From patchwork Mon Aug 12 09:55:17 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Tianchen Ding X-Patchwork-Id: 13760379 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by smtp.lore.kernel.org (Postfix) with ESMTP id 6CA4EC3DA7F for ; Mon, 12 Aug 2024 09:55:33 +0000 (UTC) Received: by kanga.kvack.org (Postfix) id CDD006B008A; Mon, 12 Aug 2024 05:55:32 -0400 (EDT) Received: by kanga.kvack.org (Postfix, from userid 40) id C63F66B008C; Mon, 12 Aug 2024 05:55:32 -0400 (EDT) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id B04426B0092; Mon, 12 Aug 2024 05:55:32 -0400 (EDT) X-Delivered-To: linux-mm@kvack.org Received: from relay.hostedemail.com (smtprelay0016.hostedemail.com [216.40.44.16]) by kanga.kvack.org (Postfix) with ESMTP id 8ED136B008A for ; Mon, 12 Aug 2024 05:55:32 -0400 (EDT) Received: from smtpin11.hostedemail.com (a10.router.float.18 [10.200.18.1]) by unirelay03.hostedemail.com (Postfix) with ESMTP id EB829A070D for ; Mon, 12 Aug 2024 09:55:31 +0000 (UTC) X-FDA: 82443136062.11.D7744C9 Received: from out30-131.freemail.mail.aliyun.com (out30-131.freemail.mail.aliyun.com [115.124.30.131]) by imf24.hostedemail.com (Postfix) with ESMTP id 084D3180012 for ; Mon, 12 Aug 2024 09:55:28 +0000 (UTC) Authentication-Results: imf24.hostedemail.com; dkim=pass header.d=linux.alibaba.com header.s=default header.b=IUNA8MZf; dmarc=pass (policy=none) header.from=linux.alibaba.com; spf=pass (imf24.hostedemail.com: domain of dtcccc@linux.alibaba.com designates 115.124.30.131 as permitted sender) smtp.mailfrom=dtcccc@linux.alibaba.com ARC-Seal: i=1; s=arc-20220608; d=hostedemail.com; t=1723456479; a=rsa-sha256; cv=none; b=HZTudV8+13SqwL1YKd48PgsMXkbvzSg4hRJMA0BQh5aANdBJ1baaklMc/acGsVC9fOq+rZ A4wRj+jcOBy/1BzFfpTZQJdTJc5AZl3aWYVjm6rj60c3hdESajPeycgy0rY84UOaRfSl7/ cvAXknty0cV5VsaZKIfz50Qa/N6zi2I= ARC-Authentication-Results: i=1; imf24.hostedemail.com; dkim=pass header.d=linux.alibaba.com header.s=default header.b=IUNA8MZf; dmarc=pass (policy=none) header.from=linux.alibaba.com; spf=pass (imf24.hostedemail.com: domain of dtcccc@linux.alibaba.com designates 115.124.30.131 as permitted sender) smtp.mailfrom=dtcccc@linux.alibaba.com ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=hostedemail.com; s=arc-20220608; t=1723456479; h=from:from:sender:reply-to:subject:subject:date:date: message-id:message-id:to:to:cc:cc:mime-version:mime-version: content-type:content-transfer-encoding:content-transfer-encoding: in-reply-to:references:dkim-signature; bh=nZqkS7Z96TWLWnJVjQ8HLYQ0F4DVw3Kyu1YxwZNjc/M=; b=P8SeQSsa/7t3LYzkNgLI9KAXspXIwwtwl0iZacoNEweVVkZFvnyStAJZi/HlnMrEqcUUMd tKV7WRkcrqSeredlS3iztjGcZ8qwGMQihc5Mw98ZFd4JWoriRgU71bW9NGWv2PITRrfC1s rcJxxwUap1vD3tpYhO4Mn1i3eDLHDSo= DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.alibaba.com; s=default; t=1723456524; h=From:To:Subject:Date:Message-Id:MIME-Version; bh=nZqkS7Z96TWLWnJVjQ8HLYQ0F4DVw3Kyu1YxwZNjc/M=; b=IUNA8MZfe/epapVjkhn5hzWIe9HxyiX7MuFPHd3O7TWHcKL45y19IFK4IuFmBLfeRpn9SoYO48O1d/RifVxzOa6wBUBrEe+IYhMC2A3Ul1FdtR3baR8vvFxhP4KYi+U33RftuyzSYRRliIkQMsjh51RJFHJJmk+06fJlbh3aF2U= Received: from localhost.localdomain(mailfrom:dtcccc@linux.alibaba.com fp:SMTPD_---0WChYhe1_1723456517) by smtp.aliyun-inc.com; Mon, 12 Aug 2024 17:55:23 +0800 From: Tianchen Ding To: linux-kernel@vger.kernel.org Cc: Marco Elver , Alexander Potapenko , Dmitry Vyukov , Andrew Morton , kasan-dev@googlegroups.com, linux-mm@kvack.org Subject: [PATCH v2] kfence: Save freeing stack trace at calling time instead of freeing time Date: Mon, 12 Aug 2024 17:55:17 +0800 Message-Id: <20240812095517.2357-1-dtcccc@linux.alibaba.com> X-Mailer: git-send-email 2.39.3 MIME-Version: 1.0 X-Rspamd-Queue-Id: 084D3180012 X-Rspam-User: X-Rspamd-Server: rspam05 X-Stat-Signature: bcbr81mp6s16qe5zn7pubre7sw3burz9 