- 08 Jun, 2022 40 commits
-
-
Like Xu authored
The CPUID features PDCM, DS and DTES64 are required for PEBS feature. KVM would expose CPUID feature PDCM, DS and DTES64 to guest when PEBS is supported in the KVM on the Ice Lake server platforms. Originally-by:
Andi Kleen <ak@linux.intel.com> Co-developed-by:
Kan Liang <kan.liang@linux.intel.com> Signed-off-by:
Luwei Kang <luwei.kang@intel.com> Signed-off-by:
Like Xu <likexu@tencent.com> Message-Id: <20220411101946.20262-18-likexu@tencent.com> Signed-off-by:
Paolo Bonzini <pbonzini@redhat.com>
-
Like Xu authored
For the same purpose, the leagcy intel_pmu_lbr_is_compatible() can be renamed for reuse by more callers, and remove the comment about LBR use case can be deleted by the way. Signed-off-by:
Like Xu <like.xu@linux.intel.com> Acked-by:
Peter Zijlstra (Intel) <peterz@infradead.org> Message-Id: <20220411101946.20262-17-likexu@tencent.com> Signed-off-by:
-
Like Xu authored
The information obtained from the interface perf_get_x86_pmu_capability() doesn't change, so an exported "struct x86_pmu_capability" is introduced for all guests in the KVM, and it's initialized before hardware_setup(). Signed-off-by:
-
Like Xu authored
The guest PEBS will be disabled when some users try to perf KVM and its user-space through the same PEBS facility OR when the host perf doesn't schedule the guest PEBS counter in a one-to-one mapping manner (neither of these are typical scenarios). The PEBS records in the guest DS buffer are still accurate and the above two restrictions will be checked before each vm-entry only if guest PEBS is deemed to be enabled. Suggested-by:
Wei Wang <wei.w.wang@intel.com> Signed-off-by:
-
Like Xu authored
It allows this inline function to be reused by more callers in more files, such as pmu_intel.c. Signed-off-by:
-
Like Xu authored
The bit 12 represents "Processor Event Based Sampling Unavailable (RO)" : 1 = PEBS is not supported. 0 = PEBS is supported. A write to this PEBS_UNAVL available bit will bring #GP(0) when guest PEBS is enabled. Some PEBS drivers in guest may care about this bit. Signed-off-by:
-
Like Xu authored
If IA32_PERF_CAPABILITIES.PEBS_BASELINE [bit 14] is set, the adaptive PEBS is supported. The PEBS_DATA_CFG MSR and adaptive record enable bits (IA32_PERFEVTSELx.Adaptive_Record and IA32_FIXED_CTR_CTRL. FCx_Adaptive_Record) are also supported. Adaptive PEBS provides software the capability to configure the PEBS records to capture only the data of interest, keeping the record size compact. An overflow of PMCx results in generation of an adaptive PEBS record with state information based on the selections specified in MSR_PEBS_DATA_CFG.By default, the record only contain the Basic group. When guest adaptive PEBS is enabled, the IA32_PEBS_ENABLE MSR will be added to the perf_guest_switch_msr() and switched during the VMX transitions just like CORE_PERF_GLOBAL_CTRL MSR. According to Intel SDM, software is recommended to PEBS Baseline when the following is true. IA32_PERF_CAPABILITIES.PEBS_BASELINE[14] && IA32_PERF_CAPABILITIES.PEBS_FMT[11:8] ≥ 4. Co-developed-by:
-
Like Xu authored
When CPUID.01H:EDX.DS[21] is set, the IA32_DS_AREA MSR exists and points to the linear address of the first byte of the DS buffer management area, which is used to manage the PEBS records. When guest PEBS is enabled, the MSR_IA32_DS_AREA MSR will be added to the perf_guest_switch_msr() and switched during the VMX transitions just like CORE_PERF_GLOBAL_CTRL MSR. The WRMSR to IA32_DS_AREA MSR brings a #GP(0) if the source register contains a non-canonical address. Originally-by:
-
Like Xu authored
The PEBS-PDIR facility on Ice Lake server is supported on IA31_FIXED0 only. If the guest configures counter 32 and PEBS is enabled, the PEBS-PDIR facility is supposed to be used, in which case KVM adjusts attr.precise_ip to 3 and request host perf to assign the exactly requested counter or fail. The CPU model check is also required since some platforms may place the PEBS-PDIR facility in another counter index. Signed-off-by:
-
Like Xu authored
