Zachary Amsden [Fri, 20 Aug 2010 08:07:30 +0000 (22:07 -1000)]
KVM: x86: Fix a possible backwards warp of kvmclock
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.
Zachary Amsden [Fri, 20 Aug 2010 08:07:28 +0000 (22:07 -1000)]
KVM: x86: Add clock sync request to hardware enable
If there are active VCPUs which are marked as belonging to
a particular hardware CPU, request a clock sync for them when
enabling hardware; the TSC could be desynchronized on a newly
arriving CPU, and we need to recompute guests system time
relative to boot after a suspend event.
This covers both cases.
Note that it is acceptable to take the spinlock, as either
no other tasks will be running and no locks held (BSP after
resume), or other tasks will be guaranteed to drop the lock
relatively quickly (AP on CPU_STARTING).
Noting we now get clock synchronization requests for VCPUs
which are starting up (or restarting), it is tempting to
attempt to remove the arch/x86/kvm/x86.c CPU hot-notifiers
at this time, however it is not correct to do so; they are
required for systems with non-constant TSC as the frequency
may not be known immediately after the processor has started
until the cpufreq driver has had a chance to run and query
the chipset.
Updated: implement better locking semantics for hardware_enable
Removed the hack of dropping and retaking the lock by adding the
semantic that we always hold kvm_lock when hardware_enable is
called. The one place that doesn't need to worry about it is
resume, as resuming a frozen CPU, the spinlock won't be taken.
Zachary Amsden [Fri, 20 Aug 2010 08:07:26 +0000 (22:07 -1000)]
KVM: x86: Robust TSC compensation
Make the match of TSC find TSC writes that are close to each other
instead of perfectly identical; this allows the compensator to also
work in migration / suspend scenarios.
Zachary Amsden [Fri, 20 Aug 2010 08:07:25 +0000 (22:07 -1000)]
KVM: x86: Add helper functions for time computation
Add a helper function to compute the kernel time and convert nanoseconds
back to CPU specific cycles. Note that these must not be called in preemptible
context, as that would mean the kernel could enter software suspend state,
which would cause non-atomic operation.
Also, convert the KVM_SET_CLOCK / KVM_GET_CLOCK ioctls to use the kernel
time helper, these should be bootbased as well.
Zachary Amsden [Fri, 20 Aug 2010 08:07:24 +0000 (22:07 -1000)]
KVM: x86: Fix deep C-state TSC desynchronization
When CPUs with unstable TSCs enter deep C-state, TSC may stop
running. This causes us to require resynchronization. Since
we can't tell when this may potentially happen, we assume the
worst by forcing re-compensation for it at every point the VCPU
task is descheduled.
Zachary Amsden [Fri, 20 Aug 2010 08:07:21 +0000 (22:07 -1000)]
KVM: x86: Make cpu_tsc_khz updates use local CPU
This simplifies much of the init code; we can now simply always
call tsc_khz_changed, optionally passing it a new value, or letting
it figure out the existing value (while interrupts are disabled, and
thus, by inference from the rule, not raceful against CPU hotplug or
frequency updates, which will issue IPIs to the local CPU to perform
this very same task).
Zachary Amsden [Fri, 20 Aug 2010 08:07:20 +0000 (22:07 -1000)]
KVM: x86: TSC reset compensation
Attempt to synchronize TSCs which are reset to the same value. In the
case of a reliable hardware TSC, we can just re-use the same offset, but
on non-reliable hardware, we can get closer by adjusting the offset to
match the elapsed time.
Zachary Amsden [Fri, 20 Aug 2010 08:07:17 +0000 (22:07 -1000)]
KVM: x86: Move TSC offset writes to common code
Also, ensure that the storing of the offset and the reading of the TSC
are never preempted by taking a spinlock. While the lock is overkill
now, it is useful later in this patch series.
Zachary Amsden [Fri, 20 Aug 2010 08:07:16 +0000 (22:07 -1000)]
KVM: x86: Convert TSC writes to TSC offset writes
Change svm / vmx to be the same internally and write TSC offset
instead of bare TSC in helper functions. Isolated as a single
patch to contain code movement.
Zachary Amsden [Fri, 20 Aug 2010 08:07:15 +0000 (22:07 -1000)]
KVM: x86: Drop vm_init_tsc
This is used only by the VMX code, and is not done properly;
if the TSC is indeed backwards, it is out of sync, and will
need proper handling in the logic at each and every CPU change.
For now, drop this test during init as misguided.
Wei Yongjun [Mon, 23 Aug 2010 08:13:15 +0000 (16:13 +0800)]
KVM: MMU: fix missing percpu counter destroy
commit ad05c88266b4cce1c820928ce8a0fb7690912ba1
(KVM: create aggregate kvm_total_used_mmu_pages value)
introduce percpu counter kvm_total_used_mmu_pages but never
destroy it, this may cause oops when rmmod & modprobe.
