From mboxrd@z Thu Jan 1 00:00:00 1970 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 bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id A1EE5CA0FFE for ; Tue, 2 Sep 2025 07:38:12 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender:List-Subscribe:List-Help :List-Post:List-Archive:List-Unsubscribe:List-Id:Content-Transfer-Encoding: Content-Type:Mime-Version:References:In-Reply-To:Message-Id:Subject:Cc:To: From:Date:Reply-To:Content-ID:Content-Description:Resent-Date:Resent-From: Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=XZSLYhdlIbnghXt6TwaEm7+yoO4aoLtOznhHPUnvjTU=; b=F5fQR1uBR2COLqVKBom27w4rv9 UlidcwWi1hiklIvilG6eZJTlciabNh9oifcYfE4L1ZCZJ4ndeijUUlrIXMqgL6ziVdEFtCu6NiSRY KZALOcuVTH7Zjne+IK7CsOivLKk6SLQcB97K9aX4TkwxuEaBm+wY54C1x+7Cp8b/DDbw/EItweN5f c0q1OLjparONbQkEhkkdAjaq7xN7VlidO4Tj0cDi9Hg5ATc+UOFq+rCys4jq9JiliRukhirpBUA5o dCQv48Wc85QCtnMkFIw5FEFzloKrHm67iDWrsH3D3iWiOTxBLjgYrgIKl4Jzr5w3ixhWy4UKd/uoO 0BoMjw2w==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.98.2 #2 (Red Hat Linux)) id 1utLaf-0000000Fms1-1QKS; Tue, 02 Sep 2025 07:38:05 +0000 Received: from sea.source.kernel.org ([2600:3c0a:e001:78e:0:1991:8:25]) by bombadil.infradead.org with esmtps (Exim 4.98.2 #2 (Red Hat Linux)) id 1utLXy-0000000FmEG-02XU for linux-arm-kernel@lists.infradead.org; Tue, 02 Sep 2025 07:35:19 +0000 Received: from smtp.kernel.org (transwarp.subspace.kernel.org [100.75.92.58]) by sea.source.kernel.org (Postfix) with ESMTP id 9CA704348F; Tue, 2 Sep 2025 07:35:17 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 32F30C4CEED; Tue, 2 Sep 2025 07:35:16 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1756798517; bh=E/WY5o6a97fVHU8e5uy2CoWMRAn7yHxp4l1Tt05M7pQ=; h=Date:From:To:Cc:Subject:In-Reply-To:References:From; b=nymPyFHaAcgHY3kjoLFAWTfg/VeqM5171gZw/UhNyg9PSulFo70S9+Uf89ywYHI2F TP0kj0tj7CVr8uk31H7r6i8kgwp/Vz/lIMQJfsFXZ+qe5QdumAEUtB7oaB6fWS30oA ba0B1Z4l66xfrJs0oNuIbiVCvOJzTxdFntl4jR8aunxFyTXYVkITEOMVs9zZvzFT9I BksleDrnd05NzHesNhvNdLu7bpN3EkUHHVKhhYu27OCgSambrPhwWIieAEK8n3eQys 9QgBDGPKGDm6KMWAW2efFr5R/C+a3M3oSggjnqSZSBuK2CHxWS5GxKmt453CYagsKq 0tCjlMKVdJOnA== Date: Tue, 2 Sep 2025 16:35:14 +0900 From: Masami Hiramatsu (Google) To: Luo Gengkun Cc: Steven Rostedt , mathieu.desnoyers@efficios.com, linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org, Catalin Marinas , Will Deacon , linux-arm-kernel@lists.infradead.org, Mark Rutland Subject: Re: [PATCH] tracing: Fix tracing_marker may trigger page fault during preempt_disable Message-Id: <20250902163514.f877d9c96e913f08c0c6b0b1@kernel.org> In-Reply-To: References: <20250819105152.2766363-1-luogengkun@huaweicloud.com> <20250819135008.5f1ba00e@gandalf.local.home> <436e4fa7-f8c7-4c23-a28a-4e5eebe2f854@huaweicloud.com> <20250829082604.1e3fd06e@gandalf.local.home> <20250902005645.8c6436b535731a4917745c5d@kernel.org> X-Mailer: Sylpheed 3.8.0beta1 (GTK+ 2.24.33; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20250902_003518_127397_08A3C471 X-CRM114-Status: GOOD ( 27.74 