From: "Huang, Kai" <kai.huang@intel.com>
To: "kvm@vger.kernel.org" <kvm@vger.kernel.org>,
"Hansen, Dave" <dave.hansen@intel.com>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Cc: "Luck, Tony" <tony.luck@intel.com>,
"bagasdotme@gmail.com" <bagasdotme@gmail.com>,
"ak@linux.intel.com" <ak@linux.intel.com>,
"Wysocki, Rafael J" <rafael.j.wysocki@intel.com>,
"kirill.shutemov@linux.intel.com"
<kirill.shutemov@linux.intel.com>, "Christopherson,,
Sean" <seanjc@google.com>,
"Chatre, Reinette" <reinette.chatre@intel.com>,
"pbonzini@redhat.com" <pbonzini@redhat.com>,
"linux-mm@kvack.org" <linux-mm@kvack.org>,
"Yamahata, Isaku" <isaku.yamahata@intel.com>,
"tglx@linutronix.de" <tglx@linutronix.de>,
"Shahar, Sagi" <sagis@google.com>,
"imammedo@redhat.com" <imammedo@redhat.com>,
"Gao, Chao" <chao.gao@intel.com>,
"Brown, Len" <len.brown@intel.com>,
"peterz@infradead.org" <peterz@infradead.org>,
"sathyanarayanan.kuppuswamy@linux.intel.com"
<sathyanarayanan.kuppuswamy@linux.intel.com>,
"Huang, Ying" <ying.huang@intel.com>,
"Williams, Dan J" <dan.j.williams@intel.com>
Subject: Re: [PATCH v8 15/16] x86/virt/tdx: Flush cache in kexec() when TDX is enabled
Date: Wed, 11 Jan 2023 00:13:45 +0000 [thread overview]
Message-ID: <b84232220d03889321248ffb82739c64204cc4af.camel@intel.com> (raw)
In-Reply-To: <944ffd4b-3090-e068-a649-b9a84add8395@intel.com>
On Tue, 2023-01-10 at 07:27 -0800, Dave Hansen wrote:
> On 1/10/23 03:29, Huang, Kai wrote:
> > On Fri, 2023-01-06 at 16:35 -0800, Dave Hansen wrote:
> > > On 12/8/22 22:52, Kai Huang wrote:
> ...
> > > > However, this implementation doesn't convert TDX private pages back to
> > > > normal in kexec() because of below considerations:
> > > >
> > > > 1) Neither the kernel nor the TDX module has existing infrastructure to
> > > > track which pages are TDX private pages.
> > > > 2) The number of TDX private pages can be large, and converting all of
> > > > them (cache flush + using MOVDIR64B to clear the page) in kexec() can
> > > > be time consuming.
> > > > 3) The new kernel will almost only use KeyID 0 to access memory. KeyID
> > > > 0 doesn't support integrity-check, so it's OK.
> > > > 4) The kernel doesn't (and may never) support MKTME. If any 3rd party
> > > > kernel ever supports MKTME, it can/should do MOVDIR64B to clear the
> > > > page with the new MKTME KeyID (just like TDX does) before using it.
> > >
> > > Yeah, why are we getting all worked up about MKTME when there is not
> > > support?
> >
> > I am not sure whether we need to consider 3rd party kernel case?
>
> No, we don't.
Good to know.
>
> > > The only thing that matters here is dirty cacheline writeback. There
> > > are two things the kernel needs to do to mitigate that:
> > >
> > > 1. Stop accessing TDX private memory mappings
> > > 1a. Stop making TDX module calls (uses global private KeyID)
> > > 1b. Stop TDX guests from running (uses per-guest KeyID)
> > > 2. Flush any cachelines from previous private KeyID writes
> > >
> > > There are a couple of ways we can do #2. We do *NOT* need to convert
> > > *ANYTHING* back to KeyID 0. Page conversion doesn't even come into play
> > > in any way as far as I can tell.
