VMCLEAR all in-use VMCSes during a crash, even if kdump's NMI shootdown
interrupted a KVM update of the percpu in-use VMCS list.
Because NMIs are not blocked by disabling IRQs, it's possible that
crash_vmclear_local_loaded_vmcss() could be called while the percpu list
of VMCSes is being modified, e.g. in the middle of list_add() in
vmx_vcpu_load_vmcs(). This potential corner case was called out in the
original commit[*], but the analysis of its impact was wrong.
Skipping the VMCLEARs is wrong because it all but guarantees that a
loaded, and therefore cached, VMCS will live across kexec and corrupt
memory in the new kernel. Corruption will occur because the CPU's VMCS
cache is non-coherent, i.e. not snooped, and so the writeback of VMCS
memory on its eviction will overwrite random memory in the new kernel.
The VMCS will live because the NMI shootdown also disables VMX, i.e. the
in-progress VMCLEAR will #UD, and existing Intel CPUs do not flush the
VMCS cache on VMXOFF.
Furthermore, interrupting list_add() and list_del() is safe due to
crash_vmclear_local_loaded_vmcss() using forward iteration. list_add()
ensures the new entry is not visible to forward iteration unless the
entire add completes, via WRITE_ONCE(prev->next, new). A bad "prev"
pointer could be observed if the NMI shootdown interrupted list_del() or
list_add(), but list_for_each_entry() does not consume ->prev.
In addition to removing the temporary disabling of VMCLEAR, open code
loaded_vmcs_init() in __loaded_vmcs_clear() and reorder VMCLEAR so that
the VMCS is deleted from the list only after it's been VMCLEAR'd.
Deleting the VMCS before VMCLEAR would allow a race where the NMI
shootdown could arrive between list_del() and vmcs_clear() and thus
neither flow would execute a successful VMCLEAR. Alternatively, more
code could be moved into loaded_vmcs_init(), but that gets rather silly
as the only other user, alloc_loaded_vmcs(), doesn't need the smp_wmb()
and would need to work around the list_del().
Update the smp_*() comments related to the list manipulation, and
opportunistically reword them to improve clarity.
[*] https://patchwork.kernel.org/patch/1675731/#3720461
Fixes: 8f536b7697 ("KVM: VMX: provide the vmclear function and a bitmap to support VMCLEAR in kdump")
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200321193751.24985-2-sean.j.christopherson@intel.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
For CPU supporting fast short REP MOV (XF86_FEATURE_FSRM) e.g Icelake,
Tigerlake, expose it in KVM supported cpuid as well.
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Message-Id: <20200323092236.3703-1-zhenyuw@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add an alternative format that can be more easily used for further
processing later on.
Note that we add a timestamp in the first column for both, the regular
and the new csv format.
Signed-off-by: Stefan Raspl <raspl@linux.ibm.com>
Message-Id: <20200306114250.57585-5-raspl@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This now controls both, the refresh rate of the interactive mode as well
as the logging mode. Which, as a consequence, means that the default of
logging mode is now 3s, too (use command line switch '-s' to adjust to
your liking).
Signed-off-by: Stefan Raspl <raspl@linux.ibm.com>
Message-Id: <20200306114250.57585-4-raspl@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
optparse is deprecated for a while, hence switching over to argparse
(which also works with python2).
As a consequence, help output has some subtle changes, the most
significant one being that the options are all listed explicitly
instead of a universal '[options]' indicator. Also, some of the error
messages are phrased slightly different.
While at it, squashed a number of minor PEP8 issues.
Signed-off-by: Stefan Raspl <raspl@linux.ibm.com>
Message-Id: <20200306114250.57585-3-raspl@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Make sure command line arguments are sorted alphabetically
everywhere, and adjusted existing texts for interactive command 's' to
become consistent with the long form --set-delay.
Throwing in some PEP8 fixes (all cosmetics) for good measure.
