android: add mallinfo2

[SamB]

It's actually just mallinfo in a funny hat, just like in C; see https://android-review.googlesource.com/c/platform/bionic/+/1910571 for the review of when they added this to Bionic's malloc.h header.

Since it's actually just mallinfo in dsguise, there is no minimum SDK level (just like for mallinfo).

Unlike in the header, I had to copy and paste the struct, because that's easier than building/borrowing a Rust macro to duplicate the fields. (Sometimes worse really is better.)

[rustbot]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @tgross35 (or someone else) some time within the next two weeks.

Please see the contribution instructions for more information. Namely, in order to ensure the minimum review times lag, PR authors and assigned reviewers should ensure that the review label (S-waiting-on-review and S-waiting-on-author) stays updated, invoking these commands when appropriate:

• @rustbot author: the review is finished, PR author should check the comments and take action accordingly
• @rustbot review: the author is ready for a review, this PR will be queued again in the reviewer's queue

[SamB]

Hmm, I was really hoping there was Android CI because there are a lot of funny little failures when I run "cargo test" on my phone (and that's after skipping the tests for sem_open/sem_close/sem_unlink that Termux #defines to names not found in libc.so).

Lucky for me, it looks like more checks materialized while I was writing this comment.

Well, the dynamic linker looks a bit grouchy, but the tests themselves appear to pass.

See e.g.
https://github.com/rust-lang/libc/actions/runs/11006374551/job/30560642282?pr=3940#step:4:972 for the linker complaints:

linker: /checkout/target/x86_64-linux-android/debug/deps/main-5f4c44db87887603: unsupported flags DT_FLAGS_1=0x8000001

No clue why the dynamic linker on my phone is happy with the same flags ...

[tgross35]

This looks fine to me, @maurer mind double checking?

[maurer]

Seems fine from an Android POV, that function is indeed available.

That said:

1. Maybe we should use a type alias rather than a separate struct? Not a strong opinion here though.
2. What are you looking for mallinfo2 for? malloc_info is more reliable and supported from API level 23, which accounts for 98.6% of existing devices. (It is true though, that rustc still supports back to API level 21.)

[tgross35]

Labeled waiting on author since the type alias makes sense here as Matthew mentioned (pub type mallinfo2 = mallinfo). Just comment @rustbot review once that is done.

What are you looking for mallinfo2 for? malloc_info is more reliable and supported from API level 23, which accounts for 98.6% of existing devices. (It is true though, that rustc still supports back to API level 21.)

It would be good to confirm that this is actually useful as well (easy enough if you are supporting an old API).

[tgross35]

@SamB could you change this to make use of type aliases? Also confirm that this is for situations where malloc_info isn't available.

[bors]

☔ The latest upstream changes (presumably #4132) made this pull request unmergeable. Please resolve the merge conflicts.