wasmtime / PR #5020 fuzzgen: Generate compiler flags · git-wasmtime

Stream: git-wasmtime

Topic: wasmtime / PR #5020 fuzzgen: Generate compiler flags

Wasmtime GitHub notifications bot (Oct 05 2022 at 13:56):

afonso360 opened PR #5020 from fuzzgen-flags to main:

:wave: Hey,

This PR allows fuzzgen to generate compiler flags and test using those. We only allow compiler flags that shouldn't change the behaviour of the code.

I wanted to submit this now so that we could fuzz PR's such as #4953 and #5004 that require optimizations enabled.

However, when testing this it has found an issue with one of the passes (I haven't yet narrowed it down), but I don't want to start fixing it if we are going to change pretty much everything once #4953 lands. I'm opening this as a draft, and lets re test this once #4953 lands.

<details>
<summary>The actual bug manifests as following:</summary>

Fails with:
target/aarch64-unknown-linux-gnu/release/cranelift-fuzzgen: Running 1 inputs 1 time(s) each.
Running: fuzz/artifacts/cranelift-fuzzgen/crash-00e37c2858c35c573ed85d618ea5d0c75ae06fda
thread '<unnamed>' panicked at 'assertion failed: `(left != right)`
  left: `v137`,
 right: `v137`: Aliasing v137 to v134 would create a loop', cranelift/codegen/src/ir/dfg.rs:322:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
And I cannot reproduce via the CLIF file since that runs into https://github.com/bytecodealliance/wasmtime/issues/4875#issuecomment-1239314171 (which I kinda forgot about :sweat_smile: ). But I'm going to try and investigate that one in the mean time.
</details>

This also enables regalloc_checker as we discussed in #4979.

cc: @jameysharp

Wasmtime GitHub notifications bot (Oct 08 2022 at 21:52):

afonso360 updated PR #5020 from fuzzgen-flags to main.

Wasmtime GitHub notifications bot (Oct 10 2022 at 19:04):

afonso360 updated PR #5020 from fuzzgen-flags to main.

Wasmtime GitHub notifications bot (Oct 12 2022 at 10:03):

afonso360 updated PR #5020 from fuzzgen-flags to main.

Wasmtime GitHub notifications bot (Oct 17 2022 at 20:00):

cfallin submitted PR review.

Wasmtime GitHub notifications bot (Oct 17 2022 at 20:00):

cfallin submitted PR review.

Wasmtime GitHub notifications bot (Oct 17 2022 at 20:00):

cfallin created PR review comment:

It probably would make sense to leave a comment in cranelift/codegen/meta/src/shared/settings.rs where all of these options are defined to note "please add to fuzzing list ..." as a step when adding new otions?

Wasmtime GitHub notifications bot (Oct 17 2022 at 20:00):

cfallin created PR review comment:

Does it make sense to add riscv64 here too (or are you worried it would find too many issues)?

Actually in the spirit of your enum .all() method above -- perhaps we can define a cranelift_codegen::isa::all() that returns all Triples compiled into the build, and use that here? (Feel free to drop that idea though if it's more than ~20 lines of complexity or so.)

Wasmtime GitHub notifications bot (Oct 17 2022 at 20:00):

cfallin created PR review comment:

Some of these can probably be excluded, based on knowledge of what they do and likely payoff from fuzzing time:

machine_code_cfg_info generates additional metadata for the embedder but this doesn't feed back into compilation anywhere; we can leave it on unconditionally to make sure the generation doesn't panic

enable_simd, enable_float, enable_atomics will cause panics in compilation if an input uses one of those features and it isn't enabled, but don't change compilation otherwise; so we can probably turn them on unconditionally (we're not, e.g., trying to test that a SIMD input causes panics with the feature off)

Wasmtime GitHub notifications bot (Oct 17 2022 at 21:07):

afonso360 submitted PR review.

Wasmtime GitHub notifications bot (Oct 17 2022 at 21:07):

afonso360 created PR review comment:

Yeah the all() idea sounds good, I've been adding them where it think they would be useful for others.

Wasmtime GitHub notifications bot (Oct 17 2022 at 21:14):

afonso360 submitted PR review.

Wasmtime GitHub notifications bot (Oct 17 2022 at 21:14):

afonso360 created PR review comment:

Agreed, let me know if any other flags come to mind!

enable_simd, enable_float, enable_atomics will cause panics in compilation if an input uses one of those features and it isn't enabled,

Enabling them unconditionally sounds good.

But at least for enable_floats it hasn't panicked yet and we almost always insert floats for larger programs, so we might be silently allowing them anyway. (Haven't tested this yet)

Wasmtime GitHub notifications bot (Oct 20 2022 at 17:48):

afonso360 updated PR #5020 from fuzzgen-flags to main.

Wasmtime GitHub notifications bot (Oct 20 2022 at 17:49):

afonso360 submitted PR review.

Wasmtime GitHub notifications bot (Oct 20 2022 at 17:49):

afonso360 created PR review comment:

This ended up not being very easy to do, I might try it again later. But either way I've added riscv64 to the target list.

Wasmtime GitHub notifications bot (Oct 20 2022 at 22:41):

afonso360 has marked PR #5020 as ready for review.

Wasmtime GitHub notifications bot (Oct 20 2022 at 23:40):

cfallin merged PR #5020.

Last updated: Dec 23 2024 at 12:05 UTC