An innocent reader of the following x86 KVM code:
bool kvm_arch_can_inject_async_page_present(struct kvm_vcpu *vcpu)
{
if (!(vcpu->arch.apf.msr_val & KVM_ASYNC_PF_ENABLED))
return true;
...
may get very confused: if APF mechanism is not enabled, why do we report
that we 'can inject async page present'? In reality, upon injection
kvm_arch_async_page_present() will check the same condition again and,
in case APF is disabled, will just drop the item. This is fine as the
guest which deliberately disabled APF doesn't expect to get any APF
notifications.
Rename kvm_arch_can_inject_async_page_present() to
kvm_arch_can_dequeue_async_page_present() to make it clear what we are
checking: if the item can be dequeued (meaning either injected or just
dropped).
On s390 kvm_arch_can_inject_async_page_present() always returns 'true' so
the rename doesn't matter much.
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20200525144125.143875-4-vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This reverts commit 5b494aea13.
If unlocked==true then the vma pointer could be invalidated, so the 2nd
follow_pfn() is potentially racy: we do need to get out and redo
find_vma_intersection().
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Now that the 32bit KVM/arm host is a distant memory, let's move the
whole of the KVM/arm64 code into the arm64 tree.
As they said in the song: Welcome Home (Sanitarium).
Signed-off-by: Marc Zyngier <maz@kernel.org>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20200513104034.74741-1-maz@kernel.org
Two new stats for exposing halt-polling cpu usage:
halt_poll_success_ns
halt_poll_fail_ns
Thus sum of these 2 stats is the total cpu time spent polling. "success"
means the VCPU polled until a virtual interrupt was delivered. "fail"
means the VCPU had to schedule out (either because the maximum poll time
was reached or it needed to yield the CPU).
To avoid touching every arch's kvm_vcpu_stat struct, only update and
export halt-polling cpu usage stats if we're on x86.
Exporting cpu usage as a u64 and in nanoseconds means we will overflow at
~500 years, which seems reasonably large.
Signed-off-by: David Matlack <dmatlack@google.com>
Signed-off-by: Jon Cargille <jcargill@google.com>
Reviewed-by: Jim Mattson <jmattson@google.com>
Message-Id: <20200508182240.68440-1-jcargill@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
While optimizing posted-interrupt delivery especially for the timer
fastpath scenario, I measured kvm_x86_ops.deliver_posted_interrupt()
to introduce substantial latency because the processor has to perform
all vmentry tasks, ack the posted interrupt notification vector,
read the posted-interrupt descriptor etc.
This is not only slow, it is also unnecessary when delivering an
interrupt to the current CPU (as is the case for the LAPIC timer) because
PIR->IRR and IRR->RVI synchronization is already performed on vmentry
Therefore skip kvm_vcpu_trigger_posted_interrupt in this case, and
instead do vmx_sync_pir_to_irr() on the EXIT_FASTPATH_REENTER_GUEST
fastpath as well.
Tested-by: Haiwei Li <lihaiwei@tencent.com>
Cc: Haiwei Li <lihaiwei@tencent.com>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
Message-Id: <1588055009-12677-6-git-send-email-wanpengli@tencent.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
hva_to_pfn_remapped() calls fixup_user_fault(), which has already
handled the retry gracefully. Even if "unlocked" is set to true, it
means that we've got a VM_FAULT_RETRY inside fixup_user_fault(),
however the page fault has already retried and we should have the pfn
set correctly. No need to do that again.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20200416155906.267462-1-peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The '==' expression itself is bool, no need to convert it to bool again.
This fixes the following coccicheck warning:
virt/kvm/eventfd.c:724:38-43: WARNING: conversion to bool not needed
here
Signed-off-by: Jason Yan <yanaijie@huawei.com>
Message-Id: <20200420123805.4494-1-yanaijie@huawei.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The use of any sort of waitqueue (simple or regular) for
wait/waking vcpus has always been an overkill and semantically
wrong. Because this is per-vcpu (which is blocked) there is
only ever a single waiting vcpu, thus no need for any sort of
queue.
As such, make use of the rcuwait primitive, with the following
considerations:
- rcuwait already provides the proper barriers that serialize
concurrent waiter and waker.
- Task wakeup is done in rcu read critical region, with a
stable task pointer.
- Because there is no concurrency among waiters, we need
not worry about rcuwait_wait_event() calls corrupting
the wait->task. As a consequence, this saves the locking
done in swait when modifying the queue. This also applies
to per-vcore wait for powerpc kvm-hv.
The x86 tscdeadline_latency test mentioned in 8577370fb0
("KVM: Use simple waitqueue for vcpu->wq") shows that, on avg,
latency is reduced by around 15-20% with this change.
Cc: Paul Mackerras <paulus@ozlabs.org>
Cc: kvmarm@lists.cs.columbia.edu
Cc: linux-mips@vger.kernel.org
Reviewed-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Message-Id: <20200424054837.5138-6-dave@stgolabs.net>
[Avoid extra logic changes. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This allows making request to all other vcpus except the one
specified in the parameter.
Signed-off-by: Suravee Suthikulpanit <suravee.suthikulpanit@amd.com>
Message-Id: <1588771076-73790-2-git-send-email-suravee.suthikulpanit@amd.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In the unlikely event that a 32bit vcpu traps into the hypervisor
on an instruction that is located right at the end of the 32bit
range, the emulation of that instruction is going to increment
PC past the 32bit range. This isn't great, as userspace can then
observe this value and get a bit confused.
Conversly, userspace can do things like (in the context of a 64bit
guest that is capable of 32bit EL0) setting PSTATE to AArch64-EL0,
set PC to a 64bit value, change PSTATE to AArch32-USR, and observe
that PC hasn't been truncated. More confusion.
Fix both by:
- truncating PC increments for 32bit guests
- sanitizing all 32bit regs every time a core reg is changed by
userspace, and that PSTATE indicates a 32bit mode.
