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=-10.8 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED 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 8DAA6C433ED for ; Wed, 21 Apr 2021 22:01:03 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 6086F6142F for ; Wed, 21 Apr 2021 22:01:03 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235659AbhDUWBf (ORCPT ); Wed, 21 Apr 2021 18:01:35 -0400 Received: from us-smtp-delivery-124.mimecast.com ([170.10.133.124]:58870 "EHLO us-smtp-delivery-124.mimecast.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235570AbhDUWBX (ORCPT ); Wed, 21 Apr 2021 18:01:23 -0400 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1619042449; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=Nz6vYmnx1AF3iaZVPl6QF8XXcfKkHzh0encG7kXdRsI=; b=H7J8r3fba317XQBt7SN98vCD1NVRzaI/XQTeB/GBs0t9TCifrDAp/yOS4YmEavgd9G6SVN 9/1r+OCnPTYNld2lbgUud/SGnqrPu/2389ldKLQ4fC7tfEyp2LvkqFGiyhnGhlDqvUTR4u OOhUeXZYB0lEQBXrfgYwvwkk4i2SHSE= Received: from mail-ed1-f70.google.com (mail-ed1-f70.google.com [209.85.208.70]) (Using TLS) by relay.mimecast.com with ESMTP id us-mta-64-yN2bYQyCNgWtOZaX30CDhA-1; Wed, 21 Apr 2021 18:00:27 -0400 X-MC-Unique: yN2bYQyCNgWtOZaX30CDhA-1 Received: by mail-ed1-f70.google.com with SMTP id r14-20020a50d68e0000b0290385504d6e4eso5807668edi.7 for ; Wed, 21 Apr 2021 15:00:27 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:in-reply-to:references:date :message-id:mime-version:content-transfer-encoding; bh=Nz6vYmnx1AF3iaZVPl6QF8XXcfKkHzh0encG7kXdRsI=; b=l+H6mX6a41xDcnrlEjPrQQQK2jJ3KM1OyFVVElP9BOt5agZ3b9FzBt0kOeFd8D65fC 3FbYJosphQHxcVO1oHKUA5qdNjCfqxV6LAJMBcwU16YOaNwZEpx9r3pMmet1zqcsUhA0 jJJuRk44utB9PMZxsyIgfZzZchdYSwGvuWz0K0UowTCpT41sTgEnWOBE98gOuIzrVAXH mzjU905puA1WAl1p/3IqWViwsqNVbvmSyJz0k7BZO7DmUCbz5yMu6istYEbpMYvsfQsx 5LbM4DVl/9YwZisMvnUd7xLaWg6z8/kgzf1Xu5eD6vwIJLG9MBTErpKqzA1VAH6v5L3o ZV0w== X-Gm-Message-State: AOAM5307nFrm41c6TX5Wue45aayGBRQT/ic31YxaxP0bSjr4ueukVkW/ x+ugmC6/o9Ytd5WhVsTq+9813As1R9wOCMf0fP6IDfWgAmupbx5/Ysw2fQ5G8B12Jq3kK3zCJxL yX27HEIPseFCD X-Received: by 2002:a05:6402:5107:: with SMTP id m7mr82395edd.75.1619042426337; Wed, 21 Apr 2021 15:00:26 -0700 (PDT) X-Google-Smtp-Source: ABdhPJw6icC4l8bBtVy0rrIaRfzGROqWYIqgRSxkbaQSHdX8P8U6IjmiqR3l4lBRv7isiW1m/EkXiQ== X-Received: by 2002:a05:6402:5107:: with SMTP id m7mr82343edd.75.1619042425897; Wed, 21 Apr 2021 15:00:25 -0700 (PDT) Received: from alrua-x1.borgediget.toke.dk ([45.145.92.2]) by smtp.gmail.com with ESMTPSA id n11sm516891ejg.43.2021.04.21.15.00.25 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 21 Apr 2021 15:00:25 -0700 (PDT) Received: by alrua-x1.borgediget.toke.dk (Postfix, from userid 1000) id E13B4180675; Thu, 22 Apr 2021 00:00:24 +0200 (CEST) From: Toke =?utf-8?Q?H=C3=B8iland-J=C3=B8rgensen?= To: paulmck@kernel.org Cc: Martin KaFai Lau , Jesper Dangaard Brouer , Hangbin Liu , bpf@vger.kernel.org, netdev@vger.kernel.org, Jiri Benc , Eelco Chaudron , ast@kernel.org, Daniel Borkmann , Lorenzo Bianconi , David Ahern , Andrii Nakryiko , Alexei Starovoitov , John Fastabend , Maciej Fijalkowski , =?utf-8?B?QmrDtnJuIFQ=?= =?utf-8?B?w7ZwZWw=?= Subject: Re: [PATCHv7 bpf-next 1/4] bpf: run devmap xdp_prog on flush instead of bulk enqueue In-Reply-To: <20210421213019.GY975577@paulmck-ThinkPad-P17-Gen-1> References: <20210419183223.GC975577@paulmck-ThinkPad-P17-Gen-1> <877dkygeca.fsf@toke.dk> <20210419214117.GE975577@paulmck-ThinkPad-P17-Gen-1> <87y2ddgbsn.fsf@toke.dk> <20210419223143.GG975577@paulmck-ThinkPad-P17-Gen-1> <87wnsvhg0m.fsf@toke.dk> <20210421145951.GT975577@paulmck-ThinkPad-P17-Gen-1> <87r1j3h0hw.fsf@toke.dk> <20210421205146.GX975577@paulmck-ThinkPad-P17-Gen-1> <87im4fgx81.fsf@toke.dk> <20210421213019.GY975577@paulmck-ThinkPad-P17-Gen-1> X-Clacks-Overhead: GNU Terry Pratchett Date: Thu, 22 Apr 2021 00:00:24 +0200 Message-ID: <87bla7gux3.fsf@toke.dk> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: bpf@vger.kernel.org "Paul E. McKenney" writes: > On Wed, Apr 21, 2021 at 11:10:38PM +0200, Toke