From: "Manwaring, Derek" <derekmn@amazon.com>
To: <seanjc@google.com>, <andrew.cooper3@citrix.com>,
<dave.hansen@linux.intel.com>
Cc: <ackerleytng@google.com>, <ajones@ventanamicro.com>,
<anup@brainfault.org>, <bfoster@redhat.com>, <brauner@kernel.org>,
<david@redhat.com>, <derekmn@amazon.com>, <erdemaktas@google.com>,
<fan.du@intel.com>, <fvdl@google.com>, <haibo1.xu@intel.com>,
<isaku.yamahata@intel.com>, <jgg@nvidia.com>,
<jgowans@amazon.com>, <jhubbard@nvidia.com>,
<jthoughton@google.com>, <jun.miao@intel.com>,
<kalyazin@amazon.co.uk>, <kent.overstreet@linux.dev>,
<kvm@vger.kernel.org>, <linux-fsdevel@kvack.org>,
<linux-kernel@vger.kernel.org>, <linux-kselftest@vger.kernel.org>,
<linux-mm@kvack.org>, <maciej.wieczor-retman@intel.com>,
<mike.kravetz@oracle.com>, <muchun.song@linux.dev>,
<oliver.upton@linux.dev>, <pbonzini@redhat.com>,
<peterx@redhat.com>, <pgonda@google.com>, <pvorel@suse.cz>,
<qperret@google.com>, <quic_eberman@quicinc.com>,
<richard.weiyang@gmail.com>, <rientjes@google.com>,
<roypat@amazon.co.uk>, <rppt@kernel.org>, <shuah@kernel.org>,
<tabba@google.com>, <vannapurve@google.com>,
<vkuznets@redhat.com>, <willy@infradead.org>,
<zhiquan1.li@intel.com>, <graf@amazon.de>, <mlipp@amazon.at>,
<canellac@amazon.at>
Subject: Re: [RFC PATCH 30/39] KVM: guest_memfd: Handle folio preparation for guest_memfd mmap
Date: Tue, 8 Oct 2024 20:51:13 -0700 [thread overview]
Message-ID: <ac337485-f8ab-45a4-b223-eb846e21c762@amazon.com> (raw)
In-Reply-To: <ZwWOfXd9becAm4lH@google.com>
On 2024-10-08 at 19:56+0000 Sean Christopherson wrote:
> Another (slightly crazy) approach would be use protection keys to provide the
> security properties that you want, while giving KVM (and userspace) a quick-and-easy
> override to access guest memory.
>
> 1. mmap() guest_memfd into userpace with RW protections
> 2. Configure PKRU to make guest_memfd memory inaccessible by default
> 3. Swizzle PKRU on-demand when intentionally accessing guest memory
>
> It's essentially the same idea as SMAP+STAC/CLAC, just applied to guest memory
> instead of to usersepace memory.
>
> The benefit of the PKRU approach is that there are no PTE modifications, and thus
> no TLB flushes, and only the CPU that is access guest memory gains temporary
> access. The big downside is that it would be limited to modern hardware, but
> that might be acceptable, especially if it simplifies KVM's implementation.
Yeah this might be worth it if it simplifies significantly. Jenkins et
al. showed MPK worked for stopping in-process Spectre V1 [1]. While
future hardware bugs are always possible, the host kernel would still
offer better protection overall since discovery of additional Spectre
approaches and gadgets in the kernel is more likely (I think it's a
bigger surface area than hardware-specific MPK transient execution
issues).
Patrick, we talked about this a couple weeks ago and ended up focusing
on within-userspace protection, but I see keys can also be used to stop
kernel access like Andrew's project he mentioned during Dave's MPK
session at LPC [2]. Andrew, could you share that here?
