From: Sean Christopherson <sean.j.christopherson@intel.com>
To: Ben Gardon <bgardon@google.com>
Cc: kvm@vger.kernel.org, Paolo Bonzini <pbonzini@redhat.com>,
Peter Feiner <pfeiner@google.com>,
Peter Shier <pshier@google.com>,
Junaid Shahid <junaids@google.com>,
Jim Mattson <jmattson@google.com>
Subject: Re: [RFC PATCH 00/28] kvm: mmu: Rework the x86 TDP direct mapped case
Date: Thu, 17 Oct 2019 11:50:02 -0700 [thread overview]
Message-ID: <20191017185002.GG20903@linux.intel.com> (raw)
In-Reply-To: <20190926231824.149014-1-bgardon@google.com>
On Thu, Sep 26, 2019 at 04:17:56PM -0700, Ben Gardon wrote:
> Over the years, the needs for KVM's x86 MMU have grown from running small
> guests to live migrating multi-terabyte VMs with hundreds of vCPUs. Where
> we previously depended upon shadow paging to run all guests, we now have
> the use of two dimensional paging (TDP). This RFC proposes and
> demonstrates two major changes to the MMU. First, an iterator abstraction
> that simplifies traversal of TDP paging structures when running an L1
> guest. This abstraction takes advantage of the relative simplicity of TDP
> to simplify the implementation of MMU functions. Second, this RFC changes
> the synchronization model to enable more parallelism than the monolithic
> MMU lock. This "direct mode" MMU is currently in use at Google and has
> given us the performance necessary to live migrate our 416 vCPU, 12TiB
> m2-ultramem-416 VMs.
>
> The primary motivation for this work was to handle page faults in
> parallel. When VMs have hundreds of vCPUs and terabytes of memory, KVM's
> MMU lock suffers from extreme contention, resulting in soft-lockups and
> jitter in the guest. To demonstrate this I also written, and will submit
> a demand paging test to KVM selftests. The test creates N vCPUs, which
> each touch disjoint regions of memory. Page faults are picked up by N
> user fault FD handlers, one for each vCPU. Over a 1 second profile of
> the demand paging test, with 416 vCPUs and 4G per vCPU, 98% of the
> execution time was spent waiting for the MMU lock! With this patch
> series the total execution time for the test was reduced by 89% and the
> execution was dominated by get_user_pages and the user fault FD ioctl.
> As a secondary benefit, the iterator-based implementation does not use
> the rmap or struct kvm_mmu_pages, saving ~0.2% of guest memory in KVM
> overheads.
>
> The goal of this RFC is to demonstrate and gather feedback on the
> iterator pattern, the memory savings it enables for the "direct case"
> and the changes to the synchronization model. Though they are interwoven
> in this series, I will separate the iterator from the synchronization
> changes in a future series. I recognize that some feature work will be
> needed to make this patch set ready for merging. That work is detailed
> at the end of this cover letter.
Diving into this series is on my todo list, but realistically that's not
going to happen until after KVM forum. Sorry I can't provide timely
feedback.
