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/ObjCARC/contract-storestrong.ll

262 lines
8.8 KiB
LLVM
Raw Normal View History

The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; RUN: opt -objc-arc-contract -S < %s | FileCheck %s
target datalayout = "e-p:64:64:64"
declare i8* @objc_retain(i8*)
declare void @objc_release(i8*)
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
declare void @use_pointer(i8*)
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
@x = external global i8*
Catch more CHECK that can be converted to CHECK-LABEL in Transforms for easier debugging. No functionality change. This conversion 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/;\(.*\)\([A-Za-z0-9_]*\):\( *\)define\([^@]*\)@$FUNC\([( ]*\)\$/;\1\2-LABEL:\3define\4@$FUNC(/g" $TEMP done mv $TEMP $NAME fi done llvm-svn: 186269
2013-07-14 09:50:49 +08:00
; CHECK-LABEL: define void @test0(
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; CHECK: entry:
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: tail call void @objc_storeStrong(i8** @x, i8* %p) [[NUW:#[0-9]+]]
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; CHECK-NEXT: ret void
[objc-arc] Test cleanups. Mainly adding paranoid checks for the closing brace of a function to help with FileCheck error readability. Also some other minor changes. No actual CHECK changes. llvm-svn: 180668
2013-04-27 13:25:54 +08:00
; CHECK-NEXT: }
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
define void @test0(i8* %p) {
entry:
%0 = tail call i8* @objc_retain(i8* %p) nounwind
[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 i8*, i8** @x, align 8
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
store i8* %0, i8** @x, align 8
tail call void @objc_release(i8* %tmp) nounwind
ret void
}
; Don't do this if the load is volatile.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define void @test1(i8* %p) {
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; CHECK-NEXT: entry:
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: %0 = tail call i8* @objc_retain(i8* %p) [[NUW]]
[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
; CHECK-NEXT: %tmp = load volatile i8*, i8** @x, align 8
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; CHECK-NEXT: store i8* %0, i8** @x, align 8
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: tail call void @objc_release(i8* %tmp) [[NUW]]
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; CHECK-NEXT: ret void
; CHECK-NEXT: }
define void @test1(i8* %p) {
entry:
%0 = tail call i8* @objc_retain(i8* %p) nounwind
[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 volatile i8*, i8** @x, align 8
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
store i8* %0, i8** @x, align 8
tail call void @objc_release(i8* %tmp) nounwind
ret void
}
; Don't do this if the store is volatile.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define void @test2(i8* %p) {
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; CHECK-NEXT: entry:
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: %0 = tail call i8* @objc_retain(i8* %p) [[NUW]]
[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
; CHECK-NEXT: %tmp = load i8*, i8** @x, align 8
Move "atomic" and "volatile" designations on instructions after the opcode of the instruction. Note that this change affects the existing non-atomic load and store instructions; the parser now accepts both forms, and the change is noted in the release notes. llvm-svn: 137527
2011-08-13 06:50:01 +08:00
; CHECK-NEXT: store volatile i8* %0, i8** @x, align 8
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: tail call void @objc_release(i8* %tmp) [[NUW]]
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
; CHECK-NEXT: ret void
; CHECK-NEXT: }
define void @test2(i8* %p) {
entry:
%0 = tail call i8* @objc_retain(i8* %p) nounwind
[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 i8*, i8** @x, align 8
Upgrade syntax of tests using volatile instructions to use 'load volatile' instead of 'volatile load', which is archaic. llvm-svn: 145171
2011-11-27 14:54:59 +08:00
store volatile i8* %0, i8** @x, align 8
The ARC language-specific optimizer. Credit to Dan Gohman. llvm-svn: 133108
2011-06-16 07:37:01 +08:00
tail call void @objc_release(i8* %tmp) nounwind
ret void
}
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
Fixed a typo llvm-svn: 156471
2012-05-09 22:43:50 +08:00
; Don't do this if there's a use of the old pointer value between the store
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
; and the release.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define void @test3(i8* %newValue) {
; CHECK-NEXT: entry:
; CHECK-NEXT: %x0 = tail call i8* @objc_retain(i8* %newValue) [[NUW]]
[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
; CHECK-NEXT: %x1 = load i8*, i8** @x, align 8
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-NEXT: store i8* %x0, i8** @x, align 8
; CHECK-NEXT: tail call void @use_pointer(i8* %x1), !clang.arc.no_objc_arc_exceptions !0
