Since commit b009031f74 ("KVM: PPC: Book3S HV: Take out virtual
core piggybacking code", 2016-09-15), we only have at most one
vcore per subcore. Previously, the fact that there might be more
than one vcore per subcore meant that we had the notion of a
"master vcore", which was the vcore that controlled thread 0 of
the subcore. We also needed a list per subcore in the core_info
struct to record which vcores belonged to each subcore. Now that
there can only be one vcore in the subcore, we can replace the
list with a simple pointer and get rid of the notion of the
master vcore (and in fact treat every vcore as a master vcore).
We can also get rid of the subcore_vm[] field in the core_info
struct since it is never read.
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This adds a few last pieces of the support for radix guests:
* Implement the backends for the KVM_PPC_CONFIGURE_V3_MMU and
KVM_PPC_GET_RMMU_INFO ioctls for radix guests
* On POWER9, allow secondary threads to be on/off-lined while guests
are running.
* Set up LPCR and the partition table entry for radix guests.
* Don't allocate the rmap array in the kvm_memory_slot structure
on radix.
* Don't try to initialize the HPT for radix guests, since they don't
have an HPT.
* Take out the code that prevents the HV KVM module from
initializing on radix hosts.
At this stage, we only support radix guests if the host is running
in radix mode, and only support HPT guests if the host is running in
HPT mode. Thus a guest cannot switch from one mode to the other,
which enables some simplifications.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This adds code to keep track of dirty pages when requested (that is,
when memslot->dirty_bitmap is non-NULL) for radix guests. We use the
dirty bits in the PTEs in the second-level (partition-scoped) page
tables, together with a bitmap of pages that were dirty when their
PTE was invalidated (e.g., when the page was paged out). This bitmap
is stored in the first half of the memslot->dirty_bitmap area, and
kvm_vm_ioctl_get_dirty_log_hv() now uses the second half for the
bitmap that gets returned to userspace.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This adapts our implementations of the MMU notifier callbacks
(unmap_hva, unmap_hva_range, age_hva, test_age_hva, set_spte_hva)
to call radix functions when the guest is using radix. These
implementations are much simpler than for HPT guests because we
have only one PTE to deal with, so we don't need to traverse
rmap chains.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This adds the code to construct the second-level ("partition-scoped" in
architecturese) page tables for guests using the radix MMU. Apart from
the PGD level, which is allocated when the guest is created, the rest
of the tree is all constructed in response to hypervisor page faults.
As well as hypervisor page faults for missing pages, we also get faults
for reference/change (RC) bits needing to be set, as well as various
other error conditions. For now, we only set the R or C bit in the
guest page table if the same bit is set in the host PTE for the
backing page.
This code can take advantage of the guest being backed with either
transparent or ordinary 2MB huge pages, and insert 2MB page entries
into the guest page tables. There is no support for 1GB huge pages
yet.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This adds a field in struct kvm_arch and an inline helper to
indicate whether a guest is a radix guest or not, plus a new file
to contain the radix MMU code, which currently contains just a
translate function which knows how to traverse the guest page
tables to translate an address.
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
POWER8 has one virtual timebase (VTB) register per subcore, not one
per CPU thread. The HV KVM code currently treats VTB as a per-thread
register, which can lead to spurious soft lockup messages from guests
which use the VTB as the time source for the soft lockup detector.
(CPUs before POWER8 did not have the VTB register.)
For HV KVM, this fixes the problem by making only the primary thread
in each virtual core save and restore the VTB value. With this,
the VTB state becomes part of the kvmppc_vcore structure. This
also means that "piggybacking" of multiple virtual cores onto one
subcore is not possible on POWER8, because then the virtual cores
would share a single VTB register.
PR KVM emulates a VTB register, which is per-vcpu because PR KVM
has no notion of CPU threads or SMT. For PR KVM we move the VTB
state into the kvmppc_vcpu_book3s struct.
Cc: stable@vger.kernel.org # v3.14+
Reported-by: Thomas Huth <thuth@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Add a module parameter kvm_irq_bypass for kvm_hv.ko to
disable IRQ bypass for passthrough interrupts. The default
value of this tunable is 1 - that is enable the feature.
Since the tunable is used by built-in kernel code, we use
the module_param_cb macro to achieve this.
Signed-off-by: Suresh Warrier <warrier@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
This patch introduces new halt polling functionality into the kvm_hv kernel
module. When a vcore is idle it will poll for some period of time before
scheduling itself out.