Cc: stable@vger.kernel.org
Acked-by: Will Deacon <will@kernel.org>
Signed-off-by: Marc Zyngier <maz@kernel.org>
KVM now expects to be able to use HW-accelerated delivery of vSGIs
as soon as the guest has enabled thm. Unfortunately, we only
initialize the GICv4 context if we have a virtual ITS exposed to
the guest.
Fix it by always initializing the GICv4.1 context if it is
available on the host.
Fixes: 2291ff2f2a ("KVM: arm64: GICv4.1: Plumb SGI implementation selection in the distributor")
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
KVM_CAP_HALT_POLL is a per-VM capability that lets userspace
control the halt-polling time, allowing halt-polling to be tuned or
disabled on particular VMs.
With dynamic halt-polling, a VM's VCPUs can poll from anywhere from
[0, halt_poll_ns] on each halt. KVM_CAP_HALT_POLL sets the
upper limit on the poll time.
Signed-off-by: David Matlack <dmatlack@google.com>
Signed-off-by: Jon Cargille <jcargill@google.com>
Reviewed-by: Jim Mattson <jmattson@google.com>
Message-Id: <20200417221446.108733-1-jcargill@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
If we're going to fail out the vgic_add_lpi(), let's make sure the
allocated vgic_irq memory is also freed. Though it seems that both
cases are unlikely to fail.
Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200414030349.625-3-yuzenghui@huawei.com
It's likely that the vcpu fails to handle all virtual interrupts if
userspace decides to destroy it, leaving the pending ones stay in the
ap_list. If the un-handled one is a LPI, its vgic_irq structure will
be eventually leaked because of an extra refcount increment in
vgic_queue_irq_unlock().
This was detected by kmemleak on almost every guest destroy, the
backtrace is as follows:
unreferenced object 0xffff80725aed5500 (size 128):
comm "CPU 5/KVM", pid 40711, jiffies 4298024754 (age 166366.512s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 08 01 a9 73 6d 80 ff ff ...........sm...
c8 61 ee a9 00 20 ff ff 28 1e 55 81 6c 80 ff ff .a... ..(.U.l...
backtrace:
[<000000004bcaa122>] kmem_cache_alloc_trace+0x2dc/0x418
[<0000000069c7dabb>] vgic_add_lpi+0x88/0x418
[<00000000bfefd5c5>] vgic_its_cmd_handle_mapi+0x4dc/0x588
[<00000000cf993975>] vgic_its_process_commands.part.5+0x484/0x1198
[<000000004bd3f8e3>] vgic_its_process_commands+0x50/0x80
[<00000000b9a65b2b>] vgic_mmio_write_its_cwriter+0xac/0x108
[<0000000009641ebb>] dispatch_mmio_write+0xd0/0x188
[<000000008f79d288>] __kvm_io_bus_write+0x134/0x240
[<00000000882f39ac>] kvm_io_bus_write+0xe0/0x150
[<0000000078197602>] io_mem_abort+0x484/0x7b8
[<0000000060954e3c>] kvm_handle_guest_abort+0x4cc/0xa58
[<00000000e0d0cd65>] handle_exit+0x24c/0x770
[<00000000b44a7fad>] kvm_arch_vcpu_ioctl_run+0x460/0x1988
[<0000000025fb897c>] kvm_vcpu_ioctl+0x4f8/0xee0
[<000000003271e317>] do_vfs_ioctl+0x160/0xcd8
[<00000000e7f39607>] ksys_ioctl+0x98/0xd8
Fix it by retiring all pending LPIs in the ap_list on the destroy path.
p.s. I can also reproduce it on a normal guest shutdown. It is because
userspace still send LPIs to vcpu (through KVM_SIGNAL_MSI ioctl) while
the guest is being shutdown and unable to handle it. A little strange
though and haven't dig further...
Reviewed-by: James Morse <james.morse@arm.com>
Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
[maz: moved the distributor deallocation down to avoid an UAF splat]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200414030349.625-2-yuzenghui@huawei.com
There is no point in accessing the HW when writing to any of the
ISPENDR/ICPENDR registers from userspace, as only the guest should
be allowed to change the HW state.
Introduce new userspace-specific accessors that deal solely with
the virtual state. Note that the API differs from that of GICv3,
where userspace exclusively uses ISPENDR to set the state. Too
bad we can't reuse it.
Fixes: 82e40f558d ("KVM: arm/arm64: vgic-v2: Handle SGI bits in GICD_I{S,C}PENDR0 as WI")
Reviewed-by: James Morse <james.morse@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
There is no point in accessing the HW when writing to any of the
ISENABLER/ICENABLER registers from userspace, as only the guest
should be allowed to change the HW state.
Introduce new userspace-specific accessors that deal solely with
the virtual state.
Reported-by: James Morse <james.morse@arm.com>
Tested-by: James Morse <james.morse@arm.com>
Reviewed-by: James Morse <james.morse@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
When a guest tries to read the active state of its interrupts,
we currently just return whatever state we have in memory. This
means that if such an interrupt lives in a List Register on another
CPU, we fail to obsertve the latest active state for this interrupt.
In order to remedy this, stop all the other vcpus so that they exit
and we can observe the most recent value for the state. This is
similar to what we are doing for the write side of the same
registers, and results in new MMIO handlers for userspace (which
do not need to stop the guest, as it is supposed to be stopped
already).
Reported-by: Julien Grall <julien@xen.org>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
When a nested page fault is taken from an address that does not have
a memslot associated to it, kvm_mmu_do_page_fault returns RET_PF_EMULATE
(via mmu_set_spte) and kvm_mmu_page_fault then invokes svm_need_emulation_on_page_fault.
The default answer there is to return false, but in this case this just
causes the page fault to be retried ad libitum. Since this is not a
fast path, and the only other case where it is taken is an erratum,
just stick a kvm_vcpu_gfn_to_memslot check in there to detect the
common case where the erratum is not happening.
