Re: rust binder: many pr_warn!() / pr_err!() with missing newlines

[Alice Ryhl <aliceryhl@google.com>]

On Thu, Feb 12, 2026 at 11:30:01PM +0100, Jann Horn wrote:
> Hi!
> 
> I've noticed that in many of the rust files under
> drivers/android/binder/, there are pr_warn!() and pr_err!() statements
> that lack the terminating newline, wrongly indicating that the log
> lines are supposed to be followed by continuations:
> 
> $ git grep 'pr_.*("' drivers/android/ | grep -v '\\n"'
> drivers/android/binder.c: pr_info_ratelimited("%pV", &vaf);
> drivers/android/binder.c: pr_info_ratelimited("%pV", &vaf);
> drivers/android/binder/allocation.rs:
> pr_warn!("Failed to clear data on free: {:?}", e);
> drivers/android/binder/context.rs:
> pr_err!("Context::register_process called on the wrong context.");
> drivers/android/binder/context.rs:
> pr_err!("Context::deregister_process called on the wrong context.");
> drivers/android/binder/context.rs:
> pr_warn!("BINDER_SET_CONTEXT_MGR already set");
> drivers/android/binder/node.rs:            pr_err!("inc_ref_done
> called when no active inc_refs");
> drivers/android/binder/node.rs:                pr_err!("Failure:
> refcount underflow!");
> drivers/android/binder/page_range.rs:            pr_err!("set_page
> called when there is already a page");
> drivers/android/binder/page_range.rs:            pr_debug!("Failed to
> register with vma: invalid vma->vm_mm");
> drivers/android/binder/page_range.rs:            pr_debug!("Failed to
> register with vma: size zero");
> drivers/android/binder/page_range.rs:            pr_debug!("Failed to
> register with vma: already registered");
> drivers/android/binder/page_range.rs:
> pr_warn!("Error in use_page_slow: {:?}", err);
> drivers/android/binder/page_range.rs:                pr_warn!("Page is null!");
> drivers/android/binder/process.rs:            pr_warn!("Notification
> added to `delivered_deaths` twice.");
> drivers/android/binder/process.rs:
> pr_err!("get_current_thread was called from the wrong process.");
> drivers/android/binder/process.rs:                pr_err!("Cannot
> create two threads with the same id.");
> drivers/android/binder/process.rs:
> pr_err!("Detected mismatch between handle_is_present and by_handle");
> drivers/android/binder/process.rs:                pr_warn!("use_range
> failure {:?}", err);
> drivers/android/binder/process.rs:
> pr_err!("outstanding_txns underflow");
> drivers/android/binder/process.rs:            pr_err!("Same thread
> registered with `ready_threads` twice.");
> drivers/android/binder/range_alloc/tree.rs:
> pr_warn!("ENOSPC from range_alloc.reserve_new - size: {}", size);
> drivers/android/binder/rust_binder_main.rs:        pr_warn!("Loaded
> Rust Binder.");
> drivers/android/binder/thread.rs:            pr_warn!("Thread reply
> work is already in use.");
> drivers/android/binder/thread.rs:            pr_warn!("Thread return
> work is already in use.");
> drivers/android/binder/thread.rs:
> pr_err!("UserSlice::read_all returned wrong length in
> BINDER_TYPE_FDA");
> drivers/android/binder/thread.rs:                    pr_warn!("Failed
> copying into alloc: {:?}", err);
> drivers/android/binder/thread.rs:
> pr_warn!("Failed copying ptr into alloc: {:?}", err);
> drivers/android/binder/thread.rs:                    pr_warn!("Failed
> skipping {} from reader: {:?}", fixup_len, err);
> drivers/android/binder/thread.rs:                pr_warn!("Failed
> copying remainder into alloc: {:?}", err);
> drivers/android/binder/thread.rs:                    pr_warn!("Failed
> to get security ctx for id {}: {:?}", secid, err);
> drivers/android/binder/thread.rs:                    pr_warn!("Got
> transaction with invalid offset.");
> drivers/android/binder/thread.rs:
> pr_warn!("Error while translating object.");
> drivers/android/binder/thread.rs:                pr_warn!("Failure in
> apply_sg: {:?}", err);
> drivers/android/binder/thread.rs:                pr_warn!("Failed to
> write security context: {:?}", err);
> drivers/android/binder/thread.rs:                pr_warn!("got new
> transaction with bad transaction stack");
> drivers/android/binder/thread.rs:                pr_warn!("Transaction
> stack changed during transaction!");
> drivers/android/binder/thread.rs:                pr_warn!("Failure
> when writing BR_NOOP at beginning of buffer.");
> drivers/android/binder/thread.rs:
> pr_warn!("Failure in work getter: {:?}", err);
> drivers/android/binder/transaction.rs:
> pr_warn!("Failure in copy_transaction_data: {:?}", err);
> drivers/android/binder/transaction.rs:                pr_warn!("Oneway
> transaction should not be in a transaction stack.");
> drivers/android/binder/transaction.rs:
> pr_warn!("Failure in copy_transaction_data: {:?}", err);
> drivers/android/binder/transaction.rs:                pr_err!("Failed
> to submit oneway transaction to node.");
> 
> My understanding is that in practice, this likely mostly has the
> effect that these lines will be printed with an unintended delay.

Yes, that sounds right.

> I'm not going to send a patch for this, I think the binder maintainers
> probably have a better idea of how/when to fix that in a way that
> won't lead to ugly merge conflicts somewhere.

Thanks for reporting this. I think all that needs to happen is a patch
fixing them, and we backport it whereever is required.

Alice