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 <xiaoguangrong@cn.fujitsu.com>
Acked-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
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: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
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: Takuya Yoshikawa <yoshikawa.takuya@oss.ntt.co.jp>
Signed-off-by: Fernando Luis Vazquez Cao <fernando@oss.ntt.co.jp>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
If guest indicates that it can handle async pf in kernel mode too send
it, but only if interrupts are enabled.
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
If guest can detect that it runs in non-preemptable context it can
handle async PFs at any time, so let host know that it can send async
PF even if guest cpu is not in userspace.
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Send async page fault to a PV guest if it accesses swapped out memory.
Guest will choose another task to run upon receiving the fault.
Allow async page fault injection only when guest is in user mode since
otherwise guest may be in non-sleepable context and will not be able
to reschedule.
Vcpu will be halted if guest will fault on the same page again or if
vcpu executes kernel code.
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Keep track of memslots changes by keeping generation number in memslots
structure. Provide kvm_write_guest_cached() function that skips
gfn_to_hva() translation if memslots was not changed since previous
invocation.
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
When page is swapped in it is mapped into guest memory only after guest
tries to access it again and generate another fault. To save this fault
we can map it immediately since we know that guest is going to access
the page. Do it only when tdp is enabled for now. Shadow paging case is
more complicated. CR[034] and EFER registers should be switched before
doing mapping and then switched back.
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
If a guest accesses swapped out memory do not swap it in from vcpu thread
context. Schedule work to do swapping and put vcpu into halted state
instead.
Interrupts will still be delivered to the guest and if interrupt will
cause reschedule guest will continue to run another task.
[avi: remove call to get_user_pages_noio(), nacked by Linus; this
makes everything synchrnous again]
Acked-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
The only bit of EFER that affects the mmu is NX, and this is already
accounted for (LME only takes effect when changing cr0).
Based on a patch by Hillf Danton.
Signed-off-by: Avi Kivity <avi@redhat.com>
To support xsave properly for the guest the SVM module need
software support for it. As long as this is not present do
not report the xsave as supported feature in cpuid.
As a side-effect this patch moves the bit() helper function
into the x86.h file so that it can be used in svm.c too.
KVM-Stable-Tag.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
CPUID's OSXSAVE is a mirror of CR4.OSXSAVE bit. We need to update the CPUID
after migration.
KVM-Stable-Tag.
Signed-off-by: Sheng Yang <sheng@linux.intel.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
smp_call_function_many is specified to be called only with preemption
disabled. Fulfill this requirement.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Structures kvm_vcpu_events, kvm_debugregs, kvm_pit_state2 and
kvm_clock_data are copied to userland with some padding and reserved
fields unitialized. It leads to leaking of contents of kernel stack
memory. We have to initialize them to zero.
In patch v1 Jan Kiszka suggested to fill reserved fields with zeros
instead of memset'ting the whole struct. It makes sense as these
fields are explicitly marked as padding. No more fields need zeroing.
KVM-Stable-Tag.
Signed-off-by: Vasiliy Kulikov <segooon@gmail.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
I have observed the following bug trigger:
1. userspace calls GET_DIRTY_LOG
2. kvm_mmu_slot_remove_write_access is called and makes a page ro
3. page fault happens and makes the page writeable
fault is logged in the bitmap appropriately
4. kvm_vm_ioctl_get_dirty_log swaps slot pointers
a lot of time passes
5. guest writes into the page
6. userspace calls GET_DIRTY_LOG
At point (5), bitmap is clean and page is writeable,
thus, guest modification of memory is not logged
and GET_DIRTY_LOG returns an empty bitmap.
The rule is that all pages are either dirty in the current bitmap,
or write-protected, which is violated here.
It seems that just moving kvm_mmu_slot_remove_write_access down
to after the slot pointer swap should fix this bug.
