Failing to set any of the ICS property should really never happen:
- object_property_add_child() always succeed unless the child object
already has a parent, which isn't the case here obviously since the
ICS has just been created with object_new()
- the ICS has an "nr-irqs" property than can be set as long as the ICS
isn't realized
In both cases, an error indicates there is a bug in QEMU. Propagating the
error, ie. exiting QEMU since spapr_irq_init() is called with &error_fatal
doesn't make much sense. Abort instead. This is consistent with what is
done with XIVE : both qdev_create() and qdev_prop_set_uint32() abort QEMU
on error.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <157403285265.409804.8683093665795248192.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>
Traditional PCI INTx for vfio devices can only perform well if using
an in-kernel irqchip. Therefore, vfio_intx_update() issues a warning
if an in kernel irqchip is not available.
We usually do have an in-kernel irqchip available for pseries machines
on POWER hosts. However, because the platform allows feature
negotiation of what interrupt controller model to use, we don't
currently initialize it until machine reset. vfio_intx_update() is
called (first) from vfio_realize() before that, so it can issue a
spurious warning, even if we will have an in kernel irqchip by the
time we need it.
To workaround this, make a call to spapr_irq_update_active_intc() from
spapr_irq_init() which is called at machine realize time, before the
vfio realize. This call will be pretty much obsoleted by the later
call at reset time, but it serves to suppress the spurious warning
from VFIO.
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
pseries machine type can have one of two different interrupt controllers in
use depending on feature negotiation with the guest. Usually this is
invisible to devices, because they route to a common set of qemu_irqs which
in turn dispatch to the correct back end.
VFIO passthrough devices, however, wire themselves up directly to the KVM
irqchip for performance, which means they are affected by this change in
interrupt controller. To get them to adjust correctly for the change in
irqchip, we need to fire the kvm irqchip change notifier.
Cc: Alex Williamson <alex.williamson@redhat.com>
Cc: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Acked-by: Alex Williamson <alex.williamson@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>
For the benefit of peripheral device allocation, the number of available
irqs really wants to be the same on a given machine type version,
regardless of what irq backends we are using. That's the case now, but
only because we make sure the different SpaprIrq instances have the same
value except for the special legacy one.
Since this really only depends on machine type version, move the value to
SpaprMachineClass instead of SpaprIrq. This also puts the code to set it
to the lower value on old machine types right next to setting
legacy_irq_allocation, which needs to go hand in hand.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
The nr_msis value we use here has to line up with whether we're using
legacy or modern irq allocation. Therefore it's safer to derive it based
on legacy_irq_allocation rather than having SpaprIrq contain a canned
value.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
The remaining logic in the post_load hook really belongs to the interrupt
controller backends, and just needs to be called on the active controller
(after the active controller is set to the right thing based on the
incoming migration in the generic spapr_irq_post_load() logic).
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
It turns out that all the logic in the SpaprIrq::reset hooks (and some in
the SpaprIrq::post_load hooks) isn't really related to resetting the irq
backend (that's handled by the backends' own reset routines). Rather its
about getting the backend ready to be the active interrupt controller or
stopping being the active interrupt controller - reset (and post_load) is
just the only time that changes at present.
To make this flow clearer, move the logic into the explicit backend
activate and deactivate hooks.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
This hook is a bit odd. The only caller is spapr_irq_init_kvm(), but
it explicitly takes an SpaprIrq *, so it's never really called through the
current SpaprIrq. Essentially this is just a way of passing through a
function pointer so that spapr_irq_init_kvm() can handle some
configuration and error handling logic without duplicating it between the
xics and xive reset paths.
So, make it just take that function pointer. Because of earlier reworks
to the KVM connect/disconnect code in the xics and xive backends we can
also eliminate some wrapper functions and streamline error handling a bit.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Both XICS and XIVE have routines to connect and disconnect KVM with
similar but not identical signatures. This adjusts them to match
exactly, which will be useful for further cleanups later.
While we're there, we add an explicit return value to the connect path
to streamline error reporting in the callers. We remove error
reporting the disconnect path. In the XICS case this wasn't used at
all. In the XIVE case the only error case was if the KVM device was
set up, but KVM didn't have the capability to do so which is pretty
obviously impossible.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
This method depends only on the active irq controller. Now that we've
formalized the notion of active controller we can dispatch directly
through that, rather than dispatching via SpaprIrq with the dual
version having to do a second conditional dispatch.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
This method depends only on the active irq controller. Now that we've
formalized the notion of active controller we can dispatch directly
through that, rather than dispatching via SpaprIrq with the dual
version having to do a second conditional dispatch.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
This method depends only on the active irq controller. Now that we've
formalized the notion of active controller we can dispatch directly through
that, rather than dispatching via SpaprIrq with the dual version having
to do a second conditional dispatch.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
spapr now has the mechanism of constructing both XICS and XIVE instances of
the SpaprInterruptController interface. However, only one of the interrupt
controllers will actually be active at any given time, depending on feature
negotiation with the guest. This is handled in the current code via
spapr_irq_current() which checks the OV5 vector from feature negotiation to
determine the current backend.
