Skip to content

Conversation

@luke-gruber
Copy link
Contributor

@luke-gruber luke-gruber commented Nov 6, 2025

We were seeing errors like:

```
* thread #8, stop reason = EXC_BAD_ACCESS (code=1, address=0x803)
  * frame #0: 0x00000001001fe944 ruby`rb_st_lookup(tab=0x00000000000007fb, key=1, value=0x00000001305b7490) at st.c:1066:22
    frame #1: 0x000000010002d658 ruby`remove_class_from_subclasses [inlined] class_get_subclasses_for_ns(tbl=0x00000000000007fb, ns_id=1) at class.c:604:9
    frame #2: 0x000000010002d650 ruby`remove_class_from_subclasses(tbl=0x00000000000007fb, ns_id=1, klass=4754039232) at class.c:620:34
    frame #3: 0x000000010002c8a8 ruby`rb_class_classext_free_subclasses(ext=0x000000011b5ce1d8, klass=4754039232, replacing=<unavailable>) at class.c:700:9
    frame #4: 0x000000010002c760 ruby`rb_class_classext_free(klass=4754039232, ext=0x000000011b5ce1d8, is_prime=true) at class.c:105:5
    frame #5: 0x00000001000e770c ruby`classext_free(ext=<unavailable>, is_prime=<unavailable>, namespace=<unavailable>, arg=<unavailable>) at gc.c:1231:5 [artificial]
    frame #6: 0x000000010002d178 ruby`rb_class_classext_foreach(klass=<unavailable>, func=(ruby`classext_free at gc.c:1228), arg=0x00000001305b75c0) at class.c:518:5
    frame #7: 0x00000001000e745c ruby`rb_gc_obj_free(objspace=0x000000012500c400, obj=4754039232) at gc.c:1282:9
    frame #8: 0x00000001000e70d4 ruby`gc_sweep_plane(objspace=0x000000012500c400, heap=<unavailable>, p=4754039232, bitset=4095, ctx=0x00000001305b76e8) at default.c:3482:21
    frame #9: 0x00000001000e6e9c ruby`gc_sweep_page(objspace=0x000000012500c400, heap=0x000000012500c540, ctx=0x00000001305b76e8) at default.c:3567:13
    frame #10: 0x00000001000e51d0 ruby`gc_sweep_step(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:3848:9
    frame #11: 0x00000001000e1880 ruby`gc_continue [inlined] gc_sweep_continue(objspace=0x000000012500c400, sweep_heap=0x000000012500c540) at default.c:3931:13
    frame #12: 0x00000001000e1754 ruby`gc_continue(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2037:9
    frame #13: 0x00000001000e10bc ruby`newobj_cache_miss [inlined] heap_prepare(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2056:5
    frame #14: 0x00000001000e1074 ruby`newobj_cache_miss [inlined] heap_next_free_page(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2280:9
    frame #15: 0x00000001000e106c ruby`newobj_cache_miss(objspace=0x000000012500c400, cache=0x0000600001b00300, heap_idx=2, vm_locked=false) at default.c:2387:38
    frame #16: 0x00000001000e0d28 ruby`newobj_alloc(objspace=<unavailable>, cache=<unavailable>, heap_idx=<unavailable>, vm_locked=<unavailable>) at default.c:2411:15 [artificial]
    frame #17: 0x00000001000d7214 ruby`newobj_of [inlined] rb_gc_impl_new_obj(objspace_ptr=<unavailable>, cache_ptr=<unavailable>, klass=<unavailable>, flags=<unavailable>, wb_protected=<unavailable>, alloc_size=<unavailable>) at default.c:2490:15
    frame #18: 0x00000001000d719c ruby`newobj_of(cr=<unavailable>, klass=4313971728, flags=258, wb_protected=<unavailable>, size=<unavailable>) at gc.c:995:17
    frame #19: 0x00000001000d73ec ruby`rb_wb_protected_newobj_of(ec=<unavailable>, klass=<unavailable>, flags=<unavailable>, size=<unavailable>) at gc.c:1044:12 [artificial]
    frame #20: 0x0000000100032d34 ruby`class_alloc0(type=<unavailable>, klass=4313971728, namespaceable=<unavailable>) at class.c:803:5
```

