Add virtio-fs (experimental)
cover Add virtio-fs (experimental)
From: "Dr. David Alan Gilbert" <[email protected]>
Hi,
This pair of patches adds the core of the virtio-fs support to qemu;
it's marked experimental since the kernel patch and spec changes aren't in yet; but they're bubbling along.
While the spec change is still in progress; the ID number is already
reserved.
A future set of patches will add the optional DAX mapping support.
The actual qemu change is pretty minimal, since it's really only
a virtio device with some queues.
Some links:
Mailing list: https://www.redhat.com/mailman/listinfo/virtio-fs
Dev tree: Including filesystem daemon: https://gitlab.com/virtio-fs/qemu
kernel: https://gitlab.com/virtio-fs/linux
virtio spec changes: https://lists.oasis-open.org/archives/virtio-dev/201908/msg00056.html
Dr. David Alan Gilbert (2):
virtio: add vhost-user-fs base device
virtio: add vhost-user-fs-pci device
configure | 13 + hw/virtio/Makefile.objs | 2 + hw/virtio/vhost-user-fs-pci.c | 79 ++++++ hw/virtio/vhost-user-fs.c | 297 ++++++++++++++++++++ include/hw/virtio/vhost-user-fs.h | 45 +++ include/standard-headers/linux/virtio_fs.h | 41 +++ include/standard-headers/linux/virtio_ids.h | 1 + 7 files changed, 478 insertions(+) create mode 100644 hw/virtio/vhost-user-fs-pci.c create mode 100644 hw/virtio/vhost-user-fs.c create mode 100644 include/hw/virtio/vhost-user-fs.h create mode 100644 include/standard-headers/linux/virtio_fs.h
Signature · 1 line
2.21.0
PATCH 1/2 virtio: add vhost-user-fs base device
From: "Dr. David Alan Gilbert" <[email protected]>
The virtio-fs virtio device provides shared file system access using the FUSE protocol carried ovew virtio.
The actual file server is implemented in an external vhost-user-fs device backend process.
--- configure | 13 + hw/virtio/Makefile.objs | 1 + hw/virtio/vhost-user-fs.c | 297 ++++++++++++++++++++ include/hw/virtio/vhost-user-fs.h | 45 +++ include/standard-headers/linux/virtio_fs.h | 41 +++ include/standard-headers/linux/virtio_ids.h | 1 + 6 files changed, 398 insertions(+) create mode 100644 hw/virtio/vhost-user-fs.c create mode 100644 include/hw/virtio/vhost-user-fs.h create mode 100644 include/standard-headers/linux/virtio_fs.h
configure +13 -0
@@ -382,6 +382,7 @@ vhost_crypto="" vhost_scsi="" vhost_vsock="" vhost_user="" +vhost_user_fs="" kvm="no" hax="no" hvf="no"
@@ -1316,6 +1317,10 @@ for opt do ;; --enable-vhost-vsock) vhost_vsock="yes" ;; + --disable-vhost-user-fs) vhost_user_fs="no" + ;; + --enable-vhost-user-fs) vhost_user_fs="yes" + ;; --disable-opengl) opengl="no" ;; --enable-opengl) opengl="yes"
@@ -2269,6 +2274,10 @@ test "$vhost_crypto" = "" && vhost_crypto=$vhost_user if test "$vhost_crypto" = "yes" && test "$vhost_user" = "no"; then error_exit "--enable-vhost-crypto requires --enable-vhost-user" fi +test "$vhost_user_fs" = "" && vhost_user_fs=$vhost_user +if test "$vhost_user_fs" = "yes" && test "$vhost_user" = "no"; then + error_exit "--enable-vhost-user-fs requires --enable-vhost-user" +fi # OR the vhost-kernel and vhost-user values for simplicity if test "$vhost_net" = ""; then
@@ -6425,6 +6434,7 @@ echo "vhost-crypto support $vhost_crypto" echo "vhost-scsi support $vhost_scsi" echo "vhost-vsock support $vhost_vsock" echo "vhost-user support $vhost_user" +echo "vhost-user-fs support $vhost_user_fs" echo "Trace backends $trace_backends" if have_backend "simple"; then echo "Trace output file $trace_file-<pid>"
@@ -6921,6 +6931,9 @@ fi if test "$vhost_user" = "yes" ; then echo "CONFIG_VHOST_USER=y" >> $config_host_mak fi +if test "$vhost_user_fs" = "yes" ; then + echo "CONFIG_VHOST_USER_FS=y" >> $config_host_mak +fi if test "$blobs" = "yes" ; then echo "INSTALL_BLOBS=yes" >> $config_host_mak fi
hw/virtio/Makefile.objs +1 -0
@@ -11,6 +11,7 @@ common-obj-$(CONFIG_VIRTIO_PCI) += virtio-pci.o common-obj-$(CONFIG_VIRTIO_MMIO) += virtio-mmio.o obj-$(CONFIG_VIRTIO_BALLOON) += virtio-balloon.o obj-$(CONFIG_VIRTIO_CRYPTO) += virtio-crypto.o +obj-$(CONFIG_VHOST_USER_FS) += vhost-user-fs.o obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-pci.o obj-$(CONFIG_VIRTIO_PMEM) += virtio-pmem.o common-obj-$(call land,$(CONFIG_VIRTIO_PMEM),$(CONFIG_VIRTIO_PCI)) += virtio-pmem-pci.o
hw/virtio/vhost-user-fs.c new file +297 -0
297 lines not on this page · raw message
include/hw/virtio/vhost-user-fs.h new file +45 -0
45 lines not on this page · raw message
include/standard-headers/linux/virtio_fs.h new file +41 -0
41 lines not on this page · raw message
include/standard-headers/linux/virtio_ids.h +1 -0
@@ -43,6 +43,7 @@#define VIRTIO_ID_INPUT 18/* virtio input */ #define VIRTIO_ID_VSOCK 19/* virtio vsock transport */ #define VIRTIO_ID_CRYPTO 20/* virtio crypto */ +#define VIRTIO_ID_FS 26/* virtio filesystem */ #define VIRTIO_ID_PMEM 27/* virtio pmem */ #endif /* _LINUX_VIRTIO_IDS_H */
Signature · 1 line
2.21.0
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
Quoted 220 lines · Dr. David Alan Gilbert (git)
From: "Dr. David Alan Gilbert" <[email protected]>
The virtio-fs virtio device provides shared file system access using the FUSE protocol carried ovew virtio.
The actual file server is implemented in an external vhost-user-fs device backend process.Signed-off-by Stefan Hajnoczi <[email protected]>Signed-off-by Sebastien Boeuf <[email protected]>Signed-off-by Dr. David Alan Gilbert <[email protected]>--- configure | 13 + hw/virtio/Makefile.objs | 1 + hw/virtio/vhost-user-fs.c | 297 ++++++++++++++++++++ include/hw/virtio/vhost-user-fs.h | 45 +++ include/standard-headers/linux/virtio_fs.h | 41 +++ include/standard-headers/linux/virtio_ids.h | 1 + 6 files changed, 398 insertions(+) create mode 100644 hw/virtio/vhost-user-fs.c create mode 100644 include/hw/virtio/vhost-user-fs.h create mode 100644 include/standard-headers/linux/virtio_fs.hconfigure +13
-0@@ -382,6 +382,7 @@ vhost_crypto="" vhost_scsi="" vhost_vsock="" vhost_user="" +vhost_user_fs="" kvm="no" hax="no" hvf="no"@@ -1316,6 +1317,10 @@ for opt do ;; --enable-vhost-vsock) vhost_vsock="yes" ;; + --disable-vhost-user-fs) vhost_user_fs="no" + ;; + --enable-vhost-user-fs) vhost_user_fs="yes" + ;; --disable-opengl) opengl="no" ;; --enable-opengl) opengl="yes"@@ -2269,6 +2274,10 @@ test "$vhost_crypto" = "" && vhost_crypto=$vhost_user if test "$vhost_crypto" = "yes" && test "$vhost_user" = "no"; then error_exit "--enable-vhost-crypto requires --enable-vhost-user" fi +test "$vhost_user_fs" = "" && vhost_user_fs=$vhost_user +if test "$vhost_user_fs" = "yes" && test "$vhost_user" = "no"; then + error_exit "--enable-vhost-user-fs requires --enable-vhost-user" +fi # OR the vhost-kernel and vhost-user values for simplicity if test "$vhost_net" = ""; then@@ -6425,6 +6434,7 @@ echo "vhost-crypto support $vhost_crypto" echo "vhost-scsi support $vhost_scsi" echo "vhost-vsock support $vhost_vsock" echo "vhost-user support $vhost_user" +echo "vhost-user-fs support $vhost_user_fs" echo "Trace backends $trace_backends" if have_backend "simple"; then echo "Trace output file $trace_file-<pid>"@@ -6921,6 +6931,9 @@ fi if test "$vhost_user" = "yes" ; then echo "CONFIG_VHOST_USER=y" >> $config_host_mak fi +if test "$vhost_user_fs" = "yes" ; then + echo "CONFIG_VHOST_USER_FS=y" >> $config_host_mak +fi if test "$blobs" = "yes" ; then echo "INSTALL_BLOBS=yes" >> $config_host_mak fihw/virtio/Makefile.objs +1
-0@@ -11,6 +11,7 @@ common-obj-$(CONFIG_VIRTIO_PCI) += virtio-pci.o common-obj-$(CONFIG_VIRTIO_MMIO) += virtio-mmio.o obj-$(CONFIG_VIRTIO_BALLOON) += virtio-balloon.o obj-$(CONFIG_VIRTIO_CRYPTO) += virtio-crypto.o +obj-$(CONFIG_VHOST_USER_FS) += vhost-user-fs.o obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-pci.o obj-$(CONFIG_VIRTIO_PMEM) += virtio-pmem.o common-obj-$(call land,$(CONFIG_VIRTIO_PMEM),$(CONFIG_VIRTIO_PCI)) += virtio-pmem-pci.ohw/virtio/vhost-user-fs.c new file +128
-0@@ -0,0 +1,297 @@ +/* +* Vhost-user filesystem virtio device +* +* Copyright 2018 Red Hat, Inc. +* +* Authors: +* Stefan Hajnoczi <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or +* (at your option) any later version. See the COPYING file in the +* top-level directory. +*/ + +#include "qemu/osdep.h" +#include <sys/ioctl.h> +#include "standard-headers/linux/virtio_fs.h" +#include "qapi/error.h" +#include "hw/virtio/virtio-bus.h" +#include "hw/virtio/virtio-access.h" +#include "qemu/error-report.h" +#include "hw/virtio/vhost-user-fs.h" +#include "monitor/monitor.h" + +static void vuf_get_config (VirtIODevice *vdev,uint8_t *config) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +struct virtio_fs_config fscfg = {}; + +memcpy ((char *)fscfg.tag, fs->conf.tag, +MIN (strlen (fs->conf.tag) +1 ,sizeof (fscfg.tag))); + +virtio_stl_p (vdev, &fscfg.num_queues, fs->conf.num_queues); + +memcpy (config, &fscfg,sizeof (fscfg)); +} + +static void vuf_start (VirtIODevice *vdev) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +BusState *qbus =BUS (qdev_get_parent_bus (DEVICE (vdev))); +VirtioBusClass *k =VIRTIO_BUS_GET_CLASS (qbus); +int ret; +int i; + +if (!k->set_guest_notifiers) { +error_report ("binding does not support guest notifiers" ); +return ; + } + + ret =vhost_dev_enable_notifiers (&fs->vhost_dev, vdev); +if (ret <0 ) { +error_report ("Error enabling host notifiers: %d" , -ret); +return ; + } + + ret = k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, true); +if (ret <0 ) { +error_report ("Error binding guest notifier: %d" , -ret); + goto err_host_notifiers; + } + + fs->vhost_dev.acked_features = vdev->guest_features; + ret =vhost_dev_start (&fs->vhost_dev, vdev); +if (ret <0 ) { +error_report ("Error starting vhost: %d" , -ret); + goto err_guest_notifiers; + } + +/* +* guest_notifier_mask/pending not used yet, so just unmask +* everything here. virtio-pci will do the right thing by +* enabling/disabling irqfd. +*/ +for (i =0 ; i < fs->vhost_dev.nvqs; i++) { +vhost_virtqueue_mask (&fs->vhost_dev, vdev, i, false); + } + +return ; + +err_guest_notifiers: + k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, false); +err_host_notifiers: +vhost_dev_disable_notifiers (&fs->vhost_dev, vdev); +} + +static void vuf_stop (VirtIODevice *vdev) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +BusState *qbus =BUS (qdev_get_parent_bus (DEVICE (vdev))); +VirtioBusClass *k =VIRTIO_BUS_GET_CLASS (qbus); +int ret; + +if (!k->set_guest_notifiers) { +return ; + } + +vhost_dev_stop (&fs->vhost_dev, vdev); + + ret = k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, false); +if (ret <0 ) { +error_report ("vhost guest notifier cleanup failed: %d" , ret); +return ; + } + +vhost_dev_disable_notifiers (&fs->vhost_dev, vdev); +} + +static void vuf_set_status (VirtIODevice *vdev,uint8_t status) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +bool should_start = status &VIRTIO_CONFIG_S_DRIVER_OK ; + +if (!vdev->vm_running) { + should_start = false; + } + +if (fs->vhost_dev.started == should_start) { +return ; + } + +if (should_start) { +vuf_start (vdev); + }else { +vuf_stop (vdev); + } +} + +static uint64_t vuf_get_features(VirtIODevice *vdev,
hw/virtio/vhost-user-fs.c line 138
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +10
-0+uint64_t requested_features, + Error **errp) +{ + /* No feature bits used yet */ + return requested_features; +} + +static void vuf_handle_output (VirtIODevice *vdev,VirtQueue *vq) +{ +/* Do nothing */
Why is this safe? Is this because this never triggers? assert(0) then? If it triggers then backend won't be notified, which might cause it to get stuck.
Quoted 39 lines
hw/virtio/vhost-user-fs.c new file +39
-0+} + +static void vuf_guest_notifier_mask (VirtIODevice *vdev,int idx, +bool mask) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +vhost_virtqueue_mask (&fs->vhost_dev, vdev, idx, mask); +} + +static bool vuf_guest_notifier_pending (VirtIODevice *vdev,int idx) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +return vhost_virtqueue_pending (&fs->vhost_dev, idx); +} + +static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + }
hw/virtio/vhost-user-fs.c line 187
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +10
-0+ if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" ); +return ; + }
The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.
hw/virtio/vhost-user-fs.c line 192
Quoted 5 lines
hw/virtio/vhost-user-fs.c new file +5
-0+ +if (!is_power_of_2 (fs->conf.queue_size)) { +error_setg (errp,"queue-size property must be a power of 2" ); +return ; + }
Hmm packed ring allows non power of 2 ... We need to look into a generic helper to support VQ size checks.
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +9
-0+ +if (fs->conf.queue_size >VIRTQUEUE_MAX_SIZE ) { +error_setg (errp,"queue-size property must be %u or smaller" , +VIRTQUEUE_MAX_SIZE ); +return ; + } + +if (!vhost_user_init (&fs->vhost_user, &fs->conf.chardev, errp)) { +return ;
hw/virtio/vhost-user-fs.c line 211
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +10
-0+ } + +virtio_init (vdev,"vhost-user-fs" ,VIRTIO_ID_FS , +sizeof (struct virtio_fs_config )); + +/* Notifications queue */ +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output); + +/* Hiprio queue */ +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output);
Weird, spec patch v6 says:
+\item[0] hiprio
+\item[1\ldots n] request queues
where's the Notifications queue coming from?
Quoted 197 lines
hw/virtio/vhost-user-fs.c new file +85
-0+/* Request queues */ +for (i =0 ; i < fs->conf.num_queues; i++) { +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output); + } + +/* 1 high prio queue, plus the number configured */ + fs->vhost_dev.nvqs =1 + fs->conf.num_queues; + fs->vhost_dev.vqs = g_new0(struct vhost_virtqueue, fs->vhost_dev.nvqs); + ret =vhost_dev_init (&fs->vhost_dev, &fs->vhost_user, +VHOST_BACKEND_TYPE_USER ,0 ); +if (ret <0 ) { +error_setg_errno (errp, -ret,"vhost_dev_init failed" ); + goto err_virtio; + } + +return ; + +err_virtio: +vhost_user_cleanup (&fs->vhost_user); +virtio_cleanup (vdev); +g_free (fs->vhost_dev.vqs); +return ; +} + +static void vuf_device_unrealize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); + +/* This will stop vhost backend if appropriate. */ +vuf_set_status (vdev,0 ); + +vhost_dev_cleanup (&fs->vhost_dev); + +vhost_user_cleanup (&fs->vhost_user); + +virtio_cleanup (vdev); +g_free (fs->vhost_dev.vqs); + fs->vhost_dev.vqs =NULL ; +} + +static const VMStateDescription vuf_vmstate = { + .name ="vhost-user-fs" , + .unmigratable =1 , +}; + +static Property vuf_properties[] = { +DEFINE_PROP_CHR ("chardev" , VHostUserFS, conf.chardev), +DEFINE_PROP_STRING ("tag" , VHostUserFS, conf.tag), +DEFINE_PROP_UINT16 ("num-queues" , VHostUserFS, conf.num_queues,1 ), +DEFINE_PROP_UINT16 ("queue-size" , VHostUserFS, conf.queue_size,128 ), +DEFINE_PROP_STRING ("vhostfd" , VHostUserFS, conf.vhostfd), +DEFINE_PROP_END_OF_LIST (), +}; + +static void vuf_class_init (ObjectClass *klass,void *data) +{ +DeviceClass *dc =DEVICE_CLASS (klass); +VirtioDeviceClass *vdc =VIRTIO_DEVICE_CLASS (klass); + + dc->props = vuf_properties; + dc->vmsd = &vuf_vmstate; +set_bit (DEVICE_CATEGORY_STORAGE , dc->categories); + vdc->realize = vuf_device_realize; + vdc->unrealize = vuf_device_unrealize; + vdc->get_features = vuf_get_features; + vdc->get_config = vuf_get_config; + vdc->set_status = vuf_set_status; + vdc->guest_notifier_mask = vuf_guest_notifier_mask; + vdc->guest_notifier_pending = vuf_guest_notifier_pending; +} + +static const TypeInfo vuf_info = { + .name =TYPE_VHOST_USER_FS , + .parent =TYPE_VIRTIO_DEVICE , + .instance_size =sizeof (VHostUserFS), + .class_init = vuf_class_init, +}; + +static void vuf_register_types (void ) +{ +type_register_static (&vuf_info); +} + +type_init (vuf_register_types)include/hw/virtio/vhost-user-fs.h new file +45
-0@@ -0,0 +1,45 @@ +/* +* Vhost-user filesystem virtio device +* +* Copyright 2018 Red Hat, Inc. +* +* Authors: +* Stefan Hajnoczi <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or +* (at your option) any later version. See the COPYING file in the +* top-level directory. +*/ + +#ifndef _QEMU_VHOST_USER_FS_H +#define _QEMU_VHOST_USER_FS_H + +#include "hw/virtio/virtio.h" +#include "hw/virtio/vhost.h" +#include "hw/virtio/vhost-user.h" +#include "chardev/char-fe.h" + +#define TYPE_VHOST_USER_FS "x-vhost-user-fs-device" +#define VHOST_USER_FS (obj) \ + OBJECT_CHECK(VHostUserFS, (obj), TYPE_VHOST_USER_FS) + +typedef struct { +CharBackend chardev; +char *tag; +uint16_t num_queues; +uint16_t queue_size; +char *vhostfd; +}VHostUserFSConf ; + +typedef struct { +/*< private >*/ +VirtIODevice parent; +VHostUserFSConf conf; +struct vhost_virtqueue *vhost_vqs; +struct vhost_dev vhost_dev; +VhostUserState vhost_user; + +/*< public >*/ +}VHostUserFS ; + +#endif /* _QEMU_VHOST_USER_FS_H */ include/standard-headers/linux/virtio_fs.h new file +41
-0@@ -0,0 +1,41 @@ +#ifndef _LINUX_VIRTIO_FS_H +#define _LINUX_VIRTIO_FS_H +/* This header is BSD licensed so anyone can use the definitions to implement +* compatible drivers/servers. +* +* Redistribution and use in source and binary forms, with or without +* modification, are permitted provided that the following conditions +* are met: +* 1. Redistributions of source code must retain the above copyright +* notice, this list of conditions and the following disclaimer. +* 2. Redistributions in binary form must reproduce the above copyright +* notice, this list of conditions and the following disclaimer in the +* documentation and/or other materials provided with the distribution. +* 3. Neither the name of IBM nor the names of its contributors +* may be used to endorse or promote products derived from this software +* without specific prior written permission. +* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ``AS IS'' AND +* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +* ARE DISCLAIMED. IN NO EVENT SHALL IBM OR CONTRIBUTORS BE LIABLE +* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS +* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) +* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT +* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY +* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF +* SUCH DAMAGE. */ +#include "standard-headers/linux/types.h" +#include "standard-headers/linux/virtio_ids.h" +#include "standard-headers/linux/virtio_config.h" +#include "standard-headers/linux/virtio_types.h" + +struct virtio_fs_config { +/* Filesystem name (UTF-8, not NUL-terminated, padded with NULs) */ +uint8_t tag[36 ]; + +/* Number of request queues */ +uint32_t num_queues; +}QEMU_PACKED ; + +#endif /* _LINUX_VIRTIO_FS_H */ include/standard-headers/linux/virtio_ids.h +1
-0@@ -43,6 +43,7 @@#define VIRTIO_ID_INPUT 18/* virtio input */ #define VIRTIO_ID_VSOCK 19/* virtio vsock transport */ #define VIRTIO_ID_CRYPTO 20/* virtio crypto */ +#define VIRTIO_ID_FS 26/* virtio filesystem */ #define VIRTIO_ID_PMEM 27/* virtio pmem */ #endif /* _LINUX_VIRTIO_IDS_H */ Signature · 1 line
2.21.0
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
Quoted 43 lines · On Sun, 18 Aug 2019 07:08:31 -0400 "Michael S. Tsirkin" <[email protected]> wrote:
Quoted 42 lines · Dr. David Alan Gilbert (git)
From: "Dr. David Alan Gilbert" <[email protected]>
The virtio-fs virtio device provides shared file system access using the FUSE protocol carried ovew virtio.
The actual file server is implemented in an external vhost-user-fs device backend process.Signed-off-by Stefan Hajnoczi <[email protected]>Signed-off-by Sebastien Boeuf <[email protected]>Signed-off-by Dr. David Alan Gilbert <[email protected]>--- configure | 13 + hw/virtio/Makefile.objs | 1 + hw/virtio/vhost-user-fs.c | 297 ++++++++++++++++++++ include/hw/virtio/vhost-user-fs.h | 45 +++ include/standard-headers/linux/virtio_fs.h | 41 +++ include/standard-headers/linux/virtio_ids.h | 1 + 6 files changed, 398 insertions(+) create mode 100644 hw/virtio/vhost-user-fs.c create mode 100644 include/hw/virtio/vhost-user-fs.h create mode 100644 include/standard-headers/linux/virtio_fs.hhw/virtio/vhost-user-fs.c new file +14
-0@@ -0,0 +1,297 @@ +/* +* Vhost-user filesystem virtio device +* +* Copyright 2018 Red Hat, Inc. +* +* Authors: +* Stefan Hajnoczi <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or +* (at your option) any later version. See the COPYING file in the +* top-level directory. +*/ + +#include "qemu/osdep.h"
Quoted 8 lines
hw/virtio/vhost-user-fs.c line 22
Quoted 8 lines
hw/virtio/vhost-user-fs.c new file +8
-0+#include <sys/ioctl.h> +#include "standard-headers/linux/virtio_fs.h" +#include "qapi/error.h" +#include "hw/virtio/virtio-bus.h" +#include "hw/virtio/virtio-access.h" +#include "qemu/error-report.h" +#include "hw/virtio/vhost-user-fs.h" +#include "monitor/monitor.h"
JFYI, this needs to include hw/qdev-properties.h as well on the latest code level. (As does the pci part.)
Quoted 42 lines
Quoted 42 lines
hw/virtio/vhost-user-fs.c new file +42
-0+ +static void vuf_get_config (VirtIODevice *vdev,uint8_t *config) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +struct virtio_fs_config fscfg = {}; + +memcpy ((char *)fscfg.tag, fs->conf.tag, +MIN (strlen (fs->conf.tag) +1 ,sizeof (fscfg.tag))); + +virtio_stl_p (vdev, &fscfg.num_queues, fs->conf.num_queues); + +memcpy (config, &fscfg,sizeof (fscfg)); +} + +static void vuf_start (VirtIODevice *vdev) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +BusState *qbus =BUS (qdev_get_parent_bus (DEVICE (vdev))); +VirtioBusClass *k =VIRTIO_BUS_GET_CLASS (qbus); +int ret; +int i; + +if (!k->set_guest_notifiers) { +error_report ("binding does not support guest notifiers" ); +return ; + } + + ret =vhost_dev_enable_notifiers (&fs->vhost_dev, vdev); +if (ret <0 ) { +error_report ("Error enabling host notifiers: %d" , -ret); +return ; + } + + ret = k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, true); +if (ret <0 ) { +error_report ("Error binding guest notifier: %d" , -ret); + goto err_host_notifiers; + } + + fs->vhost_dev.acked_features = vdev->guest_features; + ret =vhost_dev_start (&fs->vhost_dev, vdev); +if (ret <0 ) {
Quoted 8 lines
hw/virtio/vhost-user-fs.c line 72
Quoted 8 lines
hw/virtio/vhost-user-fs.c new file +8
-0+error_report ("Error starting vhost: %d" , -ret); + goto err_guest_notifiers; + } + +/* +* guest_notifier_mask/pending not used yet, so just unmask +* everything here. virtio-pci will do the right thing by +* enabling/disabling irqfd.
Referring to 'virtio-pci' seems a bit suspicious :) Should that be 'the transport'? (And 'the right thing' is not really self-explanatory...)
(I have wired it up for virtio-ccw, but have not actually tried it out. Will send it out once I did.)
Quoted 12 lines
Quoted 12 lines
hw/virtio/vhost-user-fs.c new file +12
-0+*/ +for (i =0 ; i < fs->vhost_dev.nvqs; i++) { +vhost_virtqueue_mask (&fs->vhost_dev, vdev, i, false); + } + +return ; + +err_guest_notifiers: + k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, false); +err_host_notifiers: +vhost_dev_disable_notifiers (&fs->vhost_dev, vdev); +}
(...)
Quoted 30 lines
Quoted 30 lines
hw/virtio/vhost-user-fs.c new file +30
-0+static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + } + if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" );
Quoted 8 lines
hw/virtio/vhost-user-fs.c
Quoted 2 lines
hw/virtio/vhost-user-fs.c new file +2
-0+return ; + }The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.
I think the spec states that num_queues in the config space is the number of request queues only. Can we rename to num_request_queues? The hiprio queue is not really configurable, anyway.
Quoted 10 lines
hw/virtio/vhost-user-fs.c line 192
Quoted 6 lines
hw/virtio/vhost-user-fs.c new file +5
-0+ +if (!is_power_of_2 (fs->conf.queue_size)) { +error_setg (errp,"queue-size property must be a power of 2" ); +return ; + }Hmm packed ring allows non power of 2 ... We need to look into a generic helper to support VQ size checks.
Huh, I didn't notice before that there are several devices which already allow to configure the queue size... looking, there seem to be the following cases:
- bound checks and checks for power of 2 (blk, net)
- no checks (scsi) -- isn't that dangerous, as virtio_add_queue() will abort() for a too large value?
Anyway, if we have a non power of 2 size and the driver does not negotiate packed, we can just fail setting FEATURES_OK, so dropping the power of 2 check should be fine, at least when we add packed support.
Quoted 24 lines
Quoted 11 lines
hw/virtio/vhost-user-fs.c new file +9
-0+ +if (fs->conf.queue_size >VIRTQUEUE_MAX_SIZE ) { +error_setg (errp,"queue-size property must be %u or smaller" , +VIRTQUEUE_MAX_SIZE ); +return ; + } + +if (!vhost_user_init (&fs->vhost_user, &fs->conf.chardev, errp)) { +return ;hw/virtio/vhost-user-fs.c line 211
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +10
-0+ } + +virtio_init (vdev,"vhost-user-fs" ,VIRTIO_ID_FS , +sizeof (struct virtio_fs_config )); + +/* Notifications queue */ +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output); + +/* Hiprio queue */ +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output);Weird, spec patch v6 says:
Quoted 4 lines
+\item[0] hiprio
+\item[1\ldots n] request queueswhere's the Notifications queue coming from?
Maybe an old name of the hiprio queue?
Quoted 8 lines
hw/virtio/vhost-user-fs.c line 219
Quoted 8 lines
hw/virtio/vhost-user-fs.c new file +7
-0+/* Request queues */ +for (i =0 ; i < fs->conf.num_queues; i++) { +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output); + } + +/* 1 high prio queue, plus the number configured */ + fs->vhost_dev.nvqs =1 + fs->conf.num_queues;
Anyway, the notifications queue needs to go, or this is wrong :)
Quoted 16 lines
Quoted 16 lines
hw/virtio/vhost-user-fs.c new file +16
-0+ fs->vhost_dev.vqs = g_new0(struct vhost_virtqueue, fs->vhost_dev.nvqs); + ret =vhost_dev_init (&fs->vhost_dev, &fs->vhost_user, +VHOST_BACKEND_TYPE_USER ,0 ); +if (ret <0 ) { +error_setg_errno (errp, -ret,"vhost_dev_init failed" ); + goto err_virtio; + } + +return ; + +err_virtio: +vhost_user_cleanup (&fs->vhost_user); +virtio_cleanup (vdev); +g_free (fs->vhost_dev.vqs); +return ; +}
(...)
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
Quoted 49 lines · * Cornelia Huck ([email protected]) wrote:
Quoted 46 lines · On Sun, 18 Aug 2019 07:08:31 -0400 "Michael S. Tsirkin" <[email protected]> wrote:
Quoted 45 lines · Dr. David Alan Gilbert (git)
From: "Dr. David Alan Gilbert" <[email protected]>
The virtio-fs virtio device provides shared file system access using the FUSE protocol carried ovew virtio.
The actual file server is implemented in an external vhost-user-fs device backend process.Signed-off-by Stefan Hajnoczi <[email protected]>Signed-off-by Sebastien Boeuf <[email protected]>Signed-off-by Dr. David Alan Gilbert <[email protected]>--- configure | 13 + hw/virtio/Makefile.objs | 1 + hw/virtio/vhost-user-fs.c | 297 ++++++++++++++++++++ include/hw/virtio/vhost-user-fs.h | 45 +++ include/standard-headers/linux/virtio_fs.h | 41 +++ include/standard-headers/linux/virtio_ids.h | 1 + 6 files changed, 398 insertions(+) create mode 100644 hw/virtio/vhost-user-fs.c create mode 100644 include/hw/virtio/vhost-user-fs.h create mode 100644 include/standard-headers/linux/virtio_fs.hhw/virtio/vhost-user-fs.c new file +17
-0@@ -0,0 +1,297 @@ +/* +* Vhost-user filesystem virtio device +* +* Copyright 2018 Red Hat, Inc. +* +* Authors: +* Stefan Hajnoczi <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or +* (at your option) any later version. See the COPYING file in the +* top-level directory. +*/ + +#include "qemu/osdep.h" +#include <sys/ioctl.h> +#include "standard-headers/linux/virtio_fs.h" +#include "qapi/error.h"
Quoted 8 lines
Quoted 5 lines
hw/virtio/vhost-user-fs.c line 22
Quoted 5 lines
hw/virtio/vhost-user-fs.c new file +5
-0+#include "hw/virtio/virtio-bus.h" +#include "hw/virtio/virtio-access.h" +#include "qemu/error-report.h" +#include "hw/virtio/vhost-user-fs.h" +#include "monitor/monitor.h" JFYI, this needs to include
hw/qdev-properties.has well on the latest code level. (As does the pci part.)
Thanks! Updated my local version.
Dave
Quoted 190 lines
Quoted 42 lines
Quoted 42 lines
hw/virtio/vhost-user-fs.c new file +42
-0+ +static void vuf_get_config (VirtIODevice *vdev,uint8_t *config) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +struct virtio_fs_config fscfg = {}; + +memcpy ((char *)fscfg.tag, fs->conf.tag, +MIN (strlen (fs->conf.tag) +1 ,sizeof (fscfg.tag))); + +virtio_stl_p (vdev, &fscfg.num_queues, fs->conf.num_queues); + +memcpy (config, &fscfg,sizeof (fscfg)); +} + +static void vuf_start (VirtIODevice *vdev) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +BusState *qbus =BUS (qdev_get_parent_bus (DEVICE (vdev))); +VirtioBusClass *k =VIRTIO_BUS_GET_CLASS (qbus); +int ret; +int i; + +if (!k->set_guest_notifiers) { +error_report ("binding does not support guest notifiers" ); +return ; + } + + ret =vhost_dev_enable_notifiers (&fs->vhost_dev, vdev); +if (ret <0 ) { +error_report ("Error enabling host notifiers: %d" , -ret); +return ; + } + + ret = k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, true); +if (ret <0 ) { +error_report ("Error binding guest notifier: %d" , -ret); + goto err_host_notifiers; + } + + fs->vhost_dev.acked_features = vdev->guest_features; + ret =vhost_dev_start (&fs->vhost_dev, vdev); +if (ret <0 ) {Quoted 8 lines
hw/virtio/vhost-user-fs.c line 72
Quoted 8 lines
hw/virtio/vhost-user-fs.c new file +8
-0+error_report ("Error starting vhost: %d" , -ret); + goto err_guest_notifiers; + } + +/* +* guest_notifier_mask/pending not used yet, so just unmask +* everything here. virtio-pci will do the right thing by +* enabling/disabling irqfd. Referring to 'virtio-pci' seems a bit suspicious :) Should that be 'the transport'? (And 'the right thing' is not really self-explanatory...)
(I have wired it up for virtio-ccw, but have not actually tried it out. Will send it out once I did.)
Quoted 12 lines
Quoted 12 lines
hw/virtio/vhost-user-fs.c new file +12
-0+*/ +for (i =0 ; i < fs->vhost_dev.nvqs; i++) { +vhost_virtqueue_mask (&fs->vhost_dev, vdev, i, false); + } + +return ; + +err_guest_notifiers: + k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, false); +err_host_notifiers: +vhost_dev_disable_notifiers (&fs->vhost_dev, vdev); +}(...)
Quoted 30 lines
Quoted 30 lines
hw/virtio/vhost-user-fs.c new file +30
-0+static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + } + if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" );Quoted 8 lines
hw/virtio/vhost-user-fs.c
Quoted 2 lines
hw/virtio/vhost-user-fs.c new file +2
-0+return ; + }The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.I think the spec states that
num_queuesin the config space is the number of request queues only. Can we rename tonum_request_queues? The hiprio queue is not really configurable, anyway.Quoted 10 lines
hw/virtio/vhost-user-fs.c line 192
Quoted 6 lines
hw/virtio/vhost-user-fs.c new file +5
-0+ +if (!is_power_of_2 (fs->conf.queue_size)) { +error_setg (errp,"queue-size property must be a power of 2" ); +return ; + }Hmm packed ring allows non power of 2 ... We need to look into a generic helper to support VQ size checks.
Huh, I didn't notice before that there are several devices which already allow to configure the queue size... looking, there seem to be the following cases:
- bound checks and checks for power of 2 (blk, net)
- no checks (scsi) -- isn't that dangerous, asvirtio_add_queue()willabort()for a too large value?Anyway, if we have a non power of 2 size and the driver does not negotiate packed, we can just fail setting
FEATURES_OK, so dropping the power of 2 check should be fine, at least when we add packed support.Quoted 24 lines
Quoted 11 lines
hw/virtio/vhost-user-fs.c new file +9
-0+ +if (fs->conf.queue_size >VIRTQUEUE_MAX_SIZE ) { +error_setg (errp,"queue-size property must be %u or smaller" , +VIRTQUEUE_MAX_SIZE ); +return ; + } + +if (!vhost_user_init (&fs->vhost_user, &fs->conf.chardev, errp)) { +return ;hw/virtio/vhost-user-fs.c line 211
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +10
-0+ } + +virtio_init (vdev,"vhost-user-fs" ,VIRTIO_ID_FS , +sizeof (struct virtio_fs_config )); + +/* Notifications queue */ +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output); + +/* Hiprio queue */ +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output);Weird, spec patch v6 says:
Quoted 4 lines
+\item[0] hiprio
+\item[1\ldots n] request queueswhere's the Notifications queue coming from?
Maybe an old name of the hiprio queue?
Quoted 8 lines
hw/virtio/vhost-user-fs.c line 219
Quoted 8 lines
hw/virtio/vhost-user-fs.c new file +7
-0+/* Request queues */ +for (i =0 ; i < fs->conf.num_queues; i++) { +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output); + } + +/* 1 high prio queue, plus the number configured */ + fs->vhost_dev.nvqs =1 + fs->conf.num_queues;Anyway, the notifications queue needs to go, or this is wrong :)
Quoted 10 lines
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +10
-0+ fs->vhost_dev.vqs = g_new0(struct vhost_virtqueue, fs->vhost_dev.nvqs); + ret =vhost_dev_init (&fs->vhost_dev, &fs->vhost_user, +VHOST_BACKEND_TYPE_USER ,0 ); +if (ret <0 ) { +error_setg_errno (errp, -ret,"vhost_dev_init failed" ); + goto err_virtio; + } + +return ; +
Quoted 8 lines
Quoted 6 lines
Quoted 6 lines
hw/virtio/vhost-user-fs.c new file +6
-0+err_virtio: +vhost_user_cleanup (&fs->vhost_user); +virtio_cleanup (vdev); +g_free (fs->vhost_dev.vqs); +return ; +}(...)
--
Dr. David Alan Gilbert / [email protected] / Manchester, UK
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
In reply to Michael S. TsirkinQuoted 7 lines · * Michael S. Tsirkin ([email protected]) wrote:
Review
Quoted 5 lines · Dr. David Alan Gilbert (git)
From: "Dr. David Alan Gilbert" <[email protected]>
(quoted diff) +2
-0+ /* Hiprio queue */ + virtio_add_queue(vdev, fs->conf.queue_size, vuf_handle_output);
Quoted 6 lines
Weird, spec patch v6 says:
+\item[0] hiprio
+\item[1\ldots n] request queueswhere's the Notifications queue coming from?
Oops, that's a left over from when we used to have a notification queue; all the other parts of it are gone.
Dave
Quoted 194 lines
Quoted 194 lines
hw/virtio/vhost-user-fs.c new file +85
-0+/* Request queues */ +for (i =0 ; i < fs->conf.num_queues; i++) { +virtio_add_queue (vdev, fs->conf.queue_size, vuf_handle_output); + } + +/* 1 high prio queue, plus the number configured */ + fs->vhost_dev.nvqs =1 + fs->conf.num_queues; + fs->vhost_dev.vqs = g_new0(struct vhost_virtqueue, fs->vhost_dev.nvqs); + ret =vhost_dev_init (&fs->vhost_dev, &fs->vhost_user, +VHOST_BACKEND_TYPE_USER ,0 ); +if (ret <0 ) { +error_setg_errno (errp, -ret,"vhost_dev_init failed" ); + goto err_virtio; + } + +return ; + +err_virtio: +vhost_user_cleanup (&fs->vhost_user); +virtio_cleanup (vdev); +g_free (fs->vhost_dev.vqs); +return ; +} + +static void vuf_device_unrealize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); + +/* This will stop vhost backend if appropriate. */ +vuf_set_status (vdev,0 ); + +vhost_dev_cleanup (&fs->vhost_dev); + +vhost_user_cleanup (&fs->vhost_user); + +virtio_cleanup (vdev); +g_free (fs->vhost_dev.vqs); + fs->vhost_dev.vqs =NULL ; +} + +static const VMStateDescription vuf_vmstate = { + .name ="vhost-user-fs" , + .unmigratable =1 , +}; + +static Property vuf_properties[] = { +DEFINE_PROP_CHR ("chardev" , VHostUserFS, conf.chardev), +DEFINE_PROP_STRING ("tag" , VHostUserFS, conf.tag), +DEFINE_PROP_UINT16 ("num-queues" , VHostUserFS, conf.num_queues,1 ), +DEFINE_PROP_UINT16 ("queue-size" , VHostUserFS, conf.queue_size,128 ), +DEFINE_PROP_STRING ("vhostfd" , VHostUserFS, conf.vhostfd), +DEFINE_PROP_END_OF_LIST (), +}; + +static void vuf_class_init (ObjectClass *klass,void *data) +{ +DeviceClass *dc =DEVICE_CLASS (klass); +VirtioDeviceClass *vdc =VIRTIO_DEVICE_CLASS (klass); + + dc->props = vuf_properties; + dc->vmsd = &vuf_vmstate; +set_bit (DEVICE_CATEGORY_STORAGE , dc->categories); + vdc->realize = vuf_device_realize; + vdc->unrealize = vuf_device_unrealize; + vdc->get_features = vuf_get_features; + vdc->get_config = vuf_get_config; + vdc->set_status = vuf_set_status; + vdc->guest_notifier_mask = vuf_guest_notifier_mask; + vdc->guest_notifier_pending = vuf_guest_notifier_pending; +} + +static const TypeInfo vuf_info = { + .name =TYPE_VHOST_USER_FS , + .parent =TYPE_VIRTIO_DEVICE , + .instance_size =sizeof (VHostUserFS), + .class_init = vuf_class_init, +}; + +static void vuf_register_types (void ) +{ +type_register_static (&vuf_info); +} + +type_init (vuf_register_types)include/hw/virtio/vhost-user-fs.h new file +45
-0@@ -0,0 +1,45 @@ +/* +* Vhost-user filesystem virtio device +* +* Copyright 2018 Red Hat, Inc. +* +* Authors: +* Stefan Hajnoczi <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or +* (at your option) any later version. See the COPYING file in the +* top-level directory. +*/ + +#ifndef _QEMU_VHOST_USER_FS_H +#define _QEMU_VHOST_USER_FS_H + +#include "hw/virtio/virtio.h" +#include "hw/virtio/vhost.h" +#include "hw/virtio/vhost-user.h" +#include "chardev/char-fe.h" + +#define TYPE_VHOST_USER_FS "x-vhost-user-fs-device" +#define VHOST_USER_FS (obj) \ + OBJECT_CHECK(VHostUserFS, (obj), TYPE_VHOST_USER_FS) + +typedef struct { +CharBackend chardev; +char *tag; +uint16_t num_queues; +uint16_t queue_size; +char *vhostfd; +}VHostUserFSConf ; + +typedef struct { +/*< private >*/ +VirtIODevice parent; +VHostUserFSConf conf; +struct vhost_virtqueue *vhost_vqs; +struct vhost_dev vhost_dev; +VhostUserState vhost_user; + +/*< public >*/ +}VHostUserFS ; + +#endif /* _QEMU_VHOST_USER_FS_H */ include/standard-headers/linux/virtio_fs.h new file +41
-0@@ -0,0 +1,41 @@ +#ifndef _LINUX_VIRTIO_FS_H +#define _LINUX_VIRTIO_FS_H +/* This header is BSD licensed so anyone can use the definitions to implement +* compatible drivers/servers. +* +* Redistribution and use in source and binary forms, with or without +* modification, are permitted provided that the following conditions +* are met: +* 1. Redistributions of source code must retain the above copyright +* notice, this list of conditions and the following disclaimer. +* 2. Redistributions in binary form must reproduce the above copyright +* notice, this list of conditions and the following disclaimer in the +* documentation and/or other materials provided with the distribution. +* 3. Neither the name of IBM nor the names of its contributors +* may be used to endorse or promote products derived from this software +* without specific prior written permission. +* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS ``AS IS'' AND +* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE +* ARE DISCLAIMED. IN NO EVENT SHALL IBM OR CONTRIBUTORS BE LIABLE +* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS +* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) +* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT +* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY +* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF +* SUCH DAMAGE. */ +#include "standard-headers/linux/types.h" +#include "standard-headers/linux/virtio_ids.h" +#include "standard-headers/linux/virtio_config.h" +#include "standard-headers/linux/virtio_types.h" + +struct virtio_fs_config { +/* Filesystem name (UTF-8, not NUL-terminated, padded with NULs) */ +uint8_t tag[36 ]; + +/* Number of request queues */ +uint32_t num_queues; +}QEMU_PACKED ; + +#endif /* _LINUX_VIRTIO_FS_H */ include/standard-headers/linux/virtio_ids.h +1
-0@@ -43,6 +43,7 @@#define VIRTIO_ID_INPUT 18/* virtio input */ #define VIRTIO_ID_VSOCK 19/* virtio vsock transport */ #define VIRTIO_ID_CRYPTO 20/* virtio crypto */ +#define VIRTIO_ID_FS 26/* virtio filesystem */ #define VIRTIO_ID_PMEM 27/* virtio pmem */
Quoted 3 lines
Quoted 3 lines
include/standard-headers/linux/virtio_ids.h +0
-0#endif /* _LINUX_VIRTIO_IDS_H */ Signature · 1 line
2.21.0
--
Dr. David Alan Gilbert / [email protected] / Manchester, UK
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
In reply to Michael S. TsirkinQuoted 227 lines · * Michael S. Tsirkin ([email protected]) wrote:
Quoted 226 lines · Dr. David Alan Gilbert (git)
From: "Dr. David Alan Gilbert" <[email protected]>
The virtio-fs virtio device provides shared file system access using the FUSE protocol carried ovew virtio.
The actual file server is implemented in an external vhost-user-fs device backend process.Signed-off-by Stefan Hajnoczi <[email protected]>Signed-off-by Sebastien Boeuf <[email protected]>Signed-off-by Dr. David Alan Gilbert <[email protected]>--- configure | 13 + hw/virtio/Makefile.objs | 1 + hw/virtio/vhost-user-fs.c | 297 ++++++++++++++++++++ include/hw/virtio/vhost-user-fs.h | 45 +++ include/standard-headers/linux/virtio_fs.h | 41 +++ include/standard-headers/linux/virtio_ids.h | 1 + 6 files changed, 398 insertions(+) create mode 100644 hw/virtio/vhost-user-fs.c create mode 100644 include/hw/virtio/vhost-user-fs.h create mode 100644 include/standard-headers/linux/virtio_fs.hconfigure +13
-0@@ -382,6 +382,7 @@ vhost_crypto="" vhost_scsi="" vhost_vsock="" vhost_user="" +vhost_user_fs="" kvm="no" hax="no" hvf="no"@@ -1316,6 +1317,10 @@ for opt do ;; --enable-vhost-vsock) vhost_vsock="yes" ;; + --disable-vhost-user-fs) vhost_user_fs="no" + ;; + --enable-vhost-user-fs) vhost_user_fs="yes" + ;; --disable-opengl) opengl="no" ;; --enable-opengl) opengl="yes"@@ -2269,6 +2274,10 @@ test "$vhost_crypto" = "" && vhost_crypto=$vhost_user if test "$vhost_crypto" = "yes" && test "$vhost_user" = "no"; then error_exit "--enable-vhost-crypto requires --enable-vhost-user" fi +test "$vhost_user_fs" = "" && vhost_user_fs=$vhost_user +if test "$vhost_user_fs" = "yes" && test "$vhost_user" = "no"; then + error_exit "--enable-vhost-user-fs requires --enable-vhost-user" +fi # OR the vhost-kernel and vhost-user values for simplicity if test "$vhost_net" = ""; then@@ -6425,6 +6434,7 @@ echo "vhost-crypto support $vhost_crypto" echo "vhost-scsi support $vhost_scsi" echo "vhost-vsock support $vhost_vsock" echo "vhost-user support $vhost_user" +echo "vhost-user-fs support $vhost_user_fs" echo "Trace backends $trace_backends" if have_backend "simple"; then echo "Trace output file $trace_file-<pid>"@@ -6921,6 +6931,9 @@ fi if test "$vhost_user" = "yes" ; then echo "CONFIG_VHOST_USER=y" >> $config_host_mak fi +if test "$vhost_user_fs" = "yes" ; then + echo "CONFIG_VHOST_USER_FS=y" >> $config_host_mak +fi if test "$blobs" = "yes" ; then echo "INSTALL_BLOBS=yes" >> $config_host_mak fihw/virtio/Makefile.objs +1
-0@@ -11,6 +11,7 @@ common-obj-$(CONFIG_VIRTIO_PCI) += virtio-pci.o common-obj-$(CONFIG_VIRTIO_MMIO) += virtio-mmio.o obj-$(CONFIG_VIRTIO_BALLOON) += virtio-balloon.o obj-$(CONFIG_VIRTIO_CRYPTO) += virtio-crypto.o +obj-$(CONFIG_VHOST_USER_FS) += vhost-user-fs.o obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-pci.o obj-$(CONFIG_VIRTIO_PMEM) += virtio-pmem.o common-obj-$(call land,$(CONFIG_VIRTIO_PMEM),$(CONFIG_VIRTIO_PCI)) += virtio-pmem-pci.ohw/virtio/vhost-user-fs.c new file +134
-0@@ -0,0 +1,297 @@ +/* +* Vhost-user filesystem virtio device +* +* Copyright 2018 Red Hat, Inc. +* +* Authors: +* Stefan Hajnoczi <[email protected]> +* +* This work is licensed under the terms of the GNU GPL, version 2 or +* (at your option) any later version. See the COPYING file in the +* top-level directory. +*/ + +#include "qemu/osdep.h" +#include <sys/ioctl.h> +#include "standard-headers/linux/virtio_fs.h" +#include "qapi/error.h" +#include "hw/virtio/virtio-bus.h" +#include "hw/virtio/virtio-access.h" +#include "qemu/error-report.h" +#include "hw/virtio/vhost-user-fs.h" +#include "monitor/monitor.h" + +static void vuf_get_config (VirtIODevice *vdev,uint8_t *config) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +struct virtio_fs_config fscfg = {}; + +memcpy ((char *)fscfg.tag, fs->conf.tag, +MIN (strlen (fs->conf.tag) +1 ,sizeof (fscfg.tag))); + +virtio_stl_p (vdev, &fscfg.num_queues, fs->conf.num_queues); + +memcpy (config, &fscfg,sizeof (fscfg)); +} + +static void vuf_start (VirtIODevice *vdev) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +BusState *qbus =BUS (qdev_get_parent_bus (DEVICE (vdev))); +VirtioBusClass *k =VIRTIO_BUS_GET_CLASS (qbus); +int ret; +int i; + +if (!k->set_guest_notifiers) { +error_report ("binding does not support guest notifiers" ); +return ; + } + + ret =vhost_dev_enable_notifiers (&fs->vhost_dev, vdev); +if (ret <0 ) { +error_report ("Error enabling host notifiers: %d" , -ret); +return ; + } + + ret = k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, true); +if (ret <0 ) { +error_report ("Error binding guest notifier: %d" , -ret); + goto err_host_notifiers; + } + + fs->vhost_dev.acked_features = vdev->guest_features; + ret =vhost_dev_start (&fs->vhost_dev, vdev); +if (ret <0 ) { +error_report ("Error starting vhost: %d" , -ret); + goto err_guest_notifiers; + } + +/* +* guest_notifier_mask/pending not used yet, so just unmask +* everything here. virtio-pci will do the right thing by +* enabling/disabling irqfd. +*/ +for (i =0 ; i < fs->vhost_dev.nvqs; i++) { +vhost_virtqueue_mask (&fs->vhost_dev, vdev, i, false); + } + +return ; + +err_guest_notifiers: + k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, false); +err_host_notifiers: +vhost_dev_disable_notifiers (&fs->vhost_dev, vdev); +} + +static void vuf_stop (VirtIODevice *vdev) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +BusState *qbus =BUS (qdev_get_parent_bus (DEVICE (vdev))); +VirtioBusClass *k =VIRTIO_BUS_GET_CLASS (qbus); +int ret; + +if (!k->set_guest_notifiers) { +return ; + } + +vhost_dev_stop (&fs->vhost_dev, vdev); + + ret = k->set_guest_notifiers (qbus->parent, fs->vhost_dev.nvqs, false); +if (ret <0 ) { +error_report ("vhost guest notifier cleanup failed: %d" , ret); +return ; + } + +vhost_dev_disable_notifiers (&fs->vhost_dev, vdev); +} + +static void vuf_set_status (VirtIODevice *vdev,uint8_t status) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); +bool should_start = status &VIRTIO_CONFIG_S_DRIVER_OK ; + +if (!vdev->vm_running) { + should_start = false; + } + +if (fs->vhost_dev.started == should_start) { +return ; + } + +if (should_start) { +vuf_start (vdev); + }else { +vuf_stop (vdev); + } +} + +static uint64_t vuf_get_features (VirtIODevice *vdev, +uint64_t requested_features, +Error **errp) +{ +/* No feature bits used yet */ +return requested_features; +}
Quoted 8 lines
hw/virtio/vhost-user-fs.c line 138
Quoted 4 lines
hw/virtio/vhost-user-fs.c new file +4
-0+ +static void vuf_handle_output (VirtIODevice *vdev,VirtQueue *vq) +{ +/* Do nothing */ Why is this safe? Is this because this never triggers? assert(0) then? If it triggers then backend won't be notified, which might cause it to get stuck.
We never process these queues in qemu - always in the guest; so am I correct in thinking those shouldn't be used?
Quoted 47 lines
Quoted 47 lines
hw/virtio/vhost-user-fs.c new file +47
-0+} + +static void vuf_guest_notifier_mask (VirtIODevice *vdev,int idx, +bool mask) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +vhost_virtqueue_mask (&fs->vhost_dev, vdev, idx, mask); +} + +static bool vuf_guest_notifier_pending (VirtIODevice *vdev,int idx) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +return vhost_virtqueue_pending (&fs->vhost_dev, idx); +} + +static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + } + if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" );
Quoted 8 lines
hw/virtio/vhost-user-fs.c line 187
Quoted 2 lines
hw/virtio/vhost-user-fs.c new file +2
-0+return ; + }The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.
Should we be doing nvectors the same way virtio-scsi-pci does it; with a magic 'unspecified' default where it sets the nvectors based on the number of queues?
I think my preference is not to show the users the mandatory queues.
Quoted 9 lines
hw/virtio/vhost-user-fs.c line 192
Quoted 5 lines
hw/virtio/vhost-user-fs.c new file +5
-0+ +if (!is_power_of_2 (fs->conf.queue_size)) { +error_setg (errp,"queue-size property must be a power of 2" ); +return ; + }Hmm packed ring allows non power of 2 ... We need to look into a generic helper to support VQ size checks.
Which would also have to include the negotiation of where it's doing packaged ring?
<snip>
Dave
--
Dr. David Alan Gilbert / [email protected] / Manchester, UK
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
Quoted 10 lines · Dr. David Alan Gilbert
Quoted 8 lines · * Michael S. Tsirkin ([email protected]) wrote:
hw/virtio/vhost-user-fs.c line 138
Quoted 3 lines · Dr. David Alan Gilbert (git)
hw/virtio/vhost-user-fs.c new file +3
-0+static void vuf_handle_output (VirtIODevice *vdev,VirtQueue *vq) +{ +/* Do nothing */ Why is this safe? Is this because this never triggers? assert(0) then? If it triggers then backend won't be notified, which might cause it to get stuck.
Quoted 2 lines
We never process these queues in qemu - always in the guest; so am I correct in thinking those shouldn't be used?
s/guest/vhost-user backend process/
vuf_handle_output() should never be called.
Quoted 56 lines
Quoted 47 lines
Quoted 47 lines
hw/virtio/vhost-user-fs.c new file +47
-0+} + +static void vuf_guest_notifier_mask (VirtIODevice *vdev,int idx, +bool mask) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +vhost_virtqueue_mask (&fs->vhost_dev, vdev, idx, mask); +} + +static bool vuf_guest_notifier_pending (VirtIODevice *vdev,int idx) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +return vhost_virtqueue_pending (&fs->vhost_dev, idx); +} + +static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + } + if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" );Quoted 8 lines
Review
Quoted 2 lines
(quoted diff) +2
-0+ return; + }The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.
Quoted 5 lines
Should we be doing nvectors the same way virtio-scsi-pci does it; with a magic 'unspecified' default where it sets the nvectors based on the number of queues?
I think my preference is not to show the users the mandatory queues.
I agree. Users want to control multiqueue, not on the absolute number of virtqueues including mandatory queues.
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines
hw/virtio/vhost-user-fs.c new file +5
-0+ +if (!is_power_of_2 (fs->conf.queue_size)) { +error_setg (errp,"queue-size property must be a power of 2" ); +return ; + }
Quoted 6 lines
Quoted 3 lines
Hmm packed ring allows non power of 2 ... We need to look into a generic helper to support VQ size checks.
Which would also have to include the negotiation of where it's doing packaged ring?
It's impossible to perform this check at .realize() time since the packed virtqueue layout is negotiated via a VIRTIO feature bit. This puts us in the awkward position of either failing when the guest has already booted or rounding up the queue size for split ring layouts (with a warning message?).
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
Quoted 43 lines · On Thu, 22 Aug 2019 09:52:37 +0100 Stefan Hajnoczi <[email protected]> wrote:
Quoted 42 lines · Dr. David Alan Gilbert
Quoted 40 lines · * Michael S. Tsirkin ([email protected]) wrote:
Quoted 22 lines · Dr. David Alan Gilbert (git)
hw/virtio/vhost-user-fs.c new file +22
-0+static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + }hw/virtio/vhost-user-fs.c line 187
Quoted 10 lines
hw/virtio/vhost-user-fs.c new file +10
-0+ if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" ); +return ; + }The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.
Quoted 8 lines
Quoted 5 lines
Should we be doing nvectors the same way virtio-scsi-pci does it; with a magic 'unspecified' default where it sets the nvectors based on the number of queues?
I think my preference is not to show the users the mandatory queues.
I agree. Users want to control multiqueue, not on the absolute number of virtqueues including mandatory queues.
I agree as well, but let me advocate again for renaming this to 'num_request_queues' or similar to make it more obvious what this number actually means.
Quoted 11 lines
Quoted 11 lines
Quoted 10 lines
hw/virtio/vhost-user-fs.c line 192
Quoted 6 lines
hw/virtio/vhost-user-fs.c new file +5
-0+ +if (!is_power_of_2 (fs->conf.queue_size)) { +error_setg (errp,"queue-size property must be a power of 2" ); +return ; + }Hmm packed ring allows non power of 2 ... We need to look into a generic helper to support VQ size checks.
Quoted 8 lines
Quoted 2 lines
Which would also have to include the negotiation of where it's doing packaged ring?
It's impossible to perform this check at .realize() time since the packed virtqueue layout is negotiated via a VIRTIO feature bit. This puts us in the awkward position of either failing when the guest has already booted or rounding up the queue size for split ring layouts (with a warning message?).
I fear that is always going to be awkward if you allow to specify the queue size via a property. Basically, you can do two things: fail to accept FEATURES_OK if the queue size is not a power of 2 and the guest did not negotiate packed ring, or disallow to set a non power of 2 value here, which is what the other devices with such a property currently do (see also my other mail.) Would probably be good if all devices used the same approach (when we introduced packed ring support.)
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
Quoted 50 lines · Cornelia Huck
Quoted 47 lines · On Thu, 22 Aug 2019 09:52:37 +0100 Stefan Hajnoczi <[email protected]> wrote:
Quoted 46 lines · Dr. David Alan Gilbert
Quoted 32 lines · * Michael S. Tsirkin ([email protected]) wrote:
Quoted 30 lines · Dr. David Alan Gilbert (git)
hw/virtio/vhost-user-fs.c new file +30
-0+static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + } + if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" );Quoted 8 lines
Review
Quoted 2 lines
(quoted diff) +2
-0+ return; + }The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.Should we be doing nvectors the same way virtio-scsi-pci does it; with a magic 'unspecified' default where it sets the nvectors based on the number of queues?
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
I think my preference is not to show the users the mandatory queues.
I agree. Users want to control multiqueue, not on the absolute number of virtqueues including mandatory queues.
I agree as well, but let me advocate again for renaming this to '
num_request_queues' or similar to make it more obvious what this number actually means.
Good idea.
Re: [Qemu-devel] [PATCH 1/2] virtio: add vhost-user-fs base device
In reply to Stefan HajnocziQuoted 11 lines · * Stefan Hajnoczi ([email protected]) wrote:
Quoted 10 lines · Dr. David Alan Gilbert
Quoted 8 lines · * Michael S. Tsirkin ([email protected]) wrote:
hw/virtio/vhost-user-fs.c line 138
Quoted 3 lines · Dr. David Alan Gilbert (git)
hw/virtio/vhost-user-fs.c new file +3
-0+static void vuf_handle_output (VirtIODevice *vdev,VirtQueue *vq) +{ +/* Do nothing */ Why is this safe? Is this because this never triggers? assert(0) then? If it triggers then backend won't be notified, which might cause it to get stuck.
Quoted 6 lines
Quoted 2 lines
We never process these queues in qemu - always in the guest; so am I correct in thinking those shouldn't be used?
s/guest/vhost-user backend process/
vuf_handle_output()should never be called.
It turns out it does get called in one case during cleanup, in the case where the daemon died before qemu, virtio_bus_cleanup_host_notifier goes around the notifiers and calls all the ones where there's anything left in the eventfd.
Dave
Quoted 75 lines
Quoted 56 lines
Quoted 47 lines
Quoted 47 lines
hw/virtio/vhost-user-fs.c new file +47
-0+} + +static void vuf_guest_notifier_mask (VirtIODevice *vdev,int idx, +bool mask) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +vhost_virtqueue_mask (&fs->vhost_dev, vdev, idx, mask); +} + +static bool vuf_guest_notifier_pending (VirtIODevice *vdev,int idx) +{ +VHostUserFS *fs =VHOST_USER_FS (vdev); + +return vhost_virtqueue_pending (&fs->vhost_dev, idx); +} + +static void vuf_device_realize (DeviceState *dev,Error **errp) +{ +VirtIODevice *vdev =VIRTIO_DEVICE (dev); +VHostUserFS *fs =VHOST_USER_FS (dev); +unsigned int i; +size_t len; +int ret; + +if (!fs->conf.chardev.chr) { +error_setg (errp,"missing chardev" ); +return ; + } + +if (!fs->conf.tag) { +error_setg (errp,"missing tag property" ); +return ; + } + len =strlen (fs->conf.tag); +if (len ==0 ) { +error_setg (errp,"tag property cannot be empty" ); +return ; + } + if (len > sizeof_field(struct virtio_fs_config, tag)) { +error_setg (errp,"tag property must be %zu bytes or less" , + sizeof_field(struct virtio_fs_config, tag)); +return ; + } + +if (fs->conf.num_queues ==0 ) { +error_setg (errp,"num-queues property must be larger than 0" );Quoted 8 lines
Review
Quoted 2 lines
(quoted diff) +2
-0+ return; + }The strange thing is that actual # of queues is this number + 2. And this affects an optimal number of vectors (see patch 2). Not sure what a good solution is - include the mandatory queues in the number?
Needs to be documented in some way.Quoted 5 lines
Should we be doing nvectors the same way virtio-scsi-pci does it; with a magic 'unspecified' default where it sets the nvectors based on the number of queues?
I think my preference is not to show the users the mandatory queues.
I agree. Users want to control multiqueue, not on the absolute number of virtqueues including mandatory queues.
Quoted 10 lines
Quoted 9 lines
hw/virtio/vhost-user-fs.c line 192
Quoted 5 lines
hw/virtio/vhost-user-fs.c new file +5
-0+ +if (!is_power_of_2 (fs->conf.queue_size)) { +error_setg (errp,"queue-size property must be a power of 2" ); +return ; + }Hmm packed ring allows non power of 2 ... We need to look into a generic helper to support VQ size checks.
Quoted 8 lines
Quoted 2 lines
Which would also have to include the negotiation of where it's doing packaged ring?
It's impossible to perform this check at .realize() time since the packed virtqueue layout is negotiated via a VIRTIO feature bit. This puts us in the awkward position of either failing when the guest has already booted or rounding up the queue size for split ring layouts (with a warning message?).
--
Dr. David Alan Gilbert / [email protected] / Manchester, UK
PATCH 2/2 virtio: add vhost-user-fs-pci device
In reply to Dr. David Alan Gilbert (git)From: "Dr. David Alan Gilbert" <[email protected]>
Add the PCI version of vhost-user-fs.
Launch QEMU like this:
qemu -chardev socket,path=/tmp/vhost-fs.sock,id=chr0
-device x-vhost-user-fs-pci,tag=myfs,chardev=chr0
--- hw/virtio/Makefile.objs | 1 + hw/virtio/vhost-user-fs-pci.c | 79 +++++++++++++++++++++++++++++++++++ 2 files changed, 80 insertions(+) create mode 100644 hw/virtio/vhost-user-fs-pci.c
hw/virtio/Makefile.objs +1 -0
@@ -15,6 +15,7 @@ obj-$(CONFIG_VHOST_USER_FS) += vhost-user-fs.o obj-$(call land,$(CONFIG_VIRTIO_CRYPTO),$(CONFIG_VIRTIO_PCI)) += virtio-crypto-pci.o obj-$(CONFIG_VIRTIO_PMEM) += virtio-pmem.o common-obj-$(call land,$(CONFIG_VIRTIO_PMEM),$(CONFIG_VIRTIO_PCI)) += virtio-pmem-pci.o +obj-$(call land,$(CONFIG_VHOST_USER_FS),$(CONFIG_VIRTIO_PCI)) += vhost-user-fs-pci.o obj-$(CONFIG_VHOST_VSOCK) += vhost-vsock.o ifeq ($(CONFIG_VIRTIO_PCI),y)
hw/virtio/vhost-user-fs-pci.c new file +79 -0
79 lines not on this page · raw message
Signature · 1 line
2.21.0
Patchew URL: https://patchew.org/QEMU/[email protected]/
Patchew URL: https://patchew.org/QEMU/[email protected]/
Hi,
This series failed build test on s390x host. Please find the details below.
=== TEST SCRIPT BEGIN ===
#!/bin/bash
# Testing script will be invoked under the git checkout with # HEAD pointing to a commit that has the patches applied on top of "base"
# branch set -e
echo
echo "=== ENV ==="
env
echo
echo "=== PACKAGES ==="
rpm -qa
echo
echo "=== UNAME ==="
uname -a
CC=$HOME/bin/cc
INSTALL=$PWD/install
BUILD=$PWD/build
mkdir -p $BUILD $INSTALL
SRC=$PWD
cd $BUILD
$SRC/configure --cc=$CC --prefix=$INSTALL
make -j4
# XXX: we need reliable clean up
# make check -j4 V=1 make install === TEST SCRIPT END ===
The full log is available at
http://patchew.org/logs/[email protected]/testing.s390x/?type=message.
---
Email generated automatically by Patchew [https://patchew.org/].
Please send your feedback to [email protected]