Today we handle split real mode by mapping both instruction and data faults
into a special virtual address space that only exists during the split mode
phase.

This is good enough to catch 32bit Linux guests that use split real mode for
copy_from/to_user. In this case we're always prefixed with 0xc0000000 for our
instruction pointer and can map the user space process freely below there.

However, that approach fails when we're running KVM inside of KVM. Here the 1st
level last_inst reader may well be in the same virtual page as a 2nd level
interrupt handler.

It also fails when running Mac OS X guests. Here we have a 4G/4G split, so a
kernel copy_from/to_user implementation can easily overlap with user space
addresses.

The architecturally correct way to fix this would be to implement an instruction
interpreter in KVM that kicks in whenever we go into split real mode. This
interpreter however would not receive a great amount of testing and be a lot of
bloat for a reasonably isolated corner case.

So I went back to the drawing board and tried to come up with a way to make
split real mode work with a single flat address space. And then I realized that
we could get away with the same trick that makes it work for Linux:

Whenever we see an instruction address during split real mode that may collide,
we just move it higher up the virtual address space to a place that hopefully
does not collide (keep your fingers crossed!).

That approach does work surprisingly well. I am able to successfully run
Mac OS X guests with KVM and QEMU (no split real mode hacks like MOL) when I
apply a tiny timing probe hack to QEMU. I'd say this is a win over even more
broken split real mode :).
Signed-off-by: NAlexander Graf <agraf@suse.de>

When a page lookup failed because we're not allowed to write to the page, we
should not overwrite that value with another lookup on the second PTEG which
will return "page not found". Instead, we should just tell the caller that we
had a permission problem.

This fixes Mac OS X guests looping endlessly in page lookup code for me.
Signed-off-by: NAlexander Graf <agraf@suse.de>

When we have a page that we're not allowed to write to, xlate() will already
tell us -EPERM on lookup of that page. With the code as is we change it into
a "page missing" error which a guest may get confused about. Instead, just
tell the caller about the -EPERM directly.

This fixes Mac OS X guests when run with DCBZ32 emulation.
Signed-off-by: NAlexander Graf <agraf@suse.de>

When building KVM with a lot of vcores (NR_CPUS is big), we can potentially
get out of the ld immediate range for dereferences inside that struct.

Move the array to the end of our kvm_arch struct. This fixes compilation
issues with NR_CPUS=2048 for me.
Signed-off-by: NAlexander Graf <agraf@suse.de>

For FSL e6500 core the kernel uses power management SPR register (PWRMGTCR0)
to enable idle power down for cores and devices by setting up the idle count
period at boot time. With the host already controlling the power management
configuration the guest could simply benefit from it, so emulate guest request
as a general store.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

Now that we've fixed all the issues that HV KVM code had on little endian
hosts, we can enable it in the kernel configuration for users to play with.
Signed-off-by: NAlexander Graf <agraf@suse.de>

For code that doesn't live in modules we can just branch to the real function
names, giving us compatibility with ABIv1 and ABIv2.

Do this for the compiled-in code of HV KVM.
Signed-off-by: NAlexander Graf <agraf@suse.de>

On the exit path from the guest we check what type of interrupt we received
if we received one. This means we're doing hardware access to the XICS interrupt
controller.

However, when running on a little endian system, this access is byte reversed.

So let's make sure to swizzle the bytes back again and virtually make XICS
accesses big endian.
Signed-off-by: NAlexander Graf <agraf@suse.de>

Some data structures are always stored in big endian. Among those are the LPPACA
fields as well as the shadow slb. These structures might be shared with a
hypervisor.

So whenever we access those fields, make sure we do so in big endian byte order.
Signed-off-by: NAlexander Graf <agraf@suse.de>

There are a few shared data structures between the host and the guest. Most
of them get registered through the VPA interface.

These data structures are defined to always be in big endian byte order, so
let's make sure we always access them in big endian.
Signed-off-by: NAlexander Graf <agraf@suse.de>

When running on an LE host all data structures are kept in little endian
byte order. However, the HTAB still needs to be maintained in big endian.