Signed-off-by: Stefan Raspl <raspl@linux.ibm.com>
Message-Id: <20200306114250.57585-2-raspl@linux.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The steal_time test's timespec stop condition was wrong and should have
used the timespec functions instead to avoid being wrong, but
timespec_diff had a strange interface. Rework all the timespec API and
its use.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In kvm_arch_dev_ioctl(), the brackets of case KVM_X86_GET_MCE_CAP_SUPPORTED
accidently encapsulates case KVM_GET_MSR_FEATURE_INDEX_LIST and case
KVM_GET_MSRS. It doesn't affect functionality but it's misleading.
Remove unnecessary brackets and opportunistically add a "break" in the
default path.
Suggested-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Xiaoyao Li <xiaoyao.li@intel.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Tack on "used max basic" at the end of the CPUID tracepoint when the
output values correspond to the max basic leaf, i.e. when emulating
Intel's out-of-range CPUID behavior. Observing "cpuid entry not found"
in the tracepoint with non-zero output values is confusing for users
that aren't familiar with the out-of-range semantics, and qualifying the
"not found" case hopefully makes it clear that "found" means "found the
exact entry".
Suggested-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Output the requested index when tracing CPUID emulation; it's basically
mandatory for leafs where the index is meaningful, and is helpful for
verifying KVM correctness even when the index isn't meaningful, e.g. the
trace for a Linux guest's hypervisor_cpuid_base() probing appears to
be broken (returns all zeroes) at first glance, but is correct because
the index is non-zero, i.e. the output values correspond to a random
index in the maximum basic leaf.
Suggested-by: Xiaoyao Li <xiaoyao.li@intel.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
EFER is set for L2 using svm_set_efer, which hardcodes EFER_SVME to 1 and hides
an incorrect value for EFER.SVME in the L1 VMCB. Perform the check manually
to detect invalid guest state.
Reported-by: Krish Sadhukhan <krish.sadhukhan@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
On Tigerlake new AVX512 VP2INTERSECT feature is available.
This allows to expose it via KVM_GET_SUPPORTED_CPUID.
Cc: "Zhong, Yang" <yang.zhong@intel.com>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The name of nested_vmx_exit_reflected suggests that it's purely
a test, but it actually marks VMCS12 pages as dirty. Move this to
vmx_handle_exit, observing that the initial nested_run_pending check in
nested_vmx_exit_reflected is pointless---nested_run_pending has just
been cleared in vmx_vcpu_run and won't be set until handle_vmlaunch
or handle_vmresume.
Suggested-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Registers in "regs" array are indexed as rax/rcx/rdx/.../rsi/rdi/r8/...
Reorder access to "regs" array in vmenter.S to follow its natural order.
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
1. Allow to disable gisa
2. protected virtual machines
Protected VMs (PVM) are KVM VMs, where KVM can't access the VM's
state like guest memory and guest registers anymore. Instead the
PVMs are mostly managed by a new entity called Ultravisor (UV),
which provides an API, so KVM and the PV can request management
actions.
PVMs are encrypted at rest and protected from hypervisor access
while running. They switch from a normal operation into protected
mode, so we can still use the standard boot process to load a
encrypted blob and then move it into protected mode.
Rebooting is only possible by passing through the unprotected/normal
mode and switching to protected again.
One mm related patch will go via Andrews mm tree ( mm/gup/writeback:
add callbacks for inaccessible pages)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)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=x6Hn
-----END PGP SIGNATURE-----
Merge tag 'kvm-s390-next-5.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD
KVM: s390: Features and Enhancements for 5.7 part1
1. Allow to disable gisa
2. protected virtual machines
Protected VMs (PVM) are KVM VMs, where KVM can't access the VM's
state like guest memory and guest registers anymore. Instead the
PVMs are mostly managed by a new entity called Ultravisor (UV),
which provides an API, so KVM and the PV can request management
actions.
PVMs are encrypted at rest and protected from hypervisor access
while running. They switch from a normal operation into protected
mode, so we can still use the standard boot process to load a
encrypted blob and then move it into protected mode.