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org On Tue, 2 Sep 2025 11:47:32 +0800 Luo Gengkun wrote: > > On 2025/9/1 23:56, Masami Hiramatsu (Google) wrote: > > On Fri, 29 Aug 2025 08:26:04 -0400 > > Steven Rostedt wrote: > > > >> [ Adding arm64 maintainers ] > >> > >> On Fri, 29 Aug 2025 16:29:07 +0800 > >> Luo Gengkun wrote: > >> > >>> On 2025/8/20 1:50, Steven Rostedt wrote: > >>>> On Tue, 19 Aug 2025 10:51:52 +0000 > >>>> Luo Gengkun wrote: > >>>> > >>>>> Both tracing_mark_write and tracing_mark_raw_write call > >>>>> __copy_from_user_inatomic during preempt_disable. But in some case, > >>>>> __copy_from_user_inatomic may trigger page fault, and will call schedule() > >>>>> subtly. And if a task is migrated to other cpu, the following warning will > >>>> Wait! What? > >>>> > >>>> __copy_from_user_inatomic() is allowed to be called from in atomic context. > >>>> Hence the name it has. How the hell can it sleep? If it does, it's totally > >>>> broken! > >>>> > >>>> Now, I'm not against using nofault() as it is better named, but I want to > >>>> know why you are suggesting this change. Did you actually trigger a bug here? > >>> yes, I trigger this bug in arm64. > >> And I still think this is an arm64 bug. > > I think it could be. > > > >>>> > >>>>> be trigger: > >>>>> if (RB_WARN_ON(cpu_buffer, > >>>>> !local_read(&cpu_buffer->committing))) > >>>>> > >>>>> An example can illustrate this issue: > > You've missed an important part. > > > >>>>> process flow CPU > >>>>> --------------------------------------------------------------------- > >>>>> > >>>>> tracing_mark_raw_write(): cpu:0 > >>>>> ... > >>>>> ring_buffer_lock_reserve(): cpu:0 > >>>>> ... > > preempt_disable_notrace(); --> this is unlocked by ring_buffer_unlock_commit() > > > >>>>> cpu = raw_smp_processor_id() cpu:0 > >>>>> cpu_buffer = buffer->buffers[cpu] cpu:0 > >>>>> ... > >>>>> ... > >>>>> __copy_from_user_inatomic(): cpu:0 > > So this is called under preempt-disabled. > > > >>>>> ... > >>>>> # page fault > >>>>> do_mem_abort(): cpu:0 > >>>> Sounds to me that arm64 __copy_from_user_inatomic() may be broken. > >>>> > >>>>> ... > >>>>> # Call schedule > >>>>> schedule() cpu:0 > > If this does not check the preempt flag, it is a problem. > > Maybe arm64 needs to do fixup and abort instead of do_mem_abort()? > > My kernel was built without CONFIG_PREEMPT_COUNT, so the preempt_disable() > does nothing more than act as a barrier. In this case, it can pass the > check by schedule(). Perhaps this is another issue? OK, I got it. Indeed, in that case, we have no way to check this happens in the preempt critical section. Anyway, as in discussed here, __copy_from_user_inatomic() is for the internal function, so I'm also OK to this patch. Reviewed-by: Masami Hiramatsu (Google) BTW, currently we just write a fault message if the __copy_from_user_*() hits a fault, but I think we can retry with normal __copy_from_user() to a kernel buffer and copy it in the ring buffer as slow path. Thank you, -- Masami Hiramatsu (Google)