maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/atomic-minmax-i6432.ll

301 lines
9.1 KiB
LLVM
Raw Normal View History

[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
[X86] Add FeatureCMOV explicitly to all CPUs that support it. Remove FeatureCMOV implication from Feature64Bit and FeatureSSE1 Summary: Previously most CPUs inherited cmov support through Feature64Bit(or FeatureCMPXCHG16HB implying Feature64Bit) or FeatureSSE1. This has the surprising side effect that -mattr=-cmov causes an assert to fire in 64-bit mode because it clears the Feature64Bit. Or in 32-bit mode, -mattr=-cmov disables any sse/avx features which seems surprising. This patch removes the implication and instead updates hasCMOV in X86Subtarget to check SSE1 or is64Bit in addition to the regular cmov flag. This should keep most things working the way they did before. I don't believe there is a way to specific "-cmov" directly from clang so this should only effect our lower level tools. This does stop -mattr=cx16(cmpxchg16b) from implying cmov is enabled via the 64bit flag as you can see from one of the changed tests. But that was a 32-bit test so I don't know why it enabled cx16 anyway. For the other test I had to add -sse to override the new sse check in hasCMOV. Reviewers: RKSimon, DavidKreitzer, spatel Reviewed By: RKSimon Subscribers: llvm-commits, jfb Differential Revision: https://reviews.llvm.org/D51228 llvm-svn: 340707
2018-08-27 02:29:33 +08:00
; RUN: llc -mattr=+cmov -mtriple=i386-pc-linux -verify-machineinstrs < %s | FileCheck %s -check-prefix=LINUX
; RUN: llc -mattr=+cmov -mtriple=i386-macosx -relocation-model=pic -verify-machineinstrs < %s | FileCheck %s -check-prefix=PIC
When expanding atomic load arith instructions, do not lose target flags. rdar://12453106 llvm-svn: 165568
2012-10-10 07:48:33 +08:00
Add missing i64 max/min/umax/umin on 32-bit target - Turn on atomic6432.ll and add specific test case as well llvm-svn: 164616
2012-09-26 02:08:13 +08:00
@sc64 = external global i64
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
define i64 @atomic_max_i64() nounwind {
; LINUX-LABEL: atomic_max_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX: # %bb.0: # %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: pushl %ebx
; LINUX-NEXT: pushl %esi
; LINUX-NEXT: movl sc64+4, %edx
; LINUX-NEXT: movl sc64, %eax
; LINUX-NEXT: movl $4, %esi
; LINUX-NEXT: .p2align 4, 0x90
; LINUX-NEXT: .LBB0_1: # %atomicrmw.start
; LINUX-NEXT: # =>This Inner Loop Header: Depth=1
; LINUX-NEXT: cmpl %eax, %esi
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: movl $0, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: sbbl %edx, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: movl $0, %ecx
; LINUX-NEXT: cmovll %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: movl $5, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: cmovll %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: lock cmpxchg8b sc64
; LINUX-NEXT: jne .LBB0_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX-NEXT: # %bb.2: # %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: popl %esi
; LINUX-NEXT: popl %ebx
; LINUX-NEXT: retl
;
; PIC-LABEL: atomic_max_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC: ## %bb.0: ## %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: pushl %ebx
; PIC-NEXT: pushl %edi
; PIC-NEXT: pushl %esi
; PIC-NEXT: calll L0$pb
; PIC-NEXT: L0$pb:
; PIC-NEXT: popl %eax
; PIC-NEXT: movl L_sc64$non_lazy_ptr-L0$pb(%eax), %esi
; PIC-NEXT: movl (%esi), %eax
; PIC-NEXT: movl 4(%esi), %edx
; PIC-NEXT: movl $4, %edi
; PIC-NEXT: .p2align 4, 0x90
; PIC-NEXT: LBB0_1: ## %atomicrmw.start
; PIC-NEXT: ## =>This Inner Loop Header: Depth=1
; PIC-NEXT: cmpl %eax, %edi
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: movl $0, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: sbbl %edx, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: movl $0, %ecx
; PIC-NEXT: cmovll %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: movl $5, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: cmovll %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: lock cmpxchg8b (%esi)
