[all-commits] [llvm/llvm-project] 08dd00: BOLT may never run the tentative code layout for c...

[Paschalis Mpeis via All-commits]

  Branch: refs/heads/users/paschalis-mpeis/lite-split-functions-cold-layout
  Home:   https://github.com/llvm/llvm-project
  Commit: 08dd00fd7e25114e456acaa650d56d4202303518
      https://github.com/llvm/llvm-project/commit/08dd00fd7e25114e456acaa650d56d4202303518
  Author: Paschalis Mpeis <Paschalis.Mpeis at arm.com>
  Date:   2024-10-01 (Tue, 01 Oct 2024)

  Changed paths:
    M bolt/lib/Passes/LongJmp.cpp
    A bolt/test/AArch64/split-funcs-lite.test

  Log Message:
  -----------
  BOLT may never run the tentative code layout for cold BBs.

However, it might later require the addresses of cold BBs to compute
PC-relative jumps between hot and cold fragments of a function. Since
the addresses were never estimated, it uses 0x0 for cold BBs, resulting
in incorrect estimated addresses for any instructions within them.

As a result, LongJump often expands a branch that would otherwise fit in
a single instruction into a short jump.

For example, `relaxStub` might expand from:
```armasm
b       .Ltmp1234
```

to:
```armasm
adrp    x16, .Ltmp1234
add     x16, x16, :lo12:.Ltmp1234
br      x16
```

While this expansion is not wrong, it is unnecessary. Moreover, in some
large binaries this expansion has lead to runtime crashes. This should
be investigated further, as such expansions should not cause crashes.


  Commit: 31dd42ab65e14670cb89713c497fcf2b4cf098b8
      https://github.com/llvm/llvm-project/commit/31dd42ab65e14670cb89713c497fcf2b4cf098b8
  Author: Paschalis Mpeis <Paschalis.Mpeis at arm.com>
  Date:   2024-10-01 (Tue, 01 Oct 2024)

  Changed paths:
    M bolt/lib/Passes/LongJmp.cpp
    M bolt/test/AArch64/split-funcs-lite.test

  Log Message:
  -----------
  [AArch64][BOLT] Ensure tentative code layout for cold BBs runs.

When BOLT is not processing all functions (e.g., when lite mode is
enabled or when multiple functions are ignored via the 'skip-funcs' flag),
we might skip the tentative code layout estimation for cold basic blocks.
However, due to the `-split-functions`` flag, we still need to compute
PC-relative distances between hot and cold basic blocks.

In such cases, BOLT will use '0x0' as the first address of the basic
block, leading to incorrect estimations of the necessary PC-relative
addresses. Consequently, the relaxStub method expands all those branches
to the short-jump sequence unnecessarily.

Such an unnecessary expansion by `relaxStub` is from:
```armasm
b       .Ltmp1234
```

to:
```armasm
adrp    x16, .Ltmp1234
add     x16, x16, :lo12:.Ltmp1234
br      x16
```


  Commit: 690d4857e8d99fe4b9ada1e64de94574f8869a40
      https://github.com/llvm/llvm-project/commit/690d4857e8d99fe4b9ada1e64de94574f8869a40
  Author: Paschalis Mpeis <Paschalis.Mpeis at arm.com>
  Date:   2024-10-01 (Tue, 01 Oct 2024)

  Changed paths:
    M bolt/lib/Passes/LongJmp.cpp
    M bolt/test/AArch64/split-funcs-lite.test

  Log Message:
  -----------
  Addressing reviewers.


Compare: https://github.com/llvm/llvm-project/compare/7cd7950ee082...690d4857e8d9

To unsubscribe from these emails, change your notification settings at https://github.com/llvm/llvm-project/settings/notifications