ACPI subsystem already has all logic in place the only
thing left to eject CPU is destroy it and ammend
present CPUs counter in CMOS, do so.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
First remove VCPU from exec loop and only then remove lapic.
Signed-off-by: Chen Fan <chen.fan.fnst@cn.fujitsu.com>
Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
Signed-off-by: Zhu Guihua <zhugh.fnst@cn.fujitsu.com>
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
instance_id is generated by last_used_id + 1 for a given device type
so for QEMU with 3 CPUs instance_id for APICs is a seti of [0, 1, 2]
When CPU in the middle is hot-removed and migration started
APICs with instance_ids 0 and 2 are transferred in migration stream.
However target starts with 2 CPUs and APICs' instance_ids are
generated from scratch [0, 1] hence migration fails with error
Unknown savevm section or instance 'apic' 2
Fix issue by manually registering APIC's vmsd with apic_id as
instance_id, in this case instance_id on target will always
match instance_id on source as apic_id is the same for a given
cpu instance.
Reported-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Callbacks will do necessary cleanups before APIC device is deleted
Signed-off-by: Chen Fan <chen.fan.fnst@cn.fujitsu.com>
Signed-off-by: Gu Zheng <guz.fnst@cn.fujitsu.com>
Signed-off-by: Zhu Guihua <zhugh.fnst@cn.fujitsu.com>
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
kvm-apic.io_memory memory region had its parent set to NULL at
memory_region_init_io() time, so it ended up as a child in
/unattached contaner.
As result when kvm-apic instance was deleted, the child property
/unattached/kvm-apic-msi[XXX] contained a reference to
kvm-apic.io_memory address which was freed as part of kvm-apic.
Do the same as 'apic' and make kvm-apic instance the owner
of the memory region so that it won't end up in /unattached
and gets cleanly released along with related kvm-apic instance.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
local_apics[] is sized to contain all APIC ID supported in xAPIC mode,
so use APIC ID as index in it instead of constantly increasing counter idx.
Fixes error "apic initialization failed" when a CPU hotplugged and
unplugged more times than there are free slots in local_apics[].
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Radim Krčmář <rkrcmar@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
MAX_APICS is only used by child 'apic' class and not
by its parent TYPE_APIC_COMMON or any other derived
class.
Move check into end user 'apic' class so it won't
get in the way of other APIC implementations
if they support more then MAX_APICS.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Radim Krčmář <rkrcmar@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
it returns a list of present/possible to hotplug CPU
objects with a list of properties to use with
device_add.
in PC case returned list would looks like:
-> { "execute": "query-hotpluggable-cpus" }
<- {"return": [
{
"type": "qemu64-x86_64-cpu", "vcpus-count": 1,
"props": {"core-id": 0, "socket-id": 1, "thread-id": 0}
},
{
"qom-path": "/machine/unattached/device[0]",
"type": "qemu64-x86_64-cpu", "vcpus-count": 1,
"props": {"core-id": 0, "socket-id": 0, "thread-id": 0}
}
]}
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
It will still allow us to use cpu_index as migration instance_id
since when CPUs are added contiguously (from the first to the last)
and removed in opposite order, cpu_index stays stable and it's
reproducible on destination side.
While there is work in progress to support migration when there
are holes in cpu_index range resulting from out-of-order plug or
unplug, this patch is intended as an interim solution until
cpu_index usage is cleaned up.
As result of this patch it would be possible to plug/unplug CPUs,
but in limited order that doesn't break migration.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Boot CPU is assumed to always present in QEMU code, so
untile that assumptions are gone, deny removal request,
In another words QEMU won't support BSP hot-unplug.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Consolidate possible_cpus array management in pc_cpu_plug() for
smp_cpus, coldplugged with -device and hotplugged with
device_add.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Currently present CPUs counter in CMOS only contains
smp_cpus (i.e. initial CPUs specified with -smp X) and
doesn't account for CPUs created with -device.
If VM is started with additional CPUs added with
-device, it will hang in BIOS waiting for condition
smp_cpus == counted_cpus
forever as counted_cpus will include -device CPUs as well
and be more than smp_cpus.
Make present CPUs counter in CMOS to count all CPUs
(initial and coldplugged with -device) by delaying
it to machine done time when it possible to count
CPUs added with -device.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
CPU added with device_add help won't have APIC ID set,
so set it according to socket/core/thread ids provided
with device_add command.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
object_property_add_child() silently fails with error that it can't
create duplicate propery 'apic' as we already have 'apic' property
registered for 'apic' feature. As result generic device_realize puts
apic into unattached container.
