Signed-off-by: NAdam Buchbinder <adam.buchbinder@gmail.com>
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>

The enable_kernel_*() functions leave the relevant MSR bits enabled
until we exit the kernel sometime later. Create disable versions
that wrap the kernel use of FP, Altivec VSX or SPE.

While we don't want to disable it normally for performance reasons
(MSR writes are slow), it will be used for a debug boot option that
does this and catches bad uses in other areas of the kernel.
Signed-off-by: NAnton Blanchard <anton@samba.org>
Signed-off-by: NMichael Ellerman <mpe@ellerman.id.au>

This new statistic can help diagnosing VCPUs that, for any reason,
trigger bad behavior of halt_poll_ns autotuning.

For example, say halt_poll_ns = 480000, and wakeups are spaced exactly
like 479us, 481us, 479us, 481us. Then KVM always fails polling and wastes
10+20+40+80+160+320+480 = 1110 microseconds out of every
479+481+479+481+479+481+479 = 3359 microseconds. The VCPU then
is consuming about 30% more CPU than it would use without
polling.  This would show as an abnormally high number of
attempted polling compared to the successful polls.

Acked-by: Christian Borntraeger <borntraeger@de.ibm.com<
Reviewed-by: NDavid Matlack <dmatlack@google.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

On this switch branch the regs initialization
doesn't happen so add it.
This was found with the help of a static
code analysis tool.
Signed-off-by: NLaurentiu Tudor <Laurentiu.Tudor@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This lets the function access the new memory slot without going through
kvm_memslots and id_to_memslot.  It will simplify the code when more
than one address space will be supported.

Unfortunately, the "const"ness of the new argument must be casted
away in two places.  Fixing KVM to accept const struct kvm_memory_slot
pointers would require modifications in pretty much all architectures,
and is left for later.
Reviewed-by: NRadim Krcmar <rkrcmar@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Architecture-specific helpers are not supposed to muck with
struct kvm_userspace_memory_region contents.  Add const to
enforce this.

In order to eliminate the only write in __kvm_set_memory_region,
the cleaning of deleted slots is pulled up from update_memslots
to __kvm_set_memory_region.
Reviewed-by: NTakuya Yoshikawa <yoshikawa_takuya_b1@lab.ntt.co.jp>
Reviewed-by: NRadim Krcmar <rkrcmar@redhat.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

This patch introduces a new module parameter for the KVM module; when it
is present, KVM attempts a bit of polling on every HLT before scheduling
itself out via kvm_vcpu_block.

This parameter helps a lot for latency-bound workloads---in particular
I tested it with O_DSYNC writes with a battery-backed disk in the host.
In this case, writes are fast (because the data doesn't have to go all
the way to the platters) but they cannot be merged by either the host or
the guest.  KVM's performance here is usually around 30% of bare metal,
or 50% if you use cache=directsync or cache=writethrough (these
parameters avoid that the guest sends pointless flush requests, and
at the same time they are not slow because of the battery-backed cache).
The bad performance happens because on every halt the host CPU decides
to halt itself too.  When the interrupt comes, the vCPU thread is then
migrated to a new physical CPU, and in general the latency is horrible
because the vCPU thread has to be scheduled back in.

With this patch performance reaches 60-65% of bare metal and, more
important, 99% of what you get if you use idle=poll in the guest.  This
means that the tunable gets rid of this particular bottleneck, and more
work can be done to improve performance in the kernel or QEMU.

Of course there is some price to pay; every time an otherwise idle vCPUs
is interrupted by an interrupt, it will poll unnecessarily and thus
impose a little load on the host.  The above results were obtained with
a mostly random value of the parameter (500000), and the load was around
1.5-2.5% CPU usage on one of the host's core for each idle guest vCPU.

The patch also adds a new stat, /sys/kernel/debug/kvm/halt_successful_poll,
that can be used to tune the parameter.  It counts how many HLT
instructions received an interrupt during the polling period; each
successful poll avoids that Linux schedules the VCPU thread out and back
in, and may also avoid a likely trip to C1 and back for the physical CPU.

While the VM is idle, a Linux 4 VCPU VM halts around 10 times per second.
Of these halts, almost all are failed polls.  During the benchmark,
instead, basically all halts end within the polling period, except a more
or less constant stream of 50 per second coming from vCPUs that are not
running the benchmark.  The wasted time is thus very low.  Things may
be slightly different for Windows VMs, which have a ~10 ms timer tick.

