llvm-project/llvm/test/CodeGen/ARM/2010-06-21-nondarwin-tc.ll

147 lines
11 KiB
LLVM
Raw Normal View History

; RUN: llc < %s -mtriple=armv4t-unknown-linux-gnueabi | FileCheck %s
; PR 7433
; XFAIL: *
%0 = type { i8*, i8* }
%1 = type { i8*, i8*, i8* }
%"class.llvm::Record" = type { i32, %"class.std::basic_string", %"class.llvm::SMLoc", %"class.std::vector", %"class.std::vector", %"class.std::vector" }
%"class.llvm::RecordVal" = type { %"class.std::basic_string", %"struct.llvm::Init"*, i32, %"struct.llvm::Init"* }
%"class.llvm::SMLoc" = type { i8* }
%"class.llvm::StringInit" = type { [8 x i8], %"class.std::basic_string" }
%"class.std::basic_string" = type { %"class.llvm::SMLoc" }
%"class.std::vector" = type { [12 x i8] }
%"struct.llvm::Init" = type { i32 (...)** }
@_ZTIN4llvm5RecTyE = external constant %0 ; <%0*> [#uses=1]
@_ZTIN4llvm4InitE = external constant %0 ; <%0*> [#uses=1]
@_ZTIN4llvm11RecordRecTyE = external constant %1 ; <%1*> [#uses=1]
@.str8 = external constant [47 x i8] ; <[47 x i8]*> [#uses=1]
@_ZTIN4llvm9UnsetInitE = external constant %1 ; <%1*> [#uses=1]
@.str51 = external constant [45 x i8] ; <[45 x i8]*> [#uses=1]
@__PRETTY_FUNCTION__._ZNK4llvm7VarInit12getFieldInitERNS_6RecordEPKNS_9RecordValERKSs = external constant [116 x i8] ; <[116 x i8]*> [#uses=1]
[opaque pointer type] Add textual IR support for explicit type parameter for global aliases update.py: import fileinput import sys import re alias_match_prefix = r"(.*(?:=|:|^)\s*(?:external |)(?:(?:private|internal|linkonce|linkonce_odr|weak|weak_odr|common|appending|extern_weak|available_externally) )?(?:default |hidden |protected )?(?:dllimport |dllexport )?(?:unnamed_addr |)(?:thread_local(?:\([a-z]*\))? )?alias" plain = re.compile(alias_match_prefix + r" (.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|addrspacecast|\[\[[a-zA-Z]|\{\{).*$)") cast = re.compile(alias_match_prefix + r") ((?:bitcast|inttoptr|addrspacecast)\s*\(.* to (.*?)(| addrspace\(\d+\) *)\*\)\s*(?:;.*)?$)") gep = re.compile(alias_match_prefix + r") ((?:getelementptr)\s*(?:inbounds)?\s*\((?P<type>.*), (?P=type)(?:\s*addrspace\(\d+\)\s*)?\* .*\)\s*(?:;.*)?$)") def conv(line): m = re.match(cast, line) if m: return m.group(1) + " " + m.group(3) + ", " + m.group(2) m = re.match(gep, line) if m: return m.group(1) + " " + m.group(3) + ", " + m.group(2) m = re.match(plain, line) if m: return m.group(1) + ", " + m.group(2) + m.group(3) + "*" + m.group(4) + "\n" return line for line in sys.stdin: sys.stdout.write(conv(line)) apply.sh: for name in "$@" 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 llvm-svn: 247378
2015-09-11 11:22:04 +08:00
@_ZN4llvm9RecordValC1ERKSsPNS_5RecTyEj = alias void (%"class.llvm::RecordVal"*, %"class.std::basic_string"*, %"struct.llvm::Init"*, i32), void (%"class.llvm::RecordVal"*, %"class.std::basic_string"*, %"struct.llvm::Init"*, i32)* @_ZN4llvm9RecordValC2ERKSsPNS_5RecTyEj ; <void (%"class.llvm::RecordVal"*, %"class.std::basic_string"*, %"struct.llvm::Init"*, i32)*> [#uses=0]