This fixes an infinite loop in the new set_memory_region_test.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In earlier versions of kvm, 'kvm_run' was an independent structure
and was not included in the vcpu structure. At present, 'kvm_run'
is already included in the vcpu structure, so the parameter
'kvm_run' is redundant.
This patch simplifies the function definition, removes the extra
'kvm_run' parameter, and extracts it from the 'kvm_vcpu' structure
if necessary.
Signed-off-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Message-Id: <20200416051057.26526-1-tianjia.zhang@linux.alibaba.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Create a new function kvm_is_visible_memslot() and use it from
kvm_is_visible_gfn(); use the new function in try_async_pf() too,
to avoid an extra memslot lookup.
Opportunistically squish a multi-line comment into a single-line comment.
Note, the end result, KVM_PFN_NOSLOT, is unchanged.
Cc: Jim Mattson <jmattson@google.com>
Cc: Rick Edgecombe <rick.p.edgecombe@intel.com>
Suggested-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The placement of kvm_create_vcpu_debugfs is more or less irrelevant, since
it cannot fail and userspace should not care about the debugfs entries until
it knows the vcpu has been created. Moving it after the last failure
point removes the need to remove the directory when unwinding the creation.
Reviewed-by: Emanuele Giuseppe Esposito <eesposit@redhat.com>
Message-Id: <20200331224222.393439-1-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Implementing (and even advertising) 64bit PSCI functions to 32bit
guests is at least a bit odd, if not altogether violating the
spec which says ("5.2.1 Register usage in arguments and return values"):
"Adherence to the SMC Calling Conventions implies that any AArch32
caller of an SMC64 function will get a return code of 0xFFFFFFFF(int32).
This matches the NOT_SUPPORTED error code used in PSCI"
Tighten the implementation by pretending these functions are not
there for 32bit guests.
Reviewed-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
When a guest delibarately uses an SMC32 function number (which is allowed),
we should make sure we drop the top 32bits from the input arguments, as they
could legitimately be junk.
Reported-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
The variable r is being assigned with a value that is never read
and it is being updated later with a new value. The initialization is
redundant and can be removed.
Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Message-Id: <20200410113526.13822-1-colin.king@canonical.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When deciding whether a guest has to be stopped we check whether this
is a private interrupt or not. Unfortunately, there's an off-by-one bug
here, and we fail to recognize a whole range of interrupts as being
global (GICv2 SPIs 32-63).
Fix the condition from > to be >=.
Cc: stable@vger.kernel.org
Fixes: abd7229626 ("KVM: arm/arm64: Simplify active_change_prepare and plug race")
Reported-by: André Przywara <andre.przywara@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Pass @opaque to kvm_arch_hardware_setup() and
kvm_arch_check_processor_compat() to allow architecture specific code to
reference @opaque without having to stash it away in a temporary global
variable. This will enable x86 to separate its vendor specific callback
ops, which are passed via @opaque, into "init" and "runtime" ops without
having to stash away the "init" ops.
No functional change intended.
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Tested-by: Cornelia Huck <cohuck@redhat.com> #s390
Acked-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200321202603.19355-2-sean.j.christopherson@intel.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reset the LRU slot if it becomes invalid when deleting a memslot to fix
an out-of-bounds/use-after-free access when searching through memslots.
Explicitly check for there being no used slots in search_memslots(), and
in the caller of s390's approximation variant.
Fixes: 36947254e5 ("KVM: Dynamically size memslot array based on number of used slots")
Reported-by: Qian Cai <cai@lca.pw>
Cc: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320205546.2396-2-sean.j.christopherson@intel.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The vgic-state debugfs file could do with showing the pending state
of the HW-backed SGIs. Plug it into the low-level code.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-24-maz@kernel.org
Each time a Group-enable bit gets flipped, the state of these bits
needs to be forwarded to the hardware. This is a pretty heavy
handed operation, requiring all vcpus to reload their GICv4
configuration. It is thus implemented as a new request type.
These enable bits are programmed into the HW by setting the VGrp{0,1}En
fields of GICR_VPENDBASER when the vPEs are made resident again.
Of course, we only support Group-1 for now...
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-22-maz@kernel.org
The GICv4.1 architecture gives the hypervisor the option to let
the guest choose whether it wants the good old SGIs with an
active state, or the new, HW-based ones that do not have one.
For this, plumb the configuration of SGIs into the GICv3 MMIO
handling, present the GICD_TYPER2.nASSGIcap to the guest,
and handle the GICD_CTLR.nASSGIreq setting.
In order to be able to deal with the restore of a guest, also
apply the GICD_CTLR.nASSGIreq setting at first run so that we
can move the restored SGIs to the HW if that's what the guest
had selected in a previous life.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-21-maz@kernel.org
In order to let a guest buy in the new, active-less SGIs, we
need to be able to switch between the two modes.
Handle this by stopping all guest activity, transfer the state
from one mode to the other, and resume the guest. Nothing calls
this code so far, but a later patch will plug it into the MMIO
emulation.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-20-maz@kernel.org
Most of the GICv3 emulation code that deals with SGIs now has to be
aware of the v4.1 capabilities in order to benefit from it.
Add such support, keyed on the interrupt having the hw flag set and
being a SGI.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-19-maz@kernel.org
As GICv4.1 understands the life cycle of doorbells (instead of
just randomly firing them at the most inconvenient time), just
enable them at irq_request time, and be done with it.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Link: https://lore.kernel.org/r/20200304203330.4967-18-maz@kernel.org
In order to hide some of the differences between v4.0 and v4.1, move
the doorbell management out of the KVM code, and into the GICv4-specific
layer. This allows the calling code to ask for the doorbell when blocking,
and otherwise to leave the doorbell permanently disabled.
