Whenever a vcore state is VCORE_PREEMPT we need to be counting stolen
time for it. This currently isn't the case when we have a vcore that
no longer has any runnable threads in it but still has a runner task,
so we do an explicit call to kvmppc_core_start_stolen() in that case.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
When a vcore gets preempted, we put it on the preempted vcore list for
the current CPU. The runner task then calls schedule() and comes back
some time later and takes itself off the list. We need to be careful
to lock the list that it was put onto, which may not be the list for the
current CPU since the runner task may have moved to another CPU.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This builds on the ability to run more than one vcore on a physical
core by using the micro-threading (split-core) modes of the POWER8
chip. Previously, only vcores from the same VM could be run together,
and (on POWER8) only if they had just one thread per core. With the
ability to split the core on guest entry and unsplit it on guest exit,
we can run up to 8 vcpu threads from up to 4 different VMs, and we can
run multiple vcores with 2 or 4 vcpus per vcore.
Dynamic micro-threading is only available if the static configuration
of the cores is whole-core mode (unsplit), and only on POWER8.
To manage this, we introduce a new kvm_split_mode struct which is
shared across all of the subcores in the core, with a pointer in the
paca on each thread. In addition we extend the core_info struct to
have information on each subcore. When deciding whether to add a
vcore to the set already on the core, we now have two possibilities:
(a) piggyback the vcore onto an existing subcore, or (b) start a new
subcore.
Currently, when any vcpu needs to exit the guest and switch to host
virtual mode, we interrupt all the threads in all subcores and switch
the core back to whole-core mode. It may be possible in future to
allow some of the subcores to keep executing in the guest while
subcore 0 switches to the host, but that is not implemented in this
patch.
This adds a module parameter called dynamic_mt_modes which controls
which micro-threading (split-core) modes the code will consider, as a
bitmap. In other words, if it is 0, no micro-threading mode is
considered; if it is 2, only 2-way micro-threading is considered; if
it is 4, only 4-way, and if it is 6, both 2-way and 4-way
micro-threading mode will be considered. The default is 6.
With this, we now have secondary threads which are the primary thread
for their subcore and therefore need to do the MMU switch. These
threads will need to be started even if they have no vcpu to run, so
we use the vcore pointer in the PACA rather than the vcpu pointer to
trigger them.
It is now possible for thread 0 to find that an exit has been
requested before it gets to switch the subcore state to the guest. In
that case we haven't added the guest's timebase offset to the
timebase, so we need to be careful not to subtract the offset in the
guest exit path. In fact we just skip the whole path that switches
back to host context, since we haven't switched to the guest context.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
When running a virtual core of a guest that is configured with fewer
threads per core than the physical cores have, the extra physical
threads are currently unused. This makes it possible to use them to
run one or more other virtual cores from the same guest when certain
conditions are met. This applies on POWER7, and on POWER8 to guests
with one thread per virtual core. (It doesn't apply to POWER8 guests
with multiple threads per vcore because they require a 1-1 virtual to
physical thread mapping in order to be able to use msgsndp and the
TIR.)
The idea is that we maintain a list of preempted vcores for each
physical cpu (i.e. each core, since the host runs single-threaded).
Then, when a vcore is about to run, it checks to see if there are
any vcores on the list for its physical cpu that could be
piggybacked onto this vcore's execution. If so, those additional
vcores are put into state VCORE_PIGGYBACK and their runnable VCPU
threads are started as well as the original vcore, which is called
the master vcore.
After the vcores have exited the guest, the extra ones are put back
onto the preempted list if any of their VCPUs are still runnable and
not idle.
This means that vcpu->arch.ptid is no longer necessarily the same as
the physical thread that the vcpu runs on. In order to make it easier
for code that wants to send an IPI to know which CPU to target, we
now store that in a new field in struct vcpu_arch, called thread_cpu.
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Tested-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
When compiling the KVM code for POWER with "make C=1", sparse
complains about functions missing proper prototypes and a 64-bit
constant missing the ULL prefix. Let's fix this by making the
functions static or by including the proper header with the
prototypes, and by appending a ULL prefix to the constant
PPC_MPPE_ADDRESS_MASK.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Alexander 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: Radim Krcmar <rkrcmar@redhat.com>
Signed-off-by: Paolo 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: Takuya Yoshikawa <yoshikawa_takuya_b1@lab.ntt.co.jp>
Reviewed-by: Radim Krcmar <rkrcmar@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
kvm_memslots provides lockdep checking. Use it consistently instead of
explicit dereferencing of kvm->memslots.
