commit 9cfec6d097c607e36199cf0cfbb8cf5acbd8e9b2 upstream.
When running android emulator (which is based on QEMU 2.12) on
certain Intel hosts with kernel version 6.3-rc1 or above, guest
will freeze after loading a snapshot. This is almost 100%
reproducible. By default, the android emulator will use snapshot
to speed up the next launching of the same android guest. So
this breaks the android emulator badly.
I tested QEMU 8.0.4 from Debian 12 with an Ubuntu 22.04 guest by
running command "loadvm" after "savevm". The same issue is
observed. At the same time, none of our AMD platforms is impacted.
More experiments show that loading the KVM module with
"enable_apicv=false" can workaround it.
The issue started to show up after commit 8e6ed96cdd ("KVM: x86:
fire timer when it is migrated and expired, and in oneshot mode").
However, as is pointed out by Sean Christopherson, it is introduced
by commit 967235d320 ("KVM: vmx: clear pending interrupts on
KVM_SET_LAPIC"). commit 8e6ed96cdd ("KVM: x86: fire timer when
it is migrated and expired, and in oneshot mode") just makes it
easier to hit the issue.
Having both commits, the oneshot lapic timer gets fired immediately
inside the KVM_SET_LAPIC call when loading the snapshot. On Intel
platforms with APIC virtualization and posted interrupt processing,
this eventually leads to setting the corresponding PIR bit. However,
the whole PIR bits get cleared later in the same KVM_SET_LAPIC call
by apicv_post_state_restore. This leads to timer interrupt lost.
The fix is to move vmx_apicv_post_state_restore to the beginning of
the KVM_SET_LAPIC call and rename to vmx_apicv_pre_state_restore.
What vmx_apicv_post_state_restore does is actually clearing any
former apicv state and this behavior is more suitable to carry out
in the beginning.
Fixes: 967235d320 ("KVM: vmx: clear pending interrupts on KVM_SET_LAPIC")
Cc: stable@vger.kernel.org
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Haitao Shan <hshan@google.com>
Link: https://lore.kernel.org/r/20230913000215.478387-1-hshan@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 629d3698f6958ee6f8131ea324af794f973b12ac upstream.
When IPI virtualization is enabled, a WARN is triggered if bit12 of ICR
MSR is set after APIC-write VM-exit. The reason is kvm_apic_send_ipi()
thinks the APIC_ICR_BUSY bit should be cleared because KVM has no delay,
but kvm_apic_write_nodecode() doesn't clear the APIC_ICR_BUSY bit.
Under the x2APIC section, regarding ICR, the SDM says:
It remains readable only to aid in debugging; however, software should
not assume the value returned by reading the ICR is the last written
value.
I.e. the guest is allowed to set bit 12. However, the SDM also gives KVM
free reign to do whatever it wants with the bit, so long as KVM's behavior
doesn't confuse userspace or break KVM's ABI.
Clear bit 12 so that it reads back as '0'. This approach is safer than
"do nothing" and is consistent with the case where IPI virtualization is
disabled or not supported, i.e.,
handle_fastpath_set_x2apic_icr_irqoff() -> kvm_x2apic_icr_write()
Opportunistically replace the TODO with a comment calling out that eating
the write is likely faster than a conditional branch around the busy bit.
Link: https://lore.kernel.org/all/ZPj6iF0Q7iynn62p@google.com/
Fixes: 5413bcba7e ("KVM: x86: Add support for vICR APIC-write VM-Exits in x2APIC mode")
Cc: stable@vger.kernel.org
Signed-off-by: Tao Su <tao1.su@linux.intel.com>
Tested-by: Yi Lai <yi1.lai@intel.com>
Reviewed-by: Chao Gao <chao.gao@intel.com>
Link: https://lore.kernel.org/r/20230914055504.151365-1-tao1.su@linux.intel.com
[sean: tweak changelog, replace TODO with comment, drop local "val"]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2770d4722036d6bd24bcb78e9cd7f6e572077d03 upstream.
Hyper-V enabled Windows Server 2022 KVM VM cannot be started on Zen1 Ryzen
since it crashes at boot with SYSTEM_THREAD_EXCEPTION_NOT_HANDLED +
STATUS_PRIVILEGED_INSTRUCTION (in other words, because of an unexpected #GP
in the guest kernel).
This is because Windows tries to set bit 8 in MSR_AMD64_TW_CFG and can't
handle receiving a #GP when doing so.
Give this MSR the same treatment that commit 2e32b71906
("x86, kvm: Add MSR_AMD64_BU_CFG2 to the list of ignored MSRs") gave
MSR_AMD64_BU_CFG2 under justification that this MSR is baremetal-relevant
only.
Although apparently it was then needed for Linux guests, not Windows as in
this case.
With this change, the aforementioned guest setup is able to finish booting
successfully.
This issue can be reproduced either on a Summit Ridge Ryzen (with
just "-cpu host") or on a Naples EPYC (with "-cpu host,stepping=1" since
EPYC is ordinarily stepping 2).
Alternatively, userspace could solve the problem by using MSR filters, but
forcing every userspace to define a filter isn't very friendly and doesn't
add much, if any, value. The only potential hiccup is if one of these
"baremetal-only" MSRs ever requires actual emulation and/or has F/M/S
specific behavior. But if that happens, then KVM can still punt *that*
handling to userspace since userspace MSR filters "win" over KVM's default
handling.
Signed-off-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/1ce85d9c7c9e9632393816cf19c902e0a3f411f1.1697731406.git.maciej.szmigiero@oracle.com
[sean: call out MSR filtering alternative]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit d6800af51c76b6dae20e6023bbdc9b3da3ab5121 upstream.
Don't apply the stimer's counter side effects when modifying its
value from user-space, as this may trigger spurious interrupts.
For example:
- The stimer is configured in auto-enable mode.
- The stimer's count is set and the timer enabled.
- The stimer expires, an interrupt is injected.
- The VM is live migrated.
- The stimer config and count are deserialized, auto-enable is ON, the
stimer is re-enabled.
- The stimer expires right away, and injects an unwarranted interrupt.
Cc: stable@vger.kernel.org
Fixes: 1f4b34f825 ("kvm/x86: Hyper-V SynIC timers")
Signed-off-by: Nicolas Saenz Julienne <nsaenz@amazon.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Link: https://lore.kernel.org/r/20231017155101.40677-1-nsaenz@amazon.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
- Truncate writes to PMU counters to the counter's width to avoid spurious
overflows when emulating counter events in software.
- Set the LVTPC entry mask bit when handling a PMI (to match Intel-defined
architectural behavior).
