Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S933011AbcLGU6i (ORCPT ); Wed, 7 Dec 2016 15:58:38 -0500 Received: from p3plsmtps2ded03.prod.phx3.secureserver.net ([208.109.80.60]:57390 "EHLO p3plsmtps2ded03.prod.phx3.secureserver.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753113AbcLGU6g (ORCPT ); Wed, 7 Dec 2016 15:58:36 -0500 x-originating-ip: 72.167.245.219 From: kys@exchange.microsoft.com To: gregkh@linuxfoundation.org, linux-kernel@vger.kernel.org, devel@linuxdriverproject.org, olaf@aepfle.de, apw@canonical.com, vkuznets@redhat.com, jasowang@redhat.com, leann.ogasawara@canonical.com Cc: "K. Y. Srinivasan" Subject: [PATCH 2/2] hv: make CPU offlining prevention fine-grained Date: Wed, 7 Dec 2016 14:53:12 -0800 Message-Id: <1481151192-22440-2-git-send-email-kys@exchange.microsoft.com> X-Mailer: git-send-email 1.7.4.1 In-Reply-To: <1481151192-22440-1-git-send-email-kys@exchange.microsoft.com> References: <1481151150-22390-1-git-send-email-kys@exchange.microsoft.com> <1481151192-22440-1-git-send-email-kys@exchange.microsoft.com> Reply-To: kys@microsoft.com X-CMAE-Envelope: MS4wfDU8VUo/yZbNX8LE571wFTuTGP9swjDbIkO6oyTJsusWjG6ONntrppKvQfyFg/hoxYqdA5Fo90NO/OBV4ntbZzmxxxu4ozDn/3A1nH6F2Go5Pyw2fR6Y LxkczEfnCIKjB3cFJz3FRAogKfqAw2+wE/rNuQAh+9P4f26sby/puzZYqe/gjSweINzPx/igGuliOYl+/uyWIhw0ZYRcZoe7EQrXXRGwRqxS8kDZtWlVokES LHzh2n/U90NFlR4rL37DeUgdZFxtFGrObS9BootSHSwHN6Z/RXJB81x7VtiffYGzDHwGXtoIhd7lw4UPmGnL/C3ICDib8k4Tg/8Fxa01zUcVFT3T9Z4i5zSN Qe5H+NaJ5j8LBuq+r/3Rx5JqbRGaKN7aNMMGFSzZ/kcR0JwDIz/2ABi8713Z3NitaN/+0OOeaxer2usCmG19lfOSEABSwfGRdpNLOMQpykEjRPnJeBw= Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3515 Lines: 108 From: Vitaly Kuznetsov Since commit e513229b4c38 ("Drivers: hv: vmbus: prevent cpu offlining on newer hypervisors") cpu offlining was disabled. It is still true that we can't offline CPUs which have VMBus channels bound to them but we may have 'free' CPUs (e.v. we booted with maxcpus= parameter and onlined CPUs after VMBus was initialized), these CPUs may be disabled without issues. In future, we may even allow closing CPUs which have only sub-channels assinged to them by closing these sub-channels. All devices will continue to work. Signed-off-by: Vitaly Kuznetsov Signed-off-by: K. Y. Srinivasan --- drivers/hv/hv.c | 31 +++++++++++++++++++++++++++++++ drivers/hv/vmbus_drv.c | 9 ++++----- 2 files changed, 35 insertions(+), 5 deletions(-) diff --git a/drivers/hv/hv.c b/drivers/hv/hv.c index b871fcc..04a0fd8 100644 --- a/drivers/hv/hv.c +++ b/drivers/hv/hv.c @@ -587,10 +587,41 @@ int hv_synic_cleanup(unsigned int cpu) union hv_synic_simp simp; union hv_synic_siefp siefp; union hv_synic_scontrol sctrl; + struct vmbus_channel *channel, *sc; + bool channel_found = false; + unsigned long flags; if (!hv_context.synic_initialized) return -EFAULT; + /* + * Search for channels which are bound to the CPU we're about to + * cleanup. In case we find one and vmbus is still connected we need to + * fail, this will effectively prevent CPU offlining. There is no way + * we can re-bind channels to different CPUs for now. + */ + mutex_lock(&vmbus_connection.channel_mutex); + list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) { + if (channel->target_cpu == cpu) { + channel_found = true; + break; + } + spin_lock_irqsave(&channel->lock, flags); + list_for_each_entry(sc, &channel->sc_list, sc_list) { + if (sc->target_cpu == cpu) { + channel_found = true; + break; + } + } + spin_unlock_irqrestore(&channel->lock, flags); + if (channel_found) + break; + } + mutex_unlock(&vmbus_connection.channel_mutex); + + if (channel_found && vmbus_connection.conn_state == CONNECTED) + return -EBUSY; + /* Turn off clockevent device */ if (ms_hyperv.features & HV_X64_MSR_SYNTIMER_AVAILABLE) { clockevents_unbind_device(hv_context.clk_evt[cpu], cpu); diff --git a/drivers/hv/vmbus_drv.c b/drivers/hv/vmbus_drv.c index a16d519..565bdd1 100644 --- a/drivers/hv/vmbus_drv.c +++ b/drivers/hv/vmbus_drv.c @@ -1008,9 +1008,6 @@ static int vmbus_bus_init(void) if (ret) goto err_connect; - if (vmbus_proto_version > VERSION_WIN7) - cpu_hotplug_disable(); - /* * Only register if the crash MSRs are available */ @@ -1486,6 +1483,9 @@ static void hv_kexec_handler(void) { hv_synic_clockevents_cleanup(); vmbus_initiate_unload(false); + vmbus_connection.conn_state = DISCONNECTED; + /* Make sure conn_state is set as hv_synic_cleanup checks for it */ + mb(); cpuhp_remove_state(hyperv_cpuhp_online); hv_cleanup(false); }; @@ -1498,6 +1498,7 @@ static void hv_crash_handler(struct pt_regs *regs) * doing the cleanup for current CPU only. This should be sufficient * for kdump. */ + vmbus_connection.conn_state = DISCONNECTED; hv_synic_cleanup(smp_processor_id()); hv_cleanup(true); }; @@ -1566,8 +1567,6 @@ static void __exit vmbus_exit(void) cpuhp_remove_state(hyperv_cpuhp_online); hv_synic_free(); acpi_bus_unregister_driver(&vmbus_acpi_driver); - if (vmbus_proto_version > VERSION_WIN7) - cpu_hotplug_enable(); } -- 1.7.4.1