Enable vhost-user to be used on BSD systems
Subsystems (2)
v3 cover Enable vhost-user to be used on BSD systems
Since QEMU is already able to emulate ioeventfd using pipefd, we're already pretty close to supporting vhost-user on non-Linux systems.
This two patches bridge the gap by:
1. Adding a new event_notifier_get_wfd() to return wfd on the places where the peer is expected to write to the notifier.
2. Modifying the build system to it allows enabling vhost-user on BSD.
v1->v2:
- Drop: "Allow returning EventNotifier's wfd" (Alex Williamson)
- Add: "event_notifier: add event_notifier_get_wfd()" (Alex Williamson)
- Add: "vhost: use wfd on functions setting vring call fd"
- Rename: "Allow building vhost-user in BSD" to "configure, meson: allow enabling vhost-user on all POSIX systems"
- Instead of making possible enabling vhost-user on Linux and BSD systems, allow enabling it on all non-Windows platforms. (Paolo Bonzini)
v2->v3:
- Add a section to docs/interop/vhost-user.rst explaining how vhost-user is supported on non-Linux platforms. (Stefan Hajnoczi)
Sergio Lopez (4):
event_notifier: add event_notifier_get_wfd()
vhost: use wfd on functions setting vring call fd configure, meson: allow enabling vhost-user on all POSIX systems
docs: vhost-user: add subsection for non-Linux platforms
configure | 4 ++-- docs/interop/vhost-user.rst | 18 ++++++++++++++++++ hw/virtio/vhost.c | 6 +++--- include/qemu/event_notifier.h | 1 + meson.build | 2 +- util/event_notifier-posix.c | 5 +++++ 6 files changed, 30 insertions(+), 6 deletions(-)
Signature · 1 line
2.35.1
PATCH v3 1/4 event_notifier: add event_notifier_get_wfd()
event_notifier_get_fd(const EventNotifier *e) always returns EventNotifier's read file descriptor (rfd). This is not a problem when the EventNotifier is backed by a an eventfd, as a single file descriptor is used both for reading and triggering events (rfd == wfd).
But, when EventNotifier is backed by a pipe pair, we have two file descriptors, one that can only be used for reads (rfd), and the other only for writes (wfd).
There's, at least, one known situation in which we need to obtain wfd instead of rfd, which is when setting up the file that's going to be sent to the peer in vhost's SET_VRING_CALL.
Add a new event_notifier_get_wfd(const EventNotifier *e) that can be used to obtain wfd where needed.
--- include/qemu/event_notifier.h | 1 + util/event_notifier-posix.c | 5 +++++ 2 files changed, 6 insertions(+)
include/qemu/event_notifier.h +1 -0
@@ -38,6 +38,7 @@ int event_notifier_test_and_clear(EventNotifier *);#ifdef CONFIG_POSIX void event_notifier_init_fd (EventNotifier *,int fd);int event_notifier_get_fd (const EventNotifier *); +int event_notifier_get_wfd (const EventNotifier *);#else HANDLE event_notifier_get_handle (EventNotifier *);#endif
util/event_notifier-posix.c +5 -0
@@ -99,6 +99,11 @@ int event_notifier_get_fd(const EventNotifier *e)return e->rfd; } +int event_notifier_get_wfd (const EventNotifier *e) +{ +return e->wfd; +} +int event_notifier_set (EventNotifier *e) {static const uint64_t value =1 ;
Signature · 1 line
2.35.1
PATCH v3 2/4 vhost: use wfd on functions setting vring call fd
In reply to Sergio LopezWhen ioeventfd is emulated using qemu_pipe(), only EventNotifier's wfd can be used for writing.
Use the recently introduced event_notifier_get_wfd() function to obtain the fd that our peer must use to signal the vring.