So every time we access any HTAB we need to make sure we do so in the right
byte order. Fix up all accesses to manually byte swap.
Signed-off-by: NAlexander Graf <agraf@suse.de>

From assembly code we might not only have to explicitly BE access 64bit values,
but sometimes also 32bit ones. Add helpers that allow for easy use of lwzx/stwx
in their respective byte-reverse or native form.
Signed-off-by: NAlexander Graf <agraf@suse.de>
CC: Benjamin Herrenschmidt <benh@kernel.crashing.org>

Tlb search operation used for victim hint relies on the default tlb set by the
host. When hardware tablewalk support is enabled in the host, the default tlb is
TLB1 which leads KVM to evict the bolted entry. Set and restore the default tlb
when searching for victim hint.
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Reviewed-by: NScott Wood <scottwood@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This adds support for the H_SET_MODE hcall.  This hcall is a
multiplexer that has several functions, some of which are called
rarely, and some which are potentially called very frequently.
Here we add support for the functions that set the debug registers
CIABR (Completed Instruction Address Breakpoint Register) and
DAWR/DAWRX (Data Address Watchpoint Register and eXtension),
since they could be updated by the guest as often as every context
switch.

This also adds a kvmppc_power8_compatible() function to test to see
if a guest is compatible with POWER8 or not.  The CIABR and DAWR/X
only exist on POWER8.
Signed-off-by: NMichael Neuling <mikey@neuling.org>
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This adds code to check that when the KVM_CAP_PPC_ENABLE_HCALL
capability is used to enable or disable in-kernel handling of an
hcall, that the hcall is actually implemented by the kernel.
If not an EINVAL error is returned.

This also checks the default-enabled list of hcalls and prints a
warning if any hcall there is not actually implemented.
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NAlexander Graf <agraf@suse.de>

This provides a way for userspace controls which sPAPR hcalls get
handled in the kernel.  Each hcall can be individually enabled or
disabled for in-kernel handling, except for H_RTAS.  The exception
for H_RTAS is because userspace can already control whether
individual RTAS functions are handled in-kernel or not via the
KVM_PPC_RTAS_DEFINE_TOKEN ioctl, and because the numeric value for
H_RTAS is out of the normal sequence of hcall numbers.

Hcalls are enabled or disabled using the KVM_ENABLE_CAP ioctl for the
KVM_CAP_PPC_ENABLE_HCALL capability on the file descriptor for the VM.
The args field of the struct kvm_enable_cap specifies the hcall number
in args[0] and the enable/disable flag in args[1]; 0 means disable
in-kernel handling (so that the hcall will always cause an exit to
userspace) and 1 means enable.  Enabling or disabling in-kernel
handling of an hcall is effective across the whole VM.

The ability for KVM_ENABLE_CAP to be used on a VM file descriptor
on PowerPC is new, added by this commit.  The KVM_CAP_ENABLE_CAP_VM
capability advertises that this ability exists.

When a VM is created, an initial set of hcalls are enabled for
in-kernel handling.  The set that is enabled is the set that have
an in-kernel implementation at this point.  Any new hcall
implementations from this point onwards should not be added to the
default set without a good reason.

No distinction is made between real-mode and virtual-mode hcall
implementations; the one setting controls them both.
Signed-off-by: NPaul Mackerras <paulus@samba.org>
Signed-off-by: NAlexander Graf <agraf@suse.de>

On vcpu schedule, the condition checked for tlb pollution is too loose.
The tlb entries of a vcpu become polluted (vs stale) only when a different
vcpu within the same logical partition runs in-between. Optimize the tlb
invalidation condition keeping last_vcpu per logical partition id.

With the new invalidation condition, a guest shows 4% performance improvement
on P5020DS while running a memory stress application with the cpu oversubscribed,
the other guest running a cpu intensive workload.

