks to Duncan for the nice tiny testcase. llvm-svn: 84992
2009-10-24 13:22:15 +08:00
ret i16 %a
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0x08A1
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test11(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i16 2209
; 0xA108
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test11(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i16 -24312
fix PR5287, a serious regression from my previous patches. Thanks to Duncan for the nice tiny testcase. llvm-svn: 84992
2009-10-24 13:22:15 +08:00
}
Fix PR5551 by not ignoring the top level constantexpr when folding a load from constant. llvm-svn: 90545
2009-12-04 14:29:29 +08:00
; PR5551
@test12g = private constant [6 x i8] c"a\00b\00\00\00"
define i16 @test12() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%a = load i16, i16* getelementptr inbounds ([3 x i16], [3 x i16]* bitcast ([6 x i8]* @test12g to [3 x i16]*), i32 0, i64 1)
Fix PR5551 by not ignoring the top level constantexpr when folding a load from constant. llvm-svn: 90545
2009-12-04 14:29:29 +08:00
ret i16 %a
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; 0x0062
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test12(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i16 98
; 0x6200
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test12(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i16 25088
Fix PR5551 by not ignoring the top level constantexpr when folding a load from constant. llvm-svn: 90545
2009-12-04 14:29:29 +08:00
}
fix PR5978 by peeling the loop so that we avoid shifting the result int by 8 for the first byte. While normally harmless, if the result is smaller than a byte, this shift is invalid. llvm-svn: 93018
2010-01-09 03:02:23 +08:00
; PR5978
@g5 = constant i8 4
define i1 @test13() {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%A = load i1, i1* bitcast (i8* @g5 to i1*)
fix PR5978 by peeling the loop so that we avoid shifting the result int by 8 for the first byte. While normally harmless, if the result is smaller than a byte, this shift is invalid. llvm-svn: 93018
2010-01-09 03:02:23 +08:00
ret i1 %A
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test13(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i1 false
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test13(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i1 false
fix PR5978 by peeling the loop so that we avoid shifting the result int by 8 for the first byte. While normally harmless, if the result is smaller than a byte, this shift is invalid. llvm-svn: 93018
2010-01-09 03:02:23 +08:00
}
When loading from a constant, fold inttoptr if the integer type and the resulting pointer type both have the same size. llvm-svn: 124987
2011-02-07 04:11:56 +08:00
@g6 = constant [2 x i8*] [i8* inttoptr (i64 1 to i8*), i8* inttoptr (i64 2 to i8*)]
define i64 @test14() nounwind {
entry:
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp = load i64, i64* bitcast ([2 x i8*]* @g6 to i64*)
When loading from a constant, fold inttoptr if the integer type and the resulting pointer type both have the same size. llvm-svn: 124987
2011-02-07 04:11:56 +08:00
ret i64 %tmp
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test14(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i64 1
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test14(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i64 1
When loading from a constant, fold inttoptr if the integer type and the resulting pointer type both have the same size. llvm-svn: 124987
2011-02-07 04:11:56 +08:00
}
Teach ConstantFolding about pointer address spaces llvm-svn: 188831
2013-08-21 05:20:04 +08:00
; Check with address space pointers
@g6_as1 = constant [2 x i8 addrspace(1)*] [i8 addrspace(1)* inttoptr (i16 1 to i8 addrspace(1)*), i8 addrspace(1)* inttoptr (i16 2 to i8 addrspace(1)*)]
define i16 @test14_as1() nounwind {
entry:
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%tmp = load i16, i16* bitcast ([2 x i8 addrspace(1)*]* @g6_as1 to i16*)
Teach ConstantFolding about pointer address spaces llvm-svn: 188831
2013-08-21 05:20:04 +08:00
ret i16 %tmp
; LE: @test14_as1
; LE: ret i16 1
; BE: @test14_as1
; BE: ret i16 1
}
When loading from a constant, fold inttoptr if the integer type and the resulting pointer type both have the same size. llvm-svn: 124987
2011-02-07 04:11:56 +08:00
define i64 @test15() nounwind {
entry:
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%tmp = load i64, i64* bitcast (i8** getelementptr inbounds ([2 x i8*], [2 x i8*]* @g6, i32 0, i64 1) to i64*)
When loading from a constant, fold inttoptr if the integer type and the resulting pointer type both have the same size. llvm-svn: 124987
2011-02-07 04:11:56 +08:00
ret i64 %tmp
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; LE-LABEL: @test15(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; LE: ret i64 2
Update Transforms tests to use CHECK-LABEL for easier debugging. No functionality change. This update was done with the following bash script: find test/Transforms -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc" $NAME; then TEMP=`mktemp -t temp` cp $NAME $TEMP sed -n "s/^define [^@]*@\([A-Za-z0-9_]*\)(.*$/\1/p" < $NAME | \ while read FUNC; do sed -i '' "s/;\(.*\)\([A-Za-z0-9_]*\):\( *\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186268
2013-07-14 09:42:54 +08:00
; BE-LABEL: @test15(
llvm/ConstantFolding.cpp: Make ReadDataFromGlobal() and FoldReinterpretLoadFromConstPtr() Big-endian-aware. llvm-svn: 167595
2012-11-09 04:34:25 +08:00
; BE: ret i64 2
When loading from a constant, fold inttoptr if the integer type and the resulting pointer type both have the same size. llvm-svn: 124987
2011-02-07 04:11:56 +08:00
}
Teach the constant folder to look through bitcast constant expressions much more effectively when trying to constant fold a load of a constant. Previously, we only handled bitcasts by trying to find a totally generic byte representation of the constant and use that. Now, we look through the bitcast to see what constant we might fold the load into, and then try to form a constant expression cast of the found value that would be equivalent to loading the value. You might wonder why on earth this actually matters. Well, turns out that the Itanium ABI causes us to create a single array for a vtable where the first elements are virtual base offsets, followed by the virtual function pointers. Because the array is homogenous the element type is consistently i8* and we inttoptr the virtual base offsets into the initial elements. Then constructors bitcast these pointers to i64 pointers prior to loading them. Boom, no more constant folding of virtual base offsets. This is the first fix to LLVM to address the *insane* performance Eric Niebler discovered with Clang on his range comprehensions[1]. There is more to come though, this doesn't *really* fix the problem fully. [1]: http://ericniebler.com/2014/04/27/range-comprehensions/ llvm-svn: 208856
2014-05-15 17:56:28 +08:00
@gv7 = constant [4 x i8*] [i8* null, i8* inttoptr (i64 -14 to i8*), i8* null, i8* null]
define i64 @test16.1() {
[opaque pointer type] Add textual IR support for explicit type parameter to load instruction Essentially the same as the GEP change in r230786. A similar migration script can be used to update test cases, though a few more test case improvements/changes were required this time around: (r229269-r229278) import fileinput import sys import re pat = re.compile(r"((?:=|:|^)\s*load (?:atomic )?(?:volatile )?(.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|\[\[[a-zA-Z]|\{\{).*$)") for line in sys.stdin: sys.stdout.write(re.sub(pat, r"\1, \2\3*\4", line)) Reviewers: rafael, dexonsmith, grosser Differential Revision: http://reviews.llvm.org/D7649 llvm-svn: 230794
2015-02-28 05:17:42 +08:00
%v = load i64, i64* bitcast ([4 x i8*]* @gv7 to i64*), align 8
Teach the constant folder to look through bitcast constant expressions much more effectively when trying to constant fold a load of a constant. Previously, we only handled bitcasts by trying to find a totally generic byte representation of the constant and use that. Now, we look through the bitcast to see what constant we might fold the load into, and then try to form a constant expression cast of the found value that would be equivalent to loading the value. You might wonder why on earth this actually matters. Well, turns out that the Itanium ABI causes us to create a single array for a vtable where the first elements are virtual base offsets, followed by the virtual function pointers. Because the array is homogenous the element type is consistently i8* and we inttoptr the virtual base offsets into the initial elements. Then constructors bitcast these pointers to i64 pointers prior to loading them. Boom, no more constant folding of virtual base offsets. This is the first fix to LLVM to address the *insane* performance Eric Niebler discovered with Clang on his range comprehensions[1]. There is more to come though, this doesn't *really* fix the problem fully. [1]: http://ericniebler.com/2014/04/27/range-comprehensions/ llvm-svn: 208856
