mirrors/qemu
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
9df0b0e09c
qemu/include/hw/virtio/vhost.h

85 lines
2.5 KiB
C
Raw Normal View History

vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
#ifndef VHOST_H
#define VHOST_H
#include "hw/hw.h"
Add vhost_ops to vhost_dev struct and replace all relevant ioctls Decouple vhost from the Linux kernel by introducing vhost_ops. The intention is to provide different backends - a 'kernel' backend based on the ioctl interface, and an 'user' backend based on a UNIX domain socket and shared memory interface. Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com> Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2014-05-27 16:05:35 +04:00
#include "hw/virtio/vhost-backend.h"
hw: move headers to include/ Many of these should be cleaned up with proper qdev-/QOM-ification. Right now there are many catch-all headers in include/hw/ARCH depending on cpu.h, and this makes it necessary to compile these files per-target. However, fixing this does not belong in these patches. Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-02-05 20:06:20 +04:00
#include "hw/virtio/virtio.h"
exec: move include files to include/exec/ Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2012-12-17 21:19:49 +04:00
#include "exec/memory.h"
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
/* Generic structures common for any vhost based device. */
struct vhost_virtqueue {
int kick;
int call;
void *desc;
void *avail;
void *used;
int num;
unsigned long long used_phys;
unsigned used_size;
void *ring;
unsigned long long ring_phys;
unsigned ring_size;
vhost: backend masking support 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>
2012-12-24 19:37:01 +04:00
EventNotifier masked_notifier;
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
};
typedef unsigned long vhost_log_chunk_t;
#define VHOST_LOG_PAGE 0x1000
#define VHOST_LOG_BITS (8 * sizeof(vhost_log_chunk_t))
#define VHOST_LOG_CHUNK (VHOST_LOG_PAGE * VHOST_LOG_BITS)
vhost: add vhost_get_features and vhost_ack_features Generalize the features get/ack to be used for both vhost-net and vhost-scsi. In vhost-net add vhost_net_get_feature_bits to select the feature bit set depending on the NetClient kind. Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2014-05-27 16:04:42 +04:00
#define VHOST_INVALID_FEATURE_BIT (0xff)
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
vhost: logs sharing Currently we allocate one vhost log per vhost device. This is sub optimal when: - Guest has several device with vhost as backend - Guest has multiqueue devices In the above cases, we can avoid the memory allocation by sharing a single vhost log among all the vhost devices. This is done through: - Introducing a new vhost_log structure with refcnt inside. - Using a global pointer to vhost_log structure that will be used. And introduce helper to get the log with expected log size and helper to - drop the refcnt to the old log. - Each vhost device still keep track of a pointer to the log that was used. With above, if no resize happens, all vhost device will share a single vhost log. During resize, a new vhost_log structure will be allocated and made for the global pointer. And each vhost devices will drop the refcnt to the old log. Tested by doing scp during migration for a 2 queues virtio-net-pci. Cc: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2015-06-04 12:28:46 +03:00
struct vhost_log {
unsigned long long size;
int refcnt;
vhost_log_chunk_t log[0];
};
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
struct vhost_memory;
struct vhost_dev {
vhost: convert to MemoryListener API Drop the use of cpu_register_phys_memory_client() in favour of the new MemoryListener API. The new API simplifies the caller, since there is no need to deal with splitting and merging slots; however this is not exploited in this patch. Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-18 16:06:05 +04:00
MemoryListener memory_listener;
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
struct vhost_memory *mem;
vhost: avoid cpu_get_physical_page_desc() This reaches into the innards of the memory core, which are being changed. Switch to a memory API version. Signed-off-by: Avi Kivity <avi@redhat.com>
2011-12-19 15:18:13 +04:00
int n_mem_sections;
MemoryRegionSection *mem_sections;
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
struct vhost_virtqueue *vqs;
int nvqs;
vhost: fix typo in vq_index description Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com> Acked-by: Jason Wang <jasowang@redhat.com> Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
2015-03-26 14:10:29 +03:00
/* the first virtqueue which would be used by this vhost dev */
vhost: multiqueue support 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>
2013-01-30 15:12:35 +04:00
int vq_index;
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
unsigned long long features;
unsigned long long acked_features;
unsigned long long backend_features;
bool started;
bool log_enabled;
unsigned long long log_size;
vhost: block migration if backend does not log memory vhost user does not support LOG_ALL feature bit. Generally, we should not try to set this bit without checking that backend can support it first. Detect and block migration. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2014-06-18 18:20:42 +04:00
Error *migration_blocker;
vhost: Add vhost_commit callback for SeaBIOS ROM region re-mapping This patch follows MST's recommendation to move checks for vhost_verify_ring_mappings() -> cpu_physical_memory_map() operations from MemoryListener->region_[add,del]() -> vhost_set_memory() into final MemoryListener->commit() -> vhost_commit() callback. It addresses the case where virtio-scsi vq ioport RAM re-mapping to read-only SeaBIOS ROM triggers a cpu_physical_memory_map() NIL MemoryRegionSection pointer failure. Also save vhost_dev->mem_changed_[start,end]_addr values in vhost_set_memory() for final ranges_overlap checks. (Thanks Paolo!) Cc: Michael S. Tsirkin <mst@redhat.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Asias He <asias@redhat.com> Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org> Acked-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2013-04-03 13:15:11 +04:00
bool memory_changed;
hwaddr mem_changed_start_addr;
hwaddr mem_changed_end_addr;
Add vhost_ops to vhost_dev struct and replace all relevant ioctls Decouple vhost from the Linux kernel by introducing vhost_ops. The intention is to provide different backends - a 'kernel' backend based on the ioctl interface, and an 'user' backend based on a UNIX domain socket and shared memory interface. Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com> Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2014-05-27 16:05:35 +04:00
const VhostOps *vhost_ops;
Add vhost-backend and VhostBackendType Use vhost_set_backend_type to initialise a proper vhost_ops structure. In vhost_net_init and vhost_net_start_one call conditionally TAP related initialisation depending on the vhost backend type. Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com> Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2014-05-27 16:05:49 +04:00
void *opaque;
vhost: logs sharing Currently we allocate one vhost log per vhost device. This is sub optimal when: - Guest has several device with vhost as backend - Guest has multiqueue devices In the above cases, we can avoid the memory allocation by sharing a single vhost log among all the vhost devices. This is done through: - Introducing a new vhost_log structure with refcnt inside. - Using a global pointer to vhost_log structure that will be used. And introduce helper to get the log with expected log size and helper to - drop the refcnt to the old log. - Each vhost device still keep track of a pointer to the log that was used. With above, if no resize happens, all vhost device will share a single vhost log. During resize, a new vhost_log structure will be allocated and made for the global pointer. And each vhost devices will drop the refcnt to the old log. Tested by doing scp during migration for a 2 queues virtio-net-pci. Cc: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Jason Wang <jasowang@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2015-06-04 12:28:46 +03:00
struct vhost_log *log;
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
};
vhost_net_init will use VhostNetOptions to get all its arguments vhost_dev_init will replace devfd and devpath with a single opaque argument. This is initialised with a file descriptor. When TAP is used (through vhost_net), open /dev/vhost-net and pass the fd as an opaque parameter in VhostNetOptions. The same applies to vhost-scsi - open /dev/vhost-scsi and pass the fd. Signed-off-by: Antonios Motakis <a.motakis@virtualopensystems.com> Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2014-05-27 16:05:22 +04:00
int vhost_dev_init(struct vhost_dev *hdev, void *opaque,
vhost: enable vhost without without MSI-X We use vhostforce to enable vhost even if Guests don't have MSI-X support and we fall back to QEMU virtio-net. This gives a very small performance gain, but the disadvantage is that guest now controls which virtio code is running (qemu or vhost) so our attack surface is doubled. This patch will enable vhost unconditionally whenever it's requested. For compatibility, enable vhost when vhostforce is set, as well. Signed-off-by: Pankaj Gupta <pagupta@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Jason Wang <jasowang@redhat.com>
2015-06-16 11:18:59 +03:00
VhostBackendType backend_type);
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
void vhost_dev_cleanup(struct vhost_dev *hdev);
vhost: force vhost off for non-MSI guests 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>
2011-02-01 23:13:42 +03:00
bool vhost_dev_query(struct vhost_dev *hdev, VirtIODevice *vdev);
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
int vhost_dev_start(struct vhost_dev *hdev, VirtIODevice *vdev);
void vhost_dev_stop(struct vhost_dev *hdev, VirtIODevice *vdev);
vhost-net: cleanup host notifiers at last step When the vhost notifier is disabled, the userspace handler runs immediately: virtio_pci_set_host_notifier_internal might call virtio_queue_notify_vq. Since the VQ state and the tap backend state aren't recovered yet, this causes "Guest moved used index from XXX to YYY" assertions. The solution is to split out host notifier handling from vhost VQ setup and disable notifiers as our last step when we stop vhost-net. For symmetry enable them first thing on start. Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2011-08-11 11:21:18 +04:00
int vhost_dev_enable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev);
void vhost_dev_disable_notifiers(struct vhost_dev *hdev, VirtIODevice *vdev);
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
vhost: backend masking support 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>
2012-12-24 19:37:01 +04:00
/* Test and clear masked event pending status.
* Should be called after unmask to avoid losing events.
*/
bool vhost_virtqueue_pending(struct vhost_dev *hdev, int n);
/* Mask/unmask events from this vq.
*/
void vhost_virtqueue_mask(struct vhost_dev *hdev, VirtIODevice *vdev, int n,
bool mask);
vhost: 64 bit features Make sure that all vhost interfaces use 64 bit features, as the virtio core does, and make sure to use ULL everywhere possible to be on the safe side. Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
2015-06-04 13:34:20 +03:00
uint64_t vhost_get_features(struct vhost_dev *hdev, const int *feature_bits,
uint64_t features);
vhost: add vhost_get_features and vhost_ack_features Generalize the features get/ack to be used for both vhost-net and vhost-scsi. In vhost-net add vhost_net_get_feature_bits to select the feature bit set depending on the NetClient kind. Signed-off-by: Nikolay Nikolaev <n.nikolaev@virtualopensystems.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2014-05-27 16:04:42 +04:00
void vhost_ack_features(struct vhost_dev *hdev, const int *feature_bits,
vhost: 64 bit features Make sure that all vhost interfaces use 64 bit features, as the virtio core does, and make sure to use ULL everywhere possible to be on the safe side. Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
2015-06-04 13:34:20 +03:00
uint64_t features);
vhost: vhost net support This adds vhost net device support in qemu. Will be tied to tap device and virtio by following patches. Raw backend is currently missing, will be worked on/submitted separately. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-17 14:08:17 +03:00
#endif
Reference in New Issue Copy Permalink