As it's programming error, abort if name collision happens in future
and fix property name for apic_state to 'lapic', this way apic is
a child of cpu instance.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
These are both stored in CPUID[EAX=7,EBX=0].ECX. KVM is going to
be able to emulate both (albeit with a performance loss in the case
of RDPID, which therefore will be in KVM_GET_EMULATED_CPUID rather
than KVM_GET_SUPPORTED_CPUID).
It's also possible to implement both in TCG, but this is for 2.8.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
These properties will be used by as address where to plug
CPU with help -device/device_add commands.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Custom apic-id setter/getter doesn't do any property specific
checks anymore, so clean it up and use more compact static
property DEFINE_PROP_UINT32 instead.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Machine code knows about all possible APIC IDs so use that
instead of hack which does O(n^2) complexity duplicate
checks, interating over global CPUs list.
As result duplicate check is done only once with O(log n) complexity.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
It will be reused in the next patch at pre_plug time
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Add the host-phys-bits boolean property, if true, take phys-bits
from the hosts physical bits value, overriding either the default
or the user specified value.
We can also use the value we read from the host to check the users
explicitly set value and warn them if it doesn't match.
Note:
a) We only read the hosts value in KVM mode (because on non-x86
we get an abort if we try)
b) We don't warn about trying to use host-phys-bits in TCG mode,
we just fall back to the TCG default. This allows the machine
type to set the host-phys-bits flag if it wants and then to
work in both TCG and KVM.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
It's reverse of apicid_from_topo_ids() and will be used in follow up
patches to fill in data structures for query-hotpluggable-cpus and
for user friendly error reporting.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Redo 9886e834 (target-i386: Require APIC ID to be explicitly set before
CPU realize) in another way that doesn't use int64_t to detect
if apic-id property has been set.
Use the fact that 0xFFFFFFFF is the broadcast
value that a CPU can't have and set default
uint32_t apic_id to it instead of using int64_t.
Later uint32_t apic_id will be used to drop custom
property setter/getter in favor of static property.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Fill the bits between 51..number-of-physical-address-bits in the
MTRR_PHYSMASKn variable range mtrr masks so that they're consistent
in the migration stream irrespective of the physical address space
of the source VM in a migration.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
The CPU GPs if we try and set a bit in a variable MTRR mask above
the limit of physical address bits on the host. We hit this
when loading a migration from a host with a larger physical
address limit than our destination (e.g. a Xeon->i7 of same
generation) but previously used to get away with it
until 48e1a45 started checking that msr writes actually worked.
It seems in our case the GP probably comes from KVM emulating
that GP.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Currently QEMU sets the x86 number of physical address bits to the
magic number 40. This is only correct on some small AMD systems;
Intel systems tend to have 36, 39, 46 bits, and large AMD systems
tend to have 48.
Having the value different from your actual hardware is detectable
by the guest and in principal can cause problems;
The current limit of 40 stops TB VMs being created by those lucky
enough to have that much.
This patch lets you set the physical bits by a cpu property but
defaults to the same 40bits which matches TCGs setup.
I've removed the ancient warning about the 42 bit limit in exec.c;
I can't find that limit in there and no one else seems to know where
it is.
We use a magic value of 0 as the property default so that we can
later distinguish between the default and a user set value.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Provide a constant for the number of address bits supported under TCG.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Coverity complains that the GICR_IPRIORITYR case in gicv3_readl()
can overflow an array, because it doesn't know that the offsets
passed to that function must be word aligned. Add some assert()s
which hopefully tell Coverity that this isn't possible.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Message-id: 1468261372-17508-1-git-send-email-peter.maydell@linaro.org
Coverity complains that the exit() in gicv3_class_name()
can be unreachable, because if TARGET_AARCH64 is defined
then all code paths return before reaching it. Move the
exit() up to the error_report() that it belongs with.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Shannon Zhao <shannon.zhao@linaro.org>
Message-id: 1468260552-8400-1-git-send-email-peter.maydell@linaro.org
disas/bfd.h defines ATTRIBUTE_UNUSED, but unfortunately the
ALSA system headers also define this macro, which means that
you can get a compilation failure if building with ALSA and
any files happen to include the alsa headers before bfd.h
rather than the other way around.
This is unfortunate namespace pollution by the ALSA headers but
we can work around it. Add an #ifndef guard to bfd.h and remove
the unnecessary extra definition in disas/arm.c to fix this.
Reported-by: BALATON Zoltan <balaton@eik.bme.hu>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: 1468937076-21503-1-git-send-email-peter.maydell@linaro.org
The AArch64 Linux ABI syscall 84 is sync_file_range, not
sync_file_range2 (in the kernel it uses the asm-generic
headers and does not define __ARCH_WANT_SYNC_FILE_RANGE2).
