Multiple - even many - PCI host bridges (i.e. PCI domains) are very
common on real PAPR compliant hardware. For reasons related to the
PAPR specified IOMMU interfaces, PCI device assignment with VFIO will
generally require at least two (virtual) PHBs and possibly more
depending on which devices are assigned.
At the moment the qemu PAPR PCI code will not deal with this well,
leaving several crucial parameters of PHBs other than the default one
uninitialized. This patch reworks the code to allow this.
Every PHB needs a unique BUID (Bus Unit Identifier, the id used for
the PAPR PCI related interfaces) and a unique LIOBN (Logical IO Bus
Number, the id used for the PAPR IOMMU related interfaces). In
addition they need windows in CPU real address space to access PCI
memory space, PCI IO space and MSIs. Properties are added to the PCI
host bridge qdevice to allow configuration of all these.
To simplify configuration of multiple PHBs for common cases, a
convenience "index" property is also added. This can be set instead
of the low-level properties, and will generate suitable values for the
other parameters, different for each index value.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
Replace by SYS_BUS_DEVICE() QOM cast macro using a scripted conversion.
Avoids the old macro creeping into new code.
Resolve a Coding Style warning in openpic code.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Cc: Anthony Liguori <anthony@codemonkey.ws>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
There are lots of external users of pci_internals.h,
apparently making it an internal interface only didn't
work out. Let's stop pretending it's an internal header.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In one of the recent reworks to the XICS code, a bug was introduced where
we use the wrong sense and allocate level interrupts instead of message
interrupts for PCI MSIs. This patch fixes it.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
This reverts commit a178274efa.
Contrary to that commit's message, the users of old_portio are not all
gone. In particular VGA still uses it via portio_list_add().
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
target_phys_addr_t is unwieldly, violates the C standard (_t suffixes are
reserved) and its purpose doesn't match the name (most target_phys_addr_t
addresses are not target specific). Replace it with a finger-friendly,
standards conformant hwaddr.
Outstanding patchsets can be fixed up with the command
git rebase -i --exec 'find -name "*.[ch]"
| xargs s/target_phys_addr_t/hwaddr/g' origin
Signed-off-by: Avi Kivity <avi@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Currently the XICS interrupt controller emulation uses a custom enum to
specify whether a given interrupt is level-sensitive or message-triggered.
This enum makes life awkward for saving the state, and isn't particularly
useful since there are only two possibilities. This patch replaces the
enum with a simple bool.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
The emulated PCI host bridge on the pseries machine incorporates an IOMMU
(PAPR TCE table). Currently the mappings in this IOMMU are not cleared
when we reset the system. This patch fixes this bug. To do this it adds
a new reset function to the IOMMU emulation code. The VIO devices already
reset their TCE tables, but they do so by destroying and re-creating their
DMA context. This doesn't work for the PCI host bridge, because the
infrastructure for PCI IOMMUs has already copied/cached the DMA pointer
context into the subordinate PCI device structures.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
Use PCIHostState and PCI_HOST_BRIDGE() where appropriate.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Introduce type constant. Introduce cast macro to drop bogus busdev field
that would've broken SYS_BUS_DEVICE(). Avoid accessing parent fields
directly.
Free the identifier phb as acronym for PCI_HOST_BRIDGE.
Updated against conflicting merge from branch 'agraf/ppc-for-upstream'
(0d16fdd732), which removed busdev field
differently, moved some code around and added new occurrences of 'phb'.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
During the QOM migration they were amended with further info but this is
no longer the case. All static TypeInfos can be const these days.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Previously the only PCI bus supported was the emulated PCI bus with
fixed DMA window with start at 0 and size 1GB. As we are going to support
PCI pass through which DMA window properties are set by the host
kernel, we have to support DMA windows with parameters other than default.
This patch adds:
1. DMA window properties to sPAPRPHBState: LIOBN (bus id), start,
size of the window.
2. An additional function spapr_dma_dt() to populate DMA window
properties in the device tree which simply accepts all the parameters
and does not try to guess what kind of IOMMU is given to it.
The original spapr_dma_dt() is renamed to spapr_tcet_dma_dt().
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
This patch implements MSI and MSI-X support for the pseries PCI host
bridge. To do this it adds:
* A "config_space_address to msi_table" map, since the MSI RTAS calls
take a PCI config space address as an identifier.
* A MSIX memory region to catch msi_notify()/msix_notiry() from
virtio-pci and pass them to the guest via qemu_irq_pulse().
* RTAS call "ibm,change-msi" which sets up MSI vectors for a
device. Note that this call may configure and return lesser number of
vectors than requested.
* RTAS call "ibm,query-interrupt-source-number" which translates MSI
vector to interrupt controller (XICS) IRQ number.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[agraf: fix error case ndev < 0]
Signed-off-by: Alexander Graf <agraf@suse.de>
This adds a trace event in the pseries PCI specific set_irq() function to
assist in debugging.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[agraf: add trace.h include]
Signed-off-by: Alexander Graf <agraf@suse.de>
The pseries PCI code makes use of an internal find_dev() function which
locates a PCIDevice * given a (platform specific) bus ID and device
address. Internally this needs to first locate the host bridge on which
the device resides based on the bus ID. This patch exposes that host
bridge lookup as a separate function, which we will need later in the MSI
and VFIO code.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[agraf: drop trace.h inclusion]
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently the RTAS functions for handling PCI are registered from the
class init code for the PCI host bridge. That sort of makes sense
now, but will break in the future when vfio gives us multiple types of
host bridge for pseries (emulated and pass-through, at least). The
RTAS functions will be common across all host bridge types (and will
call out to different places internally depending on the type).
So, this patch moves the RTAS registration into its own function
called direct from the machine setup code.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently, the interfaces in the pseries machine code for assignment
and setup of interrupts pass around qemu_irq objects. That was done
in an attempt not to be too closely linked to the specific XICS
interrupt controller. However interactions with the device tree setup
made that attempt rather futile, and XICS is part of the PAPR spec
anyway, so this really just meant we had to carry both the qemu_irq
pointers and the XICS irq numbers around.
This mess will just get worse when we add upcoming PCI MSI support,
since that will require tracking a bunch more interrupt. Therefore,
this patch reworks the spapr code to just use XICS irq numbers
(roughly equivalent to GSIs on x86) and only retrieve the qemu_irq
pointers from the XICS code when we need them (a trivial lookup).
This is a reworked and generalized version of an earlier spapr_pci
specific patch from Alexey Kardashevskiy.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[agraf: fix checkpath warning]
Signed-off-by: Alexander Graf <agraf@suse.de>
spapr_populate_pci_devices() populates the device tree only with bus
properties and has nothing to do with the devices on it as PCI BAR
allocation is done by the system firmware (SLOF).
New name - spapr_populate_pci_dt() - describes the functionality better.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Alexander Graf <agraf@suse.de>
The PCIHostState struct already contains SysBusDevice so
the one in sPAPRPHBState has to go.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Alexander Graf <agraf@suse.de>
By default mingw-gcc is trying to pack structures the way to
preserve binary compatibility with MS Visual C what leads to
incorrect and unexpected padding in the PCI bus ranges property of
the sPAPR PHB.
The patch replaces __attribute__((packed)) with more strict QEMU_PACKED
which actually is __attribute__((gcc_struct, packed)) on Windows.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Tested-by: Stefan Weil <sw@weilnetz.de>
Reviewed-by: Stefan Weil <sw@weilnetz.de>
Acked-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Currently the pseries machine emulation does not support DMA for emulated
PCI devices, because the PAPR spec always requires a (guest visible,
paravirtualized) IOMMU which was not implemented. Now that we have
infrastructure for IOMMU emulation, we can correct this and allow PCI DMA
for pseries.
With the existing PAPR IOMMU code used for VIO devices, this is almost
trivial. We use a single DMAContext for each (virtual) PCI host bridge,
which is the usual configuration on real PAPR machines (which often have
_many_ PCI host bridges).