> >
> > May I ask why? When I was writing this patch I was not sure whether kexec()
> > should give the new kernel a clean slate. SGX driver doesn't EREMOVE all EPC
> > during kexec() but depends on the new kernel to do that too, but I don't know
> > what's the general guide of supporting kexec().
>
> Think about it this way: kexec() is modifying persistent (across kexec)
> state to get the system ready for the new kernel. The caches are
> persistent state. Devices have persistent state. Memory state persists
> across kexec(). The memory integrity metadata persists.
>
> What persistent state does a conversion to KeyID-0 affect? It resets
> the integrity metadata and the memory contents.
>
> Kexec leaves memory contents in place and doesn't zero them, so memory
> contents don't matter. The integrity metadata also doesn't matter
> because the memory will be used as KeyID-0 and that KeyID doesn't read
> the integrity metadata.
Right. So I guess we just need to call out the new kernel will use memory as
KeyID-0?
>
> What practical impact does a conversion back to KeyID-0 serve? What
> persistent state does it affect that matters?
If we can be sure the new kernel will use KeyID-0, then we don't need to
convert. In the 3) and 4) in my changelog, I actually was trying to convery
this.
>
> > > I think you're also saying that since all CPUs go through this path and
> > > there is no TDX activity between the WBINVD and the native_halt() that
> > > 1a and 1b basically happen for "free" without needing to do theme
> > > explicitly.
> >
> > Yes. Should we mention this part in changelog?
>
> That would be nice.
>
Will do.
next prev parent reply other threads:[~2023-01-11 0:14 UTC|newest]
Thread overview: 84+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-12-09 6:52 [PATCH v8 00/16] TDX host kernel support Kai Huang
2022-12-09 6:52 ` [PATCH v8 01/16] x86/tdx: Define TDX supported page sizes as macros Kai Huang
2023-01-06 19:04 ` Dave Hansen
2022-12-09 6:52 ` [PATCH v8 02/16] x86/virt/tdx: Detect TDX during kernel boot Kai Huang
2023-01-06 17:09 ` Dave Hansen
2023-01-08 22:25 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 03/16] x86/virt/tdx: Make INTEL_TDX_HOST depend on X86_X2APIC Kai Huang
2023-01-06 19:04 ` Dave Hansen
2022-12-09 6:52 ` [PATCH v8 04/16] x86/virt/tdx: Add skeleton to initialize TDX on demand Kai Huang
2023-01-06 17:14 ` Dave Hansen
2023-01-08 22:26 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 05/16] x86/virt/tdx: Implement functions to make SEAMCALL Kai Huang
2023-01-06 17:29 ` Dave Hansen
2023-01-09 10:30 ` Huang, Kai
2023-01-09 19:54 ` Dave Hansen
2023-01-09 22:10 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 06/16] x86/virt/tdx: Get information about TDX module and TDX-capable memory Kai Huang
2023-01-06 17:46 ` Dave Hansen
2023-01-09 10:25 ` Huang, Kai
2023-01-09 19:52 ` Dave Hansen
2023-01-09 22:07 ` Huang, Kai
2023-01-09 22:11 ` Dave Hansen
2022-12-09 6:52 ` [PATCH v8 07/16] x86/virt/tdx: Use all system memory when initializing TDX module as TDX memory Kai Huang
2023-01-06 18:18 ` Dave Hansen
2023-01-09 11:48 ` Huang, Kai
2023-01-09 16:51 ` Dave Hansen
2023-01-10 12:09 ` Huang, Kai
2023-01-10 16:18 ` Dave Hansen
