Devices may have component devices and buses.
Device realization may fail. Realization is recursive: a device's
realize() method realizes its components, and device_set_realized()
realizes its buses (which should in turn realize the devices on that
bus, except bus_set_realized() doesn't implement that, yet).
When realization of a component or bus fails, we need to roll back:
unrealize everything we realized so far. If any of these unrealizes
failed, the device would be left in an inconsistent state. Must not
happen.
device_set_realized() lets it happen: it ignores errors in the roll
back code starting at label child_realize_fail.
Since realization is recursive, unrealization must be recursive, too.
But how could a partly failed unrealize be rolled back? We'd have to
re-realize, which can fail. This design is fundamentally broken.
device_set_realized() does not roll back at all. Instead, it keeps
unrealizing, ignoring further errors.
It can screw up even for a device with no buses: if the lone
dc->unrealize() fails, it still unregisters vmstate, and calls
listeners' unrealize() callback.
bus_set_realized() does not roll back either. Instead, it stops
unrealizing.
Fortunately, no unrealize method can fail, as we'll see below.
To fix the design error, drop parameter @errp from all the unrealize
methods.
Any unrealize method that uses @errp now needs an update. This leads
us to unrealize() methods that can fail. Merely passing it to another
unrealize method cannot cause failure, though. Here are the ones that
do other things with @errp:
* virtio_serial_device_unrealize()
Fails when qbus_set_hotplug_handler() fails, but still does all the
other work. On failure, the device would stay realized with its
resources completely gone. Oops. Can't happen, because
qbus_set_hotplug_handler() can't actually fail here. Pass
&error_abort to qbus_set_hotplug_handler() instead.
* hw/ppc/spapr_drc.c's unrealize()
Fails when object_property_del() fails, but all the other work is
already done. On failure, the device would stay realized with its
vmstate registration gone. Oops. Can't happen, because
object_property_del() can't actually fail here. Pass &error_abort
to object_property_del() instead.
* spapr_phb_unrealize()
Fails and bails out when remove_drcs() fails, but other work is
already done. On failure, the device would stay realized with some
of its resources gone. Oops. remove_drcs() fails only when
chassis_from_bus()'s object_property_get_uint() fails, and it can't
here. Pass &error_abort to remove_drcs() instead.
Therefore, no unrealize method can fail before this patch.
device_set_realized()'s recursive unrealization via bus uses
object_property_set_bool(). Can't drop @errp there, so pass
&error_abort.
We similarly unrealize with object_property_set_bool() elsewhere,
always ignoring errors. Pass &error_abort instead.
Several unrealize methods no longer handle errors from other unrealize
methods: virtio_9p_device_unrealize(),
virtio_input_device_unrealize(), scsi_qdev_unrealize(), ...
Much of the deleted error handling looks wrong anyway.
One unrealize methods no longer ignore such errors:
usb_ehci_pci_exit().
Several realize methods no longer ignore errors when rolling back:
v9fs_device_realize_common(), pci_qdev_unrealize(),
spapr_phb_realize(), usb_qdev_realize(), vfio_ccw_realize(),
virtio_device_realize().
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200505152926.18877-17-armbru@redhat.com>
The only way object_property_add() can fail is when a property with
the same name already exists. Since our property names are all
hardcoded, failure is a programming error, and the appropriate way to
handle it is passing &error_abort.
Same for its variants, except for object_property_add_child(), which
additionally fails when the child already has a parent. Parentage is
also under program control, so this is a programming error, too.
We have a bit over 500 callers. Almost half of them pass
&error_abort, slightly fewer ignore errors, one test case handles
errors, and the remaining few callers pass them to their own callers.
The previous few commits demonstrated once again that ignoring
programming errors is a bad idea.
Of the few ones that pass on errors, several violate the Error API.
The Error ** argument must be NULL, &error_abort, &error_fatal, or a
pointer to a variable containing NULL. Passing an argument of the
latter kind twice without clearing it in between is wrong: if the
first call sets an error, it no longer points to NULL for the second
call. ich9_pm_add_properties(), sparc32_ledma_realize(),
sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize()
are wrong that way.
