From: "Michael S. Tsirkin" <mst@redhat.com>
To: Laszlo Ersek <lersek@redhat.com>
Cc: qemu-devel@nongnu.org, Peter Maydell <peter.maydell@linaro.org>,
Eugenio Perez Martin <eperezma@redhat.com>,
German Maglione <gmaglione@redhat.com>,
Liu Jiang <gerry@linux.alibaba.com>,
Sergio Lopez Pascual <slp@redhat.com>,
Stefano Garzarella <sgarzare@redhat.com>
Subject: Re: [PULL 30/63] vhost-user: call VHOST_USER_SET_VRING_ENABLE synchronously
Date: Wed, 4 Oct 2023 08:53:14 -0400 [thread overview]
Message-ID: <20231004085242-mutt-send-email-mst@kernel.org> (raw)
In-Reply-To: <d7c0aa1f-8158-1688-9ce3-d6b5b04d8ae5@redhat.com>
On Wed, Oct 04, 2023 at 12:11:44PM +0200, Laszlo Ersek wrote:
> On 10/4/23 10:44, Michael S. Tsirkin wrote:
> > From: Laszlo Ersek <lersek@redhat.com>
> >
> > (1) The virtio-1.2 specification
> > <http://docs.oasis-open.org/virtio/virtio/v1.2/virtio-v1.2.html> writes:
> >
> >> 3 General Initialization And Device Operation
> >> 3.1 Device Initialization
> >> 3.1.1 Driver Requirements: Device Initialization
> >>
> >> [...]
> >>
> >> 7. Perform device-specific setup, including discovery of virtqueues for
> >> the device, optional per-bus setup, reading and possibly writing the
> >> device’s virtio configuration space, and population of virtqueues.
> >>
> >> 8. Set the DRIVER_OK status bit. At this point the device is “live”.
> >
> > and
> >
> >> 4 Virtio Transport Options
> >> 4.1 Virtio Over PCI Bus
> >> 4.1.4 Virtio Structure PCI Capabilities
> >> 4.1.4.3 Common configuration structure layout
> >> 4.1.4.3.2 Driver Requirements: Common configuration structure layout
> >>
> >> [...]
> >>
> >> The driver MUST configure the other virtqueue fields before enabling the
> >> virtqueue with queue_enable.
> >>
> >> [...]
> >
> > (The same statements are present in virtio-1.0 identically, at
> > <http://docs.oasis-open.org/virtio/virtio/v1.0/virtio-v1.0.html>.)
> >
> > These together mean that the following sub-sequence of steps is valid for
> > a virtio-1.0 guest driver:
> >
> > (1.1) set "queue_enable" for the needed queues as the final part of device
> > initialization step (7),
> >
> > (1.2) set DRIVER_OK in step (8),
> >
> > (1.3) immediately start sending virtio requests to the device.
> >
> > (2) When vhost-user is enabled, and the VHOST_USER_F_PROTOCOL_FEATURES
> > special virtio feature is negotiated, then virtio rings start in disabled
> > state, according to
> > <https://qemu-project.gitlab.io/qemu/interop/vhost-user.html#ring-states>.
> > In this case, explicit VHOST_USER_SET_VRING_ENABLE messages are needed for
> > enabling vrings.
> >
> > Therefore setting "queue_enable" from the guest (1.1) is a *control plane*
> > operation, which travels from the guest through QEMU to the vhost-user
> > backend, using a unix domain socket.
> >
> > Whereas sending a virtio request (1.3) is a *data plane* operation, which
> > evades QEMU -- it travels from guest to the vhost-user backend via
> > eventfd.
> >
> > This means that steps (1.1) and (1.3) travel through different channels,
> > and their relative order can be reversed, as perceived by the vhost-user
> > backend.
> >
> > That's exactly what happens when OVMF's virtiofs driver (VirtioFsDxe) runs
> > against the Rust-language virtiofsd version 1.7.2. (Which uses version
> > 0.10.1 of the vhost-user-backend crate, and version 0.8.1 of the vhost
> > crate.)
> >
> > Namely, when VirtioFsDxe binds a virtiofs device, it goes through the
> > device initialization steps (i.e., control plane operations), and
> > immediately sends a FUSE_INIT request too (i.e., performs a data plane
> > operation). In the Rust-language virtiofsd, this creates a race between
> > two components that run *concurrently*, i.e., in different threads or
> > processes:
> >
> > - Control plane, handling vhost-user protocol messages:
> >
> > The "VhostUserSlaveReqHandlerMut::set_vring_enable" method
> > [crates/vhost-user-backend/src/handler.rs] handles
> > VHOST_USER_SET_VRING_ENABLE messages, and updates each vring's "enabled"
> > flag according to the message processed.
