Edivaldo reports a problem that the array of NetClientState in NICState is too
large - MAX_QUEUE_NUM(1024) which will wastes memory even if multiqueue is not
used.
Instead of static arrays, solving this issue by allocating the queues on demand
for both the NetClientState array in NICState and VirtIONetQueue array in
VirtIONet.
Tested by myself, with single virtio-net-pci device. The memory allocation is
almost the same as when multiqueue is not merged.
Cc: Edivaldo de Araujo Pereira <edivaldoapereira@yahoo.com.br>
Cc: qemu-stable@nongnu.org
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Currently, the config size for virtio devices is hard coded. When a new
feature is added that changes the config size, drivers that assume a static
config size will break. For purposes of backward compatibility, there needs
to be a way to inform drivers of the config size needed to accommodate the
set of features enabled.
aliguori: merged in
- hw/virtio-net: use existing macros to implement endof
- hw/virtio-net: fix config_size data type
Signed-off-by: Jesse Larrew <jlarrew@linux.vnet.ibm.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This patch add migration support for multiqueue virtio-net. Instead of bumping
the version, we conditionally send the info of multiqueue only when the device
support more than one queue to maintain the backward compatibility.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This patch implements both userspace and vhost support for multiple queue
virtio-net (VIRTIO_NET_F_MQ). This is done by introducing an array of
VirtIONetQueue to VirtIONet.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
To support multiqueue virtio-net, the first step is to separate the virtqueue
related fields from VirtIONet to a new structure VirtIONetQueue. The following
patches will add an array of VirtIONetQueue to VirtIONet based on this patch.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This patch lets vhost support multiqueue. The idea is simple, just launching
multiple threads of vhost and let each of vhost thread processing a subset of
the virtqueues of the device. After this change each emulated device can have
multiple vhost threads as its backend.
To do this, a virtqueue index were introduced to record to first virtqueue that
will be handled by this vhost_net device. Based on this and nvqs, vhost could
calculate its relative index to setup vhost_net device.
Since we may have many vhost/net devices for a virtio-net device. The setting of
guest notifiers were moved out of the starting/stopping of a specific vhost
thread. The vhost_net_{start|stop}() were renamed to
vhost_net_{start|stop}_one(), and a new vhost_net_{start|stop}() were introduced
to configure the guest notifiers and start/stop all vhost/vhost_net devices.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
To support multiqueue nic, this patch separate the nic destructor from
qemu_del_net_client() to a new helper qemu_del_nic() since the mapping bettween
NiCState and NetClientState were not 1:1 in multiqueue. The following patches
would refactor this function to support multiqueue nic.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
To support multiqueue, this patch introduces a helper qemu_get_nic() to get
NICState from a NetClientState. The following patches would refactor this helper
to support multiqueue.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
To support multiqueue, the patch introduce a helper qemu_get_queue()
which is used to get the NetClientState of a device. The following patches would
refactor this helper to support multiqueue.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
This patch makes rx commands consistent with specification.
Signed-off-by: Amos Kong <akong@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In virtio-net guest driver, currently we write MAC address to
pci config space byte by byte, this means that we have an
intermediate step where mac is wrong. This patch introduced
a new control command to set MAC address, it's atomic.
VIRTIO_NET_F_CTRL_MAC_ADDR is a new feature bit for compatibility.
"mac" field will be set to read-only when VIRTIO_NET_F_CTRL_MAC_ADDR
is acked.
Signed-off-by: Amos Kong <akong@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Virtio-net code makes assumption about virtqueue descriptor layout
(e.g. sg[0] is the header, sg[1] is the data buffer).
This patch makes code not rely on the layout of descriptors.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Amos Kong <akong@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Once guest overrides virtio net primary mac,
it retains the value set until qemu exit.
This is inconsistent with standard nic behaviour.
To fix, revert the mac to the original value on reset.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Support backend guest notifier masking in vhost-net:
create eventfd at device init, when masked,
make vhost use that as eventfd instead of
sending an interrupt.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
As vhost started is cleared last thing on stop,
set it first things on start. This makes it
possible to use vhost_started while start is in
progress which is used by follow-up patches.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Move public headers to include/net, and leave private headers in net/.
Put the virtio headers in include/net/tap.h, removing the multiple copies
that existed. Leave include/net/tap.h as the interface for NICs, and
net/tap_int.h as the interface for OS-specific parts of the tap backend.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Modern linux supports arbitrary header size,
which makes it possible to pass mrg buf header
to tap directly without iovec mangling.
Use this capability when it is there.
This removes the need to deal with it in
vhost-net as we do now.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
There's no reason to query header support at random
times: at load or feature query.
Driver also might not query functions.
Cleaner to do it at device init.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Remove code duplication using guest header length that we track.
Drop specific layout requirement for rx buffers: things work
using generic iovec functions in any case.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Now that we know host hdr length, we don't need to
duplicate the logic in receive_hdr: caller can
figure out the offset itself.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Avoid magling iov manually: use safe iov operations
for processing packets incoming to guest.
This also removes the requirement for virtio header to
fit the first s/g entry exactly.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Avoid tweaking iovec during receive. This removes
the need to copy the vector.
Note: we currently have an evil cast in work_around_broken_dhclient
and unfortunately this patch does not fix it - just
pushes the evil cast to another place.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
qemu_fseek() is known to be wrong. Would be removed on the next
commit. This code should never been used (value has been
MAC_TABLE_ENTRIES since 2009).
Signed-off-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
nc.link_down could not be migrated, this patch updates link_down in
virtio_post_load() to keep it coincident with real link status.
Signed-off-by: Amos Kong <akong@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@gmail.com>
virtio-net has code to flush the queue and notify the iothread
whenever new receive buffers are added by the guest. That is
fine, and indeed we need to do the same in all other drivers.
However, notifying the iothread should be work for the network
subsystem. And since we are at it we can add a little smartness:
if some of the queued packets already could not be delivered,
there is no need to notify the iothread.
Reported-by: Luigi Rizzo <rizzo@iet.unipi.it>
Cc: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Cc: Jan Kiszka <jan.kiszka@siemens.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Amos Kong <akong@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Another step in moving the vlan feature out of net core. Users only
deal with NetClientState and therefore qemu_del_vlan_client() should be
named qemu_del_net_client().
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Zhi Yong Wu <wuzhy@linux.vnet.ibm.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
The vlan feature is no longer part of net core. Rename VLANClientState
to NetClientState because net clients are not explicitly associated with
a vlan at all, instead they have a peer net client to which they are
connected.
This patch is a mechanical search-and-replace except for a few
whitespace fixups where changing VLANClientState to NetClientState
misaligned whitespace.
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Zhi Yong Wu <wuzhy@linux.vnet.ibm.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reorder arguments to be more natural, readable and
consistent with other iov_* functions, and change
argument names, from:
iov_from_buf(iov, iov_cnt, buf, iov_off, size)
to
iov_from_buf(iov, iov_cnt, offset, buf, bytes)
The result becomes natural English:
copy data to this `iov' vector with `iov_cnt'
elements starting at byte offset `offset'
from memory buffer `buf', processing `bytes'
bytes max.
(Try to read the original prototype this way).
Also change iov_clear() to more general iov_memset()
(it uses memset() internally anyway).
While at it, add comments to the header file
describing what the routines actually does.
The patch only renames argumens in the header, but
keeps old names in the implementation. The next
patch will touch actual code to match.
Now, it might look wrong to pay so much attention
to so small things. But we've so many badly designed
interfaces already so the whole thing becomes rather
confusing or error prone. One example of this is
previous commit and small discussion which emerged
from it, with an outcome that the utility functions
like these aren't well-understdandable, leading to
strange usage cases. That's why I paid quite some
attention to this set of functions and a few
others in subsequent patches.
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
Right now, DeviceInfo acts as the class for qdev. In order to switch to a
proper ObjectClass derivative, we need to ween all of the callers off of
interacting directly with the info pointer.
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
virtio_cleanup() will be changed by the following patch to remove the
VirtIONet struct that gets allocated via virtio_common_init(). Ensure
we don't dereference the structure after calling the cleanup function.
Signed-off-by: Amit Shah <amit.shah@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
iov_to_buf() has an 'offset' parameter, iov_from_buf() hasn't.
This patch adds the missing parameter to iov_from_buf().
It also renames the 'offset' parameter to 'iov_off' to
emphasize it's the offset into the iovec and not the buffer.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Acked-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This was done with:
sed -i 's/qemu_get_clock\>/qemu_get_clock_ns/' \
$(git grep -l 'qemu_get_clock\>' )
sed -i 's/qemu_new_timer\>/qemu_new_timer_ns/' \
$(git grep -l 'qemu_new_timer\>' )
after checking that get_clock and new_timer never occur twice
on the same line. There were no missed occurrences; however, even
if there had been, they would have been caught by the compiler.
There was exactly one false positive in qemu_run_timers:
- current_time = qemu_get_clock (clock);
+ current_time = qemu_get_clock_ns (clock);
which is of course not in this patch.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
A pointer to a size_t variable was passed as the void * pointer to
lduw_p() in virtio_net_receive(). Instead of acting on the 16-bit value
this caused failure on big-endian hosts.
Avoid this issue in the future by using stw_p() instead. In general we
should use ld*_p() for loading from target memory and st*_p() for
storing to target memory anyway, not the other way around.
Also tighten up a correct use of lduw_p() when stw_p() should be used
instead in virtio_net_get_config().
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
qemu makes it possible to disable link at tap which is not communicated
to the guest but causes all packets to be dropped.
When vhost-net is enabled, vhost needs to be aware of both the virtio
link_down and the peer link_down. we switch to userspace emulation when
either is down.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reported-by: pradeep <psuriset@linux.vnet.ibm.com>
Acked-by: Alex Williamson <alex.williamson@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
When MSI is off, each interrupt needs to be bounced through the io
thread when it's set/cleared, so vhost-net causes more context switches and
higher CPU utilization than userspace virtio which handles networking in
the same thread.
We'll need to fix this by adding level irq support in kvm irqfd,
for now disable vhost-net in these configurations.
Added a vhostforce flag to force vhost-net back on.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
virtio-net used to work on cross-endianness configurations, but doesn't
anymore with recent guest kernels, as the new features don't handle
endianness correctly.
This patch fixes wrong conversion, and add missing ones to make
virtio-net working. Tested on the following configurations:
- i386 guest on x86_64 host
- ppc guest on x86_64 host
- i386 guest on mips host
- ppc guest on mips host
Cc: Anthony Liguori <aliguori@us.ibm.com>
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Move tracking vmstate change from virtio-net to virtio.c
as it is going to be used by virito-blk and virtio-pci
for the ioeventfd support.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
