From nobody Wed Jan 12 21:40:33 2022 X-Original-To: dev-commits-src-all@mlmmj.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mlmmj.nyi.freebsd.org (Postfix) with ESMTP id 5602F195EBD9; Wed, 12 Jan 2022 21:40:38 +0000 (UTC) (envelope-from markjdb@gmail.com) Received: from mail-qk1-x72c.google.com (mail-qk1-x72c.google.com [IPv6:2607:f8b0:4864:20::72c]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256 client-signature RSA-PSS (2048 bits) client-digest SHA256) (Client CN "smtp.gmail.com", Issuer "GTS CA 1D4" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 4JZ1Fd27kLz4sMq; Wed, 12 Jan 2022 21:40:37 +0000 (UTC) (envelope-from markjdb@gmail.com) Received: by mail-qk1-x72c.google.com with SMTP id z10so531284qkf.7; Wed, 12 Jan 2022 13:40:37 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=sender:date:from:to:cc:subject:message-id:references:mime-version :content-disposition:in-reply-to; bh=xr8D6ixJl5gvWwxHKcuRdPyCWSG/X4KY4iBoKOWPvpE=; b=Plx4b3KpkwxD5st7wAnMFm5A95t5UbBBN/L17g/M2DUNSqrwIHJoJrwt4W62+7yMAL Tk09e79NREWajk6QCzTAe1vT0v20XICFqUnVshMlqwLKd4yWW7jEqGvLHJcqjKxFnVDm eW+xovQaTVy6jagfBqJbphbyrT3cdUlMdwzachiOAWhGl4KpQs8nHHy6JX6NqsustA4Q 7T510gul2Ii5nkaqyeZtV9zpDgFLJ9LO1UWoJPgX6d+wesgAj3r2l6Z6WBxPT7EsXn+R qWmCg0nLQDqgWFG+9bQ3JW9vD++AiNBITR7w1wBSLFrDCAvv+wZffUH0eJSOU82VWYfS WKeQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=x-gm-message-state:sender:date:from:to:cc:subject:message-id :references:mime-version:content-disposition:in-reply-to; bh=xr8D6ixJl5gvWwxHKcuRdPyCWSG/X4KY4iBoKOWPvpE=; b=cnBYZtVMTAV0zStQC7w062dsEJI4A53IDbbuCDQcKI5qjphwyDY9Mnq+1GOTp0khAw B2vMURrGW7/XznLW35pS2l387pCb3HzghLU3wwpwvTHaXMyJqiuL+dPQ0gEVmjgv0p4+ 92YWcWiKK3DWDLT9PmiE7icxT4K9zd24k1P43bc3CilLUNRWBC99fgonDsMFO/lkfMSW UqE0ERset1I4lKUUiFR3UiRsShUAyi++arGSTTcLnzXRzadCkdTLUwSsXGI9IyGrkhoF 6QLFwDl7Agi1jTt18eFAQfSui9YmgAg53hs5aZjRlmSiAHUUdafuVzegtFyuRmzyxAP2 tGjw== X-Gm-Message-State: AOAM533+O7vBWMU3GuhZXROa3Aq69iWh74QpN94yZyQi+IeWhCzVOF2Z BVjjXrmeScMnvs/xCm0D6hgULGghvbg= X-Google-Smtp-Source: ABdhPJxiV0R35YsvEjVFD3hDUQsZYRnPiklEE1h6CEmvZIvk8/FGHF9gn016SAdXRa2RIRkpmHluvg== X-Received: by 2002:a05:620a:124e:: with SMTP id a14mr1311524qkl.779.1642023636230; Wed, 12 Jan 2022 13:40:36 -0800 (PST) Received: from nuc (198-84-189-58.cpe.teksavvy.com. [198.84.189.58]) by smtp.gmail.com with ESMTPSA id t11sm551437qkm.67.2022.01.12.13.40.35 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 12 Jan 2022 13:40:35 -0800 (PST) Sender: Mark Johnston Date: Wed, 12 Jan 2022 16:40:33 -0500 From: Mark Johnston To: Wojciech Macek Cc: src-committers@freebsd.org, dev-commits-src-all@freebsd.org, dev-commits-src-main@freebsd.org Subject: Re: git: 2e72208b6c62 - main - ip_mroute: do not call epoch_waitwhen lock is taken Message-ID: References: <202201111019.20BAJjCg058259@gitrepo.freebsd.org> List-Id: Commit messages for all branches of the src repository List-Archive: https://lists.freebsd.org/archives/dev-commits-src-all List-Help: List-Post: List-Subscribe: List-Unsubscribe: Sender: owner-dev-commits-src-all@freebsd.org X-BeenThere: dev-commits-src-all@freebsd.org MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <202201111019.20BAJjCg058259@gitrepo.freebsd.org> X-Rspamd-Queue-Id: 4JZ1Fd27kLz4sMq X-Spamd-Bar: + Authentication-Results: mx1.freebsd.org; dkim=pass header.d=gmail.com header.s=20210112 header.b=Plx4b3Kp; dmarc=none; spf=pass (mx1.freebsd.org: domain of markjdb@gmail.com designates 2607:f8b0:4864:20::72c as permitted sender) smtp.mailfrom=markjdb@gmail.com X-Spamd-Result: default: False [1.07 / 