- 12 Jan, 2011 40 commits
-
-
Xiao Guangrong authored
If reserved bit is set, we need inject the #PF with PFEC.RSVD=1, but shadow_notrap_nonpresent_pte injects #PF with PFEC.RSVD=0 only Signed-off-by:
Xiao Guangrong <xiaoguangrong@cn.fujitsu.com> Signed-off-by:
Marcelo Tosatti <mtosatti@redhat.com>
-
Jan Kiszka authored
Adds API documentation for KVM_[DE]ASSIGN_PCI_DEVICE, KVM_[DE]ASSIGN_DEV_IRQ, KVM_SET_GSI_ROUTING, KVM_ASSIGN_SET_MSIX_NR, and KVM_ASSIGN_SET_MSIX_ENTRY. Acked-by:
Alex Williamson <alex.williamson@redhat.com> Acked-by:
Michael S. Tsirkin <mst@redhat.com> Signed-off-by:
Jan Kiszka <jan.kiszka@siemens.com> Signed-off-by:
-
Jan Kiszka authored
Any arch not supporting device assigment will also not build assigned-dev.c. So testing for KVM_CAP_DEVICE_DEASSIGNMENT is pointless. KVM_CAP_ASSIGN_DEV_IRQ is unconditinally set. Moreover, add a default case for dispatching the ioctl. Acked-by:
-
Jan Kiszka authored
The guest may change states that pci_reset_function does not touch. So we better save/restore the assigned device across guest usage. Acked-by:
-
Jan Kiszka authored
Cosmetic change, but it helps to correlate IRQs with PCI devices. Acked-by:
-
Jan Kiszka authored
This improves the IRQ forwarding for assigned devices: By using the kernel's threaded IRQ scheme, we can get rid of the latency-prone work queue and simplify the code in the same run. Moreover, we no longer have to hold assigned_dev_lock while raising the guest IRQ, which can be a lenghty operation as we may have to iterate over all VCPUs. The lock is now only used for synchronizing masking vs. unmasking of INTx-type IRQs, thus is renames to intx_lock. Acked-by:
-
Jan Kiszka authored
When we deassign a guest IRQ, clear the potentially asserted guest line. There might be no chance for the guest to do this, specifically if we switch from INTx to MSI mode. Acked-by:
-
Avi Kivity authored
This allows Linux to mask cpuid bits if, for example, nx is enabled on only some cpus. Signed-off-by:
Avi Kivity <avi@redhat.com> Signed-off-by:
-
Avi Kivity authored
Instead of querying cpuid directly, use the Linux accessors (boot_cpu_has, etc.). This allows the things like the clearcpuid kernel command line to work (when it's fixed wrt scattered cpuid bits). Acked-by:
Joerg Roedel <joerg.roedel@amd.com> Signed-off-by:
-
Xiao Guangrong authored
If apf is generated in L2 guest and is completed in L1 guest, it will prefault this apf in L1 guest's mmu context. Signed-off-by:
-
Xiao Guangrong authored
Let's support apf for nonpaing guest Signed-off-by:
-
Xiao Guangrong authored
If CR0.PG is changed, the page fault cann't be avoid when the prefault address is accessed later And it also fix a bug: it can retry a page enabled #PF in page disabled context if mmu is shadow page This idear is from Gleb Natapov Signed-off-by:
-
Xiao Guangrong authored
Add AUDIT_POST_SYNC audit for long mode shadow page Signed-off-by:
-
Jan Kiszka authored
IA64 support forces us to abstract the allocation of the kvm structure. But instead of mixing this up with arch-specific initialization and doing the same on destruction, split both steps. This allows to move generic destruction calls into generic code. It also fixes error clean-up on failures of kvm_create_vm for IA64. Signed-off-by:
-
Tracey Dent authored
Changed makefile to use the ccflags-y option instead of EXTRA_CFLAGS. Signed-off-by:
Tracey Dent <tdent48227@gmail.com> Signed-off-by:
-
Xiao Guangrong authored
Remove the declaration of kvm_mmu_set_base_ptes() Signed-off-by:
-
Jan Kiszka authored
Signed-off-by:
-
Avi Kivity authored
While not mandated by the spec, Linux relies on NMI being blocked by an IF-enabling STI. VMX also refuses to enter a guest in this state, at least on some implementations. Disallow NMI while blocked by STI by checking for the condition, and requesting an interrupt window exit if it occurs. Signed-off-by:
-
Xiao Guangrong authored
In kvm_async_pf_wakeup_all(), we add a dummy apf to vcpu->async_pf.done without holding vcpu->async_pf.lock, it will break if we are handling apfs at this time. Also use 'list_empty_careful()' instead of 'list_empty()' Signed-off-by:
Gleb Natapov <gleb@redhat.com> Signed-off-by:
-
Xiao Guangrong authored
If it's no need to inject async #PF to PV guest we can handle more completed apfs at one time, so we can retry guest #PF as early as possible Signed-off-by:
-
Xiao Guangrong authored
In current code, it checks async pf completion out of the wait context, like this: if (vcpu->arch.mp_state == KVM_MP_STATE_RUNNABLE && !vcpu->arch.apf.halted) r = vcpu_enter_guest(vcpu); else { ...... kvm_vcpu_block(vcpu) ^- waiting until 'async_pf.done' is not empty } kvm_check_async_pf_completion(vcpu) ^- delete list from async_pf.done So, if we check aysnc pf completion first, it can be blocked at kvm_vcpu_block Fixed by mark the vcpu is unhalted in kvm_check_async_pf_completion() path Signed-off-by: -
Xiao Guangrong authored
Don't search later slots if the slot is empty Acked-by:
-
Xiao Guangrong authored
