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 F0354C2BA1A for ; Fri, 21 Jun 2024 10:12:37 +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:In-Reply-To:Content-Type: MIME-Version:References:Message-ID:Subject:Cc:To:From:Date:Reply-To: Content-Transfer-Encoding:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=1DWtkynomLva1GjC/L7HABHgy10osB8Eyz7PKb62PxA=; b=r8S7oAM4rGqUk0qmVUplW3SmoD LwrpKoTTTDoozEOcMXEp634HzFDYU+QBDJKBlFO0ebWBSc6F+zOPmzd7hNXkRywDhSvR/UvJtYxmt 2qSMmW/zfYM7dvhVx9TJWneNTFHXD/bTIBQBW91E16CO0tAhKG2PtJLL+NILFRYosRCaZct5SaGzT 50nCW1BqBoQ9RUiRos7WLD6bvLkh7Q402EgPkE6haI7EglQR8erTlw6PIihT+CXmm1XsFBQeJGqMe hBArF+ZDz/bxkrzyDKP0cQjaUetKSACCHRCNA2r4iTYMrROvp0Q5sm0bS6n4J1DCy5TsRtkPjs5sT HTT+keJw==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.97.1 #2 (Red Hat Linux)) id 1sKbFn-00000008hOi-3pTb; Fri, 21 Jun 2024 10:12:23 +0000 Received: from foss.arm.com ([217.140.110.172]) by bombadil.infradead.org with esmtp (Exim 4.97.1 #2 (Red Hat Linux)) id 1sKbFk-00000008hNR-1n9q for linux-arm-kernel@lists.infradead.org; Fri, 21 Jun 2024 10:12:21 +0000 Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 7695EDA7; Fri, 21 Jun 2024 03:12:39 -0700 (PDT) Received: from J2N7QTR9R3 (usa-sjc-imap-foss1.foss.arm.com [10.121.207.14]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id 9DC1D3F6A8; Fri, 21 Jun 2024 03:12:12 -0700 (PDT) Date: Fri, 21 Jun 2024 11:12:06 +0100 From: Mark Rutland To: Lingyue Cc: catalin.marinas@arm.com, will@kernel.org, dianders@chromium.org, swboyd@chromium.org, frederic@kernel.org, james.morse@arm.com, scott@os.amperecomputing.com, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, huangshaobo3@xiaomi.com, huangjun7@xiaomi.com Subject: Re: [PATCH] arm64: smp: do not allocate CPU IDs to invalid CPU nodes Message-ID: References: <20240621075045.249798-1-lingyue@xiaomi.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20240621075045.249798-1-lingyue@xiaomi.com> X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20240621_031220_589866_8186CFA8 X-CRM114-Status: GOOD ( 25.66 ) 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 Fri, Jun 21, 2024 at 03:50:45PM +0800, Lingyue wrote: > Many modules, such as arch topology, rely on num_possible_cpus() to > allocate memory and then access the allocated space using CPU IDs. > These modules assume that there are no gaps in cpu_possible_mask. Is there any documented requirement that cpu_possible_mask has no gaps? It looks like other architectures can have gaps in their cpu_possible_mask, there's no documented requiremetns AFAICT, and there are a bunch of commits handling cpu_possible_mask having gaps, e.g. bc75e99983df1efd ("rcu: Correctly handle sparse possible cpus") 3da43104d3187184 ("ARC: Adjust cpuinfo for non-continuous cpu ids") 72917235fd5f0863 ("tracing: Fix for non-continuous cpu ids") ... so I don't think that the topology code should assume that there are no gaps in cpu_possible_mask. > However, in of_parse_and_init_cpus(), CPU IDs are still allocated > for invalid CPU nodes, leading to gaps in cpu_possible_mask and > resulting in out-of-bounds memory access. So it is crucial to avoid > allocating CPU IDs to invalid CPU nodes. AFAICT the topology code could use 'nr_cpu_ids' instead of 'nr_possible_cpus()', like the tracing commit above, or it could use a per-cpu allocation to avoid this. > This issue can be reproduced easily on QEMU with KASAN enabled, by > modifing reg property of a CPU node to 0xFFFFFFFF > > [ 0.197756] BUG: KASAN: slab-out-of-bounds in topology_normalize_cpu_scale.part.0+0x2cc/0x34c > [ 0.199518] Read of size 4 at addr ffff000007ebe924 by task swapper/0/1 > [ 0.200087] > [ 0.200739] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 6.10.0-rc4 #3 > [ 0.201647] Hardware name: linux,dummy-virt (DT) > [ 0.203067] Call trace: > [ 0.203404] dump_backtrace+0x90/0xe8 > [ 0.203974] show_stack+0x18/0x24 > [ 0.204424] dump_stack_lvl+0x78/0x90 > [ 0.205090] print_report+0x114/0x5cc > [ 0.205908] kasan_report+0xa4/0xf0 > [ 0.206488] __asan_report_load4_noabort+0x20/0x2c > [ 0.207427] topology_normalize_cpu_scale.part.0+0x2cc/0x34c > [ 0.208275] init_cpu_topology+0x254/0x430 > [ 0.209518] smp_prepare_cpus+0x20/0x25c > [ 0.210824] kernel_init_freeable+0x1dc/0x4fc > [ 0.212047] kernel_init+0x24/0x1ec > [ 0.213143] ret_from_fork+0x10/0x20 > > Signed-off-by: Lingyue > --- > arch/arm64/kernel/smp.c | 6 +++--- > 1 file changed, 3 insertions(+), 3 deletions(-) > > diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c > index 31c8b3094dd7..5b4178145920 100644 > --- a/arch/arm64/kernel/smp.c > +++ b/arch/arm64/kernel/smp.c > @@ -638,12 +638,12 @@ static void __init of_parse_and_init_cpus(void) > u64 hwid = of_get_cpu_hwid(dn, 0); > > if (hwid & ~MPIDR_HWID_BITMASK) > - goto next; > + continue; > > if (is_mpidr_duplicate(cpu_count, hwid)) { > pr_err("%pOF: duplicate cpu reg properties in the DT\n", > dn); > - goto next; > + continue; > } > > /* > @@ -656,7 +656,7 @@ static void __init of_parse_and_init_cpus(void) > if (bootcpu_valid) { > pr_err("%pOF: duplicate boot cpu reg property in DT\n", > dn); > - goto next; > + continue; > } > People get very upset when CPU numbering changes, so I'd prefer to avoid this if possible. Mark. > bootcpu_valid = true; > -- > 2.34.1 >