Revert "CFI: Skip non-passed arguments"

[rcvalle]

Since we're now using an approach which is a variant of #123082 (that transforms closures into dynamic Fn traits but isolating it to the Fn call methods only) instead of #121962 or #122573, skipping non-passed arguments shouldn't be necessary for KCFI anymore and we can claim back the reduced granularity.

This should be tested with and merged after the remaining KCFI-related patches are merged.

This reverts commit f2f0d25 (#122456).

This was split off from #116404.

cc @compiler-errors @workingjubilee @maurer

[rustbot]

r? @wesleywiser

rustbot has assigned @wesleywiser.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rcvalle]

r? @compiler-errors

[rust-log-analyzer]

The job x86_64-gnu-llvm-17 failed! Check out the build log: (web) (plain)

Click to see the possible cause of the failure (guessed by this bot)

#16 exporting to docker image format
#16 sending tarball 30.5s done
#16 DONE 34.6s
##[endgroup]
Setting extra environment values for docker:  --env ENABLE_GCC_CODEGEN=1 --env GCC_EXEC_PREFIX=/usr/lib/gcc/
[CI_JOB_NAME=x86_64-gnu-llvm-17]
---
sccache: Starting the server...
##[group]Configure the build
configure: processing command line
configure: 
configure: build.configure-args := ['--build=x86_64-unknown-linux-gnu', '--llvm-root=/usr/lib/llvm-17', '--enable-llvm-link-shared', '--set', 'rust.thin-lto-import-instr-limit=10', '--set', 'change-id=99999999', '--enable-verbose-configure', '--enable-sccache', '--disable-manage-submodules', '--enable-locked-deps', '--enable-cargo-native-static', '--set', 'rust.codegen-units-std=1', '--set', 'dist.compression-profile=balanced', '--dist-compression-formats=xz', '--disable-dist-src', '--release-channel=nightly', '--enable-debug-assertions', '--enable-overflow-checks', '--enable-llvm-assertions', '--set', 'rust.verify-llvm-ir', '--set', 'rust.codegen-backends=llvm,cranelift,gcc', '--set', 'llvm.static-libstdcpp', '--enable-new-symbol-mangling']
configure: target.x86_64-unknown-linux-gnu.llvm-config := /usr/lib/llvm-17/bin/llvm-config
configure: llvm.link-shared     := True
configure: rust.thin-lto-import-instr-limit := 10
configure: change-id            := 99999999
---
failures:

---- [ui] tests/ui/sanitizer/cfi-closure-fn-ptr-cast.rs#cfi stdout ----

error in revision `cfi`: test run failed!
status: signal: 4 (SIGILL) (core dumped)
command: cd "/checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/sanitizer/cfi-closure-fn-ptr-cast.cfi" && RUST_TEST_THREADS="16" "/checkout/obj/build/x86_64-unknown-linux-gnu/test/ui/sanitizer/cfi-closure-fn-ptr-cast.cfi/a"
stderr: none

[workingjubilee]

We can choose to ship this eventually, since as we've discussed, while I find eliding () to be a fine implementation, it's not quite as "Rust-incorrect" per se as other problems we are managing with CFI... but it does need to actually build and run code, yeah, so it's gonna wait until enough things land that it can work first. ^^;

[rcvalle]

It doesn't just elide (), but also function items, and more (e.g., the extern function and the constructor that had to be wrapped in a reference in the tests so they aren't gone). For example, this made functions that had only one parameter that was a function item and otherwise were each one placed in its own group to be grouped together with all functions that were just fn func(), which was what happened with these examples in the tests. And this is just an example, any function that was disambiguated by a ZST had a similar grouping (or ungrouping).

But yes, we should wait the remaining patches to be merged and test this with them before merging. If we still find an issue, we should look into it and fix it, instead of work around it with this.

[bors]

☔ The latest upstream changes (presumably #123455) made this pull request unmergeable. Please resolve the merge conflicts.