Guest - old invalidation condition
  real 3.89
  user 3.87
  sys 0.01

Guest - enhanced invalidation condition
  real 3.75
  user 3.73
  sys 0.01

Host
  real 3.70
  user 1.85
  sys 0.00

The memory stress application accesses 4KB pages backed by 75% of available
TLB0 entries:

char foo[ENTRIES][4096] __attribute__ ((aligned (4096)));

int main()
{
	char bar;
	int i, j;

	for (i = 0; i < ITERATIONS; i++)
        	for (j = 0; j < ENTRIES; j++)
            		bar = foo[j][0];

	return 0;
}
Signed-off-by: NMihai Caraman <mihai.caraman@freescale.com>
Reviewed-by: NScott Wood <scottwood@freescale.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

While sending sparse with endian checks over the code base, it triggered at
some places that were missing casts or had wrong types. Fix them up.
Signed-off-by: NAlexander Graf <agraf@suse.de>

We switched to ABIv2 on Little Endian systems now which gets rid of the
dotted function names. Branch to the actual functions when we see such
a system.
Signed-off-by: NAlexander Graf <agraf@suse.de>

Both kvmppc_hv_entry_trampoline and kvmppc_entry_trampoline are
assembly functions that are exported to modules and also require
a valid r2.

As such we need to use _GLOBAL_TOC so we provide a global entry
point that establishes the TOC (r2).
Signed-off-by: NAnton Blanchard <anton@samba.org>
Signed-off-by: NAlexander Graf <agraf@suse.de>

To establish addressability quickly, ABIv2 requires the target
address of the function being called to be in r12.
Signed-off-by: NAnton Blanchard <anton@samba.org>
Signed-off-by: NAlexander Graf <agraf@suse.de>

If we're running PR KVM in HV mode, we may get hypervisor doorbell interrupts.
Handle those the same way we treat normal doorbells.
Signed-off-by: NAlexander Graf <agraf@suse.de>

Some compilers complain about uninitialized variables in the compute_tlbie_rb
function. When you follow the code path you'll realize that we'll never get
to that point, but the compiler isn't all that smart.

So just default to 4k page sizes for everything, making the compiler happy
and the code slightly easier to read.
Signed-off-by: NAlexander Graf <agraf@suse.de>
Acked-by: NPaul Mackerras <paulus@samba.org>

When we're using PR KVM we must not allow the CPU to take interrupts
in virtual mode, as the SLB does not contain host kernel mappings
when running inside the guest context.

To make sure we get good performance for non-KVM tasks but still
properly functioning PR KVM, let's just disable AIL whenever a vcpu
is scheduled in.

This is fundamentally different from how we deal with AIL on pSeries
type machines where we disable AIL for the whole machine as soon as
a single KVM VM is up.

The reason for that is easy - on pSeries we do not have control over
per-cpu configuration of AIL. We also don't want to mess with CPU hotplug
races and AIL configuration, so setting it per CPU is easier and more
flexible.

This patch fixes running PR KVM on POWER8 bare metal for me.
Signed-off-by: NAlexander Graf <agraf@suse.de>
Acked-by: NPaul Mackerras <paulus@samba.org>

Writing to IC is not allowed in the privileged mode.
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

virtual time base register is a per VM, per cpu register that needs
to be saved and restored on vm exit and entry. Writing to VTB is not
allowed in the privileged mode.
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
[agraf: fix compile error]
Signed-off-by: NAlexander Graf <agraf@suse.de>

We use time base for PURR and SPURR emulation with PR KVM since we
are emulating a single threaded core. When using time base
we need to make sure that we don't accumulate time spent in the host
in PURR and SPURR value.

Also we don't need to emulate mtspr because both the registers are
hypervisor resource.
Signed-off-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NAlexander Graf <agraf@suse.de>

The rule to create the final images uses a zImage.% pattern.
Unfortunately, this also matches the names of the zImage.*.lds linker
scripts, which appear as a dependency of the final images. This somehow
worked when $(srctree) used to be an absolute path, but now the pattern
matches too much. List only the images from $(image-y) as the target of
the rule, to avoid the circular dependency.
Reported-and-tested-by: NMike Qiu <qiudayu@linux.vnet.ibm.com>
Signed-off-by: NMichal Marek <mmarek@suse.cz>

Commit 2749a2f2 (powerpc/book3s: Fix machine check handling for
unhandled errors) introduced a few ABIv2 issues.