> >
> > - Data plane, handling virtio / FUSE requests:
> >
> > The "VringEpollHandler::handle_event" method
> > [crates/vhost-user-backend/src/event_loop.rs] handles the incoming
> > virtio / FUSE request, consuming the virtio kick at the same time. If
> > the vring's "enabled" flag is set, the virtio / FUSE request is
> > processed genuinely. If the vring's "enabled" flag is clear, then the
> > virtio / FUSE request is discarded.
> >
> > Note that OVMF enables the queue *first*, and sends FUSE_INIT *second*.
> > However, if the data plane processor in virtiofsd wins the race, then it
> > sees the FUSE_INIT *before* the control plane processor took notice of
> > VHOST_USER_SET_VRING_ENABLE and green-lit the queue for the data plane
> > processor. Therefore the latter drops FUSE_INIT on the floor, and goes
> > back to waiting for further virtio / FUSE requests with epoll_wait.
> > Meanwhile OVMF is stuck waiting for the FUSET_INIT response -- a deadlock.
> >
> > The deadlock is not deterministic. OVMF hangs infrequently during first
> > boot. However, OVMF hangs almost certainly during reboots from the UEFI
> > shell.
> >
> > The race can be "reliably masked" by inserting a very small delay -- a
> > single debug message -- at the top of "VringEpollHandler::handle_event",
> > i.e., just before the data plane processor checks the "enabled" field of
> > the vring. That delay suffices for the control plane processor to act upon
> > VHOST_USER_SET_VRING_ENABLE.
> >
> > We can deterministically prevent the race in QEMU, by blocking OVMF inside
> > step (1.1) -- i.e., in the write to the "queue_enable" register -- until
> > VHOST_USER_SET_VRING_ENABLE actually *completes*. That way OVMF's VCPU
> > cannot advance to the FUSE_INIT submission before virtiofsd's control
> > plane processor takes notice of the queue being enabled.
> >
> > Wait for VHOST_USER_SET_VRING_ENABLE completion by:
> >
> > - setting the NEED_REPLY flag on VHOST_USER_SET_VRING_ENABLE, and waiting
> > for the reply, if the VHOST_USER_PROTOCOL_F_REPLY_ACK vhost-user feature
> > has been negotiated, or
> >
> > - performing a separate VHOST_USER_GET_FEATURES *exchange*, which requires
> > a backend response regardless of VHOST_USER_PROTOCOL_F_REPLY_ACK.
> >
> > Cc: "Michael S. Tsirkin" <mst@redhat.com> (supporter:vhost)
> > Cc: Eugenio Perez Martin <eperezma@redhat.com>
> > Cc: German Maglione <gmaglione@redhat.com>
> > Cc: Liu Jiang <gerry@linux.alibaba.com>
> > Cc: Sergio Lopez Pascual <slp@redhat.com>
> > Cc: Stefano Garzarella <sgarzare@redhat.com>
> > Signed-off-by: Laszlo Ersek <lersek@redhat.com>
> > Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
> > Message-Id: <20230830134055.106812-8-lersek@redhat.com>
> > Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > Reviewed-by: Eugenio Pérez <eperezma@redhat.com>
> > ---
> > hw/virtio/vhost-user.c | 16 +++++++++++++++-
> > 1 file changed, 15 insertions(+), 1 deletion(-)
> >
> > diff --git a/hw/virtio/vhost-user.c b/hw/virtio/vhost-user.c
> > index ae0734d461..eb983ae295 100644
> > --- a/hw/virtio/vhost-user.c
> > +++ b/hw/virtio/vhost-user.c
> > @@ -1214,7 +1214,21 @@ static int vhost_user_set_vring_enable(struct vhost_dev *dev, int enable)
> > .num = enable,
> > };
> >
> > - ret = vhost_set_vring(dev, VHOST_USER_SET_VRING_ENABLE, &state, false);
> > + /*
> > + * SET_VRING_ENABLE travels from guest to QEMU to vhost-user backend /
> > + * control plane thread via unix domain socket. Virtio requests travel
> > + * from guest to vhost-user backend / data plane thread via eventfd.
