Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [PATCH 2/2] KVM: Scalable memslots implementation
To:     Sean Christopherson <seanjc@google.com>,
        "Maciej S. Szmigiero" <mail@maciej.szmigiero.name>
Cc:     Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Igor Mammedov <imammedo@redhat.com>,
        Marc Zyngier <maz@kernel.org>,
        James Morse <james.morse@arm.com>,
        Julien Thierry <julien.thierry.kdev@gmail.com>,
        Suzuki K Poulose <suzuki.poulose@arm.com>,
        Huacai Chen <chenhuacai@kernel.org>,
        Aleksandar Markovic <aleksandar.qemu.devel@gmail.com>,
        Paul Mackerras <paulus@ozlabs.org>,
        Christian Borntraeger <borntraeger@de.ibm.com>,
        Janosch Frank <frankja@linux.ibm.com>,
        David Hildenbrand <david@redhat.com>,
        Cornelia Huck <cohuck@redhat.com>,
        Claudio Imbrenda <imbrenda@linux.ibm.com>,
        Joerg Roedel <joro@8bytes.org>, kvm@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <ceb96527b6f7bb662eec813f05b897a551ebd0b2.1612140117.git.maciej.szmigiero@oracle.com>
 <4d748e0fd50bac68ece6952129aed319502b6853.1612140117.git.maciej.szmigiero@oracle.com>
 <YBisBkSYPoaOM42F@google.com>
 <9e6ca093-35c3-7cca-443b-9f635df4891d@maciej.szmigiero.name>
 <YBnjso2OeX1/r3Ls@google.com>
From:   Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <9caed824-a3ab-cd23-457e-cb35bb913f91@redhat.com>
Date:   Wed, 3 Feb 2021 14:43:57 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101
 Thunderbird/78.6.0
MIME-Version: 1.0
In-Reply-To: <YBnjso2OeX1/r3Ls@google.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 03/02/21 00:43, Sean Christopherson wrote:
>> But overall, this solution (and the one with id_to_index moved into the
>> main array, too) is still O(n) per memslot operation as you still need to
>> copy the array to either make space for the new memslot or to remove the
>> hole from the removed memslot.
> 
> Yes, but that problem that can be solved separately from the performance issue
> with hva lookups.

I know the world is not just QEMU, but I'll note that in QEMU all 
memslot operations are roughly O(n) anyway.

> Dumping everything into a single patch makes bisecting nearly worthless, e.g. if
> fast hva lookups breaks a non-x86 architecture, we should able to bisect to
> exactly that, not a massive patch that completely rewrites all of the memslot
> code in one fell swoop.
> 
> Mega patches with multiple logical changes are also extremely difficult to
> review.

Indeed.  This patch does get a +1 for having associated selftests, but 
the diffstat puts it back at the bottom of the pile. :)

Paolo