When a guest counter is configured as a PEBS counter through IA32_PEBS_ENABLE, a guest PEBS event will be reprogrammed by configuring a non-zero precision level in the perf_event_attr. The guest PEBS overflow PMI bit would be set in the guest GLOBAL_STATUS MSR when PEBS facility generates a PEBS overflow PMI based on guest IA32_DS_AREA MSR. Even with the same counter index and the same event code and mask, guest PEBS events will not be reused for non-PEBS events. Originally-by:
-
Like Xu authored
If IA32_PERF_CAPABILITIES.PEBS_BASELINE [bit 14] is set, the IA32_PEBS_ENABLE MSR exists and all architecturally enumerated fixed and general-purpose counters have corresponding bits in IA32_PEBS_ENABLE that enable generation of PEBS records. The general-purpose counter bits start at bit IA32_PEBS_ENABLE[0], and the fixed counter bits start at bit IA32_PEBS_ENABLE[32]. When guest PEBS is enabled, the IA32_PEBS_ENABLE MSR will be added to the perf_guest_switch_msr() and atomically switched during the VMX transitions just like CORE_PERF_GLOBAL_CTRL MSR. Based on whether the platform supports x86_pmu.pebs_ept, it has also refactored the way to add more msrs to arr[] in intel_guest_get_msrs() for extensibility. Originally-by:
-
Peter Zijlstra (Intel) authored
The value of pebs_counter_mask will be accessed frequently for repeated use in the intel_guest_get_msrs(). So it can be optimized instead of endlessly mucking about with branches. Signed-off-by:
-
Like Xu authored
The mask value of fixed counter control register should be dynamic adjusted with the number of fixed counters. This patch introduces a variable that includes the reserved bits of fixed counter control registers. This is a generic code refactoring. Co-developed-by:
-
Like Xu authored
On Intel platforms, the software can use the IA32_MISC_ENABLE[7] bit to detect whether the processor supports performance monitoring facility. It depends on the PMU is enabled for the guest, and a software write operation to this available bit will be ignored. The proposal to ignore the toggle in KVM is the way to go and that behavior matches bare metal. Signed-off-by:
-
Like Xu authored
Splitting the logic for determining the guest values is unnecessarily confusing, and potentially fragile. Perf should have full knowledge and control of what values are loaded for the guest. If we change .guest_get_msrs() to take a struct kvm_pmu pointer, then it can generate the full set of guest values by grabbing guest ds_area and pebs_data_cfg. Alternatively, .guest_get_msrs() could take the desired guest MSR values directly (ds_area and pebs_data_cfg), but kvm_pmu is vendor agnostic, so we don't see any reason to not just pass the pointer. Suggested-by:
Sean Christopherson <seanjc@google.com> Signed-off-by:
-
Like Xu authored
With PEBS virtualization, the guest PEBS records get delivered to the guest DS, and the host pmi handler uses perf_guest_cbs->is_in_guest() to distinguish whether the PMI comes from the guest code like Intel PT. No matter how many guest PEBS counters are overflowed, only triggering one fake event is enough. The fake event causes the KVM PMI callback to be called, thereby injecting the PEBS overflow PMI into the guest. KVM may inject the PMI with BUFFER_OVF set, even if the guest DS is empty. That should really be harmless. Thus guest PEBS handler would retrieve the correct information from its own PEBS records buffer. Cc: linux-perf-users@vger.kernel.org Originally-by:
-
Like Xu authored
Add support for EPT-Friendly PEBS, a new CPU feature that enlightens PEBS to translate guest linear address through EPT, and facilitates handling VM-Exits that occur when accessing PEBS records. More information can be found in the December 2021 release of Intel's SDM, Volume 3, 18.9.5 "EPT-Friendly PEBS". This new hardware facility makes sure the guest PEBS records will not be lost, which is available on Intel Ice Lake Server platforms (and later). KVM will check this field through perf_get_x86_pmu_capability() instead of hard coding the CPU models in the KVM code. If it is supported, the guest PEBS capability will be exposed to the guest. Guest PEBS can be enabled when and only when "EPT-Friendly PEBS" is supported and EPT is enabled. Cc: linux-perf-users@vger.kernel.org Signed-off-by:
