qemu.git
10 months agox86: cpuhp: prevent guest crash on CPU hotplug when broadcast SMI is in use
Igor Mammedov [Wed, 23 Sep 2020 09:46:41 +0000 (05:46 -0400)] 
x86: cpuhp: prevent guest crash on CPU hotplug when broadcast SMI is in use

There were reports of guest crash on CPU hotplug, when using q35 machine
type and OVMF with SMM, due to hotplugged CPU trying to process SMI at
default SMI handler location without it being relocated by firmware first.

Fix it by refusing hotplug if firmware hasn't negotiated CPU hotplug with
SMI support while SMI broadcast is in use.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20200923094650.1301166-3-imammedo@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agox86: lpc9: let firmware negotiate 'CPU hotplug with SMI' features
Igor Mammedov [Wed, 23 Sep 2020 09:46:40 +0000 (05:46 -0400)] 
x86: lpc9: let firmware negotiate 'CPU hotplug with SMI' features

It will allow firmware to notify QEMU that firmware requires SMI
being triggered on CPU hot[un]plug, so that it would be able to account
for hotplugged CPU and relocate it to new SMM base and/or safely remove
CPU on unplug.

Using negotiated features, follow up patches will insert SMI upcall
into AML code, to make sure that firmware processes hotplug before
guest OS would attempt to use new CPU.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Tested-by: Laszlo Ersek <lersek@redhat.com>
Message-Id: <20200923094650.1301166-2-imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovirtio: update MemoryRegionCaches when guest set bad features
Li Qiang [Sat, 19 Sep 2020 08:27:06 +0000 (01:27 -0700)] 
virtio: update MemoryRegionCaches when guest set bad features

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: db812c4073c7 ("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>
10 months agovhost-vsock-ccw: force virtio version 1
Stefano Garzarella [Mon, 21 Sep 2020 12:25:06 +0000 (14:25 +0200)] 
vhost-vsock-ccw: force virtio version 1

virtio-vsock was introduced after the release of VIRTIO 1.0
specifications, so it should be 'modern-only'.

This patch forces virtio version 1 as done for vhost-vsock-pci.

To avoid migration issues, we force virtio version 1 only when
legacy check is enabled in the new machine types (>= 5.1).

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: <20200921122506.82515-5-sgarzare@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovhost-user-vsock-pci: force virtio version 1
Stefano Garzarella [Mon, 21 Sep 2020 12:25:05 +0000 (14:25 +0200)] 
vhost-user-vsock-pci: force virtio version 1

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>
10 months agovhost-vsock-pci: force virtio version 1
Stefano Garzarella [Mon, 21 Sep 2020 12:25:04 +0000 (14:25 +0200)] 
vhost-vsock-pci: force virtio version 1

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>
10 months agovirtio: skip legacy support check on machine types less than 5.1
Stefano Garzarella [Mon, 21 Sep 2020 12:25:03 +0000 (14:25 +0200)] 
virtio: skip legacy support check on machine types less than 5.1

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>
10 months agoconfigure: Fix build dependencies with vhost-vdpa.
Laurent Vivier [Thu, 24 Sep 2020 21:00:23 +0000 (23:00 +0200)] 
configure: Fix build dependencies with vhost-vdpa.

Following the same logic as for vhost-net-user and vhost-kernel,
enable vhost-net if vhost-net-vdpa is enabled and vhost-net is not
explicitly disabled.
See 299e6f19b3e2 ("vhost-net: revamp configure logic")

Autoselect VHOST if VHOST_VDPA is set
See 21c6b0c87e85 ("configure: simplify vhost condition with Kconfig")
See 2becc36a3e53 ("meson: infrastructure for building emulators"

Problems can be triggered using;
... --enable-vhost-vdpa --disable-vhost-user --disable-vhost-kernel ...

Fixes: 108a64818e69 ("vhost-vdpa: introduce vhost-vdpa backend")
Cc: lulu@redhat.com
Cc: pbonzini@redhat.com
Cc: marcandre.lureau@redhat.com
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20200924210023.160679-1-lvivier@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovhost-vdpa: add trace-events
Laurent Vivier [Fri, 25 Sep 2020 09:10:55 +0000 (11:10 +0200)] 
vhost-vdpa: add trace-events

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>
10 months agoutil/hexdump: introduce qemu_hexdump_line()
Laurent Vivier [Fri, 25 Sep 2020 09:10:54 +0000 (11:10 +0200)] 
util/hexdump: introduce qemu_hexdump_line()

Dumping one line of hexadecimal/ASCII from a buffer is often needed.
Move this part from qemu_hexdump() and use it

Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20200925091055.186023-2-lvivier@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovirtio-pmem-pci: force virtio version 1
Pankaj Gupta [Fri, 25 Sep 2020 10:22:51 +0000 (12:22 +0200)] 
virtio-pmem-pci: force virtio version 1

 Qemu fails with below error when trying to run with virtio pmem:

 (qemu) qemu-system-x86_64: -device virtio-pmem-pci,memdev=mem1,id=nv1:
  device is modern-only, use disable-legacy=on

 This patch fixes this by forcing virtio 1 with virtio-pmem.

fixes: adf0748a49 ("virtio-pci: Proxy for virtio-pmem")
Signed-off-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Message-Id: <20200925102251.7216-1-pankaj.gupta.linux@gmail.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovirtio-iommu-pci: force virtio version 1
Eric Auger [Tue, 8 Sep 2020 19:33:09 +0000 (21:33 +0200)] 
virtio-iommu-pci: force virtio version 1

Commit 9b3a35ec82 ("virtio: verify that legacy support is not
accidentally on") added a safety check that requires to set
'disable-legacy=on' on virtio-iommu-pci:

qemu-system-aarch64: -device virtio-iommu-pci: device is modern-only,
use disable-legacy=on

virtio-iommu 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 removing the need to specify 'disable-legacy=on' on
virtio-iommu-pci device.

Cc: qemu-stable@nongnu.org
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20200908193309.20569-3-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovirtio-iommu: Check gtrees are non null before destroying them
Eric Auger [Tue, 8 Sep 2020 19:33:08 +0000 (21:33 +0200)] 
virtio-iommu: Check gtrees are non null before destroying them

If realize fails, domains and endpoints trees may be NULL. On
unrealize(), this produces assertions:

"GLib: g_tree_destroy: assertion 'tree != NULL' failed"

Check that the trees are non NULL before destroying them.

Cc: qemu-stable@nongnu.org
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20200908193309.20569-2-eric.auger@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agocphp: remove deprecated cpu-add command(s)
Igor Mammedov [Tue, 15 Sep 2020 12:04:03 +0000 (08:04 -0400)] 
cphp: remove deprecated cpu-add command(s)

These were deprecated since 4.0, remove both HMP and QMP variants.

Users should use device_add command instead. To get list of
possible CPUs and options, use 'info hotpluggable-cpus' HMP
or query-hotpluggable-cpus QMP command.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Acked-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20200915120403.1074579-1-imammedo@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agotests/qtest/vhost-user-test: prepare the tests for adding new dev class
Dima Stepanov [Fri, 11 Sep 2020 08:39:45 +0000 (11:39 +0300)] 
tests/qtest/vhost-user-test: prepare the tests for adding new dev class

For now only vhost-user-net device is supported by the test. Other
vhost-user devices are not tested. As a first step make source code
refactoring so new devices can reuse the same test routines. To make
this provide a new vhost_user_ops structure with the methods to
initialize device, its command line or make a proper vhost-user
responses.

Signed-off-by: Dima Stepanov <dimastep@yandex-team.ru>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <a48b60fb96fef230b75fff72a2d77040dcb5ef7c.1599813294.git.dimastep@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovhost: check queue state in the vhost_dev_set_log routine
Dima Stepanov [Fri, 11 Sep 2020 08:39:44 +0000 (11:39 +0300)] 
vhost: check queue state in the vhost_dev_set_log routine

If the vhost-user-blk daemon provides only one virtqueue, but device was
added with several queues, then QEMU will send more VHOST-USER command
than expected by daemon side. The vhost_virtqueue_start() routine
handles such case by checking the return value from the
virtio_queue_get_desc_addr() function call. Add the same check to the
vhost_dev_set_log() routine.

