Pull kvm fixes from Paolo Bonzini:
"x86 host:
- Expose KVM_CAP_ENABLE_CAP since it is supported
- Disable KVM_HC_CLOCK_PAIRING in TSC catchup mode
- Ensure async page fault token is nonzero
- Fix lockdep false negative
- Fix FPU migration regression from the AMX changes
x86 guest:
- Don't use PV TLB/IPI/yield on uniprocessor guests
PPC:
- reserve capability id (topic branch for ppc/kvm)"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: x86: nSVM: disallow userspace setting of MSR_AMD64_TSC_RATIO to non default value when tsc scaling disabled
KVM: x86/mmu: make apf token non-zero to fix bug
KVM: PPC: reserve capability 210 for KVM_CAP_PPC_AIL_MODE_3
x86/kvm: Don't use pv tlb/ipi/sched_yield if on 1 vCPU
x86/kvm: Fix compilation warning in non-x86_64 builds
x86/kvm/fpu: Remove kvm_vcpu_arch.guest_supported_xcr0
x86/kvm/fpu: Limit guest user_xfeatures to supported bits of XCR0
kvm: x86: Disable KVM_HC_CLOCK_PAIRING if tsc is in always catchup mode
KVM: Fix lockdep false negative during host resume
KVM: x86: Add KVM_CAP_ENABLE_CAP to x86
If nested tsc scaling is disabled, MSR_AMD64_TSC_RATIO should
never have non default value.
Due to way nested tsc scaling support was implmented in qemu,
it would set this msr to 0 when nested tsc scaling was disabled.
Ignore that value for now, as it causes no harm.
Fixes: 5228eb96a4 ("KVM: x86: nSVM: implement nested TSC scaling")
Cc: stable@vger.kernel.org
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20220223115649.319134-1-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In current async pagefault logic, when a page is ready, KVM relies on
kvm_arch_can_dequeue_async_page_present() to determine whether to deliver
a READY event to the Guest. This function test token value of struct
kvm_vcpu_pv_apf_data, which must be reset to zero by Guest kernel when a
READY event is finished by Guest. If value is zero meaning that a READY
event is done, so the KVM can deliver another.
But the kvm_arch_setup_async_pf() may produce a valid token with zero
value, which is confused with previous mention and may lead the loss of
this READY event.
This bug may cause task blocked forever in Guest:
INFO: task stress:7532 blocked for more than 1254 seconds.
Not tainted 5.10.0 #16
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
task:stress state:D stack: 0 pid: 7532 ppid: 1409
flags:0x00000080
Call Trace:
__schedule+0x1e7/0x650
schedule+0x46/0xb0
kvm_async_pf_task_wait_schedule+0xad/0xe0
? exit_to_user_mode_prepare+0x60/0x70
__kvm_handle_async_pf+0x4f/0xb0
? asm_exc_page_fault+0x8/0x30
exc_page_fault+0x6f/0x110
? asm_exc_page_fault+0x8/0x30
asm_exc_page_fault+0x1e/0x30
RIP: 0033:0x402d00
RSP: 002b:00007ffd31912500 EFLAGS: 00010206
RAX: 0000000000071000 RBX: ffffffffffffffff RCX: 00000000021a32b0
RDX: 000000000007d011 RSI: 000000000007d000 RDI: 00000000021262b0
RBP: 00000000021262b0 R08: 0000000000000003 R09: 0000000000000086
R10: 00000000000000eb R11: 00007fefbdf2baa0 R12: 0000000000000000
R13: 0000000000000002 R14: 000000000007d000 R15: 0000000000001000
Signed-off-by: Liang Zhang <zhangliang5@huawei.com>
Message-Id: <20220222031239.1076682-1-zhangliang5@huawei.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pull x86 fixes from Borislav Petkov:
- Fix the ptrace regset xfpregs_set() callback to behave according to
the ABI
- Handle poisoned pages properly in the SGX reclaimer code
* tag 'x86_urgent_for_v5.17_rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/ptrace: Fix xfpregs_set()'s incorrect xmm clearing
x86/sgx: Fix missing poison handling in reclaimer
xfpregs_set() handles 32-bit REGSET_XFP and 64-bit REGSET_FP. The actual
code treats these regsets as modern FX state (i.e. the beginning part of
XSTATE). The declarations of the regsets thought they were the legacy
i387 format. The code thought they were the 32-bit (no xmm8..15) variant
of XSTATE and, for good measure, made the high bits disappear by zeroing
the wrong part of the buffer. The latter broke ptrace, and everything
else confused anyone trying to understand the code. In particular, the
nonsense definitions of the regsets confused me when I wrote this code.