- Treat KVM_REQ_PMI as a wake event instead of queueing host IRQ work to
kick the guest out of emulated halt.
-----BEGIN PGP SIGNATURE-----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=g8g3
-----END PGP SIGNATURE-----
Merge tag 'kvm-x86-pmu-6.6-fixes' of https://github.com/kvm-x86/linux into HEAD
KVM x86/pmu fixes for 6.6:
- Truncate writes to PMU counters to the counter's width to avoid spurious
overflows when emulating counter events in software.
- Set the LVTPC entry mask bit when handling a PMI (to match Intel-defined
architectural behavior).
- Treat KVM_REQ_PMI as a wake event instead of queueing host IRQ work to
kick the guest out of emulated halt.
Commit 916e3e5f26 ("KVM: SVM: Do not use user return MSR support for
virtualized TSC_AUX") introduced a local variable used for the rdmsr()
function for the high 32-bits of the MSR value. This variable is not used
after being set and triggers a warning or error, when treating warnings
as errors, when the unused-but-set-variable flag is set. Mark this
variable as __maybe_unused to fix this.
Fixes: 916e3e5f26 ("KVM: SVM: Do not use user return MSR support for virtualized TSC_AUX")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Message-Id: <0da9874b6e9fcbaaa5edeb345d7e2a7c859fc818.1696271334.git.thomas.lendacky@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
svm_leave_nested() similar to a nested VM exit, get the vCPU out of nested
mode and thus should end the local inhibition of AVIC on this vCPU.
Failure to do so, can lead to hangs on guest reboot.
Raise the KVM_REQ_APICV_UPDATE request to refresh the AVIC state of the
current vCPU in this case.
Fixes: f44509f849 ("KVM: x86: SVM: allow AVIC to co-exist with a nested guest running")
Cc: stable@vger.kernel.org
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230928173354.217464-4-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In later revisions of AMD's APM, there is a new 'incomplete IPI' exit code:
"Invalid IPI Vector - The vector for the specified IPI was set to an
illegal value (VEC < 16)"
Note that tests on Zen2 machine show that this VM exit doesn't happen and
instead AVIC just does nothing.
Add support for this exit code by doing nothing, instead of filling
the kernel log with errors.
Also replace an unthrottled 'pr_err()' if another unknown incomplete
IPI exit happens with vcpu_unimpl()
(e.g in case AMD adds yet another 'Invalid IPI' exit reason)
Cc: <stable@vger.kernel.org>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230928173354.217464-3-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The following problem exists since x2avic was enabled in the KVM:
svm_set_x2apic_msr_interception is called to enable the interception of
the x2apic msrs.
In particular it is called at the moment the guest resets its apic.
Assuming that the guest's apic was in x2apic mode, the reset will bring
it back to the xapic mode.
The svm_set_x2apic_msr_interception however has an erroneous check for
'!apic_x2apic_mode()' which prevents it from doing anything in this case.
As a result of this, all x2apic msrs are left unintercepted, and that
exposes the bare metal x2apic (if enabled) to the guest.
Oops.
Remove the erroneous '!apic_x2apic_mode()' check to fix that.
This fixes CVE-2023-5090
Fixes: 4d1d7942e3 ("KVM: SVM: Introduce logic to (de)activate x2AVIC mode")
Cc: stable@vger.kernel.org
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Tested-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230928173354.217464-2-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Mask off xfeatures that aren't exposed to the guest only when saving guest
state via KVM_GET_XSAVE{2} instead of modifying user_xfeatures directly.
Preserving the maximal set of xfeatures in user_xfeatures restores KVM's
ABI for KVM_SET_XSAVE, which prior to commit ad856280dd ("x86/kvm/fpu:
Limit guest user_xfeatures to supported bits of XCR0") allowed userspace
to load xfeatures that are supported by the host, irrespective of what
xfeatures are exposed to the guest.
There is no known use case where userspace *intentionally* loads xfeatures
that aren't exposed to the guest, but the bug fixed by commit ad856280dd
was specifically that KVM_GET_SAVE{2} would save xfeatures that weren't
exposed to the guest, e.g. would lead to userspace unintentionally loading
guest-unsupported xfeatures when live migrating a VM.
Restricting KVM_SET_XSAVE to guest-supported xfeatures is especially
problematic for QEMU-based setups, as QEMU has a bug where instead of
terminating the VM if KVM_SET_XSAVE fails, QEMU instead simply stops
loading guest state, i.e. resumes the guest after live migration with
incomplete guest state, and ultimately results in guest data corruption.
Note, letting userspace restore all host-supported xfeatures does not fix
setups where a VM is migrated from a host *without* commit ad856280dd,
to a target with a subset of host-supported xfeatures. However there is
no way to safely address that scenario, e.g. KVM could silently drop the
unsupported features, but that would be a clear violation of KVM's ABI and
so would require userspace to opt-in, at which point userspace could
simply be updated to sanitize the to-be-loaded XSAVE state.
Reported-by: Tyler Stachecki <stachecki.tyler@gmail.com>
Closes: https://lore.kernel.org/all/20230914010003.358162-1-tstachecki@bloomberg.net
Fixes: ad856280dd ("x86/kvm/fpu: Limit guest user_xfeatures to supported bits of XCR0")
Cc: stable@vger.kernel.org
Cc: Leonardo Bras <leobras@redhat.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Message-Id: <20230928001956.924301-3-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Plumb an xfeatures mask into __copy_xstate_to_uabi_buf() so that KVM can
constrain which xfeatures are saved into the userspace buffer without
having to modify the user_xfeatures field in KVM's guest_fpu state.
KVM's ABI for KVM_GET_XSAVE{2} is that features that are not exposed to
guest must not show up in the effective xstate_bv field of the buffer.
Saving only the guest-supported xfeatures allows userspace to load the
saved state on a different host with a fewer xfeatures, so long as the
target host supports the xfeatures that are exposed to the guest.
KVM currently sets user_xfeatures directly to restrict KVM_GET_XSAVE{2} to
the set of guest-supported xfeatures, but doing so broke KVM's historical
ABI for KVM_SET_XSAVE, which allows userspace to load any xfeatures that
are supported by the *host*.
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230928001956.924301-2-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When the irq_work callback, kvm_pmi_trigger_fn(), is invoked during a
VM-exit that also invokes __kvm_perf_overflow() as a result of
instruction emulation, kvm_pmu_deliver_pmi() will be called twice
before the next VM-entry.
