2015-09-02 07:59:58 +00:00
|
|
|
DPDK Release 2.2
|
|
|
|
================
|
|
|
|
|
|
|
|
New Features
|
|
|
|
------------
|
|
|
|
|
2015-10-27 12:51:43 +00:00
|
|
|
* **Added API in ethdev to retrieve RX/TX queue information.**
|
|
|
|
|
|
|
|
* Add the ability for the upper layer to query RX/TX queue information.
|
|
|
|
* Add into rte_eth_dev_info new fields to represent information about
|
|
|
|
RX/TX descriptors min/max/align numbers per queue for the device.
|
|
|
|
|
2015-10-30 14:25:48 +00:00
|
|
|
* **Added RSS dynamic configuration to bonding.**
|
|
|
|
|
2015-10-29 09:18:38 +00:00
|
|
|
* **Added e1000 Rx interrupt support.**
|
|
|
|
|
2015-10-21 07:55:26 +00:00
|
|
|
* **Added igb TSO support for both PF and VF.**
|
|
|
|
|
2015-10-16 13:05:38 +00:00
|
|
|
* **RSS enhancement on Intel x550 NIC**
|
|
|
|
|
|
|
|
* Support 512 entries RSS redirection table.
|
2015-10-16 13:05:40 +00:00
|
|
|
* Support per VF RSS redirection table.
|
2015-10-16 13:05:38 +00:00
|
|
|
|
2015-10-26 05:27:33 +00:00
|
|
|
* **Flow director enhancement on Intel x550 NIC**
|
|
|
|
|
|
|
|
* Add 2 new flow director modes on x550.
|
|
|
|
* One is MAC VLAN mode, the other is tunnel mode.
|
|
|
|
|
2015-10-30 14:16:38 +00:00
|
|
|
* **Added i40e vector RX/TX.**
|
|
|
|
|
2015-10-30 04:57:35 +00:00
|
|
|
* **Added i40e flow control support.**
|
|
|
|
|
2015-10-31 15:57:25 +00:00
|
|
|
* **Added DCB support to i40e PF driver.**
|
|
|
|
|
2015-10-30 08:02:54 +00:00
|
|
|
* **Added fm10k vector RX/TX.**
|
|
|
|
|
2015-10-12 06:37:16 +00:00
|
|
|
* **Added fm10k TSO support for both PF and VF.**
|
|
|
|
|
2015-10-31 02:44:43 +00:00
|
|
|
* **Added fm10k VMDQ support.**
|
|
|
|
|
2015-10-26 04:53:55 +00:00
|
|
|
* **New NIC Boulder Rapid support.**
|
|
|
|
|
|
|
|
Boulder Rapid is a new NIC of Intel's fm10k family.
|
|
|
|
|
2015-10-08 13:46:05 +00:00
|
|
|
* **Enhanced support for the Chelsio CXGBE driver.**
|
|
|
|
|
2015-10-08 13:46:06 +00:00
|
|
|
* Added support for Jumbo Frames.
|
2015-10-08 13:46:05 +00:00
|
|
|
* Optimize forwarding performance for Chelsio T5 40GbE cards.
|
|
|
|
|
2015-10-30 18:52:42 +00:00
|
|
|
* **Added support for Mellanox ConnectX-4 adapters (mlx5).**
|
|
|
|
|
|
|
|
The mlx5 poll-mode driver implements support for Mellanox ConnectX-4 EN
|
|
|
|
and Mellanox ConnectX-4 Lx EN families of 10/25/40/50/100 Gb/s adapters.
|
|
|
|
|
|
|
|
Like mlx4, this PMD is only available for Linux and is disabled by default
|
|
|
|
due to external dependencies (libibverbs and libmlx5).
|
|
|
|
|
virtio: optimize ring layout
In DPDK based switching environment, mostly vhost runs on a dedicated core
while virtio processing in guest VMs runs on different cores.
Take RX for example, with generic implementation, for each guest buffer,
a) virtio driver allocates a descriptor from free descriptor list
b) modify the entry of avail ring to point to allocated descriptor
c) after packet is received, free the descriptor
When vhost fetches the avail ring, it need to fetch the modified L1 cache from
virtio core, which is a heavy cost in current CPU implementation.
This idea of this optimization is:
allocate the fixed descriptor for each entry of avail ring, so avail ring will
always be the same during the run.
This removes L1M cache transfer from virtio core to vhost core for avail ring.
(Note we couldn't avoid the cache transfer for descriptors).
Besides, descriptor allocation and free operation is eliminated.
This also makes vector procesing possible to further accelerate the processing.