; PIC-NEXT: jne LBB0_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC-NEXT: ## %bb.2: ## %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: popl %esi
; PIC-NEXT: popl %edi
; PIC-NEXT: popl %ebx
; PIC-NEXT: retl
; PIC-NEXT: ## -- End function
entry:
%max = atomicrmw max i64* @sc64, i64 5 acquire
ret i64 %max
}
define i64 @atomic_min_i64() nounwind {
; LINUX-LABEL: atomic_min_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX: # %bb.0: # %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: pushl %ebx
; LINUX-NEXT: movl sc64+4, %edx
; LINUX-NEXT: movl sc64, %eax
; LINUX-NEXT: .p2align 4, 0x90
; LINUX-NEXT: .LBB1_1: # %atomicrmw.start
; LINUX-NEXT: # =>This Inner Loop Header: Depth=1
; LINUX-NEXT: cmpl $7, %eax
; LINUX-NEXT: movl %edx, %ecx
; LINUX-NEXT: sbbl $0, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: movl $0, %ecx
; LINUX-NEXT: cmovll %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: movl $6, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: cmovll %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: lock cmpxchg8b sc64
; LINUX-NEXT: jne .LBB1_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX-NEXT: # %bb.2: # %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: popl %ebx
; LINUX-NEXT: retl
;
; PIC-LABEL: atomic_min_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC: ## %bb.0: ## %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: pushl %ebx
; PIC-NEXT: pushl %esi
; PIC-NEXT: calll L1$pb
; PIC-NEXT: L1$pb:
; PIC-NEXT: popl %eax
; PIC-NEXT: movl L_sc64$non_lazy_ptr-L1$pb(%eax), %esi
; PIC-NEXT: movl (%esi), %eax
; PIC-NEXT: movl 4(%esi), %edx
; PIC-NEXT: .p2align 4, 0x90
; PIC-NEXT: LBB1_1: ## %atomicrmw.start
; PIC-NEXT: ## =>This Inner Loop Header: Depth=1
; PIC-NEXT: cmpl $7, %eax
; PIC-NEXT: movl %edx, %ecx
; PIC-NEXT: sbbl $0, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: movl $0, %ecx
; PIC-NEXT: cmovll %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: movl $6, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: cmovll %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: lock cmpxchg8b (%esi)
; PIC-NEXT: jne LBB1_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC-NEXT: ## %bb.2: ## %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: popl %esi
; PIC-NEXT: popl %ebx
; PIC-NEXT: retl
; PIC-NEXT: ## -- End function
entry:
%min = atomicrmw min i64* @sc64, i64 6 acquire
ret i64 %min
}
define i64 @atomic_umax_i64() nounwind {
; LINUX-LABEL: atomic_umax_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX: # %bb.0: # %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: pushl %ebx
; LINUX-NEXT: pushl %esi
; LINUX-NEXT: movl sc64+4, %edx
; LINUX-NEXT: movl sc64, %eax
; LINUX-NEXT: movl $7, %esi
; LINUX-NEXT: .p2align 4, 0x90
; LINUX-NEXT: .LBB2_1: # %atomicrmw.start
; LINUX-NEXT: # =>This Inner Loop Header: Depth=1
; LINUX-NEXT: cmpl %eax, %esi
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: movl $0, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: sbbl %edx, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: movl $0, %ecx
; LINUX-NEXT: cmovbl %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: movl $7, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: cmovbl %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: lock cmpxchg8b sc64
; LINUX-NEXT: jne .LBB2_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX-NEXT: # %bb.2: # %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: popl %esi
; LINUX-NEXT: popl %ebx
; LINUX-NEXT: retl
;
; PIC-LABEL: atomic_umax_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC: ## %bb.0: ## %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: pushl %ebx
; PIC-NEXT: pushl %edi
; PIC-NEXT: pushl %esi
; PIC-NEXT: calll L2$pb
; PIC-NEXT: L2$pb:
; PIC-NEXT: popl %eax
; PIC-NEXT: movl L_sc64$non_lazy_ptr-L2$pb(%eax), %esi
; PIC-NEXT: movl (%esi), %eax
; PIC-NEXT: movl 4(%esi), %edx
; PIC-NEXT: movl $7, %edi
; PIC-NEXT: .p2align 4, 0x90
; PIC-NEXT: LBB2_1: ## %atomicrmw.start
; PIC-NEXT: ## =>This Inner Loop Header: Depth=1
