Re: [PATCH slab/for-next 03/14] lib/xarray: specify SLAB_MAY_ACCOUNT for radix_tree_node cache

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] lib/xarray: undeclared SLAB_MAY_ACCOUNT in tools/include/linux/slab.h
--

commit 93954c4312f84244c94152321c79f4a0b093067d
Author: Harry Yoo (Meta) <harry@kernel.org>

lib/xarray: specify SLAB_MAY_ACCOUNT for radix_tree_node cache

This commit adds SLAB_MAY_ACCOUNT when creating the radix_tree_node slab
cache in radix_tree_init() so that nodes allocated with __GFP_ACCOUNT can be
charged to a memcg.

> diff --git a/lib/radix-tree.c b/lib/radix-tree.c
> index 976b9bd02a1b5..2c7531deb0df7 100644
> --- a/lib/radix-tree.c
> +++ b/lib/radix-tree.c
> @@ -1600,7 +1600,7 @@ void __init radix_tree_init(void)
>  	BUILD_BUG_ON(XA_CHUNK_SIZE > 255);
>  	radix_tree_node_cachep = kmem_cache_create("radix_tree_node",
>  			sizeof(struct radix_tree_node), 0,
> -			SLAB_PANIC | SLAB_RECLAIM_ACCOUNT,
> +			SLAB_PANIC | SLAB_RECLAIM_ACCOUNT | SLAB_MAY_ACCOUNT,

[Severity: Medium]
Does adding SLAB_MAY_ACCOUNT to radix_tree_init() in lib/radix-tree.c break
the userspace test suite builds under tools/testing/ (such as radix-tree,
rbtree, and vma)?

When building those test suites, tools/testing/shared/shared.mk generates
radix-tree.c from lib/radix-tree.c:

tools/testing/shared/shared.mk:
radix-tree.c: ../../../lib/radix-tree.c
	sed -e 's/^static //' -e 's/__always_inline //' -e 's/inline //' < $< > $@

and compiles radix-tree.o against tools/include/linux/slab.h rather than the
kernel's include/linux/slab.h.

However, tools/include/linux/slab.h does not define SLAB_MAY_ACCOUNT:

tools/include/linux/slab.h:
enum _slab_flag_bits {
	_SLAB_KMALLOC,
	_SLAB_HWCACHE_ALIGN,
	_SLAB_PANIC,
	_SLAB_TYPESAFE_BY_RCU,
	_SLAB_ACCOUNT,
	_SLAB_FLAGS_LAST_BIT
};

Should SLAB_MAY_ACCOUNT also be defined in tools/include/linux/slab.h so
that compiling radix-tree.o in the userspace test harness succeeds?

>  			radix_tree_node_ctor);
>  	ret = cpuhp_setup_state_nocalls(CPUHP_RADIX_DEAD, "lib/radix:dead",
>  					NULL, radix_tree_cpu_dead);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261021-memalloc-prof-dont-waste-slabobj-ext-metadata-v1-0-096800aa3330@kernel.org?part=3