When the one appropriate choice of argument is &error_abort, letting
users pick the argument is a bad idea.
Drop parameter @errp and assert the preconditions instead.
There's one exception to "duplicate property name is a programming
error": the way object_property_add() implements the magic (and
undocumented) "automatic arrayification". Don't drop @errp there.
Instead, rename object_property_add() to object_property_try_add(),
and add the obvious wrapper object_property_add().
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200505152926.18877-15-armbru@redhat.com>
[Two semantic rebase conflicts resolved]
This is a model of the PCIe Host Bridge (PHB3) found on a POWER8
processor. It includes the PowerBus logic interface (PBCQ), IOMMU
support, a single PCIe Gen.3 Root Complex, and support for MSI and LSI
interrupt sources as found on a POWER8 system using the XICS interrupt
controller.
The POWER8 processor comes in different flavors: Venice, Murano,
Naple, each having a different number of PHBs. To make things simpler,
the models provides 3 PHB3 per chip. Some platforms, like the
Firestone, can also couple PHBs on the first chip to provide more
bandwidth but this is too specific to model in QEMU.
XICS requires some adjustment to support the PHB3 MSI. The changes are
provided here but they could be decoupled in prereq patches.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20200127144506.11132-3-clg@kaod.org>
[dwg: Use device_class_set_props()]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The correct way to do this is to deassert the input pins on the CPU side.
This is the case since a previous change.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157548862298.3650476.1228720391270249433.stgit@bahia.lan>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The ICP object has both a pointer and an ICP_PROP_CPU property pointing
to the cpu. Confusing bugs could arise if these ever go out of sync.
Change the property definition so that it explicitly sets the pointer.
The property isn't optional : not being able to set the link is a bug
and QEMU should rather abort than exit in this case.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157403284709.409804.16142099083325945141.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The ICP object has both a pointer and an ICP_PROP_XICS property pointing
to the XICS fabric. Confusing bugs could arise if these ever go out of
sync.
Change the property definition so that it explicitly sets the pointer.
The property isn't optional : not being able to set the link is a bug
and QEMU should rather abort than exit in this case.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157403284152.409804.17114564311521923733.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The ICS object has both a pointer and an ICS_PROP_XICS property pointing
to the XICS fabric. Confusing bugs could arise if these ever go out of
sync.
Change the property definition so that it explicitely sets the pointer.
The property isn't optional : not being able to set the link is a bug
and QEMU should rather abort than exit in this case.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157403283596.409804.17347207690271971987.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
CPU_FOREACH() can race with vCPU hotplug/unplug on sPAPR machines, ie.
we may try to print out info about a vCPU with a NULL presenter pointer.
Check that in order to prevent QEMU from crashing.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157192725327.3146912.12047076483178652551.stgit@bahia.lan>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
When a VCPU gets connected to the XIVE interrupt controller, we add a
const link targetting the CPU object to the TCTX object. Similar links
are added to the ICP object when using the XICS interrupt controller.
As explained in <qom/object.h>:
* The caller must ensure that @target stays alive as long as
* this property exists. In the case @target is a child of @obj,
* this will be the case. Otherwise, the caller is responsible for
* taking a reference.
We're in the latter case for both XICS and XIVE. Add the missing
calls to object_ref() and object_unref().
This doesn't fix any known issue because the life cycle of the TCTX or
ICP happens to be shorter than the one of the CPU or XICS fabric, but
better safe than sorry.
Signed-off-by: Greg Kurz <groug@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <157192724770.3146912.15400869269097231255.stgit@bahia.lan>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
SpaprInterruptControllerClass and PnvChipClass have an intc_create() method
that calls the appropriate routine, ie. icp_create() or xive_tctx_create(),
to establish the link between the VCPU and the presenter component of the
interrupt controller during realize.
There aren't any symmetrical call to be called when the VCPU gets unrealized
though. It is assumed that object_unparent() is the only thing to do.
This is questionable because the parenting logic around the CPU and
presenter objects is really an implementation detail of the interrupt
controller. It shouldn't be open-coded in the machine code.