Calling kvm_pmu_deliver_pmi() twice is unlikely to be problematic now that
KVM sets the LVTPC mask bit when delivering a PMI. But using IRQ work to
trigger the PMI is still broken, albeit very theoretically.
E.g. if the self-IPI to trigger IRQ work is be delayed long enough for the
vCPU to be migrated to a different pCPU, then it's possible for
kvm_pmi_trigger_fn() to race with the kvm_pmu_deliver_pmi() from
KVM_REQ_PMI and still generate two PMIs.
KVM could set the mask bit using an atomic operation, but that'd just be
piling on unnecessary code to workaround what is effectively a hack. The
*only* reason KVM uses IRQ work is to ensure the PMI is treated as a wake
event, e.g. if the vCPU just executed HLT.
Remove the irq_work callback for synthesizing a PMI, and all of the
logic for invoking it. Instead, to prevent a vcpu from leaving C0 with
a PMI pending, add a check for KVM_REQ_PMI to kvm_vcpu_has_events().
Fixes: 9cd803d496 ("KVM: x86: Update vPMCs when retiring instructions")
Signed-off-by: Jim Mattson <jmattson@google.com>
Tested-by: Mingwei Zhang <mizhang@google.com>
Tested-by: Dapeng Mi <dapeng1.mi@linux.intel.com>
Signed-off-by: Mingwei Zhang <mizhang@google.com>
Link: https://lore.kernel.org/r/20230925173448.3518223-2-mizhang@google.com
[sean: massage changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Per the SDM, "When the local APIC handles a performance-monitoring
counters interrupt, it automatically sets the mask flag in the LVT
performance counter register." Add this behavior to KVM's local APIC
emulation.
Failure to mask the LVTPC entry results in spurious PMIs, e.g. when
running Linux as a guest, PMI handlers that do a "late_ack" spew a large
number of "dazed and confused" spurious NMI warnings.
Fixes: f5132b0138 ("KVM: Expose a version 2 architectural PMU to a guests")
Cc: stable@vger.kernel.org
Signed-off-by: Jim Mattson <jmattson@google.com>
Tested-by: Mingwei Zhang <mizhang@google.com>
Signed-off-by: Mingwei Zhang <mizhang@google.com>
Link: https://lore.kernel.org/r/20230925173448.3518223-3-mizhang@google.com
[sean: massage changelog, correct Fixes]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Performance counters are defined to have width less than 64 bits. The
vPMU code maintains the counters in u64 variables but assumes the value
to fit within the defined width. However, for Intel non-full-width
counters (MSR_IA32_PERFCTRx) the value receieved from the guest is
truncated to 32 bits and then sign-extended to full 64 bits. If a
negative value is set, it's sign-extended to 64 bits, but then in
kvm_pmu_incr_counter() it's incremented, truncated, and compared to the
previous value for overflow detection.
That previous value is not truncated, so it always evaluates bigger than
the truncated new one, and a PMI is injected. If the PMI handler writes
a negative counter value itself, the vCPU never quits the PMI loop.
Turns out that Linux PMI handler actually does write the counter with
the value just read with RDPMC, so when no full-width support is exposed
via MSR_IA32_PERF_CAPABILITIES, and the guest initializes the counter to
a negative value, it locks up.
This has been observed in the field, for example, when the guest configures
atop to use perfevents and runs two instances of it simultaneously.
To address the problem, maintain the invariant that the counter value
always fits in the defined bit width, by truncating the received value
in the respective set_msr methods. For better readability, factor the
out into a helper function, pmc_write_counter(), shared by vmx and svm
parts.
Fixes: 9cd803d496 ("KVM: x86: Update vPMCs when retiring instructions")
Cc: stable@vger.kernel.org
Signed-off-by: Roman Kagan <rkagan@amazon.de>
Link: https://lore.kernel.org/all/20230504120042.785651-1-rkagan@amazon.de
Tested-by: Like Xu <likexu@tencent.com>
[sean: tweak changelog, s/set/write in the helper]
Signed-off-by: Sean Christopherson <seanjc@google.com>
When the TSC_AUX MSR is virtualized, the TSC_AUX value is swap type "B"
within the VMSA. This means that the guest value is loaded on VMRUN and
the host value is restored from the host save area on #VMEXIT.
Since the value is restored on #VMEXIT, the KVM user return MSR support
for TSC_AUX can be replaced by populating the host save area with the
current host value of TSC_AUX. And, since TSC_AUX is not changed by Linux
post-boot, the host save area can be set once in svm_hardware_enable().
This eliminates the two WRMSR instructions associated with the user return
MSR support.
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Message-Id: <d381de38eb0ab6c9c93dda8503b72b72546053d7.1694811272.git.thomas.lendacky@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The checks for virtualizing TSC_AUX occur during the vCPU reset processing
path. However, at the time of initial vCPU reset processing, when the vCPU
is first created, not all of the guest CPUID information has been set. In
this case the RDTSCP and RDPID feature support for the guest is not in
place and so TSC_AUX virtualization is not established.
This continues for each vCPU created for the guest. On the first boot of
an AP, vCPU reset processing is executed as a result of an APIC INIT
event, this time with all of the guest CPUID information set, resulting
in TSC_AUX virtualization being enabled, but only for the APs. The BSP
always sees a TSC_AUX value of 0 which probably went unnoticed because,
at least for Linux, the BSP TSC_AUX value is 0.
Move the TSC_AUX virtualization enablement out of the init_vmcb() path and
into the vcpu_after_set_cpuid() path to allow for proper initialization of
the support after the guest CPUID information has been set.
With the TSC_AUX virtualization support now in the vcpu_set_after_cpuid()
path, the intercepts must be either cleared or set based on the guest
CPUID input.
Fixes: 296d5a17e7 ("KVM: SEV-ES: Use V_TSC_AUX if available instead of RDTSC/MSR_TSC_AUX intercepts")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Message-Id: <4137fbcb9008951ab5f0befa74a0399d2cce809a.1694811272.git.thomas.lendacky@amd.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
svm_recalc_instruction_intercepts() is always called at least once
before the vCPU is started, so the setting or clearing of the RDTSCP
intercept can be dropped from the TSC_AUX virtualization support.
Extracted from a patch by Tom Lendacky.
Cc: stable@vger.kernel.org
Fixes: 296d5a17e7 ("KVM: SEV-ES: Use V_TSC_AUX if available instead of RDTSC/MSR_TSC_AUX intercepts")
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Stop zapping invalidate TDP MMU roots via work queue now that KVM
preserves TDP MMU roots until they are explicitly invalidated. Zapping
roots asynchronously was effectively a workaround to avoid stalling a vCPU
for an extended during if a vCPU unloaded a root, which at the time
happened whenever the guest toggled CR0.WP (a frequent operation for some
guest kernels).
