Version: GnuPG v1
iQEcBAABAgAGBQJXT9DWAAoJEO8Ells5jWIRgFAH/1ZDXm8V523AMDOEvBAWgqur
Dj8ZaIwFkqJp7xtLdhS0yKF3xW+vtgx9k+Qftk0S8qEiFKPbThR8iB5VNuesErwd
AZhWo4bnVhKwtWyMw3BDRDK1N4huAWPMZEva1xovR/Cc9v5IG5mx57/K3Zz5C8ec
Jsn4DsLKN0q7W0D0dlnbEOkSjl6iKJchvfPCR6UfvrU7BxfXaCZ9Z7Sfh8ec6tfr
iMgcV9u3A3Zs72gTM9/jdKx8vOrWtdKJufJ8s2Bctc7CyfBNWwnV8PjndhEe3Xvs
vlYeJopdpDPsdMkMtYD6cevtEgvD5yhOBndJ7et807jjuCvUf837tMhodKkFk9M=
=SjIZ
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/jasowang/tags/net-pull-request' into staging
# gpg: Signature made Thu 02 Jun 2016 07:23:18 BST using RSA key ID 398D6211
# gpg: Good signature from "Jason Wang (Jason Wang on RedHat) <jasowang@redhat.com>"
# gpg: WARNING: This key is not certified with sufficiently trusted signatures!
# gpg: It is not certain that the signature belongs to the owner.
# Primary key fingerprint: 215D 46F4 8246 689E C77F 3562 EF04 965B 398D 6211
* remotes/jasowang/tags/net-pull-request: (31 commits)
Add ENET device to i.MX6 SOC.
Add ENET/Gbps Ethernet support to FEC device
i.MX: move FEC device to a register array structure.
i.MX: Rename i.MX FEC defines to ENET_XXX
i.MX: reset TX/RX descriptors when FEC is disabled.
i.MX: Fix FEC code for ECR register reset value.
i.MX: Fix FEC code for MDIO address selection
i.MX: Fix FEC code for MDIO operation selection
net: handle optional VLAN header in checksum computation.
net: improve UDP/TCP checksum computation.
e1000e: Introduce qtest for e1000e device
net: Introduce e1000e device emulation
e1000: Move out code that will be reused in e1000e
e1000_regs: Add definitions for Intel 82574-specific bits
vmxnet3: Use pci_dma_* API instead of cpu_physical_memory_*
net_pkt: Extend packet abstraction as required by e1000e functionality
rtl8139: Move more TCP definitions to common header
net_pkt: Name vmxnet3 packet abstractions more generic
vmxnet3: Use common MAC address tracing macros
net: Add macros for MAC address tracing
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
The ENET device (present in i.MX6) is "derived" from FEC and backward
compatible with it.
This patch adds the necessary support of the added feature in the ENET
device to allow Linux to use it (on supported processors).
Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
This is to prepare for the ENET Gb device of the i.MX6.
Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
According to the FEC chapter of i.MX25 reference manual
RX adn TX descriptors are reseted when the FEC device is disabled through ECR.
Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
According to the FEC chapter of i.MX25 reference manual ECR register is
initialized at 0xf0000000 at reset time.
We fix the value.
Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
According to the FEC chapter of i.MX25 reference manual
When writing to MMFR register, the MDIO device and adress are selected by
bit 27 to 23 and bit 22 to 18 respectively. This is a total of 10 bits
that need to be used by the Phy chip/address decoding function.
This patch fixes the number of bits used from 9 to 10.
Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
According to the FEC chapter of i.MX25 reference manual
When writing the MMFR register, bit 29 and 28 select the requested operation.
* 10 means read operation with valid MII mgmt frame
* 11 means read operation with non compliant MII mgmt frame
* 01 means write operation with valid MII mgmt frame
* 00 means write operation with non compliant MII mgmt frame
So while bit 28 does change beween read/write for valid MII mgmt frame, the
mening is inverted for non compliant MII mgmt frame.