; CHECK-NEXT: tail call void @objc_release(i8* %x1) [[NUW]], !clang.imprecise_release !0
; CHECK-NEXT: ret void
; CHECK-NEXT: }
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
define void @test3(i8* %newValue) {
entry:
%x0 = tail call i8* @objc_retain(i8* %newValue) nounwind
[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
%x1 = load i8*, i8** @x, align 8
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
store i8* %newValue, i8** @x, align 8
tail call void @use_pointer(i8* %x1), !clang.arc.no_objc_arc_exceptions !0
tail call void @objc_release(i8* %x1) nounwind, !clang.imprecise_release !0
ret void
}
; Like test3, but with an icmp use instead of a call, for good measure.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define i1 @test4(i8* %newValue, i8* %foo) {
; CHECK-NEXT: entry:
; CHECK-NEXT: %x0 = tail call i8* @objc_retain(i8* %newValue) [[NUW]]
[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
; CHECK-NEXT: %x1 = load i8*, i8** @x, align 8
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-NEXT: store i8* %x0, i8** @x, align 8
; CHECK-NEXT: %t = icmp eq i8* %x1, %foo
; CHECK-NEXT: tail call void @objc_release(i8* %x1) [[NUW]], !clang.imprecise_release !0
; CHECK-NEXT: ret i1 %t
; CHECK-NEXT: }
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
define i1 @test4(i8* %newValue, i8* %foo) {
entry:
%x0 = tail call i8* @objc_retain(i8* %newValue) nounwind
[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
%x1 = load i8*, i8** @x, align 8
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
store i8* %newValue, i8** @x, align 8
%t = icmp eq i8* %x1, %foo
tail call void @objc_release(i8* %x1) nounwind, !clang.imprecise_release !0
ret i1 %t
}
; Do form an objc_storeStrong here, because the use is before the store.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define i1 @test5(i8* %newValue, i8* %foo) {
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
; CHECK: %t = icmp eq i8* %x1, %foo
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK: tail call void @objc_storeStrong(i8** @x, i8* %newValue) [[NUW]]
[objc-arc] Test cleanups. Mainly adding paranoid checks for the closing brace of a function to help with FileCheck error readability. Also some other minor changes. No actual CHECK changes. llvm-svn: 180668
2013-04-27 13:25:54 +08:00
; CHECK: }
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
define i1 @test5(i8* %newValue, i8* %foo) {
entry:
%x0 = tail call i8* @objc_retain(i8* %newValue) nounwind
[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
%x1 = load i8*, i8** @x, align 8
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
%t = icmp eq i8* %x1, %foo
store i8* %newValue, i8** @x, align 8
tail call void @objc_release(i8* %x1) nounwind, !clang.imprecise_release !0
ret i1 %t
}
; Like test5, but the release is before the store.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define i1 @test6(i8* %newValue, i8* %foo) {
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
; CHECK: %t = icmp eq i8* %x1, %foo
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK: tail call void @objc_storeStrong(i8** @x, i8* %newValue) [[NUW]]
[objc-arc] Test cleanups. Mainly adding paranoid checks for the closing brace of a function to help with FileCheck error readability. Also some other minor changes. No actual CHECK changes. llvm-svn: 180668
2013-04-27 13:25:54 +08:00
; CHECK: }
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
define i1 @test6(i8* %newValue, i8* %foo) {
entry:
%x0 = tail call i8* @objc_retain(i8* %newValue) nounwind
[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
%x1 = load i8*, i8** @x, align 8
Fix objc_storeStrong pattern matching to catch a potential use of the old value after the store but before it is released. This fixes rdar:/11116986. llvm-svn: 156442
2012-05-09 07:34:08 +08:00
tail call void @objc_release(i8* %x1) nounwind, !clang.imprecise_release !0
%t = icmp eq i8* %x1, %foo
store i8* %newValue, i8** @x, align 8
ret i1 %t
}
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
; Like test0, but there's no store, so don't form an objc_storeStrong.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define void @test7(
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
; CHECK-NEXT: entry:
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: %0 = tail call i8* @objc_retain(i8* %p) [[NUW]]
[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
; CHECK-NEXT: %tmp = load i8*, i8** @x, align 8
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: tail call void @objc_release(i8* %tmp) [[NUW]]
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
; CHECK-NEXT: ret void
; CHECK-NEXT: }
define void @test7(i8* %p) {
entry:
%0 = tail call i8* @objc_retain(i8* %p) nounwind
[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 i8*, i8** @x, align 8
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
tail call void @objc_release(i8* %tmp) nounwind
ret void
}
; Like test0, but there's no retain, so don't form an objc_storeStrong.
