Skip to content

L
linux
Switch branch/tag
  1. 27 Nov, 2020 4 commits
    • Vitaly Kuznetsov's avatar
      kvm: x86/mmu: Fix get_mmio_spte() on CPUs supporting 5-level PT · 9a2a0d3c
      Vitaly Kuznetsov authored 
      Commit 95fb5b02 ("kvm: x86/mmu: Support MMIO in the TDP MMU") caused
the following WARNING on an Intel Ice Lake CPU:

 get_mmio_spte: detect reserved bits on spte, addr 0xb80a0, dump hierarchy:
 ------ spte 0xb80a0 level 5.
 ------ spte 0xfcd210107 level 4.
 ------ spte 0x1004c40107 level 3.
 ------ spte 0x1004c41107 level 2.
 ------ spte 0x1db00000000b83b6 level 1.
 WARNING: CPU: 109 PID: 10254 at arch/x86/kvm/mmu/mmu.c:3569 kvm_mmu_page_fault.cold.150+0x54/0x22f [kvm]
...
 Call Trace:
  ? kvm_io_bus_get_first_dev+0x55/0x110 [kvm]
  vcpu_enter_guest+0xaa1/0x16a0 [kvm]
  ? vmx_get_cs_db_l_bits+0x17/0x30 [kvm_intel]
  ? skip_emulated_instruction+0xaa/0x150 [kvm_intel]
  kvm_arch_vcpu_ioctl_run+0xca/0x520 [kvm]

The guest triggering this crashes. Note, this happens with the traditional
MMU and EPT enabled, not with the newly introduced TDP MMU. Turns out,
there was a subtle change in the above mentioned commit. Previously,
walk_shadow_page_get_mmio_spte() was setting 'root' to 'iterator.level'
which is returned by shadow_walk_init() and this equals to
'vcpu->arch.mmu->shadow_root_level'. Now, get_mmio_spte() sets it to
'int root = vcpu->arch.mmu->root_level'.

The difference between 'root_level' and 'shadow_root_level' on CPUs
supporting 5-level page tables is that in some case we don't want to
use 5-level, in particular when 'cpuid_maxphyaddr(vcpu) <= 48'
kvm_mmu_get_tdp_level() returns '4'. In case upper layer is not used,
the corresponding SPTE will fail '__is_rsvd_bits_set()' check.

Revert to using 'shadow_root_level'.

Fixes: 95fb5b02 ("kvm: x86/mmu: Support MMIO in the TDP MMU")
Signed-off-by: default avatarVitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20201126110206.2118959-1-vkuznets@redhat.com>
Signed-off-by: default avatarPaolo Bonzini <pbonzini@redhat.com>
      9a2a0d3c
    • Paolo Bonzini's avatar
      KVM: x86: Fix split-irqchip vs interrupt injection window request · 71cc849b
      Paolo Bonzini authored 
      kvm_cpu_accept_dm_intr and kvm_vcpu_ready_for_interrupt_injection are
a hodge-podge of conditions, hacked together to get something that
more or less works.  But what is actually needed is much simpler;
in both cases the fundamental question is, do we have a place to stash
an interrupt if userspace does KVM_INTERRUPT?

In userspace irqchip mode, that is !vcpu->arch.interrupt.injected.
Currently kvm_event_needs_reinjection(vcpu) covers it, but it is
unnecessarily restrictive.

In split irqchip mode it's a bit more complicated, we need to check
kvm_apic_accept_pic_intr(vcpu) (the IRQ window exit is basically an INTACK
cycle and thus requires ExtINTs not to be masked) as well as
!pending_userspace_extint(vcpu).  However, there is no need to
check kvm_event_needs_reinjection(vcpu), since split irqchip keeps
pending ExtINT state separate from event injection state, and checking
kvm_cpu_has_interrupt(vcpu) is wrong too since ExtINT has higher
priority than APIC interrupts.  In fact the latter fixes a bug:
when userspace requests an IRQ window vmexit, an interrupt in the
local APIC can cause kvm_cpu_has_interrupt() to be true and thus
kvm_vcpu_ready_for_interrupt_injection() to return false.  When this
happens, vcpu_run does not exit to userspace but the interrupt window
vmexits keep occurring.  The VM loops without any hope of making progress.