--- hw/virtio/vhost.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-)
hw/virtio/vhost.c +3 -3
@@ -1287,7 +1287,7 @@ static int vhost_virtqueue_init(struct vhost_dev *dev,return r; } - file.fd =event_notifier_get_fd (&vq->masked_notifier); + file.fd =event_notifier_get_wfd (&vq->masked_notifier); r = dev->vhost_ops->vhost_set_vring_call (dev, &file);if (r) {VHOST_OPS_DEBUG (r,"vhost_set_vring_call failed" );
@@ -1542,9 +1542,9 @@ void vhost_virtqueue_mask(struct vhost_dev *hdev, VirtIODevice *vdev, int n,if (mask) {assert (vdev->use_guest_notifier_mask); - file.fd =event_notifier_get_fd (&hdev->vqs[index].masked_notifier); + file.fd =event_notifier_get_wfd (&hdev->vqs[index].masked_notifier); }else { - file.fd =event_notifier_get_fd (virtio_queue_get_guest_notifier (vvq)); + file.fd =event_notifier_get_wfd (virtio_queue_get_guest_notifier (vvq)); } file.index = hdev->vhost_ops->vhost_get_vq_index (hdev, n);
Signature · 1 line
2.35.1
PATCH v3 3/4 configure, meson: allow enabling vhost-user on all POSIX systems
In reply to Sergio LopezWith the possibility of using a pipe pair via qemu_pipe() as a replacement on operating systems that doesn't support eventfd, vhost-user can also work on all POSIX systems.
This change allows enabling vhost-user on all non-Windows platforms and makes libvhost_user (which still depends on eventfd) a linux-only feature.
--- configure | 4 ++-- meson.build | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-)
configure +2 -2
@@ -1659,8 +1659,8 @@ fi # vhost interdependencies and host support # vhost backends -if test "$vhost_user" = "yes" && test "$linux" != "yes"; then - error_exit "vhost-user is only available on Linux" +if test "$vhost_user" = "yes" && test "$mingw32" = "yes"; then + error_exit "vhost-user is not available on Windows" fi test "$vhost_vdpa" = "" && vhost_vdpa=$linux if test "$vhost_vdpa" = "yes" && test "$linux" != "yes"; then
meson.build +1 -1
@@ -2701,7 +2701,7 @@ if have_system or have_user endif vhost_user = not_found -if 'CONFIG_VHOST_USER' in config_host +if targetos == 'linux' and 'CONFIG_VHOST_USER' in config_host libvhost_user = subproject('libvhost-user') vhost_user = libvhost_user.get_variable('vhost_user_dep') endif
Signature · 1 line
2.35.1
PATCH v3 4/4 docs: vhost-user: add subsection for non-Linux platforms
In reply to Sergio LopezAdd a section explaining how vhost-user is supported on platforms other than Linux.
--- docs/interop/vhost-user.rst | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)
docs/interop/vhost-user.rst +18 -0
@@ -38,6 +38,24 @@ conventions <backend_conventions>`. *Master* and *slave* can be either a client (i.e. connecting) or server (listening) in the socket communication. +Support for platforms other than Linux +-------------------------------------- + +While vhost-user was initially developed targeting Linux, nowadays is +supported on any platform that provides the following features: + +- The ability to share a mapping injected into the guest between + multiple processes, so both QEMU and the vhost-user daemon servicing + the device can access simultaneously the memory regions containing + the virtqueues and the data associated with each request. + +- AF_UNIX sockets with SCM_RIGHTS, so QEMU can communicate with the + vhost-user daemon and send it file descriptors when needed. + +- Either eventfd or pipe/pipe2. On platforms where eventfd is not + available, QEMU will automatically fallback to pipe2 or, as a last + resort, pipe. + Message Specification =====================
Signature · 1 line
2.35.1
Re: [PATCH v3 4/4] docs: vhost-user: add subsection for non-Linux platforms
Quoted 13 lines · Sergio Lopez
Add a section explaining how vhost-user is supported on platforms other than Linux.
Signed-off-by Sergio Lopez <[email protected]>--- docs/interop/vhost-user.rst | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)docs/interop/vhost-user.rst +0
-0
docs/interop/vhost-user.rst conventions <backend_conventions>`. · line 44
Quoted 7 lines
docs/interop/vhost-user.rst +4
-0@@ -38,6 +38,24 @@ conventions <backend_conventions>`. *Master* and *slave* can be either a client (i.e. connecting) or server (listening) in the socket communication. +Support for platforms other than Linux +-------------------------------------- + +While vhost-user was initially developed targeting Linux, nowadays is
s/is/it is/
docs/interop/vhost-user.rst conventions <backend_conventions>`. · line 50
Quoted 6 lines
docs/interop/vhost-user.rst +6
-0+supported on any platform that provides the following features: + +- The ability to share a mapping injected into the guest between + multiple processes, so both QEMU and the vhost-user daemon servicing + the device can access simultaneously the memory regions containing + the virtqueues and the data associated with each request.
Please generalize this statement since there are other vhost-user protocol features aside from guest RAM access that involve shared memory:
1. VHOST_USER_SET_LOG_BASE
2. VHOST_USER_SET_INFLIGHT_FD
The exact requirement is:
The vhost-user protocol relies on shared memory represented by a file descriptor so it can be passed over a UNIX domain socket and then mapped by the other process.
docs/interop/vhost-user.rst conventions <backend_conventions>`. · line 56
Quoted 6 lines
docs/interop/vhost-user.rst +6
-0+ +- AF_UNIX sockets with SCM_RIGHTS, so QEMU can communicate with the + vhost-user daemon and send it file descriptors when needed. + +- Either eventfd or pipe/pipe2. On platforms where eventfd is not + available, QEMU will automatically fallback to pipe2 or, as a last
The noun is "fallback", the verb form is "fall back":
s/fallback/fall back/
It's worth mentioning that events are sent over pipe fds by writing an 8-byte value. The 8-byte value has no meaning and should not be interpreted.
Re: [PATCH v3 4/4] docs: vhost-user: add subsection for non-Linux platforms
In reply to Sergio LopezQuoted 13 lines · Sergio Lopez
Add a section explaining how vhost-user is supported on platforms other than Linux.
Signed-off-by Sergio Lopez <[email protected]>--- docs/interop/vhost-user.rst | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)docs/interop/vhost-user.rst +0
-0
docs/interop/vhost-user.rst conventions <backend_conventions>`. · line 41
Quoted 4 lines
docs/interop/vhost-user.rst +1
-0@@ -38,6 +38,24 @@ conventions <backend_conventions>`. *Master* and *slave* can be either a client (i.e. connecting) or server (listening) in the socket communication. +Support for platforms other than Linux
It's not just Linux - any platform without eventfd.
So I think we should have a section explaining that whereever spec says eventfd it can be a pipe if system does not support creating eventfd.
Quoted 22 lines
docs/interop/vhost-user.rst +17
-0+-------------------------------------- + +While vhost-user was initially developed targeting Linux, nowadays is +supported on any platform that provides the following features: + +- The ability to share a mapping injected into the guest between + multiple processes, so both QEMU and the vhost-user daemon servicing + the device can access simultaneously the memory regions containing + the virtqueues and the data associated with each request. + +- AF_UNIX sockets with SCM_RIGHTS, so QEMU can communicate with the + vhost-user daemon and send it file descriptors when needed. + +- Either eventfd or pipe/pipe2. On platforms where eventfd is not + available, QEMU will automatically fallback to pipe2 or, as a last + resort, pipe. + Message Specification =====================Signature · 1 line
2.35.1
Re: [PATCH v3 4/4] docs: vhost-user: add subsection for non-Linux platforms
Quoted 17 lines · Michael S. Tsirkin
Quoted 16 lines · Sergio Lopez
Add a section explaining how vhost-user is supported on platforms other than Linux.
Signed-off-by Sergio Lopez <[email protected]>--- docs/interop/vhost-user.rst | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+)docs/interop/vhost-user.rst +0
-0@@ -38,6 +38,24 @@ conventions <backend_conventions>`. *Master* and *slave* can be either a client (i.e. connecting) or server (listening) in the socket communication.
Quoted 8 lines
docs/interop/vhost-user.rst conventions <backend_conventions>`. · line 41
Quoted 1 line
docs/interop/vhost-user.rst +1
-0+Support for platforms other than LinuxIt's not just Linux - any platform without eventfd.
So I think we should have a section explaining that whereever spec says eventfd it can be a pipe if system does not support creating eventfd.
I'm confused. This is exactly what this subsection intends to do...
Thanks,
Sergio.
Quoted 23 lines
Quoted 22 lines
docs/interop/vhost-user.rst +17
-0+-------------------------------------- + +While vhost-user was initially developed targeting Linux, nowadays is +supported on any platform that provides the following features: + +- The ability to share a mapping injected into the guest between + multiple processes, so both QEMU and the vhost-user daemon servicing + the device can access simultaneously the memory regions containing + the virtqueues and the data associated with each request. + +- AF_UNIX sockets with SCM_RIGHTS, so QEMU can communicate with the + vhost-user daemon and send it file descriptors when needed. + +- Either eventfd or pipe/pipe2. On platforms where eventfd is not + available, QEMU will automatically fallback to pipe2 or, as a last + resort, pipe. + Message Specification =====================Signature · 1 line
2.35.1
Quoted 29 lines · Sergio Lopez
Since QEMU is already able to emulate ioeventfd using pipefd, we're already pretty close to supporting vhost-user on non-Linux systems.
This two patches bridge the gap by:
1. Adding a new
event_notifier_get_wfd()to return wfd on the places where the peer is expected to write to the notifier.2. Modifying the build system to it allows enabling vhost-user on BSD.
v1->v2:
- Drop: "Allow returning EventNotifier's wfd" (Alex Williamson)
- Add: "event_notifier: addevent_notifier_get_wfd()" (Alex Williamson)
- Add: "vhost: use wfd on functions setting vring call fd"
- Rename: "Allow building vhost-user in BSD" to "configure, meson: allow enabling vhost-user on all POSIX systems"
- Instead of making possible enabling vhost-user on Linux and BSD systems, allow enabling it on all non-Windows platforms. (Paolo Bonzini)
v2->v3:
- Add a section todocs/interop/vhost-user.rstexplaining how vhost-user is supported on non-Linux platforms. (Stefan Hajnoczi)Sergio Lopez (4):
event_notifier: addevent_notifier_get_wfd()
vhost: use wfd on functions setting vring call fd configure, meson: allow enabling vhost-user on all POSIX systems
docs: vhost-user: add subsection for non-Linux platforms
Quoted 12 lines
configure | 4 ++-- docs/interop/vhost-user.rst | 18 ++++++++++++++++++ hw/virtio/vhost.c | 6 +++--- include/qemu/event_notifier.h | 1 + meson.build | 2 +- util/event_notifier-posix.c | 5 +++++ 6 files changed, 30 insertions(+), 6 deletions(-)Signature · 1 line
2.35.1
I posted comments on the vhost-user.rst patch. Otherwise:
Quoted 10 lines · Sergio Lopez
Since QEMU is already able to emulate ioeventfd using pipefd, we're already pretty close to supporting vhost-user on non-Linux systems.
This two patches bridge the gap by:
1. Adding a new
event_notifier_get_wfd()to return wfd on the places where the peer is expected to write to the notifier.2. Modifying the build system to it allows enabling vhost-user on BSD.
Quoted 8 lines
v1->v2:
- Drop: "Allow returning EventNotifier's wfd" (Alex Williamson)
- Add: "event_notifier: addevent_notifier_get_wfd()" (Alex Williamson)
- Add: "vhost: use wfd on functions setting vring call fd"
- Rename: "Allow building vhost-user in BSD" to "configure, meson: allow enabling vhost-user on all POSIX systems"
- Instead of making possible enabling vhost-user on Linux and BSD systems, allow enabling it on all non-Windows platforms. (Paolo Bonzini)
I picked 1,2.
Waiting on updated doc patch to apply 3,4.
Quoted 21 lines
v2->v3:
- Add a section todocs/interop/vhost-user.rstexplaining how vhost-user is supported on non-Linux platforms. (Stefan Hajnoczi)Sergio Lopez (4):
event_notifier: addevent_notifier_get_wfd()
vhost: use wfd on functions setting vring call fd configure, meson: allow enabling vhost-user on all POSIX systems
docs: vhost-user: add subsection for non-Linux platformsconfigure | 4 ++-- docs/interop/vhost-user.rst | 18 ++++++++++++++++++ hw/virtio/vhost.c | 6 +++--- include/qemu/event_notifier.h | 1 + meson.build | 2 +- util/event_notifier-posix.c | 5 +++++ 6 files changed, 30 insertions(+), 6 deletions(-)Signature · 1 line
2.35.1