Rebooting is only possible by passing through the unprotected/normal
mode and switching to protected again.
One mm related patch will go via Andrews mm tree ( mm/gup/writeback:
add callbacks for inaccessible pages)
Check that guest doesn't hang when an invalid eVMCS GPA is specified.
Testing that #UD is injected would probably be better but selftests lack
the infrastructure currently.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Check that VMfailInvalid happens when eVMCS revision is is invalid.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM allows to use revision_id from MSR_IA32_VMX_BASIC as eVMCS revision_id
to workaround a bug in genuine Hyper-V (see the comment in
nested_vmx_handle_enlightened_vmptrld()), this shouldn't be used by
default. Switch to using KVM_EVMCS_VERSION(1).
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
nested_vmx_handle_enlightened_vmptrld() fails in two cases:
- when we fail to kvm_vcpu_map() the supplied GPA
- when revision_id is incorrect.
Genuine Hyper-V raises #UD in the former case (at least with *some*
incorrect GPAs) and does VMfailInvalid() in the later. KVM doesn't do
anything so L1 just gets stuck retrying the same faulty VMLAUNCH.
nested_vmx_handle_enlightened_vmptrld() has two call sites:
nested_vmx_run() and nested_get_vmcs12_pages(). The former needs to queue
do much: the failure there happens after migration when L2 was running (and
L1 did something weird like wrote to VP assist page from a different vCPU),
just kill L1 with KVM_EXIT_INTERNAL_ERROR.
Reported-by: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
[Squash kbuild autopatch. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When vmx_set_nested_state() happens, we may not have all the required
data to map enlightened VMCS: e.g. HV_X64_MSR_VP_ASSIST_PAGE MSR may not
yet be restored so we need a postponed action. Currently, we (ab)use
need_vmcs12_to_shadow_sync/nested_sync_vmcs12_to_shadow() for that but
this is not ideal:
- We may not need to sync anything if L2 is running
- It is hard to propagate errors from nested_sync_vmcs12_to_shadow()
as we call it from vmx_prepare_switch_to_guest() which happens just
before we do VMLAUNCH, the code is not ready to handle errors there.
Move eVMCS mapping to nested_get_vmcs12_pages() and request
KVM_REQ_GET_VMCS12_PAGES, it seems to be is less abusive in nature.
It would probably be possible to introduce a specialized KVM_REQ_EVMCS_MAP
but it is undesirable to propagate eVMCS specifics all the way up to x86.c
Note, we don't need to request KVM_REQ_GET_VMCS12_PAGES from
vmx_set_nested_state() directly as nested_vmx_enter_non_root_mode() already
does that. Requesting KVM_REQ_GET_VMCS12_PAGES is done to document the
(non-obvious) side-effect and to be future proof.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Changed all tests and utilities to use TEST_FAIL macro
instead of TEST_ASSERT(false,...).
Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Some tests/utilities use the TEST_ASSERT(false, ...) pattern to
indicate a failure and stop execution.
This change introduces the TEST_FAIL macro which is a wrap around
TEST_ASSERT(false, ...) and so provides a direct alternative for
failing a test.
Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Normal reset and initial CPU reset do not clear all registers. Add a
test that those registers are NOT changed.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
We should not only test the oneregs or the get_(x)regs interfaces but
also the sync_regs. Those are usually the canonical place for register
content.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The guest crashes very early due to changes in the control registers
used by dynamic address translation. Let us use different registers
that will not crash the guest.
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The steal-time test confirms what is reported to the guest as stolen
time is consistent with the run_delay reported for the VCPU thread
on the host. Both x86_64 and AArch64 have the concept of steal/stolen
time so this test is introduced for both architectures.
While adding the test we ensure .gitignore has all tests listed
(it was missing s390x/resets) and that the Makefile has all tests
listed in alphabetical order (not really necessary, but it almost
was already...). We also extend the common API with a new num-guest-
pages call and a new timespec call.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Also correct the comment and prototype for vm_create_default(),
as it takes a number of pages, not a size.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Use the format attribute to enable printf format warnings, and
then fix them all.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
acrs are 32 bit and not 64 bit.