When all of the runnable vcpus on a vcore have ceded (and thus the vcore is
idle) we schedule ourselves out to allow something else to run. In the
event that we need to wake up very quickly (for example an interrupt
arrives), we are required to wait until we get scheduled again.
Implement halt polling so that when a vcore is idle, and before scheduling
ourselves, we poll for vcpus in the runnable_threads list which have
pending exceptions or which leave the ceded state. If we poll successfully
then we can get back into the guest very quickly without ever scheduling
ourselves, otherwise we schedule ourselves out as before.
There exists generic halt_polling code in virt/kvm_main.c, however on
powerpc the polling conditions are different to the generic case. It would
be nice if we could just implement an arch specific kvm_check_block()
function, but there is still other arch specific things which need to be
done for kvm_hv (for example manipulating vcore states) which means that a
separate implementation is the best option.
Testing of this patch with a TCP round robin test between two guests with
virtio network interfaces has found a decrease in round trip time of ~15us
on average. A performance gain is only seen when going out of and
back into the guest often and quickly, otherwise there is no net benefit
from the polling. The polling interval is adjusted such that when we are
often scheduled out for long periods of time it is reduced, and when we
often poll successfully it is increased. The rate at which the polling
interval increases or decreases, and the maximum polling interval, can
be set through module parameters.
Based on the implementation in the generic kvm module by Wanpeng Li and
Paolo Bonzini, and on direction from Paul Mackerras.
Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
The struct kvmppc_vcore is a structure used to store various information
about a virtual core for a kvm guest. The runnable_threads element of the
struct provides a list of all of the currently runnable vcpus on the core
(those in the KVMPPC_VCPU_RUNNABLE state). The previous implementation of
this list was a linked_list. The next patch requires that the list be able
to be iterated over without holding the vcore lock.
Reimplement the runnable_threads list in the kvmppc_vcore struct as an
array. Implement function to iterate over valid entries in the array and
update access sites accordingly.
Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
The next commit will introduce a member to the kvmppc_vcore struct which
references MAX_SMT_THREADS which is defined in kvm_book3s_asm.h, however
this file isn't included in kvm_host.h directly. Thus compiling for
certain platforms such as pmac32_defconfig and ppc64e_defconfig with KVM
fails due to MAX_SMT_THREADS not being defined.
Move the struct kvmppc_vcore definition to kvm_book3s.h which explicitly
includes kvm_book3s_asm.h.
Signed-off-by: Suraj Jitindar Singh <sjitindarsingh@gmail.com>
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
To date, we have implemented two I/O usage models for persistent memory,
PMEM (a persistent "ram disk") and DAX (mmap persistent memory into
userspace). This series adds a third, DAX-GUP, that allows DAX mappings
to be the target of direct-i/o. It allows userspace to coordinate
DMA/RDMA from/to persistent memory.
The implementation leverages the ZONE_DEVICE mm-zone that went into
4.3-rc1 (also discussed at kernel summit) to flag pages that are owned
and dynamically mapped by a device driver. The pmem driver, after
mapping a persistent memory range into the system memmap via
devm_memremap_pages(), arranges for DAX to distinguish pfn-only versus
page-backed pmem-pfns via flags in the new pfn_t type.
The DAX code, upon seeing a PFN_DEV+PFN_MAP flagged pfn, flags the
resulting pte(s) inserted into the process page tables with a new
_PAGE_DEVMAP flag. Later, when get_user_pages() is walking ptes it keys
off _PAGE_DEVMAP to pin the device hosting the page range active.
Finally, get_page() and put_page() are modified to take references
against the device driver established page mapping.
Finally, this need for "struct page" for persistent memory requires
memory capacity to store the memmap array. Given the memmap array for a
large pool of persistent may exhaust available DRAM introduce a
mechanism to allocate the memmap from persistent memory. The new
"struct vmem_altmap *" parameter to devm_memremap_pages() enables
arch_add_memory() to use reserved pmem capacity rather than the page
allocator.
This patch (of 18):
The core has developed a need for a "pfn_t" type [1]. Move the existing
pfn_t in KVM to kvm_pfn_t [2].
[1]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002199.html
[2]: https://lists.01.org/pipermail/linux-nvdimm/2015-September/002218.html
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: Christoffer Dall <christoffer.dall@linaro.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
In 64 bit kernels, the Fixed Point Exception Register (XER) is a 64
bit field (e.g. in kvm_regs and kvm_vcpu_arch) and in most places it is
accessed as such.
This patch corrects places where it is accessed as a 32 bit field by a
64 bit kernel. In some cases this is via a 32 bit load or store
instruction which, depending on endianness, will cause either the
lower or upper 32 bits to be missed. In another case it is cast as a
u32, causing the upper 32 bits to be cleared.
This patch corrects those places by extending the access methods to
64 bits.
Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Tested-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
This fixes a bug in the tracking of pages that get modified by the
guest. If the guest creates a large-page HPTE, writes to memory
somewhere within the large page, and then removes the HPTE, we only
record the modified state for the first normal page within the large
page, when in fact the guest might have modified some other normal
page within the large page.
To fix this we use some unused bits in the rmap entry to record the
order (log base 2) of the size of the page that was modified, when
removing an HPTE. Then in kvm_test_clear_dirty_npages() we use that
order to return the correct number of modified pages.
The same thing could in principle happen when removing a HPTE at the
host's request, i.e. when paging out a page, except that we never
page out large pages, and the guest can only create large-page HPTEs
if the guest RAM is backed by large pages. However, we also fix
this case for the sake of future-proofing.
The reference bit is also subject to the same loss of information. We
don't make the same fix here for the reference bit because there isn't
an interface for userspace to find out which pages the guest has
referenced, whereas there is one for userspace to find out which pages
the guest has modified. Because of this loss of information, the
kvm_age_hva_hv() and kvm_test_age_hva_hv() functions might incorrectly
say that a page has not been referenced when it has, but that doesn't
matter greatly because we never page or swap out large pages.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Book3S HV only (debugging aids, minor performance improvements and some
cleanups). But there are also bug fixes and small cleanups for ARM,
x86 and s390.
The task_migration_notifier revert and real fix is still pending review,
but I'll send it as soon as possible after -rc1.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
iQEcBAABAgAGBQJVOONLAAoJEL/70l94x66DbsMIAIpZPsaqgXOC1sDEiZuYay+6
rD4n4id7j8hIAzcf3AlZdyf5XgLlr6I1Zyt62s1WcoRq/CCnL7k9EljzSmw31WFX
P2y7/J0iBdkn0et+PpoNThfL2GsgTqNRCLOOQlKgEQwMP9Dlw5fnUbtC1UchOzTg
eAMeBIpYwufkWkXhdMw4PAD4lJ9WxUZ1eXHEBRzJb0o0ZxAATJ1tPZGrFJzoUOSM
WsVNTuBsNd7upT02kQdvA1TUo/OPjseTOEoksHHwfcORt6bc5qvpctL3jYfcr7sk
/L6sIhYGVNkjkuredjlKGLfT2DDJjSEdJb1k2pWrDRsY76dmottQubAE9J9cDTk=
=OAi2
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Pull second batch of KVM changes from Paolo Bonzini:
"This mostly includes the PPC changes for 4.1, which this time cover
Book3S HV only (debugging aids, minor performance improvements and
some cleanups). But there are also bug fixes and small cleanups for
ARM, x86 and s390.
The task_migration_notifier revert and real fix is still pending
review, but I'll send it as soon as possible after -rc1"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (29 commits)
KVM: arm/arm64: check IRQ number on userland injection
KVM: arm: irqfd: fix value returned by kvm_irq_map_gsi
KVM: VMX: Preserve host CR4.MCE value while in guest mode.
KVM: PPC: Book3S HV: Use msgsnd for signalling threads on POWER8
KVM: PPC: Book3S HV: Translate kvmhv_commence_exit to C
KVM: PPC: Book3S HV: Streamline guest entry and exit
KVM: PPC: Book3S HV: Use bitmap of active threads rather than count
KVM: PPC: Book3S HV: Use decrementer to wake napping threads
KVM: PPC: Book3S HV: Don't wake thread with no vcpu on guest IPI
KVM: PPC: Book3S HV: Get rid of vcore nap_count and n_woken
KVM: PPC: Book3S HV: Move vcore preemption point up into kvmppc_run_vcpu
KVM: PPC: Book3S HV: Minor cleanups
KVM: PPC: Book3S HV: Simplify handling of VCPUs that need a VPA update
KVM: PPC: Book3S HV: Accumulate timing information for real-mode code
KVM: PPC: Book3S HV: Create debugfs file for each guest's HPT
KVM: PPC: Book3S HV: Add ICP real mode counters
KVM: PPC: Book3S HV: Move virtual mode ICP functions to real-mode
KVM: PPC: Book3S HV: Convert ICS mutex lock to spin lock
KVM: PPC: Book3S HV: Add guest->host real mode completion counters
KVM: PPC: Book3S HV: Add helpers for lock/unlock hpte
...
On POWER, storage caching is usually configured via the MMU - attributes
such as cache-inhibited are stored in the TLB and the hashed page table.
This makes correctly performing cache inhibited IO accesses awkward when
the MMU is turned off (real mode). Some CPU models provide special
registers to control the cache attributes of real mode load and stores but
this is not at all consistent. This is a problem in particular for SLOF,
the firmware used on KVM guests, which runs entirely in real mode, but
which needs to do IO to load the kernel.
To simplify this qemu implements two special hypercalls, H_LOGICAL_CI_LOAD
and H_LOGICAL_CI_STORE which simulate a cache-inhibited load or store to
a logical address (aka guest physical address). SLOF uses these for IO.
However, because these are implemented within qemu, not the host kernel,
these bypass any IO devices emulated within KVM itself. The simplest way
to see this problem is to attempt to boot a KVM guest from a virtio-blk
device with iothread / dataplane enabled. The iothread code relies on an
in kernel implementation of the virtio queue notification, which is not
triggered by the IO hcalls, and so the guest will stall in SLOF unable to
load the guest OS.
This patch addresses this by providing in-kernel implementations of the
2 hypercalls, which correctly scan the KVM IO bus. Any access to an
address not handled by the KVM IO bus will cause a VM exit, hitting the
qemu implementation as before.
Note that a userspace change is also required, in order to enable these
new hcall implementations with KVM_CAP_PPC_ENABLE_HCALL.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[agraf: fix compilation]
Signed-off-by: Alexander Graf <agraf@suse.de>
These don't seem to be used anywhere.
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Alexander Graf <agraf@suse.de>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Will deacon <will.deacon@arm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Luiz Capitulino <lcapitulino@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
This removes the code that was added to enable HV KVM to work
on PPC970 processors. The PPC970 is an old CPU that doesn't
support virtualizing guest memory. Removing PPC970 support also
lets us remove the code for allocating and managing contiguous
real-mode areas, the code for the !kvm->arch.using_mmu_notifiers
case, the code for pinning pages of guest memory when first
accessed and keeping track of which pages have been pinned, and
the code for handling H_ENTER hypercalls in virtual mode.
Book3S HV KVM is now supported only on POWER7 and POWER8 processors.
The KVM_CAP_PPC_RMA capability now always returns 0.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
We handle FSCR feature bits (well, TAR only really today) lazily when the guest
starts using them. So when a guest activates the bit and later uses that feature
we enable it for real in hardware.
However, when the guest stops using that bit we don't stop setting it in
hardware. That means we can potentially lose a trap that the guest expects to
happen because it thinks a feature is not active.
This patch adds support to drop TAR when then guest turns it off in FSCR. While
at it it also restricts FSCR access to 64bit systems - 32bit ones don't have it.
Signed-off-by: Alexander Graf <agraf@suse.de>
We use kvmppc_ld and kvmppc_st to emulate load/store instructions that may as
well access the magic page. Special case it out so that we can properly access
it.
Signed-off-by: Alexander Graf <agraf@suse.de>
We have enough common infrastructure now to resolve GVA->GPA mappings at
runtime. With this we can move our book3s specific helpers to load / store
in guest virtual address space to common code as well.
Signed-off-by: Alexander 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: Mihai Caraman <mihai.caraman@freescale.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
The magic page is defined as a 4k page of per-vCPU data that is shared
between the guest and the host to accelerate accesses to privileged
registers.
However, when the host is using 64k page size granularity we weren't quite
as strict about that rule anymore. Instead, we partially treated all of the
upper 64k as magic page and mapped only the uppermost 4k with the actual
magic contents.
This works well enough for Linux which doesn't use any memory in kernel
space in the upper 64k, but Mac OS X got upset. So this patch makes magic
page actually stay in a 4k range even on 64k page size hosts.
This patch fixes magic page usage with Mac OS X (using MOL) on 64k PAGE_SIZE
hosts for me.
