6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-28 04:33:05 +00:00
Code Issues Projects Releases Wiki Activity
2e3524ec17
llvm-6502/test/CodeGen/Mips/atomic.ll

448 lines
17 KiB
LLVM
Raw Normal View History

[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; RUN: llc -march=mipsel --disable-machine-licm -mcpu=mips32 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32-ANY -check-prefix=NO-SEB-SEH -check-prefix=CHECK-EL -check-prefix=NOT-MICROMIPS
; RUN: llc -march=mipsel --disable-machine-licm -mcpu=mips32r2 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32-ANY -check-prefix=HAS-SEB-SEH -check-prefix=CHECK-EL -check-prefix=NOT-MICROMIPS
; RUN: llc -march=mipsel --disable-machine-licm -mcpu=mips32r6 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32-ANY -check-prefix=HAS-SEB-SEH -check-prefix=CHECK-EL -check-prefix=NOT-MICROMIPS
; RUN: llc -march=mips64el --disable-machine-licm -mcpu=mips4 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS64-ANY -check-prefix=NO-SEB-SEH -check-prefix=CHECK-EL -check-prefix=NOT-MICROMIPS
; RUN: llc -march=mips64el --disable-machine-licm -mcpu=mips64 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS64-ANY -check-prefix=NO-SEB-SEH -check-prefix=CHECK-EL -check-prefix=NOT-MICROMIPS
; RUN: llc -march=mips64el --disable-machine-licm -mcpu=mips64r2 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS64-ANY -check-prefix=HAS-SEB-SEH -check-prefix=CHECK-EL -check-prefix=NOT-MICROMIPS
; RUN: llc -march=mips64el --disable-machine-licm -mcpu=mips64r6 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS64-ANY -check-prefix=HAS-SEB-SEH -check-prefix=CHECK-EL -check-prefix=NOT-MICROMIPS
; RUN: llc -march=mipsel --disable-machine-licm -mcpu=mips32r2 -mattr=micromips < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32-ANY -check-prefix=HAS-SEB-SEH -check-prefix=CHECK-EL -check-prefix=MICROMIPS
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; Keep one big-endian check so that we don't reduce testing, but don't add more
; since endianness doesn't affect the body of the atomic operations.
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; RUN: llc -march=mips --disable-machine-licm -mcpu=mips32 < %s | FileCheck %s -check-prefix=ALL -check-prefix=MIPS32-ANY -check-prefix=NO-SEB-SEH -check-prefix=CHECK-EB -check-prefix=NOT-MICROMIPS
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
@x = common global i32 0, align 4
[mips] Promote i32 arguments to i64 for the N32/N64 ABI and fix <64-bit structs... Summary: ... and after all that refactoring, it's possible to distinguish softfloat floating point values from integers so this patch no longer breaks softfloat to do it. Remove direct handling of i32's in the N32/N64 ABI by promoting them to i64. This more closely reflects the ABI documentation and also fixes problems with stack arguments on big-endian targets. We now rely on signext/zeroext annotations (already generated by clang) and the Assert[SZ]ext nodes to avoid the introduction of unnecessary sign/zero extends. It was not possible to convert three tests to use signext/zeroext. These tests are bswap.ll, ctlz-v.ll, ctlz-v.ll. It's not possible to put signext on a vector type so we just accept the sign extends here for now. These tests don't pass the vectors the same way clang does (clang puts multiple elements in the same argument, these map 1 element to 1 argument) so we don't need to worry too much about it. With this patch, all known N32/N64 bugs should be fixed and we now pass the first 10,000 tests generated by ABITest.py. Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D6117 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221534 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-07 16:54:21 +00:00
define i32 @AtomicLoadAdd32(i32 signext %incr) nounwind {
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
entry:
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw add i32* @x, i32 %incr monotonic
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i32 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicLoadAdd32:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(x)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(x)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R1:[0-9]+]], 0($[[R0]])
; ALL: addu $[[R2:[0-9]+]], $[[R1]], $4
; ALL: sc $[[R2]], 0($[[R0]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R2]], $[[BB0]]
; MICROMIPS: beqzc $[[R2]], $[[BB0]]
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