Once we try to fix these with something like

     return kvm_arch_interrupt_allowed(vcpu) &&
-        !kvm_cpu_has_interrupt(vcpu) &&
-        !kvm_event_needs_reinjection(vcpu) &&
-        kvm_cpu_accept_dm_intr(vcpu);
+        (!lapic_in_kernel(vcpu)
+         ? !vcpu->arch.interrupt.injected
+         : (kvm_apic_accept_pic_intr(vcpu)
+            && !pending_userspace_extint(v)));

we realize two things.  First, thanks to the previous patch the complex
conditional can reuse !kvm_cpu_has_extint(vcpu).  Second, the interrupt
window request in vcpu_enter_guest()

        bool req_int_win =
                dm_request_for_irq_injection(vcpu) &&
                kvm_cpu_accept_dm_intr(vcpu);

should be kept in sync with kvm_vcpu_ready_for_interrupt_injection():
it is unnecessary to ask the processor for an interrupt window
if we would not be able to return to userspace.  Therefore,
kvm_cpu_accept_dm_intr(vcpu) is basically !kvm_cpu_has_extint(vcpu)
ANDed with the existing check for masked ExtINT.  It all makes sense:

- we can accept an interrupt from userspace if there is a place
  to stash it (and, for irqchip split, ExtINTs are not masked).
  Interrupts from userspace _can_ be accepted even if right now
  EFLAGS.IF=0.

- in order to tell userspace we will inject its interrupt ("IRQ
  window open" i.e. kvm_vcpu_ready_for_interrupt_injection), both
  KVM and the vCPU need to be ready to accept the interrupt.

... and this is what the patch implements.
Reported-by: default avatarDavid Woodhouse <dwmw@amazon.co.uk>
Analyzed-by: default avatarDavid Woodhouse <dwmw@amazon.co.uk>
Cc: stable@vger.kernel.org
Signed-off-by: default avatarPaolo Bonzini <pbonzini@redhat.com>
Reviewed-by: default avatarNikos Tsironis <ntsironis@arrikto.com>
Reviewed-by: default avatarDavid Woodhouse <dwmw@amazon.co.uk>
Tested-by: default avatarDavid Woodhouse <dwmw@amazon.co.uk>
      71cc849b
    • Paolo Bonzini's avatar
      KVM: x86: handle !lapic_in_kernel case in kvm_cpu_*_extint · 72c3bcdc
      Paolo Bonzini authored 
      Centralize handling of interrupts from the userspace APIC
in kvm_cpu_has_extint and kvm_cpu_get_extint, since
userspace APIC interrupts are handled more or less the
same as ExtINTs are with split irqchip.  This removes
duplicated code from kvm_cpu_has_injectable_intr and
kvm_cpu_has_interrupt, and makes the code more similar
between kvm_cpu_has_{extint,interrupt} on one side
and kvm_cpu_get_{extint,interrupt} on the other.

Cc: stable@vger.kernel.org
Reviewed-by: default avatarFilippo Sironi <sironi@amazon.de>
Reviewed-by: default avatarDavid Woodhouse <dwmw@amazon.co.uk>
Tested-by: default avatarDavid Woodhouse <dwmw@amazon.co.uk>
Signed-off-by: default avatarPaolo Bonzini <pbonzini@redhat.com>
      72c3bcdc
    • Paolo Bonzini's avatar
      Merge tag 'kvmarm-fixes-5.10-4' of... · 545f6394
      Paolo Bonzini authored 
      Merge tag 'kvmarm-fixes-5.10-4' of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into kvm-master