declare i8* @__dynamic_cast(i8*, i8*, i8*, i32)
declare void @__assert_fail(i8*, i8*, i32, i8*) noreturn
declare void @_ZN4llvm9RecordValC2ERKSsPNS_5RecTyEj(%"class.llvm::RecordVal"*, %"class.std::basic_string"*, %"struct.llvm::Init"*, i32) align 2
define %"struct.llvm::Init"* @_ZNK4llvm7VarInit12getFieldInitERNS_6RecordEPKNS_9RecordValERKSs(%"class.llvm::StringInit"* %this, %"class.llvm::Record"* %R, %"class.llvm::RecordVal"* %RV, %"class.std::basic_string"* %FieldName) align 2 {
;CHECK: ldmia sp!, {r4, r5, r6, r7, r8, lr}
;CHECK: bx r12 @ TAILCALL
entry:
%.loc = alloca i32 ; <i32*> [#uses=2]
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%tmp.i = getelementptr inbounds %"class.llvm::StringInit", %"class.llvm::StringInit"* %this, i32 0, i32 0, i32 4 ; <i8*> [#uses=1]
%0 = bitcast i8* %tmp.i to %"struct.llvm::Init"** ; <%"struct.llvm::Init"**> [#uses=1]
%tmp2.i = load %"struct.llvm::Init"*, %"struct.llvm::Init"** %0 ; <%"struct.llvm::Init"*> [#uses=2]
%1 = icmp eq %"struct.llvm::Init"* %tmp2.i, null ; <i1> [#uses=1]
br i1 %1, label %entry.return_crit_edge, label %tmpbb
entry.return_crit_edge: ; preds = %entry
br label %return
tmpbb: ; preds = %entry
%2 = bitcast %"struct.llvm::Init"* %tmp2.i to i8* ; <i8*> [#uses=1]
%3 = tail call i8* @__dynamic_cast(i8* %2, i8* bitcast (%0* @_ZTIN4llvm5RecTyE to i8*), i8* bitcast (%1* @_ZTIN4llvm11RecordRecTyE to i8*), i32 -1) ; <i8*> [#uses=1]
%phitmp = icmp eq i8* %3, null ; <i1> [#uses=1]
br i1 %phitmp, label %.return_crit_edge, label %if.then
.return_crit_edge: ; preds = %tmpbb
br label %return
if.then: ; preds = %tmpbb
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%tmp2.i.i.i.i = getelementptr inbounds %"class.llvm::StringInit", %"class.llvm::StringInit"* %this, i32 0, i32 1, i32 0, i32 0 ; <i8**> [#uses=1]
%tmp3.i.i.i.i = load i8*, i8** %tmp2.i.i.i.i ; <i8*> [#uses=2]
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%arrayidx.i.i.i.i = getelementptr inbounds i8, i8* %tmp3.i.i.i.i, i32 -12 ; <i8*> [#uses=1]
%tmp.i.i.i = bitcast i8* %arrayidx.i.i.i.i to i32* ; <i32*> [#uses=1]
%tmp2.i.i.i = load i32, i32* %tmp.i.i.i ; <i32> [#uses=1]
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%tmp.i5 = getelementptr inbounds %"class.llvm::Record", %"class.llvm::Record"* %R, i32 0, i32 4 ; <%"class.std::vector"*> [#uses=1]
%tmp2.i.i = getelementptr inbounds %"class.llvm::Record", %"class.llvm::Record"* %R, i32 0, i32 4, i32 0, i32 4 ; <i8*> [#uses=1]
%4 = bitcast i8* %tmp2.i.i to %"class.llvm::RecordVal"** ; <%"class.llvm::RecordVal"**> [#uses=1]
%tmp3.i.i6 = load %"class.llvm::RecordVal"*, %"class.llvm::RecordVal"** %4 ; <%"class.llvm::RecordVal"*> [#uses=1]
%tmp5.i.i = bitcast %"class.std::vector"* %tmp.i5 to %"class.llvm::RecordVal"** ; <%"class.llvm::RecordVal"**> [#uses=1]