> > + * Even if the guest enables the ring first, and pushes its first virtio
> > + * request second (conforming to the virtio spec), the data plane thread
> > + * in the backend may see the virtio request before the control plane
> > + * thread sees the queue enablement. This causes (in fact, requires) the
> > + * data plane thread to discard the virtio request (it arrived on a
> > + * seemingly disabled queue). To prevent this out-of-order delivery,
> > + * don't let the guest proceed to pushing the virtio request until the
> > + * backend control plane acknowledges enabling the queue -- IOW, pass
> > + * wait_for_reply=true below.
> > + */
> > + ret = vhost_set_vring(dev, VHOST_USER_SET_VRING_ENABLE, &state, true);
> > if (ret < 0) {
> > /*
> > * Restoring the previous state is likely infeasible, as well as
>
> This is not the latest version (v3) of this set -- please see
> <https://patchwork.ozlabs.org/project/qemu-devel/cover/20231002203221.17241-1-lersek@redhat.com/>.
>
> Thanks,
> Laszlo
Ouch. OK I will drop. Feel free to send v4 tweaking commit message - I
think you wanted to do it anyway right?
next prev parent reply other threads:[~2023-10-04 12:53 UTC|newest]
Thread overview: 99+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-04 8:43 [PULL 00/63] virtio,pci: features, cleanups Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 01/63] pci: SLT must be RO Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 02/63] hw/virtio: Propagate page_mask to vhost_vdpa_listener_skipped_section() Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 03/63] hw/virtio: Propagate page_mask to vhost_vdpa_section_end() Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 04/63] hw/virtio/vhost-vdpa: Inline TARGET_PAGE_ALIGN() macro Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 05/63] hw/virtio/vhost-vdpa: Use target-agnostic qemu_target_page_mask() Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 06/63] hw/virtio: Build vhost-vdpa.o once Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 07/63] hw/virtio/meson: Rename softmmu_virtio_ss[] -> system_virtio_ss[] Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 08/63] virtio: add vhost-user-base and a generic vhost-user-device Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 09/63] hw/virtio: add config support to vhost-user-device Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 10/63] virtio-net: do not reset vlan filtering at set_features Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 11/63] virtio-net: Expose MAX_VLAN Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 12/63] vdpa: Restore vlan filtering state Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 13/63] vdpa: Allow VIRTIO_NET_F_CTRL_VLAN in SVQ Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 14/63] virtio: don't zero out memory region cache for indirect descriptors Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 15/63] vdpa: use first queue SVQ state for CVQ default Michael S. Tsirkin
2023-10-04 8:43 ` [PULL 16/63] vdpa: export vhost_vdpa_set_vring_ready Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 17/63] vdpa: rename vhost_vdpa_net_load to vhost_vdpa_net_cvq_load Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 18/63] vdpa: move vhost_vdpa_set_vring_ready to the caller Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 19/63] vdpa: remove net cvq migration blocker Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 20/63] vhost: Add count argument to vhost_svq_poll() Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 21/63] qmp: remove virtio_list, search QOM tree instead Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 22/63] qmp: update virtio feature maps, vhost-user-gpio introspection Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 23/63] vhost-user: move VhostUserProtocolFeature definition to header file Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 24/63] vhost-user: strip superfluous whitespace Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 25/63] vhost-user: tighten "reply_supported" scope in "set_vring_addr" Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 26/63] vhost-user: factor out "vhost_user_write_sync" Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 27/63] vhost-user: flatten "enforce_reply" into "vhost_user_write_sync" Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 28/63] vhost-user: hoist "write_sync", "get_features", "get_u64" Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 29/63] vhost-user: allow "vhost_set_vring" to wait for a reply Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 30/63] vhost-user: call VHOST_USER_SET_VRING_ENABLE synchronously Michael S. Tsirkin
2023-10-04 10:11 ` Laszlo Ersek
2023-10-04 12:53 ` Michael S. Tsirkin [this message]
2023-10-04 13:28 ` Laszlo Ersek
2023-10-04 8:44 ` [PULL 31/63] hw/isa/ich9: Add comment on imperfect emulation of PIC vs. I/O APIC routing Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 32/63] tests/acpi: Allow update of DSDT.cxl Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 33/63] hw/cxl: Add QTG _DSM support for ACPI0017 device Michael S. Tsirkin
2023-10-04 17:46 ` Thomas Huth
2023-10-04 22:14 ` Michael S. Tsirkin
2023-10-04 23:09 ` [PATCH v3] " Dave Jiang
2023-10-05 3:36 ` Michael S. Tsirkin
2023-10-05 16:11 ` Dave Jiang
2023-10-05 16:32 ` Michael S. Tsirkin
2023-10-06 12:09 ` Jonathan Cameron via
2023-10-06 12:09 ` Jonathan Cameron
2023-10-06 17:50 ` Dan Williams
2023-10-06 22:15 ` [PATCH v4] " Dave Jiang
2023-10-09 15:47 ` Jonathan Cameron via
2023-10-09 15:47 ` Jonathan Cameron
2023-10-09 16:06 ` Dave Jiang
2023-10-09 15:44 ` [PATCH v3] " Jonathan Cameron via
2023-10-09 15:44 ` Jonathan Cameron
2023-10-07 21:17 ` Michael S. Tsirkin
2023-10-09 15:40 ` Jonathan Cameron via
2023-10-09 15:40 ` Jonathan Cameron
2023-10-05 17:00 ` Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 34/63] tests/acpi: Update DSDT.cxl with QTG DSM Michael S. Tsirkin
2023-10-04 8:44 ` [PULL 35/63] hw/i386/acpi-build: Use pc_madt_cpu_entry() directly Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 36/63] hw/acpi/cpu: Have build_cpus_aml() take a build_madt_cpu_fn callback Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 37/63] hw/acpi/acpi_dev_interface: Remove now unused madt_cpu virtual method Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 38/63] hw/acpi/acpi_dev_interface: Remove now unused #include "hw/boards.h" Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 39/63] hw/i386: Remove now redundant TYPE_ACPI_GED_X86 Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 40/63] hw/i386/acpi-build: Determine SMI command port just once Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 41/63] hw/acpi: Trace GPE access in all device models, not just PIIX4 Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 42/63] hw/acpi/core: Trace enable and status registers of GPE separately Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 43/63] vdpa: fix gcc cvq_isolated uninitialized variable warning Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 44/63] vdpa net: zero vhost_vdpa iova_tree pointer at cleanup Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 45/63] hw/cxl: Push cxl_decoder_count_enc() and cxl_decode_ig() into .c Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 46/63] hw/cxl: Add utility functions decoder interleave ways and target count Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 47/63] hw/cxl: Fix and use same calculation for HDM decoder block size everywhere Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 48/63] hw/cxl: Support 4 HDM decoders at all levels of topology Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 49/63] hw/pci-bridge/cxl-upstream: Add serial number extended capability support Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 50/63] vdpa net: fix error message setting virtio status Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 51/63] vdpa net: stop probing if cannot set features Michael S. Tsirkin
2023-10-04 8:45 ` [PULL 52/63] vdpa net: follow VirtIO initialization properly at cvq isolation probing Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 53/63] amd_iommu: Fix APIC address check Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 54/63] hw/i386/pc: improve physical address space bound check for 32-bit x86 systems Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 55/63] pcie_sriov: unregister_vfs(): fix error path Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 56/63] libvhost-user.c: add assertion to vu_message_read_default Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 57/63] virtio: use shadow_avail_idx while checking number of heads Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 58/63] virtio: remove unnecessary thread fence while reading next descriptor Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 59/63] virtio: remove unused next argument from virtqueue_split_read_next_desc() Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 60/63] util/uuid: add a hash function Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 61/63] hw/display: introduce virtio-dmabuf Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 62/63] vhost-user: add shared_object msg Michael S. Tsirkin
2023-10-04 8:46 ` [PULL 63/63] libvhost-user: handle " Michael S. Tsirkin
2023-10-04 8:54 ` [PULL 00/63] virtio,pci: features, cleanups Philippe Mathieu-Daudé
2023-10-04 9:08 ` Michael S. Tsirkin
2023-10-04 8:58 ` Michael S. Tsirkin
2023-10-04 16:26 ` Michael S. Tsirkin
2023-10-04 16:50 ` Stefan Hajnoczi
2023-10-04 17:04 ` Michael S. Tsirkin
2023-10-04 17:40 ` Thomas Huth
2023-10-04 22:23 ` Michael S. Tsirkin
2023-10-05 6:10 ` Thomas Huth
2023-10-04 18:23 ` Stefan Hajnoczi
2023-10-04 17:20 ` Michael S. Tsirkin
2023-10-04 18:29 ` Stefan Hajnoczi
2023-10-04 22:16 ` Michael S. Tsirkin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20231004085242-mutt-send-email-mst@kernel.org \
--to=mst@redhat.com \
--cc=eperezma@redhat.com \
--cc=gerry@linux.alibaba.com \
--cc=gmaglione@redhat.com \
--cc=lersek@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=sgarzare@redhat.com \
--cc=slp@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).