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 X-Spam-Level: X-Spam-Status: No, score=-7.7 required=3.0 tests=DATE_IN_PAST_12_24, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_PASS,URIBL_BLOCKED,USER_AGENT_MUTT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 0A768C43381 for ; Wed, 27 Mar 2019 14:57:06 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id CFB9F2082F for ; Wed, 27 Mar 2019 14:57:05 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1727488AbfC0O5F (ORCPT ); Wed, 27 Mar 2019 10:57:05 -0400 Received: from mx0b-001b2d01.pphosted.com ([148.163.158.5]:51448 "EHLO mx0a-001b2d01.pphosted.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1726469AbfC0O5F (ORCPT ); Wed, 27 Mar 2019 10:57:05 -0400 Received: from pps.filterd (m0098413.ppops.net [127.0.0.1]) by mx0b-001b2d01.pphosted.com (8.16.0.27/8.16.0.27) with SMTP id x2REuxiU021259 for ; Wed, 27 Mar 2019 10:57:03 -0400 Received: from e16.ny.us.ibm.com (e16.ny.us.ibm.com [129.33.205.206]) by mx0b-001b2d01.pphosted.com with ESMTP id 2rgb8frarj-1 (version=TLSv1.2 cipher=AES256-GCM-SHA384 bits=256 verify=NOT) for ; Wed, 27 Mar 2019 10:57:00 -0400 Received: from localhost by e16.ny.us.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Wed, 27 Mar 2019 14:55:44 -0000 Received: from b01cxnp23032.gho.pok.ibm.com (9.57.198.27) by e16.ny.us.ibm.com (146.89.104.203) with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted; (version=TLSv1/SSLv3 cipher=AES256-GCM-SHA384 bits=256/256) Wed, 27 Mar 2019 14:55:40 -0000 Received: from b01ledav003.gho.pok.ibm.com (b01ledav003.gho.pok.ibm.com [9.57.199.108]) by b01cxnp23032.gho.pok.ibm.com (8.14.9/8.14.9/NCO v10.0) with ESMTP id x2REtdHP18612292 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=OK); Wed, 27 Mar 2019 14:55:39 GMT Received: from b01ledav003.gho.pok.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id 464F7B206A; Wed, 27 Mar 2019 14:55:39 +0000 (GMT) Received: from b01ledav003.gho.pok.ibm.com (unknown [127.0.0.1]) by IMSVA (Postfix) with ESMTP id 1725FB2064; Wed, 27 Mar 2019 14:55:39 +0000 (GMT) Received: from paulmck-ThinkPad-W541 (unknown [9.70.82.188]) by b01ledav003.gho.pok.ibm.com (Postfix) with ESMTP; Wed, 27 Mar 2019 14:55:39 +0000 (GMT) Received: by paulmck-ThinkPad-W541 (Postfix, from userid 1000) id D233216C60F7; Tue, 26 Mar 2019 19:47:51 -0700 (PDT) Date: Tue, 26 Mar 2019 19:47:51 -0700 From: "Paul E. McKenney" To: "Joel Fernandes (Google)" Cc: linux-kernel@vger.kernel.org, byungchul.park@lge.com, kernel-team@android.com, rcu@vger.kernel.org, Ingo Molnar , Josh Triplett , Lai Jiangshan , linux-kselftest@vger.kernel.org, Mathieu Desnoyers , Peter Zijlstra , Shuah Khan , Steven Rostedt , Will Deacon Subject: Re: [PATCH v2 2/4] rcutree: Add checks for dynticks counters in rcu_is_cpu_rrupt_from_idle Reply-To: paulmck@linux.ibm.com References: <20190326192411.198070-1-joel@joelfernandes.org> <20190326192411.198070-2-joel@joelfernandes.