Xiaotian Feng [Tue, 24 Aug 2010 02:31:07 +0000 (10:31 +0800)]
KVM: MMU: fix regression from rework mmu_shrink() code
Latest kvm mmu_shrink code rework makes kernel changes kvm->arch.n_used_mmu_pages/
kvm->arch.n_max_mmu_pages at kvm_mmu_free_page/kvm_mmu_alloc_page, which is called
by kvm_mmu_commit_zap_page. So the kvm->arch.n_used_mmu_pages or
kvm_mmu_available_pages(vcpu->kvm) is unchanged after kvm_mmu_prepare_zap_page(),
This caused kvm_mmu_change_mmu_pages/__kvm_mmu_free_some_pages loops forever.
Moving kvm_mmu_commit_zap_page would make the while loop performs as normal.
Reported-by: Avi Kivity <avi@redhat.com> Signed-off-by: Xiaotian Feng <dfeng@redhat.com> Tested-by: Avi Kivity <avi@redhat.com> Cc: Marcelo Tosatti <mtosatti@redhat.com> Cc: Dave Hansen <dave@linux.vnet.ibm.com> Cc: Tim Pepper <lnxninja@linux.vnet.ibm.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
So, if you do anything stupid in your shrinker, the VM will doubly
punish you.
The mmu_shrink() function takes the global kvm_lock, then acquires
every VM's kvm->mmu_lock in sequence. If we have 100 VMs, then
we're going to take 101 locks. We do it twice, so each call takes
202 locks. If we're under memory pressure, we can have each cpu
trying to do this. It can get really hairy, and we've seen lock
spinning in mmu_shrink() be the dominant entry in profiles.
This is guaranteed to optimize at least half of those lock
aquisitions away. It removes the need to take any of the locks
when simply trying to count objects.
A 'percpu_counter' can be a large object, but we only have one
of these for the entire system. There are not any better
alternatives at the moment, especially ones that handle CPU
hotplug.
Signed-off-by: Dave Hansen <dave@linux.vnet.ibm.com> Signed-off-by: Tim Pepper <lnxninja@linux.vnet.ibm.com> Signed-off-by: Avi Kivity <avi@redhat.com>
Dave Hansen [Fri, 20 Aug 2010 01:11:28 +0000 (18:11 -0700)]
KVM: replace x86 kvm n_free_mmu_pages with n_used_mmu_pages
Doing this makes the code much more readable. That's
borne out by the fact that this patch removes code. "used"
also happens to be the number that we need to return back to
the slab code when our shrinker gets called. Keeping this
value as opposed to free makes the next patch simpler.
So, 'struct kvm' is kzalloc()'d. 'struct kvm_arch' is a
structure member (and not a pointer) of 'struct kvm'. That
means they start out zeroed. I _think_ they get initialized
properly by kvm_mmu_change_mmu_pages(). But, that only happens
via kvm ioctls.
Another benefit of storing 'used' intead of 'free' is
that the values are consistent from the moment the structure is
allocated: no negative "used" value.
Signed-off-by: Dave Hansen <dave@linux.vnet.ibm.com> Signed-off-by: Tim Pepper <lnxninja@linux.vnet.ibm.com> Signed-off-by: Avi Kivity <avi@redhat.com>
Dave Hansen [Fri, 20 Aug 2010 01:11:14 +0000 (18:11 -0700)]
KVM: rename x86 kvm->arch.n_alloc_mmu_pages
arch.n_alloc_mmu_pages is a poor choice of name. This value truly
means, "the number of pages which _may_ be allocated". But,
reading the name, "n_alloc_mmu_pages" implies "the number of allocated
mmu pages", which is dead wrong.
It's really the high watermark, so let's give it a name to match:
nr_max_mmu_pages. This change will make the next few patches
much more obvious and easy to read.
Signed-off-by: Dave Hansen <dave@linux.vnet.ibm.com> Signed-off-by: Tim Pepper <lnxninja@linux.vnet.ibm.com> Signed-off-by: Avi Kivity <avi@redhat.com>
Dave Hansen [Fri, 20 Aug 2010 01:11:05 +0000 (18:11 -0700)]
KVM: abstract kvm x86 mmu->n_free_mmu_pages
"free" is a poor name for this value. In this context, it means,
"the number of mmu pages which this kvm instance should be able to
allocate." But "free" implies much more that the objects are there
and ready for use. "available" is a much better description, especially
when you see how it is calculated.
In this patch, we abstract its use into a function. We'll soon
replace the function's contents by calculating the value in a
different way.
All of the reads of n_free_mmu_pages are taken care of in this
patch. The modification sites will be handled in a patch
later in the series.