H=C3=B8iland-J=C3=B8rgense= n wrote: >> "Paul E. McKenney" writes: >>=20 >> > On Wed, Apr 21, 2021 at 09:59:55PM +0200, Toke H=C3=B8iland-J=C3=B8rge= nsen wrote: >> >> "Paul E. McKenney" writes: >> >>=20 >> >> > On Wed, Apr 21, 2021 at 04:24:41PM +0200, Toke H=C3=B8iland-J=C3=B8= rgensen wrote: >> >> >> "Paul E. McKenney" writes: >> >> >>=20 >> >> >> > On Tue, Apr 20, 2021 at 12:16:40AM +0200, Toke H=C3=B8iland-J=C3= =B8rgensen wrote: >> >> >> >> "Paul E. McKenney" writes: >> >> >> >>=20 >> >> >> >> > On Mon, Apr 19, 2021 at 11:21:41PM +0200, Toke H=C3=B8iland-J= =C3=B8rgensen wrote: >> >> >> >> >> "Paul E. McKenney" writes: >> >> >> >> >>=20 >> >> >> >> >> > On Mon, Apr 19, 2021 at 08:12:27PM +0200, Toke H=C3=B8ilan= d-J=C3=B8rgensen wrote: >> >> >> >> >> >> "Paul E. McKenney" writes: >> >> >> >> >> >>=20 >> >> >> >> >> >> > On Sat, Apr 17, 2021 at 02:27:19PM +0200, Toke H=C3=B8i= land-J=C3=B8rgensen wrote: >> >> >> >> >> >> >> "Paul E. McKenney" writes: >> >> >> >> >> >> >>=20 >> >> >> >> >> >> >> > On Fri, Apr 16, 2021 at 11:22:52AM -0700, Martin KaF= ai Lau wrote: >> >> >> >> >> >> >> >> On Fri, Apr 16, 2021 at 03:45:23PM +0200, Jesper Da= ngaard Brouer wrote: >> >> >> >> >> >> >> >> > On Thu, 15 Apr 2021 17:39:13 -0700 >> >> >> >> >> >> >> >> > Martin KaFai Lau wrote: >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > > On Thu, Apr 15, 2021 at 10:29:40PM +0200, Toke = H=C3=B8iland-J=C3=B8rgensen wrote: >> >> >> >> >> >> >> >> > > > Jesper Dangaard Brouer wr= ites: >> >> >> >> >> >> >> >> > > >=20=20=20 >> >> >> >> >> >> >> >> > > > > On Thu, 15 Apr 2021 10:35:51 -0700 >> >> >> >> >> >> >> >> > > > > Martin KaFai Lau wrote: >> >> >> >> >> >> >> >> > > > >=20=20 >> >> >> >> >> >> >> >> > > > >> On Thu, Apr 15, 2021 at 11:22:19AM +0200, = Toke H=C3=B8iland-J=C3=B8rgensen wrote:=20=20 >> >> >> >> >> >> >> >> > > > >> > Hangbin Liu write= s: >> >> >> >> >> >> >> >> > > > >> >=20=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> > > On Wed, Apr 14, 2021 at 05:17:11PM -07= 00, Martin KaFai Lau wrote:=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> > >> > static void bq_xmit_all(struct xdp= _dev_bulk_queue *bq, u32 flags) >> >> >> >> >> >> >> >> > > > >> > >> > { >> >> >> >> >> >> >> >> > > > >> > >> > struct net_device *dev =3D bq->de= v; >> >> >> >> >> >> >> >> > > > >> > >> > - int sent =3D 0, err =3D 0; >> >> >> >> >> >> >> >> > > > >> > >> > + int sent =3D 0, drops =3D 0, err = =3D 0; >> >> >> >> >> >> >> >> > > > >> > >> > + unsigned int cnt =3D bq->count; >> >> >> >> >> >> >> >> > > > >> > >> > + int to_send =3D cnt; >> >> >> >> >> >> >> >> > > > >> > >> > int i; >> >> >> >> >> >> >> >> > > > >> > >> >=20=20 >> >> >> >> >> >> >> >> > > > >> > >> > - if (unlikely(!bq->count)) >> >> >> >> >> >> >> >> > > > >> > >> > + if (unlikely(!cnt)) >> >> >> >> >> >> >> >> > > > >> > >> > return; >> >> >> >> >> >> >> >> > > > >> > >> >=20=20 >> >> >> >> >> >> >> >> > > > >> > >> > - for (i =3D 0; i < bq->count; i++)= { >> >> >> >> >> >> >> >> > > > >> > >> > + for (i =3D 0; i < cnt; i++) { >> >> >> >> >> >> >> >> > > > >> > >> > struct xdp_frame *xdpf =3D bq->q= [i]; >> >> >> >> >> >> >> >> > > > >> > >> >=20=20 >> >> >> >> >> >> >> >> > > > >> > >> > prefetch(xdpf); >> >> >> >> >> >> >> >> > > > >> > >> > } >> >> >> >> >> >> >> >> > > > >> > >> >=20=20 >> >> >> >> >> >> >> >> > > > >> > >> > - sent =3D dev->netdev_ops->ndo_xdp= _xmit(dev, bq->count, bq->q, flags); >> >> >> >> >> >> >> >> > > > >> > >> > + if (bq->xdp_prog) {=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> > >> bq->xdp_prog is used here >> >> >> >> >> >> >> >> > > > >> > >>=20=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> > >> > + to_send =3D dev_map_bpf_prog_run= (bq->xdp_prog, bq->q, cnt, dev); >> >> >> >> >> >> >> >> > > > >> > >> > + if (!to_send) >> >> >> >> >> >> >> >> > > > >> > >> > + goto out; >> >> >> >> >> >> >> >> > > > >> > >> > + >> >> >> >> >> >> >> >> > > > >> > >> > + drops =3D cnt - to_send; >> >> >> >> >> >> >> >> > > > >> > >> > + } >> >> >> >> >> >> >> >> > > > >> > >> > +=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> > >>=20 >> >> >> >> >> >> >> >> > > > >> > >> [ ... ] >> >> >> >> >> >> >> >> > > > >> > >>=20=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> > >> > static void bq_enqueue(struct net_= device *dev, struct xdp_frame *xdpf, >> >> >> >> >> >> >> >> > > > >> > >> > - struct net_device *dev_rx) >> >> >> >> >> >> >> >> > > > >> > >> > + struct net_device *dev_rx= , struct bpf_prog *xdp_prog) >> >> >> >> >> >> >> >> > > > >> > >> > { >> >> >> >> >> >> >> >> > > > >> > >> > struct list_head *flush_list =3D = this_cpu_ptr(&dev_flush_list); >> >> >> >> >> >> >> >> > > > >> > >> > struct xdp_dev_bulk_queue *bq =3D= this_cpu_ptr(dev->xdp_bulkq); >> >> >> >> >> >> >> >> > > > >> > >> > @@ -412,18 +466,22 @@ static void b= q_enqueue(struct net_device *dev, struct xdp_frame *xdpf, >> >> >> >> >> >> >> >> > > > >> > >> > /* Ingress dev_rx will be the sam= e for all xdp_frame's in >> >> >> >> >> >> >> >> > > > >> > >> > * bulk_queue, because bq stored = per-CPU and must be flushed >> >> >> >> >> >> >> >> > > > >> > >> > * from net_device drivers NAPI f= unc end. >> >> >> >> >> >> >> >> > > > >> > >> > + * >> >> >> >> >> >> >> >> > > > >> > >> > + * Do the same with xdp_prog and = flush_list since these fields >> >> >> >> >> >> >> >> > > > >> > >> > + * are only ever modified togethe= r. >> >> >> >> >> >> >> >> > > > >> > >> > */ >> >> >> >> >> >> >> >> > > > >> > >> > - if (!bq->dev_rx) >> >> >> >> >> >> >> >> > > > >> > >> > + if (!bq->dev_rx) { >> >> >> >> >> >> >> >> > > > >> > >> > bq->dev_rx =3D dev_rx; >> >> >> >> >> >> >> >> > > > >> > >> > + bq->xdp_prog =3D xdp_prog;=20=20= =20=20 >> >> >> >> >> >> >> >> > > > >> > >> bp->xdp_prog is assigned here and cou= ld be used later in bq_xmit_all(). >> >> >> >> >> >> >> >> > > > >> > >> How is bq->xdp_prog protected? Are th= ey all under one rcu_read_lock()? >> >> >> >> >> >> >> >> > > > >> > >> It is not very obvious after taking a= quick look at xdp_do_flush[_map]. >> >> >> >> >> >> >> >> > > > >> > >>=20 >> >> >> >> >> >> >> >> > > > >> > >> e.g. what if the devmap elem gets del= eted.=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> > > >> >> >> >> >> >> >> >> > > > >> > > Jesper knows better than me. From my v= eiw, based on the description of >> >> >> >> >> >> >> >> > > > >> > > __dev_flush(): >> >> >> >> >> >> >> >> > > > >> > > >> >> >> >> >> >> >> >> > > > >> > > On devmap tear down we ensure the flus= h list is empty before completing to >> >> >> >> >> >> >> >> > > > >> > > ensure all flush operations have compl= eted. When drivers update the bpf >> >> >> >> >> >> >> >> > > > >> > > program they may need to ensure any fl= ush ops are also complete.=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> >> >> >> >> >> >> >> >> > > > >> AFAICT, the bq->xdp_prog is not from the d= ev. It is from a devmap's elem. >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > The bq->xdp_prog comes form the devmap "dev" elem= ent, and it is stored >> >> >> >> >> >> >> >> > in temporarily in the "bq" structure that is only= valid for this >> >> >> >> >> >> >> >> > softirq NAPI-cycle. I'm slightly worried that we= copied this pointer >> >> >> >> >> >> >> >> > the the xdp_prog here, more below (and Q for Paul= ). >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > > > >> >=20 >> >> >> >> >> >> >> >> > > > >> > Yeah, drivers call xdp_do_flush() before= exiting their NAPI poll loop, >> >> >> >> >> >> >> >> > > > >> > which also runs under one big rcu_read_l= ock(). So the storage in the >> >> >> >> >> >> >> >> > > > >> > bulk queue is quite temporary, it's just= used for bulking to increase >> >> >> >> >> >> >> >> > > > >> > performance :)=20=20=20=20 >> >> >> >> >> >> >> >> > > > >> >> >> >> >> >> >> >> >> > > > >> I am missing the one big rcu_read_lock() p= art. For example, in i40e_txrx.c, >> >> >> >> >> >> >> >> > > > >> i40e_run_xdp() has its own rcu_read_lock/u= nlock(). dst->xdp_prog used to run >> >> >> >> >> >> >> >> > > > >> in i40e_run_xdp() and it is fine. >> >> >> >> >> >> >> >> > > > >>=20 >> >> >> >> >> >> >> >> > > > >> In this patch, dst->xdp_prog is run outsid= e of i40e_run_xdp() where the >> >> >> >> >> >> >> >> > > > >> rcu_read_unlock() has already done. It is= now run in xdp_do_flush_map(). >> >> >> >> >> >> >> >> > > > >> or I missed the big rcu_read_lock() in i40= e_napi_poll()? >> >> >> >> >> >> >> >> > > > >> >> >> >> >> >> >> >> >> > > > >> I do see the big rcu_read_lock() in mlx5e_= napi_poll().