[llvm] 89baeae - Reapply "RegAllocFast: Rewrite and improve"

Florian Hahn via llvm-commits llvm-commits at lists.llvm.org
Thu Feb 4 07:17:14 PST 2021 

To close the loop, I put up https://reviews.llvm.org/D96015 <https://reviews.llvm.org/D96015> with a potential fix.

> On Feb 4, 2021, at 11:04, Florian Hahn <florian_hahn at apple.com> wrote:
> 
> 
> Hi Matt,
> 
> 
> We are seeing an issue where regallocfast runs out of registers with `llc -O0 -verify-machineinstrs` on X86 on the snippet below. Any ideas what could be going wrong here?
> 
> 
> target datalayout = "e-m:o-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
> target triple = "x86_64-unknown-linux"
> 
> define void @widget(i32* %arg, i1 %arg1) align 2 personality i8* undef {
> bb:
>  br i1 %arg1, label %bb2, label %bb3
> 
> bb2:                                              ; preds = %bb
>  %tmp = alloca i8, i64 6144, align 32
>  br label %bb3
> 
> bb3:                                              ; preds = %bb2, %bb
>  %tmp4 = load i32, i32* %arg, align 4
>  br label %bb5
> 
> bb5:                                              ; preds = %bb3
>  %tmp6 = lshr i32 %tmp4, 8
>  %tmp7 = and i32 %tmp6, 255
>  invoke void (i8*, i32, i8*, ...) @baz(i8* undef, i32 3427, i8* undef, double undef, double undef, i32 undef, i32 %tmp7)
>          to label %bb8 unwind label %bb9
> 
> bb8:                                              ; preds = %bb5
>  ret void
> 
> bb9:                                              ; preds = %bb5
>  %tmp10 = landingpad { i8*, i32 }
>          cleanup
>  ret void
> }
> 
> declare void @baz(i8*, i32, i8*, …)
> 
> 
> `llc -O0 -verify-machineinstrs`  produces:
> 
> error: ran out of registers during register allocation
> 
>> On Sep 30, 2020, at 15:35, Matt Arsenault via llvm-commits <llvm-commits at lists.llvm.org> wrote:
>> 
>> 
>> Author: Matt Arsenault
>> Date: 2020-09-30T10:35:25-04:00
>> New Revision: 89baeaef2fa9a2441d087a218ac82e11a5d4e548
>> 
>> URL: https://github.com/llvm/llvm-project/commit/89baeaef2fa9a2441d087a218ac82e11a5d4e548
>> DIFF: https://github.com/llvm/llvm-project/commit/89baeaef2fa9a2441d087a218ac82e11a5d4e548.diff
>> 
>> LOG: Reapply "RegAllocFast: Rewrite and improve"
>> 
>> This reverts commit 73a6a164b84a8195defbb8f5eeb6faecfc478ad4.