X-HE-Tag: 1723456528-694153 X-HE-Meta: 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 amjuHZ+D eHJLeusswzyDSrvfz3oINmmWxuCIvQcl/27foSHZobE5xeb4YbN6fvtKR2BGGd/wwoDrmvNin1Wt0yjiKW7KpkB/VrlVnfAaSa3cmIG4LZSRypCSlE20vCl0TXjTcZ4KjSpfSLFoI6ixEWMnRc6IVoi2Q1ZzYMsG35QvqELGfiLAatcb9+bgMAkz5GRJwxD/5/UOnSSAKnw6tHKSluzjb6GWvajQXlb04bqDjj/9qHpZ+hBr8UpwXarM2JAGQ/roUbyg9fD54+xgttWW2kvHtqDrxKomOwsFL/e6+ X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: List-Subscribe: List-Unsubscribe: For kmem_cache with SLAB_TYPESAFE_BY_RCU, the freeing trace stack at calling kmem_cache_free() is more useful. While the following stack is meaningless and provides no help: freed by task 46 on cpu 0 at 656.840729s: rcu_do_batch+0x1ab/0x540 nocb_cb_wait+0x8f/0x260 rcu_nocb_cb_kthread+0x25/0x80 kthread+0xd2/0x100 ret_from_fork+0x34/0x50 ret_from_fork_asm+0x1a/0x30 Signed-off-by: Tianchen Ding Reviewed-by: Marco Elver --- v2: Rename and inline tiny helper kfence_obj_allocated(). Improve code style and comments. v1: https://lore.kernel.org/all/20240812065947.6104-1-dtcccc@linux.alibaba.com/ --- mm/kfence/core.c | 39 +++++++++++++++++++++++++++++---------- mm/kfence/kfence.h | 1 + mm/kfence/report.c | 7 ++++--- 3 files changed, 34 insertions(+), 13 deletions(-) diff --git a/mm/kfence/core.c b/mm/kfence/core.c index c3ef7eb8d4dc..67fc321db79b 100644 --- a/mm/kfence/core.c +++ b/mm/kfence/core.c @@ -273,6 +273,13 @@ static inline unsigned long metadata_to_pageaddr(const struct kfence_metadata *m return pageaddr; } +static inline bool kfence_obj_allocated(const struct kfence_metadata *meta) +{ + enum kfence_object_state state = READ_ONCE(meta->state); + + return state == KFENCE_OBJECT_ALLOCATED || state == KFENCE_OBJECT_RCU_FREEING; +} + /* * Update the object's metadata state, including updating the alloc/free stacks * depending on the state transition. @@ -282,10 +289,14 @@ metadata_update_state(struct kfence_metadata *meta, enum kfence_object_state nex unsigned long *stack_entries, size_t num_stack_entries) { struct kfence_track *track = - next == KFENCE_OBJECT_FREED ? &meta->free_track : &meta->alloc_track; + next == KFENCE_OBJECT_ALLOCATED ? &meta->alloc_track : &meta->free_track; lockdep_assert_held(&meta->lock); + /* Stack has been saved when calling rcu, skip. */ + if (READ_ONCE(meta->state) == KFENCE_OBJECT_RCU_FREEING) + goto out; + if (stack_entries) { memcpy(track->stack_entries, stack_entries, num_stack_entries * sizeof(stack_entries[0])); @@ -301,6 +312,7 @@ metadata_update_state(struct kfence_metadata *meta, enum kfence_object_state nex track->cpu = raw_smp_processor_id(); track->ts_nsec = local_clock(); /* Same source as printk timestamps. */ +out: /* * Pairs with READ_ONCE() in * kfence_shutdown_cache(), @@ -506,7 +518,7 @@ static void kfence_guarded_free(void *addr, struct kfence_metadata *meta, bool z raw_spin_lock_irqsave(&meta->lock, flags); - if (meta->state != KFENCE_OBJECT_ALLOCATED || meta->addr != (unsigned long)addr) { + if (!kfence_obj_allocated(meta) || meta->addr != (unsigned long)addr) { /* Invalid or double-free, bail out. */ atomic_long_inc(&counters[KFENCE_COUNTER_BUGS]); kfence_report_error((unsigned long)addr, false, NULL, meta, @@ -784,7 +796,7 @@ static void kfence_check_all_canary(void) for (i = 0; i < CONFIG_KFENCE_NUM_OBJECTS; i++) { struct kfence_metadata *meta = &kfence_metadata[i]; - if (meta->state == KFENCE_OBJECT_ALLOCATED) + if (kfence_obj_allocated(meta)) check_canary(meta); } } @@ -1010,12 +1022,11 @@ void kfence_shutdown_cache(struct kmem_cache *s) * the lock will not help, as different critical section * serialization will have the same outcome. */ - if (READ_ONCE(meta->cache) != s || - READ_ONCE(meta->state) != KFENCE_OBJECT_ALLOCATED) + if (READ_ONCE(meta->cache) != s || !kfence_obj_allocated(meta)) continue; raw_spin_lock_irqsave(&meta->lock, flags); - in_use = meta->cache == s && meta->state == KFENCE_OBJECT_ALLOCATED; + in_use = meta->cache == s && kfence_obj_allocated(meta); raw_spin_unlock_irqrestore(&meta->lock, flags); if (in_use) { @@ -1160,11 +1171,19 @@ void __kfence_free(void *addr) * the object, as the object page may be recycled for other-typed * objects once it has been freed. meta->cache may be NULL if the cache * was destroyed. + * Save the stack trace here so that reports show where the user freed + * the object. */ - if (unlikely(meta->cache && (meta->cache->flags & SLAB_TYPESAFE_BY_RCU))) + if (unlikely(meta->cache && (meta->cache->flags & SLAB_TYPESAFE_BY_RCU))) { + unsigned long flags; + + raw_spin_lock_irqsave(&meta->lock, flags); + metadata_update_state(meta, KFENCE_OBJECT_RCU_FREEING, NULL, 0); + raw_spin_unlock_irqrestore(&meta->lock, flags); call_rcu(&meta->rcu_head, rcu_guarded_free); - else + } else { kfence_guarded_free(addr, meta, false); + } } bool kfence_handle_page_fault(unsigned long addr, bool is_write, struct pt_regs *regs) @@ -1188,14 +1207,14 @@ bool kfence_handle_page_fault(unsigned long addr, bool is_write, struct pt_regs int distance = 0; meta = addr_to_metadata(addr - PAGE_SIZE); - if (meta && READ_ONCE(meta->state) == KFENCE_OBJECT_ALLOCATED) { + if (meta && kfence_obj_allocated(meta)) { to_report = meta; /* Data race ok; distance calculation approximate. */ distance = addr - data_race(meta->addr + meta->size); } meta = addr_to_metadata(addr + PAGE_SIZE); - if (meta && READ_ONCE(meta->state) == KFENCE_OBJECT_ALLOCATED) { + if (meta && kfence_obj_allocated(meta)) { /* Data race ok; distance calculation approximate. */ if (!to_report || distance > data_race(meta->addr) - addr) to_report = meta; diff --git a/mm/kfence/kfence.h b/mm/kfence/kfence.h index db87a05047bd..dfba5ea06b01 100644 --- a/mm/kfence/kfence.h +++ b/mm/kfence/kfence.h @@ -38,6 +38,7 @@ enum kfence_object_state { KFENCE_OBJECT_UNUSED, /* Object is unused. */ KFENCE_OBJECT_ALLOCATED, /* Object is currently allocated. */ + KFENCE_OBJECT_RCU_FREEING, /* Object was allocated, and then being freed by rcu. */ KFENCE_OBJECT_FREED, /* Object was allocated, and then freed. */ }; diff --git a/mm/kfence/report.c b/mm/kfence/report.c index 73a6fe42845a..451991a3a8f2 100644 --- a/mm/kfence/report.c +++ b/mm/kfence/report.c @@ -114,7 +114,8 @@ static void kfence_print_stack(struct seq_file *seq, const struct kfence_metadat /* Timestamp matches printk timestamp format. */ seq_con_printf(seq, "%s by task %d on cpu %d at %lu.%06lus (%lu.%06lus ago):\n", - show_alloc ? "allocated" : "freed", track->pid, + show_alloc ? "allocated" : meta->state == KFENCE_OBJECT_RCU_FREEING ? + "rcu freeing" : "freed", track->pid, track->cpu, (unsigned long)ts_sec, rem_nsec / 1000, (unsigned long)interval_nsec, rem_interval_nsec / 1000); @@ -149,7 +150,7 @@ void kfence_print_object(struct seq_file *seq, const struct kfence_metadata *met kfence_print_stack(seq, meta, true); - if (meta->state == KFENCE_OBJECT_FREED) { + if (meta->state == KFENCE_OBJECT_FREED || meta->state == KFENCE_OBJECT_RCU_FREEING) { seq_con_printf(seq, "\n"); kfence_print_stack(seq, meta, false); } @@ -318,7 +319,7 @@ bool __kfence_obj_info(struct kmem_obj_info *kpp, void *object, struct slab *sla kpp->kp_slab_cache = meta->cache; kpp->kp_objp = (void *)meta->addr; kfence_to_kp_stack(&meta->alloc_track, kpp->kp_stack); - if (meta->state == KFENCE_OBJECT_FREED) + if (meta->state == KFENCE_OBJECT_FREED || meta->state == KFENCE_OBJECT_RCU_FREEING) kfence_to_kp_stack(&meta->free_track, kpp->kp_free_stack); /* get_stack_skipnr() ensures the first entry is outside allocator. */ kpp->kp_ret = kpp->kp_stack[0];