Not checking this can lead to invalid dev->vdev member access in
vhost_device_iotlb_miss if backend issue an iotlb message in a bad
timing, either maliciously or by a bug.
Reproduced rebooting a guest with testpmd in txonly forward mode.
#0 0x0000559ffff94394 in vhost_device_iotlb_miss (
dev=dev@entry=0x55a0012f6680, iova=10245279744, write=1)
at ../hw/virtio/vhost.c:1013
#1 0x0000559ffff9ac31 in vhost_backend_handle_iotlb_msg (
imsg=0x7ffddcfd32c0, dev=0x55a0012f6680)
at ../hw/virtio/vhost-backend.c:411
#2 vhost_backend_handle_iotlb_msg (dev=dev@entry=0x55a0012f6680,
imsg=imsg@entry=0x7ffddcfd32c0)
at ../hw/virtio/vhost-backend.c:404
#3 0x0000559fffeded7b in slave_read (opaque=0x55a0012f6680)
at ../hw/virtio/vhost-user.c:1464
#4 0x000055a0000c541b in aio_dispatch_handler (
ctx=ctx@entry=0x55a0010a2120, node=0x55a0012d9e00)
at ../util/aio-posix.c:329
Fixes: 020e571b8b ("vhost: rework IOTLB messaging")
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20210129090728.831208-1-eperezma@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This property was only required for compatibility reasons in the
pc-1.0 machine type and earlier. Now that these machine types have
been removed, the property is not useful anymore.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <20210203171832.483176-4-thuth@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: David Hildenbrand <david@redhat.com>
Previous work on dev-iotlb message broke vhost on either SMMU or virtio-iommu
since dev-iotlb (or PCIe ATS) is not yet supported for those archs.
An initial idea is that we can let IOMMU to export this information to vhost so
that vhost would know whether the vIOMMU would support dev-iotlb, then vhost
can conditionally register to dev-iotlb or the old iotlb way. We can work
based on some previous patch to introduce PCIIOMMUOps as Yi Liu proposed [1].
However it's not as easy as I thought since vhost_iommu_region_add() does not
have a PCIDevice context at all since it's completely a backend. It seems
non-trivial to pass over a PCI device to the backend during init. E.g. when
the IOMMU notifier registered hdev->vdev is still NULL.
To make the fix smaller and easier, this patch goes the other way to leverage
the flag_changed() hook of vIOMMUs so that SMMU and virtio-iommu can trap the
dev-iotlb registration and fail it. Then vhost could try the fallback solution
as using UNMAP invalidation for it's translations.
[1] https://lore.kernel.org/qemu-devel/1599735398-6829-4-git-send-email-yi.l.liu@intel.com/
Reported-by: Eric Auger <eric.auger@redhat.com>
Fixes: b68ba1ca57
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Tested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20210204191228.187550-1-peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This patch adds trace events for virtio-pmem functionality.
Adding trace events for virtio pmem request, reponse and host
side fsync functionality.
Signed-off-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Message-Id: <20201117115705.32195-1-pankaj.gupta.linux@gmail.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Address space is destroyed without proper removal of its listeners with
current code. They are expected to be removed in
virtio_device_instance_finalize [1], but qemu calls it through
object_deinit, after address_space_destroy call through
device_set_realized [2].
Move it to virtio_device_unrealize, called before device_set_realized
[3] and making it symmetric with memory_listener_register in
virtio_device_realize.
v2: Delete no-op call of virtio_device_instance_finalize.
Add backtraces.