While a clever hack, zapping roots via an unbound worker had subtle,
unintended consequences on host scheduling, especially when zapping
multiple roots, e.g. as part of a memslot. Because the work of zapping a
root is no longer bound to the task that initiated the zap, things like
the CPU affinity and priority of the original task get lost. Losing the
affinity and priority can be especially problematic if unbound workqueues
aren't affined to a small number of CPUs, as zapping multiple roots can
cause KVM to heavily utilize the majority of CPUs in the system, *beyond*
the CPUs KVM is already using to run vCPUs.
When deleting a memslot via KVM_SET_USER_MEMORY_REGION, the async root
zap can result in KVM occupying all logical CPUs for ~8ms, and result in
high priority tasks not being scheduled in in a timely manner. In v5.15,
which doesn't preserve unloaded roots, the issues were even more noticeable
as KVM would zap roots more frequently and could occupy all CPUs for 50ms+.
Consuming all CPUs for an extended duration can lead to significant jitter
throughout the system, e.g. on ChromeOS with virtio-gpu, deleting memslots
is a semi-frequent operation as memslots are deleted and recreated with
different host virtual addresses to react to host GPU drivers allocating
and freeing GPU blobs. On ChromeOS, the jitter manifests as audio blips
during games due to the audio server's tasks not getting scheduled in
promptly, despite the tasks having a high realtime priority.
Deleting memslots isn't exactly a fast path and should be avoided when
possible, and ChromeOS is working towards utilizing MAP_FIXED to avoid the
memslot shenanigans, but KVM is squarely in the wrong. Not to mention
that removing the async zapping eliminates a non-trivial amount of
complexity.
Note, one of the subtle behaviors hidden behind the async zapping is that
KVM would zap invalidated roots only once (ignoring partial zaps from
things like mmu_notifier events). Preserve this behavior by adding a flag
to identify roots that are scheduled to be zapped versus roots that have
already been zapped but not yet freed.
Add a comment calling out why kvm_tdp_mmu_invalidate_all_roots() can
encounter invalid roots, as it's not at all obvious why zapping
invalidated roots shouldn't simply zap all invalid roots.
Reported-by: Pattara Teerapong <pteerapong@google.com>
Cc: David Stevens <stevensd@google.com>
Cc: Yiwei Zhang<zzyiwei@google.com>
Cc: Paul Hsia <paulhsia@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230916003916.2545000-4-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
All callers except the MMU notifier want to process all address spaces.
Remove the address space ID argument of for_each_tdp_mmu_root_yield_safe()
and switch the MMU notifier to use __for_each_tdp_mmu_root_yield_safe().
Extracted out of a patch by Sean Christopherson <seanjc@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The mmu_notifier path is a bit of a special snowflake, e.g. it zaps only a
single address space (because it's per-slot), and can't always yield.
Because of this, it calls kvm_tdp_mmu_zap_leafs() in ways that no one
else does.
Iterate manually over the leafs in response to an mmu_notifier
invalidation, instead of invoking kvm_tdp_mmu_zap_leafs(). Drop the
@can_yield param from kvm_tdp_mmu_zap_leafs() as its sole remaining
caller unconditionally passes "true".
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20230916003916.2545000-2-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
* Clean up vCPU targets, always returning generic v8 as the preferred target
* Trap forwarding infrastructure for nested virtualization (used for traps
that are taken from an L2 guest and are needed by the L1 hypervisor)
* FEAT_TLBIRANGE support to only invalidate specific ranges of addresses
when collapsing a table PTE to a block PTE. This avoids that the guest
refills the TLBs again for addresses that aren't covered by the table PTE.
* Fix vPMU issues related to handling of PMUver.
* Don't unnecessary align non-stack allocations in the EL2 VA space
* Drop HCR_VIRT_EXCP_MASK, which was never used...
* Don't use smp_processor_id() in kvm_arch_vcpu_load(),
but the cpu parameter instead
* Drop redundant call to kvm_set_pfn_accessed() in user_mem_abort()
* Remove prototypes without implementations
RISC-V:
* Zba, Zbs, Zicntr, Zicsr, Zifencei, and Zihpm support for guest
* Added ONE_REG interface for SATP mode
* Added ONE_REG interface to enable/disable multiple ISA extensions
* Improved error codes returned by ONE_REG interfaces
* Added KVM_GET_REG_LIST ioctl() implementation for KVM RISC-V
* Added get-reg-list selftest for KVM RISC-V
s390:
* PV crypto passthrough enablement (Tony, Steffen, Viktor, Janosch)
Allows a PV guest to use crypto cards. Card access is governed by
the firmware and once a crypto queue is "bound" to a PV VM every
other entity (PV or not) looses access until it is not bound
anymore. Enablement is done via flags when creating the PV VM.
* Guest debug fixes (Ilya)
x86:
* Clean up KVM's handling of Intel architectural events
* Intel bugfixes
* Add support for SEV-ES DebugSwap, allowing SEV-ES guests to use debug
registers and generate/handle #DBs
* Clean up LBR virtualization code
* Fix a bug where KVM fails to set the target pCPU during an IRTE update
* Fix fatal bugs in SEV-ES intrahost migration
* Fix a bug where the recent (architecturally correct) change to reinject
#BP and skip INT3 broke SEV guests (can't decode INT3 to skip it)
* Retry APIC map recalculation if a vCPU is added/enabled
* Overhaul emergency reboot code to bring SVM up to par with VMX, tie the
"emergency disabling" behavior to KVM actually being loaded, and move all of
the logic within KVM
* Fix user triggerable WARNs in SVM where KVM incorrectly assumes the TSC
ratio MSR cannot diverge from the default when TSC scaling is disabled
up related code
* Add a framework to allow "caching" feature flags so that KVM can check if
the guest can use a feature without needing to search guest CPUID
* Rip out the ancient MMU_DEBUG crud and replace the useful bits with
CONFIG_KVM_PROVE_MMU
* Fix KVM's handling of !visible guest roots to avoid premature triple fault
injection
* Overhaul KVM's page-track APIs, and KVMGT's usage, to reduce the API surface
that is needed by external users (currently only KVMGT), and fix a variety
of issues in the process
This last item had a silly one-character bug in the topic branch that
was sent to me. Because it caused pretty bad selftest failures in
some configurations, I decided to squash in the fix. So, while the
exact commit ids haven't been in linux-next, the code has (from the
kvm-x86 tree).