The effect is also visible on Marcelo's recently-introduced latency
test for the TSC deadline timer.  Though of course a non-RT kernel has
awful latency bounds, the latency of the timer is around 8000-10000 clock
cycles compared to 20000-120000 without setting halt_poll_ns.  For the TSC
deadline timer, thus, the effect is both a smaller average latency and
a smaller variance.
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The kvmppc_get_last_inst function recently received a facelift that allowed
us to pass an enum of the type of instruction we want to read into it rather
than an unreadable boolean.

Unfortunately, not all callers ended up passing the enum. This wasn't really
an issue as "true" and "false" happen to match the two enum values we have,
but it's still hard to read.

Update all callers of kvmppc_get_last_inst() to follow the new calling
convention.
Signed-off-by: NAlexander Graf <agraf@suse.de>

This patch extends the use of illegal instruction as software
breakpoint instruction across the ppc platform. Patch extends
booke program interrupt code to support software breakpoint.
Signed-off-by: NMadhavan Srinivasan <maddy@linux.vnet.ibm.com>
[agraf: Fix bookehv]
Signed-off-by: NAlexander Graf <agraf@suse.de>

Powerpc timer implementation is a copycat version of s390. Now that they removed
the tasklet with commit ea74c0ea follow this
optimization.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NBogdan Purcareata <bogdan.purcareata@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This patch emulates debug registers and debug exception
to support guest using debug resource. This enables running
gdb/kgdb etc in guest.

On BOOKE architecture we cannot share debug resources between QEMU and
guest because:
    When QEMU is using debug resources then debug exception must
    be always enabled. To achieve this we set MSR_DE and also set
    MSRP_DEP so guest cannot change MSR_DE.

    When emulating debug resource for guest we want guest
    to control MSR_DE (enable/disable debug interrupt on need).

    So above mentioned two configuration cannot be supported
    at the same time. So the result is that we cannot share
    debug resources between QEMU and Guest on BOOKE architecture.

In the current design QEMU gets priority over guest, this means that if
QEMU is using debug resources then guest cannot use them and if guest is
using debug resource then QEMU can overwrite them.
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Make ONE_REG generic for server and embedded architectures by moving
kvm_vcpu_ioctl_get_one_reg() and kvm_vcpu_ioctl_set_one_reg() functions
to powerpc layer.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Add AltiVec support in KVM for Book3e. FPU support gracefully reuse host
infrastructure so follow the same approach for AltiVec.

Book3e specification defines shared interrupt numbers for SPE and AltiVec
units. Still SPE is present in e200/e500v2 cores while AltiVec is present in
e6500 core. So we can currently decide at compile-time which of the SPE or
AltiVec units to support exclusively by using CONFIG_SPE_POSSIBLE and
CONFIG_PPC_E500MC defines. As Alexander Graf suggested, keep SPE and AltiVec
exception handlers distinct to improve code readability.

Guests have the privilege to enable AltiVec, so we always need to support
AltiVec in KVM and implicitly in host to reflect interrupts and to save/restore
the unit context. KVM will be loaded on cores with AltiVec unit only if
CONFIG_ALTIVEC is defined. Use this define to guard KVM AltiVec logic.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Increase FPU laziness by loading the guest state into the unit before entering
the guest instead of doing it on each vcpu schedule. Without this improvement
an interrupt may claim floating point corrupting guest state.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Guest visible debug register and hardware visible debug registers are
same, so ther is no need to have arch->shadow_dbg_reg, instead use
arch->dbg_reg.
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Dbsr is not visible to userspace and we do not think any need to
expose this to userspace because:
  Userspace cannot inject debug interrupt to guest (as this
  does not know guest ability to handle debug interrupt), so
  userspace will always clear DBSR.
  Now if userspace has to always clear DBSR in KVM_EXIT_DEBUG
  handling then clearing dbsr in kernel looks simple as this
  avoid doing SET_SREGS/set_one_reg() to clear DBSR
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Debug interrupt can be either "critical level" or "debug level".
There are separate set of save/restore registers used for different level.
Example: DSRR0/DSRR1 are used for "debug level" and CSRR0/CSRR1
are used for critical level debug interrupt.

