[GIT PULL] KVM: Fixes for 7.3-rcN

[GIT PULL] KVM: Fixes for 7.3-rcN

[Sean Christopherson]

Please pull a variety of fixes.  Nothing is super urgent, but all of these can
be fatal to the host and/or guest (except for the selftest change, which fixes
a regression that was introduced in 7.3).

The following changes since commit d599822bdb66aeec5ec76297b0fc6efaaeefe07c:

  Merge tag 'kvm-s390-master-7.3-1' of https://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD (2026-09-04 17:34:28 +0200)

are available in the Git repository at:

  https://github.com/kvm-x86/linux.git tags/kvm-x86-fixes-7.3-rc5

for you to fetch changes up to 119e9db233ad0f4cbd4cfb9f9385a7bca378b37d:

  KVM: Don't pre-reserve xarray entries when storing empty/NULL attributes (2026-09-22 08:36:16 -0700)

----------------------------------------------------------------
KVM fixes for 7.3-rcN

 - Fix a brown paper bag bug where KVM would incorrectly treat Intel PMU MSRs
   as valid on AMD.

 - Fix a regression in the hardware disable selftest where it checked the wrong
   macro when detecting glibc support (breaks at least musl).

 - Never clear KVM_REQ_VM_DEAD so that dead VMs stay dead, which is especially
   important for KVM_BUG_ON() flows, which often guard more dangerous bugs.

 - Re-pend GET_NESTED_STATE_PAGES if getting the pages fails, to fix a bug
   where KVM would let userspace run a broken setup with stale vmcs12 pages.

 - Fix a class of bugs where KVM would fail to fill kvm_run exit fields if
   getting nested pages failed.

 - Treat reserved entries in the memory attributes xarray as "no attributes",
   to fix false positives when checking for mixed attributes.

 - Fix memcg accounting for the memory attributes xarray (the xarray library
   subtly requires the xarray to be configured for accounting upfront; the gfp
   flags taken at runtime are used only rarely).

 - Don't pre-reserve xarray entries when storing empty attributes, as storing
   NULL must not require memory allocation (KVM and other subsystems heavily
   rely on this behavior).

----------------------------------------------------------------
David Ballesteros (1):
      KVM: Ensure memory attributes xarray nodes are accounted to the caller's memcg

Jim Mattson (1):
      KVM: x86/pmu: Move Intel PMU global MSRs to intel_is_valid_msr()

Sean Christopherson (4):
      KVM: selftests: Use __GLIBC__, not _GNU_SOURCE, to detect actual glibc
      KVM: Never clear KVM_REQ_VM_DEAD from a vCPU's requests
      KVM: x86: Re-pend GET_NESTED_STATE_PAGES if getting said pages fails
      KVM: x86: Fill kvm_run exit fields in common get_nested_state_pages() error paths

Zeng Chi (2):
      KVM: Don't treat reserved xarray entries as having memory attributes
      KVM: Don't pre-reserve xarray entries when storing empty/NULL attributes

 arch/arm64/kvm/arm.c                               |  2 +-
 arch/x86/kvm/mmu/mmu.c                             |  2 +-
 arch/x86/kvm/pmu.c                                 |  8 -----
 arch/x86/kvm/svm/nested.c                          |  7 +----
 arch/x86/kvm/vmx/nested.c                          | 15 ++++------
 arch/x86/kvm/vmx/pmu_intel.c                       |  3 ++
 arch/x86/kvm/vmx/tdx.c                             |  2 +-
 arch/x86/kvm/x86.c                                 |  6 +++-
 include/linux/kvm_host.h                           |  9 ++++--
 .../testing/selftests/kvm/hardware_disable_test.c  |  6 ++--
 virt/kvm/kvm_main.c                                | 35 ++++++++++++++++++----
 11 files changed, 56 insertions(+), 39 deletions(-)

Re: [GIT PULL] KVM: Fixes for 7.3-rcN

[Paolo Bonzini]

On Sat, Sep 26, 2026 at 1:33 AM Sean Christopherson <seanjc@google.com> wrote:
>
> Please pull a variety of fixes.  Nothing is super urgent, but all of these can
> be fatal to the host and/or guest (except for the selftest change, which fixes
> a regression that was introduced in 7.3).
>
> The following changes since commit d599822bdb66aeec5ec76297b0fc6efaaeefe07c:
>
>   Merge tag 'kvm-s390-master-7.3-1' of https://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD (2026-09-04 17:34:28 +0200)
>
> are available in the Git repository at:
>
>   https://github.com/kvm-x86/linux.git tags/kvm-x86-fixes-7.3-rc5
>
> for you to fetch changes up to 119e9db233ad0f4cbd4cfb9f9385a7bca378b37d:
>
>   KVM: Don't pre-reserve xarray entries when storing empty/NULL attributes (2026-09-22 08:36:16 -0700)

Pulled, thanks. I added on top the SEV cache writeback patches which
are not present here.