Cc: Alex Graf <agraf@suse.de>
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Make qbus children show up as link<> properties. There is no stable
addressing for qbus children so we use an unstable naming convention.
This is okay in QOM though because the composition name is expected to
be what's stable.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andreas Färber <afaerber@suse.de>
Currently the pseries PCI code uses a somewhat strange scheme of PCI irq
allocation - one per slot up to a maximum that's greater than the usual 4.
This scheme more or less worked, because we were able to tell the guest the
irq mapping in the device tree, however it's a bit odd and may break
assumptions in the future. Worse, the array used to construct the dev
tree interrupt map was mis-sized, we got away with it only because it
happened that our SPAPR_PCI_NUM_LSI value was greater than 7.
This patch changes the pseries PCI code to use the same interrupt swizzling
scheme as is standardized for PCI to PCI bridges. This makes for better
consistency, deals better with any devices which use multiple interrupt
pins and will make life easier in the future when we add passthrough of
what may be either a host bridge or a PCI to PCI bridge. This won't break
existing guests, because they don't assume a particular mapping scheme for
host bridges, but just follow what we tell them in the device tree (also
updated to match, of course). This patch also fixes the allocation of the
irq map.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
On the pseries platform, access to PCI config space is via RTAS calls(
which go to the hypervisor) rather than MMIO. This means we don't use
the same code path as nearly everyone else which goes through pci_host.c
and we're missing some of the parameter checking along the way.
We do have some parameter checking in the RTAS calls, but it's not enough.
It checks for overruns, but does not check for unaligned accesses,
oversized accesses (which means the guest could trigger an assertion
failure from pci_host_config_{read,write}_common(). Worse it doesn't do
the basic checking for the number of RTAS arguments and results before
accessing them.
This patch fixes these bugs.
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
[AF: Fix typos spotted by mst]
Signed-off-by: Andreas Färber <afaerber@suse.de>
Currently, the function spapr_create_phb() uses its parameters to
initialize the correct memory windows for the new PCI Host Bridge
(PHB). This is not the way things are supposed to be done with qdevs,
and means you can't create extra PHBs easily using -device.
Since pSeries machines can and do have many PHBs with various
configurations, this is a real limitation, not just a theoretical.
This patch, therefore, alters the PHB initialization code to use qdev
properties to set these parameters of the new bridge, moving most of
the code from spapr_create_phb() to spapr_phb_init().
While we're at it, we change the naming of each PCI bus and its
associated memory regions to be less arbitrary and make it easier to
relate the guest and qemu views of memory to each other.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Signed-off-by: Alexander Graf <agraf@suse.de>
The pseries "xics" interrupt controller, like most interrupt
controllers can support both message (i.e. edge sensitive) interrupts
and level sensitive interrupts, but it needs to know which are which.
When I implemented the xics emulation for qemu, the only devices we
supported were the PAPR virtual IO devices. These devices only use
message interrupts, so they were the only ones I implemented in xics.
Since then, however, we have added support for PCI devices, which use
level sensitive interrupts. It turns out the message interrupt logic
still actually works most of the time for these, but there are
circumstances where we can lost interrupts due to the incorrect
interrupt logic.
This patch, therefore, implements the correct xics level-sensitive
interrupt logic. The type of the interrupt is set when a device
allocates a new xics interrupt.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
The sPAPR PCI code defines a PCI device "spapr-pci-host-bridge-pci" which
is never used. This came over from the earlier bridge driver we used as
a template. Some other bridges appear on their own PCI bus as a device,
but that is not true of pSeries bridges, which are pure host to PCI with
no visible presence on the PCI side.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
The 'bars' constant array was used in experimental device allocation code
which is no longer necessary now that we always run the SLOF firmware.
This patch removes the now redundant variable.
Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
Most MemoryRegionOps already had the const attribute.
This patch adds it to the remaining ones.
Signed-off-by: Stefan Weil <sw@weilnetz.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
Replace device_init() with generalized type_init().
While at it, unify naming convention: type_init([$prefix_]register_types)
Also, type_init() is a function, so add preceding blank line where
necessary and don't put a semicolon after the closing brace.