Using CPU_FTR_DEBUG_LVL_EXC to decide which interrupt level to be used.
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

DCR handling was only needed for 440 KVM. Since we removed it, we can also
remove handling of DCR accesses.
Signed-off-by: NAlexander Graf <agraf@suse.de>

We're going to implement guest code interpretation in KVM for some rare
corner cases. This code needs to be able to inject data and instruction
faults into the guest when it encounters them.

Expose generic APIs to do this in a reasonably subarch agnostic fashion.
Signed-off-by: NAlexander Graf <agraf@suse.de>

We have a nice API to find the translated GPAs of a GVA including protection
flags. So far we only use it on Book3S, but there's no reason the same shouldn't
be used on BookE as well.

Implement a kvmppc_xlate() version for BookE and clean it up to make it more
readable in general.
Signed-off-by: NAlexander Graf <agraf@suse.de>

On book3e, KVM uses load external pid (lwepx) dedicated instruction to read
guest last instruction on the exit path. lwepx exceptions (DTLB_MISS, DSI
and LRAT), generated by loading a guest address, needs to be handled by KVM.
These exceptions are generated in a substituted guest translation context
(EPLC[EGS] = 1) from host context (MSR[GS] = 0).

Currently, KVM hooks only interrupts generated from guest context (MSR[GS] = 1),
doing minimal checks on the fast path to avoid host performance degradation.
lwepx exceptions originate from host state (MSR[GS] = 0) which implies
additional checks in DO_KVM macro (beside the current MSR[GS] = 1) by looking
at the Exception Syndrome Register (ESR[EPID]) and the External PID Load Context
Register (EPLC[EGS]). Doing this on each Data TLB miss exception is obvious
too intrusive for the host.

Read guest last instruction from kvmppc_load_last_inst() by searching for the
physical address and kmap it. This address the TODO for TLB eviction and
execute-but-not-read entries, and allow us to get rid of lwepx until we are
able to handle failures.

A simple stress benchmark shows a 1% sys performance degradation compared with
previous approach (lwepx without failure handling):

time for i in `seq 1 10000`; do /bin/echo > /dev/null; done

real    0m 8.85s
user    0m 4.34s
sys     0m 4.48s

vs

real    0m 8.84s
user    0m 4.36s
sys     0m 4.44s

A solution to use lwepx and to handle its exceptions in KVM would be to temporary
highjack the interrupt vector from host. This imposes additional synchronizations
for cores like FSL e6500 that shares host IVOR registers between hardware threads.
This optimized solution can be later developed on top of this patch.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

On book3e, guest last instruction is read on the exit path using load
external pid (lwepx) dedicated instruction. This load operation may fail
due to TLB eviction and execute-but-not-read entries.

This patch lay down the path for an alternative solution to read the guest
last instruction, by allowing kvmppc_get_lat_inst() function to fail.
Architecture specific implmentations of kvmppc_load_last_inst() may read
last guest instruction and instruct the emulation layer to re-execute the
guest in case of failure.

Make kvmppc_get_last_inst() definition common between architectures.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

kvmppc_set_epr() is already defined in asm/kvm_ppc.h, So
rename and move get_epr helper function to same file.
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
[agraf: remove duplicate return]
Signed-off-by: NAlexander Graf <agraf@suse.de>

Use kvmppc_set_sprg[0-7]() and kvmppc_get_sprg[0-7]() helper
functions
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Add and use kvmppc_set_esr() and kvmppc_get_esr() helper functions
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Uses kvmppc_set_dar() and kvmppc_get_dar() helper functions
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Use kvmppc_set_srr0/srr1() and kvmppc_get_srr0/srr1() helper functions
Signed-off-by: NBharat Bhushan <Bharat.Bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Simplify the handling of lazy EE by going directly from fully-enabled
to hard-disabled.  This replaces the lazy_irq_pending() check
(including its misplaced kvm_guest_exit() call).

As suggested by Tiejun Chen, move the interrupt disabling into
kvmppc_prepare_to_enter() rather than have each caller do it.  Also
move the IRQ enabling on heavyweight exit into
kvmppc_prepare_to_enter().
Signed-off-by: NScott Wood <scottwood@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Rather than calling hard_irq_disable() when we're back in C code
we can just call RECONCILE_IRQ_STATE to soft disable IRQs while
we're already in hard disabled state.