Use 'DECLARE_EVENT_CLASS' to cleanup async_pf tracepoints Acked-by:
-
Xiao Guangrong authored
Tracing 'async' and *pfn is useless, since 'async' is always true, and '*pfn' is always "fault_pfn' We can trace 'gva' and 'gfn' instead, it can help us to see the life-cycle of an async_pf Signed-off-by:
-
Takuya Yoshikawa authored
Let's use newly introduced vzalloc(). Signed-off-by:
Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp> Signed-off-by:
Jesper Juhl <jj@chaosbits.net> Signed-off-by:
-
Gleb Natapov authored
Currently the exit is unhandled, so guest halts with error if it tries to execute INVD instruction. Call into emulator when INVD instruction is executed by a guest instead. This instruction is not needed by ordinary guests, but firmware (like OpenBIOS) use it and fail. Signed-off-by:
-
Jan Kiszka authored
Micro optimization to avoid calling wbinvd twice on the CPU that has to emulate it. As we might be preempted between smp_call_function_many and the local wbinvd, the cache might be filled again so that real work could be done uselessly. Signed-off-by:
-
Heiko Carstens authored
Fixes this: CC arch/s390/kvm/../../../virt/kvm/kvm_main.o arch/s390/kvm/../../../virt/kvm/kvm_main.c: In function 'kvm_dev_ioctl_create_vm': arch/s390/kvm/../../../virt/kvm/kvm_main.c:1828:10: warning: unused variable 'r' Signed-off-by:
Heiko Carstens <heiko.carstens@de.ibm.com> Signed-off-by:
-
Heiko Carstens authored
Fixes this: CC arch/s390/kvm/../../../virt/kvm/kvm_main.o arch/s390/kvm/../../../virt/kvm/kvm_main.c: In function 'kvm_clear_guest_page': arch/s390/kvm/../../../virt/kvm/kvm_main.c:1224:2: warning: passing argument 3 of 'kvm_write_guest_page' makes pointer from integer without a cast arch/s390/kvm/../../../virt/kvm/kvm_main.c:1185:5: note: expected 'const void *' but argument is of type 'long unsigned int' Signed-off-by:
-
Takuya Yoshikawa authored
Currently we are using vmalloc() for all dirty bitmaps even if they are small enough, say less than K bytes. We use kmalloc() if dirty bitmap size is less than or equal to PAGE_SIZE so that we can avoid vmalloc area usage for VGA. This will also make the logging start/stop faster. Signed-off-by:
-
Takuya Yoshikawa authored
Currently x86's kvm_vm_ioctl_get_dirty_log() needs to allocate a bitmap by vmalloc() which will be used in the next logging and this has been causing bad effect to VGA and live-migration: vmalloc() consumes extra systime, triggers tlb flush, etc. This patch resolves this issue by pre-allocating one more bitmap and switching between two bitmaps during dirty logging. Performance improvement: I measured performance for the case of VGA update by trace-cmd. The result was 1.5 times faster than the original one. In the case of live migration, the improvement ratio depends on the workload and the guest memory size. In general, the larger the memory size is the more benefits we get. Note: This does not change other architectures's logic but the allocation size becomes twice. This will increase the actual memory consumption only when the new size changes the number of pages allocated by vmalloc(). Signed-off-by:
Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp> Signed-off-by:
-
Takuya Yoshikawa authored
This makes it easy to change the way of allocating/freeing dirty bitmaps. Signed-off-by:
-
Gleb Natapov authored
Add tracepoint for userspace exit. Signed-off-by:
-
Marcelo Tosatti authored
As suggested by Andrea, pass r/w error code to gup(), upgrading read fault to writable if host pte allows it. Signed-off-by:
-
Marcelo Tosatti authored
This can happen in the following scenario: vcpu0 vcpu1 read fault gup(.write=0) gup(.write=1) reuse swap cache, no COW set writable spte use writable spte set read-only spte Signed-off-by:
-
Marcelo Tosatti authored
Unused. Signed-off-by:
-
Marcelo Tosatti authored
The EPT present/writable bits use the same position as normal pagetable bits. Since direct_map passes ACC_ALL to mmu_set_spte, thus always setting the writable bit on sptes, use the generic PT_PRESENT shadow_base_pte. Also pass present/writable error code information from EPT violation to generic pagefault handler. Signed-off-by:
-
Avi Kivity authored
After an interrupt injection, the PPR changes, and we have to reflect that into the vapic. This causes a KVM_REQ_EVENT to be set, which causes the whole interrupt injection routine to be run again (harmlessly). Optimize by only setting KVM_REQ_EVENT if the ppr was lowered; otherwise there is no chance that a new injection is needed. Signed-off-by: -
Avi Kivity authored
This abstraction only serves to obfuscate. Remove. Signed-off-by:
-
Avi Kivity authored
ldt is never used in the kernel context; same goes for fs (x86_64) and gs (i386). So save/restore them in the heavyweight exit path instead of the lightweight path. By itself, this doesn't buy us much, but it paves the way for moving vmload and vmsave to the heavyweight exit path, since they modify the same registers. [jan: fix copy/pase mistake on i386] Signed-off-by:
-