[1]
#0 virtio_device_instance_finalize (obj=0x555557de5120)
at /home/qemu/include/hw/virtio/virtio.h:71
#1 0x0000555555b703c9 in object_deinit (type=0x555556639860,
obj=<optimized out>) at ../qom/object.c:671
#2 object_finalize (data=0x555557de5120) at ../qom/object.c:685
#3 object_unref (objptr=0x555557de5120) at ../qom/object.c:1184
#4 0x0000555555b4de9d in bus_free_bus_child (kid=0x555557df0660)
at ../hw/core/qdev.c:55
#5 0x0000555555c65003 in call_rcu_thread (opaque=opaque@entry=0x0)
at ../util/rcu.c:281
Queued by:
#0 bus_remove_child (bus=0x555557de5098,
child=child@entry=0x555557de5120) at ../hw/core/qdev.c:60
#1 0x0000555555b4ee31 in device_unparent (obj=<optimized out>)
at ../hw/core/qdev.c:984
#2 0x0000555555b70465 in object_finalize_child_property (
obj=<optimized out>, name=<optimized out>, opaque=0x555557de5120)
at ../qom/object.c:1725
#3 0x0000555555b6fa17 in object_property_del_child (
child=0x555557de5120, obj=0x555557ddcf90) at ../qom/object.c:645
#4 object_unparent (obj=0x555557de5120) at ../qom/object.c:664
#5 0x0000555555b4c071 in bus_unparent (obj=<optimized out>)
at ../hw/core/bus.c:147
#6 0x0000555555b70465 in object_finalize_child_property (
obj=<optimized out>, name=<optimized out>, opaque=0x555557de5098)
at ../qom/object.c:1725
#7 0x0000555555b6fa17 in object_property_del_child (
child=0x555557de5098, obj=0x555557ddcf90) at ../qom/object.c:645
#8 object_unparent (obj=0x555557de5098) at ../qom/object.c:664
#9 0x0000555555b4ee19 in device_unparent (obj=<optimized out>)
at ../hw/core/qdev.c:981
#10 0x0000555555b70465 in object_finalize_child_property (
obj=<optimized out>, name=<optimized out>, opaque=0x555557ddcf90)
at ../qom/object.c:1725
#11 0x0000555555b6fa17 in object_property_del_child (
child=0x555557ddcf90, obj=0x55555685da10) at ../qom/object.c:645
#12 object_unparent (obj=0x555557ddcf90) at ../qom/object.c:664
#13 0x00005555558dc331 in pci_for_each_device_under_bus (
opaque=<optimized out>, fn=<optimized out>, bus=<optimized out>)
at ../hw/pci/pci.c:1654
[2]
Optimizer omits pci_qdev_unrealize, called by device_set_realized, and
do_pci_unregister_device, called by pci_qdev_unrealize and caller of
address_space_destroy.
#0 address_space_destroy (as=0x555557ddd1b8)
at ../softmmu/memory.c:2840
#1 0x0000555555b4fc53 in device_set_realized (obj=0x555557ddcf90,
value=<optimized out>, errp=0x7fffeea8f1e0)
at ../hw/core/qdev.c:850
#2 0x0000555555b6eaa6 in property_set_bool (obj=0x555557ddcf90,
v=<optimized out>, name=<optimized out>, opaque=0x555556650ba0,
errp=0x7fffeea8f1e0) at ../qom/object.c:2255
#3 0x0000555555b70e07 in object_property_set (
obj=obj@entry=0x555557ddcf90,
name=name@entry=0x555555db99df "realized",
v=v@entry=0x7fffe46b7500,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1400
#4 0x0000555555b73c5f in object_property_set_qobject (
obj=obj@entry=0x555557ddcf90,
name=name@entry=0x555555db99df "realized",
value=value@entry=0x7fffe44f6180,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/qom-qobject.c:28
#5 0x0000555555b71044 in object_property_set_bool (
obj=0x555557ddcf90, name=0x555555db99df "realized",
value=<optimized out>, errp=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1470
#6 0x0000555555921cb7 in pcie_unplug_device (bus=<optimized out>,
dev=0x555557ddcf90,
opaque=<optimized out>) at /home/qemu/include/hw/qdev-core.h:17
#7 0x00005555558dc331 in pci_for_each_device_under_bus (
opaque=<optimized out>, fn=<optimized out>,
bus=<optimized out>) at ../hw/pci/pci.c:1654
[3]
#0 virtio_device_unrealize (dev=0x555557de5120)
at ../hw/virtio/virtio.c:3680
#1 0x0000555555b4fc63 in device_set_realized (obj=0x555557de5120,
value=<optimized out>, errp=0x7fffee28df90)
at ../hw/core/qdev.c:850