[mips] Promote i32 arguments to i64 for the N32/N64 ABI and fix <64-bit structs... Summary: ... and after all that refactoring, it's possible to distinguish softfloat floating point values from integers so this patch no longer breaks softfloat to do it. Remove direct handling of i32's in the N32/N64 ABI by promoting them to i64. This more closely reflects the ABI documentation and also fixes problems with stack arguments on big-endian targets. We now rely on signext/zeroext annotations (already generated by clang) and the Assert[SZ]ext nodes to avoid the introduction of unnecessary sign/zero extends. It was not possible to convert three tests to use signext/zeroext. These tests are bswap.ll, ctlz-v.ll, ctlz-v.ll. It's not possible to put signext on a vector type so we just accept the sign extends here for now. These tests don't pass the vectors the same way clang does (clang puts multiple elements in the same argument, these map 1 element to 1 argument) so we don't need to worry too much about it. With this patch, all known N32/N64 bugs should be fixed and we now pass the first 10,000 tests generated by ABITest.py. Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D6117 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221534 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-07 16:54:21 +00:00
define i32 @AtomicLoadNand32(i32 signext %incr) nounwind {
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
entry:
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw nand i32* @x, i32 %incr monotonic
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i32 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicLoadNand32:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(x)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(x)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R1:[0-9]+]], 0($[[R0]])
; ALL: and $[[R3:[0-9]+]], $[[R1]], $4
; ALL: nor $[[R2:[0-9]+]], $zero, $[[R3]]
; ALL: sc $[[R2]], 0($[[R0]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R2]], $[[BB0]]
; MICROMIPS: beqzc $[[R2]], $[[BB0]]
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
[mips] Promote i32 arguments to i64 for the N32/N64 ABI and fix <64-bit structs... Summary: ... and after all that refactoring, it's possible to distinguish softfloat floating point values from integers so this patch no longer breaks softfloat to do it. Remove direct handling of i32's in the N32/N64 ABI by promoting them to i64. This more closely reflects the ABI documentation and also fixes problems with stack arguments on big-endian targets. We now rely on signext/zeroext annotations (already generated by clang) and the Assert[SZ]ext nodes to avoid the introduction of unnecessary sign/zero extends. It was not possible to convert three tests to use signext/zeroext. These tests are bswap.ll, ctlz-v.ll, ctlz-v.ll. It's not possible to put signext on a vector type so we just accept the sign extends here for now. These tests don't pass the vectors the same way clang does (clang puts multiple elements in the same argument, these map 1 element to 1 argument) so we don't need to worry too much about it. With this patch, all known N32/N64 bugs should be fixed and we now pass the first 10,000 tests generated by ABITest.py. Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D6117 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221534 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-07 16:54:21 +00:00
define i32 @AtomicSwap32(i32 signext %newval) nounwind {
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
entry:
Set mayLoad or mayStore flags for SC and LL in order to prevent LICM from moving them out of the loop. Previously, stores and loads to a stack frame object were inserted to accomplish this. Remove the code that was needed to do this. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135415 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-18 18:52:12 +00:00
%newval.addr = alloca i32, align 4
store i32 %newval, i32* %newval.addr, align 4
%tmp = load i32* %newval.addr, align 4
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw xchg i32* @x, i32 %tmp monotonic
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i32 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicSwap32:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(x)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(x)
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll ${{[0-9]+}}, 0($[[R0]])
; ALL: sc $[[R2:[0-9]+]], 0($[[R0]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R2]], $[[BB0]]
; MICROMIPS: beqzc $[[R2]], $[[BB0]]
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
[mips] Promote i32 arguments to i64 for the N32/N64 ABI and fix <64-bit structs... Summary: ... and after all that refactoring, it's possible to distinguish softfloat floating point values from integers so this patch no longer breaks softfloat to do it. Remove direct handling of i32's in the N32/N64 ABI by promoting them to i64. This more closely reflects the ABI documentation and also fixes problems with stack arguments on big-endian targets. We now rely on signext/zeroext annotations (already generated by clang) and the Assert[SZ]ext nodes to avoid the introduction of unnecessary sign/zero extends. It was not possible to convert three tests to use signext/zeroext. These tests are bswap.ll, ctlz-v.ll, ctlz-v.ll. It's not possible to put signext on a vector type so we just accept the sign extends here for now. These tests don't pass the vectors the same way clang does (clang puts multiple elements in the same argument, these map 1 element to 1 argument) so we don't need to worry too much about it. With this patch, all known N32/N64 bugs should be fixed and we now pass the first 10,000 tests generated by ABITest.py. Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D6117 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221534 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-07 16:54:21 +00:00
