virtio_queue_get_desc_addr returns 64-bit hwaddr while int is usually 32-bit.
If returned hwaddr is not equal to 0 but least-significant 32 bits are
equal to 0 then this code will not actually stop running queue.
Signed-off-by: Yury Kotov <yury-kotov@yandex-team.ru>
Acked-by: Jia He <hejianet@gmail.com>
Cc: qemu-stable@nongnu.org
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Check region type first before casting the memory region
to IOMMUMemoryRegion. Otherwise QEMU will abort with below
error message when casting non-IOMMU memory region:
vhost_iommu_region_add: Object 0x561f28bce4f0 is not an
instance of type qemu:iommu-memory-region
Fixes: cb1efcf462 ("iommu: Add IOMMU index argument to notifier APIs")
Cc: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add support for multiple IOMMU indexes to the IOMMU notifier APIs.
When initializing a notifier with iommu_notifier_init(), the caller
must pass the IOMMU index that it is interested in. When a change
happens, the IOMMU implementation must pass
memory_region_notify_iommu() the IOMMU index that has changed and
that notifiers must be called for.
IOMMUs which support only a single index don't need to change.
Callers which only really support working with IOMMUs with a single
index can use the result of passing MEMTXATTRS_UNSPECIFIED to
memory_region_iommu_attrs_to_index().
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-id: 20180604152941.20374-3-peter.maydell@linaro.org
vDPA support, fix to vhost blk RO bit handling, some include path
cleanups, NFIT ACPI table.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
-----BEGIN PGP SIGNATURE-----
iQEcBAABAgAGBQJbEXNvAAoJECgfDbjSjVRpc8gH/R8xrcFrV+k9wwbgYcOcGb6Y
LWjseE31pqJcxRV80vLOdzYEuLStZQKQQY7xBDMlA5vdyvZxIA6FLO2IsiJSbFAk
EK8pclwhpwQAahr8BfzenabohBv2UO7zu5+dqSvuJCiMWF3jGtPAIMxInfjXaOZY
odc1zY2D2EgsC7wZZ1hfraRbISBOiRaez9BoGDKPOyBY9G1ASEgxJgleFgoBLfsK
a1XU+fDM6hAVdxftfkTm0nibyf7PWPDyzqghLqjR9WXLvZP3Cqud4p8N29mY51pR
KSTjA4FYk6Z9EVMltyBHfdJs6RQzglKjxcNGdlrvacDfyFi79fGdiosVllrjfJM=
=3+V0
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/mst/tags/for_upstream' into staging
acpi, vhost, misc: fixes, features
vDPA support, fix to vhost blk RO bit handling, some include path
cleanups, NFIT ACPI table.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
# gpg: Signature made Fri 01 Jun 2018 17:25:19 BST
# gpg: using RSA key 281F0DB8D28D5469
# gpg: Good signature from "Michael S. Tsirkin <mst@kernel.org>"
# gpg: aka "Michael S. Tsirkin <mst@redhat.com>"
# Primary key fingerprint: 0270 606B 6F3C DF3D 0B17 0970 C350 3912 AFBE 8E67
# Subkey fingerprint: 5D09 FD08 71C8 F85B 94CA 8A0D 281F 0DB8 D28D 5469
* remotes/mst/tags/for_upstream: (31 commits)
vhost-blk: turn on pre-defined RO feature bit
ACPI testing: test NFIT platform capabilities
nvdimm, acpi: support NFIT platform capabilities
tests/.gitignore: add entry for generated file
arch_init: sort architectures
ui: use local path for local headers
qga: use local path for local headers
colo: use local path for local headers
migration: use local path for local headers
usb: use local path for local headers
sd: fix up include
vhost-scsi: drop an unused include
ppc: use local path for local headers
rocker: drop an unused include
e1000e: use local path for local headers
ioapic: fix up includes
ide: use local path for local headers
display: use local path for local headers
trace: use local path for local headers
migration: drop an unused include
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
As part of plumbing MemTxAttrs down to the IOMMU translate method,
add MemTxAttrs as an argument to address_space_get_iotlb_entry().
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20180521140402.23318-12-peter.maydell@linaro.org
This patch introduces a vhost op for vhost backends to allow
them to filter the memory sections that they can handle.
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add some trace points for IOTLB translation for vhost. After vhost-user
is setup, the only IO path that QEMU will participate should be the
IOMMU translation, so it'll be good we can track this with explicit
timestamps when needed to see how long time we take to do the
translation, and whether there's anything stuck inside. It might be
useful for triaging vhost-user problems.