This matches the v4.1 code perfectly, and only results in a minor
refactoring of the v4.0 code.
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Link: https://lore.kernel.org/r/20200304203330.4967-14-maz@kernel.org
Drop largepages_enabled, kvm_largepages_enabled() and
kvm_disable_largepages() now that all users are gone.
Note, largepages_enabled was an x86-only flag that got left in common
KVM code when KVM gained support for multiple architectures.
No functional change intended.
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It could take kvm->mmu_lock for an extended period of time when
enabling dirty log for the first time. The main cost is to clear
all the D-bits of last level SPTEs. This situation can benefit from
manual dirty log protect as well, which can reduce the mmu_lock
time taken. The sequence is like this:
1. Initialize all the bits of the dirty bitmap to 1 when enabling
dirty log for the first time
2. Only write protect the huge pages
3. KVM_GET_DIRTY_LOG returns the dirty bitmap info
4. KVM_CLEAR_DIRTY_LOG will clear D-bit for each of the leaf level
SPTEs gradually in small chunks
Under the Intel(R) Xeon(R) Gold 6152 CPU @ 2.10GHz environment,
I did some tests with a 128G windows VM and counted the time taken
of memory_global_dirty_log_start, here is the numbers:
VM Size Before After optimization
128G 460ms 10ms
Signed-off-by: Jay Zhou <jianjay.zhou@huawei.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove includes of asm/kvm_host.h from files that already include
linux/kvm_host.h to make it more obvious that there is no ordering issue
between the two headers. linux/kvm_host.h includes asm/kvm_host.h to
pick up architecture specific settings, and this will never change, i.e.
including asm/kvm_host.h after linux/kvm_host.h may seem problematic,
but in practice is simply redundant.
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Now that the memslot logic doesn't assume memslots are always non-NULL,
dynamically size the array of memslots instead of unconditionally
allocating memory for the maximum number of memslots.
Note, because a to-be-deleted memslot must first be invalidated, the
array size cannot be immediately reduced when deleting a memslot.
However, consecutive deletions will realize the memory savings, i.e.
a second deletion will trim the entry.
Tested-by: Christoffer Dall <christoffer.dall@arm.com>
Tested-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Refactor memslot handling to treat the number of used slots as the de
facto size of the memslot array, e.g. return NULL from id_to_memslot()
when an invalid index is provided instead of relying on npages==0 to
detect an invalid memslot. Rework the sorting and walking of memslots
in advance of dynamically sizing memslots to aid bisection and debug,
e.g. with luck, a bug in the refactoring will bisect here and/or hit a
WARN instead of randomly corrupting memory.
Alternatively, a global null/invalid memslot could be returned, i.e. so
callers of id_to_memslot() don't have to explicitly check for a NULL
memslot, but that approach runs the risk of introducing difficult-to-
debug issues, e.g. if the global null slot is modified. Constifying
the return from id_to_memslot() to combat such issues is possible, but
would require a massive refactoring of arch specific code and would
still be susceptible to casting shenanigans.
Add function comments to update_memslots() and search_memslots() to
explicitly (and loudly) state how memslots are sorted.
Opportunistically stuff @hva with a non-canonical value when deleting a
private memslot on x86 to detect bogus usage of the freed slot.
No functional change intended.
Tested-by: Christoffer Dall <christoffer.dall@arm.com>
Tested-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Rework kvm_get_dirty_log() so that it "returns" the associated memslot
on success. A future patch will rework memslot handling such that
id_to_memslot() can return NULL, returning the memslot makes it more
obvious that the validity of the memslot has been verified, i.e.
precludes the need to add validity checks in the arch code that are
technically unnecessary.
To maintain ordering in s390, move the call to kvm_arch_sync_dirty_log()
from s390's kvm_vm_ioctl_get_dirty_log() to the new kvm_get_dirty_log().
This is a nop for PPC, the only other arch that doesn't select
KVM_GENERIC_DIRTYLOG_READ_PROTECT, as its sync_dirty_log() is empty.
Ideally, moving the sync_dirty_log() call would be done in a separate
patch, but it can't be done in a follow-on patch because that would
temporarily break s390's ordering. Making the move in a preparatory
patch would be functionally correct, but would create an odd scenario
where the moved sync_dirty_log() would operate on a "different" memslot
due to consuming the result of a different id_to_memslot(). The
memslot couldn't actually be different as slots_lock is held, but the
code is confusing enough as it is, i.e. moving sync_dirty_log() in this
patch is the lesser of all evils.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Move the implementations of KVM_GET_DIRTY_LOG and KVM_CLEAR_DIRTY_LOG
for CONFIG_KVM_GENERIC_DIRTYLOG_READ_PROTECT into common KVM code.
The arch specific implemenations are extremely similar, differing
only in whether the dirty log needs to be sync'd from hardware (x86)
and how the TLBs are flushed. Add new arch hooks to handle sync
and TLB flush; the sync will also be used for non-generic dirty log
support in a future patch (s390).
The ulterior motive for providing a common implementation is to
eliminate the dependency between arch and common code with respect to
the memslot referenced by the dirty log, i.e. to make it obvious in the
code that the validity of the memslot is guaranteed, as a future patch
will rework memslot handling such that id_to_memslot() can return NULL.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Clean up __kvm_set_memory_region() to achieve several goals:
- Remove local variables that serve no real purpose
- Improve the readability of the code
- Better show the relationship between the 'old' and 'new' memslot
- Prepare for dynamically sizing memslots
- Document subtle gotchas (via comments)
Note, using 'tmp' to hold the initial memslot is not strictly necessary
at this juncture, e.g. 'old' could be directly copied from
id_to_memslot(), but keep the pointer usage as id_to_memslot() will be
able to return a NULL pointer once memslots are dynamically sized.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Now that all callers of kvm_free_memslot() pass NULL for @dont, remove
the param from the top-level routine and all arch's implementations.
No functional change intended.
