linux-kernel - [PATCH v2 00/15] KVM: x86: VMX: Fix MSR namespacing

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: <20200923180409.32255-1-sean.j.christopherson@intel.com>
Date:   Wed, 23 Sep 2020 11:03:54 -0700
From:   Sean Christopherson <sean.j.christopherson@...el.com>
To:     Paolo Bonzini <pbonzini@...hat.com>
Cc:     Sean Christopherson <sean.j.christopherson@...el.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
Subject: [PATCH v2 00/15] KVM: x86: VMX: Fix MSR namespacing

This series attempts to clean up VMX's MSR namespacing, which is in
unimitigated disaster (keeping things PG).

There are a variety of ways VMX saves and restores guest MSRs, all with
unique properties and mechanisms, but with haphazard namespacing (assuming
there is any namespacing at all).  Some fun collisions:

  __find_msr_index(), find_msr_entry() and vmx_find_msr_index()

  vmx_set_guest_msr() and vmx_set_msr()

  structs vmx_msrs, vmx_msr_entry, shared_msr_entry, kvm_shared_msrs and
  kvm_shared_msrs_values

  vcpu_vmx fields guest_msrs, msr_autoload.guest and msr_autostore.guest

Probably the most infurating/confusing nomenclature is "index", which can
mean MSR's ECX index, index into one of several VMX arrays, or index into
a common x86 array.  __find_msr_index() even manages to mix at least three
different meanings in about as many lines of code.

The biggest change is to rename the "shared MSRs" mechanism to "user
return MSRs" (details in patch 1), most everything else is either derived
from that rename or is fairly straightforward cleanup.

No true functional changes, although the update_transition_efer() change
in patch 10 dances pretty close to being a functional change.

v2:
  - Rebased to kvm/queue, commit e1ba1a15af73 ("KVM: SVM: Enable INVPCID
    feature on AMD").

Sean Christopherson (15):
  KVM: x86: Rename "shared_msrs" to "user_return_msrs"
  KVM: VMX: Prepend "MAX_" to MSR array size defines
  KVM: VMX: Rename "vmx_find_msr_index" to "vmx_find_loadstore_msr_slot"
  KVM: VMX: Rename the "shared_msr_entry" struct to "vmx_uret_msr"
  KVM: VMX: Rename vcpu_vmx's "nmsrs" to "nr_uret_msrs"
  KVM: VMX: Rename vcpu_vmx's "save_nmsrs" to "nr_active_uret_msrs"
  KVM: VMX: Rename vcpu_vmx's "guest_msrs_ready" to
    "guest_uret_msrs_loaded"
  KVM: VMX: Rename "__find_msr_index" to "__vmx_find_uret_msr"
  KVM: VMX: Check guest support for RDTSCP before processing MSR_TSC_AUX
  KVM: VMX: Move uret MSR lookup into update_transition_efer()
  KVM: VMX: Add vmx_setup_uret_msr() to handle lookup and swap
  KVM: VMX: Rename "find_msr_entry" to "vmx_find_uret_msr"
  KVM: VMX: Rename "vmx_set_guest_msr" to "vmx_set_guest_uret_msr"
  KVM: VMX: Rename "vmx_msr_index" to "vmx_uret_msrs_list"
  KVM: VMX: Rename vmx_uret_msr's "index" to "slot"

 arch/x86/include/asm/kvm_host.h |   4 +-
 arch/x86/kvm/vmx/nested.c       |  22 ++--
 arch/x86/kvm/vmx/vmx.c          | 184 ++++++++++++++++----------------
 arch/x86/kvm/vmx/vmx.h          |  24 ++---
 arch/x86/kvm/x86.c              | 101 +++++++++---------
 5 files changed, 168 insertions(+), 167 deletions(-)

-- 
2.28.0