KVM-Stable-Tag.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
* 'kvm-updates/2.6.37' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (321 commits)
KVM: Drop CONFIG_DMAR dependency around kvm_iommu_map_pages
KVM: Fix signature of kvm_iommu_map_pages stub
KVM: MCE: Send SRAR SIGBUS directly
KVM: MCE: Add MCG_SER_P into KVM_MCE_CAP_SUPPORTED
KVM: fix typo in copyright notice
KVM: Disable interrupts around get_kernel_ns()
KVM: MMU: Avoid sign extension in mmu_alloc_direct_roots() pae root address
KVM: MMU: move access code parsing to FNAME(walk_addr) function
KVM: MMU: audit: check whether have unsync sps after root sync
KVM: MMU: audit: introduce audit_printk to cleanup audit code
KVM: MMU: audit: unregister audit tracepoints before module unloaded
KVM: MMU: audit: fix vcpu's spte walking
KVM: MMU: set access bit for direct mapping
KVM: MMU: cleanup for error mask set while walk guest page table
KVM: MMU: update 'root_hpa' out of loop in PAE shadow path
KVM: x86 emulator: Eliminate compilation warning in x86_decode_insn()
KVM: x86: Fix constant type in kvm_get_time_scale
KVM: VMX: Add AX to list of registers clobbered by guest switch
KVM guest: Move a printk that's using the clock before it's ready
KVM: x86: TSC catchup mode
...
Now we have MCG_SER_P (and corresponding SRAO/SRAR MCE) support in
kernel and QEMU-KVM, the MCG_SER_P should be added into
KVM_MCE_CAP_SUPPORTED to make all these code really works.
Reported-by: Dean Nelson <dnelson@redhat.com>
Signed-off-by: Huang Ying <ying.huang@intel.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
get_kernel_ns() wants preemption disabled. It doesn't make a lot of sense
during the get/set ioctls (no way to make them non-racy) but the callee wants
it.
Signed-off-by: Avi Kivity <avi@redhat.com>
Older gcc versions complain about the improper type (for x86-32), 4.5
seems to fix this silently. However, we should better use the right type
initially.
Signed-off-by: Jan Kiszka <jan.kiszka@siemens.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Negate the effects of AN TYM spell while kvm thread is preempted by tracking
conversion factor to the highest TSC rate and catching the TSC up when it has
fallen behind the kernel view of time. Note that once triggered, we don't
turn off catchup mode.
A slightly more clever version of this is possible, which only does catchup
when TSC rate drops, and which specifically targets only CPUs with broken
TSC, but since these all are considered unstable_tsc(), this patch covers
all necessary cases.
Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
This just changes some names to better reflect the usage they
will be given. Separated out to keep confusion to a minimum.
Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
The math in kvm_get_time_scale relies on the fact that
NSEC_PER_SEC < 2^32. To use the same function to compute
arbitrary time scales, we must extend the first reduction
step to shrink the base rate to a 32-bit value, and
possibly reduce the scaled rate into a 32-bit as well.
Note we must take care to avoid an arithmetic overflow
when scaling up the tps32 value (this could not happen
with the fixed scaled value of NSEC_PER_SEC, but can
happen with scaled rates above 2^31.
Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
This adds a wrapper function kvm_inject_realmode_interrupt() around the
emulator function emulate_int_real() to allow real mode interrupt injection.
[avi: initialize operand and address sizes before emulating interrupts]
[avi: initialize rip for real mode interrupt injection]
[avi: clear interrupt pending flag after emulating interrupt injection]
Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
The PIC code used to be called from preempt_disable() context, which
wasn't very good for PREEMPT_RT. That is no longer the case, so move
back from raw_spinlock_t to spinlock_t.
Signed-off-by: Avi Kivity <avi@redhat.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
If preempted after kvmclock values are updated, but before hardware
virtualization is entered, the last tsc time as read by the guest is
never set. It underflows the next time kvmclock is updated if there
has not yet been a successful entry / exit into hardware virt.
Fix this by simply setting last_tsc to the newly read tsc value so
that any computed nsec advance of kvmclock is nulled.
Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
This patch moves the detection whether a page-fault was
nested or not out of the error code and moves it into a
separate variable in the fault struct.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Change the interrupt injection code to work from preemptible, interrupts
enabled context. This works by adding a ->cancel_injection() operation
that undoes an injection in case we were not able to actually enter the guest
(this condition could never happen with atomic injection).
Signed-off-by: Avi Kivity <avi@redhat.com>
Instead of blindly attempting to inject an event before each guest entry,
check for a possible event first in vcpu->requests. Sites that can trigger
event injection are modified to set KVM_REQ_EVENT:
- interrupt, nmi window opening
- ppr updates
- i8259 output changes
- local apic irr changes
- rflags updates
- gif flag set
- event set on exit
This improves non-injecting entry performance, and sets the stage for
non-atomic injection.
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch fixes a bug in KVM where it _always_ reports the
support of the SVM feature to userspace. But KVM only
supports SVM on AMD hardware and only when it is enabled in
the kernel module. This patch fixes the wrong reporting.
Cc: stable@kernel.org
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This function need to be able to load the pdptrs from any
mmu context currently in use. So change this function to
take an kvm_mmu parameter to fit these needs.
As a side effect this patch also moves the cached pdptrs
from vcpu_arch into the kvm_mmu struct.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
KVM currently ignores fetch faults in the instruction
emulator. With nested-npt we could have such faults. This
patch adds the code to handle these.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch implements logic to make sure that either a
page-fault/page-fault-vmexit or a nested-page-fault-vmexit
is propagated back to the guest.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch introduces the init_kvm_nested_mmu() function
which is used to re-initialize the nested mmu when the l2
guest changes its paging mode.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch introduces the kvm_read_guest_page_x86 function
which reads from the physical memory of the guest. If the
guest is running in guest-mode itself with nested paging
enabled it will read from the guest's guest physical memory
instead.
The patch also changes changes the code to use this function
where it is necessary.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch adds a function which can read from the guests
physical memory or from the guest's guest physical memory.
This will be used in the two-dimensional page table walker.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch introduces the walk_mmu pointer which points to
the mmu-context currently used for gva_to_gpa translations.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch introduces a mmu-callback to translate gpa
addresses in the walk_addr code. This is later used to
translate l2_gpa addresses into l1_gpa addresses.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
This patch introduces a struct with two new fields in
vcpu_arch for x86:
* fault.address
* fault.error_code
This will be used to correctly propagate page faults back
into the guest when we could have either an ordinary page
fault or a nested page fault. In the case of a nested page
fault the fault-address is different from the original
address that should be walked. So we need to keep track
about the real fault-address.
Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
Some operating systems store data about the host processor at the
time of installation, and when booted on a more uptodate cpu tries
to read MSR_EBC_FREQUENCY_ID. This has been found with XP.
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
MSR_K7_CLK_CTL is a no longer documented MSR, which is only relevant
on said old AMD CPU models. This change returns the expected value,
which the Linux kernel is expecting to avoid writing back the MSR,
plus it ignores all writes to the MSR.
Signed-off-by: Jes Sorensen <Jes.Sorensen@redhat.com>
Signed-off-by: Avi Kivity <avi@redhat.com>
x86_emulate_insn() will return 1 if instruction can be restarted
without re-entering a guest.
Signed-off-by: Gleb Natapov <gleb@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
Kernel time, which advances in discrete steps may progress much slower
than TSC. As a result, when kvmclock is adjusted to a new base, the
apparent time to the guest, which runs at a much higher, nsec scaled
rate based on the current TSC, may have already been observed to have
a larger value (kernel_ns + scaled tsc) than the value to which we are
setting it (kernel_ns + 0).
We must instead compute the clock as potentially observed by the guest
for kernel_ns to make sure it does not go backwards.
Signed-off-by: Zachary Amsden <zamsden@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>