Signed-off-by: Alexander Graf <agraf@suse.de>
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: Alexander 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: Alexander 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: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander 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: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander 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: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
The shared (magic) page is a data structure that contains often used
supervisor privileged SPRs accessible via memory to the user to reduce
the number of exits we have to take to read/write them.
When we actually share this structure with the guest we have to maintain
it in guest endianness, because some of the patch tricks only work with
native endian load/store operations.
Since we only share the structure with either host or guest in little
endian on book3s_64 pr mode, we don't have to worry about booke or book3s hv.
For booke, the shared struct stays big endian. For book3s_64 hv we maintain
the struct in host native endian, since it never gets shared with the guest.
For book3s_64 pr we introduce a variable that tells us which endianness the
shared struct is in and route every access to it through helper inline
functions that evaluate this variable.
Signed-off-by: Alexander Graf <agraf@suse.de>
When the guest does an MMIO write which is handled successfully by an
ioeventfd, ioeventfd_write() returns 0 (success) and
kvmppc_handle_store() returns EMULATE_DONE. Then
kvmppc_emulate_mmio() converts EMULATE_DONE to RESUME_GUEST_NV and
this causes an exit from the loop in kvmppc_vcpu_run_hv(), causing an
exit back to userspace with a bogus exit reason code, typically
causing userspace (e.g. qemu) to crash with a message about an unknown
exit code.
This adds handling of RESUME_GUEST_NV in kvmppc_vcpu_run_hv() in order
to fix that. For generality, we define a helper to check for either
of the return-to-guest codes we use, RESUME_GUEST and RESUME_GUEST_NV,
to make it easy to check for either and provide one place to update if
any other return-to-guest code gets defined in future.
Since it only affects Book3S HV for now, the helper is added to
the kvm_book3s.h header file.
We use the helper in two places in kvmppc_run_core() as well for
future-proofing, though we don't see RESUME_GUEST_NV in either place
at present.
[paulus@samba.org - combined 4 patches into one, rewrote description]
Suggested-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
MMIO emulation reads the last instruction executed by the guest
and then emulates. If the guest is running in Little Endian order,
or more generally in a different endian order of the host, the
instruction needs to be byte-swapped before being emulated.
This patch adds a helper routine which tests the endian order of
the host and the guest in order to decide whether a byteswap is
needed or not. It is then used to byteswap the last instruction
of the guest in the endian order of the host before MMIO emulation
is performed.
Finally, kvmppc_handle_load() of kvmppc_handle_store() are modified
to reverse the endianness of the MMIO if required.
Signed-off-by: Cédric Le Goater <clg@fr.ibm.com>
[agraf: add booke handling]
Signed-off-by: Alexander Graf <agraf@suse.de>
We had code duplication between the inline functions to get our last
instruction on normal interrupts and system call interrupts. Unify
both helper functions towards a single implementation.
Signed-off-by: Alexander Graf <agraf@suse.de>
The load_up_fpu and load_up_altivec functions were never intended to
be called from C, and do things like modifying the MSR value in their
callers' stack frames, which are assumed to be interrupt frames. In
addition, on 32-bit Book S they require the MMU to be off.
This makes KVM use the new load_fp_state() and load_vr_state() functions
instead of load_up_fpu/altivec. This means we can remove the assembler
glue in book3s_rmhandlers.S, and potentially fixes a bug on Book E,
where load_up_fpu was called directly from C.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
The kvmppc_copy_{to,from}_svcpu functions are publically visible,
so we should also export them in a header for others C files to
consume.
So far we didn't need this because we only called it from asm code.
The next patch will introduce a C caller.
Signed-off-by: Alexander Graf <agraf@suse.de>
This help us to identify whether we are running with hypervisor mode KVM
enabled. The change is needed so that we can have both HV and PR kvm
enabled in the same kernel.
If both HV and PR KVM are included, interrupts come in to the HV version
of the kvmppc_interrupt code, which then jumps to the PR handler,
renamed to kvmppc_interrupt_pr, if the guest is a PR guest.
Allowing both PR and HV in the same kernel required some changes to
kvm_dev_ioctl_check_extension(), since the values returned now can't
be selected with #ifdefs as much as previously. We look at is_hv_enabled
to return the right value when checking for capabilities.For capabilities that
are only provided by HV KVM, we return the HV value only if
is_hv_enabled is true. For capabilities provided by PR KVM but not HV,
we return the PR value only if is_hv_enabled is false.