It's not clear to me how reliably the kernel prevents its own access to
such pages. I see a few papers that warrant more investigation:
"we found multiple interfaces that Linux, by design, provides for
accessing process memory that ignore PKU domains on a page." [3]
"Though Connor et al. demonstrate that existing MPK protections can be
bypassed by using the kernel as a confused deputy, compelling recent
work indicates that MPK operations can be made secure." [4]
Dave and others, if you're aware of resources clarifying how strong the
boundaries are, that would be helpful.
Derek
[1] https://www.cs.dartmouth.edu/~sws/pubs/jas2020.pdf
[2] https://www.youtube.com/watch?v=gEUeMfrNH94&t=1028s
[3] https://www.usenix.org/system/files/sec20-connor.pdf
[4] https://ics.uci.edu/~dabrowsa/kirth-eurosys22-pkru.pdf
next prev parent reply other threads:[~2024-10-09 3:51 UTC|newest]
Thread overview: 130+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-10 23:43 [RFC PATCH 00/39] 1G page support for guest_memfd Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 01/39] mm: hugetlb: Simplify logic in dequeue_hugetlb_folio_vma() Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 02/39] mm: hugetlb: Refactor vma_has_reserves() to should_use_hstate_resv() Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 03/39] mm: hugetlb: Remove unnecessary check for avoid_reserve Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 04/39] mm: mempolicy: Refactor out policy_node_nodemask() Ackerley Tng
2024-09-11 16:46 ` Gregory Price
2024-09-10 23:43 ` [RFC PATCH 05/39] mm: hugetlb: Refactor alloc_buddy_hugetlb_folio_with_mpol() to interpret mempolicy instead of vma Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 06/39] mm: hugetlb: Refactor dequeue_hugetlb_folio_vma() to use mpol Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 07/39] mm: hugetlb: Refactor out hugetlb_alloc_folio Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 08/39] mm: truncate: Expose preparation steps for truncate_inode_pages_final Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 09/39] mm: hugetlb: Expose hugetlb_subpool_{get,put}_pages() Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 10/39] mm: hugetlb: Add option to create new subpool without using surplus Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 11/39] mm: hugetlb: Expose hugetlb_acct_memory() Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 12/39] mm: hugetlb: Move and expose hugetlb_zero_partial_page() Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 13/39] KVM: guest_memfd: Make guest mem use guest mem inodes instead of anonymous inodes Ackerley Tng
2025-04-02 4:01 ` Yan Zhao
2025-04-23 20:22 ` Ackerley Tng
2025-04-24 3:53 ` Yan Zhao
2024-09-10 23:43 ` [RFC PATCH 14/39] KVM: guest_memfd: hugetlb: initialization and cleanup Ackerley Tng
2024-09-20 9:17 ` Vishal Annapurve
2024-10-01 23:00 ` Ackerley Tng
2024-12-01 17:59 ` Peter Xu
2025-02-13 9:47 ` Ackerley Tng
2025-02-26 18:55 ` Ackerley Tng
2025-03-06 17:33 ` Peter Xu
2024-09-10 23:43 ` [RFC PATCH 15/39] KVM: guest_memfd: hugetlb: allocate and truncate from hugetlb Ackerley Tng
2024-09-13 22:26 ` Elliot Berman
2024-10-03 20:23 ` Ackerley Tng
2024-10-30 9:01 ` Jun Miao
2025-02-11 1:21 ` Ackerley Tng
2024-12-01 17:55 ` Peter Xu
2025-02-13 7:52 ` Ackerley Tng
2025-02-13 16:48 ` Peter Xu