Clean this all up. Change the declarations to match reality (which
shouldn't change the generated code, let alone the ABI) and fix
xfpregs_set() to clear the correct bits and to only do so for 32-bit
callers.
Fixes: 6164331d15 ("x86/fpu: Rewrite xfpregs_set()")
Reported-by: Luís Ferreira <contact@lsferreira.net>
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: <stable@vger.kernel.org>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=215524
Link: https://lore.kernel.org/r/YgpFnZpF01WwR8wU@zn.tnic
On non-x86_64 builds, helpers gtod_is_based_on_tsc() and
kvm_guest_supported_xfd() are defined but never used. Because these are
static inline but are in a .c file, some compilers do warn for them with
-Wunused-function, which becomes an error if -Werror is present.
Add #ifdef so they are only defined in x86_64 builds.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Leonardo Bras <leobras@redhat.com>
Message-Id: <20220218034100.115702-1-leobras@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The SGX reclaimer code lacks page poison handling in its main
free path. This can lead to avoidable machine checks if a
poisoned page is freed and reallocated instead of being
isolated.
A troublesome scenario is:
1. Machine check (#MC) occurs (asynchronous, !MF_ACTION_REQUIRED)
2. arch_memory_failure() is eventually called
3. (SGX) page->poison set to 1
4. Page is reclaimed
5. Page added to normal free lists by sgx_reclaim_pages()
^ This is the bug (poison pages should be isolated on the
sgx_poison_page_list instead)
6. Page is reallocated by some innocent enclave, a second (synchronous)
in-kernel #MC is induced, probably during EADD instruction.
^ This is the fallout from the bug
(6) is unfortunate and can be avoided by replacing the open coded
enclave page freeing code in the reclaimer with sgx_free_epc_page()
to obtain support for poison page handling that includes placing the
poisoned page on the correct list.
Fixes: d6d261bded ("x86/sgx: Add new sgx_epc_page flag bit to mark free pages")
Fixes: 992801ae92 ("x86/sgx: Initial poison handling for dirty and free pages")
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Link: https://lkml.kernel.org/r/dcc95eb2aaefb042527ac50d0a50738c7c160dac.1643830353.git.reinette.chatre@intel.com
kvm_vcpu_arch currently contains the guest supported features in both
guest_supported_xcr0 and guest_fpu.fpstate->user_xfeatures field.
Currently both fields are set to the same value in
kvm_vcpu_after_set_cpuid() and are not changed anywhere else after that.
Since it's not good to keep duplicated data, remove guest_supported_xcr0.
To keep the code more readable, introduce kvm_guest_supported_xcr()
and kvm_guest_supported_xfd() to replace the previous usages of
guest_supported_xcr0.
Signed-off-by: Leonardo Bras <leobras@redhat.com>
Message-Id: <20220217053028.96432-3-leobras@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
During host/guest switch (like in kvm_arch_vcpu_ioctl_run()), the kernel
swaps the fpu between host/guest contexts, by using fpu_swap_kvm_fpstate().
When xsave feature is available, the fpu swap is done by:
- xsave(s) instruction, with guest's fpstate->xfeatures as mask, is used
to store the current state of the fpu registers to a buffer.
- xrstor(s) instruction, with (fpu_kernel_cfg.max_features &
XFEATURE_MASK_FPSTATE) as mask, is used to put the buffer into fpu regs.
For xsave(s) the mask is used to limit what parts of the fpu regs will
be copied to the buffer. Likewise on xrstor(s), the mask is used to
limit what parts of the fpu regs will be changed.
The mask for xsave(s), the guest's fpstate->xfeatures, is defined on
kvm_arch_vcpu_create(), which (in summary) sets it to all features
supported by the cpu which are enabled on kernel config.
This means that xsave(s) will save to guest buffer all the fpu regs
contents the cpu has enabled when the guest is paused, even if they
are not used.
This would not be an issue, if xrstor(s) would also do that.
xrstor(s)'s mask for host/guest swap is basically every valid feature
contained in kernel config, except XFEATURE_MASK_PKRU.
Accordingto kernel src, it is instead switched in switch_to() and
flush_thread().
Then, the following happens with a host supporting PKRU starts a
guest that does not support it:
1 - Host has XFEATURE_MASK_PKRU set. 1st switch to guest,
2 - xsave(s) fpu regs to host fpustate (buffer has XFEATURE_MASK_PKRU)
3 - xrstor(s) guest fpustate to fpu regs (fpu regs have XFEATURE_MASK_PKRU)
4 - guest runs, then switch back to host,
5 - xsave(s) fpu regs to guest fpstate (buffer now have XFEATURE_MASK_PKRU)
6 - xrstor(s) host fpstate to fpu regs.
7 - kvm_vcpu_ioctl_x86_get_xsave() copy guest fpstate to userspace (with
XFEATURE_MASK_PKRU, which should not be supported by guest vcpu)
On 5, even though the guest does not support PKRU, it does have the flag
set on guest fpstate, which is transferred to userspace via vcpu ioctl
KVM_GET_XSAVE.
This becomes a problem when the user decides on migrating the above guest
to another machine that does not support PKRU: the new host restores
guest's fpu regs to as they were before (xrstor(s)), but since the new
host don't support PKRU, a general-protection exception ocurs in xrstor(s)
and that crashes the guest.
This can be solved by making the guest's fpstate->user_xfeatures hold
a copy of guest_supported_xcr0. This way, on 7 the only flags copied to
userspace will be the ones compatible to guest requirements, and thus
there will be no issue during migration.
As a bonus, it will also fail if userspace tries to set fpu features
(with the KVM_SET_XSAVE ioctl) that are not compatible to the guest
configuration. Such features will never be returned by KVM_GET_XSAVE
or KVM_GET_XSAVE2.
Also, since kvm_vcpu_after_set_cpuid() now sets fpstate->user_xfeatures,
there is not need to set it in kvm_check_cpuid(). So, change
fpstate_realloc() so it does not touch fpstate->user_xfeatures if a
non-NULL guest_fpu is passed, which is the case when kvm_check_cpuid()
calls it.
Signed-off-by: Leonardo Bras <leobras@redhat.com>
Message-Id: <20220217053028.96432-2-leobras@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
If vcpu has tsc_always_catchup set each request updates pvclock data.
KVM_HC_CLOCK_PAIRING consumers such as ptp_kvm_x86 rely on tsc read on
host's side and do hypercall inside pvclock_read_retry loop leading to
infinite loop in such situation.
v3:
Removed warn
Changed return code to KVM_EFAULT
v2:
Added warn
Signed-off-by: Anton Romanov <romanton@google.com>
Message-Id: <20220216182653.506850-1-romanton@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Follow the precedent set by other architectures that support the VCPU
ioctl, KVM_ENABLE_CAP, and advertise the VM extension, KVM_CAP_ENABLE_CAP.
This way, userspace can ensure that KVM_ENABLE_CAP is available on a
vcpu before using it.
Fixes: 5c919412fe ("kvm/x86: Hyper-V synthetic interrupt controller")
Signed-off-by: Aaron Lewis <aaronlewis@google.com>
Message-Id: <20220214212950.1776943-1-aaronlewis@google.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pull kvm fixes from Paolo Bonzini:
"ARM:
- Read HW interrupt pending state from the HW
x86:
- Don't truncate the performance event mask on AMD
- Fix Xen runstate updates to be atomic when preempting vCPU
- Fix for AMD AVIC interrupt injection race
- Several other AMD fixes"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: x86/pmu: Use AMD64_RAW_EVENT_MASK for PERF_TYPE_RAW
KVM: x86/pmu: Don't truncate the PerfEvtSeln MSR when creating a perf event
KVM: SVM: fix race between interrupt delivery and AVIC inhibition
KVM: SVM: set IRR in svm_deliver_interrupt
KVM: SVM: extract avic_ring_doorbell
selftests: kvm: Remove absent target file
KVM: arm64: vgic: Read HW interrupt pending state from the HW
KVM: x86/xen: Fix runstate updates to be atomic when preempting vCPU
KVM: x86: SVM: move avic definitions from AMD's spec to svm.h
KVM: x86: lapic: don't touch irr_pending in kvm_apic_update_apicv when inhibiting it
KVM: x86: nSVM: deal with L1 hypervisor that intercepts interrupts but lets L2 control them
KVM: x86: nSVM: expose clean bit support to the guest
KVM: x86: nSVM/nVMX: set nested_run_pending on VM entry which is a result of RSM
KVM: x86: nSVM: mark vmcb01 as dirty when restoring SMM saved state
KVM: x86: nSVM: fix potential NULL derefernce on nested migration
KVM: x86: SVM: don't passthrough SMAP/SMEP/PKE bits in !NPT && !gCR0.PG case
Revert "svm: Add warning message for AVIC IPI invalid target"
AMD's event select is 3 nybbles, with the high nybble in bits 35:32 of
a PerfEvtSeln MSR. Don't drop the high nybble when setting up the
config field of a perf_event_attr structure for a call to
perf_event_create_kernel_counter().
Fixes: ca724305a2 ("KVM: x86/vPMU: Implement AMD vPMU code for KVM")
Reported-by: Stephane Eranian <eranian@google.com>
Signed-off-by: Jim Mattson <jmattson@google.com>
Message-Id: <20220203014813.2130559-1-jmattson@google.com>
Reviewed-by: David Dunn <daviddunn@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pull objtool fix from Borislav Petkov:
"Fix a case where objtool would mistakenly warn about instructions
being unreachable"
* tag 'objtool_urgent_for_v5.17_rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/bug: Merge annotate_reachable() into _BUG_FLAGS() asm
Pull x86 fix from Borislav Petkov:
"Prevent softlockups when tearing down large SGX enclaves"
* tag 'x86_urgent_for_v5.17_rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/sgx: Silence softlockup detection when releasing large enclaves
Pull xen fixes from Juergen Gross:
- Two small cleanups
- Another fix for addressing the EFI framebuffer above 4GB when running
as Xen dom0
- A patch to let Xen guests use reserved bits in MSI- and IO-APIC-
registers for extended APIC-IDs the same way KVM guests are doing it
already
* tag 'for-linus-5.17a-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen/pci: Make use of the helper macro LIST_HEAD()
xen/x2apic: Fix inconsistent indenting
xen/x86: detect support for extended destination ID
xen/x86: obtain full video frame buffer address for Dom0 also under EFI
If svm_deliver_avic_intr is called just after the target vcpu's AVIC got
inhibited, it might read a stale value of vcpu->arch.apicv_active
which can lead to the target vCPU not noticing the interrupt.
To fix this use load-acquire/store-release so that, if the target vCPU
is IN_GUEST_MODE, we're guaranteed to see a previous disabling of the
AVIC. If AVIC has been disabled in the meanwhile, proceed with the
KVM_REQ_EVENT-based delivery.
Incomplete IPI vmexit has the same races as svm_deliver_avic_intr, and
in fact it can be handled in exactly the same way; the only difference
lies in who has set IRR, whether svm_deliver_interrupt or the processor.
Therefore, svm_complete_interrupt_delivery can be used to fix incomplete
IPI vmexits as well.
Co-developed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
SVM has to set IRR for both the AVIC and the software-LAPIC case,
so pull it up to the common function that handles both configurations.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The check on the current CPU adds an extra level of indentation to
svm_deliver_avic_intr and conflates documentation on what happens
if the vCPU exits (of interest to svm_deliver_avic_intr) and migrates
(only of interest to avic_ring_doorbell, which calls get/put_cpu()).
Extract the wrmsr to a separate function and rewrite the
comment in svm_deliver_avic_intr().
Co-developed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Vijay reported that the "unclobbered_vdso_oversubscribed" selftest
triggers the softlockup detector.
Actual SGX systems have 128GB of enclave memory or more. The
"unclobbered_vdso_oversubscribed" selftest creates one enclave which
consumes all of the enclave memory on the system. Tearing down such a
large enclave takes around a minute, most of it in the loop where
the EREMOVE instruction is applied to each individual 4k enclave page.
Spending one minute in a loop triggers the softlockup detector.
Add a cond_resched() to give other tasks a chance to run and placate
the softlockup detector.
Cc: stable@vger.kernel.org
Fixes: 1728ab54b4 ("x86/sgx: Add a page reclaimer")
Reported-by: Vijay Dhanraj <vijay.dhanraj@intel.com>
Signed-off-by: Reinette Chatre <reinette.chatre@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Tested-by: Jarkko Sakkinen <jarkko@kernel.org> (kselftest as sanity check)
Link: https://lkml.kernel.org/r/ced01cac1e75f900251b0a4ae1150aa8ebd295ec.1644345232.git.reinette.chatre@intel.com
There are circumstances whem kvm_xen_update_runstate_guest() should not
sleep because it ends up being called from __schedule() when the vCPU
is preempted:
[ 222.830825] kvm_xen_update_runstate_guest+0x24/0x100
[ 222.830878] kvm_arch_vcpu_put+0x14c/0x200
[ 222.830920] kvm_sched_out+0x30/0x40
[ 222.830960] __schedule+0x55c/0x9f0
To handle this, make it use the same trick as __kvm_xen_has_interrupt(),
of using the hva from the gfn_to_hva_cache directly. Then it can use
pagefault_disable() around the accesses and just bail out if the page
is absent (which is unlikely).
I almost switched to using a gfn_to_pfn_cache here and bailing out if
kvm_map_gfn() fails, like kvm_steal_time_set_preempted() does — but on
closer inspection it looks like kvm_map_gfn() will *always* fail in
atomic context for a page in IOMEM, which means it will silently fail
to make the update every single time for such guests, AFAICT. So I
didn't do it that way after all. And will probably fix that one too.
Cc: stable@vger.kernel.org
Fixes: 30b5c851af ("KVM: x86/xen: Add support for vCPU runstate information")
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Message-Id: <b17a93e5ff4561e57b1238e3e7ccd0b613eb827e.camel@infradead.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Xen allows the usage of some previously reserved bits in the IO-APIC
RTE and the MSI address fields in order to store high bits for the
target APIC ID. Such feature is already implemented by QEMU/KVM and
HyperV, so in order to enable it just add the handler that checks for
it's presence.
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/20220120152527.7524-3-roger.pau@citrix.com
Signed-off-by: Juergen Gross <jgross@suse.com>
Commit 7f7b4236f2 ("x86/PCI: Ignore E820 reservations for bridge windows
on newer systems") fixes the touchpad not working on laptops like
the Lenovo IdeaPad 3 15IIL05 and the Lenovo IdeaPad 5 14IIL05, as well as
fixing thunderbolt hotplug issues on the Lenovo Yoga C940.
Unfortunately it turns out that this is causing issues with suspend/resume
on Lenovo ThinkPad X1 Carbon Gen 2 laptops. So, per the no regressions
policy, rever this. Note I'm looking into another fix for the issues this
fixed.
Fixes: 7f7b4236f2 ("x86/PCI: Ignore E820 reservations for bridge windows on newer systems")
BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=2029207
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
asm/svm.h is the correct place for all values that are defined in
the SVM spec, and that includes AVIC.
Also add some values from the spec that were not defined before
and will be soon useful.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20220207155447.840194-10-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm_apic_update_apicv is called when AVIC is still active, thus IRR bits
can be set by the CPU after it is called, and don't cause the irr_pending
to be set to true.
Also logic in avic_kick_target_vcpu doesn't expect a race with this
function so to make it simple, just keep irr_pending set to true and
let the next interrupt injection to the guest clear it.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20220207155447.840194-9-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Fix a corner case in which the L1 hypervisor intercepts
interrupts (INTERCEPT_INTR) and either doesn't set
virtual interrupt masking (V_INTR_MASKING) or enters a
nested guest with EFLAGS.IF disabled prior to the entry.
In this case, despite the fact that L1 intercepts the interrupts,
KVM still needs to set up an interrupt window to wait before
injecting the INTR vmexit.
Currently the KVM instead enters an endless loop of 'req_immediate_exit'.
Exactly the same issue also happens for SMIs and NMI.
Fix this as well.
Note that on VMX, this case is impossible as there is only
'vmexit on external interrupts' execution control which either set,
in which case both host and guest's EFLAGS.IF
are ignored, or not set, in which case no VMexits are delivered.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20220207155447.840194-8-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM already honours few clean bits thus it makes sense
to let the nested guest know about it.
Note that KVM also doesn't check if the hardware supports
clean bits, and therefore nested KVM was
already setting clean bits and L0 KVM
was already honouring them.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20220207155447.840194-6-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
While RSM induced VM entries are not full VM entries,
they still need to be followed by actual VM entry to complete it,
unlike setting the nested state.
This patch fixes boot of hyperv and SMM enabled
windows VM running nested on KVM, which fail due
to this issue combined with lack of dirty bit setting.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Cc: stable@vger.kernel.org
Message-Id: <20220207155447.840194-5-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
While usually, restoring the smm state makes the KVM enter
the nested guest thus a different vmcb (vmcb02 vs vmcb01),
KVM should still mark it as dirty, since hardware
can in theory cache multiple vmcbs.
Failure to do so, combined with lack of setting the
nested_run_pending (which is fixed in the next patch),
might make KVM re-enter vmcb01, which was just exited from,
with completely different set of guest state registers
(SMM vs non SMM) and without proper dirty bits set,
which results in the CPU reusing stale IDTR pointer
which leads to a guest shutdown on any interrupt.
On the real hardware this usually doesn't happen,
but when running nested, L0's KVM does check and
honour few dirty bits, causing this issue to happen.
This patch fixes boot of hyperv and SMM enabled
windows VM running nested on KVM.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Cc: stable@vger.kernel.org
Message-Id: <20220207155447.840194-4-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Turns out that due to review feedback and/or rebases
I accidentally moved the call to nested_svm_load_cr3 to be too early,
before the NPT is enabled, which is very wrong to do.
KVM can't even access guest memory at that point as nested NPT
is needed for that, and of course it won't initialize the walk_mmu,
which is main issue the patch was addressing.
Fix this for real.
Fixes: 232f75d3b4 ("KVM: nSVM: call nested_svm_load_cr3 on nested state load")
Cc: stable@vger.kernel.org
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-Id: <20220207155447.840194-3-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When the guest doesn't enable paging, and NPT/EPT is disabled, we
use guest't paging CR3's as KVM's shadow paging pointer and
we are technically in direct mode as if we were to use NPT/EPT.