Signed-off-by: Dima Stepanov <dimastep@yandex-team.ru>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <6232946d5af09e9775076645909964a6539b8ab5.1599813294.git.dimastep@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovhost: recheck dev state in the vhost_migration_log routine
Dima Stepanov [Fri, 11 Sep 2020 08:39:43 +0000 (11:39 +0300)] 
vhost: recheck dev state in the vhost_migration_log routine

vhost-user devices can get a disconnect in the middle of the VHOST-USER
handshake on the migration start. If disconnect event happened right
before sending next VHOST-USER command, then the vhost_dev_set_log()
call in the vhost_migration_log() function will return error. This error
will lead to the assert() and close the QEMU migration source process.
For the vhost-user devices the disconnect event should not break the
migration process, because:
  - the device will be in the stopped state, so it will not be changed
    during migration
  - if reconnect will be made the migration log will be reinitialized as
    part of reconnect/init process:
    #0  vhost_log_global_start (listener=0x563989cf7be0)
    at hw/virtio/vhost.c:920
    #1  0x000056398603d8bc in listener_add_address_space (listener=0x563989cf7be0,
        as=0x563986ea4340 <address_space_memory>)
    at softmmu/memory.c:2664
    #2  0x000056398603dd30 in memory_listener_register (listener=0x563989cf7be0,
        as=0x563986ea4340 <address_space_memory>)
    at softmmu/memory.c:2740
    #3  0x0000563985fd6956 in vhost_dev_init (hdev=0x563989cf7bd8,
        opaque=0x563989cf7e30, backend_type=VHOST_BACKEND_TYPE_USER,
        busyloop_timeout=0)
    at hw/virtio/vhost.c:1385
    #4  0x0000563985f7d0b8 in vhost_user_blk_connect (dev=0x563989cf7990)
    at hw/block/vhost-user-blk.c:315
    #5  0x0000563985f7d3f6 in vhost_user_blk_event (opaque=0x563989cf7990,
        event=CHR_EVENT_OPENED)
    at hw/block/vhost-user-blk.c:379
Update the vhost-user-blk device with the internal started_vu field which
will be used for initialization (vhost_user_blk_start) and clean up
(vhost_user_blk_stop). This additional flag in the VhostUserBlk structure
will be used to track whether the device really needs to be stopped and
cleaned up on a vhost-user level.
The disconnect event will set the overall VHOST device (not vhost-user) to
the stopped state, so it can be used by the general vhost_migration_log
routine.
Such approach could be propogated to the other vhost-user devices, but
better idea is just to make the same connect/disconnect code for all the
vhost-user devices.

This migration issue was slightly discussed earlier:
  - https://lists.gnu.org/archive/html/qemu-devel/2020-05/msg01509.html
  - https://lists.gnu.org/archive/html/qemu-devel/2020-05/msg05241.html

Signed-off-by: Dima Stepanov <dimastep@yandex-team.ru>
Reviewed-by: Raphael Norwitz <raphael.norwitz@nutanix.com>
Message-Id: <9fbfba06791a87813fcee3e2315f0b904cc6789a.1599813294.git.dimastep@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agopc: fix auto_enable_numa_with_memhp/auto_enable_numa_with_memdev for the 5.0 machine
David Hildenbrand [Thu, 20 Aug 2020 09:48:28 +0000 (11:48 +0200)] 
pc: fix auto_enable_numa_with_memhp/auto_enable_numa_with_memdev for the 5.0 machine

Unfortunately, a typo sneeked in: we want to set
auto_enable_numa_with_memdev to false, not auto_enable_numa_with_memhp.

Cc: qemu-stable@nongnu.org # v5.1
Fixes: 195784a0cfad (numa: Auto-enable NUMA when any memory devices are possible)
Reported-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20200820094828.30348-1-david@redhat.com>
Reviewed-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>
10 months agovirtio-mem: detach the element from the virtqueue when error occurs
Li Qiang [Sun, 16 Aug 2020 14:22:45 +0000 (07:22 -0700)] 
virtio-mem: detach the element from the virtqueue when error occurs

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: <20200816142245.17556-1-liq3ea@163.com>
Fixes: 910b25766b ("virtio-mem: Paravirtualized memory hot(un)plug")
Acked-by: David Hildenbrand <david@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovhost-vdpa: batch updating IOTLB mappings
Jason Wang [Mon, 7 Sep 2020 10:49:03 +0000 (18:49 +0800)] 
vhost-vdpa: batch updating IOTLB mappings

To speed up the memory mapping updating between vhost-vDPA and vDPA
device driver, this patch passes the IOTLB batching flags via IOTLB
API. Two new flags was introduced, VHOST_IOTLB_BATCH_BEGIN is a hint
that a bathced IOTLB updating may be initiated from the
userspace. VHOST_IOTLB_BATCH_END is a hint that userspace has finished
the updating:

VHOST_IOTLB_BATCH_BEGIN
VHOST_IOTLB_UPDATE/VHOST_IOTLB_INVALIDATE
...
VHOST_IOTLB_BATCH_END

Vhost-vDPA can then know that all mappings has been set and can do
optimization like passing all the mappings to the vDPA device driver.

Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20200907104903.31551-4-jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agovhost: switch to use IOTLB v2 format
Jason Wang [Mon, 7 Sep 2020 10:49:02 +0000 (18:49 +0800)] 
vhost: switch to use IOTLB v2 format

This patch tries to switch to use new kernel IOTLB format V2. Previous
version may have inconsistent ABI between 32bit and 64bit machines
because of the hole after type field. Refer kernel commit
("429711aec282 vhost: switch to use new message format") for more
information.

To enable this feature, qemu need to use a new ioctl
VHOST_SET_BACKEND_FEATURE with VHOST_BACKEND_F_IOTLB_MSG_V2 bit. A new
vhost setting backend features ops was introduced. And when we try to
set features for vhost dev, we will examine the support of new IOTLB
format and enable it. This process is total transparent to guest,
which means we can have different IOTLB message type in src and dst
during migration.

The conversion of IOTLB message is straightforward, just check the
type and behave accordingly.

Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20200907104903.31551-3-jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agolinux headers: sync to 5.9-rc4
Jason Wang [Mon, 7 Sep 2020 10:49:01 +0000 (18:49 +0800)] 
linux headers: sync to 5.9-rc4

Update against Linux 5.9-rc4.

Cc: Cornelia Huck <cohuck@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20200907104903.31551-2-jasowang@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
10 months agoMerge remote-tracking branch 'remotes/alistair/tags/pull-register-20200927' into...
Peter Maydell [Mon, 28 Sep 2020 15:49:10 +0000 (16:49 +0100)] 
Merge remote-tracking branch 'remotes/alistair/tags/pull-register-20200927' into staging

Two small patches. One with a fix for the register API instance_size
and one for removing unused address variables from load_elf.

# gpg: Signature made Sun 27 Sep 2020 14:45:06 BST
# gpg:                using RSA key F6C4AC46D4934868D3B8CE8F21E10D29DF977054
# gpg: Good signature from "Alistair Francis <alistair@alistair23.me>" [full]
# Primary key fingerprint: F6C4 AC46 D493 4868 D3B8  CE8F 21E1 0D29 DF97 7054

* remotes/alistair/tags/pull-register-20200927:
  core/register: Specify instance_size in the TypeInfo
  load_elf: Remove unused address variables from callers

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agoMerge remote-tracking branch 'remotes/vivier2/tags/trivial-branch-for-5.2-pull-reques...
Peter Maydell [Mon, 28 Sep 2020 13:03:09 +0000 (14:03 +0100)] 
Merge remote-tracking branch 'remotes/vivier2/tags/trivial-branch-for-5.2-pull-request' into staging

Trivial Patches Pull request 20200928

# gpg: Signature made Mon 28 Sep 2020 10:15:00 BST
# gpg:                using RSA key CD2F75DDC8E3A4DC2E4F5173F30C38BD3F2FBE3C
# gpg:                issuer "laurent@vivier.eu"
# gpg: Good signature from "Laurent Vivier <lvivier@redhat.com>" [full]
# gpg:                 aka "Laurent Vivier <laurent@vivier.eu>" [full]
# gpg:                 aka "Laurent Vivier (Red Hat) <lvivier@redhat.com>" [full]
# Primary key fingerprint: CD2F 75DD C8E3 A4DC 2E4F  5173 F30C 38BD 3F2F BE3C

* remotes/vivier2/tags/trivial-branch-for-5.2-pull-request:
  docs/system/deprecated: Move lm32 and unicore32 to the right section
  migration/multifd: Remove superfluous semicolons
  timer: Fix timer_mod_anticipate() documentation
  vhost-vdpa: remove useless variable
  Add *.pyc back to the .gitignore file
  virtio: vdpa: omit check return of g_malloc
  meson: fix static flag summary
  vhost-vdpa: fix indentation in vdpa_ops

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agocore/register: Specify instance_size in the TypeInfo
Alistair Francis [Thu, 20 Aug 2020 22:46:11 +0000 (15:46 -0700)] 
core/register: Specify instance_size in the TypeInfo

Reported-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <4cf1beb7dafb9143c261d266557d3173bf160524.1598376594.git.alistair.francis@wdc.com>

10 months agoload_elf: Remove unused address variables from callers
BALATON Zoltan [Sun, 5 Jul 2020 17:22:11 +0000 (19:22 +0200)] 
load_elf: Remove unused address variables from callers

Several callers of load_elf() pass pointers for lowaddr and highaddr
parameters which are then not used for anything. This may stem from a
misunderstanding that load_elf need a value here but in fact it can
take NULL to ignore these values. Remove such unused variables and
pass NULL instead from callers that don't need these.

Signed-off-by: BALATON Zoltan <balaton@eik.bme.hu>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Acked-by: David Gibson <david@gibson.dropbear.id.au>
Acked-by: Max Filippov <jcmvbkbc@gmail.com>
Message-Id: <20200705174020.BDD0174633F@zero.eik.bme.hu>
Signed-off-by: Alistair Francis <alistair.francis@wdc.com>
10 months agoMerge remote-tracking branch 'remotes/dgilbert/tags/pull-migration-20200925a' into...
Peter Maydell [Fri, 25 Sep 2020 13:46:18 +0000 (14:46 +0100)] 
Merge remote-tracking branch 'remotes/dgilbert/tags/pull-migration-20200925a' into staging

Migration and virtiofsd pull

Chuan Zheng's Dirtyrate and TLS changes, with small fixes from Dov and
Luarent.
Small virtiofs changes from Harry, Stefan, Vivek and Jiachen.
One HMP/monitor rework from me.

# gpg: Signature made Fri 25 Sep 2020 13:03:50 BST
# gpg:                using RSA key 45F5C71B4A0CB7FB977A9FA90516331EBC5BFDE7
# gpg: Good signature from "Dr. David Alan Gilbert (RH2) <dgilbert@redhat.com>" [full]
# Primary key fingerprint: 45F5 C71B 4A0C B7FB 977A  9FA9 0516 331E BC5B FDE7

* remotes/dgilbert/tags/pull-migration-20200925a: (26 commits)
  virtiofsd: Add -o allow_direct_io|no_allow_direct_io options
  virtiofsd: Used glib "shared" thread pool
  virtiofsd: document cache=auto default
  monitor: Use LOCK_GUARD macros
  migration/tls: add trace points for multifd-tls
  migration/tls: add support for multifd tls-handshake
  migration/tls: extract cleanup function for common-use
  migration/tls: add tls_hostname into MultiFDSendParams
  migration/tls: extract migration_tls_client_create for common-use
  migration/tls: save hostname into MigrationState
  migration: increase max-bandwidth to 128 MiB/s (1 Gib/s)
  migration: Truncate state file in xen-save-devices-state
  migration/dirtyrate: Add trace_calls to make it easier to debug
  migration/dirtyrate: Implement qmp_cal_dirty_rate()/qmp_get_dirty_rate() function
  migration/dirtyrate: Implement calculate_dirtyrate() function
  migration/dirtyrate: Implement set_sample_page_period() and is_sample_period_valid()
  migration/dirtyrate: skip sampling ramblock with size below MIN_RAMBLOCK_SIZE
  migration/dirtyrate: Compare page hash results for recorded sampled page
  migration/dirtyrate: Record hash results for each sampled page
  migration/dirtyrate: move RAMBLOCK_FOREACH_MIGRATABLE into ram.h
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agovirtiofsd: Add -o allow_direct_io|no_allow_direct_io options
Jiachen Zhang [Mon, 24 Aug 2020 10:59:57 +0000 (18:59 +0800)] 
virtiofsd: Add -o allow_direct_io|no_allow_direct_io options

Due to the commit 65da4539803373ec4eec97ffc49ee90083e56efd, the O_DIRECT
open flag of guest applications will be discarded by virtiofsd. While
this behavior makes it consistent with the virtio-9p scheme when guest
applications use direct I/O, we no longer have any chance to bypass the
host page cache.

Therefore, we add a flag 'allow_direct_io' to lo_data. If '-o
 no_allow_direct_io' option is added, or none of '-o allow_direct_io' or
 '-o no_allow_direct_io' is added, the 'allow_direct_io' will be set to
 0, and virtiofsd discards O_DIRECT as before. If '-o allow_direct_io'
is added to the starting command-line, 'allow_direct_io' will be set to
1, so that the O_DIRECT flags will be retained and host page cache can
be bypassed.

Signed-off-by: Jiachen Zhang <zhangjiachen.jaycee@bytedance.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200824105957.61265-1-zhangjiachen.jaycee@bytedance.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agovirtiofsd: Used glib "shared" thread pool
Vivek Goyal [Mon, 21 Sep 2020 21:32:16 +0000 (17:32 -0400)] 
virtiofsd: Used glib "shared" thread pool

glib offers thread pools and it seems to support "exclusive" and "shared"
thread pools.

https://developer.gnome.org/glib/stable/glib-Thread-Pools.html#g-thread-pool-new

Currently we use "exlusive" thread pools but its performance seems to be
poor. I tried using "shared" thread pools and performance seems much
better. I posted performance results here.

https://www.redhat.com/archives/virtio-fs/2020-September/msg00080.html

So lets switch to shared thread pools. We can think of making it optional
once somebody can show in what cases exclusive thread pools offer better
results. For now, my simple performance tests across the board see
better results with shared thread pools.

Signed-off-by: Vivek Goyal <vgoyal@redhat.com>
Message-Id: <20200921213216.GE13362@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
  With seccomp fix from Miklos

10 months agovirtiofsd: document cache=auto default
Harry G. Coin [Wed, 16 Sep 2020 11:22:50 +0000 (12:22 +0100)] 
virtiofsd: document cache=auto default

The virtiofsd --help output documents the cache=auto default value but
the man page does not. Fix this.

Signed-off-by: Harry G. Coin <hgcoin@gmail.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200916112250.760245-1-stefanha@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomonitor: Use LOCK_GUARD macros
Dr. David Alan Gilbert [Tue, 22 Sep 2020 09:57:41 +0000 (10:57 +0100)] 
monitor: Use LOCK_GUARD macros

Use the lock guard macros in monitor/misc.c - saves
a lot of unlocks in error paths, and the occasional goto.

Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20200922095741.101911-1-dgilbert@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/tls: add trace points for multifd-tls
Chuan Zheng [Tue, 15 Sep 2020 03:04:02 +0000 (11:04 +0800)] 
migration/tls: add trace points for multifd-tls

add trace points for multifd-tls for debug.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: Yan Jin <jinyan12@huawei.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <1600139042-104593-7-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/tls: add support for multifd tls-handshake
Chuan Zheng [Tue, 15 Sep 2020 03:04:01 +0000 (11:04 +0800)] 
migration/tls: add support for multifd tls-handshake

Similar like migration main thread, we need to do handshake
for each multifd thread.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: Yan Jin <jinyan12@huawei.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <1600139042-104593-6-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/tls: extract cleanup function for common-use
Chuan Zheng [Tue, 15 Sep 2020 03:04:00 +0000 (11:04 +0800)] 
migration/tls: extract cleanup function for common-use

multifd channel cleanup is need if multifd handshake failed,
let's extract it.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: Yan Jin <jinyan12@huawei.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <1600139042-104593-5-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/tls: add tls_hostname into MultiFDSendParams
Chuan Zheng [Tue, 15 Sep 2020 03:03:59 +0000 (11:03 +0800)] 
migration/tls: add tls_hostname into MultiFDSendParams

Since multifd creation is async with migration_channel_connect, we should
pass the hostname from MigrationState to MultiFDSendParams.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: Yan Jin <jinyan12@huawei.com>
Message-Id: <1600139042-104593-4-git-send-email-zhengchuan@huawei.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/tls: extract migration_tls_client_create for common-use
Chuan Zheng [Tue, 15 Sep 2020 03:03:58 +0000 (11:03 +0800)] 
migration/tls: extract migration_tls_client_create for common-use

migration_tls_client_create will be used in multifd-tls, let's
extract it.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: Yan Jin <jinyan12@huawei.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <1600139042-104593-3-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/tls: save hostname into MigrationState
Chuan Zheng [Tue, 15 Sep 2020 03:03:57 +0000 (11:03 +0800)] 
migration/tls: save hostname into MigrationState

hostname is need in multifd-tls, save hostname into MigrationState.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: Yan Jin <jinyan12@huawei.com>
Message-Id: <1600139042-104593-2-git-send-email-zhengchuan@huawei.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration: increase max-bandwidth to 128 MiB/s (1 Gib/s)
Laurent Vivier [Mon, 21 Sep 2020 14:49:57 +0000 (16:49 +0200)] 
migration: increase max-bandwidth to 128 MiB/s (1 Gib/s)

max-bandwidth is set by default to 32 MiB/s (256 Mib/s)
since 2008 (5bb7910af031c).

Most of the CPUs can dirty memory faster than that now,
and this is clearly a problem with POWER where the page
size is 64 kiB and not 4 KiB.

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Message-Id: <20200921144957.979989-1-lvivier@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration: Truncate state file in xen-save-devices-state
Dov Murik [Mon, 21 Sep 2020 09:48:30 +0000 (09:48 +0000)] 
migration: Truncate state file in xen-save-devices-state

When running the xen-save-devices-state QMP command, if the filename
already exists it will be truncated before dumping the devices' state
into it.

Signed-off-by: Dov Murik <dovmurik@linux.vnet.ibm.com>
Message-Id: <20200921094830.114028-1-dovmurik@linux.vnet.ibm.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: Add trace_calls to make it easier to debug
Chuan Zheng [Wed, 16 Sep 2020 06:22:07 +0000 (14:22 +0800)] 
migration/dirtyrate: Add trace_calls to make it easier to debug

Add trace_calls to make it easier to debug

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Message-Id: <1600237327-33618-13-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: Implement qmp_cal_dirty_rate()/qmp_get_dirty_rate() function
Chuan Zheng [Wed, 16 Sep 2020 06:22:06 +0000 (14:22 +0800)] 
migration/dirtyrate: Implement qmp_cal_dirty_rate()/qmp_get_dirty_rate() function

Implement qmp_cal_dirty_rate()/qmp_get_dirty_rate() function which could be called

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Message-Id: <1600237327-33618-12-git-send-email-zhengchuan@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
  atomic function fixup
  Wording fixup in migration.json based on Eric's review

10 months agomigration/dirtyrate: Implement calculate_dirtyrate() function
Chuan Zheng [Wed, 16 Sep 2020 06:22:05 +0000 (14:22 +0800)] 
migration/dirtyrate: Implement calculate_dirtyrate() function

Implement calculate_dirtyrate() function.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: YanYing Zhuang <ann.zhuangyanying@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-11-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: Implement set_sample_page_period() and is_sample_period_valid()
Chuan Zheng [Wed, 16 Sep 2020 06:22:04 +0000 (14:22 +0800)] 
migration/dirtyrate: Implement set_sample_page_period() and is_sample_period_valid()

Implement is_sample_period_valid() to check if the sample period is vaild and
do set_sample_page_period() to sleep specific time between sample actions.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-10-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: skip sampling ramblock with size below MIN_RAMBLOCK_SIZE
Chuan Zheng [Wed, 16 Sep 2020 06:22:03 +0000 (14:22 +0800)] 
migration/dirtyrate: skip sampling ramblock with size below MIN_RAMBLOCK_SIZE

In order to sample real RAM, skip ramblock with size below MIN_RAMBLOCK_SIZE
which is set as 128M.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-9-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: Compare page hash results for recorded sampled page
Chuan Zheng [Wed, 16 Sep 2020 06:22:02 +0000 (14:22 +0800)] 
migration/dirtyrate: Compare page hash results for recorded sampled page

Compare page hash results for recorded sampled page.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: YanYing Zhuang <ann.zhuangyanying@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-8-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: Record hash results for each sampled page
Chuan Zheng [Wed, 16 Sep 2020 06:22:01 +0000 (14:22 +0800)] 
migration/dirtyrate: Record hash results for each sampled page

Record hash results for each sampled page, crc32 is taken to calculate
hash results for each sampled length in TARGET_PAGE_SIZE.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: YanYing Zhuang <ann.zhuangyanying@huawei.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-7-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: move RAMBLOCK_FOREACH_MIGRATABLE into ram.h
Chuan Zheng [Wed, 16 Sep 2020 06:22:00 +0000 (14:22 +0800)] 
migration/dirtyrate: move RAMBLOCK_FOREACH_MIGRATABLE into ram.h

RAMBLOCK_FOREACH_MIGRATABLE is need in dirtyrate measure,
move the existing definition up into migration/ram.h

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-6-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: Add dirtyrate statistics series functions
Chuan Zheng [Wed, 16 Sep 2020 06:21:59 +0000 (14:21 +0800)] 
migration/dirtyrate: Add dirtyrate statistics series functions

Add dirtyrate statistics functions to record/update dirtyrate info.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-5-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: Add RamblockDirtyInfo to store sampled page info
Chuan Zheng [Wed, 16 Sep 2020 06:21:58 +0000 (14:21 +0800)] 
migration/dirtyrate: Add RamblockDirtyInfo to store sampled page info

Add RamblockDirtyInfo to store sampled page info of each ramblock.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-4-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration/dirtyrate: add DirtyRateStatus to denote calculation status
Chuan Zheng [Wed, 16 Sep 2020 06:21:57 +0000 (14:21 +0800)] 
migration/dirtyrate: add DirtyRateStatus to denote calculation status

add DirtyRateStatus to denote calculating status.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-3-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
  atomic name fixup

10 months agomigration/dirtyrate: setup up query-dirtyrate framwork
Chuan Zheng [Wed, 16 Sep 2020 06:21:56 +0000 (14:21 +0800)] 
migration/dirtyrate: setup up query-dirtyrate framwork

Add get_dirtyrate_thread() functions to setup query-dirtyrate
framework.

Signed-off-by: Chuan Zheng <zhengchuan@huawei.com>
Signed-off-by: YanYing Zhuang <ann.zhuangyanying@huawei.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: David Edmondson <david.edmondson@oracle.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <1600237327-33618-2-git-send-email-zhengchuan@huawei.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration: Rework migrate_send_rp_req_pages() function
Peter Xu [Tue, 8 Sep 2020 20:30:18 +0000 (16:30 -0400)] 
migration: Rework migrate_send_rp_req_pages() function

We duplicated the logic of maintaining the last_rb variable at both callers of
this function.  Pass *rb pointer into the function so that we can avoid
duplicating the logic.  Also, when we have the rb pointer, it's also easier to
remove the original 2nd & 4th parameters, because both of them (name of the
ramblock when needed, or the page size) can be fetched from the ramblock
pointer too.

Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20200908203022.341615-3-peterx@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agomigration: Properly destroy variables on incoming side
Peter Xu [Tue, 8 Sep 2020 20:30:17 +0000 (16:30 -0400)] 
migration: Properly destroy variables on incoming side

In migration_incoming_state_destroy(), we've got a few variables that aren't
destroyed properly, namely:

    main_thread_load_event
    postcopy_pause_sem_dst
    postcopy_pause_sem_fault
    rp_mutex

Destroy them properly.

Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20200908203022.341615-2-peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
10 months agoMerge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging
Peter Maydell [Thu, 24 Sep 2020 17:48:45 +0000 (18:48 +0100)] 
Merge remote-tracking branch 'remotes/stefanha/tags/block-pull-request' into staging

Pull request

This includes the atomic_ -> qatomic_ rename that touches many files and is
prone to conflicts.

# gpg: Signature made Wed 23 Sep 2020 17:08:43 BST
# gpg:                using RSA key 8695A8BFD3F97CDAAC35775A9CA4ABB381AB73C8
# gpg: Good signature from "Stefan Hajnoczi <stefanha@redhat.com>" [full]
# gpg:                 aka "Stefan Hajnoczi <stefanha@gmail.com>" [full]
# Primary key fingerprint: 8695 A8BF D3F9 7CDA AC35  775A 9CA4 ABB3 81AB 73C8

* remotes/stefanha/tags/block-pull-request:
  qemu/atomic.h: rename atomic_ to qatomic_
  tests: add test-fdmon-epoll
  fdmon-poll: reset npfd when upgrading to fdmon-epoll
  gitmodules: add qemu.org vbootrom submodule
  gitmodules: switch to qemu.org meson mirror
  gitmodules: switch to qemu.org qboot mirror
  docs/system: clarify deprecation schedule
  virtio-crypto: don't modify elem->in/out_sg
  virtio-blk: undo destructive iov_discard_*() operations
  util/iov: add iov_discard_undo()
  virtio: add vhost-user-fs-ccw device
  libvhost-user: handle endianness as mandated by the spec
  MAINTAINERS: add Stefan Hajnoczi as block/nvme.c maintainer

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agoMerge remote-tracking branch 'remotes/kraxel/tags/audio-20200923-pull-request' into...
Peter Maydell [Thu, 24 Sep 2020 16:22:17 +0000 (17:22 +0100)] 
Merge remote-tracking branch 'remotes/kraxel/tags/audio-20200923-pull-request' into staging

audio: various buffering fixes.
audio: build spiceaudio as module.

# gpg: Signature made Wed 23 Sep 2020 10:09:46 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/audio-20200923-pull-request:
  audio: build spiceaudio as module
  audio: remove qemu_spice_audio_init()
  audio: run downstream playback queue unconditionally
  audio: align audio_generic_write with audio_pcm_hw_run_out
  audio: remove unnecessary calls to put_buffer_in
  audio: align audio_generic_read with audio_pcm_hw_run_in
  audio/spiceaudio: always rate limit playback stream
  audio/audio: fix video playback slowdown with spiceaudio
  audio: handle buf == NULL in put_buffer_out()

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agoMerge remote-tracking branch 'remotes/ehabkost/tags/machine-next-pull-request' into...
Peter Maydell [Thu, 24 Sep 2020 14:28:26 +0000 (15:28 +0100)] 
Merge remote-tracking branch 'remotes/ehabkost/tags/machine-next-pull-request' into staging

machine + QOM queue, 2020-09-22

QOM cleanups:
* Convert instance properties to class properties
  (Eduardo Habkost)
* simplify object_find_property / object_class_find_property
  (Daniel P. Berrangé)

Deprecated feature removal:
* Drop support for invalid topologies (Igor Mammedov)

# gpg: Signature made Tue 22 Sep 2020 23:25:01 BST
# gpg:                using RSA key 5A322FD5ABC4D3DBACCFD1AA2807936F984DC5A6
# gpg:                issuer "ehabkost@redhat.com"
# gpg: Good signature from "Eduardo Habkost <ehabkost@redhat.com>" [full]
# Primary key fingerprint: 5A32 2FD5 ABC4 D3DB ACCF  D1AA 2807 936F 984D C5A6

* remotes/ehabkost/tags/machine-next-pull-request:
  sifive_u: Register "start-in-flash" as class property
  sifive_e: Register "revb" as class property
  i440fx: Register i440FX-pcihost properties as class properties
  machine: Register "memory-backend" as class property
  xlnx-zcu102: Register properties as class properties
  cpu/core: Register core-id and nr-threads as class properties
  s390x: Register all CPU properties as class properties
  cryptodev-backend: Register "chardev" as class property
  cryptodev-vhost-user: Register "chardev" as class property
  smp: drop support for deprecated (invalid topologies)
  qom: simplify object_find_property / object_class_find_property

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agodocs/system/deprecated: Move lm32 and unicore32 to the right section
Thomas Huth [Wed, 23 Sep 2020 08:00:15 +0000 (10:00 +0200)] 
docs/system/deprecated: Move lm32 and unicore32 to the right section

lm32 and unicore32 are softmmut targets, and not linux-user targets.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200923080015.77373-1-thuth@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agomigration/multifd: Remove superfluous semicolons
Philippe Mathieu-Daudé [Mon, 21 Sep 2020 04:02:31 +0000 (06:02 +0200)] 
migration/multifd: Remove superfluous semicolons

checkpatch.pl report superfluous semicolons since commit
ee0f3c09e01, but this one was missed:

  scripts/checkpatch.pl d32ca5ad798~..d32ca5ad798
  ERROR: superfluous trailing semicolon
  #498: FILE: migration/multifd.c:308:
  +    ram_counters.transferred += transferred;;
  total: 1 errors, 1 warnings, 2073 lines checked

Fixes: d32ca5ad798 ("multifd: Split multifd code into its own file")
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <20200921040231.437653-1-f4bug@amsat.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agotimer: Fix timer_mod_anticipate() documentation
Philippe Mathieu-Daudé [Sun, 20 Sep 2020 15:50:42 +0000 (17:50 +0200)] 
timer: Fix timer_mod_anticipate() documentation

timer_mod_anticipate() will be scaled to the timer unit,
which is not always nanosecond. Fix the documentation.

Fixes: add40e9777d ("timer: add timer_mod_anticipate*")
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20200920155042.400737-1-f4bug@amsat.org>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agovhost-vdpa: remove useless variable
Laurent Vivier [Sun, 20 Sep 2020 15:20:24 +0000 (17:20 +0200)] 
vhost-vdpa: remove useless variable

in vhost_vdpa_listener_region_del(), try_unmap is always true and so,
vhost_vdpa_dma_unmap() is always called. We can remove the variable

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <20200920152024.860172-1-lvivier@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agoAdd *.pyc back to the .gitignore file
Thomas Huth [Sat, 19 Sep 2020 10:18:59 +0000 (12:18 +0200)] 
Add *.pyc back to the .gitignore file

Python still dumps its bytecode into the source directory, so
we should continue to ignore the *.pyc files.

Fixes: 0e72b7df4d ("Simplify the .gitignore file")
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20200919101859.28739-1-thuth@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agovirtio: vdpa: omit check return of g_malloc
Li Qiang [Wed, 19 Aug 2020 14:43:09 +0000 (07:43 -0700)] 
virtio: vdpa: omit check return of g_malloc

If g_malloc fails, the application will be terminated.
No need to check the return value of g_malloc.

Signed-off-by: Li Qiang <liq3ea@163.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Message-Id: <20200819144309.67579-1-liq3ea@163.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agomeson: fix static flag summary
Laurent Vivier [Thu, 17 Sep 2020 14:07:00 +0000 (16:07 +0200)] 
meson: fix static flag summary

'static build:' must display value of CONFIG_STATIC rather than value of
CONFIG_TOOLS.

Signed-off-by: Laurent Vivier <lvivier@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200917140700.673171-1-lvivier@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agovhost-vdpa: fix indentation in vdpa_ops
Stefano Garzarella [Wed, 16 Sep 2020 15:26:34 +0000 (17:26 +0200)] 
vhost-vdpa: fix indentation in vdpa_ops

This patch fixes wrong indentation of some vdpa_ops fields introduced
with the initial commit 108a64818e ("vhost-vdpa: introduce vhost-vdpa
backend")

Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Reviewed-by: Laurent Vivier <laurent@vivier.eu>
Message-Id: <20200916152634.56917-1-sgarzare@redhat.com>
Signed-off-by: Laurent Vivier <laurent@vivier.eu>
10 months agoqemu/atomic.h: rename atomic_ to qatomic_
Stefan Hajnoczi [Wed, 23 Sep 2020 10:56:46 +0000 (11:56 +0100)] 
qemu/atomic.h: rename atomic_ to qatomic_

clang's C11 atomic_fetch_*() functions only take a C11 atomic type
pointer argument. QEMU uses direct types (int, etc) and this causes a
compiler error when a QEMU code calls these functions in a source file
that also included <stdatomic.h> via a system header file:

  $ CC=clang CXX=clang++ ./configure ... && make
  ../util/async.c:79:17: error: address argument to atomic operation must be a pointer to _Atomic type ('unsigned int *' invalid)

Avoid using atomic_*() names in QEMU's atomic.h since that namespace is
used by <stdatomic.h>. Prefix QEMU's APIs with 'q' so that atomic.h
and <stdatomic.h> can co-exist. I checked /usr/include on my machine and
searched GitHub for existing "qatomic_" users but there seem to be none.

This patch was generated using:

  $ git grep -h -o '\<atomic\(64\)\?_[a-z0-9_]\+' include/qemu/atomic.h | \
    sort -u >/tmp/changed_identifiers
  $ for identifier in $(</tmp/changed_identifiers); do
        sed -i "s%\<$identifier\>%q$identifier%g" \
            $(git grep -I -l "\<$identifier\>")
    done

I manually fixed line-wrap issues and misaligned rST tables.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200923105646.47864-1-stefanha@redhat.com>

10 months agoMerge remote-tracking branch 'remotes/ericb/tags/pull-bitmaps-2020-09-21' into staging
Peter Maydell [Wed, 23 Sep 2020 14:11:38 +0000 (15:11 +0100)] 
Merge remote-tracking branch 'remotes/ericb/tags/pull-bitmaps-2020-09-21' into staging

bitmaps patches for 2020-09-21

- Eric Blake: Improve 'qemu-img bitmap --merge' by not opening backing images

# gpg: Signature made Tue 22 Sep 2020 01:58:31 BST
# gpg:                using RSA key 71C2CC22B1C4602927D2F3AAA7A16B4A2527436A
# gpg: Good signature from "Eric Blake <eblake@redhat.com>" [full]
# gpg:                 aka "Eric Blake (Free Software Programmer) <ebb9@byu.net>" [full]
# gpg:                 aka "[jpeg image of size 6874]" [full]
# Primary key fingerprint: 71C2 CC22 B1C4 6029 27D2  F3AA A7A1 6B4A 2527 436A

* remotes/ericb/tags/pull-bitmaps-2020-09-21:
  qemu-img: Support bitmap --merge into backing image

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agotests: add test-fdmon-epoll
Stefan Hajnoczi [Tue, 15 Sep 2020 12:03:39 +0000 (13:03 +0100)] 
tests: add test-fdmon-epoll

Test aio_disable_external(), which switches from fdmon-epoll back to
fdmon-poll. This resulted in an assertion failure that was fixed in the
previous patch.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200915120339.702938-3-stefanha@redhat.com>

10 months agofdmon-poll: reset npfd when upgrading to fdmon-epoll
Stefan Hajnoczi [Tue, 15 Sep 2020 12:03:38 +0000 (13:03 +0100)] 
fdmon-poll: reset npfd when upgrading to fdmon-epoll

npfd keeps track of how many pollfds are currently being monitored. It
must be reset to 0 when fdmon_poll_wait() returns.

When npfd reaches a treshold we switch to fdmon-epoll because it scales
better.

This patch resets npfd in the case where we switch to fdmon-epoll.
Forgetting to do so results in the following assertion failure:

  util/fdmon-poll.c:65: fdmon_poll_wait: Assertion `npfd == 0' failed.

Fixes: 1f050a4690f62a1e7dabc4f44141e9f762c3769f ("aio-posix: extract ppoll(2) and epoll(7) fd monitoring")
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1869952
Message-Id: <20200915120339.702938-2-stefanha@redhat.com>

10 months agogitmodules: add qemu.org vbootrom submodule
Stefan Hajnoczi [Wed, 9 Sep 2020 10:09:37 +0000 (11:09 +0100)] 
gitmodules: add qemu.org vbootrom submodule

The vbootrom module is needed for the new NPCM7xx ARM SoCs. The
vbootrom.git repo is now mirrored on qemu.org. QEMU mirrors third-party
code to ensure that users can always build QEMU even if the dependency
goes offline and so QEMU meets its responsibilities to provide full
source code under software licenses.

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Cc: Havard Skinnemoen <hskinnemoen@google.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20200915130834.706758-4-stefanha@redhat.com>

10 months agogitmodules: switch to qemu.org meson mirror
Stefan Hajnoczi [Tue, 15 Sep 2020 13:08:33 +0000 (14:08 +0100)] 
gitmodules: switch to qemu.org meson mirror

QEMU now hosts a mirror of meson.git. QEMU mirrors third-party code to
ensure that users can always build QEMU even if the dependency goes
offline and so QEMU meets its responsibilities to provide full source
code under software licenses.

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Cc: Marc-André Lureau <marcandre.lureau@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200915130834.706758-3-stefanha@redhat.com>

10 months agogitmodules: switch to qemu.org qboot mirror
Stefan Hajnoczi [Tue, 15 Sep 2020 13:08:32 +0000 (14:08 +0100)] 
gitmodules: switch to qemu.org qboot mirror

QEMU now hosts a mirror of qboot.git. QEMU mirrors third-party code to
ensure that users can always build QEMU even if the dependency goes
offline and so QEMU meets its responsibilities to provide full source
code under software licenses.

Suggested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200915130834.706758-2-stefanha@redhat.com>

10 months agodocs/system: clarify deprecation schedule
Stefan Hajnoczi [Tue, 15 Sep 2020 15:07:34 +0000 (16:07 +0100)] 
docs/system: clarify deprecation schedule

The sentence explaining the deprecation schedule is ambiguous. Make it
clear that a feature deprecated in the Nth release is guaranteed to
remain available in the N+1th release. Removal can occur in the N+2nd
release or later.

As an example of this in action, see commit
25956af3fe5dd0385ad8017bc768a6afe41e2a74 ("block: Finish deprecation of
'qemu-img convert -n -o'"). The feature was deprecated in QEMU 4.2.0. It
was present in the 5.0.0 release and removed in the 5.1.0 release.

Suggested-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200915150734.711426-1-stefanha@redhat.com>

10 months agovirtio-crypto: don't modify elem->in/out_sg
Stefan Hajnoczi [Thu, 17 Sep 2020 09:44:55 +0000 (10:44 +0100)] 
virtio-crypto: don't modify elem->in/out_sg

A number of iov_discard_front/back() operations are made by
virtio-crypto. The elem->in/out_sg iovec arrays are modified by these
operations, resulting virtqueue_unmap_sg() calls on different addresses
than were originally mapped.

This is problematic because dirty memory may not be logged correctly,
MemoryRegion refcounts may be leaked, and the non-RAM bounce buffer can
be leaked.

Take a copy of the elem->in/out_sg arrays so that the originals are
preserved. The iov_discard_undo() API could be used instead (with better
performance) but requires careful auditing of the code, so do the simple
thing instead.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <20200917094455.822379-4-stefanha@redhat.com>

10 months agovirtio-blk: undo destructive iov_discard_*() operations
Stefan Hajnoczi [Thu, 17 Sep 2020 09:44:54 +0000 (10:44 +0100)] 
virtio-blk: undo destructive iov_discard_*() operations

Fuzzing discovered that virtqueue_unmap_sg() is being called on modified
req->in/out_sg iovecs. This means dma_memory_map() and
dma_memory_unmap() calls do not have matching memory addresses.

Fuzzing discovered that non-RAM addresses trigger a bug:

  void address_space_unmap(AddressSpace *as, void *buffer, hwaddr len,
                           bool is_write, hwaddr access_len)
  {
      if (buffer != bounce.buffer) {
          ^^^^^^^^^^^^^^^^^^^^^^^

A modified iov->iov_base is no longer recognized as a bounce buffer and
the wrong branch is taken.

There are more potential bugs: dirty memory is not tracked correctly and
MemoryRegion refcounts can be leaked.

Use the new iov_discard_undo() API to restore elem->in/out_sg before
virtqueue_push() is called.

Fixes: 827805a2492c1bbf1c0712ed18ee069b4ebf3dd6 ("virtio-blk: Convert VirtIOBlockReq.out to structrue")
Reported-by: Alexander Bulekov <alxndr@bu.edu>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Buglink: https://bugs.launchpad.net/qemu/+bug/1890360
Message-Id: <20200917094455.822379-3-stefanha@redhat.com>

10 months agoutil/iov: add iov_discard_undo()
Stefan Hajnoczi [Thu, 17 Sep 2020 09:44:53 +0000 (10:44 +0100)] 
util/iov: add iov_discard_undo()

The iov_discard_front/back() operations are useful for parsing iovecs
but they modify the array elements. If the original array is needed
after parsing finishes there is currently no way to restore it.

Although g_memdup() can be used before performing destructive
iov_discard_front/back() operations, this is inefficient.

Introduce iov_discard_undo() to restore the array to the state prior to
an iov_discard_front/back() operation.

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Li Qiang <liq3ea@gmail.com>
Message-Id: <20200917094455.822379-2-stefanha@redhat.com>

10 months agovirtio: add vhost-user-fs-ccw device
Halil Pasic [Tue, 1 Sep 2020 15:00:18 +0000 (17:00 +0200)] 
virtio: add vhost-user-fs-ccw device

Wire up the CCW device for vhost-user-fs.

Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Halil Pasic <pasic@linux.ibm.com>
Message-id: 20200901150019.29229-2-mhartmay@linux.ibm.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
10 months agolibvhost-user: handle endianness as mandated by the spec
Marc Hartmayer [Tue, 1 Sep 2020 15:00:19 +0000 (17:00 +0200)] 
libvhost-user: handle endianness as mandated by the spec

Since virtio existed even before it got standardized, the virtio
standard defines the following types of virtio devices:

 + legacy device (pre-virtio 1.0)
 + non-legacy or VIRTIO 1.0 device
 + transitional device (which can act both as legacy and non-legacy)

Virtio 1.0 defines the fields of the virtqueues as little endian,
while legacy uses guest's native endian [1]. Currently libvhost-user
does not handle virtio endianness at all, i.e. it works only if the
native endianness matches with whatever is actually needed. That means
things break spectacularly on big-endian targets. Let us handle virtio
endianness for non-legacy as required by the virtio specification [1]
and fence legacy virtio, as there is no safe way to figure out the
needed endianness conversions for all cases. The fencing of legacy
virtio devices is done in `vu_set_features_exec`.

[1] https://docs.oasis-open.org/virtio/virtio/v1.1/cs01/virtio-v1.1-cs01.html#x1-210003

Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Message-id: 20200901150019.29229-3-mhartmay@linux.ibm.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
10 months agoMAINTAINERS: add Stefan Hajnoczi as block/nvme.c maintainer
Stefan Hajnoczi [Mon, 7 Sep 2020 11:16:32 +0000 (12:16 +0100)] 
MAINTAINERS: add Stefan Hajnoczi as block/nvme.c maintainer

Development of the userspace NVMe block driver picked up again recently.
After talking with Fam I am stepping up as block/nvme.c maintainer.
Patches will be merged through my 'block' tree.

Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Klaus Jensen <k.jensen@samsung.com>
Cc: Fam Zheng <fam@euphon.net>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Klaus Jensen <k.jensen@samsung.com>
Acked-by: Fam Zheng <fam@euphon.net>
Message-id: 20200907111632.90499-1-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
10 months agoaudio: build spiceaudio as module
Gerd Hoffmann [Wed, 16 Sep 2020 08:41:17 +0000 (10:41 +0200)] 
audio: build spiceaudio as module

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20200916084117.21828-3-kraxel@redhat.com

10 months agoaudio: remove qemu_spice_audio_init()
Gerd Hoffmann [Wed, 16 Sep 2020 08:41:16 +0000 (10:41 +0200)] 
audio: remove qemu_spice_audio_init()

Handle the spice special case in audio_init instead.

With the qemu_spice_audio_init() symbol dependency being
gone we can build spiceaudio as module.

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Message-id: 20200916084117.21828-2-kraxel@redhat.com

10 months agoaudio: run downstream playback queue unconditionally
Volker Rümelin [Sun, 20 Sep 2020 17:17:27 +0000 (19:17 +0200)] 
audio: run downstream playback queue unconditionally

Run the downstream playback queue even if there are no samples
in the mixing engine buffer. The downstream queue may still have
queued samples.

Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 20200920171729.15861-7-vr_qemu@t-online.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 months agoaudio: align audio_generic_write with audio_pcm_hw_run_out
Volker Rümelin [Sun, 20 Sep 2020 17:17:26 +0000 (19:17 +0200)] 
audio: align audio_generic_write with audio_pcm_hw_run_out

The function audio_generic_write should work exactly like
audio_pcm_hw_run_out. It's a very similar function working on a
different buffer.

This patch significantly reduces the number of drop-outs with
the DirectSound backend. To hear the difference start qemu with
-audiodev dsound,id=audio0,out.mixing-engine=off and play a
song in the guest with and without this patch.

Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 20200920171729.15861-6-vr_qemu@t-online.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 months agoaudio: remove unnecessary calls to put_buffer_in
Volker Rümelin [Sun, 20 Sep 2020 17:17:25 +0000 (19:17 +0200)] 
audio: remove unnecessary calls to put_buffer_in

This patch removes unnecessary calls to the pcm_ops function
put_buffer_in(). No audio backend needs this call if the
returned length of pcm_ops function get_buffer_in() is zero.

For the DirectSound backend this prevents a call to
dsound_unlock_in() without a preceding call to dsound_lock_in().
While Windows doesn't complain it seems wrong anyway.

Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 20200920171729.15861-5-vr_qemu@t-online.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 months agoaudio: align audio_generic_read with audio_pcm_hw_run_in
Volker Rümelin [Sun, 20 Sep 2020 17:17:24 +0000 (19:17 +0200)] 
audio: align audio_generic_read with audio_pcm_hw_run_in

The function audio_generic_read should work exactly like
audio_pcm_hw_run_in. It's a very similar function working
on a different buffer.

Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 20200920171729.15861-4-vr_qemu@t-online.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 months agoaudio/spiceaudio: always rate limit playback stream
Volker Rümelin [Sun, 20 Sep 2020 17:17:23 +0000 (19:17 +0200)] 
audio/spiceaudio: always rate limit playback stream

The playback rate with the spiceaudio backend is currently too
fast if there's no spice client connected or the spice client
can't play audio. Rate limit the audio playback stream in all
cases. To calculate the rate correctly the limiter has to know
the maximum buffer size.

Fixes: 8c198ff065 ("spiceaudio: port to the new audio backend api")
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 20200920171729.15861-3-vr_qemu@t-online.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 months agoaudio/audio: fix video playback slowdown with spiceaudio
Volker Rümelin [Sun, 20 Sep 2020 17:17:22 +0000 (19:17 +0200)] 
audio/audio: fix video playback slowdown with spiceaudio

This patch allows the audio backends get_buffer_out() functions
to drop audio data and mitigates a bug reported on the qemu-devel
mailing list.

https://lists.nongnu.org/archive/html/qemu-devel/2020-09/msg03832.html

The new rules for the variables buf and size returned by
get_buffer_out() are:
size == 0: Downstream playback buffer is full. Retry later.
size > 0, buf != NULL: Copy size bytes to buf for playback.
size > 0, buf == NULL: Drop size bytes.

The audio playback rate with spiceaudio for the no audio case is
too fast, but that's what we had before commit fb35c2cec5
"audio/dsound: fix invalid parameters error". The complete fix
comes with the next patch.

Reported-by: Qi Zhou <atmgnd@outlook.com>
Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 20200920171729.15861-2-vr_qemu@t-online.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 months agoaudio: handle buf == NULL in put_buffer_out()
Volker Rümelin [Sun, 20 Sep 2020 17:17:21 +0000 (19:17 +0200)] 
audio: handle buf == NULL in put_buffer_out()

With the next patch all audio backends put_buffer_out() functions
have to handle the buf == NULL case, provided the get_buffer_out()
function may return buf = NULL and size > 0.

It turns out that all audio backends get_buffer_out() functions
either can't return buf = NULL or return buf = NULL and size = 0
at the same time. The only exception is the spiceaudio backend
where size may be uninitialized.

Signed-off-by: Volker Rümelin <vr_qemu@t-online.de>
Message-id: 20200920171729.15861-1-vr_qemu@t-online.de
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
10 months agosifive_u: Register "start-in-flash" as class property
Eduardo Habkost [Mon, 21 Sep 2020 22:10:45 +0000 (18:10 -0400)] 
sifive_u: Register "start-in-flash" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20200921221045.699690-25-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agosifive_e: Register "revb" as class property
Eduardo Habkost [Mon, 21 Sep 2020 22:10:44 +0000 (18:10 -0400)] 
sifive_e: Register "revb" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20200921221045.699690-24-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agoi440fx: Register i440FX-pcihost properties as class properties
Eduardo Habkost [Mon, 21 Sep 2020 22:10:43 +0000 (18:10 -0400)] 
i440fx: Register i440FX-pcihost properties as class properties

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20200921221045.699690-23-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agomachine: Register "memory-backend" as class property
Eduardo Habkost [Mon, 21 Sep 2020 22:10:41 +0000 (18:10 -0400)] 
machine: Register "memory-backend" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20200921221045.699690-21-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agoxlnx-zcu102: Register properties as class properties
Eduardo Habkost [Mon, 21 Sep 2020 22:10:40 +0000 (18:10 -0400)] 
xlnx-zcu102: Register properties as class properties

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20200921221045.699690-20-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agocpu/core: Register core-id and nr-threads as class properties
Eduardo Habkost [Mon, 21 Sep 2020 22:10:38 +0000 (18:10 -0400)] 
cpu/core: Register core-id and nr-threads as class properties

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20200921221045.699690-18-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agos390x: Register all CPU properties as class properties
Eduardo Habkost [Mon, 21 Sep 2020 22:10:33 +0000 (18:10 -0400)] 
s390x: Register all CPU properties as class properties

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Message-Id: <20200921221045.699690-13-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agocryptodev-backend: Register "chardev" as class property
Eduardo Habkost [Mon, 21 Sep 2020 22:10:23 +0000 (18:10 -0400)] 
cryptodev-backend: Register "chardev" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Gonglei <arei.gonglei@huawei.com>
Message-Id: <20200921221045.699690-3-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agocryptodev-vhost-user: Register "chardev" as class property
Eduardo Habkost [Mon, 21 Sep 2020 22:10:22 +0000 (18:10 -0400)] 
cryptodev-vhost-user: Register "chardev" as class property

Class properties make QOM introspection simpler and easier, as
they don't require an object to be instantiated.

Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Gonglei <arei.gonglei@huawei.com>
Message-Id: <20200921221045.699690-2-ehabkost@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agosmp: drop support for deprecated (invalid topologies)
Igor Mammedov [Fri, 11 Sep 2020 13:32:02 +0000 (09:32 -0400)] 
smp: drop support for deprecated (invalid topologies)

it's was deprecated since 3.1

Support for invalid topologies is removed, the user must ensure
that topologies described with -smp include all possible cpus,
i.e. (sockets * cores * threads) == maxcpus or QEMU will
exit with error.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Message-Id: <20200911133202.938754-1-imammedo@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agoqom: simplify object_find_property / object_class_find_property
Daniel P. Berrangé [Mon, 14 Sep 2020 13:56:17 +0000 (14:56 +0100)] 
qom: simplify object_find_property / object_class_find_property

When debugging QEMU it is often useful to put a breakpoint on the
error_setg_internal method impl.

Unfortunately the object_property_add / object_class_property_add
methods call object_property_find / object_class_property_find methods
to check if a property exists already before adding the new property.

As a result there are a huge number of calls to error_setg_internal
on startup of most QEMU commands, making it very painful to set a
breakpoint on this method.

Most callers of object_find_property and object_class_find_property,
however, pass in a NULL for the Error parameter. This simplifies the
methods to remove the Error parameter entirely, and then adds some
new wrapper methods that are able to raise an Error when needed.

Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200914135617.1493072-1-berrange@redhat.com>
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
10 months agoMerge remote-tracking branch 'remotes/kraxel/tags/input-20200921-pull-request' into...
Peter Maydell [Tue, 22 Sep 2020 20:11:10 +0000 (21:11 +0100)] 
Merge remote-tracking branch 'remotes/kraxel/tags/input-20200921-pull-request' into staging

input: tsc2xxx fix.

# gpg: Signature made Mon 21 Sep 2020 12:09:17 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/input-20200921-pull-request:
  hw/input/tsc2xxx: Reduce MouseTransformInfo structure exposure

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
10 months agoMerge remote-tracking branch 'remotes/kraxel/tags/usb-20200921-pull-request' into...
Peter Maydell [Tue, 22 Sep 2020 15:40:56 +0000 (16:40 +0100)] 
Merge remote-tracking branch 'remotes/kraxel/tags/usb-20200921-pull-request' into staging

usb: fix u2f build
usb: fix ohci oob access and loop issues

# gpg: Signature made Mon 21 Sep 2020 09:58:06 BST
# gpg:                using RSA key 4CB6D8EED3E87138
# gpg: Good signature from "Gerd Hoffmann (work) <kraxel@redhat.com>" [full]
# gpg:                 aka "Gerd Hoffmann <gerd@kraxel.org>" [full]
# gpg:                 aka "Gerd Hoffmann (private) <kraxel@gmail.com>" [full]
# Primary key fingerprint: A032 8CFF B93A 17A7 9901  FE7D 4CB6 D8EE D3E8 7138

* remotes/kraxel/tags/usb-20200921-pull-request:
  hw: usb: hcd-ohci: check for processed TD before retire
  hw: usb: hcd-ohci: check len and frame_number variables
  usb: fix u2f build

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>