2024-09-10 23:43 ` [RFC PATCH 16/39] KVM: guest_memfd: Add page alignment check for hugetlb guest_memfd Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 17/39] KVM: selftests: Add basic selftests for hugetlb-backed guest_memfd Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 18/39] KVM: selftests: Support various types of backing sources for private memory Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 19/39] KVM: selftests: Update test for various private memory backing source types Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 20/39] KVM: selftests: Add private_mem_conversions_test.sh Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 21/39] KVM: selftests: Test that guest_memfd usage is reported via hugetlb Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 22/39] mm: hugetlb: Expose vmemmap optimization functions Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 23/39] mm: hugetlb: Expose HugeTLB functions for promoting/demoting pages Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 24/39] mm: hugetlb: Add functions to add/move/remove from hugetlb lists Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 25/39] KVM: guest_memfd: Split HugeTLB pages for guest_memfd use Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 26/39] KVM: guest_memfd: Track faultability within a struct kvm_gmem_private Ackerley Tng
2024-10-10 16:06 ` Peter Xu
2024-10-11 23:32 ` Ackerley Tng
2024-10-15 21:34 ` Peter Xu
2024-10-15 23:42 ` Ackerley Tng
2024-10-16 8:45 ` David Hildenbrand
2024-10-16 20:16 ` Peter Xu
2024-10-16 22:51 ` Jason Gunthorpe
2024-10-16 23:49 ` Peter Xu
2024-10-16 23:54 ` Jason Gunthorpe
2024-10-17 14:58 ` Peter Xu
2024-10-17 16:47 ` Jason Gunthorpe
2024-10-17 17:05 ` Peter Xu
2024-10-17 17:10 ` Jason Gunthorpe
2024-10-17 19:11 ` Peter Xu
2024-10-17 19:18 ` Jason Gunthorpe
2024-10-17 19:29 ` David Hildenbrand
2024-10-18 7:15 ` Patrick Roy
2024-10-18 7:50 ` David Hildenbrand
2024-10-18 9:34 ` Patrick Roy
2024-10-17 17:11 ` David Hildenbrand
2024-10-17 17:16 ` Jason Gunthorpe
2024-10-17 17:55 ` David Hildenbrand
2024-10-17 18:26 ` Vishal Annapurve
2024-10-17 14:56 ` David Hildenbrand
2024-10-17 15:02 ` David Hildenbrand
2024-10-16 8:50 ` David Hildenbrand
2024-10-16 10:48 ` Vishal Annapurve
2024-10-16 11:54 ` David Hildenbrand
2024-10-16 11:57 ` Jason Gunthorpe
2025-02-25 20:37 ` Peter Xu
2025-04-23 22:07 ` Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 27/39] KVM: guest_memfd: Allow mmapping guest_memfd files Ackerley Tng
2025-01-20 22:42 ` Peter Xu
2025-04-23 20:25 ` Ackerley Tng
2025-03-04 23:24 ` Peter Xu
2025-04-02 4:07 ` Yan Zhao
2025-04-23 20:28 ` Ackerley Tng
2024-09-10 23:43 ` [RFC PATCH 28/39] KVM: guest_memfd: Use vm_type to determine default faultability Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 29/39] KVM: Handle conversions in the SET_MEMORY_ATTRIBUTES ioctl Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 30/39] KVM: guest_memfd: Handle folio preparation for guest_memfd mmap Ackerley Tng
2024-09-16 20:00 ` Elliot Berman
2024-10-03 21:32 ` Ackerley Tng
2024-10-03 23:43 ` Ackerley Tng
2024-10-08 19:30 ` Sean Christopherson
2024-10-07 15:56 ` Patrick Roy
2024-10-08 18:07 ` Ackerley Tng
2024-10-08 19:56 ` Sean Christopherson
2024-10-09 3:51 ` Manwaring, Derek [this message]
2024-10-09 13:52 ` Andrew Cooper
2024-10-10 16:21 ` Patrick Roy
2024-10-10 19:27 ` Manwaring, Derek
2024-10-17 23:16 ` Ackerley Tng
2024-10-18 7:10 ` Patrick Roy