Bit 29 on the other hand means read/write whatever the type of mgmt frame
involved.
So this patch change the operation selection from bit 28 to bit 29 as it is
more generic.
Signed-off-by: Jean-Christophe Dubois <jcd@tribudubois.net>
Signed-off-by: Jason Wang <jasowang@redhat.com>
This patch introduces emulation for the Intel 82574 adapter, AKA e1000e.
This implementation is derived from the e1000 emulation code, and
utilizes the TX/RX packet abstractions that were initially developed for
the vmxnet3 device. Although some parts of the introduced code may be
shared with e1000, the differences are substantial enough so that the
only shared resources for the two devices are the definitions in
hw/net/e1000_regs.h.
Similarly to vmxnet3, the new device uses virtio headers for task
offloads (for backends that support virtio extensions). Usage of
virtio headers may be forcibly disabled via a boolean device property
"vnet" (which is enabled by default). In such case task offloads
will be performed in software, in the same way it is done on
backends that do not support virtio headers.
The device code is split into two parts:
1. hw/net/e1000e.c: QEMU-specific code for a network device;
2. hw/net/e1000e_core.[hc]: Device emulation according to the spec.
The new device name is e1000e.
Intel specifications for the 82574 controller are available at:
http://www.intel.com/content/dam/doc/datasheet/82574l-gbe-controller-datasheet.pdf
Throughput measurement results (iperf2):
Fedora 22 guest, TCP, RX
4 ++------------------------------------------+
| |
| X X X X X
3.5 ++ X X X X |
| X |
| |
3 ++ |
G | X |
b | |
/ 2.5 ++ |
s | |
| |
2 ++ |
| |
| |
1.5 X+ |
| |
+ + + + + + + + + + + +
1 ++--+---+---+---+---+---+---+---+---+---+---+
32 64 128 256 512 1 2 4 8 16 32 64
B B B B B KB KB KB KB KB KB KB
Buffer size
Fedora 22 guest, TCP, TX
18 ++-------------------------------------------+
| X |
16 ++ X X X X X
| X |
14 ++ |
| |
12 ++ |
G | X |
b 10 ++ |
/ | |
s 8 ++ |
| |
6 ++ X |
| |
4 ++ |
| X |
2 ++ X |
X + + + + + + + + + + +
0 ++--+---+---+---+---+----+---+---+---+---+---+
32 64 128 256 512 1 2 4 8 16 32 64
B B B B B KB KB KB KB KB KB KB
Buffer size
Fedora 22 guest, UDP, RX
3 ++------------------------------------------+
| X
| |
2.5 ++ |
| |
| |
2 ++ X |
G | |
b | |
/ 1.5 ++ |
s | X |
| |
1 ++ |
| |
| X |
0.5 ++ |
| X |
X + + + + +
0 ++-------+--------+-------+--------+--------+
32 64 128 256 512 1
B B B B B KB
Datagram size
Fedora 22 guest, UDP, TX
1 ++------------------------------------------+
| X
0.9 ++ |
| |
0.8 ++ |
0.7 ++ |
| |
G 0.6 ++ |
b | |
/ 0.5 ++ |
s | X |
0.4 ++ |
| |
0.3 ++ |
0.2 ++ X |
| |
0.1 ++ X |
X X + + + +
0 ++-------+--------+-------+--------+--------+
32 64 128 256 512 1
B B B B B KB
Datagram size
Windows 2012R2 guest, TCP, RX
3.2 ++------------------------------------------+
| X |
3 ++ |
| |
2.8 ++ |
| |
2.6 ++ X |
G | X X X X X
b 2.4 ++ X X |
/ | |
s 2.2 ++ |
| |
2 ++ |
| X X |
1.8 ++ |
| |
1.6 X+ |
+ + + + + + + + + + + +
1.4 ++--+---+---+---+---+---+---+---+---+---+---+
32 64 128 256 512 1 2 4 8 16 32 64
B B B B B KB KB KB KB KB KB KB
Buffer size
Windows 2012R2 guest, TCP, TX
