For the MMIO and PCI buses, drivers typically allocate an IRQ line for
their device before registering the device with the device tree for the
relevant bus.
This patch moves the IRQ allocation into the bus code, which is then
called directly by the device tree when a new device is registered.
IOPORT devices, however, tend to use hardcoded IRQs for legacy reasons,
so they are still required to deal with their interrupts (which also
require remapping for non-x86 architectures).
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
Rather than performing all config accesses via ioports, map in a 24-bit
memory-mapped configuration space directly below the PCI MMIO region.
This will allow architectures to support PCI without having to support
legacy ioports in the guest kernel. Instead, kvm tool can forward the
config accesses directly to the relevant ioport config callbacks.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
PCI devices are currently registered into the pci_devices array via the
pci__register function, which can then be indexed later by architecture
code to construct device tree nodes. For MMIO devices, there is no such
utility.
Rather than invent a similar mechanism for MMIO, this patch creates a
global device registration mechanism, which allows the device type to be
specified when registered or indexing a device. Current users of the pci
registration code are migrated to the new infrastructure and virtio MMIO
devices are registered at init time.
As part of the device registration, allocation of the device number is
moved out of irq__register_device and performed when adding the device
header to the relevant bus tree, allowing us to maintain separate device
numberspaces for each bus.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
Commit b50cac5 ("PCI/sysfs: add per pci device msi[x] irq listing (v5)") added
"struct kobject" dependency to <linux/msi.h> which breaks KVM tool build. As
the header file is no longer about MSI hardware, make a new header file for
"struct msi_ms".
Reported-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
This allows config space access in a more natural manner than clunky x86 IO ports,
and is useful for other architectures. Internally, the x86 IO port access uses
these new config space interfaces.
Signed-off-by: Matt Evans <matt@ozlabs.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
vesa, pci-shmem and virtio-pci devices need to set up config space with
little-endian conversions (as config space is LE). The pci_config_address
bitfield also needs to be reversed when building on BE systems.
Signed-off-by: Matt Evans <matt@ozlabs.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
Other pieces of kvmtool may be interested in PCI_MAX_DEVICES.
Signed-off-by: Matt Evans <matt@ozlabs.org>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
This patch simplifies passing around msi messages by using
'struct kvm_irq_routing_msi' for storing of msi messages instead
of passing all msi parameters around.
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
The patch adds an ivshmem device which can be used to share memory between
guests on the same host.
This implementation is lacking inter-guest communication which should be
implemented later once information regarding the client-server protocol is
gathered, though infrastructure used to add and remove clients already exists
in the patch (but isn't used anywhere).
Patch is based on David Evansky's shmem device.
Original-patch-by: David Evensky <evensky@sandia.gov>
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
Several fixes in this patch:
* Don't ignore function level and per-vector masking. We're not
supposed to signal when masked and not doing so will improve
performance a bit (in addition to behaving correctly).
* Implement the missing PBA array. 'lspci -vv' will now show the correct
output:
Capabilities: [40] MSI-X: Enable+ Count=5 Masked-
Vector table: BAR=1 offset=00000000
PBA: BAR=3 offset=00000000
* Checking whether MSI-X is enabled or not is done by probing the
corresponding bit within the PCI header instead of trying to track
the status ourselves.
* Fallback to INTx if MSI-X isn't enabled.
* Use correct max size for MSI-X table.
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
This patch adds a missing check of mask during probing of PCI BARs.
The missing check manifested in wrong address values for the BARs after the
initial probe.
Reported-by: David Evensky <evensky@dancer.ca.sandia.gov>
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
This patch implements basic MSI-X support for virtio-rng.
The device uses the virtio preferred method of working with MSI-X by
creating one vector for configuration and one vector for each vq in the
device.
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
PCI BAR probing is done in four steps:
1. Read address (and flags).
2. Mask BAR.
3. Read BAR again - Now the expected result is the size of the BAR.
4. Mask BAR with address.
So far, we have only took care of the first step. This means that the kernel
was using address as the size, causing a PCI allocation blunder.
This patch fixes the issue by passing a proper size after masking.
Signed-off-by: Sasha Levin <levinsasha928@gmail.com>
Signed-off-by: Pekka Enberg <penberg@kernel.org>
It still doesn't work because we need to fill .class
properly. But at least virtio probe function passes fine now.
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
This patch adds simple PCI Configuration Mechanism 1 probing emulation. It
doesn't expose any actual PCI devices yet.
Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>
Will Deacon