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20190326192411.198070-2-joel@joelfernandes.org> User-Agent: Mutt/1.5.21 (2010-09-15) X-TM-AS-GCONF: 00 x-cbid: 19032714-0072-0000-0000-000004111730 X-IBM-SpamModules-Scores: X-IBM-SpamModules-Versions: BY=3.00010823; HX=3.00000242; KW=3.00000007; PH=3.00000004; SC=3.00000282; SDB=6.01180445; UDB=6.00617760; IPR=6.00961162; MB=3.00026180; MTD=3.00000008; XFM=3.00000015; UTC=2019-03-27 14:55:43 X-IBM-AV-DETECTION: SAVI=unused REMOTE=unused XFE=unused x-cbparentid: 19032714-0073-0000-0000-00004B9F285F Message-Id: <20190327024751.GV4102@linux.ibm.com> X-Proofpoint-Virus-Version: vendor=fsecure engine=2.50.10434:,, definitions=2019-03-27_09:,, signatures=0 X-Proofpoint-Spam-Details: rule=outbound_notspam policy=outbound score=0 priorityscore=1501 malwarescore=0 suspectscore=0 phishscore=0 bulkscore=0 spamscore=0 clxscore=1015 lowpriorityscore=0 mlxscore=0 impostorscore=0 mlxlogscore=999 adultscore=0 classifier=spam adjust=0 reason=mlx scancount=1 engine=8.0.1-1810050000 definitions=main-1903270105 Sender: rcu-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: rcu@vger.kernel.org On Tue, Mar 26, 2019 at 03:24:09PM -0400, Joel Fernandes (Google) wrote: > In the future we would like to combine the dynticks and dynticks_nesting > counters thus leading to simplifying the code. At the moment we cannot > do that due to concerns about usermode upcalls appearing to RCU as half > of an interrupt. Byungchul tried to do it in [1] but the > "half-interrupt" concern was raised. It is half because, what RCU > expects is rcu_irq_enter() and rcu_irq_exit() pairs when the usermode > exception happens. However, only rcu_irq_enter() is observed. This > concern may not be valid anymore, but at least it used to be the case. > > Out of abundance of caution, Paul added warnings [2] in the RCU code > which if not fired by 2021 may allow us to assume that such > half-interrupt scenario cannot happen any more, which can lead to > simplification of this code. > > Summary of the changes are the following: > > (1) In preparation for this combination of counters in the future, we > first need to first be sure that rcu_rrupt_from_idle cannot be called > from anywhere but a hard-interrupt because previously, the comments > suggested otherwise so let us be sure. We discussed this here [3]. We > use the services of lockdep to accomplish this. > > (2) Further rcu_rrupt_from_idle() is not explicit about how it is using > the counters which can lead to weird future bugs. This patch therefore > makes it more explicit about the specific counter values being tested > > (3) Lastly, we check for counter underflows just to be sure these are > not happening, because the previous code in rcu_rrupt_from_idle() was > allowing the case where the counters can underflow, and the function > would still return true. Now we are checking for specific values so let > us be confident by additional checking, that such underflows don't > happen. Any case, if they do, we should fix them and the screaming > warning is appropriate. All these checks checks are NOOPs if PROVE_RCU > and PROVE_LOCKING are disabled. > > [1] https://lore.kernel.org/patchwork/patch/952349/ > [2] Commit e11ec65cc8d6 ("rcu: Add warning to detect half-interrupts") > [3] https://lore.kernel.org/lkml/20190312150514.GB249405@google.com/ > > Cc: byungchul.park@lge.com > Cc: kernel-team@android.com > Cc: rcu@vger.kernel.org > Signed-off-by: Joel Fernandes (Google) Color me stupid: [ 48.845724] ------------[ cut here ]------------ [ 48.846619] Not in hardirq as