6502/llvm-6502
1
0
Fork 0
mirror of https://github.com/c64scene-ar/llvm-6502.git synced 2024-12-28 19:31:58 +00:00
Code Issues Projects Releases Wiki Activity
fa2e88da1c
llvm-6502/test/CodeGen/X86/atomic6432.ll

209 lines
4.5 KiB
LLVM
Raw Normal View History

Fix PR15355 - Clear 'mayStore' flag when loading from the atomic variable before the spin loop - Clear kill flag from one use to multiple use in registers forming the address to that atomic variable - don't use a physical register as live-in register in BB (neither entry nor landing pad.) by copying it into virtual register (patch by Cameron Zwarich) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@176538 91177308-0d34-0410-b5e6-96231b3b80d8
2013-03-06 00:17:04 +00:00
; RUN: llc < %s -O0 -march=x86 -mcpu=corei7 -verify-machineinstrs | FileCheck %s --check-prefix X32
Re-work X86 code generation of atomic ops with spin-loop - Rewrite/merge pseudo-atomic instruction emitters to address the following issue: * Reduce one unnecessary load in spin-loop previously the spin-loop looks like thisMBB: newMBB: ld t1 = [bitinstr.addr] op t2 = t1, [bitinstr.val] not t3 = t2 (if Invert) mov EAX = t1 lcs dest = [bitinstr.addr], t3 [EAX is implicit] bz newMBB fallthrough -->nextMBB the 'ld' at the beginning of newMBB should be lift out of the loop as lcs (or CMPXCHG on x86) will load the current memory value into EAX. This loop is refined as: thisMBB: EAX = LOAD [MI.addr] mainMBB: t1 = OP [MI.val], EAX LCMPXCHG [MI.addr], t1, [EAX is implicitly used & defined] JNE mainMBB sinkMBB: * Remove immopc as, so far, all pseudo-atomic instructions has all-register form only, there is no immedidate operand. * Remove unnecessary attributes/modifiers in pseudo-atomic instruction td * Fix issues in PR13458 - Add comprehensive tests on atomic ops on various data types. NOTE: Some of them are turned off due to missing functionality. - Revise tests due to the new spin-loop generated. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@164281 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-20 03:06:15 +00:00
@sc64 = external global i64
define void @atomic_fetch_add64() nounwind {
; X32: atomic_fetch_add64
entry:
%t1 = atomicrmw add i64* @sc64, i64 1 acquire
; X32: addl
; X32: adcl
; X32: lock
; X32: cmpxchg8b
%t2 = atomicrmw add i64* @sc64, i64 3 acquire
; X32: addl
; X32: adcl
; X32: lock
; X32: cmpxchg8b
%t3 = atomicrmw add i64* @sc64, i64 5 acquire
; X32: addl
; X32: adcl
; X32: lock
; X32: cmpxchg8b
%t4 = atomicrmw add i64* @sc64, i64 %t3 acquire
; X32: addl
; X32: adcl
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_sub64() nounwind {
; X32: atomic_fetch_sub64
%t1 = atomicrmw sub i64* @sc64, i64 1 acquire
; X32: subl
; X32: sbbl
; X32: lock
; X32: cmpxchg8b
%t2 = atomicrmw sub i64* @sc64, i64 3 acquire
; X32: subl
; X32: sbbl
; X32: lock
; X32: cmpxchg8b
%t3 = atomicrmw sub i64* @sc64, i64 5 acquire
; X32: subl
; X32: sbbl
; X32: lock
; X32: cmpxchg8b
%t4 = atomicrmw sub i64* @sc64, i64 %t3 acquire
; X32: subl
; X32: sbbl
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_and64() nounwind {
; X32: atomic_fetch_and64
%t1 = atomicrmw and i64* @sc64, i64 3 acquire
; X32: andl
; X32: andl
; X32: lock
; X32: cmpxchg8b
%t2 = atomicrmw and i64* @sc64, i64 5 acquire
; X32: andl
; X32: andl
; X32: lock
; X32: cmpxchg8b
%t3 = atomicrmw and i64* @sc64, i64 %t2 acquire
; X32: andl
; X32: andl