define i32 @AtomicCmpSwap32(i32 signext %oldval, i32 signext %newval) nounwind {
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
entry:
Set mayLoad or mayStore flags for SC and LL in order to prevent LICM from moving them out of the loop. Previously, stores and loads to a stack frame object were inserted to accomplish this. Remove the code that was needed to do this. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135415 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-18 18:52:12 +00:00
%newval.addr = alloca i32, align 4
store i32 %newval, i32* %newval.addr, align 4
%tmp = load i32* %newval.addr, align 4
IR: add a second ordering operand to cmpxhg for failure The syntax for "cmpxchg" should now look something like: cmpxchg i32* %addr, i32 42, i32 3 acquire monotonic where the second ordering argument gives the required semantics in the case that no exchange takes place. It should be no stronger than the first ordering constraint and cannot be either "release" or "acq_rel" (since no store will have taken place). rdar://problem/15996804 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@203559 91177308-0d34-0410-b5e6-96231b3b80d8
2014-03-11 10:48:52 +00:00
%0 = cmpxchg i32* @x, i32 %oldval, i32 %tmp monotonic monotonic
IR: add "cmpxchg weak" variant to support permitted failure. This commit adds a weak variant of the cmpxchg operation, as described in C++11. A cmpxchg instruction with this modifier is permitted to fail to store, even if the comparison indicated it should. As a result, cmpxchg instructions must return a flag indicating success in addition to their original iN value loaded. Thus, for uniformity *all* cmpxchg instructions now return "{ iN, i1 }". The second flag is 1 when the store succeeded. At the DAG level, a new ATOMIC_CMP_SWAP_WITH_SUCCESS node has been added as the natural representation for the new cmpxchg instructions. It is a strong cmpxchg. By default this gets Expanded to the existing ATOMIC_CMP_SWAP during Legalization, so existing backends should see no change in behaviour. If they wish to deal with the enhanced node instead, they can call setOperationAction on it. Beware: as a node with 2 results, it cannot be selected from TableGen. Currently, no use is made of the extra information provided in this patch. Test updates are almost entirely adapting the input IR to the new scheme. Summary for out of tree users: ------------------------------ + Legacy Bitcode files are upgraded during read. + Legacy assembly IR files will be invalid. + Front-ends must adapt to different type for "cmpxchg". + Backends should be unaffected by default. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210903 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-13 14:24:07 +00:00
%1 = extractvalue { i32, i1 } %0, 0
ret i32 %1
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicCmpSwap32:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(x)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(x)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $2, 0($[[R0]])
; ALL: bne $2, $4, $[[BB1:[A-Z_0-9]+]]
; ALL: sc $[[R2:[0-9]+]], 0($[[R0]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R2]], $[[BB0]]
; MICROMIPS: beqzc $[[R2]], $[[BB0]]
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: $[[BB1]]:
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
@y = common global i8 0, align 1
define signext i8 @AtomicLoadAdd8(i8 signext %incr) nounwind {
entry:
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw add i8* @y, i8 %incr monotonic
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i8 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicLoadAdd8:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(y)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(y)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: addiu $[[R1:[0-9]+]], $zero, -4
; ALL: and $[[R2:[0-9]+]], $[[R0]], $[[R1]]
; ALL: andi $[[R3:[0-9]+]], $[[R0]], 3
; CHECK-EB: xori $[[R4:[0-9]+]], $[[R3]], 3
; CHECK-EB: sll $[[R5:[0-9]+]], $[[R4]], 3
; CHECK-EL: sll $[[R5:[0-9]+]], $[[R3]], 3
; ALL: ori $[[R6:[0-9]+]], $zero, 255
; ALL: sllv $[[R7:[0-9]+]], $[[R6]], $[[R5]]
; ALL: nor $[[R8:[0-9]+]], $zero, $[[R7]]
; ALL: sllv $[[R9:[0-9]+]], $4, $[[R5]]
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R10:[0-9]+]], 0($[[R2]])
; ALL: addu $[[R11:[0-9]+]], $[[R10]], $[[R9]]
; ALL: and $[[R12:[0-9]+]], $[[R11]], $[[R7]]
; ALL: and $[[R13:[0-9]+]], $[[R10]], $[[R8]]
; ALL: or $[[R14:[0-9]+]], $[[R13]], $[[R12]]
; ALL: sc $[[R14]], 0($[[R2]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R14]], $[[BB0]]
; MICROMIPS: beqzc $[[R14]], $[[BB0]]
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: and $[[R15:[0-9]+]], $[[R10]], $[[R7]]
; ALL: srlv $[[R16:[0-9]+]], $[[R15]], $[[R5]]
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; NO-SEB-SEH: sll $[[R17:[0-9]+]], $[[R16]], 24
; NO-SEB-SEH: sra $2, $[[R17]], 24
; HAS-SEB-SEH: seb $2, $[[R16]]
