Commit Graph

32 Commits

Author SHA1 Message Date
Alexandre Raymond
9bf0960a9a Fix compilation warning due to missing header for sigaction (followup)
This patch removes all references to signal.h when qemu-common.h is included
as they become redundant.

Signed-off-by: Alexandre Raymond <cerbere@gmail.com>
Signed-off-by: Stefan Hajnoczi <stefanha@linux.vnet.ibm.com>
2011-06-08 09:04:29 +01:00
mst@redhat.com
5430a28fe4 vhost: force vhost off for non-MSI guests
When MSI is off, each interrupt needs to be bounced through the io
thread when it's set/cleared, so vhost-net causes more context switches and
higher CPU utilization than userspace virtio which handles networking in
the same thread.

We'll need to fix this by adding level irq support in kvm irqfd,
for now disable vhost-net in these configurations.

Added a vhostforce flag to force vhost-net back on.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2011-02-01 16:50:44 -06:00
Michael S. Tsirkin
27a6375de3 tap: make set_offload a nop after netdev cleanup
virtio-net expects set_offload to succeed after
peer cleanup.
Since we don't have an open fd anymore, make it so.
Fixes warning about the failure of offload setting.

Reported-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2010-11-16 08:40:07 -06:00
Michael S. Tsirkin
43849424cf tap: clear vhost_net backend on cleanup
Frontends calling tap_get_vhost_net get an invalid pointer after the
peer backend has been deleted. Jason Wang <jasowang@redhat.com> reports
this leading to a crash in ack_features when we remove the vhost-net
bakend of a virtio nic.

The fix is simply to clear the backend pointer.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2010-11-16 08:40:07 -06:00
Michael S. Tsirkin
445d892f43 tap: add APIs for vnet header length
Add APIs to control host header length. First user
will be vhost-net.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2010-09-07 20:27:33 +03:00
Michael S. Tsirkin
ef4252b149 tap: generalize code for different vnet header len
Make host vnet header length a structure field in
preparation for using this support in linux kernel.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
2010-09-07 20:27:28 +03:00
Blue Swirl
9678d9501b Remove dead assignments in various common files, spotted by clang analyzer
Value stored is never read.

Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2010-04-25 18:35:52 +00:00
Michael S. Tsirkin
b202554cf7 tap: add API to retrieve vhost net header
will be used by virtio-net for vhost net support

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-04-01 13:56:43 -05:00
Michael S. Tsirkin
82b0d80ef6 tap: add vhost/vhostfd options
This adds vhost binary option to tap, to enable vhost net accelerator.
Default is off for now, we'll be able to make default on long term
when we know it's stable.

vhostfd option can be used by management, to pass in the fd. Assigning
vhostfd implies vhost=on.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-04-01 13:56:43 -05:00
Michael S. Tsirkin
95d528a2fe tap: add interface to get device fd
Will be used by vhost to attach/detach to backend.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-04-01 13:56:43 -05:00
Anthony Liguori
31dc63dfa1 Revert "tap: invoke downscript when we exit abnormally"
This reverts commit 8af8ce4d61.
2010-03-21 14:13:34 -05:00
Anthony Liguori
8af8ce4d61 tap: invoke downscript when we exit abnormally
Right now, downscript is not invoked reliably.  If you execute 'quit' from the
monitor, it won't be invoked.

This fixes that by converting tap to use an exit_notifier to execute the
downscript.  In this case, allowing an exit notifier to include state is
critically important for the conversion.

Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-03-19 15:27:38 -05:00
Markus Armbruster
1ecda02b24 error: Replace qemu_error() by error_report()
error_report() terminates the message with a newline.  Strip it it
from its arguments.

This fixes a few error messages lacking a newline:
net_handle_fd_param()'s "No file descriptor named %s found", and
tap_open()'s "vnet_hdr=1 requested, but no kernel support for
IFF_VNET_HDR available" (all three versions).

There's one place that passes arguments without newlines
intentionally: load_vmstate().  Fix it up.
2010-03-16 16:58:32 +01:00
Markus Armbruster
2f7920166d error: Move qemu_error & friends into their own header 2010-03-16 16:55:05 +01:00
Markus Armbruster
62112d181c net: Fix bogus "Warning: vlan 0 with no nics" with -device
net_check_clients() prints this when an VLAN has host devices, but no
guest devices.  It uses VLANState members nb_guest_devs and
nb_host_devs to keep track of these devices.  However, -device does
not update nb_guest_devs, only net_init_nic() does that, for -net nic.

Check the VLAN clients directly, and remove the counters.

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-02-19 12:12:46 -06:00
Michael S. Tsirkin
ceb696159d net: add API to disable/enable polling
When vhost is bound to a backend device, we need to stop polling it when
vhost is started, and restart polling when vhost is stopped.
Add an API for that for use by vhost, and implement in tap backend.

Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2010-01-08 09:58:40 -06:00
Mark McLoughlin
df6c2a0fb2 net: initialize vnet_hdr in net_init_tap()
net_tap_init() always sets vnet_hdr using qemu_opt_get_bool(), but
initialize it in net_init_tap() just to reduce confusion.

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-12-03 09:41:35 -06:00
Mark McLoughlin
665a3b071d net: remove VLANClientState members now in NetClientInfo
Add a NetClientInfo pointer to VLANClientState and use that
for the typecode and function pointers.

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-12-03 09:41:34 -06:00
Mark McLoughlin
3e35ba93ec net: convert tap to NetClientInfo
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-12-03 09:41:29 -06:00
Juergen Lock
929fe49721 Avoid segfault on net_tap_init() failure
Check for fd == -1 there.

Signed-off-by: Juergen Lock <nox@jelal.kn-bremen.de>
Signed-off-by: Blue Swirl <blauwirbel@gmail.com>
2009-11-21 09:32:15 +00:00
Mark McLoughlin
5819c91806 tap: drain queue in tap_send()
Okay, let's try re-enabling the drain-entire-queue behaviour, with a
difference - before each subsequent packet, use qemu_can_send_packet()
to check that we can send it. This is similar to how we check before
polling the tap fd and avoids having to drop a packet if the receiver
cannot handle it.

This patch should be a performance improvement since we no longer have
to go through the mainloop for each packet.

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-11-09 08:43:02 -06:00
Mark McLoughlin
be1636b3ab tap: disable draining queue in one go
If qemu_send_packet_async() returns zero, it means the packet has been
queued and the sent callback will be invoked once it has been flushed.

This is only possible where the NIC's receive() handler returns zero
and promises to notify the networking core that room is available in its
queue again.

In the case where the receive handler does not have this capability
(and its queue fills up) it returns -1 and the networking core does not
queue up the packet. This condition is indicated by a -1 return from
qemu_send_packet_async().

Currently, tap handles this condition simply by dropping the packet. It
should do its best to avoid getting into this situation by checking such
NIC's have room for a packet before copying the packet from the tap
interface.

tap_send() used to achieve this by only reading a single packet before
returning to the mainloop. That way, tap_can_send() is called before
reading each packet.

tap_send() was changed to completely drain the tap interface queue
without taking into account the situation where the NIC returns an
error and the packet is not queued. Let's start fixing this by
reverting to the previous behaviour of reading one packet at a time.

Reported-by: Scott Tsai <scottt.tw@gmail.com>
Tested-by: Sven Rudolph <Sven_Rudolph@drewag.de>
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-11-09 08:43:02 -06:00
Alexander Graf
71f4effce7 Unbreak tap compilation on OS X
Currently compiling the tap sources breaks on Mac OS X. This is because of:

  1) tap-linux.h requiring Linux includes
  2) typos
  3) missing #includes

This patch adds what's necessary to compile tap happily on Mac OS X.
I haven't tested if using tap actually works, but I don't think that's a
major issue as that code was probably seriously untested before already.

I didn't split the patch, because it's only a few lines of code and
splitting is probably not worth the effort here.

Signed-off-by: Alexander Graf <agraf@suse.de>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 16:38:15 -05:00
Mark McLoughlin
9c282718aa net: move UFO support detection to tap-linux.c
Only supported on Linux

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:29 -05:00
Mark McLoughlin
1faac1f7d4 net: move tap_set_offload() code into tap-linux.c
TUNSETOFFLOAD is only available on Linux

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:28 -05:00
Mark McLoughlin
dc69004c7d net: move tap_probe_vnet_hdr() to tap-linux.c
Only Linux has support for IFF_VNET_HDR

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:28 -05:00
Mark McLoughlin
15ac913bfe net: move tap_set_sndbuf() to tap-linux.c
TUNSETSNDBUF is only available on linux

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:28 -05:00
Mark McLoughlin
c28b1c1007 net: move linux code into net/tap-linux.c
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:28 -05:00
Mark McLoughlin
88b3ec02b2 net: move AIX code into net/tap-aix.c
Okay, this makes the tap options available on AIX even though there's
no support, but if we want to do it right we should have not compile
the tap code at all on AIX using e.g. CONFIG_TAP.

Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:28 -05:00
Mark McLoughlin
966ea5ec4f net: move solaris code to net/tap-solaris.c
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:27 -05:00
Mark McLoughlin
e7e92325d9 net: split BSD tap_open() out into net/tap-bsd.c
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:27 -05:00
Mark McLoughlin
5281d757ef net: split all the tap code out into net/tap.c
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Anthony Liguori <aliguori@us.ibm.com>
2009-10-30 08:39:27 -05:00