We can maintain ABIv1 and ABIv2 compatibility by branching to the
function rather than the dot symbol.

Fixes: 2749a2f2 ("powerpc/book3s: Fix machine check handling for unhandled errors")
Signed-off-by: NAnton Blanchard <anton@samba.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Currently we forward MCEs to guest which have been recovered by guest.
And for unhandled errors we do not deliver the MCE to guest. It looks like
with no support of FWNMI in qemu, guest just panics whenever we deliver the
recovered MCEs to guest. Also, the existig code used to return to host for
unhandled errors which was casuing guest to hang with soft lockups inside
guest and makes it difficult to recover guest instance.

This patch now forwards all fatal MCEs to guest causing guest to crash/panic.
And, for recovered errors we just go back to normal functioning of guest
instead of returning to host. This fixes soft lockup issues in guest.
This patch also fixes an issue where guest MCE events were not logged to
host console.
Signed-off-by: NMahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

We don't see MCE counter getting increased in /proc/interrupts which gives
false impression of no MCE occurred even when there were MCE events.
The machine check early handling was added for PowerKVM and we missed to
increment the MCE count in the early handler.

We also increment mce counters in the machine_check_exception call, but
in most cases where we handle the error hypervisor never reaches there
unless its fatal and we want to crash. Only during fatal situation we may
see double increment of mce count. We need to fix that. But for
now it always good to have some count increased instead of zero.
Signed-off-by: NMahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Currently machine check handler does not check for stack overflow for
nested machine check. If we hit another MCE while inside the machine check
handler repeatedly from same address then we get into risk of stack
overflow which can cause huge memory corruption. This patch limits the
nested MCE level to 4 and panic when we cross level 4.
Signed-off-by: NMahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Current code does not check for unhandled/unrecovered errors and return from
interrupt if it is recoverable exception which in-turn triggers same machine
check exception in a loop causing hypervisor to be unresponsive.

This patch fixes this situation and forces hypervisor to panic for
unhandled/unrecovered errors.

This patch also fixes another issue where unrecoverable_exception routine
was called in real mode in case of unrecoverable exception (MSR_RI = 0).
This causes another exception vector 0x300 (data access) during system crash
leading to confusion while debugging cause of the system crash.

Also turn ME bit off while going down, so that when another MCE is hit during
panic path, system will checkstop and hypervisor will get restarted cleanly
by SP.

With the above fixes we now throw correct console messages (see below) while
crashing the system in case of unhandled/unrecoverable machine checks.

--------------
Severe Machine check interrupt [[Not recovered]
  Initiator: CPU
  Error type: UE [Instruction fetch]
    Effective address: 0000000030002864
Oops: Machine check, sig: 7 [#1]
SMP NR_CPUS=2048 NUMA PowerNV
Modules linked in: bork(O) bridge stp llc kvm [last unloaded: bork]
CPU: 36 PID: 55162 Comm: bash Tainted: G           O 3.14.0mce #1
task: c000002d72d022d0 ti: c000000007ec0000 task.ti: c000002d72de4000
NIP: 0000000030002864 LR: 00000000300151a4 CTR: 000000003001518c
REGS: c000000007ec3d80 TRAP: 0200   Tainted: G           O  (3.14.0mce)
MSR: 9000000000041002 <SF,HV,ME,RI>  CR: 28222848  XER: 20000000
CFAR: 0000000030002838 DAR: d0000000004d0000 DSISR: 00000000 SOFTE: 1
GPR00: 000000003001512c 0000000031f92cb0 0000000030078af0 0000000030002864
GPR04: d0000000004d0000 0000000000000000 0000000030002864 ffffffffffffffc9
GPR08: 0000000000000024 0000000030008af0 000000000000002c c00000000150e728
GPR12: 9000000000041002 0000000031f90000 0000000010142550 0000000040000000
GPR16: 0000000010143cdc 0000000000000000 00000000101306fc 00000000101424dc
GPR20: 00000000101424e0 000000001013c6f0 0000000000000000 0000000000000000
GPR24: 0000000010143ce0 00000000100f6440 c000002d72de7e00 c000002d72860250
GPR28: c000002d72860240 c000002d72ac0038 0000000000000008 0000000000040000
NIP [0000000030002864] 0x30002864
LR [00000000300151a4] 0x300151a4
Call Trace:
Instruction dump:
XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX
XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX XXXXXXXX
---[ end trace 7285f0beac1e29d3 ]---