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
define signext i8 @AtomicLoadSub8(i8 signext %incr) nounwind {
entry:
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw sub i8* @y, i8 %incr monotonic
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i8 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicLoadSub8:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(y)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(y)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: addiu $[[R1:[0-9]+]], $zero, -4
; ALL: and $[[R2:[0-9]+]], $[[R0]], $[[R1]]
; ALL: andi $[[R3:[0-9]+]], $[[R0]], 3
; CHECK-EL: sll $[[R5:[0-9]+]], $[[R3]], 3
[mips] Big-endian code generation for atomic instructions. Patch by Jyun-Yan You. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@182984 91177308-0d34-0410-b5e6-96231b3b80d8
2013-05-31 03:25:44 +00:00
; CHECK-EB: xori $[[R4:[0-9]+]], $[[R3]], 3
; CHECK-EB: sll $[[R5:[0-9]+]], $[[R4]], 3
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: ori $[[R6:[0-9]+]], $zero, 255
; ALL: sllv $[[R7:[0-9]+]], $[[R6]], $[[R5]]
; ALL: nor $[[R8:[0-9]+]], $zero, $[[R7]]
; ALL: sllv $[[R9:[0-9]+]], $4, $[[R5]]
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R10:[0-9]+]], 0($[[R2]])
; ALL: subu $[[R11:[0-9]+]], $[[R10]], $[[R9]]
; ALL: and $[[R12:[0-9]+]], $[[R11]], $[[R7]]
; ALL: and $[[R13:[0-9]+]], $[[R10]], $[[R8]]
; ALL: or $[[R14:[0-9]+]], $[[R13]], $[[R12]]
; ALL: sc $[[R14]], 0($[[R2]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R14]], $[[BB0]]
; MICROMIPS: beqzc $[[R14]], $[[BB0]]
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: and $[[R15:[0-9]+]], $[[R10]], $[[R7]]
; ALL: srlv $[[R16:[0-9]+]], $[[R15]], $[[R5]]
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; NO-SEB-SEH: sll $[[R17:[0-9]+]], $[[R16]], 24
; NO-SEB-SEH: sra $2, $[[R17]], 24
; HAS-SEB-SEH:seb $2, $[[R16]]
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
define signext i8 @AtomicLoadNand8(i8 signext %incr) nounwind {
entry:
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw nand i8* @y, i8 %incr monotonic
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i8 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicLoadNand8:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(y)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(y)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: addiu $[[R1:[0-9]+]], $zero, -4
; ALL: and $[[R2:[0-9]+]], $[[R0]], $[[R1]]
; ALL: andi $[[R3:[0-9]+]], $[[R0]], 3
; CHECK-EL: sll $[[R5:[0-9]+]], $[[R3]], 3
; CHECK-EB: xori $[[R4:[0-9]+]], $[[R3]], 3
; CHECK-EB: sll $[[R5:[0-9]+]], $[[R4]], 3
; ALL: ori $[[R6:[0-9]+]], $zero, 255
; ALL: sllv $[[R7:[0-9]+]], $[[R6]], $[[R5]]
; ALL: nor $[[R8:[0-9]+]], $zero, $[[R7]]
; ALL: sllv $[[R9:[0-9]+]], $4, $[[R5]]
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R10:[0-9]+]], 0($[[R2]])
; ALL: and $[[R18:[0-9]+]], $[[R10]], $[[R9]]
; ALL: nor $[[R11:[0-9]+]], $zero, $[[R18]]
; ALL: and $[[R12:[0-9]+]], $[[R11]], $[[R7]]
; ALL: and $[[R13:[0-9]+]], $[[R10]], $[[R8]]
; ALL: or $[[R14:[0-9]+]], $[[R13]], $[[R12]]
; ALL: sc $[[R14]], 0($[[R2]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R14]], $[[BB0]]
; MICROMIPS: beqzc $[[R14]], $[[BB0]]
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: and $[[R15:[0-9]+]], $[[R10]], $[[R7]]
; ALL: srlv $[[R16:[0-9]+]], $[[R15]], $[[R5]]
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; NO-SEB-SEH: sll $[[R17:[0-9]+]], $[[R16]], 24
; NO-SEB-SEH: sra $2, $[[R17]], 24
; HAS-SEB-SEH: seb $2, $[[R16]]
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
Set mayLoad or mayStore flags for SC and LL in order to prevent LICM from moving them out of the loop. Previously, stores and loads to a stack frame object were inserted to accomplish this. Remove the code that was needed to do this. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135415 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-18 18:52:12 +00:00
define signext i8 @AtomicSwap8(i8 signext %newval) nounwind {
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
entry:
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw xchg i8* @y, i8 %newval monotonic
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i8 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicSwap8:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(y)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(y)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: addiu $[[R1:[0-9]+]], $zero, -4
; ALL: and $[[R2:[0-9]+]], $[[R0]], $[[R1]]
; ALL: andi $[[R3:[0-9]+]], $[[R0]], 3
; CHECK-EL: sll $[[R5:[0-9]+]], $[[R3]], 3
; CHECK-EB: xori $[[R4:[0-9]+]], $[[R3]], 3
; CHECK-EB: sll $[[R5:[0-9]+]], $[[R4]], 3
; ALL: ori $[[R6:[0-9]+]], $zero, 255
; ALL: sllv $[[R7:[0-9]+]], $[[R6]], $[[R5]]
; ALL: nor $[[R8:[0-9]+]], $zero, $[[R7]]
; ALL: sllv $[[R9:[0-9]+]], $4, $[[R5]]
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R10:[0-9]+]], 0($[[R2]])
; ALL: and $[[R18:[0-9]+]], $[[R9]], $[[R7]]
; ALL: and $[[R13:[0-9]+]], $[[R10]], $[[R8]]
; ALL: or $[[R14:[0-9]+]], $[[R13]], $[[R18]]
; ALL: sc $[[R14]], 0($[[R2]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R14]], $[[BB0]]