=20=20 >> >> >> >> >> >> >> >> > > > > >> >> >> >> >> >> >> >> > > > > I believed/assumed xdp_do_flush_map() was a= lready protected under an >> >> >> >> >> >> >> >> > > > > rcu_read_lock. As the devmap and cpumap, w= hich get called via >> >> >> >> >> >> >> >> > > > > __dev_flush() and __cpu_map_flush(), have m= ultiple RCU objects that we >> >> >> >> >> >> >> >> > > > > are operating on.=20=20 >> >> >> >> >> >> >> >> > > >> >> >> >> >> >> >> >> > > What other rcu objects it is using during flush? >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > Look at code: >> >> >> >> >> >> >> >> > kernel/bpf/cpumap.c >> >> >> >> >> >> >> >> > kernel/bpf/devmap.c >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > The devmap is filled with RCU code and complicate= d take-down steps.=20=20 >> >> >> >> >> >> >> >> > The devmap's elements are also RCU objects and th= e BPF xdp_prog is >> >> >> >> >> >> >> >> > embedded in this object (struct bpf_dtab_netdev).= The call_rcu >> >> >> >> >> >> >> >> > function is __dev_map_entry_free(). >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > > > > Perhaps it is a bug in i40e?=20=20 >> >> >> >> >> >> >> >> > > >> >> >> >> >> >> >> >> > > A quick look into ixgbe falls into the same buc= ket. >> >> >> >> >> >> >> >> > > didn't look at other drivers though. >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > Intel driver are very much in copy-paste mode. >> >> >> >> >> >> >> >> >=20=20 >> >> >> >> >> >> >> >> > > > > >> >> >> >> >> >> >> >> > > > > We are running in softirq in NAPI context, = when xdp_do_flush_map() is >> >> >> >> >> >> >> >> > > > > call, which I think means that this CPU wil= l not go-through a RCU grace >> >> >> >> >> >> >> >> > > > > period before we exit softirq, so in-practi= ce it should be safe.=20=20 >> >> >> >> >> >> >> >> > > >=20 >> >> >> >> >> >> >> >> > > > Yup, this seems to be correct: rcu_softirq_qs= () is only called between >> >> >> >> >> >> >> >> > > > full invocations of the softirq handler, whic= h for networking is >> >> >> >> >> >> >> >> > > > net_rx_action(), and so translates into full = NAPI poll cycles.=20=20 >> >> >> >> >> >> >> >> > > >> >> >> >> >> >> >> >> > > I don't know enough to comment on the rcu/softi= rq part, may be someone >> >> >> >> >> >> >> >> > > can chime in. There is also a recent napi_thre= aded_poll(). >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > CC added Paul. (link to patch[1][2] for context) >> >> >> >> >> >> >> >> Updated Paul's email address. >> >> >> >> >> >> >> >>=20 >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > > If it is the case, then some of the existing rc= u_read_lock() is unnecessary? >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > Well, in many cases, especially depending on how = kernel is compiled, >> >> >> >> >> >> >> >> > that is true. But we want to keep these, as they= also document the >> >> >> >> >> >> >> >> > intend of the programmer. And allow us to make t= he kernel even more >> >> >> >> >> >> >> >> > preempt-able in the future. >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > > At least, it sounds incorrect to only make an e= xception here while keeping >> >> >> >> >> >> >> >> > > other rcu_read_lock() as-is. >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > Let me be clear: I think you have spotted a prob= lem, and we need to >> >> >> >> >> >> >> >> > add rcu_read_lock() at least around the invocatio= n of >> >> >> >> >> >> >> >> > bpf_prog_run_xdp() or before