From: Vlastimil Babka <vbabka@suse.cz>
To: David Wang <00107082@163.com>,
akpm@linux-foundation.org, surenb@google.com, mhocko@suse.com,
jackmanb@google.com, hannes@cmpxchg.org, ziy@nvidia.com
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3] mm/codetag: move tag retrieval back upfront in __free_pages()
Date: Tue, 6 May 2025 09:58:15 +0200 [thread overview]
Message-ID: <a8fd8cb3-c3e5-48be-a034-7a5c9f3ce778@suse.cz> (raw)
In-Reply-To: <20250505193034.91682-1-00107082@163.com>
On 5/5/25 21:30, David Wang wrote:
> Commit 51ff4d7486f0 ("mm: avoid extra mem_alloc_profiling_enabled()
> checks") introduces a possible use-after-free scenario, when page
> is non-compound, page[0] could be released by other thread right
> after put_page_testzero failed in current thread, pgalloc_tag_sub_pages
> afterwards would manipulate an invalid page for accounting remaining
> pages:
>
> [timeline] [thread1] [thread2]
> | alloc_page non-compound
> V
> | get_page, rf counter inc
> V
> | in ___free_pages
> | put_page_testzero fails
> V
> | put_page, page released
> V
> | in ___free_pages,
> | pgalloc_tag_sub_pages
> | manipulate an invalid page
> V
>
> Restore __free_pages() to its state before, retrieve alloc tag
> beforehand.
>
> Fixes: 51ff4d7486f0 ("mm: avoid extra mem_alloc_profiling_enabled() checks")
That's a 6.15-rc1 commit, so no cc: stable needed, but should go to mm-hotfixes.
> Signed-off-by: David Wang <00107082@163.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Thanks!
> ---
> include/linux/pgalloc_tag.h | 8 ++++++++
> mm/page_alloc.c | 15 ++++++---------
> 2 files changed, 14 insertions(+), 9 deletions(-)
>
> diff --git a/include/linux/pgalloc_tag.h b/include/linux/pgalloc_tag.h
> index c74077977830..8a7f4f802c57 100644
> --- a/include/linux/pgalloc_tag.h
> +++ b/include/linux/pgalloc_tag.h
> @@ -188,6 +188,13 @@ static inline struct alloc_tag *__pgalloc_tag_get(struct page *page)
> return tag;
> }
>
> +static inline struct alloc_tag *pgalloc_tag_get(struct page *page)
> +{
> + if (mem_alloc_profiling_enabled())
> + return __pgalloc_tag_get(page);
> + return NULL;
> +}
> +
> void pgalloc_tag_split(struct folio *folio, int old_order, int new_order);
> void pgalloc_tag_swap(struct folio *new, struct folio *old);
>
> @@ -199,6 +206,7 @@ static inline void clear_page_tag_ref(struct page *page) {}
> static inline void alloc_tag_sec_init(void) {}
> static inline void pgalloc_tag_split(struct folio *folio, int old_order, int new_order) {}
> static inline void pgalloc_tag_swap(struct folio *new, struct folio *old) {}
> +static inline struct alloc_tag *pgalloc_tag_get(struct page *page) { return NULL; }
>
> #endif /* CONFIG_MEM_ALLOC_PROFILING */
>
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 5669baf2a6fe..1b00e14a9780 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -1151,14 +1151,9 @@ static inline void pgalloc_tag_sub(struct page *page, unsigned int nr)
> __pgalloc_tag_sub(page, nr);
> }
>
> -static inline void pgalloc_tag_sub_pages(struct page *page, unsigned int nr)
> +/* When tag is not NULL, assuming mem_alloc_profiling_enabled */
> +static inline void pgalloc_tag_sub_pages(struct alloc_tag *tag, unsigned int nr)
> {
> - struct alloc_tag *tag;
> -
> - if (!mem_alloc_profiling_enabled())
> - return;
> -
> - tag = __pgalloc_tag_get(page);
> if (tag)
> this_cpu_sub(tag->counters->bytes, PAGE_SIZE * nr);
> }
> @@ -1168,7 +1163,7 @@ static inline void pgalloc_tag_sub_pages(struct page *page, unsigned int nr)
> static inline void pgalloc_tag_add(struct page *page, struct task_struct *task,
> unsigned int nr) {}
> static inline void pgalloc_tag_sub(struct page *page, unsigned int nr) {}
> -static inline void pgalloc_tag_sub_pages(struct page *page, unsigned int nr) {}
> +static inline void pgalloc_tag_sub_pages(struct alloc_tag *tag, unsigned int nr) {}
>
> #endif /* CONFIG_MEM_ALLOC_PROFILING */
>
> @@ -5065,11 +5060,13 @@ static void ___free_pages(struct page *page, unsigned int order,
> {
> /* get PageHead before we drop reference */
> int head = PageHead(page);
> + /* get alloc tag in case the page is released by others */
> + struct alloc_tag *tag = pgalloc_tag_get(page);
>
> if (put_page_testzero(page))
> __free_frozen_pages(page, order, fpi_flags);
> else if (!head) {
> - pgalloc_tag_sub_pages(page, (1 << order) - 1);
> + pgalloc_tag_sub_pages(tag, (1 << order) - 1);
> while (order-- > 0)
> __free_frozen_pages(page + (1 << order), order,
> fpi_flags);
prev parent reply other threads:[~2025-05-06 7:58 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-04 6:19 [PATCH] mm/codetag: sub in advance when free non-compound high order pages David Wang
2025-05-05 13:12 ` Vlastimil Babka
2025-05-05 14:31 ` David Wang
2025-05-05 14:55 ` Vlastimil Babka
2025-05-05 15:33 ` Suren Baghdasaryan
2025-05-05 16:42 ` David Wang
2025-05-05 16:53 ` Suren Baghdasaryan
2025-05-05 18:34 ` [PATCH v2] mm/codetag: move tag retrieval back upfront in __free_pages() David Wang
2025-05-05 19:17 ` David Wang
2025-05-05 19:30 ` [PATCH v3] " David Wang
2025-05-05 20:32 ` Suren Baghdasaryan
2025-05-06 7:58 ` Vlastimil Babka [this message]
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=a8fd8cb3-c3e5-48be-a034-7a5c9f3ce778@suse.cz \
--to=vbabka@suse.cz \
--cc=00107082@163.com \
--cc=akpm@linux-foundation.org \
--cc=hannes@cmpxchg.org \
--cc=jackmanb@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=surenb@google.com \
--cc=ziy@nvidia.com \
/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