This is the layout for the avail ring(take 256 ring entries for example), with
each entry pointing to the descriptor with the same index.
avail
idx
+
|
+----+----+---+-------------+------+
| 0 | 1 | 2 | ... | 254 | 255 | avail ring
+-+--+-+--+-+-+---------+---+--+---+
| | | | | |
| | | | | |
v v v | v v
+-+--+-+--+-+-+---------+---+--+---+
| 0 | 1 | 2 | ... | 254 | 255 | desc ring
+----+----+---+-------------+------+
|
|
+----+----+---+-------------+------+
| 0 | 1 | 2 | | 254 | 255 | used ring
+----+----+---+-------------+------+
|
+
This is the ring layout for TX.
As we need one virtio header for each xmit packet, we have 128 slots available.
++
||
||
+-----+-----+-----+--------------+------+------+------+
| 0 | 1 | ... | 127 || 128 | 129 | ... | 255 | avail ring
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| | | || | | |
v v v || v v v
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| 128 | 129 | ... | 255 || 128 | 129 | ... | 255 | desc ring for virtio_net_hdr
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| | | || | | |
v v v || v v v
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| 0 | 1 | ... | 127 || 0 | 1 | ... | 127 | desc ring for tx dat
+-----+-----+-----+--------------+------+------+------+
||
||
++
Signed-off-by: Huawei Xie <huawei.xie@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>
2015-10-29 14:53:24 +00:00
|
|
|
* **Enhanced support for virtio driver.**
|
|
|
|
|
|
|
|
* Virtio ring layout optimization (fixed avail ring)
|
2015-10-29 14:53:26 +00:00
|
|
|
* Vector RX
|
virtio: optimize ring layout
In DPDK based switching environment, mostly vhost runs on a dedicated core
while virtio processing in guest VMs runs on different cores.
Take RX for example, with generic implementation, for each guest buffer,
a) virtio driver allocates a descriptor from free descriptor list
b) modify the entry of avail ring to point to allocated descriptor
c) after packet is received, free the descriptor
When vhost fetches the avail ring, it need to fetch the modified L1 cache from
virtio core, which is a heavy cost in current CPU implementation.
This idea of this optimization is:
allocate the fixed descriptor for each entry of avail ring, so avail ring will
always be the same during the run.
This removes L1M cache transfer from virtio core to vhost core for avail ring.
(Note we couldn't avoid the cache transfer for descriptors).
Besides, descriptor allocation and free operation is eliminated.
This also makes vector procesing possible to further accelerate the processing.
This is the layout for the avail ring(take 256 ring entries for example), with
each entry pointing to the descriptor with the same index.
avail
idx
+
|
+----+----+---+-------------+------+
| 0 | 1 | 2 | ... | 254 | 255 | avail ring
+-+--+-+--+-+-+---------+---+--+---+
| | | | | |
| | | | | |
v v v | v v
+-+--+-+--+-+-+---------+---+--+---+
| 0 | 1 | 2 | ... | 254 | 255 | desc ring
+----+----+---+-------------+------+
|
|
+----+----+---+-------------+------+
| 0 | 1 | 2 | | 254 | 255 | used ring
+----+----+---+-------------+------+
|
+
This is the ring layout for TX.
As we need one virtio header for each xmit packet, we have 128 slots available.
++
||
||
+-----+-----+-----+--------------+------+------+------+
| 0 | 1 | ... | 127 || 128 | 129 | ... | 255 | avail ring
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| | | || | | |
v v v || v v v
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| 128 | 129 | ... | 255 || 128 | 129 | ... | 255 | desc ring for virtio_net_hdr
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| | | || | | |
v v v || v v v
+--+--+--+--+-----+---+------+---+--+---+------+--+---+
| 0 | 1 | ... | 127 || 0 | 1 | ... | 127 | desc ring for tx dat
+-----+-----+-----+--------------+------+------+------+
||
||
++
Signed-off-by: Huawei Xie <huawei.xie@intel.com>
Acked-by: Jianfeng Tan <jianfeng.tan@intel.com>
2015-10-29 14:53:24 +00:00
|
|
|
|
2015-10-22 12:35:55 +00:00
|
|
|
* **Added vhost-user multiple queue support.**
|
|
|
|
|
2015-10-27 16:03:13 +00:00
|
|
|
* **Added port hotplug support to xenvirt.**
|
|
|
|
|
|
|
|
|
2015-09-02 07:59:58 +00:00
|
|
|
Resolved Issues
|
|
|
|
---------------
|
|
|
|
|
2015-09-24 16:23:40 +00:00
|
|
|
EAL
|
|
|
|
~~~
|
|
|
|
|
|
|
|
* **eal/linux: Fixed epoll timeout.**
|
|
|
|
|
|
|
|
Fixed issue where the ``rte_epoll_wait()`` function didn't return when the
|
|
|
|
underlying call to ``epoll_wait()`` timed out.