Tested-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Move memslot deletion into its own routine so that the success path for
other memslot updates does not need to use kvm_free_memslot(), i.e. can
explicitly destroy the dirty bitmap when necessary. This paves the way
for dropping @dont from kvm_free_memslot(), i.e. all callers now pass
NULL for @dont.
Add a comment above the code to make a copy of the existing memslot
prior to deletion, it is not at all obvious that the pointer will become
stale during sorting and/or installation of new memslots.
Note, kvm_arch_commit_memory_region() allows an architecture to free
resources when moving a memslot or changing its flags, e.g. x86 frees
its arch specific memslot metadata during commit_memory_region().
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Tested-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Drop the "const" attribute from @old in kvm_arch_commit_memory_region()
to allow arch specific code to free arch specific resources in the old
memslot without having to cast away the attribute. Freeing resources in
kvm_arch_commit_memory_region() paves the way for simplifying
kvm_free_memslot() by eliminating the last usage of its @dont param.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Split out the core functionality of setting a memslot into a separate
helper in preparation for moving memslot deletion into its own routine.
Tested-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Replace a big pile o' gotos with returns to make it more obvious what
error code is being returned, and to prepare for refactoring the
functional, i.e. post-checks, portion of __kvm_set_memory_region().
Reviewed-by: Janosch Frank <frankja@linux.ibm.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Explicitly free an allocated-but-unused dirty bitmap instead of relying
on kvm_free_memslot() if an error occurs in __kvm_set_memory_region().
There is no longer a need to abuse kvm_free_memslot() to free arch
specific resources as arch specific code is now called only after the
common flow is guaranteed to succeed. Arch code can still fail, but
it's responsible for its own cleanup in that case.
Eliminating the error path's abuse of kvm_free_memslot() paves the way
for simplifying kvm_free_memslot(), i.e. dropping its @dont param.
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove kvm_arch_create_memslot() now that all arch implementations are
effectively nops. Removing kvm_arch_create_memslot() eliminates the
possibility for arch specific code to allocate memory prior to setting
a memslot, which sets the stage for simplifying kvm_free_memslot().
Cc: Janosch Frank <frankja@linux.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The two implementations of kvm_arch_create_memslot() in x86 and PPC are
both good citizens and free up all local resources if creation fails.
Return immediately (via a superfluous goto) instead of calling
kvm_free_memslot().
Note, the call to kvm_free_memslot() is effectively an expensive nop in
this case as there are no resources to be freed.
No functional change intended.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reinstall the old memslots if preparing the new memory region fails
after invalidating a to-be-{re}moved memslot.
Remove the superfluous 'old_memslots' variable so that it's somewhat
clear that the error handling path needs to free the unused memslots,
not simply the 'old' memslots.
Fixes: bc6678a33d ("KVM: introduce kvm->srcu and convert kvm_set_memory_region to SRCU update")
Reviewed-by: Christoffer Dall <christoffer.dall@arm.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Use the physical timer structure when reading the physical counter
instead of using the virtual timer structure. Thankfully, nothing is
accessing this code path yet (at least not until we enable save/restore
of the physical counter). It doesn't hurt for this to be correct though.
Signed-off-by: KarimAllah Ahmed <karahmed@amazon.de>
[maz: amended commit log]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Zenghui Yu <yuzenghui@huawei.com>
Fixes: 84135d3d18 ("KVM: arm/arm64: consolidate arch timer trap handlers")
Link: https://lore.kernel.org/r/1584351546-5018-1-git-send-email-karahmed@amazon.de
- Fix compilation on 32bit
- Move VHE guest entry/exit into the VHE-specific entry code
- Make sure all functions called by the non-VHE HYP code is tagged as __always_inline
-----BEGIN PGP SIGNATURE-----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=NpPo
-----END PGP SIGNATURE-----
Merge tag 'kvmarm-fixes-5.6-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into HEAD
KVM/arm fixes for 5.6, take #1
- Fix compilation on 32bit
- Move VHE guest entry/exit into the VHE-specific entry code
- Make sure all functions called by the non-VHE HYP code is tagged as __always_inline
With VHE, running a vCPU always requires the sequence:
1. kvm_arm_vhe_guest_enter();
2. kvm_vcpu_run_vhe();
3. kvm_arm_vhe_guest_exit()
... and as we invoke this from the shared arm/arm64 KVM code, 32-bit arm
has to provide stubs for all three functions.
To simplify the common code, and make it easier to make further
modifications to the arm64-specific portions in the near future, let's
fold kvm_arm_vhe_guest_enter() and kvm_arm_vhe_guest_exit() into
kvm_vcpu_run_vhe().
The 32-bit stubs for kvm_arm_vhe_guest_enter() and
kvm_arm_vhe_guest_exit() are removed, as they are no longer used. The
32-bit stub for kvm_vcpu_run_vhe() is left as-is.
There should be no functional change as a result of this patch.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200210114757.2889-1-mark.rutland@arm.com
Accessing a per-cpu variable only makes sense when preemption is
disabled (and the kernel does check this when the right debug options
are switched on).
For kvm_get_running_vcpu(), it is fine to return the value after
re-enabling preemption, as the preempt notifiers will make sure that
this is kept consistent across task migration (the comment above the
function hints at it, but lacks the crucial preemption management).
While we're at it, move the comment from the ARM code, which explains
why the whole thing works.
Fixes: 7495e22bb1 ("KVM: Move running VCPU from ARM to common code").
Cc: Paolo Bonzini <pbonzini@redhat.com>
Reported-by: Zenghui Yu <yuzenghui@huawei.com>
Tested-by: Zenghui Yu <yuzenghui@huawei.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/318984f6-bc36-33a3-abc6-bf2295974b06@huawei.com
Message-id: <20200207163410.31276-1-maz@kernel.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Fedora kernel builds on armv7hl began failing recently because
kvm_arm_exception_type and kvm_arm_exception_class were undeclared in
trace.h. Add the missing include.