14 ++-------------------------------------------+
| |
| X X
12 ++ |
| |
10 ++ |
| |
G | |
b 8 ++ |
/ | X |
s 6 ++ |
| |
| |
4 ++ X |
| |
2 ++ |
| X X X |
+ X X + + X X + + + + +
0 X+--+---+---+---+---+----+---+---+---+---+---+
32 64 128 256 512 1 2 4 8 16 32 64
B B B B B KB KB KB KB KB KB KB
Buffer size
Windows 2012R2 guest, UDP, RX
1.6 ++------------------------------------------X
| |
1.4 ++ |
| |
1.2 ++ |
| X |
| |
G 1 ++ |
b | |
/ 0.8 ++ |
s | |
0.6 ++ X |
| |
0.4 ++ |
| X |
| |
0.2 ++ X |
X + + + + +
0 ++-------+--------+-------+--------+--------+
32 64 128 256 512 1
B B B B B KB
Datagram size
Windows 2012R2 guest, UDP, TX
0.6 ++------------------------------------------+
| X
| |
0.5 ++ |
| |
| |
0.4 ++ |
G | |
b | |
/ 0.3 ++ X |
s | |
| |
0.2 ++ |
| |
| X |
0.1 ++ |
| X |
X X + + + +
0 ++-------+--------+-------+--------+--------+
32 64 128 256 512 1
B B B B B KB
Datagram size
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Code that will be shared moved to a separate files.
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
To make this device and network packets
abstractions ready for IOMMU.
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
This patch extends the TX/RX packet abstractions with features that will
be used by the e1000e device implementation.
Changes are:
1. Support iovec lists for RX buffers
2. Deeper RX packets parsing
3. Loopback option for TX packets
4. Extended VLAN headers handling
5. RSS processing for RX packets
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
This patch drops "vmx" prefix from packet abstractions names
to emphasize the fact they are generic and not tied to any
specific network device.
These abstractions will be reused by e1000e emulation implementation
introduced by following patches so their names need generalization.
This patch (except renamed files, adjusted comments and changes in MAINTAINTERS)
was produced by:
git grep -lz 'vmxnet_tx_pkt' | xargs -0 perl -i'' -pE "s/vmxnet_tx_pkt/net_tx_pkt/g"
git grep -lz 'vmxnet_rx_pkt' | xargs -0 perl -i'' -pE "s/vmxnet_rx_pkt/net_rx_pkt/g"
git grep -lz 'VmxnetTxPkt' | xargs -0 perl -i'' -pE "s/VmxnetTxPkt/NetTxPkt/g"
git grep -lz 'VMXNET_TX_PKT' | xargs -0 perl -i'' -pE "s/VMXNET_TX_PKT/NET_TX_PKT/g"
git grep -lz 'VmxnetRxPkt' | xargs -0 perl -i'' -pE "s/VmxnetRxPkt/NetRxPkt/g"
git grep -lz 'VMXNET_RX_PKT' | xargs -0 perl -i'' -pE "s/VMXNET_RX_PKT/NET_RX_PKT/g"
sed -ie 's/VMXNET_/NET_/g' hw/net/vmxnet_rx_pkt.c
sed -ie 's/VMXNET_/NET_/g' hw/net/vmxnet_tx_pkt.c
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Dmitry Fleytman <dmitry.fleytman@ravellosystems.com>
Signed-off-by: Leonid Bloch <leonid.bloch@ravellosystems.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
The last 8 bytes of the receive buffer list page (that has been supplied
by the guest with the H_REGISTER_LOGICAL_LAN call) contain a counter
for frames that have been dropped because there was no suitable receive
buffer available. This patch introduces code to use this field to
provide the information about dropped rx packets to the guest.
There it can be queried with "ethtool -S eth0 | grep rx_no_buffer".