This should be functionally equivalent to the code before, but
cleaner and faster.
Signed-off-by: NTiejun Chen <tiejun.chen@windriver.com>
[agraf: fix comment, commit message]
Signed-off-by: NAlexander Graf <agraf@suse.de>

KVM uses same WIM tlb attributes as the corresponding qemu pte.
For this we now search the linux pte for the requested page and
get these cache caching/coherency attributes from pte.
Signed-off-by: NBharat Bhushan <bharat.bhushan@freescale.com>
Reviewed-by: NScott Wood <scottwood@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Now that we have the vcpu floating-point and vector state stored in
the same type of struct as the main kernel uses, we can load that
state directly from the vcpu struct instead of having extra copies
to/from the thread_struct.  Similarly, when the guest state needs to
be saved, we can have it saved it directly to the vcpu struct by
setting the current->thread.fp_save_area and current->thread.vr_save_area
pointers.  That also means that we don't need to back up and restore
userspace's FP/vector state.  This all makes the code simpler and
faster.

Note that it's not necessary to save or modify current->thread.fpexc_mode,
since nothing in KVM uses or is affected by its value.  Nor is it
necessary to touch used_vr or used_vsr.
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This uses struct thread_fp_state and struct thread_vr_state to store
the floating-point, VMX/Altivec and VSX state, rather than flat arrays.
This makes transferring the state to/from the thread_struct simpler
and allows us to unify the get/set_one_reg implementations for the
VSX registers.
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Commit ce11e48b ("KVM: PPC: E500: Add
userspace debug stub support") added "struct thread_struct" to the
stack of kvmppc_vcpu_run().  thread_struct is 1152 bytes on my build,
compared to 48 bytes for the recently-introduced "struct debug_reg".
Use the latter instead.

This fixes the following error:

cc1: warnings being treated as errors
arch/powerpc/kvm/booke.c: In function 'kvmppc_vcpu_run':
arch/powerpc/kvm/booke.c:760:1: error: the frame size of 1424 bytes is larger than 1024 bytes
make[2]: *** [arch/powerpc/kvm/booke.o] Error 1
make[1]: *** [arch/powerpc/kvm] Error 2
make[1]: *** Waiting for unfinished jobs....
Signed-off-by: NScott Wood <scottwood@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This moves the kvmppc_ops callbacks to be a per VM entity. This
enables us to select HV and PR mode when creating a VM. We also
allow both kvm-hv and kvm-pr kernel module to be loaded. To
achieve this we move /dev/kvm ownership to kvm.ko module. Depending on
which KVM mode we select during VM creation we take a reference
count on respective module
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
[agraf: fix coding style]
Signed-off-by: NAlexander Graf <agraf@suse.de>

We will use that in the later patch to find the kvm ops handler
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This patch add a new callback kvmppc_ops. This will help us in enabling
both HV and PR KVM together in the same kernel. The actual change to
enable them together is done in the later patch in the series.
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
[agraf: squash in booke changes]
Signed-off-by: NAlexander Graf <agraf@suse.de>

This patch adds the debug stub support on booke/bookehv.
Now QEMU debug stub can use hw breakpoint, watchpoint and
software breakpoint to debug guest.

This is how we save/restore debug register context when switching
between guest, userspace and kernel user-process:

When QEMU is running
 -> thread->debug_reg == QEMU debug register context.
 -> Kernel will handle switching the debug register on context switch.
 -> no vcpu_load() called

QEMU makes ioctls (except RUN)
 -> This will call vcpu_load()
 -> should not change context.
 -> Some ioctls can change vcpu debug register, context saved in vcpu->debug_regs

QEMU Makes RUN ioctl
 -> Save thread->debug_reg on STACK
 -> Store thread->debug_reg == vcpu->debug_reg
 -> load thread->debug_reg
 -> RUN VCPU ( So thread points to vcpu context )

Context switch happens When VCPU running
 -> makes vcpu_load() should not load any context
 -> kernel loads the vcpu context as thread->debug_regs points to vcpu context.

On heavyweight_exit
 -> Load the context saved on stack in thread->debug_reg

Currently we do not support debug resource emulation to guest,
On debug exception, always exit to user space irrespective of
user space is expecting the debug exception or not. If this is
unexpected exception (breakpoint/watchpoint event not set by
userspace) then let us leave the action on user space. This
is similar to what it was before, only thing is that now we
have proper exit state available to user space.
Signed-off-by: NBharat Bhushan <bharat.bhushan@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>