[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ZPWBVxs77rtDWfvx@yzhao56-desk.sh.intel.com>
Date: Mon, 4 Sep 2023 15:03:51 +0800
From: Yan Zhao <yan.y.zhao@...el.com>
To: liulongfang <liulongfang@...wei.com>
CC: <linux-mm@...ck.org>, <linux-kernel@...r.kernel.org>,
<kvm@...r.kernel.org>, <pbonzini@...hat.com>, <seanjc@...gle.com>,
<mike.kravetz@...cle.com>, <apopple@...dia.com>, <jgg@...dia.com>,
<rppt@...nel.org>, <akpm@...ux-foundation.org>,
<kevin.tian@...el.com>
Subject: Re: [RFC PATCH 3/3] KVM: x86/mmu: skip zap maybe-dma-pinned pages
for NUMA migration
> I have a question. The numa id of the cpu can be reconfigured in the VM.
Are you talking about vNUMA and the numa balancing in guest?
> Will the page table migration operations initiated by the numa balance in the
> VM and the numa balance in the host conflict with each other after this setting?
There's no page table migration in host. Only page table is modified and
page is migrated in a single host.
Live migration also doesn't copy the secondary page tables (i.e.
EPT/NPT) to the target. Instead, it's rebuilt in the target side.
So, I don't quite understand your question here.
Powered by blists - more mailing lists