Fix this by adding an intc_destroy() method that undoes what was done in
intc_create(). Also NULLify the presenter pointers to avoid having
stale pointers around. This will allow to reliably check if a vCPU has
a valid presenter.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157192724208.3146912.7254684777515287626.stgit@bahia.lan>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
On the sPAPR machine and PowerNV machine, the interrupt presenters are
created by a machine handler at the core level and are reset
independently. This is not consistent and it raises issues when it
comes to handle hot-plugged CPUs. In that case, the presenters are not
reset. This is less of an issue in XICS, although a zero MFFR could
be a concern, but in XIVE, the OS CAM line is not set and this breaks
the presenting algorithm. The current code has workarounds which need
a global cleanup.
Extend the sPAPR IRQ backend and the PowerNV Chip class with a new
cpu_intc_reset() handler called by the CPU reset handler and remove
the XiveTCTX reset handler which is now redundant.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20191022163812.330-6-clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Some device types of the XICS model are exposed to the QEMU command
line:
$ ppc64-softmmu/qemu-system-ppc64 -device help | grep ic[sp]
name "icp"
name "ics"
name "ics-spapr"
name "pnv-icp", desc "PowerNV ICP"
These are internal devices that shouldn't be instantiable by the
user. By the way, they can't be because their respective realize
functions expect link properties that can't be set from the command
line:
qemu-system-ppc64: -device icp: required link 'xics' not found:
Property '.xics' not found
qemu-system-ppc64: -device ics: required link 'xics' not found:
Property '.xics' not found
qemu-system-ppc64: -device ics-spapr: required link 'xics' not found:
Property '.xics' not found
qemu-system-ppc64: -device pnv-icp: required link 'xics' not found:
Property '.xics' not found
Hide them by setting dc->user_creatable to false in the base class
"icp" and "ics" init functions.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157017826724.337875.14822177178282524024.stgit@bahia.lan>
Message-Id: <157045578962.865784.8551555523533955113.stgit@bahia.lan>
[dwg: Folded reason comment into base patch]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
TYPE_ICS_SIMPLE is the only subtype of TYPE_ICS_BASE that's ever
instantiated. The existence of different classes is mostly a hang
over from when we (misguidedly) had separate subtypes for the KVM and
non-KVM version of the device.
There could be some call for an abstract base type for ICS variants
that use a different representation of their state (PowerNV PHB3 might
want this). The current split isn't really in the right place for
that though. If we need this in future, we can re-implement it more
in line with what we actually need.
So, collapse the two classes together into just TYPE_ICS.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Currently TYPE_XICS_BASE and TYPE_XICS_SIMPLE have their own reset methods,
using the standard technique for having the subtype call the supertype's
methods before doing its own thing.
But TYPE_XICS_SIMPLE is the only subtype of TYPE_XICS_BASE ever
instantiated, so there's no point having the split here. Merge them
together into just an ics_reset() function.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
There are a number of ics_simple_*() functions that aren't actually
specific to TYPE_XICS_SIMPLE at all, and are equally valid on
TYPE_XICS_BASE. Rename them to ics_*() accordingly.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Currently ics_reject(), ics_resend() and ics_eoi() indirect through
class methods. But there's only one implementation of each method,
the one in TYPE_ICS_SIMPLE. TYPE_ICS_BASE has no implementation, but
it's never instantiated, and has no other subtypes.
So clean up by eliminating the method and just having ics_reject(),
ics_resend() and ics_eoi() contain the logic directly.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
In my "build everything" tree, changing hw/qdev-properties.h triggers
a recompile of some 2700 out of 6600 objects (not counting tests and
objects that don't depend on qemu/osdep.h).
Many places including hw/qdev-properties.h (directly or via hw/qdev.h)
actually need only hw/qdev-core.h. Include hw/qdev-core.h there
instead.
hw/qdev.h is actually pointless: all it does is include hw/qdev-core.h
and hw/qdev-properties.h, which in turn includes hw/qdev-core.h.
Replace the remaining uses of hw/qdev.h by hw/qdev-properties.h.
While there, delete a few superfluous inclusions of hw/qdev-core.h.
Touching hw/qdev-properties.h now recompiles some 1200 objects.
