BPF List
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Zhen Ni" <zhen.ni@easystack.cn>
Cc: bpf@vger.kernel.org
Subject: Re: [PATCH 5/6] mm/page_owner: add bpf_page_owner_get_memcg_info() kfunc
Date: Fri, 09 Oct 2026 12:43:20 +0000	[thread overview]
Message-ID: <sashiko-outbox-165554@kernel.org> (raw)
In-Reply-To: <20261009112308.240769-6-zhen.ni@easystack.cn>

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Critical] bpf: Missing __sz annotation allows BPF stack overflow in memcg kfunc
- [High] bpf: missing __sz annotation on size param causes stack buffer overflow
--

commit 98bc7e204a48312a6be1bc67be9ade762ffed9ff
Author: Zhen Ni <zhen.ni@easystack.cn>

mm/page_owner: add bpf_page_owner_get_memcg_info() kfunc

This commit introduces a new BPF kfunc to expose per-page memcg information,
including the cgroup path. It extracts memcg data collection into a new
helper function that the kfunc calls directly.

> diff --git a/mm/page_owner_iter.c b/mm/page_owner_iter.c
> index 907517a291443..e7a8ca1e90e8d 100644
> --- a/mm/page_owner_iter.c
> +++ b/mm/page_owner_iter.c
[ ... ]
> @@ -170,4 +170,11 @@ __bpf_kfunc int bpf_page_owner_get_nid(struct page_owner_scan *scan)
>  	return page_owner_get_nid(scan->page);
>  }
>  
> +__bpf_kfunc int bpf_page_owner_get_memcg_info(struct page_owner_scan *scan,
> +					      struct memcg_info *info,
> +					      char *buf, u32 size)

[Severity: Critical]
Does this size parameter need the __sz suffix for the verifier to track
memory bounds correctly?

Because the parameter is named size rather than buf__sz, it appears the
verifier might treat buf as a fixed 1-byte allocation and size as an
unconstrained scalar.

Could this allow a BPF program to pass a 1-byte stack buffer alongside an
arbitrarily large size value? If so, when this calls into
get_page_memcg_info() in mm/page_owner.c and executes:

get_page_memcg_info()
    ...
    if (buf && size)
        cgroup_path(memcg->css.cgroup, buf, size);

can it overwrite the BPF stack and corrupt adjacent kernel memory?

[Severity: High]
Do the buf and info parameters also need the __uninit suffix?

Without this suffix, the verifier might improperly force BPF programs to
pre-initialize these output variables just to pass verification.

> +{
> +	return get_page_memcg_info(scan->page, info, buf, size);
> +}
> +
>  __bpf_kfunc_end_defs();

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261009112308.240769-1-zhen.ni@easystack.cn?part=5

  reply	other threads:[~2026-10-09 12:43 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-10-09 11:23 [PATCH 0/6] mm/page_owner: add bpf_iter and kfuncs Zhen Ni
2026-10-09 11:23 ` [PATCH 2/6] mm/page_owner: add bpf_iter target "page_owner" Zhen Ni
2026-10-09 11:50   ` sashiko-bot
2026-10-12  6:28     ` zhen.ni
2026-10-09 11:23 ` [PATCH 3/6] mm/page_owner: add open-coded page_owner iterator Zhen Ni
2026-10-09 12:00   ` sashiko-bot
2026-10-12  6:41     ` zhen.ni
2026-10-09 11:23 ` [PATCH 4/6] mm/page_owner: add bpf_page_owner_get_nid() kfunc Zhen Ni
2026-10-09 12:20   ` sashiko-bot
2026-10-12  7:01     ` zhen.ni
2026-10-09 11:23 ` [PATCH 5/6] mm/page_owner: add bpf_page_owner_get_memcg_info() kfunc Zhen Ni
2026-10-09 12:43   ` sashiko-bot [this message]
2026-10-12  7:09     ` zhen.ni
2026-10-09 11:23 ` [PATCH 6/6] selftests/bpf: add tests for the page_owner bpf_iter target Zhen Ni
2026-10-09 12:50   ` sashiko-bot
2026-10-12  7:15     ` zhen.ni
2026-10-09 13:48   ` bot+bpf-ci
     [not found] ` <20261009112308.240769-2-zhen.ni@easystack.cn>
2026-10-09 11:35   ` [PATCH 1/6] mm/page_owner: extract page_owner_next_eligible() from read loop sashiko-bot
2026-10-10  9:40     ` zhen.ni

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-165554@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=zhen.ni@easystack.cn \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox