[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <fa85ac0cf3e6fae190dca953006d57c02fac6978.camel@intel.com>
Date: Fri, 9 May 2025 23:34:39 +0000
From: "Edgecombe, Rick P" <rick.p.edgecombe@...el.com>
To: "pbonzini@...hat.com" <pbonzini@...hat.com>, "seanjc@...gle.com"
<seanjc@...gle.com>, "Zhao, Yan Y" <yan.y.zhao@...el.com>
CC: "Shutemov, Kirill" <kirill.shutemov@...el.com>, "quic_eberman@...cinc.com"
<quic_eberman@...cinc.com>, "Li, Xiaoyao" <xiaoyao.li@...el.com>,
"kvm@...r.kernel.org" <kvm@...r.kernel.org>, "Hansen, Dave"
<dave.hansen@...el.com>, "david@...hat.com" <david@...hat.com>,
"thomas.lendacky@....com" <thomas.lendacky@....com>, "tabba@...gle.com"
<tabba@...gle.com>, "Li, Zhiquan1" <zhiquan1.li@...el.com>, "Du, Fan"
<fan.du@...el.com>, "linux-kernel@...r.kernel.org"
<linux-kernel@...r.kernel.org>, "michael.roth@....com"
<michael.roth@....com>, "Weiny, Ira" <ira.weiny@...el.com>, "vbabka@...e.cz"
<vbabka@...e.cz>, "binbin.wu@...ux.intel.com" <binbin.wu@...ux.intel.com>,
"ackerleytng@...gle.com" <ackerleytng@...gle.com>, "Yamahata, Isaku"
<isaku.yamahata@...el.com>, "Peng, Chao P" <chao.p.peng@...el.com>,
"Annapurve, Vishal" <vannapurve@...gle.com>, "jroedel@...e.de"
<jroedel@...e.de>, "Miao, Jun" <jun.miao@...el.com>, "pgonda@...gle.com"
<pgonda@...gle.com>, "x86@...nel.org" <x86@...nel.org>
Subject: Re: [RFC PATCH 18/21] KVM: x86: Split huge boundary leafs before
private to shared conversion
On Thu, 2025-04-24 at 11:08 +0800, Yan Zhao wrote:
> Before converting a GFN range from private to shared, it is necessary to
> zap the mirror page table. When huge pages are supported and the GFN range
> intersects with a huge leaf, split the huge leaf to prevent zapping GFNs
> outside the conversion range.
FALLOC_FL_PUNCH_HOLE demotion failure doesn't look like it is addressed in this
series. I noticed that mmu notifier failures are allowed to be handled by
blocking until success is possible, in most cases. KVM just doesn't need to
because it can't fail. We could think about doing retries for
FALLOC_FL_PUNCH_HOLE, while checking for signals. Or adding a ENOMEM error code
to fallocate.
Powered by blists - more mailing lists