Signed-off-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When IOMMU is enabled, we store virtqueue metadata as iova (though it
may has _phys suffix) and access them through dma helpers. Any
translation failures could be reported by IOMMU.
In this case, trying to validate iova against gpa won't work and will
cause a false error reporting. So this patch bypasses the ring
verification if IOMMU is enabled which is similar to the behavior
before 0ca1fd2d68 that calls vhost_memory_map() which is a nop when
IOMMU is enabled.
Fixes: 0ca1fd2d68 ("vhost: Simplify ring verification checks")
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
My rework of section adding combines overlapping or adjoining regions,
but checks they're actually the same underlying RAM block.
Fix the case where two blocks adjoin but don't overlap; that new region
should get added (but not combined), but my previous patch was disallowing it.
Fixes: c1ece84e7c
Reported-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
As soon as vhost-user init is done, the backend may send
VHOST_USER_SLAVE_CONFIG_CHANGE_MSG, so let's set the
notification callback before it.
Also, it will be used to know whether the device supports
the config feature to advertize it or not.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Changpeng Liu <changpeng.liu@intel.com>
Align RAMBlocks to page size alignment, and adjust the merging code
to deal with partial overlap due to that alignment.
This is needed for postcopy so that we can place/fetch whole hugepages
when under userfault.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In commit 0ca1fd2d68 ("vhost: Simplify ring verification checks"),
it checks the virtqueue desc mapping for 3 times.
Fixed: commit 0ca1fd2d68 ("vhost: Simplify ring verification checks")
Signed-off-by: Jia He <jia.he@hxt-semitech.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
In our Armv8a server, we try to configure the vhost scsi but fail
to boot up the guest (-machine virt-2.10). The guest's boot failure
is very early, even earlier than grub.
There are 3 virtqueues (ctrl, event and cmd) for virtio scsi device,
but ovmf and seabios will only set the physical address for the 3rd
one (cmd). Then in vhost_virtqueue_start(), virtio_queue_get_desc_addr
will be 0 for ctrl and event vq when qemu negotiates with ovmf. So
vhost_memory_map fails with ENOMEM.
This patch just fixs it by early quitting the virtqueue start/stop
when virtio_queue_get_desc_addr is 0.
Btw, after guest kernel starts, all the 3 queues will be initialized
and set address correctly.
Already tested on Arm64 and X86_64 qemu.
Signed-off-by: Jia He <jia.he@hxt-semitech.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Since used_memslots will be updated to the actual value after
registering memory listener for the first time, move the
memslots limit checking to the right place.
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Signed-off-by: Jay Zhou <jianjay.zhou@huawei.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Move the log_dirty check into vhost_section.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Now that the olf vhost_set_memory code is gone, the _nop and _add
callbacks are identical and can be merged. The _del callback is
no longer needed.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Remove the old update mechanism, vhost_set_memory, and the functions
and flags it used.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Compare the sections list that's just been generated, and if it's
different from the old one regenerate the region list.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
As sections are reported by the listener to the _nop and _add
methods, add them to the temporary section list but now merge them
with the previous section if the new one abuts and the backend allows.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
vhost_verify_ring_mappings() were used to verify that
rings are still accessible and related memory hasn't
been moved after flatview is updated.
It was doing checks by mapping ring's GPA+len and
checking that HVA hadn't changed with new memory map.
To avoid maybe expensive mapping call, we were
identifying address range that changed and were doing
mapping only if ring was in changed range.
However it's not neccessary to perform ring's GPA
mapping as we already have its current HVA and all
we need is to verify that ring's GPA translates to
the same HVA in updated flatview.
This will allow the following patches to simplify the range
comparison that was previously needed to avoid expensive
verify_ring_mapping calls.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
with modifications by:
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Igor spotted that there's a race, where a region that's unref'd
in a _del callback might be free'd before the set_mem_table call in
the _commit callback, and thus the vhost might end up using free memory.
Fix this by building a complete temporary sections list, ref'ing every
section (during add and nop) and then unref'ing the whole list right
at the end of commit.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The virtio_bus_set_host_notifier function no longer calls
event_notifier_cleanup when a event notifier is removed.
The commit updates the code to match the new behavior and calls
virtio_bus_cleanup_host_notifier after the notifier was de-assign
and no longer in use.
This change is a preparation to allow executing the
virtio_bus_set_host_notifier function in a memory region
transaction.
Signed-off-by: Gal Hammer <ghammer@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Tested-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This reverts commit 0750b06021.
Follow up patches are reworking the memory listeners, the new mechanism
will add its own set of traces.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This will allow callers to silence error report when the call is
allowed to failed.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180201132757.23063-2-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When unregister memory listeners, we should call, e.g.,
region_del() (and possibly other undo operations) on every existing
memory region sections there, otherwise we may leak resources that are
held during the region_add(). This patch undo the stuff for the
listeners, which emulates the case when the address space is set from
current to an empty state.
I found this problem when debugging a refcount leak issue that leads to
a device unplug event lost (please see the "Bug:" line below). In that
case, the leakage of resource is the PCI BAR memory region refcount.
And since memory regions are not keeping their own refcount but onto
their owners, so the vfio-pci device's (who is the owner of the PCI BAR
memory regions) refcount is leaked, and event missing.
We had encountered similar issues before and fixed in other
way (ee4c112846, "vhost: Release memory references on cleanup"). This
patch can be seen as a more high-level fix of similar problems that are
caused by the resource leaks from memory listeners. So now we can remove
the explicit unref of memory regions since that'll be done altogether
during unregistering of listeners now.
Bug: https://bugzilla.redhat.com/show_bug.cgi?id=1531393
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180122060244.29368-5-peterx@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Trace these operations on two memory listeners. It helps to verify the
new memory listener fix, and good to keep them there.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180122060244.29368-2-peterx@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add VHOST_USER_GET_CONFIG/VHOST_USER_SET_CONFIG messages which can be
used for live migration of vhost user devices, also vhost user devices
can benefit from the messages to get/set virtio config space from/to the
I/O target. For the purpose to support virtio config space change,
VHOST_USER_SLAVE_CONFIG_CHANGE_MSG message is added as the event notifier
in case virtio config space change in the slave I/O target.
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Since commit f1f9e6c5 "vhost: adapt vhost_verify_ring_mappings() to
virtio 1 ring layout", we check the mapping of each part (descriptor
table, available ring and used ring) of each virtqueue separately.
The checking of a part is done by the vhost_verify_ring_part_mapping()
function: it returns either 0 on success or a negative errno if the
part cannot be mapped at the same place.
Unfortunately, the vhost_verify_ring_mappings() function checks its
return value the other way round. It means that we either:
- only verify the descriptor table of the first virtqueue, and if it
is valid we ignore all the other mappings
- or ignore all broken mappings until we reach a valid one
ie, we only raise an error if all mappings are broken, and we consider
all mappings are valid otherwise (false success), which is obviously
wrong.
This patch ensures that vhost_verify_ring_mappings() only returns
success if ALL mappings are okay.
Reported-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
vhost_virtqueue_stop() gets avail index value from the backend,
except if the backend is not responding.
It happens when the backend crashes, and in this case, internal
state of the virtio queue is inconsistent, making packets
to corrupt the vring state.
With a Linux guest, it results in following error message on
backend reconnection:
[ 22.444905] virtio_net virtio0: output.0:id 0 is not a head!
[ 22.446746] net enp0s3: Unexpected TXQ (0) queue failure: -5
[ 22.476360] net enp0s3: Unexpected TXQ (0) queue failure: -5
Fixes: 283e2c2adc ("net: virtio-net discards TX data after link down")
Cc: qemu-stable@nongnu.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
vhost_log_put() is called to decomission the dirty log between qemu and
a vhost device when stopping the device. Such a call can happen from
migration_completion().
Present code sets dev->log_size to zero too early in vhost_log_put(),
causing the sync check to always return false. As a consequence, the
last pass on the dirty bitmap never happens at the end of migration.
If a vhost device was busy (writing to guest memory) until the last
moments before vhost_virtqueue_stop(), this error will result in guest
memory corruption (at least) following migrations.
Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
vhost registers a MemoryListener where it adds and removes references
to MemoryRegions as the MemoryRegionSections pass through. The
region_add callback is invoked for each existing section when the
MemoryListener is registered, but unregistering the MemoryListener
performs no reciprocal region_del callback. It's therefore the
owner of the MemoryListener's responsibility to cleanup any persistent
changes, such as these memory references, after unregistering.
The consequence of this bug is that if we have both a vhost device
and a vfio device, the vhost device will reference any mmap'd MMIO of
the vfio device via this MemoryListener. If the vhost device is then
removed, those references remain outstanding. If we then attempt to
remove the vfio device, it never gets finalized and the only way to
release the kernel file descriptors is to terminate the QEMU process.
Fixes: dfde4e6e1a ("memory: add ref/unref calls")
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-stable@nongnu.org # v1.6.0+
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
I used the clang-tidy qemu-round check to generate the fix:
https://github.com/elmarco/clang-tools-extra
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Richard Henderson <rth@twiddle.net>
This patch reworks IOTLB messaging to prepare for vhost-user
device IOTLB support.
IOTLB messages handling is extracted from vhost-kernel backend,
so that only the messages transport remains backend specifics.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Some backends might want to know when things went wrong.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This allows us to remove lots of includes of migration/migration.h
Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
In this patch, IOMMUNotifier.{start|end} are introduced to store section
information for a specific notifier. When notification occurs, we not
only check the notification type (MAP|UNMAP), but also check whether the
notified iova range overlaps with the range of specific IOMMU notifier,
and skip those notifiers if not in the listened range.
When removing an region, we need to make sure we removed the correct
VFIOGuestIOMMU by checking the IOMMUNotifier.start address as well.
This patch is solving the problem that vfio-pci devices receive
duplicated UNMAP notification on x86 platform when vIOMMU is there. The
issue is that x86 IOMMU has a (0, 2^64-1) IOMMU region, which is
splitted by the (0xfee00000, 0xfeefffff) IRQ region. AFAIK
this (splitted IOMMU region) is only happening on x86.
This patch also helps vhost to leverage the new interface as well, so
that vhost won't get duplicated cache flushes. In that sense, it's an
slight performance improvement.
Suggested-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <1491562755-23867-2-git-send-email-peterx@redhat.com>
[ehabkost: included extra vhost_iommu_region_del() change from Peter Xu]
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
We assumes the iommu_ops were attached to the root region of address
space. This may not be true for all kinds of IOMMU implementation and
especially after commit 3716d5902d ("pci: introduce a bus master
container"). So fix this by not assuming as->root has iommu_ops,
instead depending on the regions reported by memory listener through:
- register a memory listener to dma_as
- during region_add, if it's a region of IOMMU, register a specific
IOMMU notifier, and store all notifiers in a list.
- during region_del, compare and delete the IOMMU notifier from the list
This is also a must for making vhost device IOTLB works for all types
of IOMMUs. Note, since we register one notifier during each
.region_add, the IOTLB may be flushed more than one times, this is
suboptimal and could be optimized in the future.
Reported-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Fixes: 3716d5902d ("pci: introduce a bus master container")
Cc: Peter Xu <peterx@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Tested-by: Maxime Coquelin <maxime.coquelin@redhat.com>
vhost does not support RO protections on memory at the moment - adding
ROMs would mean that e.g. a buggy guest might change them in-memory - a
condition from which guest reset does not recover. Not nice.
We also definitely don't want to try logging writes into ROMs -
in particular guests set very high addresses for ROM BARs
so logging these writes would waste a lot of memory.
Maybe ROMs could be supported with the iotlb variant -
not sure, but there seems to be no good reason for virtio
to try to do DMA from ROM. So let's just skip ROM memory.
Suggested-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Tested-by: Laurent Vivier <lvivier@redhat.com>
If a migration is already in progress and somebody attempts
to add a migration blocker, this should rightly fail.
Add an errp parameter and a retcode return value to migrate_add_blocker.
Signed-off-by: John Snow <jsnow@redhat.com>
Signed-off-by: Ashijeet Acharya <ashijeetacharya@gmail.com>
Message-Id: <1484566314-3987-5-git-send-email-ashijeetacharya@gmail.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Acked-by: Greg Kurz <groug@kaod.org>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Merged with recent 'Allow invtsc migration' change
This patches implements Device IOTLB support for vhost kernel. This is
done through:
1) switch to use dma helpers when map/unmap vrings from vhost codes
2) introduce a set of VhostOps to:
- setting up device IOTLB request callback
- processing device IOTLB request
- processing device IOTLB invalidation
2) kernel support for Device IOTLB API:
- allow vhost-net to query the IOMMU IOTLB entry through eventfd
- enable the ability for qemu to update a specified mapping of vhost
- through ioctl.
- enable the ability to invalidate a specified range of iova for the
device IOTLB of vhost through ioctl. In x86/intel_iommu case this is
triggered through iommu memory region notifier from device IOTLB
invalidation descriptor processing routine.
With all the above, kernel vhost_net can co-operate with userspace
IOMMU. For vhost-user, the support could be easily done on top by
implementing the VhostOps.
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Make virtio queue suitable for push operation from qemu
after vhost was stopped.
Signed-off-by: Yuri Benditovich <yuri.benditovich@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Following the recent refactoring of virtio notifiers [1], more specifically
the patch ed08a2a0b ("virtio: use virtio_bus_set_host_notifier to
start/stop ioeventfd") that uses virtio_bus_set_host_notifier [2]
by default, core virtio code requires 'ioeventfd_started' to be set
to true/false when the host notifiers are configured.
When vhost is stopped and started, however, there is a stop followed by
another start. Since ioeventfd_started was never set to true, the 'stop'
operation triggered by virtio_bus_set_host_notifier() will not result
in a call to virtio_pci_ioeventfd_assign(assign=false). This leaves
the memory regions with stale notifiers and results on the next start
triggering the following assertion:
kvm_mem_ioeventfd_add: error adding ioeventfd: File exists
Aborted
This patch reintroduces (hopefully in a cleaner way) the concept
that was present with ioeventfd_disabled before the refactoring.
When ioeventfd_grabbed>0, ioeventfd_started tracks whether ioeventfd
should be enabled or not, but ioeventfd is actually not started at
all until vhost releases the host notifiers.
[1] http://lists.nongnu.org/archive/html/qemu-devel/2016-10/msg07748.html
[2] http://lists.nongnu.org/archive/html/qemu-devel/2016-10/msg07760.html
Reported-by: Felipe Franciosi <felipe@nutanix.com>
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Reported-by: Alex Williamson <alex.williamson@redhat.com>
Fixes: ed08a2a0b ("virtio: use virtio_bus_set_host_notifier to start/stop ioeventfd")
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Tested-by: Farhan Ali <alifm@linux.vnet.ibm.com>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The legacy vring layout is not used anymore as we use the separate
mappings even for legacy devices.
This patch simply removes it.
This also fixes a bug with virtio 1 devices when the vring descriptor table
is mapped at a higher address than the used vring because the following
function may return an insanely great value:
hwaddr virtio_queue_get_ring_size(VirtIODevice *vdev, int n)
{
return vdev->vq[n].vring.used - vdev->vq[n].vring.desc +
virtio_queue_get_used_size(vdev, n);
}
and the mapping fails.
Signed-off-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
With virtio 1, the vring layout is split in 3 separate regions of
contiguous memory for the descriptor table, the available ring and the
used ring, as opposed with legacy virtio which uses a single region.
In case of memory re-mapping, the code ensures it doesn't affect the
vring mapping. This is done in vhost_verify_ring_mappings() which assumes
the device is legacy.
This patch changes vhost_verify_ring_mappings() to check the mappings of
each part of the vring separately.
This works for legacy mappings as well.
Cc: qemu-stable@nongnu.org
Signed-off-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Commit 31190ed7 added a migration blocker in vhost_dev_init() to
check if memfd would succeed. It is better if this blocker first
checks if vhost backend requires shared log. This will avoid a
situation where a blocker is added inappropriately (e.g. shared
log allocation fails when vhost backend doesn't support it).
Signed-off-by: Rafael David Tinoco <rafael.tinoco@canonical.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Now that there is not anymore a switch from the generic ioeventfd handler
to the dataplane handler, virtio_bus_set_host_notifier(assign=true) is
always called with !bus->ioeventfd_started, hence virtio_bus_stop_ioeventfd
does nothing in this case. Move the invocation to vhost.c, which is the
only place that needs it.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This simplifies the code and removes the ioeventfd_started
and ioeventfd_set_started callback. The only difference is
in how virtio-ccw handles an error---it doesn't disable
ioeventfd forever anymore. It was the only backend to do
so, and if desired this behavior should be implemented in
virtio-bus.c.
Instead of ioeventfd_started, the ioeventfd_assign callback now
determines whether the virtio bus supports host notifiers.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
We used to set vring call fd unconditionally even if guest driver does
not use MSIX for this vritqueue at all. This will cause lots of
unnecessary userspace access and other checks for drivers does not use
interrupt at all (e.g virtio-net pmd). So check and clean vring call
fd if guest does not use any vector for this virtqueue at
all.
Perf diffs (on rx) shows lots of cpus wasted on vhost_signal() were saved:
#
28.12% -27.82% [vhost] [k] vhost_signal
14.44% -1.69% [kernel.vmlinux] [k] copy_user_generic_string
7.05% +1.53% [kernel.vmlinux] [k] __free_page_frag
6.51% +5.53% [vhost] [k] vhost_get_vq_desc
...
Pktgen tests shows 15.8% improvement on rx pps and 6.5% on tx pps.
Before: RX 2.08Mpps TX 1.35Mpps
After: RX 2.41Mpps TX 1.44Mpps
Signed-off-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