#2 0x0000555555b6eab6 in property_set_bool (obj=0x555557de5120,
v=<optimized out>, name=<optimized out>, opaque=0x555556650ba0,
errp=0x7fffee28df90) at ../qom/object.c:2255
#3 0x0000555555b70e17 in object_property_set (
obj=obj@entry=0x555557de5120,
name=name@entry=0x555555db99ff "realized",
v=v@entry=0x7ffdd8035040,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1400
#4 0x0000555555b73c6f in object_property_set_qobject (
obj=obj@entry=0x555557de5120,
name=name@entry=0x555555db99ff "realized",
value=value@entry=0x7ffdd8035020,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/qom-qobject.c:28
#5 0x0000555555b71054 in object_property_set_bool (
obj=0x555557de5120, name=name@entry=0x555555db99ff "realized",
value=value@entry=false, errp=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1470
#6 0x0000555555b4edc5 in qdev_unrealize (dev=<optimized out>)
at ../hw/core/qdev.c:403
#7 0x0000555555b4c2a9 in bus_set_realized (obj=<optimized out>,
value=<optimized out>, errp=<optimized out>)
at ../hw/core/bus.c:204
#8 0x0000555555b6eab6 in property_set_bool (obj=0x555557de5098,
v=<optimized out>, name=<optimized out>, opaque=0x555557df04c0,
errp=0x7fffee28e0a0) at ../qom/object.c:2255
#9 0x0000555555b70e17 in object_property_set (
obj=obj@entry=0x555557de5098,
name=name@entry=0x555555db99ff "realized",
v=v@entry=0x7ffdd8034f50,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1400
#10 0x0000555555b73c6f in object_property_set_qobject (
obj=obj@entry=0x555557de5098,
name=name@entry=0x555555db99ff "realized",
value=value@entry=0x7ffdd8020630,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/qom-qobject.c:28
#11 0x0000555555b71054 in object_property_set_bool (
obj=obj@entry=0x555557de5098,
name=name@entry=0x555555db99ff "realized",
value=value@entry=false, errp=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1470
#12 0x0000555555b4c725 in qbus_unrealize (
bus=bus@entry=0x555557de5098) at ../hw/core/bus.c:178
#13 0x0000555555b4fc00 in device_set_realized (obj=0x555557ddcf90,
value=<optimized out>, errp=0x7fffee28e1e0)
at ../hw/core/qdev.c:844
#14 0x0000555555b6eab6 in property_set_bool (obj=0x555557ddcf90,
v=<optimized out>, name=<optimized out>, opaque=0x555556650ba0,
errp=0x7fffee28e1e0) at ../qom/object.c:2255
#15 0x0000555555b70e17 in object_property_set (
obj=obj@entry=0x555557ddcf90,
name=name@entry=0x555555db99ff "realized",
v=v@entry=0x7ffdd8020560,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1400
#16 0x0000555555b73c6f in object_property_set_qobject (
obj=obj@entry=0x555557ddcf90,
name=name@entry=0x555555db99ff "realized",
value=value@entry=0x7ffdd8020540,
errp=errp@entry=0x5555565bbf38 <error_abort>)
at ../qom/qom-qobject.c:28
#17 0x0000555555b71054 in object_property_set_bool (
obj=0x555557ddcf90, name=0x555555db99ff "realized",
value=<optimized out>, errp=0x5555565bbf38 <error_abort>)
at ../qom/object.c:1470
#18 0x0000555555921cb7 in pcie_unplug_device (bus=<optimized out>,
dev=0x555557ddcf90, opaque=<optimized out>)
at /home/qemu/include/hw/qdev-core.h:17
#19 0x00005555558dc331 in pci_for_each_device_under_bus (
opaque=<optimized out>, fn=<optimized out>, bus=<optimized out>)
at ../hw/pci/pci.c:1654
Fixes: c611c76417 ("virtio: add MemoryListener to cache ring translations")
Buglink: https://bugs.launchpad.net/qemu/+bug/1912846
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Message-Id: <20210125192505.390554-1-eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In the kernel, virtio_gpu_init() uses virtio_get_shm_region()
since
commit 6076a9711dc5 ("drm/virtio: implement blob resources: probe for host visible region")
but vm_get_shm_region() unconditionally uses VIRTIO_MMIO_SHM_SEL to
get the address and the length of the region.