%tmp6.i.i = load %"class.llvm::RecordVal"*, %"class.llvm::RecordVal"** %tmp5.i.i ; <%"class.llvm::RecordVal"*> [#uses=5]
%sub.ptr.lhs.cast.i.i = ptrtoint %"class.llvm::RecordVal"* %tmp3.i.i6 to i32 ; <i32> [#uses=1]
%sub.ptr.rhs.cast.i.i = ptrtoint %"class.llvm::RecordVal"* %tmp6.i.i to i32 ; <i32> [#uses=1]
%sub.ptr.sub.i.i = sub i32 %sub.ptr.lhs.cast.i.i, %sub.ptr.rhs.cast.i.i ; <i32> [#uses=1]
%sub.ptr.div.i.i = ashr i32 %sub.ptr.sub.i.i, 4 ; <i32> [#uses=1]
br label %codeRepl
codeRepl: ; preds = %if.then
%targetBlock = call i1 @_ZNK4llvm7VarInit12getFieldInitERNS_6RecordEPKNS_9RecordValERKSs_for.cond.i(i32 %sub.ptr.div.i.i, %"class.llvm::RecordVal"* %tmp6.i.i, i32 %tmp2.i.i.i, i8* %tmp3.i.i.i.i, i32* %.loc) ; <i1> [#uses=1]
%.reload = load i32, i32* %.loc ; <i32> [#uses=3]
br i1 %targetBlock, label %for.cond.i.return_crit_edge, label %_ZN4llvm6Record8getValueENS_9StringRefE.exit
for.cond.i.return_crit_edge: ; preds = %codeRepl
br label %return
_ZN4llvm6Record8getValueENS_9StringRefE.exit: ; preds = %codeRepl
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%add.ptr.i.i = getelementptr inbounds %"class.llvm::RecordVal", %"class.llvm::RecordVal"* %tmp6.i.i, i32 %.reload ; <%"class.llvm::RecordVal"*> [#uses=2]
%tobool5 = icmp eq %"class.llvm::RecordVal"* %add.ptr.i.i, null ; <i1> [#uses=1]
br i1 %tobool5, label %_ZN4llvm6Record8getValueENS_9StringRefE.exit.return_crit_edge, label %if.then6
_ZN4llvm6Record8getValueENS_9StringRefE.exit.return_crit_edge: ; preds = %_ZN4llvm6Record8getValueENS_9StringRefE.exit
br label %return
if.then6: ; preds = %_ZN4llvm6Record8getValueENS_9StringRefE.exit
%cmp = icmp eq %"class.llvm::RecordVal"* %add.ptr.i.i, %RV ; <i1> [#uses=1]
br i1 %cmp, label %if.then6.if.end_crit_edge, label %land.lhs.true
if.then6.if.end_crit_edge: ; preds = %if.then6
br label %if.end
land.lhs.true: ; preds = %if.then6
%tobool10 = icmp eq %"class.llvm::RecordVal"* %RV, null ; <i1> [#uses=1]
br i1 %tobool10, label %lor.lhs.false, label %land.lhs.true.return_crit_edge
land.lhs.true.return_crit_edge: ; preds = %land.lhs.true
br label %return
lor.lhs.false: ; preds = %land.lhs.true
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%tmp.i3 = getelementptr inbounds %"class.llvm::RecordVal", %"class.llvm::RecordVal"* %tmp6.i.i, i32 %.reload, i32 3 ; <%"struct.llvm::Init"**> [#uses=1]
%tmp2.i4 = load %"struct.llvm::Init"*, %"struct.llvm::Init"** %tmp.i3 ; <%"struct.llvm::Init"*> [#uses=2]
%5 = icmp eq %"struct.llvm::Init"* %tmp2.i4, null ; <i1> [#uses=1]
br i1 %5, label %lor.lhs.false.if.end_crit_edge, label %tmpbb1
lor.lhs.false.if.end_crit_edge: ; preds = %lor.lhs.false
br label %if.end
tmpbb1: ; preds = %lor.lhs.false
%6 = bitcast %"struct.llvm::Init"* %tmp2.i4 to i8* ; <i8*> [#uses=1]
%7 = tail call i8* @__dynamic_cast(i8* %6, i8* bitcast (%0* @_ZTIN4llvm4InitE to i8*), i8* bitcast (%1* @_ZTIN4llvm9UnsetInitE to i8*), i32 -1) ; <i8*> [#uses=1]
%phitmp32 = icmp eq i8* %7, null ; <i1> [#uses=1]