In direct mode we create SPTEs with user mode permissions
because usually in the direct mode the NPT/EPT doesn't
need to restrict access based on guest CPL
(there are MBE/GMET extenstions for that but KVM doesn't use them).
In this special "use guest paging as direct" mode however,
and if CR4.SMAP/CR4.SMEP are enabled, that will make the CPU
fault on each access and KVM will enter endless loop of page faults.
Since page protection doesn't have any meaning in !PG case,
just don't passthrough these bits.
The fix is the same as was done for VMX in commit:
commit 656ec4a492 ("KVM: VMX: fix SMEP and SMAP without EPT")
This fixes the boot of windows 10 without NPT for good.
(Without this patch, BSP boots, but APs were stuck in endless
loop of page faults, causing the VM boot with 1 CPU)
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Cc: stable@vger.kernel.org
Message-Id: <20220207155447.840194-2-mlevitsk@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove a WARN on an "AVIC IPI invalid target" exit, the WARN is trivial
to trigger from guest as it will fail on any destination APIC ID that
doesn't exist from the guest's perspective.
Don't bother recording anything in the kernel log, the common tracepoint
for kvm_avic_incomplete_ipi() is sufficient for debugging.
This reverts commit 37ef0c4414.
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220204214205.3306634-2-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Pull perf fixes from Borislav Petkov:
- Intel/PT: filters could crash the kernel
- Intel: default disable the PMU for SMM, some new-ish EFI firmware has
started using CPL3 and the PMU CPL filters don't discriminate against
SMM, meaning that CPL3 (userspace only) events now also count EFI/SMM
cycles.
- Fixup for perf_event_attr::sig_data
* tag 'perf_urgent_for_v5.17_rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel/pt: Fix crash with stop filters in single-range mode
perf: uapi: Document perf_event_attr::sig_data truncation on 32 bit architectures
selftests/perf_events: Test modification of perf_event_attr::sig_data
perf: Copy perf_event_attr::sig_data on modification
x86/perf: Default set FREEZE_ON_SMI for all
Pull xen fixes from Juergen Gross:
- documentation fixes related to Xen
- enable x2apic mode when available when running as hardware
virtualized guest under Xen
- cleanup and fix a corner case of vcpu enumeration when running a
paravirtualized Xen guest
* tag 'for-linus-5.17a-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
x86/Xen: streamline (and fix) PV CPU enumeration
xen: update missing ioctl magic numers documentation
Improve docs for IOCTL_GNTDEV_MAP_GRANT_REF
xen: xenbus_dev.h: delete incorrect file name
xen/x2apic: enable x2apic mode when supported for HVM
Pull kvm fixes from Paolo Bonzini:
"ARM:
- A couple of fixes when handling an exception while a SError has
been delivered
- Workaround for Cortex-A510's single-step erratum
RISC-V:
- Make CY, TM, and IR counters accessible in VU mode
- Fix SBI implementation version
x86:
- Report deprecation of x87 features in supported CPUID
- Preparation for fixing an interrupt delivery race on AMD hardware
- Sparse fix
All except POWER and s390:
- Rework guest entry code to correctly mark noinstr areas and fix
vtime' accounting (for x86, this was already mostly correct but not
entirely; for ARM, MIPS and RISC-V it wasn't)"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
KVM: x86: Use ERR_PTR_USR() to return -EFAULT as a __user pointer
KVM: x86: Report deprecated x87 features in supported CPUID
KVM: arm64: Workaround Cortex-A510's single-step and PAC trap errata
KVM: arm64: Stop handle_exit() from handling HVC twice when an SError occurs
KVM: arm64: Avoid consuming a stale esr value when SError occur
RISC-V: KVM: Fix SBI implementation version
RISC-V: KVM: make CY, TM, and IR counters accessible in VU mode
kvm/riscv: rework guest entry logic
kvm/arm64: rework guest entry logic
kvm/x86: rework guest entry logic
kvm/mips: rework guest entry logic
kvm: add guest_state_{enter,exit}_irqoff()
KVM: x86: Move delivery of non-APICv interrupt into vendor code
kvm: Move KVM_GET_XSAVE2 IOCTL definition at the end of kvm.h
KVM/arm64 fixes for 5.17, take #2
- A couple of fixes when handling an exception while a SError has been
delivered
- Workaround for Cortex-A510's single-step[ erratum
blake2s_compress_generic is weakly aliased by blake2s_compress. The
current harness for function selection uses a function pointer, which is
ordinarily inlined and resolved at compile time. But when Clang's CFI is
enabled, CFI still triggers when making an indirect call via a weak
symbol. This seems like a bug in Clang's CFI, as though it's bucketing
weak symbols and strong symbols differently. It also only seems to
trigger when "full LTO" mode is used, rather than "thin LTO".