We were seeing errors like:

```
* thread ruby#8, stop reason = EXC_BAD_ACCESS (code=1, address=0x803)
  * frame #0: 0x00000001001fe944 ruby`rb_st_lookup(tab=0x00000000000007fb, key=1, value=0x00000001305b7490) at st.c:1066:22
    frame ruby#1: 0x000000010002d658 ruby`remove_class_from_subclasses [inlined] class_get_subclasses_for_ns(tbl=0x00000000000007fb, ns_id=1) at class.c:604:9
    frame ruby#2: 0x000000010002d650 ruby`remove_class_from_subclasses(tbl=0x00000000000007fb, ns_id=1, klass=4754039232) at class.c:620:34
    frame ruby#3: 0x000000010002c8a8 ruby`rb_class_classext_free_subclasses(ext=0x000000011b5ce1d8, klass=4754039232, replacing=<unavailable>) at class.c:700:9
    frame ruby#4: 0x000000010002c760 ruby`rb_class_classext_free(klass=4754039232, ext=0x000000011b5ce1d8, is_prime=true) at class.c:105:5
    frame ruby#5: 0x00000001000e770c ruby`classext_free(ext=<unavailable>, is_prime=<unavailable>, namespace=<unavailable>, arg=<unavailable>) at gc.c:1231:5 [artificial]
    frame ruby#6: 0x000000010002d178 ruby`rb_class_classext_foreach(klass=<unavailable>, func=(ruby`classext_free at gc.c:1228), arg=0x00000001305b75c0) at class.c:518:5
    frame ruby#7: 0x00000001000e745c ruby`rb_gc_obj_free(objspace=0x000000012500c400, obj=4754039232) at gc.c:1282:9
    frame ruby#8: 0x00000001000e70d4 ruby`gc_sweep_plane(objspace=0x000000012500c400, heap=<unavailable>, p=4754039232, bitset=4095, ctx=0x00000001305b76e8) at default.c:3482:21
    frame ruby#9: 0x00000001000e6e9c ruby`gc_sweep_page(objspace=0x000000012500c400, heap=0x000000012500c540, ctx=0x00000001305b76e8) at default.c:3567:13
    frame ruby#10: 0x00000001000e51d0 ruby`gc_sweep_step(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:3848:9
    frame ruby#11: 0x00000001000e1880 ruby`gc_continue [inlined] gc_sweep_continue(objspace=0x000000012500c400, sweep_heap=0x000000012500c540) at default.c:3931:13
    frame ruby#12: 0x00000001000e1754 ruby`gc_continue(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2037:9
    frame ruby#13: 0x00000001000e10bc ruby`newobj_cache_miss [inlined] heap_prepare(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2056:5
    frame ruby#14: 0x00000001000e1074 ruby`newobj_cache_miss [inlined] heap_next_free_page(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2280:9
    frame ruby#15: 0x00000001000e106c ruby`newobj_cache_miss(objspace=0x000000012500c400, cache=0x0000600001b00300, heap_idx=2, vm_locked=false) at default.c:2387:38
    frame ruby#16: 0x00000001000e0d28 ruby`newobj_alloc(objspace=<unavailable>, cache=<unavailable>, heap_idx=<unavailable>, vm_locked=<unavailable>) at default.c:2411:15 [artificial]
    frame ruby#17: 0x00000001000d7214 ruby`newobj_of [inlined] rb_gc_impl_new_obj(objspace_ptr=<unavailable>, cache_ptr=<unavailable>, klass=<unavailable>, flags=<unavailable>, wb_protected=<unavailable>, alloc_size=<unavailable>) at default.c:2490:15
    frame ruby#18: 0x00000001000d719c ruby`newobj_of(cr=<unavailable>, klass=4313971728, flags=258, wb_protected=<unavailable>, size=<unavailable>) at gc.c:995:17
    frame ruby#19: 0x00000001000d73ec ruby`rb_wb_protected_newobj_of(ec=<unavailable>, klass=<unavailable>, flags=<unavailable>, size=<unavailable>) at gc.c:1044:12 [artificial]
    frame #20: 0x0000000100032d34 ruby`class_alloc0(type=<unavailable>, klass=4313971728, namespaceable=<unavailable>) at class.c:803:5
```
@luke-gruber luke-gruber force-pushed the fix_ns_subclasses_refcount_with_ractors branch from 3757434 to b10d283 Compare November 6, 2025 17:51
@luke-gru luke-gru merged commit 2998c8d into ruby:master Nov 6, 2025
88 checks passed
Comment on lines +52 to +53
rb_atomic_t was = RUBY_ATOMIC_FETCH_SUB(ns_sub->refcount, 1);
if (was == 1) {
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Are we guaranteed not to be incrementing the refcount at the same time? If not we could hit a use-after-free

luke-gruber added a commit to luke-gruber/ruby that referenced this pull request Nov 11, 2025
This reverts commit 2998c8d.

We need to find a better way to fix this bug. Even with this refcount
change, errors were still being seen in CI. For now we need to remove
this failing test.
luke-gru pushed a commit that referenced this pull request Nov 11, 2025
…15138)

This reverts commit 2998c8d.