Reviewed-by: Radim Krcmar <rkrcmar@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This fixes a regression introduced in commit 25fedfca94, "KVM: PPC:
Book3S HV: Move vcore preemption point up into kvmppc_run_vcpu", which
leads to a user-triggerable oops.
In the case where we try to run a vcore on a physical core that is
not in single-threaded mode, or the vcore has too many threads for
the physical core, we iterate the list of runnable vcpus to make
each one return an EBUSY error to userspace. Since this involves
taking each vcpu off the runnable_threads list for the vcore, we
need to use list_for_each_entry_safe rather than list_for_each_entry
to traverse the list. Otherwise the kernel will crash with an oops
message like this:
Unable to handle kernel paging request for data at address 0x000fff88
Faulting instruction address: 0xd00000001e635dc8
Oops: Kernel access of bad area, sig: 11 [#2]
SMP NR_CPUS=1024 NUMA PowerNV
...
CPU: 48 PID: 91256 Comm: qemu-system-ppc Tainted: G D 3.18.0 #1
task: c00000274e507500 ti: c0000027d1924000 task.ti: c0000027d1924000
NIP: d00000001e635dc8 LR: d00000001e635df8 CTR: c00000000011ba50
REGS: c0000027d19275b0 TRAP: 0300 Tainted: G D (3.18.0)
MSR: 9000000000009033 <SF,HV,EE,ME,IR,DR,RI,LE> CR: 22002824 XER: 00000000
CFAR: c000000000008468 DAR: 00000000000fff88 DSISR: 40000000 SOFTE: 1
GPR00: d00000001e635df8 c0000027d1927830 d00000001e64c850 0000000000000001
GPR04: 0000000000000001 0000000000000001 0000000000000000 0000000000000000
GPR08: 0000000000200200 0000000000000000 0000000000000000 d00000001e63e588
GPR12: 0000000000002200 c000000007dbc800 c000000fc7800000 000000000000000a
GPR16: fffffffffffffffc c000000fd5439690 c000000fc7801c98 0000000000000001
GPR20: 0000000000000003 c0000027d1927aa8 c000000fd543b348 c000000fd543b350
GPR24: 0000000000000000 c000000fa57f0000 0000000000000030 0000000000000000
GPR28: fffffffffffffff0 c000000fd543b328 00000000000fe468 c000000fd543b300
NIP [d00000001e635dc8] kvmppc_run_core+0x198/0x17c0 [kvm_hv]
LR [d00000001e635df8] kvmppc_run_core+0x1c8/0x17c0 [kvm_hv]
Call Trace:
[c0000027d1927830] [d00000001e635df8] kvmppc_run_core+0x1c8/0x17c0 [kvm_hv] (unreliable)
[c0000027d1927a30] [d00000001e638350] kvmppc_vcpu_run_hv+0x5b0/0xdd0 [kvm_hv]
[c0000027d1927b70] [d00000001e510504] kvmppc_vcpu_run+0x44/0x60 [kvm]
[c0000027d1927ba0] [d00000001e50d4a4] kvm_arch_vcpu_ioctl_run+0x64/0x170 [kvm]
[c0000027d1927be0] [d00000001e504be8] kvm_vcpu_ioctl+0x5e8/0x7a0 [kvm]
[c0000027d1927d40] [c0000000002d6720] do_vfs_ioctl+0x490/0x780
[c0000027d1927de0] [c0000000002d6ae4] SyS_ioctl+0xd4/0xf0
[c0000027d1927e30] [c000000000009358] syscall_exit+0x0/0x98
Instruction dump:
60000000 60420000 387e1b30 38800003 38a00001 38c00000 480087d9 e8410018
ebde1c98 7fbdf040 3bdee368 419e0048 <813e1b20> 939e1b18 2f890001 409effcc
---[ end trace 8cdf50251cca6680 ]---
Fixes: 25fedfca94
Signed-off-by: Paul Mackerras <paulus@samba.org>
Reviewed-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This uses msgsnd where possible for signalling other threads within
the same core on POWER8 systems, rather than IPIs through the XICS
interrupt controller. This includes waking secondary threads to run
the guest, the interrupts generated by the virtual XICS, and the
interrupts to bring the other threads out of the guest when exiting.
Aggregated statistics from debugfs across vcpus for a guest with 32
vcpus, 8 threads/vcore, running on a POWER8, show this before the
change:
rm_entry: 3387.6ns (228 - 86600, 1008969 samples)
rm_exit: 4561.5ns (12 - 3477452, 1009402 samples)
rm_intr: 1660.0ns (12 - 553050, 3600051 samples)
and this after the change:
rm_entry: 3060.1ns (212 - 65138, 953873 samples)
rm_exit: 4244.1ns (12 - 9693408, 954331 samples)
rm_intr: 1342.3ns (12 - 1104718, 3405326 samples)
for a test of booting Fedora 20 big-endian to the login prompt.
The time taken for a H_PROD hcall (which is handled in the host
kernel) went down from about 35 microseconds to about 16 microseconds
with this change.
The noinline added to kvmppc_run_core turned out to be necessary for
good performance, at least with gcc 4.9.2 as packaged with Fedora 21
and a little-endian POWER8 host.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently, the entry_exit_count field in the kvmppc_vcore struct
contains two 8-bit counts, one of the threads that have started entering
the guest, and one of the threads that have started exiting the guest.
This changes it to an entry_exit_map field which contains two bitmaps
of 8 bits each. The advantage of doing this is that it gives us a
bitmap of which threads need to be signalled when exiting the guest.
That means that we no longer need to use the trick of setting the
HDEC to 0 to pull the other threads out of the guest, which led in
some cases to a spurious HDEC interrupt on the next guest entry.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
We can tell when a secondary thread has finished running a guest by
the fact that it clears its kvm_hstate.kvm_vcpu pointer, so there
is no real need for the nap_count field in the kvmppc_vcore struct.
This changes kvmppc_wait_for_nap to poll the kvm_hstate.kvm_vcpu
pointers of the secondary threads rather than polling vc->nap_count.
Besides reducing the size of the kvmppc_vcore struct by 8 bytes,
this also means that we can tell which secondary threads have got
stuck and thus print a more informative error message.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Rather than calling cond_resched() in kvmppc_run_core() before doing
the post-processing for the vcpus that we have just run (that is,
calling kvmppc_handle_exit_hv(), kvmppc_set_timer(), etc.), we now do
that post-processing before calling cond_resched(), and that post-
processing is moved out into its own function, post_guest_process().
The reschedule point is now in kvmppc_run_vcpu() and we define a new
vcore state, VCORE_PREEMPT, to indicate that that the vcore's runner
task is runnable but not running. (Doing the reschedule with the
vcore in VCORE_INACTIVE state would be bad because there are potentially
other vcpus waiting for the runner in kvmppc_wait_for_exec() which
then wouldn't get woken up.)
Also, we make use of the handy cond_resched_lock() function, which
unlocks and relocks vc->lock for us around the reschedule.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Previously, if kvmppc_run_core() was running a VCPU that needed a VPA
update (i.e. one of its 3 virtual processor areas needed to be pinned
in memory so the host real mode code can update it on guest entry and
exit), we would drop the vcore lock and do the update there and then.
Future changes will make it inconvenient to drop the lock, so instead
we now remove it from the list of runnable VCPUs and wake up its
VCPU task. This will have the effect that the VCPU task will exit
kvmppc_run_vcpu(), go around the do loop in kvmppc_vcpu_run_hv(), and
re-enter kvmppc_run_vcpu(), whereupon it will do the necessary call
to kvmppc_update_vpas() and then rejoin the vcore.
The one complication is that the runner VCPU (whose VCPU task is the
current task) might be one of the ones that gets removed from the
runnable list. In that case we just return from kvmppc_run_core()
and let the code in kvmppc_run_vcpu() wake up another VCPU task to be
the runner if necessary.
This all means that the VCORE_STARTING state is no longer used, so we
remove it.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This reads the timebase at various points in the real-mode guest
entry/exit code and uses that to accumulate total, minimum and
maximum time spent in those parts of the code. Currently these
times are accumulated per vcpu in 5 parts of the code:
* rm_entry - time taken from the start of kvmppc_hv_entry() until
just before entering the guest.
* rm_intr - time from when we take a hypervisor interrupt in the
guest until we either re-enter the guest or decide to exit to the
host. This includes time spent handling hcalls in real mode.
* rm_exit - time from when we decide to exit the guest until the
return from kvmppc_hv_entry().
* guest - time spend in the guest
* cede - time spent napping in real mode due to an H_CEDE hcall
while other threads in the same vcore are active.
These times are exposed in debugfs in a directory per vcpu that
contains a file called "timings". This file contains one line for
each of the 5 timings above, with the name followed by a colon and
4 numbers, which are the count (number of times the code has been
executed), the total time, the minimum time, and the maximum time,
all in nanoseconds.
The overhead of the extra code amounts to about 30ns for an hcall that
is handled in real mode (e.g. H_SET_DABR), which is about 25%. Since
production environments may not wish to incur this overhead, the new
code is conditional on a new config symbol,
CONFIG_KVM_BOOK3S_HV_EXIT_TIMING.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This creates a debugfs directory for each HV guest (assuming debugfs
is enabled in the kernel config), and within that directory, a file
by which the contents of the guest's HPT (hashed page table) can be
read. The directory is named vmnnnn, where nnnn is the PID of the
process that created the guest. The file is named "htab". This is
intended to help in debugging problems in the host's management
of guest memory.
The contents of the file consist of a series of lines like this:
3f48 4000d032bf003505 0000000bd7ff1196 00000003b5c71196
The first field is the index of the entry in the HPT, the second and
third are the HPT entry, so the third entry contains the real page
number that is mapped by the entry if the entry's valid bit is set.
The fourth field is the guest's view of the second doubleword of the
entry, so it contains the guest physical address. (The format of the
second through fourth fields are described in the Power ISA and also
in arch/powerpc/include/asm/mmu-hash64.h.)
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
We don't support real-mode areas now that 970 support is removed.
Remove the remaining details of rma from the code. Also rename
rma_setup_done to hpte_setup_done to better reflect the changes.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
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>
The VPA (virtual processor area) is defined by PAPR and is therefore
big-endian, so we need a be32_to_cpu when reading it in
kvmppc_get_yield_count(). Without this, H_CONFER always fails on a
little-endian host, causing SMP guests to waste time spinning on
spinlocks.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently the H_CONFER hcall is implemented in kernel virtual mode,
meaning that whenever a guest thread does an H_CONFER, all the threads
in that virtual core have to exit the guest. This is bad for
performance because it interrupts the other threads even if they
are doing useful work.
The H_CONFER hcall is called by a guest VCPU when it is spinning on a
spinlock and it detects that the spinlock is held by a guest VCPU that
is currently not running on a physical CPU. The idea is to give this
VCPU's time slice to the holder VCPU so that it can make progress
towards releasing the lock.
To avoid having the other threads exit the guest unnecessarily,
we add a real-mode implementation of H_CONFER that checks whether
the other threads are doing anything. If all the other threads
are idle (i.e. in H_CEDE) or trying to confer (i.e. in H_CONFER),
it returns H_TOO_HARD which causes a guest exit and allows the
H_CONFER to be handled in virtual mode.
Otherwise it spins for a short time (up to 10 microseconds) to give
other threads the chance to observe that this thread is trying to
confer. The spin loop also terminates when any thread exits the guest
or when all other threads are idle or trying to confer. If the
timeout is reached, the H_CONFER returns H_SUCCESS. In this case the
guest VCPU will recheck the spinlock word and most likely call
H_CONFER again.
This also improves the implementation of the H_CONFER virtual mode
handler. If the VCPU is part of a virtual core (vcore) which is
runnable, there will be a 'runner' VCPU which has taken responsibility
for running the vcore. In this case we yield to the runner VCPU
rather than the target VCPU.
We also introduce a check on the target VCPU's yield count: if it
differs from the yield count passed to H_CONFER, the target VCPU
has run since H_CONFER was called and may have already released
the lock. This check is required by PAPR.
Signed-off-by: Sam Bobroff <sam.bobroff@au1.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
There are two ways in which a guest instruction can be obtained from
the guest in the guest exit code in book3s_hv_rmhandlers.S. If the
exit was caused by a Hypervisor Emulation interrupt (i.e. an illegal
instruction), the offending instruction is in the HEIR register
(Hypervisor Emulation Instruction Register). If the exit was caused
by a load or store to an emulated MMIO device, we load the instruction
from the guest by turning data relocation on and loading the instruction
with an lwz instruction.
Unfortunately, in the case where the guest has opposite endianness to
the host, these two methods give results of different endianness, but
both get put into vcpu->arch.last_inst. The HEIR value has been loaded
using guest endianness, whereas the lwz will load the instruction using
host endianness. The rest of the code that uses vcpu->arch.last_inst
assumes it was loaded using host endianness.
To fix this, we define a new vcpu field to store the HEIR value. Then,
in kvmppc_handle_exit_hv(), we transfer the value from this new field to
vcpu->arch.last_inst, doing a byte-swap if the guest and host endianness
differ.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
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>
This patch adds trace points in the guest entry and exit code and also
for exceptions handled by the host in kernel mode - hypercalls and page
faults. The new events are added to /sys/kernel/debug/tracing/events
under a new subsystem called kvm_hv.
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Suresh Warrier <warrier@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently the calculations of stolen time for PPC Book3S HV guests
uses fields in both the vcpu struct and the kvmppc_vcore struct. The
fields in the kvmppc_vcore struct are protected by the
vcpu->arch.tbacct_lock of the vcpu that has taken responsibility for
running the virtual core. This works correctly but confuses lockdep,
because it sees that the code takes the tbacct_lock for a vcpu in
kvmppc_remove_runnable() and then takes another vcpu's tbacct_lock in
vcore_stolen_time(), and it thinks there is a possibility of deadlock,
causing it to print reports like this:
=============================================
[ INFO: possible recursive locking detected ]
3.18.0-rc7-kvm-00016-g8db4bc6 #89 Not tainted
---------------------------------------------
qemu-system-ppc/6188 is trying to acquire lock:
(&(&vcpu->arch.tbacct_lock)->rlock){......}, at: [<d00000000ecb1fe8>] .vcore_stolen_time+0x48/0xd0 [kvm_hv]
but task is already holding lock:
(&(&vcpu->arch.tbacct_lock)->rlock){......}, at: [<d00000000ecb25a0>] .kvmppc_remove_runnable.part.3+0x30/0xd0 [kvm_hv]
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(&(&vcpu->arch.tbacct_lock)->rlock);
lock(&(&vcpu->arch.tbacct_lock)->rlock);
*** DEADLOCK ***
May be due to missing lock nesting notation
3 locks held by qemu-system-ppc/6188:
#0: (&vcpu->mutex){+.+.+.}, at: [<d00000000eb93f98>] .vcpu_load+0x28/0xe0 [kvm]
#1: (&(&vcore->lock)->rlock){+.+...}, at: [<d00000000ecb41b0>] .kvmppc_vcpu_run_hv+0x530/0x1530 [kvm_hv]
#2: (&(&vcpu->arch.tbacct_lock)->rlock){......}, at: [<d00000000ecb25a0>] .kvmppc_remove_runnable.part.3+0x30/0xd0 [kvm_hv]
stack backtrace:
CPU: 40 PID: 6188 Comm: qemu-system-ppc Not tainted 3.18.0-rc7-kvm-00016-g8db4bc6 #89
Call Trace:
[c000000b2754f3f0] [c000000000b31b6c] .dump_stack+0x88/0xb4 (unreliable)
[c000000b2754f470] [c0000000000faeb8] .__lock_acquire+0x1878/0x2190
[c000000b2754f600] [c0000000000fbf0c] .lock_acquire+0xcc/0x1a0
[c000000b2754f6d0] [c000000000b2954c] ._raw_spin_lock_irq+0x4c/0x70
[c000000b2754f760] [d00000000ecb1fe8] .vcore_stolen_time+0x48/0xd0 [kvm_hv]
[c000000b2754f7f0] [d00000000ecb25b4] .kvmppc_remove_runnable.part.3+0x44/0xd0 [kvm_hv]
[c000000b2754f880] [d00000000ecb43ec] .kvmppc_vcpu_run_hv+0x76c/0x1530 [kvm_hv]
[c000000b2754f9f0] [d00000000eb9f46c] .kvmppc_vcpu_run+0x2c/0x40 [kvm]
[c000000b2754fa60] [d00000000eb9c9a4] .kvm_arch_vcpu_ioctl_run+0x54/0x160 [kvm]
[c000000b2754faf0] [d00000000eb94538] .kvm_vcpu_ioctl+0x498/0x760 [kvm]
[c000000b2754fcb0] [c000000000267eb4] .do_vfs_ioctl+0x444/0x770
[c000000b2754fd90] [c0000000002682a4] .SyS_ioctl+0xc4/0xe0
[c000000b2754fe30] [c0000000000092e4] syscall_exit+0x0/0x98
In order to make the locking easier to analyse, we change the code to
use a spinlock in the kvmppc_vcore struct to protect the stolen_tb and
preempt_tb fields. This lock needs to be an irq-safe lock since it is
used in the kvmppc_core_vcpu_load_hv() and kvmppc_core_vcpu_put_hv()
functions, which are called with the scheduler rq lock held, which is
an irq-safe lock.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
The kvmppc_vcore_blocked() code does not check for the wait condition
after putting the process on the wait queue. This means that it is
possible for an external interrupt to become pending, but the vcpu to
remain asleep until the next decrementer interrupt. The fix is to
make one last check for pending exceptions and ceded state before
calling schedule().
Signed-off-by: Suresh Warrier <warrier@linux.vnet.ibm.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This patch adds kernel side support for software breakpoint.
Design is that, by using an illegal instruction, we trap to hypervisor
via Emulation Assistance interrupt, where we check for the illegal instruction
and accordingly we return to Host or Guest. Patch also adds support for
software breakpoint in PR KVM.
Signed-off-by: Madhavan Srinivasan <maddy@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
Since the guest can read the machine's PVR (Processor Version Register)
directly and see the real value, we should disallow userspace from
setting any value for the guest's PVR other than the real host value.
Therefore this makes kvm_arch_vcpu_set_sregs_hv() check the supplied
PVR value and return an error if it is different from the host value,
which has been put into vcpu->arch.pvr at vcpu creation time.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Occasional failures have been seen with split-core mode and migration
where the message "KVM: couldn't grab cpu" appears. This increases
the length of time that we wait from 1ms to 10ms, which seems to
work around the issue.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
When calculating the lower bits of AVA field, use the shift
count based on the base page size. Also add the missing segment
size and remove stale comment.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
The POWER8 processor has a Micro Partition Prefetch Engine, which is
a fancy way of saying "has way to store and load contents of L2 or
L2+MRU way of L3 cache". We initiate the storing of the log (list of
addresses) using the logmpp instruction and start restore by writing
to a SPR.
The logmpp instruction takes parameters in a single 64bit register:
- starting address of the table to store log of L2/L2+L3 cache contents
- 32kb for L2
- 128kb for L2+L3
- Aligned relative to maximum size of the table (32kb or 128kb)
- Log control (no-op, L2 only, L2 and L3, abort logout)
We should abort any ongoing logging before initiating one.
To initiate restore, we write to the MPPR SPR. The format of what to write
to the SPR is similar to the logmpp instruction parameter:
- starting address of the table to read from (same alignment requirements)
- table size (no data, until end of table)
- prefetch rate (from fastest possible to slower. about every 8, 16, 24 or
32 cycles)
The idea behind loading and storing the contents of L2/L3 cache is to
reduce memory latency in a system that is frequently swapping vcores on
a physical CPU.
The best case scenario for doing this is when some vcores are doing very
cache heavy workloads. The worst case is when they have about 0 cache hits,
so we just generate needless memory operations.
This implementation just does L2 store/load. In my benchmarks this proves
to be useful.
Benchmark 1:
- 16 core POWER8
- 3x Ubuntu 14.04LTS guests (LE) with 8 VCPUs each
- No split core/SMT
- two guests running sysbench memory test.
sysbench --test=memory --num-threads=8 run
- one guest running apache bench (of default HTML page)
ab -n 490000 -c 400 http://localhost/
This benchmark aims to measure performance of real world application (apache)
where other guests are cache hot with their own workloads. The sysbench memory
benchmark does pointer sized writes to a (small) memory buffer in a loop.
In this benchmark with this patch I can see an improvement both in requests
per second (~5%) and in mean and median response times (again, about 5%).
The spread of minimum and maximum response times were largely unchanged.
benchmark 2:
- Same VM config as benchmark 1
- all three guests running sysbench memory benchmark
This benchmark aims to see if there is a positive or negative affect to this
cache heavy benchmark. Although due to the nature of the benchmark (stores) we
may not see a difference in performance, but rather hopefully an improvement
in consistency of performance (when vcore switched in, don't have to wait
many times for cachelines to be pulled in)
The results of this benchmark are improvements in consistency of performance
rather than performance itself. With this patch, the few outliers in duration
go away and we get more consistent performance in each guest.
benchmark 3:
- same 3 guests and CPU configuration as benchmark 1 and 2.
- two idle guests
- 1 guest running STREAM benchmark
This scenario also saw performance improvement with this patch. On Copy and
Scale workloads from STREAM, I got 5-6% improvement with this patch. For
Add and triad, it was around 10% (or more).
benchmark 4:
- same 3 guests as previous benchmarks
- two guests running sysbench --memory, distinctly different cache heavy
workload
- one guest running STREAM benchmark.
Similar improvements to benchmark 3.
benchmark 5:
- 1 guest, 8 VCPUs, Ubuntu 14.04
- Host configured with split core (SMT8, subcores-per-core=4)
- STREAM benchmark
In this benchmark, we see a 10-20% performance improvement across the board
of STREAM benchmark results with this patch.
Based on preliminary investigation and microbenchmarks
by Prerna Saxena <prerna@linux.vnet.ibm.com>
Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
No code changes, just split it out to a function so that with the addition
of micro partition prefetch buffer allocation (in subsequent patch) looks
neater and doesn't require excessive indentation.
Signed-off-by: Stewart Smith <stewart@linux.vnet.ibm.com>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Unfortunately, the LPCR got defined as a 32-bit register in the
one_reg interface. This is unfortunate because KVM allows userspace
to control the DPFD (default prefetch depth) field, which is in the
upper 32 bits. The result is that DPFD always get set to 0, which
reduces performance in the guest.
We can't just change KVM_REG_PPC_LPCR to be a 64-bit register ID,
since that would break existing userspace binaries. Instead we define
a new KVM_REG_PPC_LPCR_64 id which is 64-bit. Userspace can still use
the old KVM_REG_PPC_LPCR id, but it now only modifies those fields in
the bottom 32 bits that userspace can modify (ILE, TC and AIL).
If userspace uses the new KVM_REG_PPC_LPCR_64 id, it can modify DPFD
as well.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Cc: stable@vger.kernel.org
Signed-off-by: Alexander 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: Alexander 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: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
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>
Writing to IC is not allowed in the privileged mode.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander 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: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
[agraf: fix compile error]
Signed-off-by: Alexander Graf <agraf@suse.de>
Pull powerpc updates from Ben Herrenschmidt:
"Here is the bulk of the powerpc changes for this merge window. It got
a bit delayed in part because I wasn't paying attention, and in part
because I discovered I had a core PCI change without a PCI maintainer
ack in it. Bjorn eventually agreed it was ok to merge it though we'll
probably improve it later and I didn't want to rebase to add his ack.
There is going to be a bit more next week, essentially fixes that I
still want to sort through and test.
The biggest item this time is the support to build the ppc64 LE kernel
with our new v2 ABI. We previously supported v2 userspace but the
kernel itself was a tougher nut to crack. This is now sorted mostly
thanks to Anton and Rusty.
We also have a fairly big series from Cedric that add support for
64-bit LE zImage boot wrapper. This was made harder by the fact that
traditionally our zImage wrapper was always 32-bit, but our new LE
toolchains don't really support 32-bit anymore (it's somewhat there
but not really "supported") so we didn't want to rely on it. This
meant more churn that just endian fixes.
This brings some more LE bits as well, such as the ability to run in
LE mode without a hypervisor (ie. under OPAL firmware) by doing the
right OPAL call to reinitialize the CPU to take HV interrupts in the
right mode and the usual pile of endian fixes.
There's another series from Gavin adding EEH improvements (one day we
*will* have a release with less than 20 EEH patches, I promise!).
Another highlight is the support for the "Split core" functionality on
P8 by Michael. This allows a P8 core to be split into "sub cores" of
4 threads which allows the subcores to run different guests under KVM
(the HW still doesn't support a partition per thread).
And then the usual misc bits and fixes ..."
[ Further delayed by gmail deciding that BenH is a dirty spammer.