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Currently, the spapr-vlan device is trying to flush the RX queue
after each RX buffer that has been added by the guest via the
H_ADD_LOGICAL_LAN_BUFFER hypercall. In case the receive buffer pool
was empty before, we only pass single packets to the guest this
way. This can cause very bad performance if a sender is trying
to stream fragmented UDP packets to the guest. For example when
using the UDP_STREAM test from netperf with UDP packets that are
much bigger than the MTU size, almost all UDP packets are dropped
in the guest since the chances are quite high that at least one of
the fragments got lost on the way.
When flushing the receive queue, it's much better if we'd have
a bunch of receive buffers available already, so that fragmented
packets can be passed to the guest in one go. To do this, the
spapr_vlan_receive() function should return 0 instead of -1 if there
are no more receive buffers available, so that receive_disabled = 1
gets temporarily set for the receive queue, and we have to delay
the queue flushing at the end of h_add_logical_lan_buffer() a little
bit by using a timer, so that the guest gets a chance to add multiple
RX buffers before we flush the queue again.
This improves the UDP_STREAM test with the spapr-vlan device a lot:
Running
netserver -p 44444 -L <guestip> -f -D -4
in the guest, and
netperf -p 44444 -L <hostip> -H <guestip> -t UDP_STREAM -l 60 -- -m 16384
in the host, I get the following values _without_ this patch:
Socket Message Elapsed Messages
Size Size Time Okay Errors Throughput
bytes bytes secs # # 10^6bits/sec
229376 16384 60.00 1738970 0 3798.83
229376 60.00 23 0.05
That "0.05" means that almost all UDP packets got lost/discarded
at the receiving side.
With this patch applied, the value look much better:
Socket Message Elapsed Messages
Size Size Time Okay Errors Throughput
bytes bytes secs # # 10^6bits/sec
229376 16384 60.00 1789104 0 3908.35
229376 60.00 22818 49.85
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
When receiving packets over MIPSnet network device, it uses
receive buffer of size 1514 bytes. In case the controller
accepts large(MTU) packets, it could lead to memory corruption.
Add check to avoid it.
Reported by: Oleksandr Bazhaniuk <oleksandr.bazhaniuk@intel.com>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Jason Wang <jasowang@redhat.com>
open_eth_start_xmit has a huge stack usage of 65536 bytes approx.
Moving large arrays to heap to reduce stack usage.
Reduce size of a buffer allocated on stack to 0x600 bytes, which is the
maximal frame length when HUGEN bit is not set in MODER, only allocate
buffer on heap when that is too small. Thus heap is not used in typical
use case.
Signed-off-by: Zhou Jie <zhoujie2011@cn.fujitsu.com>
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Drop local definitions of MII registers and use constants from mii.h for
registers and register bits. No functional changes.
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
Reserve this to CPU state serialization.
Luckily, they were only used by sPAPR devices and these are ppc64
only. So there is no change to migration format.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When receiving packets over Stellaris ethernet controller, it
uses receive buffer of size 2048 bytes. In case the controller
accepts large(MTU) packets, it could lead to memory corruption.
Add check to avoid it.
Reported-by: Oleksandr Bazhaniuk <oleksandr.bazhaniuk@intel.com>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Message-id: 1460095428-22698-1-git-send-email-ppandit@redhat.com
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Through CP_TX_OWN and CP_RX_OWN points to the same bit, we'd better use
CP_TX_OWN for tx descriptor handling.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Indicate that autonegotiation is complete in the MII BMSR. This fixes
networking on xtfpga platform in linux v4.5.
Cc: qemu-stable@nongnu.org
Signed-off-by: Max Filippov <jcmvbkbc@gmail.com>
This reverts commit 9596ef7c7b.
This workaround in order to fix endless interrupts is no
longer needed because it was superseded by the previous patch
(e1000: Fixing interrupt pace).
Signed-off-by: Sameeh Jubran <sameeh@daynix.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
This patch introduces an upper bound for number of interrupts
per second. Without this bound an interrupt storm can occur as
it has been observed on Windows 10 when disabling the device.