-
Chao Gao authored
With IPI virtualization enabled, the processor emulates writes to APIC registers that would send IPIs. The processor sets the bit corresponding to the vector in target vCPU's PIR and may send a notification (IPI) specified by NDST and NV fields in target vCPU's Posted-Interrupt Descriptor (PID). It is similar to what IOMMU engine does when dealing with posted interrupt from devices. A PID-pointer table is used by the processor to locate the PID of a vCPU with the vCPU's APIC ID. The table size depends on maximum APIC ID assigned for current VM session from userspace. Allocating memory for PID-pointer table is deferred to vCPU creation, because irqchip mode and VM-scope maximum APIC ID is settled at that point. KVM can skip PID-pointer table allocation if !irqchip_in_kernel(). Like VT-d PI, if a vCPU goes to blocked state, VMM needs to switch its notification vector to wakeup vector. This can ensure that when an IPI for blocked vCPUs arrives, VMM can get control and wake up blocked vCPUs. And if a VCPU is preempted, its posted interrupt notification is suppressed. Note that IPI virtualization can only virualize physical-addressing, flat mode, unicast IPIs. Sending other IPIs would still cause a trap-like APIC-write VM-exit and need to be handled by VMM. Signed-off-by:
Chao Gao <chao.gao@intel.com> Signed-off-by:
Zeng Guang <guang.zeng@intel.com> Message-Id: <20220419154510.11938-1-guang.zeng@intel.com> Signed-off-by:
-
Zeng Guang authored
Basic test coverage of KVM_CAP_MAX_VCPU_ID cap. This capability can be enabled before vCPU creation and only allowed to set once. if assigned vcpu id is beyond KVM_CAP_MAX_VCPU_ID capability, vCPU creation will fail. Signed-off-by:
-
Zeng Guang authored
Introduce new max_vcpu_ids in KVM for x86 architecture. Userspace can assign maximum possible vcpu id for current VM session using KVM_CAP_MAX_VCPU_ID of KVM_ENABLE_CAP ioctl(). This is done for x86 only because the sole use case is to guide memory allocation for PID-pointer table, a structure needed to enable VMX IPI. By default, max_vcpu_ids set as KVM_MAX_VCPU_IDS. Suggested-by:
Maxim Levitsky <mlevitsk@redhat.com> Signed-off-by:
-
Zeng Guang authored
kvm_arch_vcpu_precreate() targets to handle arch specific VM resource to be prepared prior to the actual creation of vCPU. For example, x86 platform may need do per-VM allocation based on max_vcpu_ids at the first vCPU creation. It probably leads to concurrency control on this allocation as multiple vCPU creation could happen simultaneously. From the architectual point of view, it's necessary to execute kvm_arch_vcpu_precreate() under protect of kvm->lock. Currently only arm64, x86 and s390 have non-nop implementations at the stage of vCPU pre-creation. Remove the lock acquiring in s390's design and make sure all architecture can run kvm_arch_vcpu_precreate() safely under kvm->lock without recrusive lock issue. Suggested-by:
-
Zeng Guang authored
Remove the condition check cpu_has_secondary_exec_ctrls(). Calling vmx_refresh_apicv_exec_ctrl() premises secondary controls activated and VMCS fields related to APICv valid as well. If it's invoked in wrong circumstance at the worst case, VMX operation will report VMfailValid error without further harmful impact and just functions as if all the secondary controls were 0. Suggested-by:
-
Zeng Guang authored
Upcoming Intel CPUs will support virtual x2APIC MSR writes to the vICR, i.e. will trap and generate an APIC-write VM-Exit instead of intercepting the WRMSR. Add support for handling "nodecode" x2APIC writes, which were previously impossible. Note, x2APIC MSR writes are 64 bits wide. Signed-off-by:
-
Robert Hoo authored
Add tertiary_exec_control field report in dump_vmcs(). Meanwhile, reorganize the dump output of VMCS category as follows. Before change: *** Control State *** PinBased=0x000000ff CPUBased=0xb5a26dfa SecondaryExec=0x061037eb EntryControls=0000d1ff ExitControls=002befff After change: *** Control State *** CPUBased=0xb5a26dfa SecondaryExec=0x061037eb TertiaryExec=0x0000000000000010 PinBased=0x000000ff EntryControls=0000d1ff ExitControls=002befff Reviewed-by:
Robert Hoo <robert.hu@linux.intel.com> Signed-off-by:
-
Robert Hoo authored
Check VMX features on tertiary execution control in VMCS config setup. Sub-features in tertiary execution control to be enabled are adjusted according to hardware capabilities although no sub-feature is enabled in this patch. EVMCSv1 doesn't support tertiary VM-execution control, so disable it when EVMCSv1 is in use. And define the auxiliary functions for Tertiary control field here, using the new BUILD_CONTROLS_SHADOW(). Reviewed-by:
-
Robert Hoo authored
The Tertiary VM-Exec Control, different from previous control fields, is 64 bit. So extend BUILD_CONTROLS_SHADOW() by adding a 'bit' parameter, to support both 32 bit and 64 bit fields' auxiliary functions building. Suggested-by:
-
Robert Hoo authored
A new 64-bit control field "tertiary processor-based VM-execution controls", is defined [1]. It's controlled by bit 17 of the primary processor-based VM-execution controls. Different from its brother VM-execution fields, this tertiary VM- execution controls field is 64 bit. So it occupies 2 vmx_feature_leafs, TERTIARY_CTLS_LOW and TERTIARY_CTLS_HIGH. Its companion VMX capability reporting MSR,MSR_IA32_VMX_PROCBASED_CTLS3 (0x492), is also semantically different from its brothers, whose 64 bits consist of all allow-1, rather than 32-bit allow-0 and 32-bit allow-1 [1][2]. Therefore, its init_vmx_capabilities() is a little different from others. [1] ISE 6.2 "VMCS Changes" https://www.intel.com/content/www/us/en/develop/download/intel-architecture-instruction-set-extensions-programming-reference.html [2] SDM Vol3. Appendix A.3 Reviewed-by:
-
Sean Christopherson authored
Add a comment to FNAME(sync_page) to explain why the TLB flushing logic conspiculously doesn't handle the scenario of guest protections being reduced. Specifically, if synchronizing a SPTE drops execute protections, KVM will not emit a TLB flush, whereas dropping writable or clearing A/D bits does trigger a flush via mmu_spte_update(). Architecturally, until the GPTE is implicitly or explicitly flushed from the guest's perspective, KVM is not required to flush any old, stale translations. Signed-off-by:
Jim Mattson <jmattson@google.com> Message-Id: <20220513195000.99371-3-seanjc@google.com> Signed-off-by:
-
Sean Christopherson authored
All of sync_page()'s existing checks filter out only !PRESENT gPTE, because without execute-only, all upper levels are guaranteed to be at least READABLE. However, if EPT with execute-only support is in use by L1, KVM can create an SPTE that is shadow-present but guest-inaccessible (RWX=0) if the upper level combined permissions are R (or RW) and the leaf EPTE is changed from R (or RW) to X. Because the EPTE is considered present when viewed in isolation, and no reserved bits are set, FNAME(prefetch_invalid_gpte) will consider the GPTE valid, and cause a not-present SPTE to be created. The SPTE is "correct": the guest translation is inaccessible because the combined protections of all levels yield RWX=0, and KVM will just redirect any vmexits to the guest. If EPT A/D bits are disabled, KVM can mistake the SPTE for an access-tracked SPTE, but again such confusion isn't fatal, as the "saved" protections are also RWX=0. However, creating a useless SPTE in general means that KVM messed up something, even if this particular goof didn't manifest as a functional bug. So, drop SPTEs whose new protections will yield a RWX=0 SPTE, and add a WARN in make_spte() to detect creation of SPTEs that will result in RWX=0 protections. Fixes: d95c5568 ("kvm: mmu: track read permission explicitly for shadow EPT page tables") Cc: David Matlack <dmatlack@google.com> Cc: Ben Gardon <bgardon@google.com> Signed-off-by:
-
Maciej S. Szmigiero authored