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; CHECK-LABEL: define void @test8(
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
; CHECK-NEXT: 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
; CHECK-NEXT: %tmp = load i8*, i8** @x, align 8
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
; CHECK-NEXT: store i8* %p, i8** @x, align 8
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK-NEXT: tail call void @objc_release(i8* %tmp) [[NUW]]
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
; CHECK-NEXT: ret void
; CHECK-NEXT: }
define void @test8(i8* %p) {
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 i8*, i8** @x, align 8
Fix the objc_storeStrong recognizer to stop before walking off the end of a basic block if there's no store. llvm-svn: 156520
2012-05-10 07:08:33 +08:00
store i8* %p, i8** @x, align 8
tail call void @objc_release(i8* %tmp) nounwind
ret void
}
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
; Make sure that we properly handle release that *may* release our new
; value in between the retain and the store. We need to be sure that
; this we can safely move the retain to the store. This specific test
; makes sure that we properly handled a release of an unrelated
; pointer.
;
; CHECK-LABEL: define i1 @test9(i8* %newValue, i8* %foo, i8* %unrelated_ptr) {
; CHECK-NOT: objc_storeStrong
define i1 @test9(i8* %newValue, i8* %foo, i8* %unrelated_ptr) {
entry:
%x0 = tail call i8* @objc_retain(i8* %newValue) nounwind
tail call void @objc_release(i8* %unrelated_ptr) nounwind, !clang.imprecise_release !0
[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
%x1 = load i8*, i8** @x, align 8
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
tail call void @objc_release(i8* %x1) nounwind, !clang.imprecise_release !0
%t = icmp eq i8* %x1, %foo
store i8* %newValue, i8** @x, align 8
ret i1 %t
}
; Make sure that we don't perform the optimization when we just have a call.
;
; CHECK-LABEL: define i1 @test10(i8* %newValue, i8* %foo, i8* %unrelated_ptr) {
; CHECK-NOT: objc_storeStrong
define i1 @test10(i8* %newValue, i8* %foo, i8* %unrelated_ptr) {
entry:
%x0 = tail call i8* @objc_retain(i8* %newValue) nounwind
call void @use_pointer(i8* %unrelated_ptr)
[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
%x1 = load i8*, i8** @x, align 8
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
tail call void @objc_release(i8* %x1) nounwind, !clang.imprecise_release !0
%t = icmp eq i8* %x1, %foo
store i8* %newValue, i8** @x, align 8
ret i1 %t
}
; Make sure we form the store strong if the use in between the retain
; and the store does not touch reference counts.
; CHECK-LABEL: define i1 @test11(i8* %newValue, i8* %foo, i8* %unrelated_ptr) {
; CHECK: objc_storeStrong
define i1 @test11(i8* %newValue, i8* %foo, i8* %unrelated_ptr) {
entry:
%x0 = tail call i8* @objc_retain(i8* %newValue) nounwind
%t = icmp eq i8* %newValue, %foo
[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
%x1 = load i8*, i8** @x, align 8
[objc-arc-contract] We can not move retains over instructions which can not conservatively be proven to not decrement the retain's RCIdentity. I also cleaned up the code to make it more understandable for mere mortals. <rdar://problem/19853758> llvm-svn: 229937
2015-02-20 08:02:49 +08:00
tail call void @objc_release(i8* %x1) nounwind, !clang.imprecise_release !0
store i8* %newValue, i8** @x, align 8
ret i1 %t
}
Form objc_storeStrong in the presence of bitcasts. objc_storeStrong can be formed from a sequence such as %0 = tail call i8* @objc_retain(i8* %p) nounwind %tmp = load i8*, i8** @x, align 8 store i8* %0, i8** @x, align 8 tail call void @objc_release(i8* %tmp) nounwind The code was already looking through bitcasts for most of the values involved, but had missed one case where the pointer operand for the store was a bitcast. Ultimately the pointer for the load and store have to be the same value, after stripping casts. llvm-svn: 270955
2016-05-27 10:13:53 +08:00
; Make sure that we form the store strong even if there are bitcasts on
; the pointers.