Signed-off-by: Dave Hansen <dave@linux.vnet.ibm.com> Signed-off-by: Tim Pepper <lnxninja@linux.vnet.ibm.com> Signed-off-by: Avi Kivity <avi@redhat.com>
Avi Kivity [Mon, 16 Aug 2010 14:50:56 +0000 (17:50 +0300)]
KVM: x86 emulator: Use a register for ____emulate_2op() destination
Most x86 two operand instructions allow the destination to be a memory operand,
but IMUL (for example) requires that the destination be a register. Change
____emulate_2op() to take a register for both source and destination so we
can invoke IMUL.
Mohammed Gamal [Sun, 15 Aug 2010 21:47:01 +0000 (00:47 +0300)]
KVM: Separate emulation context initialization in a separate function
The code for initializing the emulation context is duplicated at two
locations (emulate_instruction() and kvm_task_switch()). Separate it
in a separate function and call it from there.
Signed-off-by: Mohammed Gamal <m.gamal005@gmail.com> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 13:04:22 +0000 (15:04 +0200)]
KVM: PPC: Move KVM trampolines before __end_interrupts
When using a relocatable kernel we need to make sure that the trampline code
and the interrupt handlers are both copied to low memory. The only way to do
this reliably is to put them in the copied section.
This patch should make relocated kernels work with KVM.
KVM-Stable-Tag Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 13:04:20 +0000 (15:04 +0200)]
KVM: PPC: Use MSR_DR for external load_up
Book3S_32 requires MSR_DR to be disabled during load_up_xxx while on Book3S_64
it's supposed to be enabled. I misread the code and disabled it in both cases,
potentially breaking the PS3 which has a really small RMA.
This patch makes KVM work on the PS3 again.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 13:04:19 +0000 (15:04 +0200)]
KVM: PPC: Add book3s_32 tlbie flush acceleration
On Book3s_32 the tlbie instruction flushed effective addresses by the mask
0x0ffff000. This is pretty hard to reflect with a hash that hashes ~0xfff, so
to speed up that target we should also keep a special hash around for it.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 13:04:17 +0000 (15:04 +0200)]
KVM: PPC: RCU'ify the Book3s MMU
So far we've been running all code without locking of any sort. This wasn't
really an issue because I didn't see any parallel access to the shadow MMU
code coming.
But then I started to implement dirty bitmapping to MOL which has the video
code in its own thread, so suddenly we had the dirty bitmap code run in
parallel to the shadow mmu code. And with that came trouble.
So I went ahead and made the MMU modifying functions as parallelizable as
I could think of. I hope I didn't screw up too much RCU logic :-). If you
know your way around RCU and locking and what needs to be done when, please
take a look at this patch.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:48:08 +0000 (14:48 +0200)]
KVM: PPC: Add get_pvinfo interface to query hypercall instructions
We need to tell the guest the opcodes that make up a hypercall through
interfaces that are controlled by userspace. So we need to add a call
for userspace to allow it to query those opcodes so it can pass them
on.
This is required because the hypercall opcodes can change based on
the hypervisor conditions. If we're running in hardware accelerated
hypervisor mode, a hypercall looks different from when we're running
without hardware acceleration.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:48:07 +0000 (14:48 +0200)]
KVM: PPC: Add Documentation about PV interface
We just introduced a new PV interface that screams for documentation. So here
it is - a shiny new and awesome text file describing the internal works of
the PPC KVM paravirtual interface.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:48:05 +0000 (14:48 +0200)]
KVM: PPC: PV mtmsrd L=0 and mtmsr
There is also a form of mtmsr where all bits need to be addressed. While the
PPC64 Linux kernel behaves resonably well here, on PPC32 we do not have an
L=1 form. It does mtmsr even for simple things like only changing EE.
So we need to hook into that one as well and check for a mask of bits that we
deem safe to change from within guest context.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:48:04 +0000 (14:48 +0200)]
KVM: PPC: PV mtmsrd L=1
The PowerPC ISA has a special instruction for mtmsr that only changes the EE
and RI bits, namely the L=1 form.
Since that one is reasonably often occuring and simple to implement, let's
go with this first. Writing EE=0 is always just a store. Doing EE=1 also
requires us to check for pending interrupts and if necessary exit back to the
hypervisor.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:48:03 +0000 (14:48 +0200)]
KVM: PPC: PV assembler helpers
When we hook an instruction we need to make sure we don't clobber any of
the registers at that point. So we write them out to scratch space in the
magic page. To make sure we don't fall into a race with another piece of
hooked code, we need to disable interrupts.
To make the later patches and code in general easier readable, let's introduce
a set of defines that save and restore r30, r31 and cr. Let's also define some
helpers to read the lower 32 bits of a 64 bit field on 32 bit systems.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:48:02 +0000 (14:48 +0200)]
KVM: PPC: Introduce branch patching helper
We will need to patch several instruction streams over to a different
code path, so we need a way to patch a single instruction with a branch
somewhere else.