Sending IPI to other CPUs
IPI complete
OPAL V3 detected !
--------------
Signed-off-by: NMahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

As Ben suggested, it's meaningful to dump PE's location code
for site engineers when hitting EEH errors. The patch introduces
function eeh_pe_loc_get() to retireve the location code from
dev-tree so that we can output it when hitting EEH errors.

If primary PE bus is root bus, the PHB's dev-node would be tried
prior to root port's dev-node. Otherwise, the upstream bridge's
dev-node of the primary PE bus will be check for the location code
directly.
Signed-off-by: NGavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

This patch enables POWER8 doorbell IPIs on powernv.

Since doorbells can only IPI within a core, we test to see when we can use
doorbells and if not we fall back to XICS.  This also enables hypervisor
doorbells to wakeup us up from nap/sleep via the LPCR PECEDH bit.

Based on tests by Anton, the best case IPI latency between two threads dropped
from 894ns to 512ns.
Signed-off-by: NMichael Neuling <mikey@neuling.org>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

On PowerNV platform, EEH errors are reported by IO accessors or poller
driven by interrupt. After the PE is isolated, we won't produce EEH
event for the PE. The current implementation has possibility of EEH
event lost in this way:

The interrupt handler queues one "special" event, which drives the poller.
EEH thread doesn't pick the special event yet. IO accessors kicks in, the
frozen PE is marked as "isolated" and EEH event is queued to the list.
EEH thread runs because of special event and purge all existing EEH events.
However, we never produce an other EEH event for the frozen PE. Eventually,
the PE is marked as "isolated" and we don't have EEH event to recover it.

The patch fixes the issue to keep EEH events for PEs that have been
marked as "isolated" with the help of additional "force" help to
eeh_remove_event().
Reported-by: NRolf Brudeseth <rolfb@us.ibm.com>
Signed-off-by: NGavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Commit b0d278b7 ("powerpc/perf_event: Reduce latency of calling
perf_event_do_pending") added a check for CONFIG_PMAC were a check for
CONFIG_PPC_PMAC was clearly intended.

Fixes: b0d278b7 ("powerpc/perf_event: Reduce latency of calling perf_event_do_pending")
Signed-off-by: NPaul Bolle <pebolle@tiscali.nl>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Commit cd64d169 ("powerpc: mtmsrd not defined") added a check for
CONFIG_PPC_CPU were a check for CONFIG_PPC_FPU was clearly intended.

Fixes: cd64d169 ("powerpc: mtmsrd not defined")
Signed-off-by: NPaul Bolle <pebolle@tiscali.nl>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

Commit cb5b242c ("powerpc/eeh: Escalate error on non-existing PE")
escalates the frozen state on non-existing PE to fenced PHB. It
was to improve kdump reliability. After that, commit 361f2a2a
("powrpc/powernv: Reset PHB in kdump kernel") was introduced to
issue complete reset on all PHBs to increase the reliability of
kdump kernel.

Commit cb5b242c becomes unuseful and it would be reverted.
Signed-off-by: NGavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>

When we have the corner case of frozen parent and child PE at the
same time, we have to handle the frozen parent PE prior to the
child. Without clearning the frozen state on parent PE, the child
PE can't be recovered successfully.

The patch searches the EEH PE hierarchy tree and returns the toppest
frozen PE to be handled. It ensures the frozen parent PE will be
handled prior to child PE.
Signed-off-by: NGavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: NBenjamin Herrenschmidt <benh@kernel.crashing.org>