commit 38e895487afc ("virtio: Implement get_shm_region for MMIO transport"
As this is not implemented in QEMU, address and length are 0 and passed
as is to devm_request_mem_region() that triggers a crash:
[drm:virtio_gpu_init] *ERROR* Could not reserve host visible region
Unable to handle kernel NULL pointer dereference at virtual address (ptrval)
According to the comments in the kernel, a non existent shared region
has a length of (u64)-1.
This is what we return now with this patch to disable the region.
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20201220163539.2255963-1-laurent@vivier.eu>
Acked-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
virtio-fs qualifies as a bootable device minimally under OVMF, but
currently the necessary "bootindex" property is missing. Add the property.
Expose the property only in the PCI device, for now. There is no boot
support for virtiofs on s390x (ccw) for the time being [1] [2], so leave
the CCW device unchanged. Add the property to the base device still,
because adding the alias to the CCW device later will be easier this way
[3].
[1] https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01745.html
[2] https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01870.html
[3] https://lists.gnu.org/archive/html/qemu-devel/2021-01/msg01751.html
Example OpenFirmware device path for the "vhost-user-fs-pci" device in the
"bootorder" fw_cfg file:
/pci@i0cf8/pci-bridge@1,6/pci1af4,105a@0/filesystem@0
Cc: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Ján Tomko <jtomko@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: virtio-fs@redhat.com
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20210112131603.12686-1-lersek@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This commit is the result of running the timer-del-timer-free.cocci
script on the whole source tree.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Corey Minyard <cminyard@mvista.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-id: 20201215154107.3255-4-peter.maydell@linaro.org
Commit 8118f0950f "migration: Append JSON description of migration
stream" needs a JSON writer. The existing qobject_to_json() wasn't a
good fit, because it requires building a QObject to convert. Instead,
migration got its very own JSON writer, in commit 190c882ce2 "QJSON:
Add JSON writer". It tacitly limits numbers to int64_t, and strings
contents to characters that don't need escaping, unlike
qobject_to_json().
The previous commit factored the JSON writer out of qobject_to_json().
Replace migration's JSON writer by it.
Cc: Juan Quintela <quintela@redhat.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20201211171152.146877-17-armbru@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Move the property types and property macros implemented in
qdev-properties-system.c to a new qdev-properties-system.h
header.
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20201211220529.2290218-16-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Generalize the qdev_hotplug variable to the different phases of
machine initialization. We would like to allow different
monitor commands depending on the phase.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Added AER capability for virtio-pci devices.
Also added property for devices, by default AER is disabled.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Message-Id: <20201203110713.204938-3-andrew@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Removed hardcoded offset for ats. Added cap offset counter
for future capabilities like AER.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Message-Id: <20201203110713.204938-2-andrew@daynix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
If we find a queue with an inconsistent guest index value, explicitly mark the
device as needing a reset - and broken - via virtio_error().
There's at least one driver implementation - the virtio-win NetKVM driver - that
is able to handle a VIRTIO_CONFIG_S_NEEDS_RESET notification and successfully
restore the device to a working state. Other implementations do not correctly
handle this, but as the VQ is not in a functional state anyway, this is still
worth doing.
Signed-off-by: John Levon <john.levon@nutanix.com>
Message-Id: <20201120185103.GA442386@sent>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This allows us to differentiate between regular IOMMU map/unmap events
and DEVIOTLB unmap. Doing so, notifiers that only need device IOTLB
invalidations will not receive regular IOMMU unmappings.
Adapt intel and vhost to use it.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20201116165506.31315-4-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This way we can tell between regular IOMMUTLBEntry (entry of IOMMU
hardware) and notifications.
In the notifications, we set explicitly if it is a MAPs or an UNMAP,
instead of trusting in entry permissions to differentiate them.
Signed-off-by: Eugenio Pérez <eperezma@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20201116165506.31315-3-eperezma@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Matthew Rosato <mjrosato@linux.ibm.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
QEMU currently truncates the mmap_offset field when sending
VHOST_USER_ADD_MEM_REG and VHOST_USER_REM_MEM_REG messages. The struct
layout looks like this:
typedef struct VhostUserMemoryRegion {
uint64_t guest_phys_addr;
uint64_t memory_size;
uint64_t userspace_addr;
uint64_t mmap_offset;
} VhostUserMemoryRegion;
typedef struct VhostUserMemRegMsg {
uint32_t padding;
/* WARNING: there is a 32-bit hole here! */
VhostUserMemoryRegion region;
} VhostUserMemRegMsg;
The payload size is calculated as follows when sending the message in
hw/virtio/vhost-user.c:
msg->hdr.size = sizeof(msg->payload.mem_reg.padding) +
sizeof(VhostUserMemoryRegion);
This calculation produces an incorrect result of only 36 bytes.
sizeof(VhostUserMemRegMsg) is actually 40 bytes.
The consequence of this is that the final field, mmap_offset, is
truncated. This breaks x86_64 TCG guests on s390 hosts. Other guest/host
combinations may get lucky if either of the following holds:
1. The guest memory layout does not need mmap_offset != 0.
2. The host is little-endian and mmap_offset <= 0xffffffff so the
truncation has no effect.
Fix this by extending the existing 32-bit padding field to 64-bit. Now
the padding reflects the actual compiler padding. This can be verified
using pahole(1).
Also document the layout properly in the vhost-user specification. The
vhost-user spec did not document the exact layout. It would be
impossible to implement the spec without looking at the QEMU source
code.
Existing vhost-user frontends and device backends continue to work after
this fix has been applied. The only change in the wire protocol is that
QEMU now sets hdr.size to 40 instead of 36. If a vhost-user
implementation has a hardcoded size check for 36 bytes, then it will
fail with new QEMUs. Both QEMU and DPDK/SPDK don't check the exact
payload size, so they continue to work.
Fixes: f1aeb14b08 ("Transmit vhost-user memory regions individually")
Cc: Raphael Norwitz <raphael.norwitz@nutanix.com>
Cc: Cornelia Huck <cohuck@redhat.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201109174355.1069147-1-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: f1aeb14b08 ("Transmit vhost-user memory regions individually")
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Virtqueue has split and packed, so before setting inflight,
you need to inform the back-end virtqueue format.
Signed-off-by: Jin Yu <jin.yu@intel.com>
Acked-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <20201103123617.28256-1-jin.yu@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This reverts commit adb29c0273.
The commit broke -device vhost-user-blk-pci because the
vhost_dev_prepare_inflight() function it introduced segfaults in
vhost_dev_set_features() when attempting to access struct vhost_dev's
vdev pointer before it has been assigned.
To reproduce the segfault simply launch a vhost-user-blk device with the
contrib vhost-user-blk device backend:
$ build/contrib/vhost-user-blk/vhost-user-blk -s /tmp/vhost-user-blk.sock -r -b /var/tmp/foo.img
$ build/qemu-system-x86_64 \
-device vhost-user-blk-pci,id=drv0,chardev=char1,addr=4.0 \
-object memory-backend-memfd,id=mem,size=1G,share=on \
-M memory-backend=mem,accel=kvm \
-chardev socket,id=char1,path=/tmp/vhost-user-blk.sock
Segmentation fault (core dumped)
Cc: Jin Yu <jin.yu@intel.com>
Cc: Raphael Norwitz <raphael.norwitz@nutanix.com>
Cc: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20201102165709.232180-1-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The virtio-iommu device can deal with arbitrary page sizes for virtual
endpoints, but for endpoints assigned with VFIO it must follow the page
granule used by the host IOMMU driver.
Implement the interface to set the vIOMMU page size mask, called by VFIO
for each endpoint. We assume that all host IOMMU drivers use the same
page granule (the host page granule). Override the page_size_mask field
in the virtio config space.
Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-10-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add notify_flag_changed() to notice when memory listeners are added and
removed.
Acked-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-7-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Implement the replay callback to setup all mappings for a new memory
region.
Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-6-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Call the memory notifiers when attaching an endpoint to a domain, to
replay existing mappings, and when detaching the endpoint, to remove all
mappings.
Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-5-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Extend VIRTIO_IOMMU_T_MAP/UNMAP request to notify memory listeners. It
will call VFIO notifier to map/unmap regions in the physical IOMMU.
Signed-off-by: Bharat Bhushan <bbhushan2@marvell.com>
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-4-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Store the memory region associated to each endpoint into the endpoint
structure, to allow efficient memory notification on map/unmap.
Acked-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-3-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Due to an invalid mask, virtio_iommu_mr() may return the wrong memory
region. It hasn't been too problematic so far because the function was
only used to test existence of an endpoint, but that is about to change.
Fixes: cfb42188b2 ("virtio-iommu: Implement attach/detach command")
Cc: QEMU Stable <qemu-stable@nongnu.org>
Acked-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Message-Id: <20201030180510.747225-2-jean-philippe@linaro.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fix uninitialized value issues reported by Coverity:
Field 'msg.reserved' is uninitialized when calling write().
While the 'struct vhost_msg' does not have a 'reserved' field,
we still initialize it to have the two parts of the function
consistent.
Reported-by: Coverity (CID 1432864: UNINIT)
Fixes: c471ad0e9b ("vhost_net: device IOTLB support")
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201103063541.2463363-1-philmd@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The block size determines the alignment requirements. Implement
get_min_alignment() of the TYPE_MEMORY_DEVICE interface.
This allows auto-assignment of a properly aligned address in guest
physical address space. For example, when specifying a 2GB block size
for a virtio-mem device with 10GB with a memory setup "-m 4G, 20G",
we'll no longer fail when realizing.
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20201008083029.9504-7-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Let's allow a minimum block size of 1 MiB in all configurations. Select
the default block size based on
- The page size of the memory backend.
- The THP size if the memory backend size corresponds to the real host
page size.
- The global minimum of 1 MiB.
and warn if something smaller is configured by the user.
VIRTIO_MEM only supports Linux (depends on LINUX), so we can probe the
THP size unconditionally.
For now we only support virtio-mem on x86-64 - there isn't a user-visible
change (x86-64 only supports 2 MiB THP on the PMD level) - the default
was, and will be 2 MiB.
If we ever have THP on the PUD level (e.g., 1 GiB THP on x86-64), we
expect it to be more transparent - e.g., to only optimize fully populated
ranges unless explicitly told /configured otherwise (in contrast to PMD
THP).
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20201008083029.9504-4-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The spec states:
"The device MUST set addr, region_size, usable_region_size, plugged_size,
requested_size to multiples of block_size."
With block sizes > 256MB, we currently wouldn't guarantee that for the
usable_region_size.
Note that we cannot exceed the region_size, as we already enforce the
alignment there properly.
Fixes: 910b25766b ("virtio-mem: Paravirtualized memory hot(un)plug")
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20201008083029.9504-3-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The spec states:
"The device MUST set addr, region_size, usable_region_size, plugged_size,
requested_size to multiples of block_size."
In some cases, we currently don't guarantee that for "addr": For example,
when starting a VM with 4 GiB boot memory and a virtio-mem device with a
block size of 2 GiB, "memaddr"/"addr" will be auto-assigned to
0x140000000 (5 GiB).
We'll try to improve auto-assignment for memory devices next, to avoid
bailing out in case memory device code selects a bad address.
Note: The Linux driver doesn't support such big block sizes yet.
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Fixes: 910b25766b ("virtio-mem: Paravirtualized memory hot(un)plug")
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Wei Yang <richardw.yang@linux.intel.com>
Cc: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Igor Mammedov <imammedo@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20201008083029.9504-2-david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
QEMU must be careful when loading device state off migration streams to
prevent a malicious source from exploiting the emulator. Overdoing these
checks has the side effect of allowing a guest to "pin itself" in cloud
environments by messing with state which is entirely in its control.
Similarly to what f3081539 achieved in usb_device_post_load(), this
commit removes such a check from virtio_load(). Worth noting, the result
of a load without this check is the same as if a guest enables a VQ with
invalid indexes to begin with. That is, the virtual device is set in a
broken state (by the datapath handler) and must be reset.
Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Message-Id: <20201028134643.110698-1-felipe@nutanix.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Virtqueue has split and packed, so before setting inflight,
you need to inform the back-end virtqueue format.
Signed-off-by: Jin Yu <jin.yu@intel.com>
Message-Id: <20200910134851.7817-1-jin.yu@intel.com>
Acked-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The first loop in vhost_get_log_size() computes the size of the dirty log
bitmap so that it allows to track changes in the entire guest memory, in
terms of GPA.
When not using a vIOMMU, the address of the vring's used structure,
vq->used_phys, is a GPA. It is thus already covered by the first loop.
When using a vIOMMU, vq->used_phys is a GIOVA that will be translated
to an HVA when the vhost backend needs to update the used structure. It
will log the corresponding GPAs into the bitmap but it certainly won't
log the GIOVA.
So in any case, vq->used_phys shouldn't be explicitly used to size the
bitmap. Drop the second loop.
This fixes a crash of the source when migrating a guest using in-kernel
vhost-net and iommu_platform=on on POWER, because DMA regions are put
over 0x800000000000000ULL. The resulting insanely huge log size causes
g_malloc0() to abort.
BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1879349
Signed-off-by: Greg Kurz <groug@kaod.org>
Message-Id: <160208823418.29027.15172801181796272300.stgit@bahia.lan>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fix uninitialized value issues reported by Coverity:
Field 'msg.reserved' is uninitialized when calling write().
Fixes: a5bd05800f ("vhost-vdpa: batch updating IOTLB mappings")
Reported-by: Coverity (CID 1432864: UNINIT)
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20201028154004.776760-1-philmd@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
vhost IOTLB API uses read()/write() to exchange iotlb messages with
the kernel module.
The QEMU implementation expects a non-blocking fd, indeed commit
c471ad0e9b ("vhost_net: device IOTLB support") set it for vhost-net.
Without this patch, if we enable iommu for the vhost-vsock device,
QEMU can hang when exchanging IOTLB messages.
As commit 894022e616 ("net: check if the file descriptor is valid
before using it") did for tap, let's use qemu_try_set_nonblock()
when fd is provided by the user.
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20201029144849.70958-1-sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
cur_mon really needs to be coroutine-local as soon as we move monitor
command handlers to coroutines and let them yield. As a first step, just
remove all direct accesses to cur_mon so that we can implement this in
the getter function later.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20201005155855.256490-4-kwolf@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Only qemu-system-FOO and qemu-storage-daemon provide QMP
monitors, therefore such declarations and definitions are
irrelevant for user-mode emulation.
Restricting the memory commands to machine.json pulls less
QAPI-generated code into user-mode.
Acked-by: Igor Mammedov <imammedo@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200913195348.1064154-7-philmd@redhat.com>
[Commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Only qemu-system-FOO and qemu-storage-daemon provide QMP
monitors, therefore such declarations and definitions are
irrelevant for user-mode emulation.
Restricting the balloon-related commands to machine.json pulls less
QAPI-generated code into user-mode.
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Markus Armbruster <armbru@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200913195348.1064154-4-philmd@redhat.com>
[Commit message tweaked]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
If error occurs while processing the virtio request we should call
'virtqueue_detach_element' to detach the element from the virtqueue
before free the elem.
Signed-off-by: Li Qiang <liq3ea@163.com>
Message-Id: <20200813165125.59928-1-liq3ea@163.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Fixes: 5f503cd9f3 ("virtio-pmem: add virtio device")
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Current the 'virtio_set_features' only update the 'MemorRegionCaches'
when the 'virtio_set_features_nocheck' return '0' which means it is
not bad features. However the guest can still trigger the access of the
used vring after set bad features. In this situation it will cause assert
failure in 'ADDRESS_SPACE_ST_CACHED'.
Buglink: https://bugs.launchpad.net/qemu/+bug/1890333
Fixes: db812c4073 ("virtio: update MemoryRegionCaches when guest negotiates features")
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Li Qiang <liq3ea@163.com>
Message-Id: <20200919082706.6703-1-liq3ea@163.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Commit 9b3a35ec82 ("virtio: verify that legacy support is not
accidentally on") added a safety check that requires to set
'disable-legacy=on' on vhost-user-vsock-pci device:
$ ./qemu-system-x86_64 ... \
-chardev socket,id=char0,reconnect=0,path=/tmp/vhost4.socket \
-device vhost-user-vsock-pci,chardev=char0
qemu-system-x86_64: -device vhost-user-vsock-pci,chardev=char0:
device is modern-only, use disable-legacy=on
virtio-vsock was introduced after the release of VIRTIO 1.0
specifications, so it should be 'modern-only'.
This patch forces virtio version 1 and removes the 'transitional_name'
property, as done for vhost-vsock-pci, removing the need to specify
'disable-legacy=on' on vhost-user-vsock-pci device.
Cc: qemu-stable@nongnu.org
Suggested-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20200921122506.82515-4-sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Commit 9b3a35ec82 ("virtio: verify that legacy support is not
accidentally on") added a safety check that requires to set
'disable-legacy=on' on vhost-vsock-pci device:
$ ./qemu-system-x86_64 ... -device vhost-vsock-pci,guest-cid=5
qemu-system-x86_64: -device vhost-vsock-pci,guest-cid=5:
device is modern-only, use disable-legacy=on
virtio-vsock was introduced after the release of VIRTIO 1.0
specifications, so it should be 'modern-only'.
In addition Cornelia verified that forcing a legacy mode on
vhost-vsock-pci device using x86-64 host and s390x guest, so with
different endianness, produces strange behaviours.
This patch forces virtio version 1 and removes the 'transitional_name'
property removing the need to specify 'disable-legacy=on' on
vhost-vsock-pci device.
To avoid migration issues, we force virtio version 1 only when
legacy check is enabled in the new machine types (>= 5.1).
As the transitional device name is not commonly used, we do not
provide compatibility handling for it.
Cc: qemu-stable@nongnu.org
Reported-by: Qian Cai <caiqian@redhat.com>
Reported-by: Qinghua Cheng <qcheng@redhat.com>
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1868449
Suggested-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20200921122506.82515-3-sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Commit 9b3a35ec82 ("virtio: verify that legacy support is not accidentally
on") added a check that returns an error if legacy support is on, but the
device does not support legacy.
Unfortunately some devices were wrongly declared legacy capable even if
they were not (e.g vhost-vsock).
To avoid migration issues, we add a virtio-device property
(x-disable-legacy-check) to skip the legacy error, printing a warning
instead, for machine types < 5.1.
Cc: qemu-stable@nongnu.org
Fixes: 9b3a35ec82 ("virtio: verify that legacy support is not accidentally on")
Suggested-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Suggested-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20200921122506.82515-2-sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add trace functions in vhost-vdpa.c.
All traces from this file can be enabled with '-trace vhost_vdpa*'.
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20200925091055.186023-3-lvivier@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