; MICROMIPS: beqzc $[[R14]], $[[BB0]]
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: and $[[R15:[0-9]+]], $[[R10]], $[[R7]]
; ALL: srlv $[[R16:[0-9]+]], $[[R15]], $[[R5]]
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; NO-SEB-SEH: sll $[[R17:[0-9]+]], $[[R16]], 24
; NO-SEB-SEH: sra $2, $[[R17]], 24
; HAS-SEB-SEH: seb $2, $[[R16]]
Replace the result usages while legalizing cmpxchg. We should update the usages to all of the results; otherwise, we might get assertion failure or SEGV during the type legalization of ATOMIC_CMP_SWAP_WITH_SUCCESS with two or more illegal types. For example, in the following sequence, both i8 and i1 might be illegal in some target, e.g. armv5, mipsel, mips64el, %0 = cmpxchg i8* %ptr, i8 %desire, i8 %new monotonic monotonic %1 = extractvalue { i8, i1 } %0, 1 Since both i8 and i1 should be legalized, the corresponding ATOMIC_CMP_SWAP_WITH_SUCCESS dag will be checked/replaced/updated twice. If we don't update the usage to *ALL* of the results in the first round, the DAG for extractvalue might be processed earlier. The GetPromotedInteger() will result in assertion failure, because its operand (i.e. the success bit of cmpxchg) is not promoted beforehand. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213569 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-21 17:33:44 +00:00
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
define signext i8 @AtomicCmpSwap8(i8 signext %oldval, i8 signext %newval) nounwind {
entry:
IR: add "cmpxchg weak" variant to support permitted failure. This commit adds a weak variant of the cmpxchg operation, as described in C++11. A cmpxchg instruction with this modifier is permitted to fail to store, even if the comparison indicated it should. As a result, cmpxchg instructions must return a flag indicating success in addition to their original iN value loaded. Thus, for uniformity *all* cmpxchg instructions now return "{ iN, i1 }". The second flag is 1 when the store succeeded. At the DAG level, a new ATOMIC_CMP_SWAP_WITH_SUCCESS node has been added as the natural representation for the new cmpxchg instructions. It is a strong cmpxchg. By default this gets Expanded to the existing ATOMIC_CMP_SWAP during Legalization, so existing backends should see no change in behaviour. If they wish to deal with the enhanced node instead, they can call setOperationAction on it. Beware: as a node with 2 results, it cannot be selected from TableGen. Currently, no use is made of the extra information provided in this patch. Test updates are almost entirely adapting the input IR to the new scheme. Summary for out of tree users: ------------------------------ + Legacy Bitcode files are upgraded during read. + Legacy assembly IR files will be invalid. + Front-ends must adapt to different type for "cmpxchg". + Backends should be unaffected by default. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210903 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-13 14:24:07 +00:00
%pair0 = cmpxchg i8* @y, i8 %oldval, i8 %newval monotonic monotonic
%0 = extractvalue { i8, i1 } %pair0, 0
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
ret i8 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: AtomicCmpSwap8:
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(y)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(y)(
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: addiu $[[R1:[0-9]+]], $zero, -4
; ALL: and $[[R2:[0-9]+]], $[[R0]], $[[R1]]
; ALL: andi $[[R3:[0-9]+]], $[[R0]], 3
; CHECK-EL: sll $[[R5:[0-9]+]], $[[R3]], 3
; CHECK-EB: xori $[[R4:[0-9]+]], $[[R3]], 3
; CHECK-EB: sll $[[R5:[0-9]+]], $[[R4]], 3
; ALL: ori $[[R6:[0-9]+]], $zero, 255
; ALL: sllv $[[R7:[0-9]+]], $[[R6]], $[[R5]]
; ALL: nor $[[R8:[0-9]+]], $zero, $[[R7]]
; ALL: andi $[[R9:[0-9]+]], $4, 255
; ALL: sllv $[[R10:[0-9]+]], $[[R9]], $[[R5]]
; ALL: andi $[[R11:[0-9]+]], $5, 255
; ALL: sllv $[[R12:[0-9]+]], $[[R11]], $[[R5]]
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R13:[0-9]+]], 0($[[R2]])
; ALL: and $[[R14:[0-9]+]], $[[R13]], $[[R7]]
; ALL: bne $[[R14]], $[[R10]], $[[BB1:[A-Z_0-9]+]]
; ALL: and $[[R15:[0-9]+]], $[[R13]], $[[R8]]
; ALL: or $[[R16:[0-9]+]], $[[R15]], $[[R12]]
; ALL: sc $[[R16]], 0($[[R2]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R16]], $[[BB0]]
; MICROMIPS: beqzc $[[R16]], $[[BB0]]
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: $[[BB1]]:
; ALL: srlv $[[R17:[0-9]+]], $[[R14]], $[[R5]]
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; NO-SEB-SEH: sll $[[R18:[0-9]+]], $[[R17]], 24
; NO-SEB-SEH: sra $2, $[[R18]], 24
; HAS-SEB-SEH: seb $2, $[[R17]]
This patch implements atomic intrinsics atomic.load.add (sub,and,or,xor, nand), atomic.swap and atomic.cmp.swap, all in i8, i16 and i32 versions. The intrinsics are implemented by creating pseudo-instructions, which are then expanded in the method MipsTargetLowering::EmitInstrWithCustomInserter. Patch by Sasa Stankovic. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@132323 91177308-0d34-0410-b5e6-96231b3b80d8
2011-05-31 02:54:07 +00:00
}