KVM/arm64 fixes for v5.10, take #4

- Fix alignment of the new HYP sections
- Fix GICR_TYPER access from userspace
      545f6394
  3. 20 Nov, 2020 1 commit
    • Sean Christopherson's avatar
      MAINTAINERS: Update email address for Sean Christopherson · c2b1209d
      Sean Christopherson authored 
      Update my email address to one provided by my new benefactor.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Jarkko Sakkinen <jarkko@kernel.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Vitaly Kuznetsov <vkuznets@redhat.com>
Cc: Wanpeng Li <wanpengli@tencent.com>
Cc: Jim Mattson <jmattson@google.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: kvm@vger.kernel.org
Signed-off-by: default avatarSean Christopherson <seanjc@google.com>
Message-Id: <20201119183707.291864-1-sean.kvm@gmail.com>
Signed-off-by: default avatarPaolo Bonzini <pbonzini@redhat.com>
      c2b1209d
  5. 18 Nov, 2020 3 commits
  7. 17 Nov, 2020 2 commits
  9. 16 Nov, 2020 3 commits
    • Ashish Kalra's avatar
      KVM: SVM: Fix offset computation bug in __sev_dbg_decrypt(). · 854c57f0
      Ashish Kalra authored 
      Fix offset computation in __sev_dbg_decrypt() to include the
source paddr before it is rounded down to be aligned to 16 bytes
as required by SEV API. This fixes incorrect guest memory dumps
observed when using qemu monitor.
Signed-off-by: default avatarAshish Kalra <ashish.kalra@amd.com>
Message-Id: <20201110224205.29444-1-Ashish.Kalra@amd.com>
Signed-off-by: default avatarPaolo Bonzini <pbonzini@redhat.com>
      854c57f0
    • Paolo Bonzini's avatar
      Merge tag 'kvm-s390-master-5.10-1' of... · d4d3c84d
      Paolo Bonzini authored 
      Merge tag 'kvm-s390-master-5.10-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into kvm-master

KVM: s390: Fixes for 5.10

- do not reset the global diag318 data for per-cpu reset
- do not mark memory as protected too early
      d4d3c84d
    • Jamie Iles's avatar
      KVM: arm64: Correctly align nVHE percpu data · 7bab16a6
      Jamie Iles authored 
      The nVHE percpu data is partially linked but the nVHE linker script did
not align the percpu section.  The PERCPU_INPUT macro would then align
the data to a page boundary:

  #define PERCPU_INPUT(cacheline)					\
  	__per_cpu_start = .;						\
  	*(.data..percpu..first)						\
  	. = ALIGN(PAGE_SIZE);						\
  	*(.data..percpu..page_aligned)					\
  	. = ALIGN(cacheline);						\
  	*(.data..percpu..read_mostly)					\
  	. = ALIGN(cacheline);						\
  	*(.data..percpu)						\
  	*(.data..percpu..shared_aligned)				\
  	PERCPU_DECRYPTED_SECTION					\
  	__per_cpu_end = .;

but then when the final vmlinux linking happens the hypervisor percpu
data is included after page alignment and so the offsets potentially
don't match.  On my build I saw that the .hyp.data..percpu section was
at address 0x20 and then the percpu data would begin at 0x1000 (because
of the page alignment in PERCPU_INPUT), but when linked into vmlinux,
everything would be shifted down by 0x20 bytes.

This manifests as one of the CPUs getting lost when running
kvm-unit-tests or starting any VM and subsequent soft lockup on a Cortex
A72 device.

Fixes: 30c95391 ("kvm: arm64: Set up hyp percpu data for nVHE")
Signed-off-by: default avatarJamie Iles <jamie@nuviainc.com>
Signed-off-by: default avatarMarc Zyngier <maz@kernel.org>
Acked-by: default avatarDavid Brazdil <dbrazdil@google.com>
Cc: David Brazdil <dbrazdil@google.com>
Cc: Marc Zyngier <maz@kernel.org>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20201113150406.14314-1-jamie@nuviainc.com
      7bab16a6
  11. 15 Nov, 2020 1 commit
    • Paolo Bonzini's avatar
      kvm: mmu: fix is_tdp_mmu_check when the TDP MMU is not in use · c887c9b9
      Paolo Bonzini authored 
      In some cases where shadow paging is in use, the root page will
be either mmu->pae_root or vcpu->arch.mmu->lm_root.  Then it will
not have an associated struct kvm_mmu_page, because it is allocated
with alloc_page instead of kvm_mmu_alloc_page.

Just return false quickly from is_tdp_mmu_root if the TDP MMU is
not in use, which also includes the case where shadow paging is
enabled.
Signed-off-by: default avatarPaolo Bonzini <pbonzini@redhat.com>
      c887c9b9
  13. 13 Nov, 2020 6 commits
  15. 12 Nov, 2020 20 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7