From nobody Thu Oct 2 10:55:47 2025 Received: from mail-lf1-f74.google.com (mail-lf1-f74.google.com [209.85.167.74]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C98FF3195E9 for ; Thu, 18 Sep 2025 14:06:41 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.167.74 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1758204405; cv=none; b=O3fVlNhkGKHNb9hL9lVn46/qJO9ga5knDi4fFFngkPhfBTwlndcVWl4UHpFgxdApNx0mmDNCWSqJ99ralPs2dxeRT6HJok6/tNZudaxoIdgXYmK8Foz/frESfYCSNW2NPI0tR9k9edQIr04YQ4n01Y/Ok7VnEoaiJcO5rl7ezWQ= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1758204405; c=relaxed/simple; bh=UlJtETuiKnavOdqQ2uIKodawMPeKGzf4YyCRIIv6Cbk=; h=Date:In-Reply-To:Mime-Version:References:Message-ID:Subject:From: To:Cc:Content-Type; b=NxnRtOj57l9UgBRDjRPq0KjhDuJMcue/Rthu3urrIuat6CAgIfbBB3hQNuBduKXqW7viTDybTCeYtvOsIPdLKhfPvhYYrPbOjePn5hWJQcnirTa0JbTs66xPfW0rXj5ahX1MNqzHowxg7/VF7UZH4gInE/ys0HhHsWwwE2jDG3Y= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--elver.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=Xpm1ceXy; arc=none smtp.client-ip=209.85.167.74 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--elver.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="Xpm1ceXy" Received: by mail-lf1-f74.google.com with SMTP id 2adb3069b0e04-55f6f4dea68so679093e87.0 for ; Thu, 18 Sep 2025 07:06:41 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1758204400; x=1758809200; darn=vger.kernel.org; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:from:to:cc:subject:date:message-id:reply-to; bh=AbpPVjMmuER0ts34LQajNdvnhoooPemuy99n7ga7GVE=; b=Xpm1ceXyUbilm0RyFmA7kcavZIKS4EvCv2LKlAjf3tZUhkW2/k5hO+hPWHT5GVmOVn mFwIbI+ksJ+xb3QLaXCzyNmJGebZgBzJT3Q46IxJHsQBLTbbGDO/TQs0GNgDKdgjjKiz yIFbuqH3asRselqyWti++ZpXxmN7Ng7vzHyMkPOrS3TbK/77YfEWxYjSoR/OoI5mnzht v6lGohNJqndeaiYUYdMFBBzneHJu/o83MiFNJvrAUcAt/tFT72Z1Ijm50BT7Pd6CrJ/i F0VkTUuf/jDSuoRyWtUalv5t6XDXxh3foCAej+FVtSzLVhOOLvk9D01vbOXI0ginUoYd oMqw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1758204400; x=1758809200; h=cc:to:from:subject:message-id:references:mime-version:in-reply-to :date:x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=AbpPVjMmuER0ts34LQajNdvnhoooPemuy99n7ga7GVE=; b=JcemOR+Var9jWwmmwrAcyqsyXD7sgAF3xtDMCFE2zIhRzSUEM+p1DYjwD6gnTLmoEa cdpzYjiHhSFKSE9i4EmQ6nAbio9O+L3I5tlN3AuSkwqQ+D8nZ6lNge26P227M/orT+fU DZPDSRa4khO1rfGzm+NegEapGFMGpYYapzxWxbT275jd/aC/WqkKzy5h4hpetcsgVJ9Y JCKSOfjqZ8Fc9lxsQGUlESnmg3keoiP7OojAzNM/jcCaJsM6e4rQ84SYEMfndsydfDhS Sgqs8B6wsHjuUKyoFZJLWAbIjEi70gpJCdMH4Re2br0gI/4q+DghPyIuR07qHbfrof0l JO6w== X-Forwarded-Encrypted: i=1; AJvYcCXbC0c/0I3unbbIZPVqkFOHhscA92VltsMLgMMOyQgnB14bpLO56+ceUQkKgk5aV4eHSlnCPhnWuYi9EqU=@vger.kernel.org X-Gm-Message-State: AOJu0Yx962TIxuVwxhYOu5+Lmc/nV9cx4fyoH8Z0cJ0hIW32ZWF7mnk8 1BBjU3CPXcXGH+eOUjPQvLVGmDUFIHC+05wjMkQTFzVTXVH1gh8VDKzjPLr5zLdqeVFZDWK//Rf pAQ== X-Google-Smtp-Source: AGHT+IEWYcZtiY85Y5TBe4RKI/DknR4t9xSRYv397N/k2+9V5y5Th2sq2ql+LNq1WKRQLl1+RDCJRaV51A== X-Received: from lfby21.prod.google.com ([2002:a19:6415:0:b0:577:abf1:7dcc]) (user=elver job=prod-delivery.src-stubby-dispatcher) by 2002:a05:6512:4041:b0:579:c485:8704 with SMTP id 2adb3069b0e04-579c4858934mr81687e87.42.1758204400078; Thu, 18 Sep 2025 07:06:40 -0700 (PDT) Date: Thu, 18 Sep 2025 15:59:38 +0200 In-Reply-To: <20250918140451.1289454-1-elver@google.