Fixes: 0e20f5e255 ("KVM: arm/arm64: Cleanup MMIO handling")
Signed-off-by: Jeremy Cline <jcline@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200205134146.82678-1-jcline@redhat.com
From Boris Ostrovsky:
The KVM hypervisor may provide a guest with ability to defer remote TLB
flush when the remote VCPU is not running. When this feature is used,
the TLB flush will happen only when the remote VPCU is scheduled to run
again. This will avoid unnecessary (and expensive) IPIs.
Under certain circumstances, when a guest initiates such deferred action,
the hypervisor may miss the request. It is also possible that the guest
may mistakenly assume that it has already marked remote VCPU as needing
a flush when in fact that request had already been processed by the
hypervisor. In both cases this will result in an invalid translation
being present in a vCPU, potentially allowing accesses to memory locations
in that guest's address space that should not be accessible.
Note that only intra-guest memory is vulnerable.
The five patches address both of these problems:
1. The first patch makes sure the hypervisor doesn't accidentally clear
a guest's remote flush request
2. The rest of the patches prevent the race between hypervisor
acknowledging a remote flush request and guest issuing a new one.
Conflicts:
arch/x86/kvm/x86.c [move from kvm_arch_vcpu_free to kvm_arch_vcpu_destroy]
__kvm_map_gfn()'s call to gfn_to_pfn_memslot() is
* relatively expensive
* in certain cases (such as when done from atomic context) cannot be called
Stashing gfn-to-pfn mapping should help with both cases.
This is part of CVE-2019-3016.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm_vcpu_(un)map operates on gfns from any current address space.
In certain cases we want to make sure we are not mapping SMRAM
and for that we can use kvm_(un)map_gfn() that we are introducing
in this patch.
This is part of CVE-2019-3016.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
According to the ARM ARM, registers CNT{P,V}_TVAL_EL0 have bits [63:32]
RES0 [1]. When reading the register, the value is truncated to the least
significant 32 bits [2], and on writes, TimerValue is treated as a signed
32-bit integer [1, 2].
When the guest behaves correctly and writes 32-bit values, treating TVAL
as an unsigned 64 bit register works as expected. However, things start
to break down when the guest writes larger values, because
(u64)0x1_ffff_ffff = 8589934591. but (s32)0x1_ffff_ffff = -1, and the
former will cause the timer interrupt to be asserted in the future, but
the latter will cause it to be asserted now. Let's treat TVAL as a
signed 32-bit register on writes, to match the behaviour described in
the architecture, and the behaviour experimentally exhibited by the
virtual timer on a non-vhe host.
[1] Arm DDI 0487E.a, section D13.8.18
[2] Arm DDI 0487E.a, section D11.2.4
Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
[maz: replaced the read-side mask with lower_32_bits]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Fixes: 8fa7616248 ("KVM: arm/arm64: arch_timer: Fix CNTP_TVAL calculation")
Link: https://lore.kernel.org/r/20200127103652.2326-1-alexandru.elisei@arm.com
Let the code never use unsupported event counters. Change
kvm_pmu_handle_pmcr() to only reset supported counters and
kvm_pmu_vcpu_reset() to only stop supported counters.
Other actions are filtered on the supported counters in
kvm/sysregs.c
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200124142535.29386-5-eric.auger@redhat.com
At the moment a SW_INCR counter always overflows on 32-bit
boundary, independently on whether the n+1th counter is
programmed as CHAIN.
Check whether the SW_INCR counter is a 64b counter and if so,
implement the 64b logic.
Fixes: 80f393a23b ("KVM: arm/arm64: Support chained PMU counters")
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200124142535.29386-4-eric.auger@redhat.com
At the moment we update the chain bitmap on type setting. This
does not take into account the enable state of the odd register.
Let's make sure a counter is never considered as chained if
the high counter is disabled.
We recompute the chain state on enable/disable and type changes.
Also let create_perf_event() use the chain bitmap and not use
kvm_pmu_idx_has_chain_evtype().
Suggested-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200124142535.29386-3-eric.auger@redhat.com
The specification says PMSWINC increments PMEVCNTR<n>_EL1 by 1
if PMEVCNTR<n>_EL0 is enabled and configured to count SW_INCR.
For PMEVCNTR<n>_EL0 to be enabled, we need both PMCNTENSET to
be set for the corresponding event counter but we also need
the PMCR.E bit to be set.
Fixes: 7a0adc7064 ("arm64: KVM: Add access handler for PMSWINC register")
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Andrew Murray <andrew.murray@arm.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200124142535.29386-2-eric.auger@redhat.com
Avoid the "writable" check in __gfn_to_hva_many(), which will always fail
on read-only memslots due to gfn_to_hva() assuming writes. Functionally,
this allows x86 to create large mappings for read-only memslots that
are backed by HugeTLB mappings.
Note, the changelog for commit 05da45583d ("KVM: MMU: large page
support") states "If the largepage contains write-protected pages, a
large pte is not used.", but "write-protected" refers to pages that are
temporarily read-only, e.g. read-only memslots didn't even exist at the
time.
Fixes: 4d8b81abc4 ("KVM: introduce readonly memslot")
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
[Redone using kvm_vcpu_gfn_to_memslot_prot. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Use kvm_vcpu_gfn_to_hva() when retrieving the host page size so that the
correct set of memslots is used when handling x86 page faults in SMM.
Fixes: 54bf36aac5 ("KVM: x86: use vcpu-specific functions to read/write/translate GFNs")
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add a helper, is_transparent_hugepage(), to explicitly check whether a
compound page is a THP and use it when populating KVM's secondary MMU.
The explicit check fixes a bug where a remapped compound page, e.g. for
an XDP Rx socket, is mapped into a KVM guest and is mistaken for a THP,
which results in KVM incorrectly creating a huge page in its secondary
MMU.
