From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759128AbZIPNwA (ORCPT ); Wed, 16 Sep 2009 09:52:00 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1759136AbZIPNv6 (ORCPT ); Wed, 16 Sep 2009 09:51:58 -0400 Received: from va3ehsobe005.messaging.microsoft.com ([216.32.180.15]:48856 "EHLO VA3EHSOBE005.bigfish.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1759123AbZIPNvx (ORCPT ); Wed, 16 Sep 2009 09:51:53 -0400 X-SpamScore: 3 X-BigFish: VPS3(zzzz1202hzzz32i203h43j66h) X-Spam-TCS-SCL: 5:0 X-FB-SS: 5, X-WSS-ID: 0KQ2H69-03-4NM-01 From: Borislav Petkov To: CC: , , , Borislav Petkov Subject: [PATCH 5/5] amd64_edac: check NB MCE bank enable on the current node properly Date: Wed, 16 Sep 2009 15:50:50 +0200 Message-ID: <1253109050-31165-6-git-send-email-borislav.petkov@amd.com> X-Mailer: git-send-email 1.6.3.3 In-Reply-To: <1253109050-31165-1-git-send-email-borislav.petkov@amd.com> References: <1253109050-31165-1-git-send-email-borislav.petkov@amd.com> X-OriginalArrivalTime: 16 Sep 2009 13:51:24.0300 (UTC) FILETIME=[C796DCC0:01CA36D4] MIME-Version: 1.0 Content-Type: text/plain Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org The old code was using smp_call_function_many which skips the current cpu if it is in the supplied cpumask. Switch to the rdmsr_on_cpus() interface which takes care of that. In addition, add get_cpus_on_this_dct_cpumask helper which computes a cpumask of all the cores on a node and thus on a DCT. Signed-off-by: Borislav Petkov --- drivers/edac/amd64_edac.c | 66 ++++++++++++++++++++++++++++++-------------- 1 files changed, 45 insertions(+), 21 deletions(-) diff --git a/drivers/edac/amd64_edac.c b/drivers/edac/amd64_edac.c index f943ad8..4e551e6 100644 --- a/drivers/edac/amd64_edac.c +++ b/drivers/edac/amd64_edac.c @@ -2741,30 +2741,53 @@ static void amd64_restore_ecc_error_reporting(struct amd64_pvt *pvt) wrmsr_on_cpus(cpumask, K8_MSR_MCGCTL, msrs); } -static void check_mcg_ctl(void *ret) +/* get all cores on this DCT */ +static void get_cpus_on_this_dct_cpumask(cpumask_t *mask, int nid) { - u64 msr_val = 0; - u8 nbe; + int cpu; - rdmsrl(MSR_IA32_MCG_CTL, msr_val); - nbe = msr_val & K8_MSR_MCGCTL_NBE; - - debugf0("core: %u, MCG_CTL: 0x%llx, NB MSR is %s\n", - raw_smp_processor_id(), msr_val, - (nbe ? "enabled" : "disabled")); - - if (!nbe) - *(int *)ret = 0; + for_each_online_cpu(cpu) + if (amd_get_nb_id(cpu) == nid) + cpumask_set_cpu(cpu, mask); } /* check MCG_CTL on all the cpus on this node */ -static int mcg_ctl_enabled_on_node(const struct cpumask *mask) +static bool amd64_nb_mce_bank_enabled_on_node(int nid) { - int ret = 1; - preempt_disable(); - smp_call_function_many(mask, check_mcg_ctl, &ret, 1); - preempt_enable(); + cpumask_t mask; + struct msr *msrs; + int cpu, nbe, idx = 0; + bool ret = false; + + cpumask_clear(&mask); + + get_cpus_on_this_dct_cpumask(&mask, nid); + + msrs = kzalloc(sizeof(struct msr) * cpumask_weight(&mask), GFP_KERNEL); + if (!msrs) { + amd64_printk(KERN_WARNING, "%s: error allocating msrs\n", + __func__); + return false; + } + + rdmsr_on_cpus(&mask, MSR_IA32_MCG_CTL, msrs); + + for_each_cpu(cpu, &mask) { + nbe = msrs[idx].l & K8_MSR_MCGCTL_NBE; + + debugf0("core: %u, MCG_CTL: 0x%llx, NB MSR is %s\n", + cpu, msrs[idx].q, + (nbe ? "enabled" : "disabled")); + + if (!nbe) + goto out; + + idx++; + } + ret = true; +out: + kfree(msrs); return ret; } @@ -2783,7 +2806,8 @@ static int amd64_check_ecc_enabled(struct amd64_pvt *pvt) { u32 value; int err = 0; - u8 ecc_enabled = 0, mcg_ctl_en = 0; + u8 ecc_enabled = 0; + bool nb_mce_en = false; err = pci_read_config_dword(pvt->misc_f3_ctl, K8_NBCFG, &value); if (err) @@ -2797,13 +2821,13 @@ static int amd64_check_ecc_enabled(struct amd64_pvt *pvt) else amd64_printk(KERN_INFO, "ECC is enabled by BIOS.\n"); - mcg_ctl_en = mcg_ctl_enabled_on_node(cpumask_of_node(pvt->mc_node_id)); - if (!mcg_ctl_en) + nb_mce_en = amd64_nb_mce_bank_enabled_on_node(pvt->mc_node_id); + if (!nb_mce_en) amd64_printk(KERN_WARNING, "NB MCE bank disabled, set MSR " "0x%08x[4] on node %d to enable.\n", MSR_IA32_MCG_CTL, pvt->mc_node_id); - if (!ecc_enabled || !mcg_ctl_en) { + if (!ecc_enabled || !nb_mce_en) { if (!ecc_enable_override) { amd64_printk(KERN_WARNING, "%s", ecc_warning); return -ENODEV; -- 1.6.3.3