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Daniel P. Berrangé" <berrange@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
Message-Id: <20190812052359.30071-22-armbru@redhat.com>
In my "build everything" tree, changing hw/hw.h triggers a recompile
of some 2600 out of 6600 objects (not counting tests and objects that
don't depend on qemu/osdep.h).
The previous commits have left only the declaration of hw_error() in
hw/hw.h. This permits dropping most of its inclusions. Touching it
now recompiles less than 200 objects.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20190812052359.30071-19-armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
In my "build everything" tree, changing migration/vmstate.h triggers a
recompile of some 2700 out of 6600 objects (not counting tests and
objects that don't depend on qemu/osdep.h).
hw/hw.h supposedly includes it for convenience. Several other headers
include it just to get VMStateDescription. The previous commit made
that unnecessary.
Include migration/vmstate.h only where it's still needed. Touching it
now recompiles only some 1600 objects.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20190812052359.30071-16-armbru@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
In my "build everything" tree, changing hw/irq.h triggers a recompile
of some 5400 out of 6600 objects (not counting tests and objects that
don't depend on qemu/osdep.h).
hw/hw.h supposedly includes it for convenience. Several other headers
include it just to get qemu_irq and.or qemu_irq_handler.
Move the qemu_irq and qemu_irq_handler typedefs from hw/irq.h to
qemu/typedefs.h, and then include hw/irq.h only where it's still
needed. Touching it now recompiles only some 500 objects.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190812052359.30071-13-armbru@redhat.com>
In my "build everything" tree, changing sysemu/reset.h triggers a
recompile of some 2600 out of 6600 objects (not counting tests and
objects that don't depend on qemu/osdep.h).
The main culprit is hw/hw.h, which supposedly includes it for
convenience.
Include sysemu/reset.h only where it's needed. Touching it now
recompiles less than 200 objects.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190812052359.30071-9-armbru@redhat.com>
This allows errors happening there to be propagated up to spapr_irq,
just like XIVE already does.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156077921763.433243.4614327010172954196.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
So that no one is tempted to drop that code, which is never called
for cold plugged CPUs.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156078063349.435533.12283208810037409702.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Recent commits changed the behavior of ics_set_irq_type() to
initialize correctly LSIs at the KVM level. ics_set_irq_type() is also
called by the realize routine of the different devices of the machine
when initial interrupts are claimed, before the ICSState device is
reseted.
In the case, the ICSIRQState priority is 0x0 and the call to
ics_set_irq_type() results in configuring the target of the
interrupt. On P9, when using the KVM XICS-on-XIVE device, the target
is configured to be server 0, priority 0 and the event queue 0 is
created automatically by KVM.
With the dual interrupt mode creating the KVM device at reset, it
leads to unexpected effects on the guest, mostly blocking IPIs. This
is wrong, fix it by reseting the ICSIRQState structure when
ics_set_irq_type() is called.
Fixes: commit 6cead90c5c ("xics: Write source state to KVM at claim time")
Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20190513084245.25755-14-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The pseries machine only uses LSIs to support legacy PCI devices. Every
PHB claims 4 LSIs at realize time. When using in-kernel XICS (or upcoming
in-kernel XIVE), QEMU synchronizes the state of all irqs, including these
LSIs, later on at machine reset.
In order to support PHB hotplug, we need a way to tell KVM about the LSIs
that doesn't require a machine reset. An easy way to do that is to always
inform KVM when an interrupt is claimed, which really isn't a performance
path.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <155059668360.1466090.5969630516627776426.stgit@bahia.lab.toulouse-stg.fr.ibm.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Adds support for the Hypervisor directed interrupts in addition to the
OS ones.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
[clg: - modified the icp_realize() and xive_tctx_realize() to take
into account explicitely the POWER9 interrupt model
- introduced a specific power9_set_irq for POWER9 ]
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20190215161648.9600-10-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
We want to use the "simple" ICS type in both KVM and non-KVM setups.
Teach the "simple" ICS how to present interrupts to KVM and adapt
sPAPR accordingly.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <155023082996.1011724.16237920586343905010.stgit@bahia.lan>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The KVM ICS reset handler simply writes the ICS state to KVM. This
doesn't need the overkill parent_reset logic we have today. Also
we want to use the same ICS type for the KVM and non-KVM case with
pseries.
