[all-commits] [llvm/llvm-project] 870094: [ELF] Decrease alignment of ThunkSection on 64-bit...

[Fangrui Song via All-commits]

  Branch: refs/heads/master
  Home:   https://github.com/llvm/llvm-project
  Commit: 870094decfc9fe80c8e0a6405421b7d09b97b02b
      https://github.com/llvm/llvm-project/commit/870094decfc9fe80c8e0a6405421b7d09b97b02b
  Author: Fangrui Song <maskray at google.com>
  Date:   2020-01-16 (Thu, 16 Jan 2020)

  Changed paths:
    M lld/ELF/SyntheticSections.cpp
    M lld/test/ELF/aarch64-call26-thunk.s
    M lld/test/ELF/aarch64-cortex-a53-843419-thunk.s
    M lld/test/ELF/aarch64-jump26-thunk.s
    M lld/test/ELF/aarch64-thunk-pi.s
    M lld/test/ELF/aarch64-thunk-script.s
    M lld/test/ELF/ppc64-dtprel.s
    M lld/test/ELF/ppc64-ifunc.s
    M lld/test/ELF/ppc64-long-branch.s
    M lld/test/ELF/ppc64-tls-gd.s
    M lld/test/ELF/ppc64-toc-restore.s

  Log Message:
  -----------
  [ELF] Decrease alignment of ThunkSection on 64-bit targets from 8 to 4

ThunkSection contains 4-byte instructions on all targets that use
thunks. Thunks should not be used in any performance sensitive places,
and locality/cache line/instruction fetching arguments should not apply.

We use 16 bytes as preferred function alignments for modern PowerPC cores.
In any case, 8 is not optimal.

Differential Revision: https://reviews.llvm.org/D72819