next prev parent reply other threads:[~2019-10-17 18:50 UTC|newest]
Thread overview: 57+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-09-26 23:17 [RFC PATCH 00/28] kvm: mmu: Rework the x86 TDP direct mapped case Ben Gardon
2019-09-26 23:17 ` [RFC PATCH 01/28] kvm: mmu: Separate generating and setting mmio ptes Ben Gardon
2019-11-27 18:15 ` Sean Christopherson
2019-09-26 23:17 ` [RFC PATCH 02/28] kvm: mmu: Separate pte generation from set_spte Ben Gardon
2019-11-27 18:25 ` Sean Christopherson
2019-09-26 23:17 ` [RFC PATCH 03/28] kvm: mmu: Zero page cache memory at allocation time Ben Gardon
2019-11-27 18:32 ` Sean Christopherson
2019-09-26 23:18 ` [RFC PATCH 04/28] kvm: mmu: Update the lpages stat atomically Ben Gardon
2019-11-27 18:39 ` Sean Christopherson
2019-12-06 20:10 ` Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 05/28] sched: Add cond_resched_rwlock Ben Gardon
2019-11-27 18:42 ` Sean Christopherson
2019-12-06 20:12 ` Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 06/28] kvm: mmu: Replace mmu_lock with a read/write lock Ben Gardon
2019-11-27 18:47 ` Sean Christopherson
2019-12-02 22:45 ` Sean Christopherson
2019-09-26 23:18 ` [RFC PATCH 07/28] kvm: mmu: Add functions for handling changed PTEs Ben Gardon
2019-11-27 19:04 ` Sean Christopherson
2019-09-26 23:18 ` [RFC PATCH 08/28] kvm: mmu: Init / Uninit the direct MMU Ben Gardon
2019-12-02 23:40 ` Sean Christopherson
2019-12-06 20:25 ` Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 09/28] kvm: mmu: Free direct MMU page table memory in an RCU callback Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 10/28] kvm: mmu: Flush TLBs before freeing direct MMU page table memory Ben Gardon
2019-12-02 23:46 ` Sean Christopherson
2019-12-06 20:31 ` Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 11/28] kvm: mmu: Optimize for freeing direct MMU PTs on teardown Ben Gardon
2019-12-02 23:54 ` Sean Christopherson
2019-09-26 23:18 ` [RFC PATCH 12/28] kvm: mmu: Set tlbs_dirty atomically Ben Gardon
2019-12-03 0:13 ` Sean Christopherson
2019-09-26 23:18 ` [RFC PATCH 13/28] kvm: mmu: Add an iterator for concurrent paging structure walks Ben Gardon
2019-12-03 2:15 ` Sean Christopherson
2019-12-18 18:25 ` Ben Gardon
2019-12-18 19:14 ` Sean Christopherson
2019-09-26 23:18 ` [RFC PATCH 14/28] kvm: mmu: Batch updates to the direct mmu disconnected list Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 15/28] kvm: mmu: Support invalidate_zap_all_pages Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 16/28] kvm: mmu: Add direct MMU page fault handler Ben Gardon
2020-01-08 17:20 ` Peter Xu
2020-01-08 18:15 ` Ben Gardon
2020-01-08 19:00 ` Peter Xu
2019-09-26 23:18 ` [RFC PATCH 17/28] kvm: mmu: Add direct MMU fast " Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 18/28] kvm: mmu: Add an hva range iterator for memslot GFNs Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 19/28] kvm: mmu: Make address space ID a property of memslots Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 20/28] kvm: mmu: Implement the invalidation MMU notifiers for the direct MMU Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 21/28] kvm: mmu: Integrate the direct mmu with the changed pte notifier Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 22/28] kvm: mmu: Implement access tracking for the direct MMU Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 23/28] kvm: mmu: Make mark_page_dirty_in_slot usable from outside kvm_main Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 24/28] kvm: mmu: Support dirty logging in the direct MMU Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 25/28] kvm: mmu: Support kvm_zap_gfn_range " Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 26/28] kvm: mmu: Integrate direct MMU with nesting Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 27/28] kvm: mmu: Lazily allocate rmap when direct MMU is enabled Ben Gardon
2019-09-26 23:18 ` [RFC PATCH 28/28] kvm: mmu: Support MMIO in the direct MMU Ben Gardon
2019-10-17 18:50 ` Sean Christopherson [this message]
2019-10-18 13:42 ` [RFC PATCH 00/28] kvm: mmu: Rework the x86 TDP direct mapped case Paolo Bonzini
2019-11-27 19:09 ` Sean Christopherson
2019-12-06 19:55 ` Ben Gardon
2019-12-06 19:57 ` Sean Christopherson
2019-12-06 20:42 ` Ben Gardon
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=20191017185002.GG20903@linux.intel.com \
--to=sean.j.christopherson@intel.com \
--cc=bgardon@google.com \
--cc=jmattson@google.com \
--cc=junaids@google.com \
--cc=kvm@vger.kernel.org \
--cc=pbonzini@redhat.com \
--cc=pfeiner@google.com \
--cc=pshier@google.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).