Call icp_set_kvm_state() from the "simple" ICS reset function.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <155023082407.1011724.1983100830860273401.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The pre_save(), post_load() and synchronize_state() methods of the
ICSStateClass type are really KVM only things. Make that obvious
by dropping the indirections and directly calling the KVM functions
instead.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <155023081817.1011724.14078777320394028836.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The realization of KVM ICP currently follows the parent_realize logic,
which is a bit overkill here. Also we want to get rid of the KVM ICP
class. Explicitely call icp_kvm_realize() from the base ICP realize
function.
Note that ICPStateClass::parent_realize is retained because powernv
needs it.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <155023080049.1011724.15423463482790260696.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The KVM ICP reset handler simply writes the ICP state to KVM. This
doesn't need the overkill parent_reset logic we have today. Call
icp_set_kvm_state() from the base ICP reset function instead.
Since there are no other users for ICPStateClass::parent_reset, and
it isn't currently expected to change, drop it as well.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <155023079461.1011724.12644984391500635645.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The pre_save(), post_load() and synchronize_state() methods of the
ICPStateClass type are really KVM only things. Make that obvious
by dropping the indirections and directly calling the KVM functions
instead.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <155023078871.1011724.3083923389814185598.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The qemu_irq array is now allocated at the machine level using a sPAPR
IRQ set_irq handler depending on the chosen interrupt mode. The use of
this handler is slightly inefficient today but it will become necessary
when the 'dual' interrupt mode is introduced.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
To support the 'dual' interrupt mode, XICS and XIVE, we plan to move
the qemu_irq array of each interrupt controller under the machine and
do the allocation under the sPAPR IRQ init method.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
From include/qapi/error.h:
* Pass an existing error to the caller with the message modified:
* error_propagate(errp, err);
* error_prepend(errp, "Could not frobnicate '%s': ", name);
Fei Li pointed out that doing error_propagate() first doesn't work
well when @errp is &error_fatal or &error_abort: the error_prepend()
is never reached.
Since I doubt fixing the documentation will stop people from getting
it wrong, introduce error_propagate_prepend(), in the hope that it
lures people away from using its constituents in the wrong order.
Update the instructions in error.h accordingly.
Convert existing error_prepend() next to error_propagate to
error_propagate_prepend(). If any of these get reached with
&error_fatal or &error_abort, the error messages improve. I didn't
check whether that's the case anywhere.
Cc: Fei Li <fli@suse.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-Id: <20181017082702.5581-2-armbru@redhat.com>
Recent cleanup in commit a028dd423e dropped the ICPStateClass::reset
handler. It is now up to child ICP classes to call the DeviceClass::reset
handler of the parent class, thanks to device_class_set_parent_reset().
This is a better object programming pattern, but unfortunately it causes
QEMU to crash during CPU hotplug:
(qemu) device_add host-spapr-cpu-core,id=core1,core-id=1
Segmentation fault (core dumped)
When the hotplug path tries to reset the ICP device, we end up calling:
static void icp_kvm_reset(DeviceState *dev)
{
ICPStateClass *icpc = ICP_GET_CLASS(dev);
icpc->parent_reset(dev);
but icpc->parent_reset is NULL... This happens because icp_kvm_class_init()
calls:
device_class_set_parent_reset(dc, icp_kvm_reset,
&icpc->parent_reset);
but dc->reset, ie, DeviceClass::reset for the TYPE_ICP type, is
itself NULL.
This patch hence sets DeviceClass::reset for the TYPE_ICP type to
point to icp_reset(). It then registers a reset handler that calls
DeviceClass::reset. If the ICP subtype has configured its own reset
handler with device_class_set_parent_reset(), this ensures it will
be called first and it can then call ICPStateClass::parent_reset
safely. This fixes the reset path for the TYPE_KVM_ICP type, which
is the only subtype that defines its own reset function.