According to the SPEC - Intel PCI/PCI-X Family of Gigabit
Ethernet Controllers Software Developer's Manual, section
13.4.18 - the Ethernet controller guarantees a maximum
observable interrupt rate of 7813 interrupts/sec. If there is
no upper bound this could lead to an interrupt storm by e1000
(when mit_delay < 500) causing interrupts to fire at a very high
pace.
Thus if mit_delay < 500 then the delay should be set to the
minimum delay possible which is 500. This can be calculated
easily as follows:
Interval = 10^9 / (7813 * 256) = 500.
Signed-off-by: Sameeh Jubran <sameeh@daynix.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
* Chardev fix from Marc-André
* config.status tweak from David
* Header file tweaks from Markus, myself and Veronia (Outreachy candidate)
* get_ticks_per_sec() removal from Rutuja (Outreachy candidate)
* Coverity fix from myself
* PKE implementation from myself, based on rth's XSAVE support
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iQEcBAABCAAGBQJW9ErPAAoJEL/70l94x66DJfEH/A/QkMpAhrgNdyVsahzsGrzE
wx5gHFIc1nBYxyr62w4apUb5jPB7zaXu0LA7EAWDeAe0pyP8hZzLT9kJyOEDsuJu
zwKN2QeLSNMtPbnbKN0I/YQ2za2xX1V5ruhSeOJoVslUI214hgnAURaGshhQNzuZ
2CluDT9KgL5cQifAnKs5kJrwhIYShYNQB+1eDC/7wk28dd/EH+sPALIoF+rqrSmt
Zu4Mdqd+9Ns+oKOjA6br9ULq/Hzg0aDfY82J+XLVVqfF3PXQe8rTDmuMf/7jTn+M
Un7ZOcei9oZF2/9vfAfKQpDCcgD9HvOUSbgqV/ubmkPPmN/LNJzeKj0fBhrRN+Y=
=K12D
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
* Log filtering from Alex and Peter
* Chardev fix from Marc-André
* config.status tweak from David
* Header file tweaks from Markus, myself and Veronia (Outreachy candidate)
* get_ticks_per_sec() removal from Rutuja (Outreachy candidate)
* Coverity fix from myself
* PKE implementation from myself, based on rth's XSAVE support
# gpg: Signature made Thu 24 Mar 2016 20:15:11 GMT using RSA key ID 78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg: aka "Paolo Bonzini <pbonzini@redhat.com>"
* remotes/bonzini/tags/for-upstream: (28 commits)
target-i386: implement PKE for TCG
config.status: Pass extra parameters
char: translate from QIOChannel error to errno
exec: fix error handling in file_ram_alloc
cputlb: modernise the debug support
qemu-log: support simple pid substitution for logs
target-arm: dfilter support for in_asm
qemu-log: dfilter-ise exec, out_asm, op and opt_op
qemu-log: new option -dfilter to limit output
qemu-log: Improve the "exec" TB execution logging
qemu-log: Avoid function call for disabled qemu_log_mask logging
qemu-log: correct help text for -d cpu
tcg: pass down TranslationBlock to tcg_code_gen
util: move declarations out of qemu-common.h
Replaced get_tick_per_sec() by NANOSECONDS_PER_SECOND
hw: explicitly include qemu-common.h and cpu.h
include/crypto: Include qapi-types.h or qemu/bswap.h instead of qemu-common.h
isa: Move DMA_transfer_handler from qemu-common.h to hw/isa/isa.h
Move ParallelIOArg from qemu-common.h to sysemu/char.h
Move QEMU_ALIGN_*() from qemu-common.h to qemu/osdep.h
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Conflicts:
scripts/clean-includes
RX buffer pools are now enabled by default for new machine types.
For older machine types, they are still disabled to avoid breaking
migration.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
tl;dr:
This patch introduces an alternate way of handling the receive
buffers of the spapr-vlan device, resulting in much better
receive performance for the guest.