; PIC-NEXT: cmpl %eax, %edi
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: movl $0, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: sbbl %edx, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: movl $0, %ecx
; PIC-NEXT: cmovbl %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: movl $7, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: cmovbl %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: lock cmpxchg8b (%esi)
; PIC-NEXT: jne LBB2_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC-NEXT: ## %bb.2: ## %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: popl %esi
; PIC-NEXT: popl %edi
; PIC-NEXT: popl %ebx
; PIC-NEXT: retl
; PIC-NEXT: ## -- End function
entry:
%umax = atomicrmw umax i64* @sc64, i64 7 acquire
ret i64 %umax
}
define i64 @atomic_umin_i64() nounwind {
; LINUX-LABEL: atomic_umin_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX: # %bb.0: # %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: pushl %ebx
; LINUX-NEXT: movl sc64+4, %edx
; LINUX-NEXT: movl sc64, %eax
; LINUX-NEXT: .p2align 4, 0x90
; LINUX-NEXT: .LBB3_1: # %atomicrmw.start
; LINUX-NEXT: # =>This Inner Loop Header: Depth=1
; LINUX-NEXT: cmpl $9, %eax
; LINUX-NEXT: movl %edx, %ecx
; LINUX-NEXT: sbbl $0, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: movl $0, %ecx
; LINUX-NEXT: cmovbl %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: movl $8, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; LINUX-NEXT: cmovbl %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: lock cmpxchg8b sc64
; LINUX-NEXT: jne .LBB3_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX-NEXT: # %bb.2: # %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: popl %ebx
; LINUX-NEXT: retl
;
; PIC-LABEL: atomic_umin_i64:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC: ## %bb.0: ## %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: pushl %ebx
; PIC-NEXT: pushl %esi
; PIC-NEXT: calll L3$pb
; PIC-NEXT: L3$pb:
; PIC-NEXT: popl %eax
; PIC-NEXT: movl L_sc64$non_lazy_ptr-L3$pb(%eax), %esi
; PIC-NEXT: movl (%esi), %eax
; PIC-NEXT: movl 4(%esi), %edx
; PIC-NEXT: .p2align 4, 0x90
; PIC-NEXT: LBB3_1: ## %atomicrmw.start
; PIC-NEXT: ## =>This Inner Loop Header: Depth=1
; PIC-NEXT: cmpl $9, %eax
; PIC-NEXT: movl %edx, %ecx
; PIC-NEXT: sbbl $0, %ecx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: movl $0, %ecx
; PIC-NEXT: cmovbl %edx, %ecx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: movl $8, %ebx
[x86] Fix an amazing goof in the handling of sub, or, and xor lowering. The comment for this code indicated that it should work similar to our handling of add lowering above: if we see uses of an instruction other than flag usage and store usage, it tries to avoid the specialized X86ISD::* nodes that are designed for flag+op modeling and emits an explicit test. Problem is, only the add case actually did this. In all the other cases, the logic was incomplete and inverted. Any time the value was used by a store, we bailed on the specialized X86ISD node. All of this appears to have been historical where we had different logic here. =/ Turns out, we have quite a few patterns designed around these nodes. We should actually form them. I fixed the code to match what we do for add, and it has quite a positive effect just within some of our test cases. The only thing close to a regression I see is using: notl %r testl %r, %r instead of: xorl -1, %r But we can add a pattern or something to fold that back out. The improvements seem more than worth this. I've also worked with Craig to update the comments to no longer be actively contradicted by the code. =[ Some of this still remains a mystery to both Craig and myself, but this seems like a large step in the direction of consistency and slightly more accurate comments. Many thanks to Craig for help figuring out this nasty stuff. Differential Revision: https://reviews.llvm.org/D37096 llvm-svn: 311737
2017-08-25 08:34:07 +08:00
; PIC-NEXT: cmovbl %eax, %ebx
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: lock cmpxchg8b (%esi)
