linux-kernel - [PATCH v2 0/4] KVM: x86/mmu: Cleanups for eager page splitting

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <20220624213039.2872507-1-seanjc@google.com>
Date:   Fri, 24 Jun 2022 21:30:35 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Paolo Bonzini <pbonzini@...hat.com>
Cc:     Sean Christopherson <seanjc@...gle.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org, David Matlack <dmatlack@...gle.com>
Subject: [PATCH v2 0/4] KVM: x86/mmu: Cleanups for eager page splitting

Eager page splitting cleanups for a few minor things that were noted in
code review but didn't make it into the committed code.

Patch 1 goes beyond the eager page splitting series, but it's still very
much relevant.

The last patch in particular is a bit more urgent than I first realized.
I had forgotten that pte_list_desc is now 128 bytes, and I also had a
brain fart and thought it was just allocating pointers, i.e. 8 bytes.
In other words, I was thinking the 513 object buffer was "only" wasting
~8kb per VM, whereas it actually costs ~64kb per VM.

v2:
 - Add helper to compute SPTE's index into its parent. [David].
 - Collect reviews [David].
 - Tweak the comment in the last patch after more analysis.

v1:
 - https://lore.kernel.org/all/20220624171808.2845941-1-seanjc@google.com

Sean Christopherson (4):
  KVM: x86/mmu: Add optimized helper to retrieve an SPTE's index
  KVM: x86/mmu: Expand quadrant comment for PG_LEVEL_4K shadow pages
  KVM: x86/mmu: Use "unsigned int", not "u32", for SPTEs' @access info
  KVM: x86/mmu: Buffer nested MMU split_desc_cache only by default
    capacity

 arch/x86/kvm/mmu/mmu.c         | 74 +++++++++++++++++++++-------------
 arch/x86/kvm/mmu/paging_tmpl.h |  4 +-
 arch/x86/kvm/mmu/spte.h        |  6 +++
 3 files changed, 53 insertions(+), 31 deletions(-)


base-commit: 4b88b1a518b337de1252b8180519ca4c00015c9e
-- 
2.37.0.rc0.161.g10f37bed90-goog