Lower memory barriers to sync instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135537 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-19 23:30:50 +00:00
[mips] Promote i32 arguments to i64 for the N32/N64 ABI and fix <64-bit structs... Summary: ... and after all that refactoring, it's possible to distinguish softfloat floating point values from integers so this patch no longer breaks softfloat to do it. Remove direct handling of i32's in the N32/N64 ABI by promoting them to i64. This more closely reflects the ABI documentation and also fixes problems with stack arguments on big-endian targets. We now rely on signext/zeroext annotations (already generated by clang) and the Assert[SZ]ext nodes to avoid the introduction of unnecessary sign/zero extends. It was not possible to convert three tests to use signext/zeroext. These tests are bswap.ll, ctlz-v.ll, ctlz-v.ll. It's not possible to put signext on a vector type so we just accept the sign extends here for now. These tests don't pass the vectors the same way clang does (clang puts multiple elements in the same argument, these map 1 element to 1 argument) so we don't need to worry too much about it. With this patch, all known N32/N64 bugs should be fixed and we now pass the first 10,000 tests generated by ABITest.py. Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D6117 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221534 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-07 16:54:21 +00:00
define i1 @AtomicCmpSwapRes8(i8* %ptr, i8 signext %oldval, i8 signext %newval) nounwind {
Replace the result usages while legalizing cmpxchg. We should update the usages to all of the results; otherwise, we might get assertion failure or SEGV during the type legalization of ATOMIC_CMP_SWAP_WITH_SUCCESS with two or more illegal types. For example, in the following sequence, both i8 and i1 might be illegal in some target, e.g. armv5, mipsel, mips64el, %0 = cmpxchg i8* %ptr, i8 %desire, i8 %new monotonic monotonic %1 = extractvalue { i8, i1 } %0, 1 Since both i8 and i1 should be legalized, the corresponding ATOMIC_CMP_SWAP_WITH_SUCCESS dag will be checked/replaced/updated twice. If we don't update the usage to *ALL* of the results in the first round, the DAG for extractvalue might be processed earlier. The GetPromotedInteger() will result in assertion failure, because its operand (i.e. the success bit of cmpxchg) is not promoted beforehand. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213569 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-21 17:33:44 +00:00
entry:
%0 = cmpxchg i8* %ptr, i8 %oldval, i8 %newval monotonic monotonic
%1 = extractvalue { i8, i1 } %0, 1
ret i1 %1
; ALL-LABEL: AtomicCmpSwapRes8
; ALL: addiu $[[R1:[0-9]+]], $zero, -4
; ALL: and $[[R2:[0-9]+]], $4, $[[R1]]
; ALL: andi $[[R3:[0-9]+]], $4, 3
; CHECK-EL: sll $[[R5:[0-9]+]], $[[R3]], 3
; CHECK-EB: xori $[[R4:[0-9]+]], $[[R3]], 3
; CHECK-EB: sll $[[R5:[0-9]+]], $[[R4]], 3
; ALL: ori $[[R6:[0-9]+]], $zero, 255
; ALL: sllv $[[R7:[0-9]+]], $[[R6]], $[[R5]]
; ALL: nor $[[R8:[0-9]+]], $zero, $[[R7]]
; ALL: andi $[[R9:[0-9]+]], $5, 255
; ALL: sllv $[[R10:[0-9]+]], $[[R9]], $[[R5]]
; ALL: andi $[[R11:[0-9]+]], $6, 255
; ALL: sllv $[[R12:[0-9]+]], $[[R11]], $[[R5]]
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R13:[0-9]+]], 0($[[R2]])
; ALL: and $[[R14:[0-9]+]], $[[R13]], $[[R7]]
; ALL: bne $[[R14]], $[[R10]], $[[BB1:[A-Z_0-9]+]]
; ALL: and $[[R15:[0-9]+]], $[[R13]], $[[R8]]
; ALL: or $[[R16:[0-9]+]], $[[R15]], $[[R12]]
; ALL: sc $[[R16]], 0($[[R2]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R16]], $[[BB0]]
; MICROMIPS: beqzc $[[R16]], $[[BB0]]
Replace the result usages while legalizing cmpxchg. We should update the usages to all of the results; otherwise, we might get assertion failure or SEGV during the type legalization of ATOMIC_CMP_SWAP_WITH_SUCCESS with two or more illegal types. For example, in the following sequence, both i8 and i1 might be illegal in some target, e.g. armv5, mipsel, mips64el, %0 = cmpxchg i8* %ptr, i8 %desire, i8 %new monotonic monotonic %1 = extractvalue { i8, i1 } %0, 1 Since both i8 and i1 should be legalized, the corresponding ATOMIC_CMP_SWAP_WITH_SUCCESS dag will be checked/replaced/updated twice. If we don't update the usage to *ALL* of the results in the first round, the DAG for extractvalue might be processed earlier. The GetPromotedInteger() will result in assertion failure, because its operand (i.e. the success bit of cmpxchg) is not promoted beforehand. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@213569 91177308-0d34-0410-b5e6-96231b3b80d8
2014-07-21 17:33:44 +00:00
; ALL: $[[BB1]]:
; ALL: srlv $[[R17:[0-9]+]], $[[R14]], $[[R5]]
; NO-SEB-SEH: sll $[[R18:[0-9]+]], $[[R17]], 24
; NO-SEB-SEH: sra $[[R19:[0-9]+]], $[[R18]], 24
; HAS-SEB-SEH: seb $[[R19:[0-9]+]], $[[R17]]
; ALL: xor $[[R20:[0-9]+]], $[[R19]], $5
; ALL: sltiu $2, $[[R20]], 1
}
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; Check one i16 so that we cover the seh sign extend
@z = common global i16 0, align 1
define signext i16 @AtomicLoadAdd16(i16 signext %incr) nounwind {
entry:
%0 = atomicrmw add i16* @z, i16 %incr monotonic