15.00]; RCVD_VIA_SMTP_AUTH(0.00)[]; ARC_NA(0.00)[]; R_DKIM_ALLOW(-0.20)[gmail.com:s=20210112]; NEURAL_HAM_MEDIUM(-1.00)[-1.000]; FROM_HAS_DN(0.00)[]; RCPT_COUNT_THREE(0.00)[4]; TO_DN_SOME(0.00)[]; TO_MATCH_ENVRCPT_ALL(0.00)[]; MIME_GOOD(-0.10)[text/plain]; MID_RHS_NOT_FQDN(0.50)[]; DMARC_NA(0.00)[freebsd.org]; R_SPF_ALLOW(-0.20)[+ip6:2607:f8b0:4000::/36]; RCVD_COUNT_THREE(0.00)[3]; NEURAL_SPAM_SHORT(0.96)[0.961]; DKIM_TRACE(0.00)[gmail.com:+]; NEURAL_SPAM_LONG(0.81)[0.808]; RCVD_IN_DNSWL_NONE(0.00)[2607:f8b0:4864:20::72c:from]; FORGED_SENDER(0.30)[markj@freebsd.org,markjdb@gmail.com]; MIME_TRACE(0.00)[0:+]; FREEMAIL_ENVFROM(0.00)[gmail.com]; ASN(0.00)[asn:15169, ipnet:2607:f8b0::/32, country:US]; FROM_NEQ_ENVFROM(0.00)[markj@freebsd.org,markjdb@gmail.com]; RCVD_TLS_ALL(0.00)[]; DWL_DNSWL_NONE(0.00)[gmail.com:dkim] X-ThisMailContainsUnwantedMimeParts: N On Tue, Jan 11, 2022 at 10:19:45AM +0000, Wojciech Macek wrote: > The branch main has been updated by wma: > > URL: https://cgit.FreeBSD.org/src/commit/?id=2e72208b6c622505323ed48dc58830fc307392b1 > > commit 2e72208b6c622505323ed48dc58830fc307392b1 > Author: Wojciech Macek > AuthorDate: 2022-01-11 10:08:35 +0000 > Commit: Wojciech Macek > CommitDate: 2022-01-11 10:19:32 +0000 > > ip_mroute: do not call epoch_waitwhen lock is taken > > mrouter_done is called with RAW IP lock taken. Some annoying > printfs are visible on the console if INVARIANTS option is enabled. > > Provide atomic-based mechanism which counts enters and exits from/to > critical section in ip_input and ip_output. > Before de-initialization of function pointers ensure (with busy-wait) > that mrouter de-initialization is visible to all readers and that we don't > remove pointers (like ip_mforward etc.) in the middle of packet processing. This doesn't address the problem that the warning is complaining about, which is that a non-sleepable lock is held while performing an expensive operation. Now we are silently spinning instead, and there's no guarantee of forward progress since threads may be frequently entering the MROUTER_RLOCK-protected section in ip_input(). And the change converts a per-CPU atomic operation (in epoch_enter()) to a global atomic operation, potentially hurting performance. The atomics are also missing requisite barriers. I think a better solution is to not hold the raw inpcb lock while calling the ip_mrouter_done() callback. That lock does not provide synchronization for anything related to ip_mroute.c. Also note that X_ip_mrouter_done() can sleep in taskqueue_drain(), which is another reason to avoid holding the lock there. > --- > sys/netinet/ip_mroute.h | 9 +++++---- > sys/netinet/raw_ip.c | 1 + > 2 files changed, 6 insertions(+), 4 deletions(-) > > diff --git a/sys/netinet/ip_mroute.h b/sys/netinet/ip_mroute.h > index 65c5bdd3a025..faf5b8c72a46 100644 > --- a/sys/netinet/ip_mroute.h > +++ b/sys/netinet/ip_mroute.h > @@ -365,11 +365,12 @@ extern int (*ip_mrouter_set)(struct socket *, struct sockopt *); > extern int (*ip_mrouter_get)(struct socket *, struct sockopt *); > extern int (*ip_mrouter_done)(void); > extern int (*mrt_ioctl)(u_long, caddr_t, int); > +extern int ip_mrouter_critical_section_cnt; > > -#define MROUTER_RLOCK_TRACKER struct epoch_tracker mrouter_et > -#define MROUTER_RLOCK() epoch_enter_preempt(net_epoch_preempt, &mrouter_et) > -#define MROUTER_RUNLOCK() epoch_exit_preempt(net_epoch_preempt, &mrouter_et) > -#define MROUTER_WAIT() epoch_wait_preempt(net_epoch_preempt) > +#define MROUTER_RLOCK_TRACKER Why keep this definition at all? > +#define MROUTER_RLOCK() atomic_add_int(&ip_mrouter_critical_section_cnt, 1) > +#define MROUTER_RUNLOCK() atomic_subtract_int(&ip_mrouter_critical_section_cnt, 1) > +#define MROUTER_WAIT() do {} while (atomic_load_int(&ip_mrouter_critical_section_cnt) != 0) > > #endif /* _KERNEL */ > > diff --git a/sys/netinet/raw_ip.c b/sys/netinet/raw_ip.c > index 7c495745806e..dc49c36f25ad 100644 > --- a/sys/netinet/raw_ip.c > +++ b/sys/netinet/raw_ip.c > @@ -120,6 +120,7 @@ VNET_DEFINE(struct socket *, ip_mrouter); > int (*ip_mrouter_set)(struct socket *, struct sockopt *); > int (*ip_mrouter_get)(struct socket *, struct sockopt *); > int (*ip_mrouter_done)(void); > +int ip_mrouter_critical_section_cnt; > int (*ip_mforward)(struct ip *, struct ifnet *, struct mbuf *, > struct ip_moptions *); > int (*mrt_ioctl)(u_long, caddr_t, int);