Generic:
* Wrap kvm_{gfn,hva}_range.pte in a union to allow mmu_notifier events to pass
action specific data without needing to constantly update the main handlers.
* Drop unused function declarations
Selftests:
* Add testcases to x86's sync_regs_test for detecting KVM TOCTOU bugs
* Add support for printf() in guest code and covert all guest asserts to use
printf-based reporting
* Clean up the PMU event filter test and add new testcases
* Include x86 selftests in the KVM x86 MAINTAINERS entry
-----BEGIN PGP SIGNATURE-----
iQFIBAABCAAyFiEE8TM4V0tmI4mGbHaCv/vSX3jHroMFAmT1m0kUHHBib256aW5p
QHJlZGhhdC5jb20ACgkQv/vSX3jHroMNgggAiN7nz6UC423FznuI+yO3TLm8tkx1
CpKh5onqQogVtchH+vrngi97cfOzZb1/AtifY90OWQi31KEWhehkeofcx7G6ERhj
5a9NFADY1xGBsX4exca/VHDxhnzsbDWaWYPXw5vWFWI6erft9Mvy3tp1LwTvOzqM
v8X4aWz+g5bmo/DWJf4Wu32tEU6mnxzkrjKU14JmyqQTBawVmJ3RYvHVJ/Agpw+n
hRtPAy7FU6XTdkmq/uCT+KUHuJEIK0E/l1js47HFAqSzwdW70UDg14GGo1o4ETxu
RjZQmVNvL57yVgi6QU38/A0FWIsWQm5IlaX1Ug6x8pjZPnUKNbo9BY4T1g==
=W+4p
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull kvm updates from Paolo Bonzini:
"ARM:
- Clean up vCPU targets, always returning generic v8 as the preferred
target
- Trap forwarding infrastructure for nested virtualization (used for
traps that are taken from an L2 guest and are needed by the L1
hypervisor)
- FEAT_TLBIRANGE support to only invalidate specific ranges of
addresses when collapsing a table PTE to a block PTE. This avoids
that the guest refills the TLBs again for addresses that aren't
covered by the table PTE.
- Fix vPMU issues related to handling of PMUver.
- Don't unnecessary align non-stack allocations in the EL2 VA space
- Drop HCR_VIRT_EXCP_MASK, which was never used...
- Don't use smp_processor_id() in kvm_arch_vcpu_load(), but the cpu
parameter instead
- Drop redundant call to kvm_set_pfn_accessed() in user_mem_abort()
- Remove prototypes without implementations
RISC-V:
- Zba, Zbs, Zicntr, Zicsr, Zifencei, and Zihpm support for guest
- Added ONE_REG interface for SATP mode
- Added ONE_REG interface to enable/disable multiple ISA extensions
- Improved error codes returned by ONE_REG interfaces
- Added KVM_GET_REG_LIST ioctl() implementation for KVM RISC-V
- Added get-reg-list selftest for KVM RISC-V
s390:
- PV crypto passthrough enablement (Tony, Steffen, Viktor, Janosch)
Allows a PV guest to use crypto cards. Card access is governed by
the firmware and once a crypto queue is "bound" to a PV VM every
other entity (PV or not) looses access until it is not bound
anymore. Enablement is done via flags when creating the PV VM.
- Guest debug fixes (Ilya)
x86:
- Clean up KVM's handling of Intel architectural events
- Intel bugfixes
- Add support for SEV-ES DebugSwap, allowing SEV-ES guests to use
debug registers and generate/handle #DBs
- Clean up LBR virtualization code
- Fix a bug where KVM fails to set the target pCPU during an IRTE
update
- Fix fatal bugs in SEV-ES intrahost migration
- Fix a bug where the recent (architecturally correct) change to
reinject #BP and skip INT3 broke SEV guests (can't decode INT3 to
skip it)
- Retry APIC map recalculation if a vCPU is added/enabled
- Overhaul emergency reboot code to bring SVM up to par with VMX, tie
the "emergency disabling" behavior to KVM actually being loaded,
and move all of the logic within KVM
- Fix user triggerable WARNs in SVM where KVM incorrectly assumes the
TSC ratio MSR cannot diverge from the default when TSC scaling is
disabled up related code
- Add a framework to allow "caching" feature flags so that KVM can
check if the guest can use a feature without needing to search
guest CPUID
- Rip out the ancient MMU_DEBUG crud and replace the useful bits with
CONFIG_KVM_PROVE_MMU
- Fix KVM's handling of !visible guest roots to avoid premature
triple fault injection
- Overhaul KVM's page-track APIs, and KVMGT's usage, to reduce the
API surface that is needed by external users (currently only
KVMGT), and fix a variety of issues in the process
Generic:
- Wrap kvm_{gfn,hva}_range.pte in a union to allow mmu_notifier
events to pass action specific data without needing to constantly
update the main handlers.
- Drop unused function declarations
Selftests:
- Add testcases to x86's sync_regs_test for detecting KVM TOCTOU bugs
- Add support for printf() in guest code and covert all guest asserts
to use printf-based reporting
- Clean up the PMU event filter test and add new testcases
- Include x86 selftests in the KVM x86 MAINTAINERS entry"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (279 commits)
KVM: x86/mmu: Include mmu.h in spte.h
KVM: x86/mmu: Use dummy root, backed by zero page, for !visible guest roots
KVM: x86/mmu: Disallow guest from using !visible slots for page tables
KVM: x86/mmu: Harden TDP MMU iteration against root w/o shadow page
KVM: x86/mmu: Harden new PGD against roots without shadow pages
KVM: x86/mmu: Add helper to convert root hpa to shadow page
drm/i915/gvt: Drop final dependencies on KVM internal details
KVM: x86/mmu: Handle KVM bookkeeping in page-track APIs, not callers
KVM: x86/mmu: Drop @slot param from exported/external page-track APIs
KVM: x86/mmu: Bug the VM if write-tracking is used but not enabled
KVM: x86/mmu: Assert that correct locks are held for page write-tracking
KVM: x86/mmu: Rename page-track APIs to reflect the new reality
KVM: x86/mmu: Drop infrastructure for multiple page-track modes
KVM: x86/mmu: Use page-track notifiers iff there are external users
KVM: x86/mmu: Move KVM-only page-track declarations to internal header
KVM: x86: Remove the unused page-track hook track_flush_slot()
drm/i915/gvt: switch from ->track_flush_slot() to ->track_remove_region()
KVM: x86: Add a new page-track hook to handle memslot deletion
drm/i915/gvt: Don't bother removing write-protection on to-be-deleted slot
KVM: x86: Reject memslot MOVE operations if KVMGT is attached
...