Determining the active controller at the point we need it like this
can be pretty confusing, because it makes it very non obvious at what
points the active controller can change. This can make it difficult
to reason about the code and where a change of active controller could
appear in sequence with other events.
Make this mechanism more explicit by adding an 'active_intc' pointer
and an explicit spapr_irq_update_active_intc() function to update it
from the CAS state. We also add hooks on the intc backend which will
get called when it is activated or deactivated.
For now we just introduce the switch and hooks, later patches will
actually start using them.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
These methods, like cpu_intc_create, really belong to the interrupt
controller, but need to be called on all possible intcs.
Like cpu_intc_create, therefore, make them methods on the intc and
always call it for all existing intcs.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
This method essentially represents code which belongs to the interrupt
controller, but needs to be called on all possible intcs, rather than
just the currently active one. The "dual" version therefore calls
into the xics and xive versions confusingly.
Handle this more directly, by making it instead a method on the intc
backend, and always calling it on every backend that exists.
While we're there, streamline the error reporting a bit.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
The SpaprIrq structure is used to represent ths spapr machine's irq
backend. Except that it kind of conflates two concepts: one is the
backend proper - a specific interrupt controller that we might or
might not be using, the other is the irq configuration which covers
the layout of irq space and which interrupt controllers are allowed.
This leads to some pretty confusing code paths for the "dual"
configuration where its hooks redirect to other SpaprIrq structures
depending on the currently active irq controller.
To clean this up, we start by introducing a new
SpaprInterruptController QOM interface to represent strictly an
interrupt controller backend, not counting anything configuration
related. We implement this interface in the XICs and XIVE interrupt
controllers, and in future we'll move relevant methods from SpaprIrq
into it.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
This method is used to set up the interrupt backends for the current
configuration. However, this means some confusing redirection between
the "dual" mode init and the init hooks for xics only and xive only modes.
Since we now have simple flags indicating whether XICS and/or XIVE are
supported, it's easier to just open code each initialization directly in
spapr_irq_init(). This will also make some future cleanups simpler.
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>
Explicitly return success or failure, rather than just relying on the
Error ** parameter. This makes handling it less verbose in the caller.
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>
SpaprIrq::ov5 stores the value for a particular byte in PAPR option vector
5 which indicates whether XICS, XIVE or both interrupt controllers are
available. As usual for PAPR, the encoding is kind of overly complicated
and confusing (though to be fair there are some backwards compat things it
has to handle).
But to make our internal code clearer, have SpaprIrq encode more directly
which backends are available as two booleans, and derive the OV5 value from
that at the point we need it.
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>
spapr_xive_irq_claim() returns a bool to indicate if it succeeded.
But most of the callers and one callee use int return values and/or an
Error * with more information instead. In any case, ints are a more
common idiom for success/failure states than bools (one never knows
what sense they'll be in).
So instead change to an int return value to indicate presence of error
+ an Error * to describe the details through that call chain.
It also didn't actually check if the irq was already claimed, which is
one of the primary purposes of the claim path, so do that.
spapr_xive_irq_free() also returned a bool... which no callers checked
and was always true, so just drop it.
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>
The irq claim and free paths for both XICS and XIVE check for some
validity conditions. Some of these represent genuine runtime failures,
however others - particularly checking that the basic irq number is in a
sane range - could only fail in the case of bugs in the callin code.
Therefore use assert()s instead of runtime failures for those.
In addition the non backend-specific part of the claim/free paths should
only be used for PAPR external irqs, that is in the range SPAPR_XIRQ_BASE
to the maximum irq number. Put assert()s for that into the top level
dispatchers as well.
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>
spapr_irq_free() can be used to free multiple irqs at once. That's useful
for its callers, but there's no need to make the individual backend hooks
handle this. We can loop across the irqs in spapr_irq_free() itself and
have the hooks just do one at time.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
These traces contain some useless information (the always-0 source#) and
have no equivalents for XIVE mode. For now just remove them, and we can
put back something more sensible if and when we need it.
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>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
This method is used to determine the name of the irq backend's node in the
device tree, so that we can find its phandle (after SLOF may have modified
it from the phandle we initially gave it).
But, in the two cases the only difference between the node name is the
presence of a unit address. Searching for a node name without considering
unit address is standard practice for the device tree, and
fdt_subnode_offset() will do exactly that, making this method unecessary.
