From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 0E2EA35AC0C for ; Thu, 21 May 2026 23:18:52 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779405534; cv=none; b=bjsZP4pjakh+KewgQI46jCyM3ReFcQH3qShoPy8t0tOs1VRtC5qUlIVI9ygohd2iTB0OtwwknO8nK2HvSvt/ptwoSNTenWJMWLuKHWyQgBAMR5KJDRQaXef56aKUlk+YfgXbm47klRJmP4M+6oIoTJpnZzsER/A8rx0Fvb16GKs= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779405534; c=relaxed/simple; bh=yLRUCrH157D8vqHzvtCQcM8LmScE+5c6CaQY9OqCZR4=; h=Date:From:To:Cc:Subject:Message-Id:In-Reply-To:References: Mime-Version:Content-Type; b=OPFkKTBXnn7bwoLNUyxlaJtbaajkdIbrPaG6cg0d4WF6s031g84eVM2tqapGwrROl+A45uSU1Q1P5aDmaNy2SpphumZhkkojTpmY5hwJovLKqBHriN/PBJn8d1E9YorTFIjlO8/vP9hpqUfY2VauB8vsbcg/SNELrAex3oTvI50= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux-foundation.org header.i=@linux-foundation.org header.b=kpV1Qz5o; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux-foundation.org header.i=@linux-foundation.org header.b="kpV1Qz5o" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 828691F000E9; Thu, 21 May 2026 23:18:52 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux-foundation.org; s=korg; t=1779405532; bh=eeLE7vM+olzD9V5LtWw4maHM2ojb43LdYkLaD8Fd2Fg=; h=Date:From:To:Cc:Subject:In-Reply-To:References; b=kpV1Qz5oagA5O1PGFoQ2GdqvVDMqKwvgOoRroM0nJ6FK3/MReuFZ768ynSfdUzCN3 QX+D6137jDXHytBgKymrl3GmeEMRJJ1Avle+XVmLdCcRArV5U7rPu2VaLHu4Mju88u fALsKDTjdJAOAD7Zqy4418KWhH8pI4wPhhyPwsKk= Date: Thu, 21 May 2026 16:18:52 -0700 From: Andrew Morton To: Jann Horn Cc: Alexander Potapenko , Dmitry Vyukov , Andrey Konovalov , kasan-dev@googlegroups.com, linux-kernel@vger.kernel.org Subject: Re: [PATCH v2] kcov: update documentation on remote coverage collection Message-Id: <20260521161852.a26bab8401b8a100f9a5e29f@linux-foundation.org> In-Reply-To: <20260520-kcov-docs-v2-1-819f78778763@google.com> References: <20260520-kcov-docs-v2-1-819f78778763@google.com> X-Mailer: Sylpheed 3.8.0beta1 (GTK+ 2.24.33; x86_64-pc-linux-gnu) Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit On Wed, 20 May 2026 20:21:29 +0200 Jann Horn wrote: > Adjust the docs on remote coverage collection to reflect the changes made > in "kcov: refactor common handle ID into kcov_common_handle_id" and > "kcov: allow simultaneous KCOV_ENABLE/KCOV_REMOTE_ENABLE". This has all become rather confusing, with fixes and new versions of fixes under different titles, etc. I did my best to piece it all together then I joined everything into a single patch, below. Please check it for accuracy, up-to-dateness and changelog truthfulness, thanks. From: Jann Horn Subject: kcov: allow simultaneous KCOV_ENABLE/KCOV_REMOTE_ENABLE Date: Tue, 05 May 2026 11:00:46 +0200 Allow the same userspace thread to simultaneously collect normal coverage in syscall context (KCOV_ENABLE) and remote coverage of asynchronous work created by the thread (KCOV_REMOTE_ENABLE). With this, remote KCOV coverage becomes useful for generic fuzzing and not just fuzzing of specific data injection interfaces. This requires that the task_struct::kcov_* fields are separated into ones that are used by the task that generates coverage, and ones that are used by the task that requested remote coverage. To split this up: - Split task_struct::kcov into kcov and kcov_remote. kcov_task_exit() now has to clean up both separately. - Only use task_struct::kcov_mode on the task that generates coverage. - Only reset task_struct::kcov_handle on the task that requested remote coverage. After this change, fields used by the task that generates coverage are: - kcov_mode - kcov_size - kcov_area - kcov - kcov_sequence - kcov_softirq Fields used by the task that requested remote coverage are: - kcov_remote - kcov_handle [jannh@google.com: remove unused constant KCOV_MODE_REMOTE, per Dmitry] Link: https://lore.kernel.org/20260515-kcov-simultaneous-remote-v2-1-56fde1cfa509@google.com [jannh@google.com: update documentation on remote coverage collection] Link: https://lore.kernel.org/20260519-kcov-docs-v1-1-5bb22f4cb20c@google.com [jannh@google.com: move and reword sentence on simultaneous normal/remote collection Link: https://lore.kernel.org/20260520-kcov-docs-v2-1-819f78778763@google.com Link: https://lore.kernel.org/20260505-kcov-simultaneous-remote-v1-1-a670ba7cefd2@google.com Signed-off-by: Jann Horn Reviewed-by: Dmitry Vyukov Cc: Alexander Potapenko Cc: Andrey Konovalov Cc: Marco Elver Signed-off-by: Andrew Morton --- Documentation/dev-tools/kcov.rst | 6 + include/linux/kcov.h | 2 include/linux/sched.h | 3 kernel/kcov.c | 94 ++++++++++++++++------------- 4 files changed, 61 insertions(+), 44 deletions(-) --- a/include/linux/sched.h~kcov-allow-simultaneous-kcov_enable-kcov_remote_enable +++ a/include/linux/sched.h @@ -1517,6 +1517,9 @@ struct task_struct { /* KCOV descriptor wired with this task or NULL: */ struct kcov *kcov; + /* KCOV descriptor for remote coverage collection from other tasks: */ + struct kcov *kcov_remote; + /* KCOV common handle for remote coverage collection: */ u64 kcov_handle; --- a/kernel/kcov.c~kcov-allow-simultaneous-kcov_enable-kcov_remote_enable +++ a/kernel/kcov.c @@ -368,6 +368,7 @@ static void kcov_start(struct task_struc WRITE_ONCE(t->kcov_mode, mode); } +/* operates on coverage-generator-owned fields */ static void kcov_stop(struct task_struct *t) { WRITE_ONCE(t->kcov_mode, KCOV_MODE_DISABLED); @@ -377,16 +378,17 @@ static void kcov_stop(struct task_struct t->kcov_area = NULL; } +/* operates on coverage-generator-owned fields */ static void kcov_task_reset(struct task_struct *t) { kcov_stop(t); t->kcov_sequence = 0; - t->kcov_handle = 0; } void kcov_task_init(struct task_struct *t) { kcov_task_reset(t); + t->kcov_remote = NULL; t->kcov_handle = current->kcov_handle; } @@ -423,11 +425,14 @@ static void kcov_remote_reset(struct kco static void kcov_disable(struct task_struct *t, struct kcov *kcov) __must_hold(&kcov->lock) { - kcov_task_reset(t); - if (kcov->remote) + if (kcov->remote) { + t->kcov_handle = 0; + t->kcov_remote = NULL; kcov_remote_reset(kcov); - else + } else { + kcov_task_reset(t); kcov_reset(kcov); + } } static void kcov_get(struct kcov *kcov) @@ -453,41 +458,47 @@ void kcov_task_exit(struct task_struct * unsigned long flags; kcov = t->kcov; - if (kcov == NULL) - return; - - spin_lock_irqsave(&kcov->lock, flags); - kcov_debug("t = %px, kcov->t = %px\n", t, kcov->t); - /* - * For KCOV_ENABLE devices we want to make sure that t->kcov->t == t, - * which comes down to: - * WARN_ON(!kcov->remote && kcov->t != t); - * - * For KCOV_REMOTE_ENABLE devices, the exiting task is either: - * - * 1. A remote task between kcov_remote_start() and kcov_remote_stop(). - * In this case we should print a warning right away, since a task - * shouldn't be exiting when it's in a kcov coverage collection - * section. Here t points to the task that is collecting remote - * coverage, and t->kcov->t points to the thread that created the - * kcov device. Which means that to detect this case we need to - * check that t != t->kcov->t, and this gives us the following: - * WARN_ON(kcov->remote && kcov->t != t); - * - * 2. The task that created kcov exiting without calling KCOV_DISABLE, - * and then again we make sure that t->kcov->t == t: - * WARN_ON(kcov->remote && kcov->t != t); - * - * By combining all three checks into one we get: - */ - if (WARN_ON(kcov->t != t)) { + if (kcov) { + spin_lock_irqsave(&kcov->lock, flags); + kcov_debug("t = %px, kcov->t = %px\n", t, kcov->t); + /* + * This could be a remote task between kcov_remote_start() and + * kcov_remote_stop(). + * In this case we should print a warning right away, since a + * task shouldn't be exiting when