2024-09-10 23:44 ` [RFC PATCH 31/39] KVM: selftests: Allow vm_set_memory_attributes to be used without asserting return value of 0 Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 32/39] KVM: selftests: Test using guest_memfd memory from userspace Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 33/39] KVM: selftests: Test guest_memfd memory sharing between guest and host Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 34/39] KVM: selftests: Add notes in private_mem_kvm_exits_test for mmap-able guest_memfd Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 35/39] KVM: selftests: Test that pinned pages block KVM from setting memory attributes to PRIVATE Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 36/39] KVM: selftests: Refactor vm_mem_add to be more flexible Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 37/39] KVM: selftests: Add helper to perform madvise by memslots Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 38/39] KVM: selftests: Update private_mem_conversions_test for mmap()able guest_memfd Ackerley Tng
2024-09-10 23:44 ` [RFC PATCH 39/39] KVM: guest_memfd: Dynamically split/reconstruct HugeTLB page Ackerley Tng
2025-04-03 12:33 ` Yan Zhao
2025-04-23 22:02 ` Ackerley Tng
2025-04-24 1:09 ` Yan Zhao
2025-04-24 4:25 ` Yan Zhao
2025-04-24 5:55 ` Chenyi Qiang
2025-04-24 8:13 ` Yan Zhao
2025-04-24 14:10 ` Vishal Annapurve
2025-04-24 18:15 ` Ackerley Tng
2025-04-25 4:02 ` Yan Zhao
2025-04-25 22:45 ` Ackerley Tng
2025-04-28 1:05 ` Yan Zhao
2025-04-28 19:02 ` Vishal Annapurve
2025-04-30 20:09 ` Ackerley Tng
2025-05-06 1:23 ` Yan Zhao
2025-05-06 19:22 ` Ackerley Tng
2025-05-07 3:15 ` Yan Zhao
2025-05-13 17:33 ` Ackerley Tng
2024-09-11 6:56 ` [RFC PATCH 00/39] 1G page support for guest_memfd Michal Hocko
2024-09-14 1:08 ` Du, Fan
2024-09-14 13:34 ` Vishal Annapurve
2025-01-28 9:42 ` Amit Shah
2025-02-03 8:35 ` Ackerley Tng
2025-02-06 11:07 ` Amit Shah
2025-02-07 6:25 ` Ackerley Tng
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=ac337485-f8ab-45a4-b223-eb846e21c762@amazon.com \
--to=derekmn@amazon.com \
--cc=ackerleytng@google.com \
--cc=ajones@ventanamicro.com \
--cc=andrew.cooper3@citrix.com \
--cc=anup@brainfault.org \
--cc=bfoster@redhat.com \
--cc=brauner@kernel.org \
--cc=canellac@amazon.at \
--cc=dave.hansen@linux.intel.com \
--cc=david@redhat.com \
--cc=erdemaktas@google.com \
--cc=fan.du@intel.com \
--cc=fvdl@google.com \
--cc=graf@amazon.de \
--cc=haibo1.xu@intel.com \
--cc=isaku.yamahata@intel.com \
--cc=jgg@nvidia.com \
--cc=jgowans@amazon.com \
--cc=jhubbard@nvidia.com \
--cc=jthoughton@google.com \
--cc=jun.miao@intel.com \
--cc=kalyazin@amazon.co.uk \
--cc=kent.overstreet@linux.dev \
--cc=kvm@vger.kernel.org \
--cc=linux-fsdevel@kvack.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=maciej.wieczor-retman@intel.com \
--cc=mike.kravetz@oracle.com \
--cc=mlipp@amazon.at \
--cc=muchun.song@linux.dev \
--cc=oliver.upton@linux.dev \
--cc=pbonzini@redhat.com \
--cc=peterx@redhat.com \
--cc=pgonda@google.com \
--cc=pvorel@suse.cz \
--cc=qperret@google.com \
--cc=quic_eberman@quicinc.com \
--cc=richard.weiyang@gmail.com \
--cc=rientjes@google.com \
--cc=roypat@amazon.co.uk \
--cc=rppt@kernel.org \
--cc=seanjc@google.com \
--cc=shuah@kernel.org \
--cc=tabba@google.com \
--cc=vannapurve@google.com \
--cc=vkuznets@redhat.com \
--cc=willy@infradead.org \
--cc=zhiquan1.li@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
all inboxes | Powered by JetHome®