; CHECK-LABEL: define void @test12(
; CHECK: entry:
; CHECK-NEXT: %p16 = bitcast i8** @x to i16**
; CHECK-NEXT: %tmp16 = load i16*, i16** %p16, align 8
; CHECK-NEXT: %tmp8 = bitcast i16* %tmp16 to i8*
; CHECK-NEXT: %p32 = bitcast i8** @x to i32**
; CHECK-NEXT: %v32 = bitcast i8* %p to i32*
; CHECK-NEXT: %0 = bitcast i16** %p16 to i8**
; CHECK-NEXT: tail call void @objc_storeStrong(i8** %0, i8* %p)
; CHECK-NEXT: ret void
; CHECK-NEXT: }
define void @test12(i8* %p) {
entry:
%retain = tail call i8* @objc_retain(i8* %p) nounwind
%p16 = bitcast i8** @x to i16**
%tmp16 = load i16*, i16** %p16, align 8
%tmp8 = bitcast i16* %tmp16 to i8*
%p32 = bitcast i8** @x to i32**
%v32 = bitcast i8* %retain to i32*
store i32* %v32, i32** %p32, align 8
tail call void @objc_release(i8* %tmp8) nounwind
ret void
}
[ObjCArc] Do not dereference an invalidated iterator. Fix a bug in ARC contract pass where an iterator that pointed to a deleted instruction was dereferenced. It appears that tryToContractReleaseIntoStoreStrong was incorrectly assuming that a call to objc_retain would not immediately follow a call to objc_release. rdar://problem/25276306 llvm-svn: 299507
2017-04-05 11:44:09 +08:00
; This used to crash.
; CHECK-LABEL: define i8* @test13(
; CHECK: tail call void @objc_storeStrong(i8** %{{.*}}, i8* %[[NEW:.*]])
; CHECK-NEXT: ret i8* %[[NEW]]
define i8* @test13(i8* %a0, i8* %a1, i8** %addr, i8* %new) {
%old = load i8*, i8** %addr, align 8
call void @objc_release(i8* %old)
%retained = call i8* @objc_retain(i8* %new)
store i8* %retained, i8** %addr, align 8
ret i8* %retained
}
IR: Make metadata typeless in assembly Now that `Metadata` is typeless, reflect that in the assembly. These are the matching assembly changes for the metadata/value split in r223802. - Only use the `metadata` type when referencing metadata from a call intrinsic -- i.e., only when it's used as a `Value`. - Stop pretending that `ValueAsMetadata` is wrapped in an `MDNode` when referencing it from call intrinsics. So, assembly like this: define @foo(i32 %v) { call void @llvm.foo(metadata !{i32 %v}, metadata !0) call void @llvm.foo(metadata !{i32 7}, metadata !0) call void @llvm.foo(metadata !1, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{metadata !3}, metadata !0) ret void, !bar !2 } !0 = metadata !{metadata !2} !1 = metadata !{i32* @global} !2 = metadata !{metadata !3} !3 = metadata !{} turns into this: define @foo(i32 %v) { call void @llvm.foo(metadata i32 %v, metadata !0) call void @llvm.foo(metadata i32 7, metadata !0) call void @llvm.foo(metadata i32* @global, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{!3}, metadata !0) ret void, !bar !2 } !0 = !{!2} !1 = !{i32* @global} !2 = !{!3} !3 = !{} I wrote an upgrade script that handled almost all of the tests in llvm and many of the tests in cfe (even handling many `CHECK` lines). I've attached it (or will attach it in a moment if you're speedy) to PR21532 to help everyone update their out-of-tree testcases. This is part of PR21532. llvm-svn: 224257
2014-12-16 03:07:53 +08:00
!0 = !{}
Use references to attribute groups on the call/invoke instructions. Listing all of the attributes for the callee of a call/invoke instruction is way too much and makes the IR unreadable. Use references to attributes instead. llvm-svn: 175877
2013-02-22 17:09:42 +08:00
; CHECK: attributes [[NUW]] = { nounwind }