around if-statement = that call >> >> >> >> >> >> >> >> > dev_map_bpf_prog_run(). (Hangbin please do this i= n V8). >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > Thank you Martin for reviewing the code carefully= enough to find this >> >> >> >> >> >> >> >> > issue, that some drivers don't have a RCU-section= around the full XDP >> >> >> >> >> >> >> >> > code path in their NAPI-loop. >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > Question to Paul. (I will attempt to describe in= generic terms what >> >> >> >> >> >> >> >> > happens, but ref real-function names). >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > We are running in softirq/NAPI context, the drive= r will call a >> >> >> >> >> >> >> >> > bq_enqueue() function for every packet (if callin= g xdp_do_redirect) , >> >> >> >> >> >> >> >> > some driver wrap this with a rcu_read_lock/unlock= () section (other have >> >> >> >> >> >> >> >> > a large RCU-read section, that include the flush = operation). >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > In the bq_enqueue() function we have a per_cpu_pt= r (that store the >> >> >> >> >> >> >> >> > xdp_frame packets) that will get flushed/send in = the call >> >> >> >> >> >> >> >> > xdp_do_flush() (that end-up calling bq_xmit_all()= ). This flush will >> >> >> >> >> >> >> >> > happen before we end our softirq/NAPI context. >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > The extension is that the per_cpu_ptr data struct= ure (after this patch) >> >> >> >> >> >> >> >> > store a pointer to an xdp_prog (which is a RCU ob= ject). In the flush >> >> >> >> >> >> >> >> > operation (which we will wrap with RCU-read secti= on), we will use this >> >> >> >> >> >> >> >> > xdp_prog pointer. I can see that it is in-princ= iple wrong to pass >> >> >> >> >> >> >> >> > this-pointer between RCU-read sections, but I con= sider this safe as we >> >> >> >> >> >> >> >> > are running under softirq/NAPI and the per_cpu_pt= r is only valid in >> >> >> >> >> >> >> >> > this short interval. >> >> >> >> >> >> >> >> >=20 >> >> >> >> >> >> >> >> > I claim a grace/quiescent RCU cannot happen betwe= en these two RCU-read >> >> >> >> >> >> >> >> > sections, but I might be wrong? (especially in th= e future or for RT). >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > If I am reading this correctly (ha!), a very high-le= vel summary of the >> >> >> >> >> >> >> > code in question is something like this: >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > void foo(void) >> >> >> >> >> >> >> > { >> >> >> >> >> >> >> > local_bh_disable(); >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > rcu_read_lock(); >> >> >> >> >> >> >> > p =3D rcu_dereference(gp); >> >> >> >> >> >> >> > do_something_with(p); >> >> >> >> >> >> >> > rcu_read_unlock(); >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > do_something_else(); >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > rcu_read_lock(); >> >> >> >> >> >> >> > do_some_other_thing(p); >> >> >> >> >> >> >> > rcu_read_unlock(); >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > local_bh_enable(); >> >> >> >> >> >> >> > } >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > void bar(struct blat *new_gp) >> >> >> >> >> >> >> > { >> >> >> >> >> >> >> > struct blat *old_gp; >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > spin_lock(my_lock); >> >> >> >> >> >> >> > old_gp =3D rcu_dereference_protected(gp, lock_held= (my_lock)); >> >> >> >> >> >> >> > rcu_assign_pointer(gp, new_gp); >> >> >> >> >> >> >> > spin_unlock(my_lock); >> >> >> >> >> >> >> > synchronize_rcu(); >> >> >> >> >> >> >> > kfree(old_gp); >> >> >> >> >> >> >> > } >> >> >> >> >> >> >>=20 >> >> >> >> >> >> >> Yeah, something like that (the object is freed using c= all_rcu() - but I >> >> >> >> >> >> >> think that's equivalent, right?). And the question is = whether we need to >> >> >> >> >> >> >> extend foo() so that is has one big rcu_read_lock() th= at covers the >> >> >> >> >> >> >> whole lifetime of