expected [ 48.847322] WARNING: CPU: 5 PID: 34 at /home/git/linux-2.6-tip/kernel/rcu/tree.c:388 rcu_is_cpu_rrupt_from_idle+0xea/0x110 [ 48.849302] Modules linked in: [ 48.849869] CPU: 5 PID: 34 Comm: cpuhp/5 Not tainted 5.1.0-rc1+ #1 [ 48.850985] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 [ 48.852436] RIP: 0010:rcu_is_cpu_rrupt_from_idle+0xea/0x110 [ 48.853455] Code: 85 c0 0f 85 59 ff ff ff 80 3d 33 55 68 01 00 0f 85 4c ff ff ff 48 c7 c7 48 d8 cc 8e 31 c0 c6 05 1d 55 68 01 01 e8 66 54 f8 ff <0f> 0b e9 30 ff ff ff 65 48 8b 05 df 58 54 72 48 85 c0 0f 94 c0 0f [ 48.856783] RSP: 0000:ffffbc46802dfdc0 EFLAGS: 00010082 [ 48.857735] RAX: 000000000000001a RBX: 0000000000022b80 RCX: 0000000000000000 [ 48.859028] RDX: 0000000000000000 RSI: 0000000000000000 RDI: ffffffff8dac906c [ 48.860313] RBP: ffffbc46802dfe20 R08: 0000000000000001 R09: 0000000000000001 [ 48.861607] R10: 000000007d53d16d R11: ffffbc46802dfb48 R12: ffff9e7d7eb62b80 [ 48.862898] R13: 0000000000000005 R14: ffffffff8dae2ac0 R15: 00000000000000c9 [ 48.864191] FS: 0000000000000000(0000) GS:ffff9e7d7eb40000(0000) knlGS:0000000000000000 [ 48.865663] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 48.866702] CR2: 0000000000000000 CR3: 0000000021022000 CR4: 00000000000006e0 [ 48.867993] Call Trace: [ 48.868450] rcu_exp_handler+0x35/0x90 [ 48.869147] generic_exec_single+0xab/0x100 [ 48.869918] ? rcu_barrier+0x240/0x240 [ 48.870607] smp_call_function_single+0x8e/0xd0 [ 48.871441] rcutree_online_cpu+0x80/0x90 [ 48.872181] cpuhp_invoke_callback+0xb5/0x890 [ 48.872979] cpuhp_thread_fun+0x172/0x210 [ 48.873722] ? cpuhp_thread_fun+0x2a/0x210 [ 48.874474] smpboot_thread_fn+0x10d/0x160 [ 48.875224] kthread+0xf3/0x130 [ 48.875804] ? sort_range+0x20/0x20 [ 48.876446] ? kthread_cancel_delayed_work_sync+0x10/0x10 [ 48.877445] ret_from_fork+0x3a/0x50 [ 48.878124] irq event stamp: 734 [ 48.878717] hardirqs last enabled at (733): [] _raw_spin_unlock_irqrestore+0x2d/0x40 [ 48.880402] hardirqs last disabled at (734): [] generic_exec_single+0x9a/0x100 [ 48.881986] softirqs last enabled at (0): [] copy_process.part.56+0x61f/0x2110 [ 48.883540] softirqs last disabled at (0): [<0000000000000000>] (null) [ 48.884840] ---[ end trace 00b4c1d2f816f4ed ]--- If a CPU invokes generic_exec_single() on itself, the "IPI handler" will be invoked directly, triggering your new lockdep check. Which is a bit wasteful. My thought is to add code to sync_rcu_exp_select_node_cpus() to check the CPU with preemption disabled, avoiding the call to smp_call_function_single() in that case. I have queued all four of your patches, and am trying the fix to the caller of smp_call_function_single() shown below. Thoughts? Thanx, Paul ------------------------------------------------------------------------ diff --git a/kernel/rcu/tree_exp.h b/kernel/rcu/tree_exp.h index 9c990df880d1..51d61028abf1 100644 --- a/kernel/rcu/tree_exp.h +++ b/kernel/rcu/tree_exp.h @@ -384,7 +384,13 @@ static void sync_rcu_exp_select_node_cpus(struct work_struct *wp) mask_ofl_test |= mask; continue; } + preempt_disable(); + if (smp_processor_id() == cpu) { + preempt_enable(); + continue; + } ret = smp_call_function_single(cpu, rcu_exp_handler, NULL, 0); + preempt_enable(); if (!ret) { mask_ofl_ipi &= ~mask; continue;