2014-05-15 17:56:28 +08:00
ret i64 %v
; LE-LABEL: @test16.1(
; LE: ret i64 0
; BE-LABEL: @test16.1(
; BE: ret i64 0
}
define i64 @test16.2() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%v = load i64, i64* bitcast (i8** getelementptr inbounds ([4 x i8*], [4 x i8*]* @gv7, i64 0, i64 1) to i64*), align 8
Teach the constant folder to look through bitcast constant expressions much more effectively when trying to constant fold a load of a constant. Previously, we only handled bitcasts by trying to find a totally generic byte representation of the constant and use that. Now, we look through the bitcast to see what constant we might fold the load into, and then try to form a constant expression cast of the found value that would be equivalent to loading the value. You might wonder why on earth this actually matters. Well, turns out that the Itanium ABI causes us to create a single array for a vtable where the first elements are virtual base offsets, followed by the virtual function pointers. Because the array is homogenous the element type is consistently i8* and we inttoptr the virtual base offsets into the initial elements. Then constructors bitcast these pointers to i64 pointers prior to loading them. Boom, no more constant folding of virtual base offsets. This is the first fix to LLVM to address the *insane* performance Eric Niebler discovered with Clang on his range comprehensions[1]. There is more to come though, this doesn't *really* fix the problem fully. [1]: http://ericniebler.com/2014/04/27/range-comprehensions/ llvm-svn: 208856
2014-05-15 17:56:28 +08:00
ret i64 %v
; LE-LABEL: @test16.2(
; LE: ret i64 -14
; BE-LABEL: @test16.2(
; BE: ret i64 -14
}
define i64 @test16.3() {
[opaque pointer type] Add textual IR support for explicit type parameter to gep operator Similar to gep (r230786) and load (r230794) changes. Similar migration script can be used to update test cases, which successfully migrated all of LLVM and Polly, but about 4 test cases needed manually changes in Clang. (this script will read the contents of stdin and massage it into stdout - wrap it in the 'apply.sh' script shown in previous commits + xargs to apply it over a large set of test cases) import fileinput import sys import re rep = re.compile(r"(getelementptr(?:\s+inbounds)?\s*\()((<\d*\s+x\s+)?([^@]*?)(|\s*addrspace\(\d+\))\s*\*(?(3)>)\s*)(?=$|%|@|null|undef|blockaddress|getelementptr|addrspacecast|bitcast|inttoptr|zeroinitializer|<|\[\[[a-zA-Z]|\{\{)", re.MULTILINE | re.DOTALL) def conv(match): line = match.group(1) line += match.group(4) line += ", " line += match.group(2) return line line = sys.stdin.read() off = 0 for match in re.finditer(rep, line): sys.stdout.write(line[off:match.start()]) sys.stdout.write(conv(match)) off = match.end() sys.stdout.write(line[off:]) llvm-svn: 232184
2015-03-14 02:20:45 +08:00
%v = load i64, i64* bitcast (i8** getelementptr inbounds ([4 x i8*], [4 x i8*]* @gv7, i64 0, i64 2) to i64*), align 8
Teach the constant folder to look through bitcast constant expressions much more effectively when trying to constant fold a load of a constant. Previously, we only handled bitcasts by trying to find a totally generic byte representation of the constant and use that. Now, we look through the bitcast to see what constant we might fold the load into, and then try to form a constant expression cast of the found value that would be equivalent to loading the value. You might wonder why on earth this actually matters. Well, turns out that the Itanium ABI causes us to create a single array for a vtable where the first elements are virtual base offsets, followed by the virtual function pointers. Because the array is homogenous the element type is consistently i8* and we inttoptr the virtual base offsets into the initial elements. Then constructors bitcast these pointers to i64 pointers prior to loading them. Boom, no more constant folding of virtual base offsets. This is the first fix to LLVM to address the *insane* performance Eric Niebler discovered with Clang on his range comprehensions[1]. There is more to come though, this doesn't *really* fix the problem fully. [1]: http://ericniebler.com/2014/04/27/range-comprehensions/ llvm-svn: 208856
2014-05-15 17:56:28 +08:00
ret i64 %v
; LE-LABEL: @test16.3(
; LE: ret i64 0
; BE-LABEL: @test16.3(
; BE: ret i64 0
}