NOTE: in later patch we replace is_hv_enabled with a static inline
function comparing kvm_ppc_ops
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander 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: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
[agraf: squash in booke changes]
Signed-off-by: Alexander Graf <agraf@suse.de>
This help ups to select the relevant code in the kernel code
when we later move HV and PR bits as seperate modules. The patch
also makes the config options for PR KVM selectable
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
With later patches supporting PR kvm as a kernel module, the changes
that has to be built into the main kernel binary to enable PR KVM module
is now selected via KVM_BOOK3S_PR_POSSIBLE
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
When the MM code is invalidating a range of pages, it calls the KVM
kvm_mmu_notifier_invalidate_range_start() notifier function, which calls
kvm_unmap_hva_range(), which arranges to flush all the existing host
HPTEs for guest pages. However, the Linux PTEs for the range being
flushed are still valid at that point. We are not supposed to establish
any new references to pages in the range until the ...range_end()
notifier gets called. The PPC-specific KVM code doesn't get any
explicit notification of that; instead, we are supposed to use
mmu_notifier_retry() to test whether we are or have been inside a
range flush notifier pair while we have been getting a page and
instantiating a host HPTE for the page.
This therefore adds a call to mmu_notifier_retry inside
kvmppc_mmu_map_page(). This call is inside a region locked with
kvm->mmu_lock, which is the same lock that is called by the KVM
MMU notifier functions, thus ensuring that no new notification can
proceed while we are in the locked region. Inside this region we
also create the host HPTE and link the corresponding hpte_cache
structure into the lists used to find it later. We cannot allocate
the hpte_cache structure inside this locked region because that can
lead to deadlock, so we allocate it outside the region and free it
if we end up not using it.
This also moves the updates of vcpu3s->hpte_cache_count inside the
regions locked with vcpu3s->mmu_lock, and does the increment in
kvmppc_mmu_hpte_cache_map() when the pte is added to the cache
rather than when it is allocated, in order that the hpte_cache_count
is accurate.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently we request write access to all pages that get mapped into the
guest, even if the guest is only loading from the page. This reduces
the effectiveness of KSM because it means that we unshare every page we
access. Also, we always set the changed (C) bit in the guest HPTE if
it allows writing, even for a guest load.
This fixes both these problems. We pass an 'iswrite' flag to the
mmu.xlate() functions and to kvmppc_mmu_map_page() to indicate whether
the access is a load or a store. The mmu.xlate() functions now only
set C for stores. kvmppc_gfn_to_pfn() now calls gfn_to_pfn_prot()
instead of gfn_to_pfn() so that it can indicate whether we need write
access to the page, and get back a 'writable' flag to indicate whether
the page is writable or not. If that 'writable' flag is clear, we then
make the host HPTE read-only even if the guest HPTE allowed writing.
This means that we can get a protection fault when the guest writes to a
page that it has mapped read-write but which is read-only on the host
side (perhaps due to KSM having merged the page). Thus we now call
kvmppc_handle_pagefault() for protection faults as well as HPTE not found
faults. In kvmppc_handle_pagefault(), if the access was allowed by the
guest HPTE and we thus need to install a new host HPTE, we then need to
remove the old host HPTE if there is one. This is done with a new
function, kvmppc_mmu_unmap_page(), which uses kvmppc_mmu_pte_vflush() to
find and remove the old host HPTE.
Since the memslot-related functions require the KVM SRCU read lock to
be held, this adds srcu_read_lock/unlock pairs around the calls to
kvmppc_handle_pagefault().
Finally, this changes kvmppc_mmu_book3s_32_xlate_pte() to not ignore
guest HPTEs that don't permit access, and to return -EPERM for accesses
that are not permitted by the page protections.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This makes PR KVM allocate its kvm_vcpu structs from the kvm_vcpu_cache
rather than having them embedded in the kvmppc_vcpu_book3s struct,
which is allocated with vzalloc. The reason is to reduce the
differences between PR and HV KVM in order to make is easier to have
them coexist in one kernel binary.
With this, the kvm_vcpu struct has a pointer to the kvmppc_vcpu_book3s
struct. The pointer to the kvmppc_book3s_shadow_vcpu struct has moved
from the kvmppc_vcpu_book3s struct to the kvm_vcpu struct, and is only
present for 32-bit, since it is only used for 32-bit.