>> 
>> Added: 
>>   llvm/test/CodeGen/AMDGPU/fast-ra-kills-vcc.mir
>>   llvm/test/CodeGen/AMDGPU/fastregalloc-illegal-subreg-physreg.mir
>>   llvm/test/CodeGen/AMDGPU/unexpected-reg-unit-state.mir
>>   llvm/test/CodeGen/PowerPC/spill-nor0.mir
>>   llvm/test/CodeGen/X86/bug47278-eflags-error.mir
>>   llvm/test/CodeGen/X86/bug47278.mir
>> 
>> Modified: 
>>   lldb/test/Shell/SymbolFile/NativePDB/disassembly.cpp
>>   llvm/lib/CodeGen/RegAllocFast.cpp
>>   llvm/test/CodeGen/AArch64/GlobalISel/builtin-return-address-pacret.ll
>>   llvm/test/CodeGen/AArch64/GlobalISel/darwin-tls-call-clobber.ll
>>   llvm/test/CodeGen/AArch64/arm64-fast-isel-br.ll
>>   llvm/test/CodeGen/AArch64/arm64-fast-isel-call.ll
>>   llvm/test/CodeGen/AArch64/arm64-fast-isel-conversion-fallback.ll
>>   llvm/test/CodeGen/AArch64/arm64-fast-isel-conversion.ll
>>   llvm/test/CodeGen/AArch64/arm64-vcvt_f.ll
>>   llvm/test/CodeGen/AArch64/arm64_32-fastisel.ll
>>   llvm/test/CodeGen/AArch64/arm64_32-null.ll
>>   llvm/test/CodeGen/AArch64/br-cond-not-merge.ll
>>   llvm/test/CodeGen/AArch64/cmpxchg-O0.ll
>>   llvm/test/CodeGen/AArch64/combine-loads.ll
>>   llvm/test/CodeGen/AArch64/fast-isel-cmpxchg.ll
>>   llvm/test/CodeGen/AArch64/popcount.ll
>>   llvm/test/CodeGen/AArch64/swift-return.ll
>>   llvm/test/CodeGen/AArch64/swifterror.ll
>>   llvm/test/CodeGen/AArch64/unwind-preserved-from-mir.mir
>>   llvm/test/CodeGen/AArch64/unwind-preserved.ll
>>   llvm/test/CodeGen/AMDGPU/GlobalISel/inline-asm.ll
>>   llvm/test/CodeGen/AMDGPU/control-flow-fastregalloc.ll
>>   llvm/test/CodeGen/AMDGPU/fastregalloc-self-loop-heuristic.mir
>>   llvm/test/CodeGen/AMDGPU/indirect-addressing-term.ll
>>   llvm/test/CodeGen/AMDGPU/mubuf-legalize-operands.ll
>>   llvm/test/CodeGen/AMDGPU/partial-sgpr-to-vgpr-spills.ll
>>   llvm/test/CodeGen/AMDGPU/reserve-vgpr-for-sgpr-spill.ll
>>   llvm/test/CodeGen/AMDGPU/spill-agpr.mir
>>   llvm/test/CodeGen/AMDGPU/spill-m0.ll
>>   llvm/test/CodeGen/AMDGPU/spill192.mir
>>   llvm/test/CodeGen/AMDGPU/wwm-reserved.ll
>>   llvm/test/CodeGen/ARM/2010-08-04-StackVariable.ll
>>   llvm/test/CodeGen/ARM/Windows/alloca.ll
>>   llvm/test/CodeGen/ARM/cmpxchg-O0-be.ll
>>   llvm/test/CodeGen/ARM/cmpxchg-O0.ll
>>   llvm/test/CodeGen/ARM/crash-greedy-v6.ll
>>   llvm/test/CodeGen/ARM/debug-info-blocks.ll
>>   llvm/test/CodeGen/ARM/fast-isel-call.ll
>>   llvm/test/CodeGen/ARM/fast-isel-intrinsic.ll
>>   llvm/test/CodeGen/ARM/fast-isel-ldr-str-thumb-neg-index.ll
>>   llvm/test/CodeGen/ARM/fast-isel-select.ll
>>   llvm/test/CodeGen/ARM/fast-isel-vararg.ll
>>   llvm/test/CodeGen/ARM/ldrd.ll
>>   llvm/test/CodeGen/ARM/legalize-bitcast.ll
>>   llvm/test/CodeGen/ARM/pr47454.ll
>>   llvm/test/CodeGen/ARM/stack-guard-reassign.ll
>>   llvm/test/CodeGen/ARM/swifterror.ll
>>   llvm/test/CodeGen/ARM/thumb-big-stack.ll
>>   llvm/test/CodeGen/Hexagon/vect/vect-load-v4i16.ll
>>   llvm/test/CodeGen/Mips/Fast-ISel/callabi.ll
>>   llvm/test/CodeGen/Mips/Fast-ISel/memtest1.ll