Reported-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
value is u64 and not string.
Reported-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Move function documentation comment blocks to the header files in
order to avoid duplicating them for each architecture. While at
it clean up and fix up the comment blocks.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add svm_vmcall_test to gitignore list, and realphabetize it.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
After test_and_set_bit() for kvm->arch.apicv_inhibit_reasons, we will
always get false when calling kvm_apicv_activated() because it's sure
apicv_inhibit_reasons do not equal to 0.
What the code wants to do, is check whether APICv was *already* active
and if so skip the costly request; we can do this using cmpxchg.
Reported-by: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
commit 5ef8acbdd6 ("KVM: nVMX: Emulate MTF when performing
instruction emulation") introduced a helper to check the MTF
VM-execution control in vmcs12. Change pre-existing check in
nested_vmx_exit_reflected() to instead use the helper.
Signed-off-by: Oliver Upton <oupton@google.com>
Reviewed-by: Krish Sadhukhan <krish.sadhukhan@oracle.com>
Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
PMU is not exposed to guest by most of products from cloud providers since the
bad performance of PMU emulation and security concern. However, it calls
perf_guest_switch_get_msrs() and clear_atomic_switch_msr() unconditionally
even if PMU is not exposed to the guest before each vmentry.
~2% vmexit time reduced can be observed by kvm-unit-tests/vmexit.flat on my
SKX server.
Before patch:
vmcall 1559
After patch:
vmcall 1529
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
s390 requires 1M aligned guest sizes. Embedding the rounding in
vm_adjust_num_guest_pages() allows us to remove it from a few
other places.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
GA Log tracepoint is useful when debugging AVIC performance
issue as it can be used with perf to count the number of times
IOMMU AVIC injects interrupts through the slow-path instead of
directly inject interrupts to the target vcpu.
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Clarify locking.rst to mention early that we're not enabling fast page
fault for indirect sps. The previous wording is confusing, in that it
seems the proposed solution has been already implemented but it has not.
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This patch reproduces for nSVM the change that was made for nVMX in
commit b5861e5cf2 ("KVM: nVMX: Fix loss of pending IRQ/NMI before
entering L2"). While I do not have a test that breaks without it, I
cannot see why it would not be necessary since all events are unblocked
by VMRUN's setting of GIF back to 1.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The current implementation of physical interrupt delivery to a nested guest
is quite broken. It relies on svm_interrupt_allowed returning false if
VINTR=1 so that the interrupt can be injected from enable_irq_window,
but this does not work for guests that do not intercept HLT or that rely
on clearing the host IF to block physical interrupts while L2 runs.
This patch can be split in two logical parts, but including only
one breaks tests so I am combining both changes together.
The first and easiest is simply to return true for svm_interrupt_allowed
if HF_VINTR_MASK is set and HIF is set. This way the semantics of
svm_interrupt_allowed are respected: svm_interrupt_allowed being false
does not mean "call enable_irq_window", it means "interrupts cannot
be injected now".
After doing this, however, we need another place to inject the
interrupt, and fortunately we already have one, check_nested_events,
which nested SVM does not implement but which is meant exactly for this
purpose. It is called before interrupts are injected, and it can
therefore do the L2->L1 switch while leaving inject_pending_event
none the wiser.
This patch was developed together with Cathy Avery, who wrote the
test and did a lot of the initial debugging.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
If a nested VM is started while an IRQ was pending and with
V_INTR_MASKING=1, the behavior of the guest depends on host IF. If it
is 1, the VM should exit immediately, before executing the first
instruction of the guest, because VMRUN sets GIF back to 1.
If it is 0 and the host has VGIF, however, at the time of the VMRUN
instruction L0 is running the guest with a pending interrupt window
request. This interrupt window request is completely irrelevant to
L2, since IF only controls virtual interrupts, so this patch drops
INTERCEPT_VINTR from the VMCB while running L2 under these circumstances.