it's in a kcov coverage + * collection section. + * + * Otherwise, this should be a task that created a local + * kcov instance and hasn't called KCOV_DISABLE. + * Make sure that t->kcov->t is consistent. + */ + if (WARN_ON(kcov->remote) || WARN_ON(kcov->t != t)) { + spin_unlock_irqrestore(&kcov->lock, flags); + return; + } + /* Just to not leave dangling references behind. */ + kcov_disable(t, kcov); spin_unlock_irqrestore(&kcov->lock, flags); - return; + kcov_put(kcov); + } + kcov = t->kcov_remote; + if (kcov) { + spin_lock_irqsave(&kcov->lock, flags); + kcov_debug("t = %px, kcov->t = %px\n", t, kcov->t); + /* + * This is a KCOV_REMOTE_ENABLE device, and the task is the + * user task which has requested remote coverage collection. + * Make sure that t->kcov->t is consistent. + */ + if (WARN_ON(!kcov->remote) || WARN_ON(kcov->t != t)) { + spin_unlock_irqrestore(&kcov->lock, flags); + return; + } + /* Just to not leave dangling references behind. */ + kcov_disable(t, kcov); + spin_unlock_irqrestore(&kcov->lock, flags); + kcov_put(kcov); } - /* Just to not leave dangling references behind. */ - kcov_disable(t, kcov); - spin_unlock_irqrestore(&kcov->lock, flags); - kcov_put(kcov); } static int kcov_mmap(struct file *filep, struct vm_area_struct *vma) @@ -629,9 +640,9 @@ static int kcov_ioctl_locked(struct kcov case KCOV_DISABLE: /* Disable coverage for the current task. */ unused = arg; - if (unused != 0 || current->kcov != kcov) - return -EINVAL; t = current; + if (unused != 0 || (kcov != t->kcov && kcov != t->kcov_remote)) + return -EINVAL; if (WARN_ON(kcov->t != t)) return -EINVAL; kcov_disable(t, kcov); @@ -641,7 +652,7 @@ static int kcov_ioctl_locked(struct kcov if (kcov->mode != KCOV_MODE_INIT || !kcov->area) return -EINVAL; t = current; - if (kcov->t != NULL || t->kcov != NULL) + if (kcov->t != NULL || t->kcov_remote != NULL) return -EBUSY; remote_arg = (struct kcov_remote_arg *)arg; mode = kcov_get_mode(remote_arg->trace_mode); @@ -651,8 +662,7 @@ static int kcov_ioctl_locked(struct kcov LONG_MAX / sizeof(unsigned long)) return -EINVAL; kcov->mode = mode; - t->kcov = kcov; - t->kcov_mode = KCOV_MODE_REMOTE; + t->kcov_remote = kcov; kcov->t = t; kcov->remote = true; kcov->remote_size = remote_arg->area_size; --- a/include/linux/kcov.h~kcov-allow-simultaneous-kcov_enable-kcov_remote_enable +++ a/include/linux/kcov.h @@ -21,8 +21,6 @@ enum kcov_mode { KCOV_MODE_TRACE_PC = 2, /* Collecting comparison operands mode. */ KCOV_MODE_TRACE_CMP = 3, - /* The process owns a KCOV remote reference. */ - KCOV_MODE_REMOTE = 4, }; #define KCOV_IN_CTXSW (1 << 30) --- a/Documentation/dev-tools/kcov.rst~kcov-allow-simultaneous-kcov_enable-kcov_remote_enable +++ a/Documentation/dev-tools/kcov.rst @@ -237,6 +237,9 @@ Both ``kcov_remote_start`` and ``kcov_re collection sections. The way a handle is used depends on the context where the matching code section executes. +A thread can use two separate KCOV instances to collect remote coverage and +normal coverage at the same time. + KCOV supports collecting remote coverage from the following contexts: 1. Global kernel background tasks. These are the tasks that are spawned during @@ -262,6 +265,9 @@ gets saved to the ``kcov_handle`` field needs to be passed to the newly spawned local tasks via custom kernel code modifications. Those tasks should in turn use the passed handle in their ``kcov_remote_start`` and ``kcov_remote_stop`` annotations. +In the kernel, common handles are wrapped in a ``kcov_common_handle_id``, which +consumes no space in builds without ``CONFIG_KCOV``; subsystems that integrate +with this mechanism should not need to use any ``#ifdef CONFIG_KCOV`` or such. KCOV follows a predefined format for both global and common handles. Each handle is a ``u64`` integer. Currently, only the one top and the lower 4 bytes _