2023-01-11 10:00 ` Huang, Kai
2023-01-12 0:56 ` Huang, Ying
2023-01-12 1:18 ` Huang, Kai
2023-01-12 1:59 ` Huang, Ying
2023-01-12 2:22 ` Huang, Kai
2023-01-12 11:33 ` Huang, Kai
2023-01-18 11:08 ` Huang, Kai
2023-01-18 13:57 ` David Hildenbrand
2023-01-18 19:38 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 08/16] x86/virt/tdx: Add placeholder to construct TDMRs to cover all TDX memory regions Kai Huang
2023-01-06 19:24 ` Dave Hansen
2023-01-10 0:40 ` Huang, Kai
2023-01-10 0:47 ` Dave Hansen
2023-01-10 2:23 ` Huang, Kai
2023-01-10 19:12 ` Dave Hansen
2023-01-11 9:23 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 09/16] x86/virt/tdx: Fill out " Kai Huang
2023-01-06 19:36 ` Dave Hansen
2023-01-10 0:45 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 10/16] x86/virt/tdx: Allocate and set up PAMTs for TDMRs Kai Huang
2023-01-06 21:53 ` Dave Hansen
2023-01-10 0:49 ` Huang, Kai
2023-01-07 0:47 ` Dave Hansen
2023-01-10 0:47 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 11/16] x86/virt/tdx: Designate reserved areas for all TDMRs Kai Huang
2023-01-06 22:07 ` Dave Hansen
2023-01-10 1:19 ` Huang, Kai
2023-01-10 1:22 ` Dave Hansen
2023-01-10 11:01 ` Huang, Kai
2023-01-10 15:19 ` Dave Hansen
2023-01-11 10:57 ` Huang, Kai
2023-01-11 16:16 ` Dave Hansen
2023-01-11 22:10 ` Huang, Kai
2023-01-10 11:01 ` Huang, Kai
2023-01-10 15:17 ` Dave Hansen
2022-12-09 6:52 ` [PATCH v8 12/16] x86/virt/tdx: Designate the global KeyID and configure the TDX module Kai Huang
2023-01-06 22:21 ` Dave Hansen
2023-01-10 10:48 ` Huang, Kai
2023-01-10 16:25 ` Dave Hansen
2023-01-10 23:33 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 13/16] x86/virt/tdx: Configure global KeyID on all packages Kai Huang
2023-01-06 22:49 ` Dave Hansen
2023-01-10 10:15 ` Huang, Kai
2023-01-10 16:53 ` Dave Hansen
2023-01-11 0:06 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 14/16] x86/virt/tdx: Initialize all TDMRs Kai Huang
2023-01-07 0:17 ` Dave Hansen
2023-01-10 10:23 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 15/16] x86/virt/tdx: Flush cache in kexec() when TDX is enabled Kai Huang
2023-01-07 0:35 ` Dave Hansen
2023-01-10 11:29 ` Huang, Kai
2023-01-10 15:27 ` Dave Hansen
2023-01-11 0:13 ` Huang, Kai [this message]
2023-01-11 0:30 ` Dave Hansen
2023-01-11 1:58 ` Huang, Kai
2022-12-09 6:52 ` [PATCH v8 16/16] Documentation/x86: Add documentation for TDX host support Kai Huang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=b84232220d03889321248ffb82739c64204cc4af.camel@intel.com \
--to=kai.huang@intel.com \
--cc=ak@linux.intel.com \
--cc=bagasdotme@gmail.com \
--cc=chao.gao@intel.com \
--cc=dan.j.williams@intel.com \
--cc=dave.hansen@intel.com \
--cc=imammedo@redhat.com \
--cc=isaku.yamahata@intel.com \
--cc=kirill.shutemov@linux.intel.com \
--cc=kvm@vger.kernel.org \
--cc=len.brown@intel.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=pbonzini@redhat.com \
--cc=peterz@infradead.org \
--cc=rafael.j.wysocki@intel.com \
--cc=reinette.chatre@intel.com \
--cc=sagis@google.com \
--cc=sathyanarayanan.kuppuswamy@linux.intel.com \
--cc=seanjc@google.com \
--cc=tglx@linutronix.de \
--cc=tony.luck@intel.com \
--cc=ying.huang@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).