While we're there, remove the XICS_NODENAME define. The name
"interrupt-controller" is required by PAPR (and IEEE1275), and a bunch of
places assume it already.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Currently spapr_qirq(), whic is used to find the qemu_irq for an spapr
global irq number, redirects through the SpaprIrq::qirq method. But
the array of qemu_irqs is allocated in the PAPR layer, not the
backends, and so the method implementations all return the same thing,
just differing in the preliminary checks they make.
So, we can remove the method, and just implement spapr_qirq() directly,
including all the relevant checks in one place. We change all those
checks into assert()s as well, since a failure here indicates an error in
the calling code.
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>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
spapr global irq numbers are different from the source numbers on the ICS
when using XICS - they're offset by XICS_IRQ_BASE (0x1000). But
spapr_irq_set_irq_xics() was passing through the global irq number to
the ICS code unmodified.
We only got away with this because of a counteracting bug - we were
incorrectly adjusting the qemu_irq we returned for a requested global irq
number.
That approach mostly worked but is very confusing, incorrectly relies on
the way the qemu_irq array is allocated, and undermines the intention of
having the global array of qemu_irqs for spapr have a consistent meaning
regardless of irq backend.
So, fix both set_irq and qemu_irq indexing. We rename some parameters at
the same time to make it clear that they are referring to spapr global
irq numbers.
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>
The only reason this parameter was needed was to work around the
inconsistent meaning of nr_irqs between xics and xive. Now that we've
fixed that, we can consistently use the number directly in the SpaprIrq
configuration.
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>
Both the XICS and XIVE interrupt backends have a "nr-irqs" property, but
it means slightly different things. For XICS (or, strictly, the ICS) it
indicates the number of "real" external IRQs. Those start at XICS_IRQ_BASE
(0x1000) and don't include the special IPI vector. For XIVE, however, it
includes the whole IRQ space, including XIVE's many IPI vectors.
The spapr code currently doesn't handle this sensibly, with the
nr_irqs value in SpaprIrq having different meanings depending on the
backend. We fix this by renaming nr_irqs to nr_xirqs and making it
always indicate just the number of external irqs, adjusting the value
we pass to XIVE accordingly. We also move to using common constants
in most of the irq configurations, to make it clearer that the IRQ
space looks the same to the guest (and emulated devices), even if the
backend is different.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
We create a subtype of TYPE_ICS specifically for sPAPR. For now all this
does is move the setup of the PAPR specific hcalls and RTAS calls to
the realize() function for this, rather than requiring the PAPR code to
explicitly call xics_spapr_init(). In future it will have some more
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>
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>
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>
It will help us to discard interrupt numbers which have not been
claimed in the next patch.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20190911133937.2716-2-clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
PHBs already take care of clearing the MSIs from the bitmap during reset
or unplug. No need to do this globally from the machine code. Rather add
an assert to ensure that PHBs have acted as expected.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156415228966.1064338.190189424190233355.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
[dwg: Fix crash in qtest case where spapr->irq_map can be NULL at the
new assert()]
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
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/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>
Just to give an indication to the user that the error condition is
handled and how.
Reported-by: Satheesh Rajendran <sathnaga@linux.vnet.ibm.com>
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156398743479.546975.14566809803480887488.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The init_emu() handles are now empty. Remove them and rename
spapr_irq_init_device() to spapr_irq_init_kvm().
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20190614165920.12670-3-clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Today, the interrupt device is fully initialized at reset when the CAS
negotiation process has completed. Depending on the KVM capabilities,
the SpaprXive memory regions (ESB, TIMA) are initialized with a host
MMIO backend or a QEMU emulated backend. This results in a complex
initialization sequence partially done at realize and later at reset,
and some memory region leaks.
To simplify this sequence and to remove of the late initialization of
the emulated device which is required to be done only once, we
introduce new memory regions specific for KVM. These regions are
mapped as overlaps on top of the emulated device to make use of the
host MMIOs. Also provide proper cleanups of these regions when the
XIVE KVM device is destroyed to fix the leaks.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20190614165920.12670-2-clg@kaod.org>
Reviewed-by: Greg Kurz <groug@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Switch to using the connect/disconnect terminology like we already do for
XIVE.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156077920102.433243.6605099291134598170.stgit@bahia.lan>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Older KVMs on POWER9 don't support destroying/recreating a KVM XICS
device, which is required by 'dual' interrupt controller mode. This
causes QEMU to emit a warning when the guest is rebooted and to fall
back on XICS emulation:
qemu-system-ppc64: warning: kernel_irqchip allowed but unavailable:
Error on KVM_CREATE_DEVICE for XICS: File exists
If kernel irqchip is required, QEMU will thus exit when the guest is
first rebooted. Failing QEMU this late may be a painful experience
for the user.
Detect that and exit at machine init instead.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156044430517.125694.6207865998817342638.stgit@bahia.lab.toulouse-stg.fr.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
QEMU may crash when running a spapr machine in 'dual' interrupt controller
mode on some older (but not that old, eg. ubuntu 18.04.2) KVMs with partial
XIVE support:
qemu-system-ppc64: hw/ppc/spapr_rtas.c:411: spapr_rtas_register:
Assertion `!name || !rtas_table[token].name' failed.
XICS is controlled by the guest thanks to a set of RTAS calls. Depending
on whether KVM XICS is used or not, the RTAS calls are handled by KVM or
QEMU. In both cases, QEMU needs to expose the RTAS calls to the guest
through the "rtas" node of the device tree.
The spapr_rtas_register() helper takes care of all of that: it adds the
RTAS call token to the "rtas" node and registers a QEMU callback to be
invoked when the guest issues the RTAS call. In the KVM XICS case, QEMU
registers a dummy callback that just prints an error since it isn't
supposed to be invoked, ever.
Historically, the XICS controller was setup during machine init and
released during final teardown. This changed when the 'dual' interrupt
controller mode was added to the spapr machine: in this case we need
to tear the XICS down and set it up again during machine reset. The
crash happens because we indeed have an incompatibility with older
KVMs that forces QEMU to fallback on emulated XICS, which tries to
re-registers the same RTAS calls.
This could be fixed by adding proper rollback that would unregister
RTAS calls on error. But since the emulated RTAS calls in QEMU can
now detect when they are mistakenly called while KVM XICS is in
use, it seems simpler to register them once and for all at machine
init. This fixes the crash and allows to remove some now useless
lines of code.
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <156044429963.125694.13710679451927268758.stgit@bahia.lab.toulouse-stg.fr.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The interrupt mode is chosen by the CAS negotiation process and
activated after a reset to take into account the required changes in
the machine. This brings new constraints on how the associated KVM IRQ
device is initialized.
Currently, each model takes care of the initialization of the KVM
device in their realize method but this is not possible anymore as the
initialization needs to be done globaly when the interrupt mode is
known, i.e. when machine is reseted. It also means that we need a way
to delete a KVM device when another mode is chosen.
Also, to support migration, the QEMU objects holding the state to
transfer should always be available but not necessarily activated.
The overall approach of this proposal is to initialize both interrupt
mode at the QEMU level to keep the IRQ number space in sync and to
allow switching from one mode to another. For the KVM side of things,
the whole initialization of the KVM device, sources and presenters, is
grouped in a single routine. The XICS and XIVE sPAPR IRQ reset
handlers are modified accordingly to handle the init and the delete
sequences of the KVM device.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20190513084245.25755-15-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
The way the XICS and the XIVE devices are initialized follows the same
pattern. First, try to connect to the KVM device and if not possible
fallback on the emulated device, unless a kernel_irqchip is required.
The spapr_irq_init_device() routine implements this sequence in
generic way using new sPAPR IRQ handlers ->init_emu() and ->init_kvm().
The XIVE init sequence is moved under the associated sPAPR IRQ
->init() handler. This will change again when KVM support is added for
the dual interrupt mode.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20190513084245.25755-12-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
All is in place for KVM now. State synchronization and migration will
come next.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20190513084245.25755-8-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
When the VM is stopped, the VM state handler stabilizes the XIVE IC
and marks the EQ pages dirty. These are then transferred to destination
before the transfer of the device vmstates starts.
The SpaprXive interrupt controller model captures the XIVE internal
tables, EAT and ENDT and the XiveTCTX model does the same for the
thread interrupt context registers.
At restart, the SpaprXive 'post_load' method restores all the XIVE
states. It is called by the sPAPR machine 'post_load' method, when all
XIVE states have been transferred and loaded.
Finally, the source states are restored in the VM change state handler
when the machine reaches the running state.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20190513084245.25755-7-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
This introduces a set of helpers when KVM is in use, which create the
KVM XIVE device, initialize the interrupt sources at a KVM level and
connect the interrupt presenters to the vCPU.
They also handle the initialization of the TIMA and the source ESB
memory regions of the controller. These have a different type under
KVM. They are 'ram device' memory mappings, similarly to VFIO, exposed
to the guest and the associated VMAs on the host are populated
dynamically with the appropriate pages using a fault handler.
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Message-Id: <20190513084245.25755-3-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
spapr_ics_create() is only called once. Merge it in spapr_irq_init_xics()
and simplify a bit the error handling by using 'error_fatal' .
Signed-off-by: Cédric Le Goater <clg@kaod.org>
Message-Id: <20190321144914.19934-13-clg@kaod.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