ret i16 %0
; ALL-LABEL: AtomicLoadAdd16:
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(z)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(z)(
; ALL: addiu $[[R1:[0-9]+]], $zero, -4
; ALL: and $[[R2:[0-9]+]], $[[R0]], $[[R1]]
; ALL: andi $[[R3:[0-9]+]], $[[R0]], 3
; CHECK-EB: xori $[[R4:[0-9]+]], $[[R3]], 2
; CHECK-EB: sll $[[R5:[0-9]+]], $[[R4]], 3
; CHECK-EL: sll $[[R5:[0-9]+]], $[[R3]], 3
; ALL: ori $[[R6:[0-9]+]], $zero, 65535
; ALL: sllv $[[R7:[0-9]+]], $[[R6]], $[[R5]]
; ALL: nor $[[R8:[0-9]+]], $zero, $[[R7]]
; ALL: sllv $[[R9:[0-9]+]], $4, $[[R5]]
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R10:[0-9]+]], 0($[[R2]])
; ALL: addu $[[R11:[0-9]+]], $[[R10]], $[[R9]]
; ALL: and $[[R12:[0-9]+]], $[[R11]], $[[R7]]
; ALL: and $[[R13:[0-9]+]], $[[R10]], $[[R8]]
; ALL: or $[[R14:[0-9]+]], $[[R13]], $[[R12]]
; ALL: sc $[[R14]], 0($[[R2]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R14]], $[[BB0]]
; MICROMIPS: beqzc $[[R14]], $[[BB0]]
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; ALL: and $[[R15:[0-9]+]], $[[R10]], $[[R7]]
; ALL: srlv $[[R16:[0-9]+]], $[[R15]], $[[R5]]
; NO-SEB-SEH: sll $[[R17:[0-9]+]], $[[R16]], 16
; NO-SEB-SEH: sra $2, $[[R17]], 16
; MIPS32R2: seh $2, $[[R16]]
}
Lower memory barriers to sync instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135537 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-19 23:30:50 +00:00
@countsint = common global i32 0, align 4
[mips] Promote i32 arguments to i64 for the N32/N64 ABI and fix <64-bit structs... Summary: ... and after all that refactoring, it's possible to distinguish softfloat floating point values from integers so this patch no longer breaks softfloat to do it. Remove direct handling of i32's in the N32/N64 ABI by promoting them to i64. This more closely reflects the ABI documentation and also fixes problems with stack arguments on big-endian targets. We now rely on signext/zeroext annotations (already generated by clang) and the Assert[SZ]ext nodes to avoid the introduction of unnecessary sign/zero extends. It was not possible to convert three tests to use signext/zeroext. These tests are bswap.ll, ctlz-v.ll, ctlz-v.ll. It's not possible to put signext on a vector type so we just accept the sign extends here for now. These tests don't pass the vectors the same way clang does (clang puts multiple elements in the same argument, these map 1 element to 1 argument) so we don't need to worry too much about it. With this patch, all known N32/N64 bugs should be fixed and we now pass the first 10,000 tests generated by ABITest.py. Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D6117 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221534 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-07 16:54:21 +00:00
define i32 @CheckSync(i32 signext %v) nounwind noinline {
Lower memory barriers to sync instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135537 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-19 23:30:50 +00:00
entry:
Convert more tests over to the new atomic instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@140559 91177308-0d34-0410-b5e6-96231b3b80d8
2011-09-26 20:27:49 +00:00
%0 = atomicrmw add i32* @countsint, i32 %v seq_cst
Lower memory barriers to sync instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135537 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-19 23:30:50 +00:00
ret i32 %0
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL-LABEL: CheckSync:
[mips] SYNC $stype instruction was added in Mips32 but SYNC with an implied operand ($stype = 0) is valid since Mips2. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211185 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-18 17:10:30 +00:00
; ALL: sync
[mips] Merge most of the big/little endian checks in atomic.ll Summary: There is very little difference between the big and little endian cases in test/CodeGen/Mips/atomic.ll. Merge them together using multiple FileCheck prefixes. Depends on D4117 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4118 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211013 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 10:25:17 +00:00
; ALL: ll
; ALL: sc
; ALL: beq
[mips] SYNC $stype instruction was added in Mips32 but SYNC with an implied operand ($stype = 0) is valid since Mips2. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211185 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-18 17:10:30 +00:00
; ALL: sync
Lower memory barriers to sync instructions. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@135537 91177308-0d34-0410-b5e6-96231b3b80d8
2011-07-19 23:30:50 +00:00
}
Do not replace operands of pseudo instructions with register $zero. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156663 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-11 23:22:18 +00:00
; make sure that this assertion in
; TwoAddressInstructionPass::TryInstructionTransform does not fail:
;
; line 1203: assert(TargetRegisterInfo::isVirtualRegister(regB) &&
;
; it failed when MipsDAGToDAGISel::ReplaceUsesWithZeroReg replaced an
; operand of an atomic instruction with register $zero.
