Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752736AbaK0UQ6 (ORCPT ); Thu, 27 Nov 2014 15:16:58 -0500 Received: from mx1.redhat.com ([209.132.183.28]:40990 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752491AbaK0UQx (ORCPT ); Thu, 27 Nov 2014 15:16:53 -0500 Date: Thu, 27 Nov 2014 21:16:42 +0100 From: Radim =?utf-8?B?S3LEjW3DocWZ?= To: Nadav Amit Cc: linux-kernel@vger.kernel.org, kvm@vger.kernel.org, Paolo Bonzini , Gleb Natapov , Nadav Amit Subject: Re: [PATCH 3/4] KVM: x86: allow 256 logical x2APICs again Message-ID: <20141127201641.GB383@potion.brq.redhat.com> References: <1417114994-25235-1-git-send-email-rkrcmar@redhat.com> <1417114994-25235-4-git-send-email-rkrcmar@redhat.com> <22F11C34-E2E4-459B-882B-A944AD0853FD@gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: <22F11C34-E2E4-459B-882B-A944AD0853FD@gmail.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org 2014-11-27 21:53+0200, Nadav Amit: > Radim Krčmář wrote: > > - new->cid_mask = (1 << KVM_X2APIC_CID_BITS) - 1; > > - new->lid_mask = 0xffff; > > + new->cid_mask = new->lid_mask = 0xffff; > You set cid_mask to 0xffff, while there are only 16 clusters. I think it is > risky (if you twist my hand would come with a scenario). Let's see :) APIC id is 8 bit, and we compute cluster part of LDR by taking four upper bits, so 16 is enough. It isn't the safest programming practice, but we already fail to check physical_map bounds and any boost to maximal APIC ID is going to require a rewrite, thus I didn't bother to do it ... All uses should be covered with the following hunk, I will add it to v2 after all reviews, diff --git a/arch/x86/kvm/lapic.c b/arch/x86/kvm/lapic.c index 6c2b8a5..30e4cc1 100644 --- a/arch/x86/kvm/lapic.c +++ b/arch/x86/kvm/lapic.c @@ -198,7 +198,7 @@ static void recalculate_apic_map(struct kvm *kvm) cid = apic_cluster_id(new, ldr); lid = apic_logical_id(new, ldr); - if (lid) + if (lid && cid < ARRAY_SIZE(map->logical_map)) new->logical_map[cid][ffs(lid) - 1] = apic; } out: > Yet, why not to set > cid_mask to (ARRAY_SIZE(map->logical_map) - 1) ? We would incorrectly deliver messages intended for high clusters, it has to be 0xffff. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/