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_or64() nounwind {
; X32: atomic_fetch_or64
%t1 = atomicrmw or i64* @sc64, i64 3 acquire
; X32: orl
; X32: orl
; X32: lock
; X32: cmpxchg8b
%t2 = atomicrmw or i64* @sc64, i64 5 acquire
; X32: orl
; X32: orl
; X32: lock
; X32: cmpxchg8b
%t3 = atomicrmw or i64* @sc64, i64 %t2 acquire
; X32: orl
; X32: orl
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_xor64() nounwind {
; X32: atomic_fetch_xor64
%t1 = atomicrmw xor i64* @sc64, i64 3 acquire
; X32: xorl
; X32: xorl
; X32: lock
; X32: cmpxchg8b
%t2 = atomicrmw xor i64* @sc64, i64 5 acquire
; X32: xorl
; X32: xorl
; X32: lock
; X32: cmpxchg8b
%t3 = atomicrmw xor i64* @sc64, i64 %t2 acquire
; X32: xorl
; X32: xorl
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_nand64(i64 %x) nounwind {
; X32: atomic_fetch_nand64
%t1 = atomicrmw nand i64* @sc64, i64 %x acquire
; X32: andl
; X32: andl
; X32: notl
; X32: notl
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_max64(i64 %x) nounwind {
%t1 = atomicrmw max i64* @sc64, i64 %x acquire
; X32: cmpl
; X32: cmpl
; X32: cmov
; X32: cmov
; X32: cmov
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_min64(i64 %x) nounwind {
%t1 = atomicrmw min i64* @sc64, i64 %x acquire
; X32: cmpl
; X32: cmpl
; X32: cmov
; X32: cmov
; X32: cmov
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_umax64(i64 %x) nounwind {
%t1 = atomicrmw umax i64* @sc64, i64 %x acquire
; X32: cmpl
; X32: cmpl
; X32: cmov
; X32: cmov
; X32: cmov
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_umin64(i64 %x) nounwind {
%t1 = atomicrmw umin i64* @sc64, i64 %x acquire
; X32: cmpl
; X32: cmpl
; X32: cmov
; X32: cmov
; X32: cmov
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_cmpxchg64() nounwind {
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
%t1 = cmpxchg i64* @sc64, i64 0, i64 1 acquire acquire
Re-work X86 code generation of atomic ops with spin-loop - Rewrite/merge pseudo-atomic instruction emitters to address the following issue: * Reduce one unnecessary load in spin-loop previously the spin-loop looks like thisMBB: newMBB: ld t1 = [bitinstr.addr] op t2 = t1, [bitinstr.val] not t3 = t2 (if Invert) mov EAX = t1 lcs dest = [bitinstr.addr], t3 [EAX is implicit] bz newMBB fallthrough -->nextMBB the 'ld' at the beginning of newMBB should be lift out of the loop as lcs (or CMPXCHG on x86) will load the current memory value into EAX. This loop is refined as: thisMBB: EAX = LOAD [MI.addr] mainMBB: t1 = OP [MI.val], EAX LCMPXCHG [MI.addr], t1, [EAX is implicitly used & defined] JNE mainMBB sinkMBB: * Remove immopc as, so far, all pseudo-atomic instructions has all-register form only, there is no immedidate operand. * Remove unnecessary attributes/modifiers in pseudo-atomic instruction td * Fix issues in PR13458 - Add comprehensive tests on atomic ops on various data types. NOTE: Some of them are turned off due to missing functionality. - Revise tests due to the new spin-loop generated. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@164281 91177308-0d34-0410-b5e6-96231b3b80d8
2012-09-20 03:06:15 +00:00
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_store64(i64 %x) nounwind {
store atomic i64 %x, i64* @sc64 release, align 8
; X32: lock
; X32: cmpxchg8b
ret void
; X32: ret
}
define void @atomic_fetch_swap64(i64 %x) nounwind {
%t1 = atomicrmw xchg i64* @sc64, i64 %x acquire
; X32: lock
; X32: xchg8b
ret void
; X32: ret
}
Reference in New Issue Copy Permalink