On Thu, 28 May 2026 00:11:44 GMT, Coleen Phillimore <[email protected]> wrote:

> Please review this mostly trivial change to clean up comments, and revert 
> some unneeded changes for valhalla.  I removed flatArrayOop.cpp because it 
> was empty.
> 
> Tested with tier1-3, tier1 testing on macosx-aarch64 in progress.
> 
> ---------
> - [x] I confirm that I make this contribution in accordance with the [OpenJDK 
> Interim AI Policy](https://openjdk.org/legal/ai).

Marked as reviewed by stefank (Committer).

src/hotspot/share/oops/instanceKlass.cpp line 575:

> 573: 
> 574: #ifdef ASSERT
> 575: void InstanceKlass::assert_bounds_check(address addr, bool edge_ok, 
> intptr_t size_in_bytes) const {

1) This function is broken w.r.t. bytes vs words in the "just after" check.
2) The "before" checks is turned off for the checks of "start_" values. That 
seems inconsistent and also makes me wonder if the before check really is 
helping here.

I would vote for nuking this function from JEP-401, but as a separate RFE.

src/hotspot/share/oops/instanceKlass.hpp line 1070:

> 1068:   inline intptr_t* end_of_itable() const;
> 1069:   inline oop static_field_base_raw();
> 1070:   void assert_bounds_check(address addr, bool edge_ok = false, intptr_t 
> size_in_bytes = -1) const;

It's a little bit surprising that this doesn't have the `NOT_DEBUG_RETURN`. If 
product code calls this we will now get a linker error instead of a compiler 
error.

-------------

PR Review: 
https://git.openjdk.org/valhalla/pull/2479#pullrequestreview-4378405817
PR Review Comment: 
https://git.openjdk.org/valhalla/pull/2479#discussion_r3315793032
PR Review Comment: 
https://git.openjdk.org/valhalla/pull/2479#discussion_r3315766394

Reply via email to