>>   llvm/test/CodeGen/Mips/Fast-ISel/pr40325.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/add.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/add_vec.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/aggregate_struct_return.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/bitreverse.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/bitwise.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/branch.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/brindirect.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/bswap.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/call.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/ctlz.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/ctpop.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/cttz.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/dyn_stackalloc.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/fcmp.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/float_constants.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/fptosi_and_fptoui.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/global_address.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/global_address_pic.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/icmp.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/jump_table_and_brjt.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/load_4_unaligned.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/load_split_because_of_memsize_or_align.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/long_ambiguous_chain_s32.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/long_ambiguous_chain_s64.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/mul.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/mul_vec.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/phi.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/rem_and_div.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/select.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/sitofp_and_uitofp.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/store_4_unaligned.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/store_split_because_of_memsize_or_align.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/sub.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/sub_vec.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/test_TypeInfoforMF.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/var_arg.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/zextLoad_and_sextLoad.ll
>>   llvm/test/CodeGen/Mips/GlobalISel/llvm-ir/zext_and_sext.ll
>>   llvm/test/CodeGen/Mips/atomic-min-max.ll
>>   llvm/test/CodeGen/Mips/atomic.ll
>>   llvm/test/CodeGen/Mips/atomic64.ll
>>   llvm/test/CodeGen/Mips/atomicCmpSwapPW.ll
>>   llvm/test/CodeGen/Mips/copy-fp64.ll
>>   llvm/test/CodeGen/Mips/implicit-sret.ll
>>   llvm/test/CodeGen/Mips/micromips-eva.mir
>>   llvm/test/CodeGen/Mips/msa/ldr_str.ll
>>   llvm/test/CodeGen/PowerPC/addegluecrash.ll
>>   llvm/test/CodeGen/PowerPC/aggressive-anti-dep-breaker-subreg.ll
>>   llvm/test/CodeGen/PowerPC/aix-overflow-toc.py
>>   llvm/test/CodeGen/PowerPC/anon_aggr.ll
>>   llvm/test/CodeGen/PowerPC/builtins-ppc-p10vsx.ll
>>   llvm/test/CodeGen/PowerPC/elf-common.ll
>>   llvm/test/CodeGen/PowerPC/fast-isel-pcrel.ll
>>   llvm/test/CodeGen/PowerPC/fp-int128-fp-combine.ll
>>   llvm/test/CodeGen/PowerPC/fp-strict-fcmp-noopt.ll
>>   llvm/test/CodeGen/PowerPC/fp64-to-int16.ll
>>   llvm/test/CodeGen/PowerPC/p9-vinsert-vextract.ll
>>   llvm/test/CodeGen/PowerPC/popcount.ll
>>   llvm/test/CodeGen/PowerPC/spill-nor0.ll