This patch adds a helper to facilitate this patching.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:48:01 +0000 (14:48 +0200)]
KVM: PPC: Introduce kvm_tmp framework
We will soon require more sophisticated methods to replace single instructions
with multiple instructions. We do that by branching to a memory region where we
write replacement code for the instruction to.
This region needs to be within 32 MB of the patched instruction though, because
that's the furthest we can jump with immediate branches.
So we keep 1MB of free space around in bss. After we're done initing we can just
tell the mm system that the unused pages are free, but until then we have enough
space to fit all our code in.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:58 +0000 (14:47 +0200)]
KVM: PPC: KVM PV guest stubs
We will soon start and replace instructions from the text section with
other, paravirtualized versions. To ease the readability of those patches
I split out the generic looping and magic page mapping code out.
This patch still only contains stubs. But at least it loops through the
text section :).
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:56 +0000 (14:47 +0200)]
KVM: Move kvm_guest_init out of generic code
Currently x86 is the only architecture that uses kvm_guest_init(). With
PowerPC we're getting a second user, but the signature is different there
and we don't need to export it, as it uses the normal kernel init framework.
So let's move the x86 specific definition of that function over to the x86
specfic header file.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:53 +0000 (14:47 +0200)]
KVM: PPC: First magic page steps
We will be introducing a method to project the shared page in guest context.
As soon as we're talking about this coupling, the shared page is colled magic
page.
This patch introduces simple defines, so the follow-up patches are easier to
read.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:52 +0000 (14:47 +0200)]
KVM: PPC: Make PAM a define
On PowerPC it's very normal to not support all of the physical RAM in real mode.
To check if we're matching on the shared page or not, we need to know the limits
so we can restrain ourselves to that range.
So let's make it a define instead of open-coding it. And while at it, let's also
increase it.
Signed-off-by: Alexander Graf <agraf@suse.de>
v2 -> v3:
- RMO -> PAM (non-magic page) Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:51 +0000 (14:47 +0200)]
KVM: PPC: Tell guest about pending interrupts
When the guest turns on interrupts again, it needs to know if we have an
interrupt pending for it. Because if so, it should rather get out of guest
context and get the interrupt.
So we introduce a new field in the shared page that we use to tell the guest
that there's a pending interrupt lying around.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:49 +0000 (14:47 +0200)]
KVM: PPC: Add PV guest critical sections
When running in hooked code we need a way to disable interrupts without
clobbering any interrupts or exiting out to the hypervisor.
To achieve this, we have an additional critical field in the shared page. If
that field is equal to the r1 register of the guest, it tells the hypervisor
that we're in such a critical section and thus may not receive any interrupts.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:47 +0000 (14:47 +0200)]
KVM: PPC: Convert SPRG[0-4] to shared page
When in kernel mode there are 4 additional registers available that are
simple data storage. Instead of exiting to the hypervisor to read and
write those, we can just share them with the guest using the page.
This patch converts all users of the current field to the shared page.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:46 +0000 (14:47 +0200)]
KVM: PPC: Convert SRR0 and SRR1 to shared page
The SRR0 and SRR1 registers contain cached values of the PC and MSR
respectively. They get written to by the hypervisor when an interrupt
occurs or directly by the kernel. They are also used to tell the rfi(d)
instruction where to jump to.
Because it only gets touched on defined events that, it's very simple to
share with the guest. Hypervisor and guest both have full r/w access.
This patch converts all users of the current field to the shared page.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:45 +0000 (14:47 +0200)]
KVM: PPC: Convert DAR to shared page.
The DAR register contains the address a data page fault occured at. This
register behaves pretty much like a simple data storage register that gets
written to on data faults. There is no hypervisor interaction required on
read or write.
This patch converts all users of the current field to the shared page.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:44 +0000 (14:47 +0200)]
KVM: PPC: Convert DSISR to shared page
The DSISR register contains information about a data page fault. It is fully
read/write from inside the guest context and we don't need to worry about
interacting based on writes of this register.
This patch converts all users of the current field to the shared page.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
Alexander Graf [Thu, 29 Jul 2010 12:47:43 +0000 (14:47 +0200)]
KVM: PPC: Convert MSR to shared page
One of the most obvious registers to share with the guest directly is the
MSR. The MSR contains the "interrupts enabled" flag which the guest has to
toggle in critical sections.
So in order to bring the overhead of interrupt en- and disabling down, let's
put msr into the shared page. Keep in mind that even though you can fully read
its contents, writing to it doesn't always update all state. There are a few
safe fields that don't require hypervisor interaction. See the documentation
for a list of MSR bits that are safe to be set from inside the guest.
Signed-off-by: Alexander Graf <agraf@suse.de> Signed-off-by: Avi Kivity <avi@redhat.com>
projects / karo-tx-linux.git / log