Explicitly include mmu.h in spte.h instead of relying on the "parent" to
include mmu.h. spte.h references a variety of macros and variables that
are defined/declared in mmu.h, and so including spte.h before (or instead
of) mmu.h will result in build errors, e.g.
arch/x86/kvm/mmu/spte.h: In function ‘is_mmio_spte’:
arch/x86/kvm/mmu/spte.h:242:23: error: ‘enable_mmio_caching’ undeclared
242 | likely(enable_mmio_caching);
| ^~~~~~~~~~~~~~~~~~~
arch/x86/kvm/mmu/spte.h: In function ‘is_large_pte’:
arch/x86/kvm/mmu/spte.h:302:22: error: ‘PT_PAGE_SIZE_MASK’ undeclared
302 | return pte & PT_PAGE_SIZE_MASK;
| ^~~~~~~~~~~~~~~~~
arch/x86/kvm/mmu/spte.h: In function ‘is_dirty_spte’:
arch/x86/kvm/mmu/spte.h:332:56: error: ‘PT_WRITABLE_MASK’ undeclared
332 | return dirty_mask ? spte & dirty_mask : spte & PT_WRITABLE_MASK;
| ^~~~~~~~~~~~~~~~
Fixes: 5a9624affe ("KVM: mmu: extract spte.h and spte.c")
Link: https://lore.kernel.org/r/20230808224059.2492476-1-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When attempting to allocate a shadow root for a !visible guest root gfn,
e.g. that resides in MMIO space, load a dummy root that is backed by the
zero page instead of immediately synthesizing a triple fault shutdown
(using the zero page ensures any attempt to translate memory will generate
a !PRESENT fault and thus VM-Exit).
Unless the vCPU is racing with memslot activity, KVM will inject a page
fault due to not finding a visible slot in FNAME(walk_addr_generic), i.e.
the end result is mostly same, but critically KVM will inject a fault only
*after* KVM runs the vCPU with the bogus root.
Waiting to inject a fault until after running the vCPU fixes a bug where
KVM would bail from nested VM-Enter if L1 tried to run L2 with TDP enabled
and a !visible root. Even though a bad root will *probably* lead to
shutdown, (a) it's not guaranteed and (b) the CPU won't read the
underlying memory until after VM-Enter succeeds. E.g. if L1 runs L2 with
a VMX preemption timer value of '0', then architecturally the preemption
timer VM-Exit is guaranteed to occur before the CPU executes any
instruction, i.e. before the CPU needs to translate a GPA to a HPA (so
long as there are no injected events with higher priority than the
preemption timer).
If KVM manages to get to FNAME(fetch) with a dummy root, e.g. because
userspace created a memslot between installing the dummy root and handling
the page fault, simply unload the MMU to allocate a new root and retry the
instruction. Use KVM_REQ_MMU_FREE_OBSOLETE_ROOTS to drop the root, as
invoking kvm_mmu_free_roots() while holding mmu_lock would deadlock, and
conceptually the dummy root has indeeed become obsolete. The only
difference versus existing usage of KVM_REQ_MMU_FREE_OBSOLETE_ROOTS is
that the root has become obsolete due to memslot *creation*, not memslot
deletion or movement.
Reported-by: Reima Ishii <ishiir@g.ecc.u-tokyo.ac.jp>
Cc: Yu Zhang <yu.c.zhang@linux.intel.com>
Link: https://lore.kernel.org/r/20230729005200.1057358-6-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Explicitly inject a page fault if guest attempts to use a !visible gfn
as a page table. kvm_vcpu_gfn_to_hva_prot() will naturally handle the
case where there is no memslot, but doesn't catch the scenario where the
gfn points at a KVM-internal memslot.
Letting the guest backdoor its way into accessing KVM-internal memslots
isn't dangerous on its own, e.g. at worst the guest can crash itself, but
disallowing the behavior will simplify fixing how KVM handles !visible
guest root gfns (immediately synthesizing a triple fault when loading the
root is architecturally wrong).
Link: https://lore.kernel.org/r/20230729005200.1057358-5-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Explicitly check that tdp_iter_start() is handed a valid shadow page
to harden KVM against bugs, e.g. if KVM calls into the TDP MMU with an
invalid or shadow MMU root (which would be a fatal KVM bug), the shadow
page pointer will be NULL.
Opportunistically stop the TDP MMU iteration instead of continuing on
with garbage if the incoming root is bogus. Attempting to walk a garbage
root is more likely to caused major problems than doing nothing.
Cc: Yu Zhang <yu.c.zhang@linux.intel.com>
Link: https://lore.kernel.org/r/20230729005200.1057358-4-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Harden kvm_mmu_new_pgd() against NULL pointer dereference bugs by sanity
checking that the target root has an associated shadow page prior to
dereferencing said shadow page. The code in question is guaranteed to
only see roots with shadow pages as fast_pgd_switch() explicitly frees the
current root if it doesn't have a shadow page, i.e. is a PAE root, and
that in turn prevents valid roots from being cached, but that's all very
subtle.
Link: https://lore.kernel.org/r/20230729005200.1057358-3-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add a dedicated helper for converting a root hpa to a shadow page in
anticipation of using a "dummy" root to handle the scenario where KVM
needs to load a valid shadow root (from hardware's perspective), but
the guest doesn't have a visible root to shadow. Similar to PAE roots,
the dummy root won't have an associated kvm_mmu_page and will need special
handling when finding a shadow page given a root.
Opportunistically retrieve the root shadow page in kvm_mmu_sync_roots()
*after* verifying the root is unsync (the dummy root can never be unsync).
Link: https://lore.kernel.org/r/20230729005200.1057358-2-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Get/put references to KVM when a page-track notifier is (un)registered
instead of relying on the caller to do so. Forcing the caller to do the
bookkeeping is unnecessary and adds one more thing for users to get
wrong, e.g. see commit 9ed1fdee9e ("drm/i915/gvt: Get reference to KVM
iff attachment to VM is successful").
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-29-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Refactor KVM's exported/external page-track, a.k.a. write-track, APIs
to take only the gfn and do the required memslot lookup in KVM proper.
Forcing users of the APIs to get the memslot unnecessarily bleeds
KVM internals into KVMGT and complicates usage of the APIs.
No functional change intended.
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-28-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Bug the VM if something attempts to write-track a gfn, but write-tracking
isn't enabled. The VM is doomed (and KVM has an egregious bug) if KVM or
KVMGT wants to shadow guest page tables but can't because write-tracking
isn't enabled.