Paolo

> ----------------------------------------------------------------
> KVM fixes for 7.3-rcN
>
>  - Fix a brown paper bag bug where KVM would incorrectly treat Intel PMU MSRs
>    as valid on AMD.
>
>  - Fix a regression in the hardware disable selftest where it checked the wrong
>    macro when detecting glibc support (breaks at least musl).
>
>  - Never clear KVM_REQ_VM_DEAD so that dead VMs stay dead, which is especially
>    important for KVM_BUG_ON() flows, which often guard more dangerous bugs.
>
>  - Re-pend GET_NESTED_STATE_PAGES if getting the pages fails, to fix a bug
>    where KVM would let userspace run a broken setup with stale vmcs12 pages.
>
>  - Fix a class of bugs where KVM would fail to fill kvm_run exit fields if
>    getting nested pages failed.
>
>  - Treat reserved entries in the memory attributes xarray as "no attributes",
>    to fix false positives when checking for mixed attributes.
>
>  - Fix memcg accounting for the memory attributes xarray (the xarray library
>    subtly requires the xarray to be configured for accounting upfront; the gfp
>    flags taken at runtime are used only rarely).
>
>  - Don't pre-reserve xarray entries when storing empty attributes, as storing
>    NULL must not require memory allocation (KVM and other subsystems heavily
>    rely on this behavior).
>
> ----------------------------------------------------------------
> David Ballesteros (1):
>       KVM: Ensure memory attributes xarray nodes are accounted to the caller's memcg
>
> Jim Mattson (1):
>       KVM: x86/pmu: Move Intel PMU global MSRs to intel_is_valid_msr()
>
> Sean Christopherson (4):
>       KVM: selftests: Use __GLIBC__, not _GNU_SOURCE, to detect actual glibc
>       KVM: Never clear KVM_REQ_VM_DEAD from a vCPU's requests
>       KVM: x86: Re-pend GET_NESTED_STATE_PAGES if getting said pages fails
>       KVM: x86: Fill kvm_run exit fields in common get_nested_state_pages() error paths
>
> Zeng Chi (2):
>       KVM: Don't treat reserved xarray entries as having memory attributes
>       KVM: Don't pre-reserve xarray entries when storing empty/NULL attributes
>
>  arch/arm64/kvm/arm.c                               |  2 +-
>  arch/x86/kvm/mmu/mmu.c                             |  2 +-
>  arch/x86/kvm/pmu.c                                 |  8 -----
>  arch/x86/kvm/svm/nested.c                          |  7 +----
>  arch/x86/kvm/vmx/nested.c                          | 15 ++++------
>  arch/x86/kvm/vmx/pmu_intel.c                       |  3 ++
>  arch/x86/kvm/vmx/tdx.c                             |  2 +-
>  arch/x86/kvm/x86.c                                 |  6 +++-
>  include/linux/kvm_host.h                           |  9 ++++--
>  .../testing/selftests/kvm/hardware_disable_test.c  |  6 ++--
>  virt/kvm/kvm_main.c                                | 35 ++++++++++++++++++----
>  11 files changed, 56 insertions(+), 39 deletions(-)
>

Re: [GIT PULL] KVM: Fixes for 7.3-rcN

[Sean Christopherson]

On Sat, Sep 26, 2026, Paolo Bonzini wrote:
> On Sat, Sep 26, 2026 at 1:33 AM Sean Christopherson <seanjc@google.com> wrote:
> >
> > Please pull a variety of fixes.  Nothing is super urgent, but all of these can
> > be fatal to the host and/or guest (except for the selftest change, which fixes
> > a regression that was introduced in 7.3).
> >
> > The following changes since commit d599822bdb66aeec5ec76297b0fc6efaaeefe07c:
> >
> >   Merge tag 'kvm-s390-master-7.3-1' of https://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD (2026-09-04 17:34:28 +0200)
> >
> > are available in the Git repository at:
> >
> >   https://github.com/kvm-x86/linux.git tags/kvm-x86-fixes-7.3-rc5
> >
> > for you to fetch changes up to 119e9db233ad0f4cbd4cfb9f9385a7bca378b37d:
> >
> >   KVM: Don't pre-reserve xarray entries when storing empty/NULL attributes (2026-09-22 08:36:16 -0700)
> 
> Pulled, thanks. I added on top the SEV cache writeback patches which
> are not present here.

Heh, because they're buggy.  Sashiko pointed out flaws in both[1][2].  I have
fixes (will send them shortly); I was late on posting v2 because it took me forever
to figure out a not-awful solution for the double-free issue due to the annoying
CONFIG_CPUMASK_OFFSTACK behavior.

[1] https://lore.kernel.org/all/20260923165304.1662E1F000FF@smtp.kernel.org
[2] https://lore.kernel.org/all/20260923165349.CAAF01F000FF@smtp.kernel.org

Re: [GIT PULL] KVM: Fixes for 7.3-rcN

[Paolo Bonzini]

On Mon, Sep 28, 2026 at 5:41 PM Sean Christopherson <seanjc@google.com> wrote:
> > Pulled, thanks. I added on top the SEV cache writeback patches which
> > are not present here.
>
> Heh, because they're buggy.  Sashiko pointed out flaws in both[1][2].

Ouch... I have just now noticed that Sashiko only replies to kvm@ and
not reply-to-all (including LKML), so depending on which duplicate an
email search took me to, I was or was not seeing its replies. Of
course for my patches I was specifically looking for those replies and
therefore not seeing the issue. Fixed now. :(

Paolo