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 References: <20250918140451.1289454-1-elver@google.com> X-Mailer: git-send-email 2.51.0.384.g4c02a37b29-goog Message-ID: <20250918140451.1289454-28-elver@google.com> Subject: [PATCH v3 27/35] kfence: Enable capability analysis From: Marco Elver To: elver@google.com, Peter Zijlstra , Boqun Feng , Ingo Molnar , Will Deacon Cc: "David S. Miller" , Luc Van Oostenryck , "Paul E. McKenney" , Alexander Potapenko , Arnd Bergmann , Bart Van Assche , Bill Wendling , Christoph Hellwig , Dmitry Vyukov , Eric Dumazet , Frederic Weisbecker , Greg Kroah-Hartman , Herbert Xu , Ian Rogers , Jann Horn , Joel Fernandes , Jonathan Corbet , Josh Triplett , Justin Stitt , Kees Cook , Kentaro Takeda , Lukas Bulwahn , Mark Rutland , Mathieu Desnoyers , Miguel Ojeda , Nathan Chancellor , Neeraj Upadhyay , Nick Desaulniers , Steven Rostedt , Tetsuo Handa , Thomas Gleixner , Thomas Graf , Uladzislau Rezki , Waiman Long , kasan-dev@googlegroups.com, linux-crypto@vger.kernel.org, linux-doc@vger.kernel.org, linux-kbuild@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, linux-security-module@vger.kernel.org, linux-sparse@vger.kernel.org, llvm@lists.linux.dev, rcu@vger.kernel.org Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Enable capability analysis for the KFENCE subsystem. Notable, kfence_handle_page_fault() required minor restructure, which also fixed a subtle race; arguably that function is more readable now. Signed-off-by: Marco Elver --- v2: * Remove disable/enable_capability_analysis() around headers. * Use __capability_unsafe() instead of __no_capability_analysis. --- mm/kfence/Makefile | 2 ++ mm/kfence/core.c | 20 +++++++++++++------- mm/kfence/kfence.h | 14 ++++++++------ mm/kfence/report.c | 4 ++-- 4 files changed, 25 insertions(+), 15 deletions(-) diff --git a/mm/kfence/Makefile b/mm/kfence/Makefile index 2de2a58d11a1..b3640bdc3c69 100644 --- a/mm/kfence/Makefile +++ b/mm/kfence/Makefile @@ -1,5 +1,7 @@ # SPDX-License-Identifier: GPL-2.0 =20 +CAPABILITY_ANALYSIS :=3D y + obj-y :=3D core.o report.o =20 CFLAGS_kfence_test.o :=3D -fno-omit-frame-pointer -fno-optimize-sibling-ca= lls diff --git a/mm/kfence/core.c b/mm/kfence/core.c index 0ed3be100963..53b81eb5f31a 100644 --- a/mm/kfence/core.c +++ b/mm/kfence/core.c @@ -132,8 +132,8 @@ struct kfence_metadata *kfence_metadata __read_mostly; static struct kfence_metadata *kfence_metadata_init __read_mostly; =20 /* Freelist with available objects. */ -static struct list_head kfence_freelist =3D LIST_HEAD_INIT(kfence_freelist= ); -static DEFINE_RAW_SPINLOCK(kfence_freelist_lock); /* Lock protecting freel= ist. */ +DEFINE_RAW_SPINLOCK(kfence_freelist_lock); /* Lock protecting freelist. */ +static struct list_head kfence_freelist __guarded_by(&kfence_freelist_lock= ) =3D LIST_HEAD_INIT(kfence_freelist); =20 /* * The static key to set up a KFENCE allocation; or if static keys are not= used @@ -253,6 +253,7 @@ static bool kfence_unprotect(unsigned long addr) } =20 static