; PIC-NEXT: jne LBB3_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC-NEXT: ## %bb.2: ## %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: popl %esi
; PIC-NEXT: popl %ebx
; PIC-NEXT: retl
; PIC-NEXT: ## -- End function
entry:
%umin = atomicrmw umin i64* @sc64, i64 8 acquire
ret i64 %umin
When expanding atomic load arith instructions, do not lose target flags. rdar://12453106 llvm-svn: 165568
2012-10-10 07:48:33 +08:00
}
@id = internal global i64 0, align 8
define void @tf_bug(i8* %ptr) nounwind {
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-LABEL: tf_bug:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX: # %bb.0: # %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: pushl %ebx
; LINUX-NEXT: pushl %esi
; LINUX-NEXT: movl {{[0-9]+}}(%esp), %esi
; LINUX-NEXT: movl id+4, %edx
; LINUX-NEXT: movl id, %eax
; LINUX-NEXT: .p2align 4, 0x90
; LINUX-NEXT: .LBB4_1: # %atomicrmw.start
; LINUX-NEXT: # =>This Inner Loop Header: Depth=1
; LINUX-NEXT: movl %eax, %ebx
; LINUX-NEXT: addl $1, %ebx
; LINUX-NEXT: movl %edx, %ecx
; LINUX-NEXT: adcl $0, %ecx
; LINUX-NEXT: lock cmpxchg8b id
; LINUX-NEXT: jne .LBB4_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; LINUX-NEXT: # %bb.2: # %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; LINUX-NEXT: addl $1, %eax
; LINUX-NEXT: adcl $0, %edx
; LINUX-NEXT: movl %eax, (%esi)
; LINUX-NEXT: movl %edx, 4(%esi)
; LINUX-NEXT: popl %esi
; LINUX-NEXT: popl %ebx
; LINUX-NEXT: retl
;
Mass update to CodeGen tests to use CHECK-LABEL for labels corresponding to function definitions for more informative error messages. No functionality change and all updated tests passed locally. This update was done with the following bash script: find test/CodeGen -name "*.ll" | \ while read NAME; do echo "$NAME" if ! grep -q "^; *RUN: *llc.*debug" $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 sed -i '' "s/;\(.*\)-LABEL-LABEL:/;\1-LABEL:/" $TEMP sed -i '' "s/;\(.*\)-NEXT-LABEL:/;\1-NEXT:/" $TEMP sed -i '' "s/;\(.*\)-NOT-LABEL:/;\1-NOT:/" $TEMP sed -i '' "s/;\(.*\)-DAG-LABEL:/;\1-DAG:/" $TEMP mv $TEMP $NAME fi done llvm-svn: 186280
2013-07-14 14:24:09 +08:00
; PIC-LABEL: tf_bug:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC: ## %bb.0: ## %entry
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: pushl %ebx
; PIC-NEXT: pushl %edi
; PIC-NEXT: pushl %esi
; PIC-NEXT: calll L4$pb
; PIC-NEXT: L4$pb:
; PIC-NEXT: popl %edi
; PIC-NEXT: movl {{[0-9]+}}(%esp), %esi
; PIC-NEXT: movl (_id-L4$pb)+4(%edi), %edx
; PIC-NEXT: movl _id-L4$pb(%edi), %eax
; PIC-NEXT: .p2align 4, 0x90
; PIC-NEXT: LBB4_1: ## %atomicrmw.start
; PIC-NEXT: ## =>This Inner Loop Header: Depth=1
; PIC-NEXT: movl %eax, %ebx
; PIC-NEXT: addl $1, %ebx
; PIC-NEXT: movl %edx, %ecx
; PIC-NEXT: adcl $0, %ecx
; PIC-NEXT: lock cmpxchg8b _id-L4$pb(%edi)
; PIC-NEXT: jne LBB4_1
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; PIC-NEXT: ## %bb.2: ## %atomicrmw.end
[x86] NFC: Clean up two tests and generate precise checks for them. Mostly this involved giving unnamed values names and running the IR through `opt` to re-format it but merging in any important comments in the original. I then deleted pointless comments and inlined the function attributes for ease of reading and editting. All of this is to make it much easier to see the instructions being generated here and evaluate any updates to the tests. llvm-svn: 311634
2017-08-24 15:38:36 +08:00
; PIC-NEXT: addl $1, %eax
; PIC-NEXT: adcl $0, %edx
; PIC-NEXT: movl %eax, (%esi)
; PIC-NEXT: movl %edx, 4(%esi)
; PIC-NEXT: popl %esi
; PIC-NEXT: popl %edi
; PIC-NEXT: popl %ebx
; PIC-NEXT: retl
; PIC-NEXT: ## -- End function
; PIC-NEXT: .zerofill __DATA,__bss,_id,8,3 ## @id
entry:
When expanding atomic load arith instructions, do not lose target flags. rdar://12453106 llvm-svn: 165568
2012-10-10 07:48:33 +08:00
%tmp1 = atomicrmw add i64* @id, i64 1 seq_cst
%tmp2 = add i64 %tmp1, 1
%tmp3 = bitcast i8* %ptr to i64*
store i64 %tmp2, i64* %tmp3, align 4
Add missing i64 max/min/umax/umin on 32-bit target - Turn on atomic6432.ll and add specific test case as well llvm-svn: 164616
2012-09-26 02:08:13 +08:00
ret void
}