>>   llvm/test/CodeGen/PowerPC/stack-guard-reassign.ll
>>   llvm/test/CodeGen/PowerPC/vsx-args.ll
>>   llvm/test/CodeGen/PowerPC/vsx.ll
>>   llvm/test/CodeGen/SPARC/fp16-promote.ll
>>   llvm/test/CodeGen/SystemZ/swift-return.ll
>>   llvm/test/CodeGen/SystemZ/swifterror.ll
>>   llvm/test/CodeGen/Thumb2/LowOverheadLoops/branch-targets.ll
>>   llvm/test/CodeGen/Thumb2/high-reg-spill.mir
>>   llvm/test/CodeGen/Thumb2/mve-vector-spill.ll
>>   llvm/test/CodeGen/X86/2009-04-14-IllegalRegs.ll
>>   llvm/test/CodeGen/X86/2010-06-28-FastAllocTiedOperand.ll
>>   llvm/test/CodeGen/X86/2013-10-14-FastISel-incorrect-vreg.ll
>>   llvm/test/CodeGen/X86/atomic-monotonic.ll
>>   llvm/test/CodeGen/X86/atomic-unordered.ll
>>   llvm/test/CodeGen/X86/atomic32.ll
>>   llvm/test/CodeGen/X86/atomic64.ll
>>   llvm/test/CodeGen/X86/atomic6432.ll
>>   llvm/test/CodeGen/X86/avx-load-store.ll
>>   llvm/test/CodeGen/X86/avx512-mask-zext-bugfix.ll
>>   llvm/test/CodeGen/X86/crash-O0.ll
>>   llvm/test/CodeGen/X86/extend-set-cc-uses-dbg.ll
>>   llvm/test/CodeGen/X86/fast-isel-cmp-branch.ll
>>   llvm/test/CodeGen/X86/fast-isel-nontemporal.ll
>>   llvm/test/CodeGen/X86/fast-isel-select-sse.ll
>>   llvm/test/CodeGen/X86/fast-isel-select.ll
>>   llvm/test/CodeGen/X86/fast-isel-x86-64.ll
>>   llvm/test/CodeGen/X86/mixed-ptr-sizes-i686.ll
>>   llvm/test/CodeGen/X86/mixed-ptr-sizes.ll
>>   llvm/test/CodeGen/X86/phys-reg-local-regalloc.ll
>>   llvm/test/CodeGen/X86/pr11415.ll
>>   llvm/test/CodeGen/X86/pr1489.ll
>>   llvm/test/CodeGen/X86/pr27591.ll
>>   llvm/test/CodeGen/X86/pr30430.ll
>>   llvm/test/CodeGen/X86/pr30813.ll
>>   llvm/test/CodeGen/X86/pr32241.ll
>>   llvm/test/CodeGen/X86/pr32284.ll
>>   llvm/test/CodeGen/X86/pr32340.ll
>>   llvm/test/CodeGen/X86/pr32345.ll
>>   llvm/test/CodeGen/X86/pr32451.ll
>>   llvm/test/CodeGen/X86/pr32484.ll
>>   llvm/test/CodeGen/X86/pr34592.ll
>>   llvm/test/CodeGen/X86/pr34653.ll
>>   llvm/test/CodeGen/X86/pr39733.ll
>>   llvm/test/CodeGen/X86/pr42452.ll
>>   llvm/test/CodeGen/X86/pr44749.ll
>>   llvm/test/CodeGen/X86/pr47000.ll
>>   llvm/test/CodeGen/X86/regalloc-fast-missing-live-out-spill.mir
>>   llvm/test/CodeGen/X86/stack-protector-msvc.ll
>>   llvm/test/CodeGen/X86/stack-protector-strong-macho-win32-xor.ll
>>   llvm/test/CodeGen/X86/swift-return.ll
>>   llvm/test/CodeGen/X86/swifterror.ll
>>   llvm/test/CodeGen/X86/volatile.ll
>>   llvm/test/CodeGen/X86/win64_eh.ll
>>   llvm/test/CodeGen/X86/x86-32-intrcc.ll
>>   llvm/test/CodeGen/X86/x86-64-intrcc.ll
>>   llvm/test/DebugInfo/AArch64/frameindices.ll
>>   llvm/test/DebugInfo/AArch64/prologue_end.ll
>>   llvm/test/DebugInfo/ARM/prologue_end.ll
>>   llvm/test/DebugInfo/Mips/delay-slot.ll
>>   llvm/test/DebugInfo/Mips/prologue_end.ll
>>   llvm/test/DebugInfo/X86/dbg-declare-arg.ll
>>   llvm/test/DebugInfo/X86/fission-ranges.ll
>>   llvm/test/DebugInfo/X86/op_deref.ll
>>   llvm/test/DebugInfo/X86/parameters.ll
>>   llvm/test/DebugInfo/X86/pieces-1.ll
>>   llvm/test/DebugInfo/X86/prologue-stack.ll
>>   llvm/test/DebugInfo/X86/reference-argument.ll
>>   llvm/test/DebugInfo/X86/spill-indirect-nrvo.ll
>>   llvm/test/DebugInfo/X86/sret.ll
>>   llvm/test/DebugInfo/X86/subreg.ll

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20210204/9f01be42/attachment.html>

More information about the llvm-commits mailing list