inline unsigned long metadata_to_pageaddr(const struct kfence_metad= ata *meta) + __must_hold(&meta->lock) { unsigned long offset =3D (meta - kfence_metadata + 1) * PAGE_SIZE * 2; unsigned long pageaddr =3D (unsigned long)&__kfence_pool[offset]; @@ -288,6 +289,7 @@ static inline bool kfence_obj_allocated(const struct kf= ence_metadata *meta) static noinline void metadata_update_state(struct kfence_metadata *meta, enum kfence_object_sta= te next, unsigned long *stack_entries, size_t num_stack_entries) + __must_hold(&meta->lock) { struct kfence_track *track =3D next =3D=3D KFENCE_OBJECT_ALLOCATED ? &meta->alloc_track : &meta->free_t= rack; @@ -485,7 +487,7 @@ static void *kfence_guarded_alloc(struct kmem_cache *ca= che, size_t size, gfp_t g alloc_covered_add(alloc_stack_hash, 1); =20 /* Set required slab fields. */ - slab =3D virt_to_slab((void *)meta->addr); + slab =3D virt_to_slab(addr); slab->slab_cache =3D cache; slab->objects =3D 1; =20 @@ -514,6 +516,7 @@ static void *kfence_guarded_alloc(struct kmem_cache *ca= che, size_t size, gfp_t g static void kfence_guarded_free(void *addr, struct kfence_metadata *meta, = bool zombie) { struct kcsan_scoped_access assert_page_exclusive; + u32 alloc_stack_hash; unsigned long flags; bool init; =20 @@ -546,9 +549,10 @@ static void kfence_guarded_free(void *addr, struct kfe= nce_metadata *meta, bool z /* Mark the object as freed. */ metadata_update_state(meta, KFENCE_OBJECT_FREED, NULL, 0); init =3D slab_want_init_on_free(meta->cache); + alloc_stack_hash =3D meta->alloc_stack_hash; raw_spin_unlock_irqrestore(&meta->lock, flags); =20 - alloc_covered_add(meta->alloc_stack_hash, -1); + alloc_covered_add(alloc_stack_hash, -1); =20 /* Check canary bytes for memory corruption. */ check_canary(meta); @@ -593,6 +597,7 @@ static void rcu_guarded_free(struct rcu_head *h) * which partial initialization succeeded. */ static unsigned long kfence_init_pool(void) + __capability_unsafe(/* constructor */) { unsigned long addr; struct page *pages; @@ -1192,6 +1197,7 @@ bool kfence_handle_page_fault(unsigned long addr, boo= l is_write, struct pt_regs { const int page_index =3D (addr - (unsigned long)__kfence_pool) / PAGE_SIZ= E; struct kfence_metadata *to_report =3D NULL; + unsigned long unprotected_page =3D 0; enum kfence_error_type error_type; unsigned long flags; =20 @@ -1225,9 +1231,8 @@ bool kfence_handle_page_fault(unsigned long addr, boo= l is_write, struct pt_regs if (!to_report) goto out; =20 - raw_spin_lock_irqsave(&to_report->lock, flags); - to_report->unprotected_page =3D addr; error_type =3D KFENCE_ERROR_OOB; + unprotected_page =3D addr; =20 /* * If the object was freed before we took the look we can still @@ -1239,7 +1244,6 @@ bool kfence_handle_page_fault(unsigned long addr, boo= l is_write, struct pt_regs if (!to_report) goto out; =20 - raw_spin_lock_irqsave(&to_report->lock, flags); error_type =3D KFENCE_ERROR_UAF; /* * We may race with __kfence_alloc(), and it is possible that a @@ -1251,6 +1255,8 @@ bool kfence_handle_page_fault(unsigned long addr, boo= l is_write, struct pt_regs =20 out: if (to_report) { + raw_spin_lock_irqsave(&to_report->lock, flags); + to_report->unprotected_page =3D unprotected_page; kfence_report_error(addr, is_write, regs, to_report, error_type); raw_spin_unlock_irqrestore(&to_report->lock, flags); } else { diff --git a/mm/kfence/kfence.h b/mm/kfence/kfence.h index dfba5ea06b01..f9caea007246 100644 --- a/mm/kfence/kfence.h +++ b/mm/kfence/kfence.h @@ -34,6 +34,8 @@ /* Maximum stack depth for reports. */ #define KFENCE_STACK_DEPTH 64 =20 +extern raw_spinlock_t kfence_freelist_lock; + /* KFENCE object states. */ enum kfence_object_state { KFENCE_OBJECT_UNUSED, /* Object is unused. */ @@ -53,7 +55,7 @@ struct kfence_track { =20 /* KFENCE metadata per guarded allocation. */ struct kfence_metadata { - struct list_head list; /* Freelist node; access under kfence_freelist_lo= ck. */ + struct list_head list __guarded_by(&kfence_freelist_lock); /* Freelist no= de. */ struct rcu_head rcu_head; /* For delayed freeing. */ =20 /* @@ -91,13 +93,13 @@ struct kfence_metadata { * In case of an invalid access, the page that was unprotected; we * optimistically only store one address. */ - unsigned long unprotected_page; + unsigned long unprotected_page __guarded_by(&lock); =20 /* Allocation and free stack information. */ - struct kfence_track alloc_track; - struct kfence_track free_track; + struct kfence_track alloc_track __guarded_by(&lock); + struct kfence_track free_track __guarded_by(&lock); /* For updating alloc_covered on frees. */ - u32 alloc_stack_hash; + u32 alloc_stack_hash __guarded_by(&lock); #ifdef CONFIG_MEMCG struct slabobj_ext obj_exts; #endif @@ -141,6 +143,6 @@ enum kfence_error_type { void kfence_report_error(unsigned long address, bool is_write, struct pt_r= egs *regs, const struct kfence_metadata *meta, enum kfence_error_type type); =20 -void kfence_print_object(struct seq_file *seq, const struct kfence_metadat= a *meta); +void kfence_print_object(struct seq_file *seq, const struct kfence_metadat= a *meta) __must_hold(&meta->lock); =20 #endif /* MM_KFENCE_KFENCE_H */ diff --git a/mm/kfence/report.c b/mm/kfence/report.c index 10e6802a2edf..787e87c26926 100644 --- a/mm/kfence/report.c +++ b/mm/kfence/report.c @@ -106,6 +106,7 @@ static int get_stack_skipnr(const unsigned long stack_e= ntries[], int num_entries =20 static void kfence_print_stack(struct seq_file *seq, const struct kfence_m= etadata *meta, bool show_alloc) + __must_hold(&meta->lock) { const struct kfence_track *track =3D show_alloc ? &meta->alloc_track : &m= eta->free_track; u64 ts_sec =3D track->ts_nsec; @@ -207,8 +208,6 @@ void kfence_report_error(unsigned long address, bool is= _write, struct pt_regs *r if (WARN_ON(type !=3D KFENCE_ERROR_INVALID && !meta)) return; =20 - if (meta) - lockdep_assert_held(&meta->lock); /* * Because we may generate reports in printk-unfriendly parts of the * kernel, such as scheduler code, the use of printk() could deadlock. @@ -263,6 +262,7 @@ void kfence_report_error(unsigned long address, bool is= _write, struct pt_regs *r stack_trace_print(stack_entries + skipnr, num_stack_entries - skipnr, 0); =20 if (meta) { + lockdep_assert_held(&meta->lock); pr_err("\n"); kfence_print_object(NULL, meta); } --=20 2.51.0.384.g4c02a37b29-goog