br i1 %phitmp32, label %.if.end_crit_edge, label %.return_crit_edge1
.return_crit_edge1: ; preds = %tmpbb1
br label %return
.if.end_crit_edge: ; preds = %tmpbb1
br label %if.end
if.end: ; preds = %.if.end_crit_edge, %lor.lhs.false.if.end_crit_edge, %if.then6.if.end_crit_edge
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%tmp.i1 = getelementptr inbounds %"class.llvm::RecordVal", %"class.llvm::RecordVal"* %tmp6.i.i, i32 %.reload, i32 3 ; <%"struct.llvm::Init"**> [#uses=1]
%tmp2.i2 = load %"struct.llvm::Init"*, %"struct.llvm::Init"** %tmp.i1 ; <%"struct.llvm::Init"*> [#uses=3]
%8 = bitcast %"class.llvm::StringInit"* %this to %"struct.llvm::Init"* ; <%"struct.llvm::Init"*> [#uses=1]
%cmp19 = icmp eq %"struct.llvm::Init"* %tmp2.i2, %8 ; <i1> [#uses=1]
br i1 %cmp19, label %cond.false, label %cond.end
cond.false: ; preds = %if.end
tail call void @__assert_fail(i8* getelementptr inbounds ([45 x i8], [45 x i8]* @.str51, i32 0, i32 0), i8* getelementptr inbounds ([47 x i8], [47 x i8]* @.str8, i32 0, i32 0), i32 1141, i8* getelementptr inbounds ([116 x i8], [116 x i8]* @__PRETTY_FUNCTION__._ZNK4llvm7VarInit12getFieldInitERNS_6RecordEPKNS_9RecordValERKSs, i32 0, i32 0)) noreturn
unreachable
cond.end: ; preds = %if.end
%9 = bitcast %"struct.llvm::Init"* %tmp2.i2 to %"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)*** ; <%"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)***> [#uses=1]
%10 = load %"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)**, %"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)*** %9 ; <%"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)**> [#uses=1]
[opaque pointer type] Add textual IR support for explicit type parameter to getelementptr instruction 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
2015-02-28 03:29:02 +08:00
%vfn = getelementptr inbounds %"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)*, %"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)** %10, i32 8 ; <%"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)**> [#uses=1]
%11 = load %"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)*, %"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)** %vfn ; <%"struct.llvm::Init"* (%"struct.llvm::Init"*, %"class.llvm::Record"*, %"class.llvm::RecordVal"*, %"class.std::basic_string"*)*> [#uses=1]
%call25 = tail call %"struct.llvm::Init"* %11(%"struct.llvm::Init"* %tmp2.i2, %"class.llvm::Record"* %R, %"class.llvm::RecordVal"* %RV, %"class.std::basic_string"* %FieldName) ; <%"struct.llvm::Init"*> [#uses=1]
ret %"struct.llvm::Init"* %call25
return: ; preds = %.return_crit_edge1, %land.lhs.true.return_crit_edge, %_ZN4llvm6Record8getValueENS_9StringRefE.exit.return_crit_edge, %for.cond.i.return_crit_edge, %.return_crit_edge, %entry.return_crit_edge
ret %"struct.llvm::Init"* null
}
declare i1 @_ZNK4llvm7VarInit12getFieldInitERNS_6RecordEPKNS_9RecordValERKSs_for.cond.i(i32, %"class.llvm::RecordVal"*, i32, i8*, i32*)