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-27-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When adding/removing gfns to/from write-tracking, assert that mmu_lock
is held for write, and that either slots_lock or kvm->srcu is held.
mmu_lock must be held for write to protect gfn_write_track's refcount,
and SRCU or slots_lock must be held to protect the memslot itself.
Tested-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-26-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Rename the page-track APIs to capture that they're all about tracking
writes, now that the facade of supporting multiple modes is gone.
Opportunstically replace "slot" with "gfn" in anticipation of removing
the @slot param from the external APIs.
No functional change intended.
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-25-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Drop "support" for multiple page-track modes, as there is no evidence
that array-based and refcounted metadata is the optimal solution for
other modes, nor is there any evidence that other use cases, e.g. for
access-tracking, will be a good fit for the page-track machinery in
general.
E.g. one potential use case of access-tracking would be to prevent guest
access to poisoned memory (from the guest's perspective). In that case,
the number of poisoned pages is likely to be a very small percentage of
the guest memory, and there is no need to reference count the number of
access-tracking users, i.e. expanding gfn_track[] for a new mode would be
grossly inefficient. And for poisoned memory, host userspace would also
likely want to trap accesses, e.g. to inject #MC into the guest, and that
isn't currently supported by the page-track framework.
A better alternative for that poisoned page use case is likely a
variation of the proposed per-gfn attributes overlay (linked), which
would allow efficiently tracking the sparse set of poisoned pages, and by
default would exit to userspace on access.
Link: https://lore.kernel.org/all/Y2WB48kD0J4VGynX@google.com
Cc: Ben Gardon <bgardon@google.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-24-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Disable the page-track notifier code at compile time if there are no
external users, i.e. if CONFIG_KVM_EXTERNAL_WRITE_TRACKING=n. KVM itself
now hooks emulated writes directly instead of relying on the page-track
mechanism.
Provide a stub for "struct kvm_page_track_notifier_node" so that including
headers directly from the command line, e.g. for testing include guards,
doesn't fail due to a struct having an incomplete type.
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-23-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Bury the declaration of the page-track helpers that are intended only for
internal KVM use in a "private" header. In addition to guarding against
unwanted usage of the internal-only helpers, dropping their definitions
avoids exposing other structures that should be KVM-internal, e.g. for
memslots. This is a baby step toward making kvm_host.h a KVM-internal
header in the very distant future.
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-22-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove ->track_remove_slot(), there are no longer any users and it's
unlikely a "flush" hook will ever be the correct API to provide to an
external page-track user.
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-21-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add a new page-track hook, track_remove_region(), that is called when a
memslot DELETE operation is about to be committed. The "remove" hook
will be used by KVMGT and will effectively replace the existing
track_flush_slot() altogether now that KVM itself doesn't rely on the
"flush" hook either.
The "flush" hook is flawed as it's invoked before the memslot operation
is guaranteed to succeed, i.e. KVM might ultimately keep the existing
memslot without notifying external page track users, a.k.a. KVMGT. In
practice, this can't currently happen on x86, but there are no guarantees
that won't change in the future, not to mention that "flush" does a very
poor job of describing what is happening.
Pass in the gfn+nr_pages instead of the slot itself so external users,
i.e. KVMGT, don't need to exposed to KVM internals (memslots). This will
help set the stage for additional cleanups to the page-track APIs.
Opportunistically align the existing srcu_read_lock_held() usage so that
the new case doesn't stand out like a sore thumb (and not aligning the
new code makes bots unhappy).
Cc: Zhenyu Wang <zhenyuw@linux.intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Co-developed-by: Sean Christopherson <seanjc@google.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-19-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Disallow moving memslots if the VM has external page-track users, i.e. if
KVMGT is being used to expose a virtual GPU to the guest, as KVMGT doesn't
correctly handle moving memory regions.
Note, this is potential ABI breakage! E.g. userspace could move regions
that aren't shadowed by KVMGT without harming the guest. However, the
only known user of KVMGT is QEMU, and QEMU doesn't move generic memory
regions. KVM's own support for moving memory regions was also broken for
multiple years (albeit for an edge case, but arguably moving RAM is
itself an edge case), e.g. see commit edd4fa37ba ("KVM: x86: Allocate
new rmap and large page tracking when moving memslot").
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-17-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Drop @vcpu from KVM's ->track_write() hook provided for external users of
the page-track APIs now that KVM itself doesn't use the page-track
mechanism.
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Reviewed-by: Zhi Wang <zhi.a.wang@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-16-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Don't use the generic page-track mechanism to handle writes to guest PTEs
in KVM's MMU. KVM's MMU needs access to information that should not be
exposed to external page-track users, e.g. KVM needs (for some definitions
of "need") the vCPU to query the current paging mode, whereas external
users, i.e. KVMGT, have no ties to the current vCPU and so should never
need the vCPU.
Moving away from the page-track mechanism will allow dropping use of the
page-track mechanism for KVM's own MMU, and will also allow simplifying
and cleaning up the page-track APIs.
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-15-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Call kvm_mmu_zap_all_fast() directly when flushing a memslot instead of
bouncing through the page-track mechanism. KVM (unfortunately) needs to
zap and flush all page tables on memslot DELETE/MOVE irrespective of
whether KVM is shadowing guest page tables.
This will allow changing KVM to register a page-track notifier on the
first shadow root allocation, and will also allow deleting the misguided
kvm_page_track_flush_slot() hook itself once KVM-GT also moves to a
different method for reacting to memslot changes.
No functional change intended.
Cc: Yan Zhao <yan.y.zhao@intel.com>
Link: https://lore.kernel.org/r/20221110014821.1548347-2-seanjc@google.com
Reviewed-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-14-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Move x86's implementation of kvm_arch_flush_shadow_{all,memslot}() into
mmu.c, and make kvm_mmu_zap_all() static as it was globally visible only
for kvm_arch_flush_shadow_all(). This will allow refactoring
kvm_arch_flush_shadow_memslot() to call kvm_mmu_zap_all() directly without
having to expose kvm_mmu_zap_all_fast() outside of mmu.c. Keeping
everything in mmu.c will also likely simplify supporting TDX, which
intends to do zap only relevant SPTEs on memslot updates.
No functional change intended.