Google knows. ]
* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (155 commits)
powerpc/powernv: Add missing include to LPC code
selftests/powerpc: Test the THP bug we fixed in the previous commit
powerpc/mm: Check paca psize is up to date for huge mappings
powerpc/powernv: Pass buffer size to OPAL validate flash call
powerpc/pseries: hcall functions are exported to modules, need _GLOBAL_TOC()
powerpc: Exported functions __clear_user and copy_page use r2 so need _GLOBAL_TOC()
powerpc/powernv: Set memory_block_size_bytes to 256MB
powerpc: Allow ppc_md platform hook to override memory_block_size_bytes
powerpc/powernv: Fix endian issues in memory error handling code
powerpc/eeh: Skip eeh sysfs when eeh is disabled
powerpc: 64bit sendfile is capped at 2GB
powerpc/powernv: Provide debugfs access to the LPC bus via OPAL
powerpc/serial: Use saner flags when creating legacy ports
powerpc: Add cpu family documentation
powerpc/xmon: Fix up xmon format strings
powerpc/powernv: Add calls to support little endian host
powerpc: Document sysfs DSCR interface
powerpc: Fix regression of per-CPU DSCR setting
powerpc: Split __SYSFS_SPRSETUP macro
arch: powerpc/fadump: Cleaning up inconsistent NULL checks
...
On recent IBM Power CPUs, while the hashed page table is looked up using
the page size from the segmentation hardware (i.e. the SLB), it is
possible to have the HPT entry indicate a larger page size. Thus for
example it is possible to put a 16MB page in a 64kB segment, but since
the hash lookup is done using a 64kB page size, it may be necessary to
put multiple entries in the HPT for a single 16MB page. This
capability is called mixed page-size segment (MPSS). With MPSS,
there are two relevant page sizes: the base page size, which is the
size used in searching the HPT, and the actual page size, which is the
size indicated in the HPT entry. [ Note that the actual page size is
always >= base page size ].
We use "ibm,segment-page-sizes" device tree node to advertise
the MPSS support to PAPR guest. The penc encoding indicates whether
we support a specific combination of base page size and actual
page size in the same segment. We also use the penc value in the
LP encoding of HPTE entry.
This patch exposes MPSS support to KVM guest by advertising the
feature via "ibm,segment-page-sizes". It also adds the necessary changes
to decode the base page size and the actual page size correctly from the
HPTE entry.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Alexander Graf <agraf@suse.de>
POWER8 introduces a new facility called the "Event Based Branch" facility.
It contains of a few registers that indicate where a guest should branch to
when a defined event occurs and it's in PR mode.
We don't want to really enable EBB as it will create a big mess with !PR guest
mode while hardware is in PR and we don't really emulate the PMU anyway.
So instead, let's just leave it at emulation of all its registers.
Signed-off-by: Alexander Graf <agraf@suse.de>
POWER8 implements a new register called TAR. This register has to be
enabled in FSCR and then from KVM's point of view is mere storage.
This patch enables the guest to use TAR.
Signed-off-by: Alexander Graf <agraf@suse.de>
POWER8 introduced a new interrupt type called "Facility unavailable interrupt"
which contains its status message in a new register called FSCR.
Handle these exits and try to emulate instructions for unhandled facilities.
Follow-on patches enable KVM to expose specific facilities into the guest.
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>
To support split core on POWER8 we need to modify various parts of the
KVM code to use threads_per_subcore instead of threads_per_core. On
systems that do not support split core threads_per_subcore ==
threads_per_core and these changes are a nop.
We use threads_per_subcore as the value reported by KVM_CAP_PPC_SMT.
This communicates to userspace that guests can only be created with
a value of threads_per_core that is less than or equal to the current
threads_per_subcore. This ensures that guests can only be created with a
thread configuration that we are able to run given the current split
core mode.
Although threads_per_subcore can change during the life of the system,
the commit that enables that will ensure that threads_per_subcore does
not change during the life of a KVM VM.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Acked-by: Alexander Graf <agraf@suse.de>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
As part of the support for split core on POWER8, we want to be able to
block splitting of the core while KVM VMs are active.
The logic to do that would be exactly the same as the code we currently
have for inhibiting onlining of secondaries.
Instead of adding an identical mechanism to block split core, rework the
secondary inhibit code to be a "HV KVM is active" check. We can then use
that in both the cpu hotplug code and the upcoming split core code.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Michael Neuling <mikey@neuling.org>
Acked-by: Alexander Graf <agraf@suse.de>
Acked-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
If an attempt is made to load the kvm-hv module on a machine which
doesn't have hypervisor mode available, return an ENODEV error,
which is the conventional thing to return to indicate that this
module is not applicable to the hardware of the current machine,
rather than EIO, which causes a warning to be printed.
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Scott Wood <scottwood@freescale.com>
This adds code to get/set_one_reg to read and write the new transactional
memory (TM) state.
Signed-off-by: Michael Neuling <mikey@neuling.org>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Acked-by: Scott Wood <scottwood@freescale.com>