@a = external global i32
define i32 @zeroreg() nounwind {
entry:
IR: add "cmpxchg weak" variant to support permitted failure. This commit adds a weak variant of the cmpxchg operation, as described in C++11. A cmpxchg instruction with this modifier is permitted to fail to store, even if the comparison indicated it should. As a result, cmpxchg instructions must return a flag indicating success in addition to their original iN value loaded. Thus, for uniformity *all* cmpxchg instructions now return "{ iN, i1 }". The second flag is 1 when the store succeeded. At the DAG level, a new ATOMIC_CMP_SWAP_WITH_SUCCESS node has been added as the natural representation for the new cmpxchg instructions. It is a strong cmpxchg. By default this gets Expanded to the existing ATOMIC_CMP_SWAP during Legalization, so existing backends should see no change in behaviour. If they wish to deal with the enhanced node instead, they can call setOperationAction on it. Beware: as a node with 2 results, it cannot be selected from TableGen. Currently, no use is made of the extra information provided in this patch. Test updates are almost entirely adapting the input IR to the new scheme. Summary for out of tree users: ------------------------------ + Legacy Bitcode files are upgraded during read. + Legacy assembly IR files will be invalid. + Front-ends must adapt to different type for "cmpxchg". + Backends should be unaffected by default. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@210903 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-13 14:24:07 +00:00
%pair0 = cmpxchg i32* @a, i32 1, i32 0 seq_cst seq_cst
%0 = extractvalue { i32, i1 } %pair0, 0
Do not replace operands of pseudo instructions with register $zero. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@156663 91177308-0d34-0410-b5e6-96231b3b80d8
2012-05-11 23:22:18 +00:00
%1 = icmp eq i32 %0, 1
%conv = zext i1 %1 to i32
ret i32 %conv
}
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
; Check that MIPS32R6 has the correct offset range.
; FIXME: At the moment, we don't seem to do addr+offset for any atomic load/store.
[mips] Promote i32 arguments to i64 for the N32/N64 ABI and fix <64-bit structs... Summary: ... and after all that refactoring, it's possible to distinguish softfloat floating point values from integers so this patch no longer breaks softfloat to do it. Remove direct handling of i32's in the N32/N64 ABI by promoting them to i64. This more closely reflects the ABI documentation and also fixes problems with stack arguments on big-endian targets. We now rely on signext/zeroext annotations (already generated by clang) and the Assert[SZ]ext nodes to avoid the introduction of unnecessary sign/zero extends. It was not possible to convert three tests to use signext/zeroext. These tests are bswap.ll, ctlz-v.ll, ctlz-v.ll. It's not possible to put signext on a vector type so we just accept the sign extends here for now. These tests don't pass the vectors the same way clang does (clang puts multiple elements in the same argument, these map 1 element to 1 argument) so we don't need to worry too much about it. With this patch, all known N32/N64 bugs should be fixed and we now pass the first 10,000 tests generated by ABITest.py. Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D6117 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@221534 91177308-0d34-0410-b5e6-96231b3b80d8
2014-11-07 16:54:21 +00:00
define i32 @AtomicLoadAdd32_OffGt9Bit(i32 signext %incr) nounwind {
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
entry:
%0 = atomicrmw add i32* getelementptr(i32* @x, i32 256), i32 %incr monotonic
ret i32 %0
; ALL-LABEL: AtomicLoadAdd32_OffGt9Bit:
; MIPS32-ANY: lw $[[R0:[0-9]+]], %got(x)
; MIPS64-ANY: ld $[[R0:[0-9]+]], %got_disp(x)(
; ALL: addiu $[[PTR:[0-9]+]], $[[R0]], 1024
; ALL: $[[BB0:[A-Z_0-9]+]]:
; ALL: ll $[[R1:[0-9]+]], 0($[[PTR]])
; ALL: addu $[[R2:[0-9]+]], $[[R1]], $4
; ALL: sc $[[R2]], 0($[[PTR]])
[mips][microMIPS] Fix bugs related to atomic SC/LL instructions Fix bugs related to atomic microMIPS SC/LL instructions: While expanding atomic operations the mips32r2 encoding was emitted instead of microMIPS. Differential Revision: http://reviews.llvm.org/D6659 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@224524 91177308-0d34-0410-b5e6-96231b3b80d8
2014-12-18 16:39:29 +00:00
; NOT-MICROMIPS: beqz $[[R2]], $[[BB0]]
; MICROMIPS: beqzc $[[R2]], $[[BB0]]
[mips][mips64r6] ll, sc, lld, and scd are re-encoded on MIPS32r6/MIPS64r6. Summary: The linked-load, store-conditional operations have been re-encoded such that have a 9-bit offset instead of the 16-bit offset they have prior to MIPS32r6/MIPS64r6. While implementing this, I noticed that the atomic load/store pseudos always emit a sign extension using sll and sra. I have improved this to use seb/seh when they are available (MIPS32r2/MIPS64r2 and above). Depends on D4118 Reviewers: jkolek, zoran.jovanovic, vmedic Reviewed By: vmedic Differential Revision: http://reviews.llvm.org/D4119 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211018 91177308-0d34-0410-b5e6-96231b3b80d8
2014-06-16 13:13:03 +00:00
}
Reference in New Issue Copy Permalink