Suggested-by: Yan Zhao <yan.y.zhao@intel.com>
Tested-by: Yongwei Ma <yongwei.ma@intel.com>
Link: https://lore.kernel.org/r/20230729013535.1070024-13-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Introduce KVM_BUG_ON_DATA_CORRUPTION() and use it in the low-level rmap
helpers to convert the existing BUG()s to WARN_ON_ONCE() when the kernel
is built with CONFIG_BUG_ON_DATA_CORRUPTION=n, i.e. does NOT want to BUG()
on corruption of host kernel data structures. Environments that don't
have infrastructure to automatically capture crash dumps, i.e. aren't
likely to enable CONFIG_BUG_ON_DATA_CORRUPTION=y, are typically better
served overall by WARN-and-continue behavior (for the kernel, the VM is
dead regardless), as a BUG() while holding mmu_lock all but guarantees
the _best_ case scenario is a panic().
Make the BUG()s conditional instead of removing/replacing them entirely as
there's a non-zero chance (though by no means a guarantee) that the damage
isn't contained to the target VM, e.g. if no rmap is found for a SPTE then
KVM may be double-zapping the SPTE, i.e. has already freed the memory the
SPTE pointed at and thus KVM is reading/writing memory that KVM no longer
owns.
Link: https://lore.kernel.org/all/20221129191237.31447-1-mizhang@google.com
Suggested-by: Mingwei Zhang <mizhang@google.com>
Cc: David Matlack <dmatlack@google.com>
Cc: Jim Mattson <jmattson@google.com>
Reviewed-by: Mingwei Zhang <mizhang@google.com>
Link: https://lore.kernel.org/r/20230729004722.1056172-13-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Plumb "struct kvm" all the way to pte_list_remove() to allow the usage of
KVM_BUG() and/or KVM_BUG_ON(). This will allow killing only the offending
VM instead of doing BUG() if the kernel is built with
CONFIG_BUG_ON_DATA_CORRUPTION=n, i.e. does NOT want to BUG() if KVM's data
structures (rmaps) appear to be corrupted.
Signed-off-by: Mingwei Zhang <mizhang@google.com>
[sean: tweak changelog]
Link: https://lore.kernel.org/r/20230729004722.1056172-12-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Use BUILD_BUG_ON_INVALID() instead of an empty do-while loop to stub out
KVM_MMU_WARN_ON() when CONFIG_KVM_PROVE_MMU=n, that way _some_ build
issues with the usage of KVM_MMU_WARN_ON() will be dected even if the
kernel is using the stubs, e.g. basic syntax errors will be detected.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Link: https://lore.kernel.org/r/20230729004722.1056172-11-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Replace MMU_DEBUG, which requires manually modifying KVM to enable the
macro, with a proper Kconfig, KVM_PROVE_MMU. Now that pgprintk() and
rmap_printk() are gone, i.e. the macro guards only KVM_MMU_WARN_ON() and
won't flood the kernel logs, enabling the option for debug kernels is both
desirable and feasible.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Link: https://lore.kernel.org/r/20230729004722.1056172-10-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Promote the ASSERT(), which is quite dead code in KVM, into a KVM_BUG_ON()
for KVM's sanity check that CR4.PAE=1 if the vCPU is in long mode when
performing a walk of guest page tables. The sanity is quite cheap since
neither EFER nor CR4.PAE requires a VMREAD, especially relative to the
cost of walking the guest page tables.
More importantly, the sanity check would have prevented the true badness
fixed by commit 112e66017b ("KVM: nVMX: add missing consistency checks
for CR0 and CR4"). The missed consistency check resulted in some versions
of KVM corrupting the on-stack guest_walker structure due to KVM thinking
there are 4/5 levels of page tables, but wiring up the MMU hooks to point
at the paging32 implementation, which only allocates space for two levels
of page tables in "struct guest_walker32".
Queue a page fault for injection if the assertion fails, as both callers,
FNAME(gva_to_gpa) and FNAME(walk_addr_generic), assume that walker.fault
contains sane info on a walk failure. E.g. not populating the fault info
could result in KVM consuming and/or exposing uninitialized stack data
before the vCPU is kicked out to userspace, which doesn't happen until
KVM checks for KVM_REQ_VM_DEAD on the next enter.
Move the check below the initialization of "pte_access" so that the
aforementioned to-be-injected page fault doesn't consume uninitialized
stack data. The information _shouldn't_ reach the guest or userspace,
but there's zero downside to being paranoid in this case.
Link: https://lore.kernel.org/r/20230729004722.1056172-9-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Convert all "runtime" assertions, i.e. assertions that can be triggered
while running vCPUs, from WARN_ON() to WARN_ON_ONCE(). Every WARN in the
MMU that is tied to running vCPUs, i.e. not contained to loading and
initializing KVM, is likely to fire _a lot_ when it does trigger. E.g. if
KVM ends up with a bug that causes a root to be invalidated before the
page fault handler is invoked, pretty much _every_ page fault VM-Exit
triggers the WARN.
If a WARN is triggered frequently, the resulting spam usually causes a lot
of damage of its own, e.g. consumes resources to log the WARN and pollutes
the kernel log, often to the point where other useful information can be
lost. In many case, the damage caused by the spam is actually worse than
the bug itself, e.g. KVM can almost always recover from an unexpectedly
invalid root.
On the flip side, warning every time is rarely helpful for debug and
triage, i.e. a single splat is usually sufficient to point a debugger in
the right direction, and automated testing, e.g. syzkaller, typically runs
with warn_on_panic=1, i.e. will never get past the first WARN anyways.
Lastly, when an assertions fails multiple times, the stack traces in KVM
are almost always identical, i.e. the full splat only needs to be captured
once. And _if_ there is value in captruing information about the failed
assert, a ratelimited printk() is sufficient and less likely to rack up a
large amount of collateral damage.
Link: https://lore.kernel.org/r/20230729004722.1056172-8-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Rename MMU_WARN_ON() to make it super obvious that the assertions are
all about KVM's MMU, not the primary MMU.
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Link: https://lore.kernel.org/r/20230729004722.1056172-7-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Massage the error message for the sanity check on SPTEs when freeing a
shadow page to be more verbose, and to print out all shadow-present SPTEs,
not just the first SPTE encountered. Printing all SPTEs can be quite
valuable for debug, e.g. highlights whether the leak is a one-off or
widepsread, or possibly the result of memory corruption (something else
in the kernel stomping on KVM's SPTEs).
Opportunistically move the MMU_WARN_ON() into the helper itself, which
will allow a future cleanup to use BUILD_BUG_ON_INVALID() as the stub for
MMU_WARN_ON(). BUILD_BUG_ON_INVALID() works as intended and results in
the compiler complaining about is_empty_shadow_page() not being declared.
Link: https://lore.kernel.org/r/20230729004722.1056172-6-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Haitao Shan