Fixes: 936a5fe6e6 ("thp: kvm mmu transparent hugepage support")
Reported-by: syzbot+c9d1fb51ac9d0d10c39d@syzkaller.appspotmail.com
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Check the result of __kvm_gfn_to_hva_cache_init() and return immediately
instead of relying on the kvm_is_error_hva() check to detect errors so
that it's abundantly clear KVM intends to immediately bail on an error.
Note, the hva check is still mandatory to handle errors on subqeuesnt
calls with the same generation. Similarly, always return -EFAULT on
error so that multiple (bad) calls for a given generation will get the
same result, e.g. on an illegal gfn wrap, propagating the return from
__kvm_gfn_to_hva_cache_init() would cause the initial call to return
-EINVAL and subsequent calls to return -EFAULT.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Barret reported a (technically benign) bug where nr_pages_avail can be
accessed without being initialized if gfn_to_hva_many() fails.
virt/kvm/kvm_main.c:2193:13: warning: 'nr_pages_avail' may be
used uninitialized in this function [-Wmaybe-uninitialized]
Rather than simply squashing the warning by initializing nr_pages_avail,
fix the underlying issues by reworking __kvm_gfn_to_hva_cache_init() to
return immediately instead of continuing on. Now that all callers check
the result and/or bail immediately on a bad hva, there's no need to
explicitly nullify the memslot on error.
Reported-by: Barret Rhoden <brho@google.com>
Fixes: f1b9dd5eb8 ("kvm: Disallow wraparound in kvm_gfn_to_hva_cache_init")
Cc: Jim Mattson <jmattson@google.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When reading/writing using the guest/host cache, check for a bad hva
before checking for a NULL memslot, which triggers the slow path for
handing cross-page accesses. Because the memslot is nullified on error
by __kvm_gfn_to_hva_cache_init(), if the bad hva is encountered after
crossing into a new page, then the kvm_{read,write}_guest() slow path
could potentially write/access the first chunk prior to detecting the
bad hva.
Arguably, performing a partial access is semantically correct from an
architectural perspective, but that behavior is certainly not intended.
In the original implementation, memslot was not explicitly nullified
and therefore the partial access behavior varied based on whether the
memslot itself was null, or if the hva was simply bad. The current
behavior was introduced as a seemingly unintentional side effect in
commit f1b9dd5eb8 ("kvm: Disallow wraparound in
kvm_gfn_to_hva_cache_init"), which justified the change with "since some
callers don't check the return code from this function, it sit seems
prudent to clear ghc->memslot in the event of an error".
Regardless of intent, the partial access is dependent on _not_ checking
the result of the cache initialization, which is arguably a bug in its
own right, at best simply weird.
Fixes: 8f964525a1 ("KVM: Allow cross page reads and writes from cached translations.")
Cc: Jim Mattson <jmattson@google.com>
Cc: Andrew Honig <ahonig@google.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
For ring-based dirty log tracking, it will be more efficient to account
writes during schedule-out or schedule-in to the currently running VCPU.
We would like to do it even if the write doesn't use the current VCPU's
address space, as is the case for cached writes (see commit 4e335d9e7d,
"Revert "KVM: Support vCPU-based gfn->hva cache"", 2017-05-02).
Therefore, add a mechanism to track the currently-loaded kvm_vcpu struct.
There is already something similar in KVM/ARM; one important difference
is that kvm_arch_vcpu_{load,put} have two callers in virt/kvm/kvm_main.c:
we have to update both the architecture-independent vcpu_{load,put} and
the preempt notifiers.
Another change made in the process is to allow using kvm_get_running_vcpu()
in preemptible code. This is allowed because preempt notifiers ensure
that the value does not change even after the VCPU thread is migrated.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It's already going to reach 2400 Bytes (which is over half of page
size on 4K page archs), so maybe it's good to have this build-time
check in case it overflows when adding new fields.
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove kvm_read_guest_atomic() because it's not used anywhere.
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Open code the allocation and freeing of the vcpu->run page in
kvm_vm_ioctl_create_vcpu() and kvm_vcpu_destroy() respectively. Doing
so allows kvm_vcpu_init() to be a pure init function and eliminates
kvm_vcpu_uninit() entirely.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Move the putting of vcpu->pid to kvm_vcpu_destroy(). vcpu->pid is
guaranteed to be NULL when kvm_vcpu_uninit() is called in the error path
of kvm_vm_ioctl_create_vcpu(), e.g. it is explicitly nullified by
kvm_vcpu_init() and is only changed by KVM_RUN.
No functional change intended.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove kvm_arch_vcpu_init() and kvm_arch_vcpu_uninit() now that all
arch specific implementations are nops.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add an arm specific hook to free the arm64-only sve_state. Doing so
eliminates the last functional code from kvm_arch_vcpu_uninit() across
all architectures and paves the way for removing kvm_arch_vcpu_init()
and kvm_arch_vcpu_uninit() entirely.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Fold init() into create() now that the two are called back-to-back by
common KVM code (kvm_vcpu_init() calls kvm_arch_vcpu_init() as its last
action, and kvm_vm_ioctl_create_vcpu() calls kvm_arch_vcpu_create()
immediately thereafter). This paves the way for removing
kvm_arch_vcpu_{un}init() entirely.
Note, there is no associated unwinding in kvm_arch_vcpu_uninit() that
needs to be relocated (to kvm_arch_vcpu_destroy()).
No functional change intended.
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove kvm_arch_vcpu_setup() now that all arch specific implementations
are nops.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Initialize the preempt notifier immediately in kvm_vcpu_init() to pave
the way for removing kvm_arch_vcpu_setup(), i.e. to allow arch specific
code to call vcpu_load() during kvm_arch_vcpu_create().
