[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

This revision was not accepted when it landed; it landed in state "Needs Review". This revision was automatically updated to reflect the committed changes. Closed by commit rG564d85e090af: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB (authored by lebedev.ri). Changed prior

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

lebedev.ri added a comment. Thanks everyone! I don't expect that there will be any further feedback, so i plan on landing this tomorrow, unless there is further feedback before then. Comment at: llvm/test/CodeGen/MIR/X86/load-with-1gb-alignment.mir:1 +# RUN: llc -run-pass=none

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

lebedev.ri updated this revision to Diff 368744. lebedev.ri marked an inline comment as done. lebedev.ri added a comment. Add `-march=x86-64` that i forgot :/ Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D108661/new/ https://reviews.llvm.org/D10866

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

arsenm added inline comments. Comment at: llvm/test/CodeGen/MIR/X86/load-with-1gb-alignment.mir:1 +# RUN: llc -run-pass=none -o - %s | FileCheck %s + Needs to add -march/-mtriple Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

lebedev.ri added a comment. @arsenm are you happy with the test coverage here? In D108661#2963519 , @craig.topper wrote: > Wild speculation. This may be a historical artifact of LoadInst and StoreInst > having their getAlignment() function written like

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

jdoerfert added a comment. Seems reasonable to me. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D108661/new/ https://reviews.llvm.org/D108661 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

lebedev.ri updated this revision to Diff 368722. lebedev.ri added a comment. Address rest of nits. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D108661/new/ https://reviews.llvm.org/D108661 Files: clang/include/clang/Sema/Sema.h clang/test/CXX

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

lebedev.ri updated this revision to Diff 368719. lebedev.ri added a comment. @arsenm i do not understand how i can avoid having IR there, but addressed other nits. thanks for taking a look! Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D108661/new/

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

arsenm added inline comments. Comment at: llvm/test/CodeGen/X86/load-with-1gb-alignment.mir:1 +# RUN: llc -run-pass machine-cp -o - %s | FileCheck %s + Should use -run-pass=none, and go in test/CodeGen/MIR/X86 Comment at: llvm/test/CodeGen/X86/

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

lebedev.ri updated this revision to Diff 368583. lebedev.ri added a comment. In D108661#2963726 , @arsenm wrote: > Could use some MIR tests to make sure that parser doesn't explode Added. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

arsenm added a comment. Could use some MIR tests to make sure that parser doesn't explode Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D108661/new/ https://reviews.llvm.org/D108661 ___ cfe-commits mailin

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

craig.topper added a comment. Wild speculation. This may be a historical artifact of LoadInst and StoreInst having their getAlignment() function written like this when this limit was created unsigned getAlignment() const { return (1 << (getSubclassDataFromInstruction() >> 1)) >> 1; } B

[PATCH] D108661: The maximal representable alignment in LLVM IR is 1GiB, not 512MiB

lebedev.ri created this revision. lebedev.ri added reviewers: jdoerfert, efriedma, aaron.ballman, nikic, MaskRay. lebedev.ri added a project: LLVM. Herald added subscribers: dexonsmith, okura, kuter, pengfei, jfb. lebedev.ri requested review of this revision. Herald added a reviewer: sstefan1. Hera