A small Rust wrapper around dl_iterate_phdr(3).
#1
Conversation
src/lib.rs
Outdated
| assert_ne!(o.addr(), 0); | ||
|
|
||
| let obj_name = o.name().clone().into_string().unwrap(); | ||
| println!("obj name: '{}'", obj_name); |
src/lib.rs
Outdated
| /// The name of the object. | ||
| name: CString, | ||
| /// Vector of program headers. | ||
| //phdrs: Vec<*const p_ffi::Elf_Phdr>, |
There was a problem hiding this comment.
Not sure I understand this line.
src/lib.rs
Outdated
| } | ||
| } | ||
|
|
||
| // The same as what `#[derive(Debug)]` would have done, but formats the fields using hex. |
There was a problem hiding this comment.
I think you can delete this comment.
src/lib.rs
Outdated
|
|
||
| impl ProgramHeader { | ||
| /// Returns the segment type (as one of the `PT_*` constants). | ||
| pub fn typ(&self) -> Elf_Word { |
There was a problem hiding this comment.
It's probably more conventional to call this type_.
| flag_strs.push("PF_MASKPROC"); | ||
| } | ||
| to_write.push_str(&format!("flags=<{}>, ", flag_strs.join("|"))); | ||
|
|
There was a problem hiding this comment.
The last 15 lines all deal with one field. It separates this from the following fields, which have one LoC per field.
src/lib.rs
Outdated
| type Item = ProgramHeader; | ||
|
|
||
| fn next(&mut self) -> Option<Self::Item> { | ||
| match self.num { |
There was a problem hiding this comment.
I think if self.num == 0 { ... } else { ... } would be clearer.
src/lib.rs
Outdated
| } | ||
|
|
||
| /// Returns a `Vec` of objects loaded into the current address space. | ||
| /// The reason this interface returns a `Vec` and not some kind of iterator, is that the C function |
There was a problem hiding this comment.
Something like "This returns a Vec because we have to read all the values from dl_iterate_phdr calls in one go (i.e. we have to ensure that calls to dl_iterate_phdr are not interspersed)."?
There was a problem hiding this comment.
The first sentence is fine, but the second doesn't make sense. C calls our callback once per header, and we have no way to interrupt this to offer the user an intermediate result. Makes sense?
There was a problem hiding this comment.
OK, then I think I would simply delete the "the reason" sentence. It's not very important detail and (if I'm being brutal) it's a bit waffly at the moment. It's probably best that we only tell the user things they really need to know, and I don't think an API is a great place for this.
|
It might be nice to compile the example program at the same point as everything else? You might be able to use a workspace (such as https://github.com/softdevteam/grmtools/blob/master/Cargo.toml). |
|
Not saying it's a bad idea, but why would that be useful? |
|
The example program might catch some compile-time bugs that aren't caught elsewhere? OK, in this case, probably not; but in things like grmtools, it can be useful (e.g. certain macro use stuff). |
|
I'd rather add |
|
Yes, but then only Travis will catch things. Anyway, in this case, it's not a big deal either way (the example is tiny), but it might be worth thinking about for other repos in the future. |
|
That's all of the comments done, I think. |
|
Ah, we should have a |
|
OK, prepare for travis testing commits. |
That's the upstream bug I mentioned. |
|
Aha! |
|
It hurt my head, but I think I found the root of the test failure. bindgen erroneously emits Reported to upstream. |
|
Just to keep this PR in the loop. A fix for the failing test was in the works, but has become blocked on an LLVM alignment issue: |
|
OK, this fixes the auto-generated test. Ready for re-review. |
.travis.yml
Outdated
| - export PATH="$PATH:$HOME/.cargo/bin" | ||
|
|
||
| script: | ||
| # By installing rustfmt, auto-generated code becomes more readable. |
There was a problem hiding this comment.
We install rustfmt but don't seem to use it?
There was a problem hiding this comment.
Edd confirms that without the use of rustfmt, auto-generated code contains lines so long that they cause Travis problems.
There was a problem hiding this comment.
A possible comment might be something like "We need to use rustfmt on auto-generated code: without it, code is put on a single line, which then causes Travis to crash because ...".
|
Give this a shot |
|
Added the rustfmt config and ran rustfmt. |
|
I think we're ready to squash? |
|
Splat. |
Let's start getting some Yorick stuff out of private repos.
Starting with the easy stuff, this is a little library to make it painless to inspect the program headers from Rust. I wrote it some time ago, and spent some time today tidying it up for code review.
In Yorick, we use this to know which object (compilation unit) a virtual address from a PT trace comes from.
One auto-generated test block us from merging this. I raised an issue for this a while back, but upstream have not yet had a chance to look: rust-lang/rust-bindgen#1370
We can still do the code review.
Cheers