[ 0.000000][ T0] Kernel panic - not syncing: CFI failure (target: blake2s_compress_generic+0x0/0x1444)
[ 0.000000][ T0] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.16.0-mainline-06981-g076c855b846e #1
[ 0.000000][ T0] Hardware name: MT6873 (DT)
[ 0.000000][ T0] Call trace:
[ 0.000000][ T0] dump_backtrace+0xfc/0x1dc
[ 0.000000][ T0] dump_stack_lvl+0xa8/0x11c
[ 0.000000][ T0] panic+0x194/0x464
[ 0.000000][ T0] __cfi_check_fail+0x54/0x58
[ 0.000000][ T0] __cfi_slowpath_diag+0x354/0x4b0
[ 0.000000][ T0] blake2s_update+0x14c/0x178
[ 0.000000][ T0] _extract_entropy+0xf4/0x29c
[ 0.000000][ T0] crng_initialize_primary+0x24/0x94
[ 0.000000][ T0] rand_initialize+0x2c/0x6c
[ 0.000000][ T0] start_kernel+0x2f8/0x65c
[ 0.000000][ T0] __primary_switched+0xc4/0x7be4
[ 0.000000][ T0] Rebooting in 5 seconds..
Nonetheless, the function pointer method isn't so terrific anyway, so
this patch replaces it with a simple boolean, which also gets inlined
away. This successfully works around the Clang bug.
In general, I'm not too keen on all of the indirection involved here; it
clearly does more harm than good. Hopefully the whole thing can get
cleaned up down the road when lib/crypto is overhauled more
comprehensively. But for now, we go with a simple bandaid.
Fixes: 6048fdcc5f ("lib/crypto: blake2s: include as built-in")
Link: https://github.com/ClangBuiltLinux/linux/issues/1567
Reported-by: Miles Chen <miles.chen@mediatek.com>
Tested-by: Miles Chen <miles.chen@mediatek.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: John Stultz <john.stultz@linaro.org>
Acked-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
CPUID.(EAX=7,ECX=0):EBX.FDP_EXCPTN_ONLY[bit 6] and
CPUID.(EAX=7,ECX=0):EBX.ZERO_FCS_FDS[bit 13] are "defeature"
bits. Unlike most of the other CPUID feature bits, these bits are
clear if the features are present and set if the features are not
present. These bits should be reported in KVM_GET_SUPPORTED_CPUID,
because if these bits are set on hardware, they cannot be cleared in
the guest CPUID. Doing so would claim guest support for a feature that
the hardware doesn't support and that can't be efficiently emulated.
Of course, any software (e.g WIN87EM.DLL) expecting these features to
be present likely predates these CPUID feature bits and therefore
doesn't know to check for them anyway.
Aaron Lewis added the corresponding X86_FEATURE macros in
commit cbb99c0f58 ("x86/cpufeatures: Add FDP_EXCPTN_ONLY and
ZERO_FCS_FDS"), with the intention of reporting these bits in
KVM_GET_SUPPORTED_CPUID, but I was unable to find a proposed patch on
the kvm list.
Opportunistically reordered the CPUID_7_0_EBX capability bits from
least to most significant.
Cc: Aaron Lewis <aaronlewis@google.com>
Signed-off-by: Jim Mattson <jmattson@google.com>
Message-Id: <20220204001348.2844660-1-jmattson@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This started out with me noticing that "dom0_max_vcpus=<N>" with <N>
larger than the number of physical CPUs reported through ACPI tables
would not bring up the "excess" vCPU-s. Addressing this is the primary
purpose of the change; CPU maps handling is being tidied only as far as
is necessary for the change here (with the effect of also avoiding the
setting up of too much per-CPU infrastructure, i.e. for CPUs which can
never come online).
Noticing that xen_fill_possible_map() is called way too early, whereas
xen_filter_cpu_maps() is called too late (after per-CPU areas were
already set up), and further observing that each of the functions serves
only one of Dom0 or DomU, it looked like it was better to simplify this.
Use the .get_smp_config hook instead, uniformly for Dom0 and DomU.
xen_fill_possible_map() can be dropped altogether, while
xen_filter_cpu_maps() is re-purposed but not otherwise changed.
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Link: https://lore.kernel.org/r/2dbd5f0a-9859-ca2d-085e-a02f7166c610@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>
In __WARN_FLAGS(), we had two asm statements (abbreviated):
asm volatile("ud2");
asm volatile(".pushsection .discard.reachable");
These pair of statements are used to trigger an exception, but then help
objtool understand that for warnings, control flow will be restored
immediately afterwards.