Add a KVM self-test that checks whether a nSVM L1 is able to successfully inject a software interrupt, a soft exception and a NMI into its L2 guest. In practice, this tests both the next_rip field consistency and L1-injected event with intervening L0 VMEXIT during its delivery: the first nested VMRUN (that's also trying to inject a software interrupt) will immediately trigger a L0 NPF. This L0 NPF will have zero in its CPU-returned next_rip field, which if incorrectly reused by KVM will trigger a #PF when trying to return to such address 0 from the interrupt handler. For NMI injection this tests whether the L1 NMI state isn't getting incorrectly mixed with the L2 NMI state if a L1 -> L2 NMI needs to be re-injected. Reviewed-by:
Maciej S. Szmigiero <maciej.szmigiero@oracle.com> Message-Id: <d5f3d56528558ad8e28a9f1e1e4187f5a1e6770a.1651440202.git.maciej.szmigiero@oracle.com> Signed-off-by:
-
Maciej S. Szmigiero authored
A NMI that L1 wants to inject into its L2 should be directly re-injected, without causing L0 side effects like engaging NMI blocking for L1. It's also worth noting that in this case it is L1 responsibility to track the NMI window status for its L2 guest. Signed-off-by:
-
Sean Christopherson authored
In the IRQ injection tracepoint, differentiate between Hard IRQs and Soft "IRQs", i.e. interrupts that are reinjected after incomplete delivery of a software interrupt from an INTn instruction. Tag reinjected interrupts as such, even though the information is usually redundant since soft interrupts are only ever reinjected by KVM. Though rare in practice, a hard IRQ can be reinjected. Signed-off-by:
-
Sean Christopherson authored
Print the error code in the exception injection tracepoint if and only if the exception has an error code. Define the entire error code sequence as a set of formatted strings, print empty strings if there's no error code, and abuse __print_symbolic() by passing it an empty array to coerce it into printing the error code as a hex string. Signed-off-by:
-
Sean Christopherson authored
Trace exceptions that are re-injected, not just those that KVM is injecting for the first time. Debugging re-injection bugs is painful enough as is, not having visibility into what KVM is doing only makes things worse. Delay propagating pending=>injected in the non-reinjection path so that the tracing can properly identify reinjected exceptions. Signed-off-by:
-
Sean Christopherson authored
Re-inject INTn software interrupts instead of retrying the instruction if the CPU encountered an intercepted exception while vectoring the INTn, e.g. if KVM intercepted a #PF when utilizing shadow paging. Retrying the instruction is architecturally wrong e.g. will result in a spurious #DB if there's a code breakpoint on the INT3/O, and lack of re-injection also breaks nested virtualization, e.g. if L1 injects a software interrupt and vectoring the injected interrupt encounters an exception that is intercepted by L0 but not L1. Signed-off-by:
-
Sean Christopherson authored
Re-inject INT3/INTO instead of retrying the instruction if the CPU encountered an intercepted exception while vectoring the software exception, e.g. if vectoring INT3 encounters a #PF and KVM is using shadow paging. Retrying the instruction is architecturally wrong, e.g. will result in a spurious #DB if there's a code breakpoint on the INT3/O, and lack of re-injection also breaks nested virtualization, e.g. if L1 injects a software exception and vectoring the injected exception encounters an exception that is intercepted by L0 but not L1. Due to, ahem, deficiencies in the SVM architecture, acquiring the next RIP may require flowing through the emulator even if NRIPS is supported, as the CPU clears next_rip if the VM-Exit is due to an exception other than "exceptions caused by the INT3, INTO, and BOUND instructions". To deal with this, "skip" the instruction to calculate next_rip (if it's not already known), and then unwind the RIP write and any side effects (RFLAGS updates). Save the computed next_rip and use it to re-stuff next_rip if injection doesn't complete. This allows KVM to do the right thing if next_rip was known prior to injection, e.g. if L1 injects a soft event into L2, and there is no backing INTn instruction, e.g. if L1 is injecting an arbitrary event. Note, it's impossible to guarantee architectural correctness given SVM's architectural flaws. E.g. if the guest executes INTn (no KVM injection), an exit occurs while vectoring the INTn, and the guest modifies the code stream while the exit is being handled, KVM will compute the incorrect next_rip due to "skipping" the wrong instruction. A future enhancement to make this less awful would be for KVM to detect that the decoded instruction is not the correct INTn and drop the to-be-injected soft event (retrying is a lesser evil compared to shoving the wrong RIP on the exception stack). Reported-by:
-
Sean Christopherson authored
If NRIPS is supported in hardware but disabled in KVM, set next_rip to the next RIP when advancing RIP as part of emulating INT3 injection. There is no flag to tell the CPU that KVM isn't using next_rip, and so leaving next_rip is left as is will result in the CPU pushing garbage onto the stack when vectoring the injected event. Reviewed-by:
-
Sean Christopherson authored
Unwind the RIP advancement done by svm_queue_exception() when injecting an INT3 ultimately "fails" due to the CPU encountering a VM-Exit while vectoring the injected event, even if the exception reported by the CPU isn't the same event that was injected. If vectoring INT3 encounters an exception, e.g. #NP, and vectoring the #NP encounters an intercepted exception, e.g. #PF when KVM is using shadow paging, then the #NP will be reported as the event that was in-progress. Note, this is still imperfect, as it will get a false positive if the INT3 is cleanly injected, no VM-Exit occurs before the IRET from the INT3 handler in the guest, the instruction following the INT3 generates an exception (directly or indirectly), _and_ vectoring that exception encounters an exception that is intercepted by KVM. The false positives could theoretically be solved by further analyzing the vectoring event, e.g. by comparing the error code against the expected error code were an exception to occur when vectoring the original injected exception, but SVM without NRIPS is a complete disaster, trying to make it 100% correct is a waste of time. Reviewed-by:
-
Maciej S. Szmigiero authored
Don't BUG/WARN on interrupt injection due to GIF being cleared, since it's trivial for userspace to force the situation via KVM_SET_VCPU_EVENTS (even if having at least a WARN there would be correct for KVM internally generated injections). kernel BUG at arch/x86/kvm/svm/svm.c:3386! invalid opcode: 0000 [#1] SMP CPU: 15 PID: 926 Comm: smm_test Not tainted 5.17.0-rc3+ #264 Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015 RIP: 0010:svm_inject_irq+0xab/0xb0 [kvm_amd] Code: <0f> 0b 0f 1f 00 0f 1f 44 00 00 80 3d ac b3 01 00 00 55 48 89 f5 53 RSP: 0018:ffffc90000b37d88 EFLAGS: 00010246 RAX: 0000000000000000 RBX: ffff88810a234ac0 RCX: 0000000000000006 RDX: 0000000000000000 RSI: ffffc90000b37df7 RDI: ffff88810a234ac0 RBP: ffffc90000b37df7 R08: ffff88810a1fa410 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000 R13: ffff888109571000 R14: ffff88810a234ac0 R15: 0000000000000000 FS: 0000000001821380(0000) GS:ffff88846fdc0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 00007f74fc550008 CR3: 000000010a6fe000 CR4: 0000000000350ea0 Call Trace: <TASK> inject_pending_event+0x2f7/0x4c0 [kvm] kvm_arch_vcpu_ioctl_run+0x791/0x17a0 [kvm] kvm_vcpu_ioctl+0x26d/0x650 [kvm] __x64_sys_ioctl+0x82/0xb0 do_syscall_64+0x3b/0xc0 entry_SYSCALL_64_after_hwframe+0x44/0xae </TASK> Fixes: 219b65dc ("KVM: SVM: Improve nested interrupt injection") Cc: stable@vger.kernel.org Co-developed-by:
-
Maciej S. Szmigiero authored
The next_rip field of a VMCB is *not* an output-only field for a VMRUN. This field value (instead of the saved guest RIP) in used by the CPU for the return address pushed on stack when injecting a software interrupt or INT3 or INTO exception. Make sure this field gets synced from vmcb12 to vmcb02 when entering L2 or loading a nested state and NRIPS is exposed to L1. If NRIPS is supported in hardware but not exposed to L1 (nrips=0 or hidden by userspace), stuff vmcb02's next_rip from the new L2 RIP to emulate a !NRIPS CPU (which saves RIP on the stack as-is). Reviewed-by:
-
