Received-SPF: pass (google.com: domain of linux-kernel+bounces-147266-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.80.249 as permitted sender) client-ip=147.75.80.249;
Precedence: bulk
MIME-Version: 1.0
References: <20240411203529.1866998-1-seanjc@google.com>
In-Reply-To: <20240411203529.1866998-1-seanjc@google.com>
From: Paolo Bonzini <pbonzini@redhat.com>
Date: Tue, 16 Apr 2024 18:55:05 +0200
Message-ID: <CABgObfanoru0cEHFbNOw9QDv8EWAkye4YPzfyDMyBmUwpJjiTA@mail.gmail.com>
Subject: Re: [GIT PULL] KVM: x86: Fixes for 6.9-rcN
To: Sean Christopherson <seanjc@google.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Thu, Apr 11, 2024 at 10:35=E2=80=AFPM Sean Christopherson <seanjc@google=
com> wrote:
>
> Please pull a big pile of fixes for 6.9.  Many of these were sent even be=
fore
> the 6.9 merge window, but I was on vacation until rc2, and things piled u=
p.
>
> The back half of the commits were _just_ rebased to drop my version of th=
e
> LVTPC masking fixes, but that's your fault. :-)  For giggles, I also push=
ed
> kvm-x86 tags/kvm-x86-fixed-6.9-rcN-unrebased if you or anyone else want a=
 paper
> trail for the pre-rebase commits.
>
> Note, there's a perf change in here that didn't get an Ack from anyone, b=
ut the
> fixes have been on-list for over a month, and I can't imagine anyone obje=
cting
> to adding a new feature flag to x86_pmu_capability, which for all intents=
 and
> purposes exists purely for KVM.
>
> Thanks!

Pulled, thanks.