Update our TARGET_NR_* definitions accordingly.
This fixes the sync_file_range syscall which otherwise
gets its arguments in the wrong order.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
The SIOCATMARK ioctl takes an argument which should be a
pointer to an integer where the kernel will write the result.
We were incorrectly declaring it as TYPE_NULL which would mean
it would always fail (with EFAULT) when it should succeed.
Correct the type.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
TIOCGPTN and related terminal control ioctls were not converted to the guest ioctl format on x86_64 targets. Convert these ioctls to enable terminal functionality on x86_64 guests.
Signed-off-by: Timothy Pearson <tpearson@raptorengineering.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Add some new blk ioctls (these are 0x12,119 through
to 0x12,127). Several of these are used by mke2fs; this silences
the warnings:
mke2fs 1.42.12 (29-Aug-2014)
Unsupported ioctl: cmd=0x127b
Unsupported ioctl: cmd=0x127a
warning: Unable to get device geometry for /dev/loop5
Unsupported ioctl: cmd=0x127c
Unsupported ioctl: cmd=0x127c
Unsupported ioctl: cmd=0x1277
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
If userspace specifies a short buffer for a target sockaddr,
the kernel will only copy in as much as it has space for
(or none at all if the length is zero) -- see the kernel
move_addr_to_user() function. Mimic this in QEMU's
host_to_target_sockaddr() routine.
In particular, this fixes a segfault running the LTP
recvfrom01 test, where the guest makes a recvfrom()
call with a bad buffer pointer and other parameters which
cause the kernel to set the addrlen to zero; because we
did not skip the attempt to swap the sa_family field we
segfaulted on the bad address.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Commit 655ed67c2a which switched synchronous signals to
benig recorded in ts->sync_signal rather than in a queue
with every other signal had a bug: we failed to clear
the flag indicating that a synchronous signal was pending
when we delivered it. This meant that we would take the signal
again and again every time the guest made a syscall.
(This is a bug introduced in my refactoring of Timothy Baldwin's
original code.)
Fix this by passing in the struct emulated_sigtable* to
handle_pending_signal(), so that we clear the pending flag
in the ts->sync_signal struct when handling a synchronous signal.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
The LOOP_GET_STATUS and LOOP_GET_STATUS64 ioctls were incorrectly
defined as IOC_W rather than IOC_R, which meant we weren't
correctly copying the information back from the kernel to the guest.
The loop_info64 structure definition was also missing a member
and using the wrong type for several 32-bit fields.
In particular, this meant that "kpartx -d image.img" didn't work
and "losetup -a" behaved strangely. Correct the ioctl type definitions.
Reported-by: Chanho Park <chanho61.park@samsung.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
The BLKSSZGET ioctl takes an argument which is a pointer to an int.
We were incorrectly declaring it to take a pointer to a long, which
meant that we would incorrectly write to memory which we should not
if the guest is a 64-bit architecture.
In particular, kpartx uses this ioctl to write to an int on the
stack, which tends to result in it crashing immediately.
Reported-by: Chanho Park <chanho61.park@samsung.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Add support for the /dev/loop-control ioctls:
LOOP_CTL_ADD
LOOP_CTL_REMOVE
LOOP_CTL_GET_FREE
[RV: fixed to apply to new header guards]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Many syscalls which take a sigset_t argument also take an argument
giving the size of the sigset_t. The kernel insists that this
matches its idea of the type size and fails EINVAL if it is not.
Implement this logic in QEMU. (This mostly just means some LTP test
cases which check error cases now pass.)
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Nested types are used by the kernel to send link information and
protocol properties.
We can see following errors with "ip link show":
Unimplemented nested type 26
Unimplemented nested type 26
Unimplemented nested type 18
Unimplemented nested type 26
Unimplemented nested type 18
Unimplemented nested type 26
This patch implements nested types 18 (IFLA_LINKINFO) and
26 (IFLA_AF_SPEC).
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
As we convert sockaddr for AF_PACKET family for sendto() (target to
host) we need also to convert this for getsockname() (host to target).
arping uses getsockname() to get the the interface address and uses
this address with sendto().
Tested with:
/sbin/arping -D -q -c2 -I eno1 192.168.122.88
...
getsockname(3, {sa_family=AF_PACKET, proto=0x806, if2,
pkttype=PACKET_HOST, addr(6)={1, 10c37b6b9a76}, [18]) = 0
...
sendto(3, "..." 28, 0,
{sa_family=AF_PACKET, proto=0x806, if2, pkttype=PACKET_HOST,
addr(6)={1, ffffffffffff}, 20) = 28
...
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Signed-off-by: Riku Voipio <riku.voipio@linaro.org>