p. >> >> >> >> >> >> > >> >> >> >> >> >> > Yes, use of call_rcu() is an asynchronous version of sy= nchronize_rcu(). >> >> >> >> >> >> > In fact, synchronize_rcu() is implemented in terms of c= all_rcu(). ;-) >> >> >> >> >> >>=20 >> >> >> >> >> >> Right, gotcha! >> >> >> >> >> >>=20 >> >> >> >> >> >> >> > I need to check up on -rt. >> >> >> >> >> >> >> > >> >> >> >> >> >> >> > But first... In recent mainline kernels, the local_b= h_disable() region >> >> >> >> >> >> >> > will look like one big RCU read-side critical sectio= n. But don't try >> >> >> >> >> >> >> > this prior to v4.20!!! In v4.19 and earlier, you wo= uld need to use >> >> >> >> >> >> >> > both synchronize_rcu() and synchronize_rcu_bh() to m= ake this work, or, >> >> >> >> >> >> >> > for less latency, synchronize_rcu_mult(call_rcu, cal= l_rcu_bh). >> >> >> >> >> >> >>=20 >> >> >> >> >> >> >> OK. Variants of this code has been around since before= then, but I >> >> >> >> >> >> >> honestly have no idea what it looked like back then ex= actly... >> >> >> >> >> >> > >> >> >> >> >> >> > I know that feeling... >> >> >> >> >> >> > >> >> >> >> >> >> >> > Except that in that case, why not just drop the inne= r rcu_read_unlock() >> >> >> >> >> >> >> > and rcu_read_lock() pair? Awkward function boundari= es or some such? >> >> >> >> >> >> >>=20 >> >> >> >> >> >> >> Well if we can just treat such a local_bh_disable()/en= able() pair as the >> >> >> >> >> >> >> equivalent of rcu_read_lock()/unlock() then I suppose = we could just get >> >> >> >> >> >> >> rid of the inner ones. What about tools like lockdep; = do they understand >> >> >> >> >> >> >> this, or are we likely to get complaints if we remove = it? >> >> >> >> >> >> > >> >> >> >> >> >> > If you just got rid of the first rcu_read_unlock() and = the second >> >> >> >> >> >> > rcu_read_lock() in the code above, lockdep will underst= and. >> >> >> >> >> >>=20 >> >> >> >> >> >> Right, but doing so entails going through all the drivers= , which is what >> >> >> >> >> >> we're trying to avoid :) >> >> >> >> >> > >> >> >> >> >> > I was afraid of that... ;-) >> >> >> >> >> > >> >> >> >> >> >> > However, if you instead get rid of -all- of the rcu_rea= d_lock() and >> >> >> >> >> >> > rcu_read_unlock() invocations in the code above, you wo= uld need to let >> >> >> >> >> >> > lockdep know by adding rcu_read_lock_bh_held(). So ins= tead of this: >> >> >> >> >> >> > >> >> >> >> >> >> > p =3D rcu_dereference(gp); >> >> >> >> >> >> > >> >> >> >> >> >> > You would do this: >> >> >> >> >> >> > >> >> >> >> >> >> > p =3D rcu_dereference_check(gp, rcu_read_lock_bh_held(= )); >> >> >> >> >> >> > >> >> >> >> >> >> > This would be needed for mainline, regardless of -rt. >> >> >> >> >> >>=20 >> >> >> >> >> >> OK. And as far as I can tell this is harmless for code pa= ths that call >> >> >> >> >> >> the same function but from a regular rcu_read_lock()-prot= ected section >> >> >> >> >> >> instead from a bh-disabled section, right? >> >> >> >> >> > >> >> >> >> >> > That is correct. That rcu_dereference_check() invocation = will make >> >> >> >> >> > lockdep be OK with rcu_read_lock() or with softirq being d= isabled. >> >> >> >> >> > Or both, for that matter. >> >> >> >> >>=20 >> >> >> >> >> OK, great, thank you for confirming my understanding! >> >> >> >> >>=20 >> >> >> >> >> >> What happens, BTW, if we *don't* get rid of all the exist= ing >> >> >> >> >> >> rcu_read_lock() sections? Going back to your foo() exampl= e above, what >> >> >> >> >> >> we're discussing is whether to add that second rcu_read_l= ock() around >> >> >> >> >> >> do_some_other_thing(p). I.e., the first one around the rc= u_dereference() >> >> >> >> >> >> is already there (in the particular driver we're discussi= ng), and the >> >> >> >> >> >> local_bh_disable/enable() pair is already there. AFAICT f= rom our >> >> >> >> >> >> discussion, there really is not much point in adding that= second >> >> >> >> >> >> rcu_read_lock/unlock(), is there? >> >> >> >> >> > >> >> >> >> >> > From an algorithmic