The problem is that volatile is not a compiler barrier. GCC explicitly
documents this:
> Note that the compiler can move even volatile asm instructions
> relative to other code, including across jump instructions.
Also, no clobbers are specified to prevent instructions from subsequent
statements from being scheduled by compiler before the second asm
statement. This can lead to instructions from subsequent statements
being emitted by the compiler before the second asm statement.
Providing a scheduling model such as via -march= options enables the
compiler to better schedule instructions with known latencies to hide
latencies from data hazards compared to inline asm statements in which
latencies are not estimated.
If an instruction gets scheduled by the compiler between the two asm
statements, then objtool will think that it is not reachable, producing
a warning.
To prevent instructions from being scheduled in between the two asm
statements, merge them.
Also remove an unnecessary unreachable() asm annotation from BUG() in
favor of __builtin_unreachable(). objtool is able to track that the ud2
from BUG() terminates control flow within the function.
Link: https://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html#Volatile
Link: https://github.com/ClangBuiltLinux/linux/issues/1483
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Link: https://lore.kernel.org/r/20220202205557.2260694-1-ndesaulniers@google.com
Add a check for !buf->single before calling pt_buffer_region_size in a
place where a missing check can cause a kernel crash.
Fixes a bug introduced by commit 670638477a ("perf/x86/intel/pt:
Opportunistically use single range output mode"), which added a
support for PT single-range output mode. Since that commit if a PT
stop filter range is hit while tracing, the kernel will crash because
of a null pointer dereference in pt_handle_status due to calling
pt_buffer_region_size without a ToPA configured.
The commit which introduced single-range mode guarded almost all uses of
the ToPA buffer variables with checks of the buf->single variable, but
missed the case where tracing was stopped by the PT hardware, which
happens when execution hits a configured stop filter.
Tested that hitting a stop filter while PT recording successfully
records a trace with this patch but crashes without this patch.
Fixes: 670638477a ("perf/x86/intel/pt: Opportunistically use single range output mode")
Signed-off-by: Tristan Hume <tristan@thume.ca>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: stable@kernel.org
Link: https://lkml.kernel.org/r/20220127220806.73664-1-tristan@thume.ca
For consistency and clarity, migrate x86 over to the generic helpers for
guest timing and lockdep/RCU/tracing management, and remove the
x86-specific helpers.
Prior to this patch, the guest timing was entered in
kvm_guest_enter_irqoff() (called by svm_vcpu_enter_exit() and
svm_vcpu_enter_exit()), and was exited by the call to
vtime_account_guest_exit() within vcpu_enter_guest().
To minimize duplication and to more clearly balance entry and exit, both
entry and exit of guest timing are placed in vcpu_enter_guest(), using
the new guest_timing_{enter,exit}_irqoff() helpers. When context
tracking is used a small amount of additional time will be accounted
towards guests; tick-based accounting is unnaffected as IRQs are
disabled at this point and not enabled until after the return from the
guest.
This also corrects (benign) mis-balanced context tracking accounting
introduced in commits:
ae95f566b3 ("KVM: X86: TSCDEADLINE MSR emulation fastpath")
26efe2fd92 ("KVM: VMX: Handle preemption timer fastpath")
Where KVM can enter a guest multiple times, calling vtime_guest_enter()
without a corresponding call to vtime_account_guest_exit(), and with
vtime_account_system() called when vtime_account_guest() should be used.
As account_system_time() checks PF_VCPU and calls account_guest_time(),
this doesn't result in any functional problem, but is unnecessarily
confusing.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jim Mattson <jmattson@google.com>
Cc: Joerg Roedel <joro@8bytes.org>
Cc: Sean Christopherson <seanjc@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vitaly Kuznetsov <vkuznets@redhat.com>
Cc: Wanpeng Li <wanpengli@tencent.com>
Message-Id: <20220201132926.3301912-4-mark.rutland@arm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Handle non-APICv interrupt delivery in vendor code, even though it means
VMX and SVM will temporarily have duplicate code. SVM's AVIC has a race
condition that requires KVM to fall back to legacy interrupt injection
_after_ the interrupt has been logged in the vIRR, i.e. to fix the race,
SVM will need to open code the full flow anyways[*]. Refactor the code
so that the SVM bug without introducing other issues, e.g. SVM would
return "success" and thus invoke trace_kvm_apicv_accept_irq() even when
delivery through the AVIC failed, and to opportunistically prepare for
using KVM_X86_OP to fill each vendor's kvm_x86_ops struct, which will
rely on the vendor function matching the kvm_x86_op pointer name.
No functional change intended.
[*] https://lore.kernel.org/all/20211213104634.199141-4-mlevitsk@redhat.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220128005208.4008533-3-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
