linux.git/drivers/vhost, branch v4.20-rc2

Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost

2018-11-01T21:42:49+00:00

Pull virtio/vhost updates from Michael Tsirkin:
 "Fixes and tweaks:

   - virtio balloon page hinting support

   - vhost scsi control queue

   - misc fixes"

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  MAINTAINERS: remove reference to bogus vsock file
  vhost/scsi: Use common handling code in request queue handler
  vhost/scsi: Extract common handling code from control queue handler
  vhost/scsi: Respond to control queue operations
  vhost/scsi: truncate T10 PI iov_iter to prot_bytes
  virtio-balloon: VIRTIO_BALLOON_F_PAGE_POISON
  mm/page_poison: expose page_poisoning_enabled to kernel modules
  virtio-balloon: VIRTIO_BALLOON_F_FREE_PAGE_HINT
  kvm_config: add CONFIG_VIRTIO_MENU

vhost: Fix Spectre V1 vulnerability

2018-10-31T19:39:15+00:00

The idx in vhost_vring_ioctl() was controlled by userspace, hence a
potential exploitation of the Spectre variant 1 vulnerability.

Fixing this by sanitizing idx before using it to index d->vqs.

Cc: Michael S. Tsirkin 
Cc: Josh Poimboeuf 
Cc: Andrea Arcangeli 
Signed-off-by: Jason Wang 
Signed-off-by: David S. Miller

vhost/scsi: Use common handling code in request queue handler

2018-10-25T01:16:13+00:00

Change the request queue handler to use common handling routines same
as the control queue handler.

Signed-off-by: Bijan Mottahedeh 
Signed-off-by: Michael S. Tsirkin

vhost/scsi: Extract common handling code from control queue handler

2018-10-25T01:16:13+00:00

Prepare to change the request queue handler to use common handling
routines.

Signed-off-by: Bijan Mottahedeh 
Signed-off-by: Michael S. Tsirkin

vhost/scsi: Respond to control queue operations

2018-10-25T01:16:13+00:00

The vhost-scsi driver currently does not handle any control queue
operations. In particular, vhost_scsi_ctl_handle_kick, merely prints out
a debug message but does nothing else. This can cause guest VMs to hang.

As part of SCSI recovery from an error, e.g., an I/O timeout, the SCSI
midlayer attempts to abort the failed operation. The SCSI virtio driver
translates the abort to a SCSI TMF request that gets put on the control
queue (virtscsi_abort -> virtscsi_tmf). The SCSI virtio driver then
waits indefinitely for this request to be completed, but it never will
because vhost-scsi never responds to that request.

To avoid a hang, always respond to control queue operations; explicitly
reject TMF requests, and return a no-op response to event requests.

Signed-off-by: Bijan Mottahedeh 
Signed-off-by: Michael S. Tsirkin

vhost/scsi: truncate T10 PI iov_iter to prot_bytes

2018-10-25T01:16:13+00:00

Commands with protection information included were not truncating the
protection iov_iter to the number of protection bytes in the command.
This resulted in vhost_scsi mis-calculating the size of the protection
SGL in vhost_scsi_calc_sgls(), and including both the protection and
data SG entries in the protection SGL.

Fixes: 09b13fa8c1a1 ("vhost/scsi: Add ANY_LAYOUT support in vhost_scsi_handle_vq")
Signed-off-by: Greg Edwards 
Signed-off-by: Michael S. Tsirkin 
Fixes: 09b13fa8c1a1093e9458549ac8bb203a7c65c62a
Cc: stable@vger.kernel.org
Reviewed-by: Paolo Bonzini

net: vhost: remove bad code line

2018-10-08T04:31:32+00:00

Signed-off-by: Tonghao Zhang 
Signed-off-by: David S. Miller

net: vhost: add rx busy polling in tx path

2018-09-27T03:25:55+00:00

This patch improves the guest receive performance.
On the handle_tx side, we poll the sock receive queue at the
same time. handle_rx do that in the same way.

We set the poll-us=100us and use the netperf to test throughput
and mean latency. When running the tests, the vhost-net kthread
of that VM, is alway 100% CPU. The commands are shown as below.

Rx performance is greatly improved by this patch. There is not
notable performance change on tx with this series though. This
patch is useful for bi-directional traffic.

netperf -H IP -t TCP_STREAM -l 20 -- -O "THROUGHPUT, THROUGHPUT_UNITS, MEAN_LATENCY"

Topology:
[Host] ->linux bridge -> tap vhost-net ->[Guest]

TCP_STREAM:
* Without the patch:  19842.95 Mbps, 6.50 us mean latency
* With the patch:     37598.20 Mbps, 3.43 us mean latency

Signed-off-by: Tonghao Zhang 
Signed-off-by: David S. Miller

net: vhost: factor out busy polling logic to vhost_net_busy_poll()

2018-09-27T03:25:55+00:00

Factor out generic busy polling logic and will be
used for in tx path in the next patch. And with the patch,
qemu can set differently the busyloop_timeout for rx queue.

To avoid duplicate codes, introduce the helper functions:
* sock_has_rx_data(changed from sk_has_rx_data)
* vhost_net_busy_poll_try_queue

Signed-off-by: Tonghao Zhang 
Signed-off-by: David S. Miller

net: vhost: replace magic number of lock annotation

2018-09-27T03:25:55+00:00

Use the VHOST_NET_VQ_XXX as a subclass for mutex_lock_nested.

Signed-off-by: Tonghao Zhang 
Acked-by: Jason Wang 
Signed-off-by: David S. Miller