Full story:
One of our testers recently discovered that the performance of the
spapr-vlan device is very poor compared to other NICs, and that
a simple "ping -i 0.2 -s 65507 someip" in the guest can result
in more than 50% lost ping packets (especially with older guest
kernels < 3.17).
After doing some analysis, it was clear that there is a problem
with the way we handle the receive buffers in spapr_llan.c: The
ibmveth driver of the guest Linux kernel tries to add a lot of
buffers into several buffer pools (with 512, 2048 and 65536 byte
sizes by default, but it can be changed via the entries in the
/sys/devices/vio/1000/pool* directories of the guest). However,
the spapr-vlan device of QEMU only tries to squeeze all receive
buffer descriptors into one single page which has been supplied
by the guest during the H_REGISTER_LOGICAL_LAN call, without
taking care of different buffer sizes. This has two bad effects:
First, only a very limited number of buffer descriptors is accepted
at all. Second, we also hand 64k buffers to the guest even if
the 2k buffers would fit better - and this results in dropped packets
in the IP layer of the guest since too much skbuf memory is used.
Though it seems at a first glance like PAPR says that we should store
the receive buffer descriptors in the page that is supplied during
the H_REGISTER_LOGICAL_LAN call, chapter 16.4.1.2 in the LoPAPR spec
declares that "the contents of these descriptors are architecturally
opaque, none of these descriptors are manipulated by code above
the architected interfaces". That means we don't have to store
the RX buffer descriptors in this page, but can also manage the
receive buffers at the hypervisor level only. This is now what we
are doing here: Introducing proper RX buffer pools which are also
sorted by size of the buffers, so we can hand out a buffer with
the best fitting size when a packet has been received.
To avoid problems with migration from/to older version of QEMU,
the old behavior is also retained and enabled by default. The new
buffer management has to be enabled via a new "use-rx-buffer-pools"
property.
Now with the new buffer pool management enabled, the problem with
"ping -s 65507" is fixed for me, and the throughput of a simple
test with wget increases from creeping 3MB/s up to 20MB/s!
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Refactor the code a little bit by extracting the code that reads
and writes the receive buffer list page into separate functions.
There should be no functional change in this patch, this is just
a preparation for the upcoming extensions that introduce receive
buffer pools.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
This patch replaces get_ticks_per_sec() calls with the macro
NANOSECONDS_PER_SECOND. Also, as there are no callers, get_ticks_per_sec()
is then removed. This replacement improves the readability and
understandability of code.
For example,
timer_mod(fdctrl->result_timer,
qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL) + (get_ticks_per_sec() / 50));
NANOSECONDS_PER_SECOND makes it obvious that qemu_clock_get_ns
matches the unit of the expression on the right side of the plus.
Signed-off-by: Rutuja Shah <rutu.shah.26@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Manually drop redundant includes that scripts/clean-includes misses,
e.g. because they're hidden in generator programs, or they use the
wrong kind of delimiter.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Commit 57cb38b included qapi/error.h into qemu/osdep.h to get the
Error typedef. Since then, we've moved to include qemu/osdep.h
everywhere. Its file comment explains: "To avoid getting into
possible circular include dependencies, this file should not include
any other QEMU headers, with the exceptions of config-host.h,
compiler.h, os-posix.h and os-win32.h, all of which are doing a
similar job to this file and are under similar constraints."
qapi/error.h doesn't do a similar job, and it doesn't adhere to
similar constraints: it includes qapi-types.h. That's in excess of
100KiB of crap most .c files don't actually need.
Add the typedef to qemu/typedefs.h, and include that instead of
qapi/error.h. Include qapi/error.h in .c files that need it and don't
get it now. Include qapi-types.h in qom/object.h for uint16List.
Update scripts/clean-includes accordingly. Update it further to match
reality: replace config.h by config-target.h, add sysemu/os-posix.h,
sysemu/os-win32.h. Update the list of includes in the qemu/osdep.h
comment quoted above similarly.
