From owner-svn-src-all@freebsd.org Wed Nov 21 17:32:11 2018 Return-Path: Delivered-To: svn-src-all@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 08173113F917; Wed, 21 Nov 2018 17:32:11 +0000 (UTC) (envelope-from markj@FreeBSD.org) Received: from mxrelay.nyi.freebsd.org (mxrelay.nyi.freebsd.org [IPv6:2610:1c1:1:606c::19:3]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client CN "mxrelay.nyi.freebsd.org", Issuer "Let's Encrypt Authority X3" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id A2BF76CB63; Wed, 21 Nov 2018 17:32:10 +0000 (UTC) (envelope-from markj@FreeBSD.org) Received: from repo.freebsd.org (repo.freebsd.org [IPv6:2610:1c1:1:6068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client did not present a certificate) by mxrelay.nyi.freebsd.org (Postfix) with ESMTPS id 6AE0E2386B; Wed, 21 Nov 2018 17:32:10 +0000 (UTC) (envelope-from markj@FreeBSD.org) Received: from repo.freebsd.org ([127.0.1.37]) by repo.freebsd.org (8.15.2/8.15.2) with ESMTP id wALHWAea065546; Wed, 21 Nov 2018 17:32:10 GMT (envelope-from markj@FreeBSD.org) Received: (from markj@localhost) by repo.freebsd.org (8.15.2/8.15.2/Submit) id wALHWAaP065545; Wed, 21 Nov 2018 17:32:10 GMT (envelope-from markj@FreeBSD.org) Message-Id: <201811211732.wALHWAaP065545@repo.freebsd.org> X-Authentication-Warning: repo.freebsd.org: markj set sender to markj@FreeBSD.org using -f From: Mark Johnston Date: Wed, 21 Nov 2018 17:32:10 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r340734 - head/sys/kern X-SVN-Group: head X-SVN-Commit-Author: markj X-SVN-Commit-Paths: head/sys/kern X-SVN-Commit-Revision: 340734 X-SVN-Commit-Repository: base MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Rspamd-Queue-Id: A2BF76CB63 X-Spamd-Result: default: False [0.45 / 15.00]; local_wl_from(0.00)[FreeBSD.org]; NEURAL_SPAM_LONG(0.00)[0.001,0]; NEURAL_SPAM_MEDIUM(0.25)[0.253,0]; ASN(0.00)[asn:11403, ipnet:2610:1c1:1::/48, country:US]; NEURAL_SPAM_SHORT(0.20)[0.198,0] X-Rspamd-Server: mx1.freebsd.org X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 21 Nov 2018 17:32:11 -0000 Author: markj Date: Wed Nov 21 17:32:09 2018 New Revision: 340734 URL: https://svnweb.freebsd.org/changeset/base/340734 Log: Avoid unsynchronized updates to kn_status. kn_status is protected by the kqueue's lock, but we were updating it without the kqueue lock held. For EVFILT_TIMER knotes, there is no knlist lock, so the knote activation could occur during the kn_status update and result in KN_QUEUED being lost, in which case we'd enqueue an already-enqueued knote, corrupting the queue. Fix the problem by setting or clearing KN_DISABLED before dropping the kqueue lock to call into the filter. KN_DISABLED is used only by the core kevent code, so there is no side effect from setting it earlier. Reported and tested by: Sylvain GALLIANO Reviewed by: kib MFC after: 3 days Sponsored by: The FreeBSD Foundation Differential Revision: https://reviews.freebsd.org/D18060 Modified: head/sys/kern/kern_event.c Modified: head/sys/kern/kern_event.c ============================================================================== --- head/sys/kern/kern_event.c Wed Nov 21 17:28:10 2018 (r340733) +++ head/sys/kern/kern_event.c Wed Nov 21 17:32:09 2018 (r340734) @@ -1533,6 +1533,8 @@ findkn: kn->kn_kevent.flags &= ~(EV_ADD | EV_DELETE | EV_ENABLE | EV_DISABLE | EV_FORCEONESHOT); kn->kn_status = KN_DETACHED; + if ((kev->flags & EV_DISABLE) != 0) + kn->kn_status |= KN_DISABLED; kn_enter_flux(kn); error = knote_attach(kn, kq); @@ -1568,6 +1570,11 @@ findkn: KNOTE_ACTIVATE(kn, 1); } + if ((kev->flags & EV_ENABLE) != 0) + kn->kn_status &= ~KN_DISABLED; + else if ((kev->flags & EV_DISABLE) != 0) + kn->kn_status |= KN_DISABLED; + /* * The user may change some filter values after the initial EV_ADD, * but doing so will not reset any filter which has already been @@ -1585,19 +1592,17 @@ findkn: kn->kn_sdata = kev->data; } +done_ev_add: /* * We can get here with kn->kn_knlist == NULL. This can happen when * the initial attach event decides that the event is "completed" - * already. i.e. filt_procattach is called on a zombie process. It - * will call filt_proc which will remove it from the list, and NULL + * already, e.g., filt_procattach() is called on a zombie process. It + * will call filt_proc() which will remove it from the list, and NULL * kn_knlist. + * + * KN_DISABLED will be stable while the knote is in flux, so the + * unlocked read will not race with an update. */ -done_ev_add: - if ((kev->flags & EV_ENABLE) != 0) - kn->kn_status &= ~KN_DISABLED; - else if ((kev->flags & EV_DISABLE) != 0) - kn->kn_status |= KN_DISABLED; - if ((kn->kn_status & KN_DISABLED) == 0) event = kn->kn_fop->f_event(kn, 0); else