point of view, the second rcu_read_loc= k() >> >> >> >> >> > and rcu_read_unlock() are redundant. Of course, there are= also >> >> >> >> >> > software-engineering considerations, including copy-pasta = issues. >> >> >> >> >> > >> >> >> >> >> >> And because that first rcu_read_lock() around the rcu_der= eference() is >> >> >> >> >> >> already there, lockdep is not likely to complain either, = so we're >> >> >> >> >> >> basically fine? Except that the code is somewhat confusin= g as-is, of >> >> >> >> >> >> course; i.e., we should probably fix it but it's not terr= ibly urgent. Or? >> >> >> >> >> > >> >> >> >> >> > I am concerned about copy-pasta-induced bugs. Someone loo= ks just at >> >> >> >> >> > the code, fails to note the fact that softirq is disabled = throughout, >> >> >> >> >> > and decides that leaking a pointer from one RCU read-side = critical >> >> >> >> >> > section to a later one is just fine. :-/ >> >> >> >> >>=20 >> >> >> >> >> Yup, totally agreed that we need to fix this for the sake of= the humans >> >> >> >> >> reading the code; just wanted to make sure my understanding = was correct >> >> >> >> >> that we don't strictly need to do anything as far as the mac= hines >> >> >> >> >> executing it are concerned :) >> >> >> >> >>=20 >> >> >> >> >> >> Hmm, looking at it now, it seems not all the lookup code = is actually >> >> >> >> >> >> doing rcu_dereference() at all, but rather just a plain R= EAD_ONCE() with >> >> >> >> >> >> a comment above it saying that RCU ensures objects won't = disappear[0]; >> >> >> >> >> >> so I suppose we're at least safe from lockdep in that sen= se :P - but we >> >> >> >> >> >> should definitely clean this up. >> >> >> >> >> >>=20 >> >> >> >> >> >> [0] Exhibit A: https://elixir.bootlin.com/linux/latest/so= urce/kernel/bpf/devmap.c#L391 >> >> >> >> >> > >> >> >> >> >> > That use of READ_ONCE() will definitely avoid lockdep comp= laints, >> >> >> >> >> > including those complaints that point out bugs. It also m= ight get you >> >> >> >> >> > sparse complaints if the RCU-protected pointer is marked w= ith __rcu. >> >> >> >> >>=20 >> >> >> >> >> It's not; it's the netdev_map member of this struct: >> >> >> >> >>=20 >> >> >> >> >> struct bpf_dtab { >> >> >> >> >> struct bpf_map map; >> >> >> >> >> struct bpf_dtab_netdev **netdev_map; /* DEVMAP type only */ >> >> >> >> >> struct list_head list; >> >> >> >> >>=20 >> >> >> >> >> /* these are only used for DEVMAP_HASH type maps */ >> >> >> >> >> struct hlist_head *dev_index_head; >> >> >> >> >> spinlock_t index_lock; >> >> >> >> >> unsigned int items; >> >> >> >> >> u32 n_buckets; >> >> >> >> >> }; >> >> >> >> >>=20 >> >> >> >> >> Will adding __rcu to such a dynamic array member do the righ= t thing when >> >> >> >> >> paired with rcu_dereference() on array members (i.e., in pla= ce of the >> >> >> >> >> READ_ONCE in the code linked above)? >> >> >> >> > >> >> >> >> > The only thing __rcu will do is provide information to the sp= arse static >> >> >> >> > analysis tool. Which will then gripe at you for applying REA= D_ONCE() >> >> >> >> > to a __rcu pointer. But it is already griping at you for app= lying >> >> >> >> > rcu_dereference() to something not marked __rcu, so... ;-) >> >> >> >>=20 >> >> >> >> Right, hence the need for a cleanup ;) >> >> >> >>=20 >> >> >> >> My question was more if it understood arrays, though. I.e., that >> >> >> >> 'netdev_map' is an array of RCU pointers, not an RCU pointer to= an >> >> >> >> array... Or am I maybe thinking that tool is way smarter than i= t is, and >> >> >> >> it just complains for any access to that field that doesn't use >> >> >> >> rcu_dereference()? >> >> >> > >> >> >> > I believe that sparse will know about the pointers being __rcu, = but >> >> >> > not the array. Unless you mark both levels. >> >> >>=20 >> >> >> Hi Paul >> >> >>=20 >> >> >> One more question, since I started adding the annotations: We are >> >> >> currently swapping out the pointers using xchg(): >> >> >> https://elixir.bootlin.com/linux/latest/source/kernel/bpf/devmap.c= #L555 >> >> >>=20 >> >> >> and even cmpxchg(): >> >> >> https://elixir.bootlin.com/linux/latest/source/kernel/bpf/devmap.c= #L831 >> >> >>=20 >> >> >> Sparse complains about these if I add the __rcu annotation to the >> >> >> definition (which otherwise works just fine with the double-pointe= r, >> >> >> BTW). Is there a way to fix that? Some kind of rcu_ macro version = of the >> >> >> atomic swaps or something? Or do we just keep the regular xchg() a= nd >> >> >> ignore those particular sparse warnings? >> >> > >> >> > Sounds like I need to supply a unrcu_pointer() macro or some such. >> >> > This would operate something like the current open-coded casts >> >> > in __rcu_dereference_protected(). >> >>=20 >> >> So with that, I would turn the existing: >> >>=20 >> >> dev =3D READ_ONCE(dtab->netdev_map[i]); >> >> if (!dev || netdev !=3D dev->dev) >> >> continue; >> >> odev =3D cmpxchg(&dtab->netdev_map[i], dev, NULL); >> >>=20 >> >> into: >> >>=20 >> >> dev =3D rcu_dereference(dtab->netdev_map[i]); >> >> if (!dev || netdev !=3D dev->dev) >> >> continue; >> >> odev =3D cmpxchg(unrcu_pointer(&dtab->netdev_map[i]), dev, NULL); >> >>=20 >> >>=20 >> >> and with a _check version: >> >>=20 >> >> old_dev =3D xchg(unrcu_pointer_check(&dtab->netdev_map[k], rcu_read_= lock_bh_held()), NULL); >> >>=20 >> >> right? >> >>=20 >> >> Or would it be: >> >> odev =3D cmpxchg(&unrcu_pointer(dtab->netdev_map[i]), dev, NULL); >> >> ? >> >>=20 >> >> > Would something like that work for you? >> >>=20 >> >> Yeah, I believe it would :) >> > >> > Except that I was forgetting that the __rcu decorates the pointed-to >> > data rather than the pointer itself. :-/ >> > >> > But that is actually easier, as you can follow the example of >> > rcu_assign_pointer(), namely using RCU_INITIALIZER(). >> > >> > So like this: >> > >> > odev =3D cmpxchg(&dtab->netdev_map[i], RCU_INITIALIZER(dev), NULL); >> > >> > I -think- that the NULL doesn't need an RCU_INITIALIZER(), but it is >> > of course sparse's opinion that matters. >> > >> > And of course like this: >> > >> > old_dev =3D xchg(&dtab->netdev_map[k], RCU_INITIALIZER(newmap)); >> > >> > Does that work, or am I still confused? >>=20 >> That gets rid of one warning, but not the other. Before (plain xchg): >>=20 >> kernel/bpf/devmap.c:657:19: warning: incorrect type in initializer (diff= erent address spaces) >> kernel/bpf/devmap.c:657:19: expected struct bpf_dtab_netdev [noderef]= __rcu *__ret >> kernel/bpf/devmap.c:657:19: got struct bpf_dtab_netdev *[assigned] dev >> kernel/bpf/devmap.c:657:17: warning: incorrect type in assignment (diffe= rent address spaces) >> kernel/bpf/devmap.c:657:17: expected struct bpf_dtab_netdev *old_dev >> kernel/bpf/devmap.c:657:17: got struct bpf_dtab_netdev [noderef] __rc= u *[assigned] __ret >>=20 >> after (RCU_INITIALIZER() on the second argument to xchg): >>=20 >> kernel/bpf/devmap.c:657:17: warning: incorrect type in assignment (diffe= rent address spaces) >> kernel/bpf/devmap.c:657:17: expected struct bpf_dtab_netdev *old_dev >> kernel/bpf/devmap.c:657:17: got struct bpf_dtab_netdev [noderef] __rc= u *[assigned] __ret >>=20 >> I can get rid of that second one by marking old_dev as __rcu, but then I >> get a new warning when dereferencing that in the subsequent >> call_rcu()... >>=20 >> So I guess we still need that unrcu_pointer(), to wrap the xchg() in? > > Well, at least this use case permits an lvalue. ;-) > > Please see below for an untested patch intended to permit the following: > > old_dev =3D unrcu_pointer(xchg(&dtab->netdev_map[k], RCU_INITIALIZER(new= map))); > > Does that do the trick? Yes, it does! With that I can mark the pointer as __rcu and get all uses of it through sparse without complaints - awesome! How do RCU patches usually make it into the kernel? Can you provide me with a proper patch I can just include along with my cleanup patches (taking it through the bpf tree)? Or do we need to go through some other tree and wait for a merge? -Toke