This reduces the number of objects depending on qapi/error.h from "all
of them" to less than a third. Unfortunately, the number depending on
qapi-types.h shrinks only a little. More work is needed for that one.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
[Fix compilation without the spice devel packages. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add property to specify rocker world. All ports will be assigned to this
world.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Until now, 0 is returned in this error case. Fix it ro return -ENOMEM.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Port to world assignment should be permitted only by qemu user. Driver
should not be able to do it, so forbid that possibility.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Ne2000 NIC uses ring buffer of NE2000_MEM_SIZE(49152)
bytes to process network packets. Registers PSTART & PSTOP
define ring buffer size & location. Setting these registers
to invalid values could lead to infinite loop or OOB r/w
access issues. Add check to avoid it.
Reported-by: Yang Hongke <yanghongke@huawei.com>
Tested-by: Yang Hongke <yanghongke@huawei.com>
Signed-off-by: Prasad J Pandit <pjp@fedoraproject.org>
Signed-off-by: Jason Wang <jasowang@redhat.com>
Clean up includes so that osdep.h is included first and headers
which it implies are not included manually.
This commit was created with scripts/clean-includes.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Eric Blake <eblake@redhat.com>
Since guest_mask_notifier can not be used in vhost-user mode due
to buffering implied by unix control socket, force
use_mask_notifier on virtio devices of vhost-user interfaces, and
send correct callfd to the guest at vhost start.
Using guest_notifier_mask function in vhost-user case may
break interrupt mask paradigm, because mask/unmask is not
really done when returning from guest_notifier_mask call, instead
message is posted in a unix socket, and processed later.
Add an option boolean flag 'use_mask_notifier' to disable the use
of guest_notifier_mask in virtio pci.
Signed-off-by: Didier Pallard <didier.pallard@6wind.com>
Signed-off-by: Victor Kaplansky <victork@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Cross-endian is now handled by the core virtio-net code.
This patch reverts:
commit 5be7d9f1b1
vhost-net: tell tap backend about the vnet endianness
and
commit cf0a628f6e81bfc9b7a944fa0b80c3594836df56
net: set endianness on all backend devices
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
When running a fully emulated device in cross-endian conditions, including
a virtio 1.0 device offered to a big endian guest, we need to fix the vnet
headers. This is currently handled by the virtio_net_hdr_swap() function
in the core virtio-net code but it should actually be handled by the net
backend.
With this patch, virtio-net now tries to configure the backend to do the
endian fixing when the device starts (i.e. drivers sets the CONFIG_OK bit).
If the backend cannot support the requested endiannes, we have to fallback
onto virtio_net_hdr_swap(): this is recorded in the needs_vnet_hdr_swap flag,
to be used in the TX and RX paths.
Note that we reset the backend to the default behaviour (guest native
endianness) when the device stops (i.e. device status had CONFIG_OK bit and
driver unsets it). This is needed, with the linux tap backend at least,
otherwise the guest may lose network connectivity if rebooted into a
different endianness.
The current vhost-net code also tries to configure net backends. This will
be no more needed and will be reverted in a subsequent patch.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Greg Kurz <gkurz@linux.vnet.ibm.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Similar to the previous patch, it's nice to have all functions
in the tree that involve a visitor and a name for conversion to
or from QAPI to consistently stick the 'name' parameter next
to the Visitor parameter.
Done by manually changing include/qom/object.h and qom/object.c,
then running this Coccinelle script and touching up the fallout
(Coccinelle insisted on adding some trailing whitespace).
@ rule1 @
identifier fn;
typedef Object, Visitor, Error;
identifier obj, v, opaque, name, errp;
@@
void fn
- (Object *obj, Visitor *v, void *opaque, const char *name,
+ (Object *obj, Visitor *v, const char *name, void *opaque,
Error **errp) { ... }
@@
identifier rule1.fn;
expression obj, v, opaque, name, errp;
@@
fn(obj, v,
- opaque, name,
+ name, opaque,
errp)
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <1454075341-13658-20-git-send-email-eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