To simplify the code, both steps of enabling the interrupt window
(setting the VINTR intercept and requesting a fake virtual interrupt
in svm_inject_irq) are grouped in the svm_set_vintr function, and
likewise for dismissing the interrupt window request in svm_clear_vintr.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Instead of touching the host intercepts so that the bitwise OR in
recalc_intercepts just works, mask away uninteresting intercepts
directly in recalc_intercepts.
This is cleaner and keeps the logic in one place even for intercepts
that can change even while L2 is running.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The set_cr3 callback is not setting the guest CR3, it is setting the
root of the guest page tables, either shadow or two-dimensional.
To make this clearer as well as to indicate that the MMU calls it
via kvm_mmu_load_cr3, rename it to load_mmu_pgd.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Similar to what kvm-intel.ko is doing, provide a single callback that
merges svm_set_cr3, set_tdp_cr3 and nested_svm_set_tdp_cr3.
This lets us unify the set_cr3 and set_tdp_cr3 entries in kvm_x86_ops.
I'm doing that in this same patch because splitting it adds quite a bit
of churn due to the need for forward declarations. For the same reason
the assignment to vcpu->arch.mmu->set_cr3 is moved to kvm_init_shadow_mmu
from init_kvm_softmmu and nested_svm_init_mmu_context.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Invert and rename the kvm_cpuid() param that controls out-of-range logic
to better reflect the semantics of the affected callers, i.e. callers
that bypass the out-of-range logic do so because they are looking up an
exact guest CPUID entry, e.g. to query the maxphyaddr.
Similarly, rename kvm_cpuid()'s internal "found" to "exact" to clarify
that it tracks whether or not the exact requested leaf was found, as
opposed to any usable leaf being found.
No functional change intended.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Move all of the out-of-range logic into a single helper,
get_out_of_range_cpuid_entry(), to avoid an extra lookup of CPUID.0.0
and to provide a single location for documenting the out-of-range
behavior.
No functional change intended.
Cc: Jim Mattson <jmattson@google.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Rework the masking in the out-of-range CPUID logic to handle the
Hypervisor sub-classes, as well as the Centaur class if the guest
virtual CPU vendor is Centaur.
Masking against 0x80000000 only handles basic and extended leafs, which
results in Hypervisor range checks being performed against the basic
CPUID class, and Centuar range checks being performed against the
Extended class. E.g. if CPUID.0x40000000.EAX returns 0x4000000A and
there is no entry for CPUID.0x40000006, then function 0x40000006 would
be incorrectly reported as out of bounds.
While there is no official definition of what constitutes a class, the
convention established for Hypervisor classes effectively uses bits 31:8
as the mask by virtue of checking for different bases in increments of
0x100, e.g. KVM advertises its CPUID functions starting at 0x40000100
when HyperV features are advertised at the default base of 0x40000000.
The bad range check doesn't cause functional problems for any known VMM
because out-of-range semantics only come into play if the exact entry
isn't found, and VMMs either support a very limited Hypervisor range,
e.g. the official KVM range is 0x40000000-0x40000001 (effectively no
room for undefined leafs) or explicitly defines gaps to be zero, e.g.
Qemu explicitly creates zeroed entries up to the Centaur and Hypervisor
limits (the latter comes into play when providing HyperV features).
The bad behavior can be visually confirmed by dumping CPUID output in
the guest when running Qemu with a stable TSC, as Qemu extends the limit
of range 0x40000000 to 0x40000010 to advertise VMware's cpuid_freq,
without defining zeroed entries for 0x40000002 - 0x4000000f.
Note, documentation of Centaur/VIA CPUs is hard to come by. Designating
0xc0000000 - 0xcfffffff as the Centaur class is a best guess as to the
behavior of a real Centaur/VIA CPU.
Fixes: 43561123ab ("kvm: x86: Improve emulation of CPUID leaves 0BH and 1FH")
Cc: Jim Mattson <jmattson@google.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