Paolo

> The following changes since commit fec50db7033ea478773b159e0e2efb135270e3=
b7:
>
>   Linux 6.9-rc3 (2024-04-07 13:22:46 -0700)
>
> are available in the Git repository at:
>
>   https://github.com/kvm-x86/linux.git tags/kvm-x86-fixes-6.9-rcN
>
> for you to fetch changes up to eefb85b3f0310c2f4149c50cb9b13094ed1dde25:
>
>   KVM: Drop unused @may_block param from gfn_to_pfn_cache_invalidate_star=
t() (2024-04-11 12:58:53 -0700)
>
> ----------------------------------------------------------------
> KVM fixes for 6.9-rcN:
>
>  - Fix a mostly benign bug in the gfn_to_pfn_cache infrastructure where K=
VM
>    would allow userspace to refresh the cache with a bogus GPA.  The bug =
has
>    existed for quite some time, but was exposed by a new sanity check add=
ed in
>    6.9 (to ensure a cache is either GPA-based or HVA-based).
>
>  - Drop an unused param from gfn_to_pfn_cache_invalidate_start() that got=
 left
>    behind during a 6.9 cleanup.
>
>  - Disable support for virtualizing adaptive PEBS, as KVM's implementatio=
n is
>    architecturally broken and can leak host LBRs to the guest.
>
>  - Fix a bug where KVM neglects to set the enable bits for general purpos=
e
>    counters in PERF_GLOBAL_CTRL when initializing the virtual PMU.  Both =
Intel
>    and AMD architectures require the bits to be set at RESET in order for=
 v2
>    PMUs to be backwards compatible with software that was written for v1 =
PMUs,
>    i.e. for software that will never manually set the global enables.
>
>  - Disable LBR virtualization on CPUs that don't support LBR callstacks, =
as
>    KVM unconditionally uses PERF_SAMPLE_BRANCH_CALL_STACK when creating t=
he
>    virtual LBR perf event, i.e. KVM will always fail to create LBR events=
 on
>    such CPUs.
>
>  - Fix a math goof in x86's hugepage logic for KVM_SET_MEMORY_ATTRIBUTES =
that
>    results in an array overflow (detected by KASAN).
>
>  - Fix a flaw in the max_guest_memory selftest that results in it exhaust=
ing
>    the supply of ucall structures when run with more than 256 vCPUs.
>
>  - Mark KVM_MEM_READONLY as supported for RISC-V in set_memory_region_tes=
t.
>
>  - Fix a bug where KVM incorrectly thinks a TDP MMU root is an indirect s=
hadow
>    root due KVM unnecessarily clobbering root_role.direct when userspace =
sets
>    guest CPUID.
>
>  - Fix a dirty logging bug in the where KVM fails to write-protect TDP MM=
U
>    SPTEs used for L2 if Page-Modification Logging is enabled for L1 and t=
he L1
>    hypervisor is NOT using EPT (if nEPT is enabled, KVM doesn't use the T=
DP MMU
>    to run L2).  For simplicity, KVM always disables PML when running L2, =
but
>    the TDP MMU wasn't accounting for root-specific conditions that force =
write-
>    protect based dirty logging.
>
> ----------------------------------------------------------------
> Andrew Jones (1):
>       KVM: selftests: fix supported_flags for riscv
>
> David Matlack (4):
>       KVM: x86/mmu: Write-protect L2 SPTEs in TDP MMU when clearing dirty=
 status
>       KVM: x86/mmu: Remove function comments above clear_dirty_{gfn_range=
,pt_masked}()
>       KVM: x86/mmu: Fix and clarify comments about clearing D-bit vs. wri=
te-protecting
>       KVM: selftests: Add coverage of EPT-disabled to vmx_dirty_log_test
>
> Maxim Levitsky (1):
>       KVM: selftests: fix max_guest_memory_test with more that 256 vCPUs
>
> Rick Edgecombe (1):
>       KVM: x86/mmu: x86: Don't overflow lpage_info when checking attribut=
es
>
> Sean Christopherson (11):
>       KVM: Add helpers to consolidate gfn_to_pfn_cache's page split check
>       KVM: Check validity of offset+length of gfn_to_pfn_cache prior to a=
ctivation
>       KVM: Explicitly disallow activatating a gfn_to_pfn_cache with INVAL=
ID_GPA
>       KVM: x86/pmu: Disable support for adaptive PEBS
>       KVM: x86/pmu: Set enable bits for GP counters in PERF_GLOBAL_CTRL a=
t "RESET"
>       KVM: selftests: Verify post-RESET value of PERF_GLOBAL_CTRL in PMCs=
 test
>       KVM: VMX: Snapshot LBR capabilities during module initialization
>       perf/x86/intel: Expose existence of callback support to KVM
>       KVM: VMX: Disable LBR virtualization if the CPU doesn't support LBR=
 callstacks
>       KVM: x86/mmu: Precisely invalidate MMU root_role during CPUID updat=
e
>       KVM: Drop unused @may_block param from gfn_to_pfn_cache_invalidate_=
start()
>
> Tao Su (1):
>       KVM: VMX: Ignore MKTME KeyID bits when intercepting #PF for allow_s=
maller_maxphyaddr
>
>  arch/x86/events/intel/lbr.c                        |  1 +
>  arch/x86/include/asm/perf_event.h                  |  1 +
>  arch/x86/kvm/mmu/mmu.c                             |  9 ++--
>  arch/x86/kvm/mmu/tdp_mmu.c                         | 51 ++++++++--------=
--
>  arch/x86/kvm/pmu.c                                 | 16 +++++-
>  arch/x86/kvm/vmx/pmu_intel.c                       |  2 +-
>  arch/x86/kvm/vmx/vmx.c                             | 41 ++++++++++++---
>  arch/x86/kvm/vmx/vmx.h                             |  6 ++-
>  .../testing/selftests/kvm/max_guest_memory_test.c  | 15 +++---
>  .../testing/selftests/kvm/set_memory_region_test.c |  2 +-
>  .../selftests/kvm/x86_64/pmu_counters_test.c       | 20 +++++++-
>  .../selftests/kvm/x86_64/vmx_dirty_log_test.c      | 60 ++++++++++++++++=
+-----
>  virt/kvm/kvm_main.c                                |  3 +-
>  virt/kvm/kvm_mm.h                                  |  6 +--
>  virt/kvm/pfncache.c                                | 50 ++++++++++++----=
--
>  15 files changed, 194 insertions(+), 89 deletions(-)
>