Signed-off-by: Andreas Färber <afaerber@suse.de>
Cc: Anthony Liguori <anthony@codemonkey.ws>
Cc: malc <av1474@comtv.ru>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This was done in a mostly automated fashion. I did it in three steps and then
rebased it into a single step which avoids repeatedly touching every file in
the tree.
The first step was a sed-based addition of the parent type to the subclass
registration functions.
The second step was another sed-based removal of subclass registration functions
while also adding virtual functions from the base class into a class_init
function as appropriate.
Finally, a python script was used to convert the DeviceInfo structures and
qdev_register_subclass functions to TypeInfo structures, class_init functions,
and type_register_static calls.
We are almost fully converted to QOM after this commit.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This converts three devices because apic and ioapic are subclasses of sysbus.
Converting subclasses independently of their base class is prohibitively hard.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
We call pci_host_config_{read,write}_common() which perform PCI config
accesses. However they don't do all limit checking the way we expect
it to.
So let's introduce a small wrapper around them, making them behave the
way we would without touching generic code.
This patch is based on a patch by David Gibson which put this logic into
the generic code.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
Currently on the pseries machine the SLOF firmware is used normally,
but we bypass it when -kernel is specified. Having these two
different boot paths can cause some confusion.
In particular at present we need to "probe" the (emulated) PCI bus and
produce device tree nodes for the PCI devices in qemu, for the -kernel
case. In the SLOF case, it takes the device tree from qemu adds some
stuff to it then passes it on to the kernel.
It's been decided that a better approach is to always boot through
SLOF, even when using -kernel. WIth this approach we can leave PCI
probing and device node creation to SLOF in all cases which removes a
bunch of code in qemu, and avoids iterating the PCI devices from the
machine specific init code which we're not supposed to do.
This patch changes qemu to always boot through SLOF, and not to create
PCI nodes. Simultaneously it updates the included version of SLOF
(submodule and binary image) to one which supports (and requires) the
new approach.
The new SLOF version also includes a number of unrelated enhancements:
support for booting from virtio-pci devices and e1000, greatly
improved FCode support and many bugfixes. It also makes SLOF ready to
be used even when specifying a kernel on the qemu command line.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
The pseries machine expects a para-virtualized guest and so supplies RTAS
functions (via a hypercall) for performing PCI config space access.
Currently the implementation of these calls into
pci_default_{read,write}_config(). However this would be incorrect for
any PCI device which overrides the default config read/write functions.
AFAICT there's only one such device today, but we should still get it
right. In addition the pci_host_config_{read,write}_common() functions
which do correctly do this dispatch, perform bounds checking on the config
space address, lack of which currently leads to an exploitable bug.
This patch corrects the problem.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
On the pseries machine (which expexts a paravirtualized guest), guest
access to PCI config space is via host-provided RTAS functions. This
patch extends these RTAS functions to permit access to PCI extended
config space, as specified in PAPR.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
spapr_populate_pci_devices() containd a loop with PCI_NUM_REGIONS (7)
iterations. However this overruns the 'bars' global array, which only has
6 elements. In fact we only want to run this loop for things listed in the
bars array, so this patch corrects the loop bounds to reflect that.
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrzej Zaborowski <andrew.zaborowski@intel.com>
This patch adds a PCI bus to the pseries machine. This instantiates
the qemu generic PCI bus code, advertises a PCI host bridge in the
guest's device tree and implements the RTAS methods specified by PAPR
to access PCI config space. It also sets up the memory regions we
need to provide windows into the PCI memory and IO space, and
advertises those to the guest.
However, because qemu can't yet emulate an IOMMU, which is mandatory on
pseries, PCI devices which use DMA (i.e. most of them) will not work with
this code alone. Still, this is enough to support the virtio_pci device
(which probably _should_ use emulated PCI DMA, but is specced to use
direct hypervisor access to guest physical memory instead).
[agraf] remove typedef which could cause compile errors
Signed-off-by: Alexey Kardashevskiy <aik@au1.ibm.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Alexander Graf <agraf@suse.de>