Signed-off-by: Paul Mackerras <paulus@samba.org>
[agraf: squash in compile fix from Aneesh]
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently, PR KVM uses 4k pages for the host-side mappings of guest
memory, regardless of the host page size. When the host page size is
64kB, we might as well use 64k host page mappings for guest mappings
of 64kB and larger pages and for guest real-mode mappings. However,
the magic page has to remain a 4k page.
To implement this, we first add another flag bit to the guest VSID
values we use, to indicate that this segment is one where host pages
should be mapped using 64k pages. For segments with this bit set
we set the bits in the shadow SLB entry to indicate a 64k base page
size. When faulting in host HPTEs for this segment, we make them
64k HPTEs instead of 4k. We record the pagesize in struct hpte_cache
for use when invalidating the HPTE.
For now we restrict the segment containing the magic page (if any) to
4k pages. It should be possible to lift this restriction in future
by ensuring that the magic 4k page is appropriately positioned within
a host 64k page.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This adds the code to interpret 64k HPTEs in the guest hashed page
table (HPT), 64k SLB entries, and to tell the guest about 64k pages
in kvm_vm_ioctl_get_smmu_info(). Guest 64k pages are still shadowed
by 4k pages.
This also adds another hash table to the four we have already in
book3s_mmu_hpte.c to allow us to find all the PTEs that we have
instantiated that match a given 64k guest page.
The tlbie instruction changed starting with POWER6 to use a bit in
the RB operand to indicate large page invalidations, and to use other
RB bits to indicate the base and actual page sizes and the segment
size. 64k pages came in slightly earlier, with POWER5++.
We use one bit in vcpu->arch.hflags to indicate that the emulated
cpu supports 64k pages, and another to indicate that it has the new
tlbie definition.
The KVM_PPC_GET_SMMU_INFO ioctl presents a bit of a problem, because
the MMU capabilities depend on which CPU model we're emulating, but it
is a VM ioctl not a VCPU ioctl and therefore doesn't get passed a VCPU
fd. In addition, commonly-used userspace (QEMU) calls it before
setting the PVR for any VCPU. Therefore, as a best effort we look at
the first vcpu in the VM and return 64k pages or not depending on its
capabilities. We also make the PVR default to the host PVR on recent
CPUs that support 1TB segments (and therefore multiple page sizes as
well) so that KVM_PPC_GET_SMMU_INFO will include 64k page and 1TB
segment support on those CPUs.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently PR-style KVM keeps the volatile guest register values
(R0 - R13, CR, LR, CTR, XER, PC) in a shadow_vcpu struct rather than
the main kvm_vcpu struct. For 64-bit, the shadow_vcpu exists in two
places, a kmalloc'd struct and in the PACA, and it gets copied back
and forth in kvmppc_core_vcpu_load/put(), because the real-mode code
can't rely on being able to access the kmalloc'd struct.
This changes the code to copy the volatile values into the shadow_vcpu
as one of the last things done before entering the guest. Similarly
the values are copied back out of the shadow_vcpu to the kvm_vcpu
immediately after exiting the guest. We arrange for interrupts to be
still disabled at this point so that we can't get preempted on 64-bit
and end up copying values from the wrong PACA.
This means that the accessor functions in kvm_book3s.h for these
registers are greatly simplified, and are same between PR and HV KVM.
In places where accesses to shadow_vcpu fields are now replaced by
accesses to the kvm_vcpu, we can also remove the svcpu_get/put pairs.
Finally, on 64-bit, we don't need the kmalloc'd struct at all any more.
With this, the time to read the PVR one million times in a loop went
from 567.7ms to 575.5ms (averages of 6 values), an increase of about
1.4% for this worse-case test for guest entries and exits. The
standard deviation of the measurements is about 11ms, so the
difference is only marginally significant statistically.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This adds the ability to have a separate LPCR (Logical Partitioning
Control Register) value relating to a guest for each virtual core,
rather than only having a single value for the whole VM. This
corresponds to what real POWER hardware does, where there is a LPCR
per CPU thread but most of the fields are required to have the same
value on all active threads in a core.
The per-virtual-core LPCR can be read and written using the
GET/SET_ONE_REG interface. Userspace can can only modify the
following fields of the LPCR value:
DPFD Default prefetch depth
ILE Interrupt little-endian
TC Translation control (secondary HPT hash group search disable)
We still maintain a per-VM default LPCR value in kvm->arch.lpcr, which
contains bits relating to memory management, i.e. the Virtualized
Partition Memory (VPM) bits and the bits relating to guest real mode.