We need to find a better way to fix this bug. Even with this refcount
change, errors were still being seen in CI. For now we need to remove
this failing test.
jacob-shops pushed a commit to Shopify/ruby that referenced this pull request Nov 12, 2025
We were seeing errors like:

```
* thread #8, stop reason = EXC_BAD_ACCESS (code=1, address=0x803)
  * frame #0: 0x00000001001fe944 ruby`rb_st_lookup(tab=0x00000000000007fb, key=1, value=0x00000001305b7490) at st.c:1066:22
    frame #1: 0x000000010002d658 ruby`remove_class_from_subclasses [inlined] class_get_subclasses_for_ns(tbl=0x00000000000007fb, ns_id=1) at class.c:604:9
    frame #2: 0x000000010002d650 ruby`remove_class_from_subclasses(tbl=0x00000000000007fb, ns_id=1, klass=4754039232) at class.c:620:34
    frame #3: 0x000000010002c8a8 ruby`rb_class_classext_free_subclasses(ext=0x000000011b5ce1d8, klass=4754039232, replacing=<unavailable>) at class.c:700:9
    frame #4: 0x000000010002c760 ruby`rb_class_classext_free(klass=4754039232, ext=0x000000011b5ce1d8, is_prime=true) at class.c:105:5
    frame #5: 0x00000001000e770c ruby`classext_free(ext=<unavailable>, is_prime=<unavailable>, namespace=<unavailable>, arg=<unavailable>) at gc.c:1231:5 [artificial]
    frame #6: 0x000000010002d178 ruby`rb_class_classext_foreach(klass=<unavailable>, func=(ruby`classext_free at gc.c:1228), arg=0x00000001305b75c0) at class.c:518:5
    frame #7: 0x00000001000e745c ruby`rb_gc_obj_free(objspace=0x000000012500c400, obj=4754039232) at gc.c:1282:9
    frame #8: 0x00000001000e70d4 ruby`gc_sweep_plane(objspace=0x000000012500c400, heap=<unavailable>, p=4754039232, bitset=4095, ctx=0x00000001305b76e8) at default.c:3482:21
    frame #9: 0x00000001000e6e9c ruby`gc_sweep_page(objspace=0x000000012500c400, heap=0x000000012500c540, ctx=0x00000001305b76e8) at default.c:3567:13
    frame #10: 0x00000001000e51d0 ruby`gc_sweep_step(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:3848:9
    frame #11: 0x00000001000e1880 ruby`gc_continue [inlined] gc_sweep_continue(objspace=0x000000012500c400, sweep_heap=0x000000012500c540) at default.c:3931:13
    frame #12: 0x00000001000e1754 ruby`gc_continue(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2037:9
    frame #13: 0x00000001000e10bc ruby`newobj_cache_miss [inlined] heap_prepare(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2056:5
    frame #14: 0x00000001000e1074 ruby`newobj_cache_miss [inlined] heap_next_free_page(objspace=0x000000012500c400, heap=0x000000012500c540) at default.c:2280:9
    frame #15: 0x00000001000e106c ruby`newobj_cache_miss(objspace=0x000000012500c400, cache=0x0000600001b00300, heap_idx=2, vm_locked=false) at default.c:2387:38
    frame #16: 0x00000001000e0d28 ruby`newobj_alloc(objspace=<unavailable>, cache=<unavailable>, heap_idx=<unavailable>, vm_locked=<unavailable>) at default.c:2411:15 [artificial]
    frame #17: 0x00000001000d7214 ruby`newobj_of [inlined] rb_gc_impl_new_obj(objspace_ptr=<unavailable>, cache_ptr=<unavailable>, klass=<unavailable>, flags=<unavailable>, wb_protected=<unavailable>, alloc_size=<unavailable>) at default.c:2490:15
    frame #18: 0x00000001000d719c ruby`newobj_of(cr=<unavailable>, klass=4313971728, flags=258, wb_protected=<unavailable>, size=<unavailable>) at gc.c:995:17
    frame #19: 0x00000001000d73ec ruby`rb_wb_protected_newobj_of(ec=<unavailable>, klass=<unavailable>, flags=<unavailable>, size=<unavailable>) at gc.c:1044:12 [artificial]
    frame #20: 0x0000000100032d34 ruby`class_alloc0(type=<unavailable>, klass=4313971728, namespaceable=<unavailable>) at class.c:803:5
```
jacob-shops pushed a commit to Shopify/ruby that referenced this pull request Nov 12, 2025
…" (ruby#15138)

This reverts commit 2998c8d.

We need to find a better way to fix this bug. Even with this refcount
change, errors were still being seen in CI. For now we need to remove
this failing test.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants