Add a new vhost-user message to give a unix socket to a vhost-user
backend for GPU display updates.
Back when I started that work, I added a new GPU channel because the
vhost-user protocol wasn't bidirectional. Since then, there is a
vhost-user-slave channel for the slave to send requests to the master.
We could extend it with GPU messages. However, the GPU protocol is
quite orthogonal to vhost-user, thus I chose to have a new dedicated
channel.
See vhost-user-gpu.rst for the protocol details.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-id: 20190524130946.31736-2-marcandre.lureau@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
That should fix the following warning:
/home/pm215/qemu/contrib/libvhost-user/libvhost-user.c: In function
‘vu_set_mem_table_exec_postcopy’:
/home/pm215/qemu/contrib/libvhost-user/libvhost-user.c:666:9: error:
format ‘%llx’ expects argument of type ‘long long unsigned int’, but
argument 5 has type ‘__u64’ [-Werror=format=]
DPRINT("%s: region %d: Registered userfault for %llx + %llx\n",
^
/home/pm215/qemu/contrib/libvhost-user/libvhost-user.c:666:9: error:
format ‘%llx’ expects argument of type ‘long long unsigned int’, but
argument 6 has type ‘__u64’ [-Werror=format=]
cc1: all warnings being treated as errors
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190514104126.6294-3-marcandre.lureau@redhat.com
{ kraxel: s/PRIu64/PRIx64/ ]
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
Mark dirty as page, the step of each call is 1.
Signed-off-by: Li Feng <fengli@smartx.com>
Message-Id: <20190420091016.213160-1-fengli@smartx.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add the config protocol feature bit if the set_config & get_config
callbacks are implemented.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-id: 20190503130034.24916-3-marcandre.lureau@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c: In function ‘vu_set_mem_table_exec_postcopy’:
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c:546:31: warning: taking address of packed member of ‘struct VhostUserMsg’ may result in an unaligned pointer value [-Waddress-of-packed-member]
546 | VhostUserMemory *memory = &vmsg->payload.memory;
| ^~~~~~~~~~~~~~~~~~~~~
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c: In function ‘vu_set_mem_table_exec’:
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c:688:31: warning: taking address of packed member of ‘struct VhostUserMsg’ may result in an unaligned pointer value [-Waddress-of-packed-member]
688 | VhostUserMemory *memory = &vmsg->payload.memory;
| ^~~~~~~~~~~~~~~~~~~~~
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c: In function ‘vu_set_vring_addr_exec’:
/home/elmarco/src/qemu/contrib/libvhost-user/libvhost-user.c:817:36: warning: taking address of packed member of ‘struct VhostUserMsg’ may result in an unaligned pointer value [-Waddress-of-packed-member]
817 | struct vhost_vring_addr *vra = &vmsg->payload.addr;
| ^~~~~~~~~~~~~~~~~~~
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-id: 20190503130034.24916-2-marcandre.lureau@redhat.com
Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
gcc_struct is for x86 only, and it generates an warning on ARM64 Clang/MinGW targets.
Signed-off-by: Cao Jiaxi <driver1998@foxmail.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Message-id: 20190503003618.10089-1-driver1998@foxmail.com
[PMM: dropped the slirp change as slirp is now a submodule]
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
This patch adds support for VHOST_USER_GET_INFLIGHT_FD and
VHOST_USER_SET_INFLIGHT_FD message to set/get shared buffer
to/from qemu. Then backend can track inflight I/O in this buffer.
Signed-off-by: Xie Yongji <xieyongji@baidu.com>
Signed-off-by: Zhang Yu <zhangyu31@baidu.com>
Message-Id: <20190228085355.9614-5-xieyongji@baidu.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Introduce vu_queue_map_desc() which should be
independent with vu_queue_pop();
Signed-off-by: Xie Yongji <xieyongji@baidu.com>
Signed-off-by: Zhang Yu <zhangyu31@baidu.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20190228085355.9614-4-xieyongji@baidu.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The vu_check_queue_msg_file() has checked the FD flag. So let's
delete the redundant check after it.
Signed-off-by: Xie Yongji <xieyongji@baidu.com>
Signed-off-by: Zhang Yu <zhangyu31@baidu.com>
Message-Id: <20190228085355.9614-3-xieyongji@baidu.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
vhost-user-input will make use of this function to undo some queue pop
in case the virtio queue does not have enough room.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20190308140454.32437-11-marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Simplify the creation of FD sources for other users. This is just
convenience to avoid duplicating similar code elsewhere.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20190308140454.32437-10-marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Since commit 2566378d6d, libvhost-user
no longer panics on disconnect (rc == 0), and instead silently ignores
an invalid VHOST_USER_NONE message.
Without extra work from the API user, this will simply busy-loop on
HUP events. The obvious thing to do is to exit(0) instead, while
additional or different work can be done by overriding
iface->process_msg().
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Jens Freimann <jfreimann@redhat.com>
Message-Id: <20190308140454.32437-5-marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Now that the VhostUserMsg.request field is used for both master &
slave requests, since commit d84599f56c820d8c1ac9928a76500dcdfbbf194d:
contrib/libvhost-user/libvhost-user.c:953:20: error: implicit conversion from enumeration type 'enum VhostUserSlaveRequest' to different enumeration type 'VhostUserRequest' (aka 'enum VhostUserRequest') [-Werror,-Wenum-conversion]
.request = VHOST_USER_SLAVE_VRING_HOST_NOTIFIER_MSG,
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20190308140454.32437-2-marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
contrib/libvhost-user is now using __u64 which is
a long long. New linux headers will use uint64_t.
These need to be printed differently: llx versus PRIx64.
For now cast to uint64_t for portability so the change
does not break the build.
Casts will go away in the future.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This patch introduces the host notifier support in
libvhost-user. A new API is added to support setting
host notifier for each queue.
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The response to a VHOST_USER_POSTCOPY_ADVISE contains a fd but doesn't
actually contain any data. FIx vu_message_write so that it doesn't
do a 0-byte write() call, since this was ending up with rc=0
that was confusing the error handling code.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Tell QEMU we understand the protocol features needed for postcopy.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This message is sent just before the end of postcopy to get the
client to stop using userfault since we wont respond to any more
requests. It should close userfaultfd so that any other pages
get mapped to the backing file automatically by the kernel, since
at this point we know we've received everything.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Clear the area and turn off THP.
PROT_NONE the area until after we've userfault advised it
to catch any unexpected changes.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
We need a better way, but at the moment we need the address of the
mappings sent back to qemu so it can interpret the messages on the
userfaultfd it reads.
This is done as a 3 stage set:
QEMU -> client
set_mem_table
mmap stuff, get addresses
client -> qemu
here are the addresses
qemu -> client
OK - now you can use them
That ensures that qemu has registered the new addresses in it's
userfault code before the client starts accessing them.
Note: We don't ask for the default 'ack' reply since we've got our own.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When new regions are sent to the client using SET_MEM_TABLE, register
them with the userfaultfd.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Split the set_mem_table routines in both qemu and libvhost-user
because the postcopy versions are going to be quite different
once changes in the later patches are added. However, this patch
doesn't produce any functional change, just the split.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Notify the vhost-user slave on reception of the 'postcopy-listen'
event from the source.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Open a userfaultfd (on a postcopy_advise) and send it back in
the reply to the qemu for it to monitor.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Allow replies with fds (for postcopy)
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Wire up a notifier to send a VHOST_USER_POSTCOPY_ADVISE
message on an incoming advise.
Later patches will fill in the behaviour/contents of the
message.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Add a vhost feature flag for postcopy support, and
use the postcopy notifier to check it before allowing postcopy.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
The sg list/indirect descriptor table may be contigious
in GPA but not in HVA address space. But libvhost-user
wasn't aware of that. This would cause out-of-bounds
access. Even a malicious guest could use it to get
information from the vhost-user backend.
Introduce a plen parameter in vu_gpa_to_va() so we can
handle this case, returning the actual mapped length.
Signed-off-by: Yongji Xie <xieyongji@baidu.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Free the mmaped memory when we need to mmap new memory
space on vu_set_mem_table_exec() and vu_set_log_base_exec() to
avoid memory leak.
Also close the corresponding fd after mmap() on
vu_set_log_base_exec() to avoid fd leak.
Signed-off-by: Yongji Xie <xieyongji@baidu.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Enable VHOST_USER_GET_CONFIG/VHOST_USER_SET_CONFIG messages in
libvhost-user library, users can implement their own I/O target
based on the library. This enable the virtio config space delivered
between QEMU host device and the I/O target.
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Allow the qemu to pass us a slave fd. We don't do anything
with it yet.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20171002191521.15748-5-dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Update the ProtocolFeature and UserRequest lists to
match hw/virtio/vhost-user.c.
Fix the text labelling in libvhost-user.c to match the list.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20171002191521.15748-4-dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Add a vu_queue_started method to complement vu_queue_enabled.
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Message-Id: <20171002191521.15748-2-dgilbert@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
This file implements a bridge from the vu_init API of libvhost-user to
GSource, so that libvhost-user can be used inside a GLib main loop.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
libvhost-user is meant to be free of glib dependency. Make sure it is
by droping qemu/osdep.h (which included glib.h)
This fixes a bad malloc()/g_free() pair.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
This is the same workaround as commit 523b018dde, which was lost
with libvhost-user transition in commit e10e798c85.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
End processing of messages when VHOST_USER_NONE
is received.
Without this we run into a vubr_panic() call and get
"PANIC: Unhandled request: 0"
Signed-off-by: Jens Freimann <jfreiman@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This commit introduces a vhost-user-scsi backend sample application. It
must be linked with libiscsi and libvhost-user.
To use it, compile with:
$ make vhost-user-scsi
And run as follows:
$ ./vhost-user-scsi -u vus.sock -i iscsi://uri_to_target/
$ qemu-system-x86_64 --enable-kvm -m 512 \
-object memory-backend-file,id=mem,size=512m,share=on,mem-path=guestmem \
-numa node,memdev=mem \
-chardev socket,id=vhost-user-scsi,path=vus.sock \
-device vhost-user-scsi-pci,chardev=vhost-user-scsi \
The application is currently limited at one LUN only and it processes
requests synchronously (therefore only achieving QD1). The purpose of
the code is to show how a backend can be implemented and to test the
vhost-user-scsi Qemu implementation.
If a different instance of this vhost-user-scsi application is executed
at a remote host, a VM can be live migrated to such a host.
Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Message-Id: <1488479153-21203-5-git-send-email-felipe@nutanix.com>
Calling libvhost-user functions like vu_queue_get_avail_bytes() when the
queue doesn't yet have addresses will result in the crashes like the
following:
Program received signal SIGSEGV, Segmentation fault.
0x000055c414112ce4 in vring_avail_idx (vq=0x55c41582fd68, vq=0x55c41582fd68)
at /home/dgilbert/git/qemu/contrib/libvhost-user/libvhost-user.c:940
940 vq->shadow_avail_idx = vq->vring.avail->idx;
(gdb) p vq
$1 = (VuVirtq *) 0x55c41582fd68
(gdb) p vq->vring
$2 = {num = 0, desc = 0x0, avail = 0x0, used = 0x0, log_guest_addr = 0, flags = 0}
at /home/dgilbert/git/qemu/contrib/libvhost-user/libvhost-user.c:940
No locals.
at /home/dgilbert/git/qemu/contrib/libvhost-user/libvhost-user.c:960
num_heads = <optimized out>
out_bytes=out_bytes@entry=0x7fffd035d7c4, max_in_bytes=max_in_bytes@entry=0,
max_out_bytes=max_out_bytes@entry=0) at /home/dgilbert/git/qemu/contrib/libvhost-user/libvhost-user.c:1034
Add a pre-condition checks on vring.avail before accessing it.
Fix documentation and return type of vu_queue_empty() while at it.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Tested-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
On gcc 3.4 and newer, simply using (void) in front of WUR functions is
not sufficient to ignore the return value. That prevents a build when
handling warnings as errors.
libvhost-user had a usage of (void)vasprintf() which triggered such a
condition. This fixes it by replacing this call with g_strdup_vprintf()
which aborts on OOM.
Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Message-Id: <1488479153-21203-2-git-send-email-felipe@nutanix.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Add a library to help implementing vhost-user backend (or slave).
Dealing with vhost-user as an application developer isn't so easy: you
have all the trouble with any protocol: validation, unix ancillary data,
shared memory, eventfd, logging, and on top of that you need to deal
with virtio queues, if possible efficiently.
qemu test has a nice vhost-user testing application vhost-user-bridge,
which implements most of vhost-user, and virtio.c which implements
virtqueues manipulation. Based on these two, I tried to make a simple
library, reusable for tests or development of new vhost-user scenarios.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
[Felipe: set used_idx copy on SET_VRING_ADDR and update shadow avail idx
on SET_VRING_BASE]
Signed-off-by: Felipe Franciosi <felipe@nutanix.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Marc-André Lureau