When this default value is updated, the update needs to be propagated
to the per-vcore values, so we add a kvmppc_update_lpcr() helper to do
that.
Signed-off-by: Paul Mackerras <paulus@samba.org>
[agraf: fix whitespace]
Signed-off-by: Alexander Graf <agraf@suse.de>
It turns out that if we exit the guest due to a hcall instruction (sc 1),
and the loading of the instruction in the guest exit path fails for any
reason, the call to kvmppc_ld() in kvmppc_get_last_inst() fetches the
instruction after the hcall instruction rather than the hcall itself.
This in turn means that the instruction doesn't get recognized as an
hcall in kvmppc_handle_exit_pr() but gets passed to the guest kernel
as a sc instruction. That usually results in the guest kernel getting
a return code of 38 (ENOSYS) from an hcall, which often triggers a
BUG_ON() or other failure.
This fixes the problem by adding a new variant of kvmppc_get_last_inst()
called kvmppc_get_last_sc(), which fetches the instruction if necessary
from pc - 4 rather than pc.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
With this, the guest can use 1TB segments as well as 256MB segments.
Since we now have the situation where a single emulated guest segment
could correspond to multiple shadow segments (as the shadow segments
are still 256MB segments), this adds a new kvmppc_mmu_flush_segment()
to scan for all shadow segments that need to be removed.
This restructures the guest HPT (hashed page table) lookup code to
use the correct hashing and matching functions for HPTEs within a
1TB segment. We use the standard hpt_hash() function instead of
open-coding the hash calculation, and we use HPTE_V_COMPARE() with
an AVPN value that has the B (segment size) field included. The
calculation of avpn is done a little earlier since it doesn't change
in the loop starting at the do_second label.
The computation in kvmppc_mmu_book3s_64_esid_to_vsid() changes so that
it returns a 256MB VSID even if the guest SLB entry is a 1TB entry.
This is because the users of this function are creating 256MB SLB
entries. We set a new VSID_1T flag so that entries created from 1T
segments don't collide with entries from 256MB segments.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This adds in-kernel emulation of the XICS (eXternal Interrupt
Controller Specification) interrupt controller specified by PAPR, for
both HV and PR KVM guests.
The XICS emulation supports up to 1048560 interrupt sources.
Interrupt source numbers below 16 are reserved; 0 is used to mean no
interrupt and 2 is used for IPIs. Internally these are represented in
blocks of 1024, called ICS (interrupt controller source) entities, but
that is not visible to userspace.
Each vcpu gets one ICP (interrupt controller presentation) entity,
used to store the per-vcpu state such as vcpu priority, pending
interrupt state, IPI request, etc.
This does not include any API or any way to connect vcpus to their
ICP state; that will be added in later patches.
This is based on an initial implementation by Michael Ellerman
<michael@ellerman.id.au> reworked by Benjamin Herrenschmidt and
Paul Mackerras.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
[agraf: fix typo, add dependency on !KVM_MPIC]
Signed-off-by: Alexander Graf <agraf@suse.de>
At present, the KVM_GET_DIRTY_LOG ioctl doesn't report modifications
done by the host to the virtual processor areas (VPAs) and dispatch
trace logs (DTLs) registered by the guest. This is because those
modifications are done either in real mode or in the host kernel
context, and in neither case does the access go through the guest's
HPT, and thus no change (C) bit gets set in the guest's HPT.
However, the changes done by the host do need to be tracked so that
the modified pages get transferred when doing live migration. In
order to track these modifications, this adds a dirty flag to the
struct representing the VPA/DTL areas, and arranges to set the flag
when the VPA/DTL gets modified by the host. Then, when we are
collecting the dirty log, we also check the dirty flags for the
VPA and DTL for each vcpu and set the relevant bit in the dirty log
if necessary. Doing this also means we now need to keep track of
the guest physical address of the VPA/DTL areas.
So as not to lose track of modifications to a VPA/DTL area when it gets
unregistered, or when a new area gets registered in its place, we need
to transfer the dirty state to the rmap chain. This adds code to
kvmppc_unpin_guest_page() to do that if the area was dirty. To simplify
that code, we now require that all VPA, DTL and SLB shadow buffer areas
fit within a single host page. Guests already comply with this
requirement because pHyp requires that these areas not cross a 4k
boundary.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