Reported-by: Satheesh Rajendran <sathnaga@linux.vnet.ibm.com>
Suggested-by: David Gibson <david@gibson.dropbear.id.au>
Fixes: a028dd423e
Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Just like for the realize handlers, this makes possible to move the
common ICSState code of the reset handlers in the ics-base class.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
This makes possible to move the common ICSState code of the realize
handlers in the ics-base class.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
This changes the ICP realize and reset handlers in DeviceRealize and
DeviceReset handlers. parent handlers are now called from the
inheriting classes which is a cleaner object pattern.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
xics_get_qirq() is only used by the sPAPR machine. Let's move it there
and change its name to reflect its scope. It will be useful for XIVE
support which will use its own set of qirqs.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The 'intc' pointer of the CPU references the interrupt presenter in
the XICS interrupt mode. When the XIVE interrupt mode is available and
activated, the machine will need to reassign this pointer to reflect
the change.
Moving this assignment under the realize routine of the CPU will ease
the process when the interrupt mode is toggled.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The sPAPR and the PowerNV core objects create the interrupt presenter
object of the CPUs in a very similar way. Let's provide a common
routine in which we use the presenter 'type' as a child identifier.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
When using the emulated XICS, the 'info pic' monitor command shows:
CPU 0 XIRR=ff000000 ((nil)) PP=ff MFRR=ff
ICS 1000..13ff 0x10040060340
1000 MSI 05 00
1001 MSI 05 00
1002 MSI 05 00
1003 MSI ff 00
1004 LSI ff 00
1005 LSI ff 00
1006 LSI ff 00
1007 LSI ff 00
1008 MSI 05 00
1009 MSI 05 00
100a MSI 05 00
100b MSI 05 00
100c MSI 05 00
but when using the in-kernel XICS with the very same guest, we get:
CPU 0 XIRR=00000000 ((nil)) PP=ff MFRR=ff
ICS 1000..13ff 0x10032e00340
1000 MSI ff 00
1001 MSI ff 00
1002 MSI ff 00
1003 MSI ff 00
1004 LSI ff 00
1005 LSI ff 00
1006 LSI ff 00
1007 LSI ff 00
1008 MSI ff 00
1009 MSI ff 00
100a MSI ff 00
100b MSI ff 00
100c MSI ff 00
ie, all irqs are masked and XIRR is null, while we should get the
same output as with the emulated XICS.
If the guest is then migrated, 'info pic' shows the expected values
on both source and destination.
The problem is that QEMU doesn't synchronize with KVM before printing
the XICS state. Migration happens to fix the output because it enforces
synchronization with KVM.
To fix the invalid output of 'info pic', this patch introduces a new
synchronize_state operation for both ICPStateClass and ICSStateClass.
The ICP operation relies on run_on_cpu() in order to kick the vCPU
and avoid sleeping on KVM_GET_ONE_REG.
Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Modify the pre_save method on VMStateDescription to return an int
rather than void so that it potentially can fail.
Changed zillions of devices to make them return 0; the only
case I've made it return non-0 is hw/intc/s390_flic_kvm.c that already
had an error_report/return case.
Note: If you add an error exit in your pre_save you must emit
an error_report to say why.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20170925112917.21340-2-dgilbert@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
If object_property_get_link() fails then it allocates an error, which
must be freed before returning. The error_get_pretty() function is
merely an accessor to the error message and doesn't free anything.
The error.h header indicates how to do it right:
* Pass an existing error to the caller with the message modified:
* error_propagate(errp, err);
* error_prepend(errp, "Could not frobnicate '%s': ", name);
Signed-off-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The ICPState objects are currently registered to vmstate as qdev objects.
Their instance ids are hence computed automatically in the migration code,
and thus depends on the order the CPU cores were plugged.
If the destination had its CPU cores plugged in a different order than the
source, then ICPState objects will have different instance_ids and load
the wrong state.
Since CPU objects have a reliable cpu_index which is already used as
instance_id in vmstate, let's use it for ICPState as well.
Please note that this doesn't break migration. Older machine types used to
allocate and realize all ICPState objects at machine init time, for the whole
lifetime of the machine. The qdev instance ids are thus 0,1,2... nr_servers
and happen to map to the vCPU indexes.
Signed-off-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>