Back when preemption support was added, the location of the call to init
the preempt notifier was perfectly sane. The overall vCPU creation flow
featured a single arch specific hook and the preempt notifer was used
immediately after its initialization (by vcpu_load()). E.g.:
vcpu = kvm_arch_ops->vcpu_create(kvm, n);
if (IS_ERR(vcpu))
return PTR_ERR(vcpu);
preempt_notifier_init(&vcpu->preempt_notifier, &kvm_preempt_ops);
vcpu_load(vcpu);
r = kvm_mmu_setup(vcpu);
vcpu_put(vcpu);
if (r < 0)
goto free_vcpu;
Today, the call to preempt_notifier_init() is sandwiched between two
arch specific calls, kvm_arch_vcpu_create() and kvm_arch_vcpu_setup(),
which needlessly forces x86 (and possibly others?) to split its vCPU
creation flow. Init the preempt notifier prior to any arch specific
call so that each arch can independently decide how best to organize
its creation flow.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Unexport kvm_vcpu_cache and kvm_vcpu_{un}init() and make them static
now that they are referenced only in kvm_main.c.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Now that all architectures tightly couple vcpu allocation/free with the
mandatory calls to kvm_{un}init_vcpu(), move the sequences verbatim to
common KVM code.
Move both allocation and initialization in a single patch to eliminate
thrash in arch specific code. The bisection benefits of moving the two
pieces in separate patches is marginal at best, whereas the odds of
introducing a transient arch specific bug are non-zero.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add kvm_vcpu_destroy() and wire up all architectures to call the common
function instead of their arch specific implementation. The common
destruction function will be used by future patches to move allocation
and initialization of vCPUs to common KVM code, i.e. to free resources
that are allocated by arch agnostic code.
No functional change intended.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add a pre-allocation arch hook to handle checks that are currently done
by arch specific code prior to allocating the vCPU object. This paves
the way for moving the allocation to common KVM code.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove the superfluous kvm_arch_vcpu_free() as it is no longer called
from commmon KVM code. Note, kvm_arch_vcpu_destroy() *is* called from
common code, i.e. choosing which function to whack is not completely
arbitrary.
Acked-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM's inject_abt64() injects an external-abort into an aarch64 guest.
The KVM_CAP_ARM_INJECT_EXT_DABT is intended to do exactly this, but
for an aarch32 guest inject_abt32() injects an implementation-defined
exception, 'Lockdown fault'.
Change this to external abort. For non-LPAE we now get the documented:
| Unhandled fault: external abort on non-linefetch (0x008) at 0x9c800f00
and for LPAE:
| Unhandled fault: synchronous external abort (0x210) at 0x9c800f00
Fixes: 74a64a9816 ("KVM: arm/arm64: Unify 32bit fault injection")
Reported-by: Beata Michalska <beata.michalska@linaro.org>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200121123356.203000-3-james.morse@arm.com
Beata reports that KVM_SET_VCPU_EVENTS doesn't inject the expected
exception to a non-LPAE aarch32 guest.
The host intends to inject DFSR.FS=0x14 "IMPLEMENTATION DEFINED fault
(Lockdown fault)", but the guest receives DFSR.FS=0x04 "Fault on
instruction cache maintenance". This fault is hooked by
do_translation_fault() since ARMv6, which goes on to silently 'handle'
the exception, and restart the faulting instruction.
It turns out, when TTBCR.EAE is clear DFSR is split, and FS[4] has
to shuffle up to DFSR[10].
As KVM only does this in one place, fix up the static values. We
now get the expected:
| Unhandled fault: lock abort (0x404) at 0x9c800f00
Fixes: 74a64a9816 ("KVM: arm/arm64: Unify 32bit fault injection")
Reported-by: Beata Michalska <beata.michalska@linaro.org>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200121123356.203000-2-james.morse@arm.com
kvm_test_age_hva() is called upon mmu_notifier_test_young(), but wrong
address range has been passed to handle_hva_to_gpa(). With the wrong
address range, no young bits will be checked in handle_hva_to_gpa().
It means zero is always returned from mmu_notifier_test_young().
This fixes the issue by passing correct address range to the underly
function handle_hva_to_gpa(), so that the hardware young (access) bit
will be visited.
Fixes: 35307b9a5f ("arm/arm64: KVM: Implement Stage-2 page aging")
Signed-off-by: Gavin Shan <gshan@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20200121055659.19560-1-gshan@redhat.com
Our MMIO handling is a bit odd, in the sense that it uses an
intermediate per-vcpu structure to store the various decoded
information that describe the access.
But the same information is readily available in the HSR/ESR_EL2
field, and we actually use this field to populate the structure.
Let's simplify the whole thing by getting rid of the superfluous
structure and save a (tiny) bit of space in the vcpu structure.
[32bit fix courtesy of Olof Johansson <olof@lixom.net>]
Signed-off-by: Marc Zyngier <maz@kernel.org>
The wrappers make it less clear that the position of the call
to kvm_arch_async_page_present depends on the architecture, and
that only one of the two call sites will actually be active.
Remove them.
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
We can store reference to kvm_stats_debugfs_item instead of copying
its values to kvm_stat_data.
This allows us to remove duplicated code and usage of temporary
kvm_stat_data inside vm_stat_get et al.
Signed-off-by: Milan Pandurov <milanpa@amazon.de>
Reviewed-by: Alexander Graf <graf@amazon.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Vitaly Kuznetsov
Paolo Bonzini
Fuad Tabba
Marc Zyngier
David Matlack
Wanpeng Li
Peter Xu
Jason Yan
Davidlohr Bueso
Suravee Suthikulpanit
Zenghui Yu
Tianjia Zhang
Colin Ian King
Sean Christopherson
Jay Zhou
KarimAllah Ahmed
Mark Rutland
Zhuang Yanying
Jeremy Cline
Boris Ostrovsky
Alexandru Elisei
Eric Auger
James Morse
Gavin Shan
Milan Pandurov