|
|
|
|
|
|
|
|
|
|
|
|
Drivers
|
|
|
|
~~~~~~~
|
|
|
|
|
2015-10-19 13:20:03 +00:00
|
|
|
* **igb: Fixed IEEE1588 frame identification in I210.**
|
|
|
|
|
|
|
|
Fixed issue where the flag PKT_RX_IEEE1588_PTP was not being set
|
|
|
|
in Intel I210 NIC, as EtherType in RX descriptor is in bits 8:10 of
|
|
|
|
Packet Type and not in the default bits 0:2.
|
|
|
|
|
2015-09-24 16:23:40 +00:00
|
|
|
* **ixgbe: Fixed issue with X550 DCB.**
|
|
|
|
|
|
|
|
Fixed a DCB issue with x550 where for 8 TCs (Traffic Classes), if a packet
|
|
|
|
with user priority 6 or 7 was injected to the NIC, then the NIC would only
|
|
|
|
put 3 packets into the queue. There was also a similar issue for 4 TCs.
|
|
|
|
|
|
|
|
* **ixgbe: Removed burst size restriction of vector RX.**
|
|
|
|
|
|
|
|
Fixed issue where a burst size less than 32 didn't receive anything.
|
|
|
|
|
|
|
|
* **i40e: Fixed base driver allocation when not using first numa node.**
|
|
|
|
|
|
|
|
Fixed i40e issue that occurred when a DPDK application didn't initialize
|
|
|
|
ports if memory wasn't available on socket 0.
|
|
|
|
|
|
|
|
* **vhost: Fixed Qemu shutdown.**
|
|
|
|
|
|
|
|
Fixed issue with libvirt ``virsh destroy`` not killing the VM.
|
|
|
|
|
2015-10-14 09:50:49 +00:00
|
|
|
* **virtio: Fixed crash after changing link state.**
|
|
|
|
|
|
|
|
Fixed io permission in the interrupt handler.
|
|
|
|
|
2015-10-20 15:37:41 +00:00
|
|
|
* **virtio: Fixed crash when releasing queue.**
|
|
|
|
|
|
|
|
Fixed issue when releasing null control queue.
|
|
|
|
|
2015-09-24 16:23:40 +00:00
|
|
|
|
|
|
|
Libraries
|
|
|
|
~~~~~~~~~
|
|
|
|
|
|
|
|
* **hash: Fixed memory allocation of Cuckoo Hash key table.**
|
|
|
|
|
|
|
|
Fixed issue where an incorrect Cuckoo Hash key table size could be
|
|
|
|
calculated limiting the size to 4GB.
|
|
|
|
|
|
|
|
|
|
|
|
Examples
|
|
|
|
~~~~~~~~
|
|
|
|
|
|
|
|
|
|
|
|
Other
|
|
|
|
~~~~~
|
|
|
|
|
2015-09-02 07:59:58 +00:00
|
|
|
|
|
|
|
Known Issues
|
|
|
|
------------
|
|
|
|
|
|
|
|
|
|
|
|
API Changes
|
|
|
|
-----------
|
|
|
|
|
2015-09-21 07:34:09 +00:00
|
|
|
* The deprecated flow director API is removed.
|
|
|
|
It was replaced by rte_eth_dev_filter_ctrl().
|
|
|
|
|
2015-10-31 15:57:23 +00:00
|
|
|
* The dcb_queue is renamed to dcb_tc in following dcb configuration
|
|
|
|
structures: rte_eth_dcb_rx_conf, rte_eth_dcb_tx_conf,
|
|
|
|
rte_eth_vmdq_dcb_conf, rte_eth_vmdq_dcb_tx_conf.
|
|
|
|
|
2015-09-01 12:58:16 +00:00
|
|
|
* The function rte_eal_pci_close_one() is removed.
|
|
|
|
It was replaced by rte_eal_pci_detach().
|
|
|
|
|
2015-09-01 14:57:31 +00:00
|
|
|
* The deprecated ACL API ipv4vlan is removed.
|
|
|
|
|
2015-09-04 09:05:40 +00:00
|
|
|
* The deprecated hash function rte_jhash2() is removed.
|
|
|
|
It was replaced by rte_jhash_32b().
|
|
|
|
|
2015-06-15 16:51:12 +00:00
|
|
|
* The deprecated KNI functions are removed:
|
|
|
|
rte_kni_create(), rte_kni_get_port_id() and rte_kni_info_get().
|
|
|
|
|
2015-06-15 16:51:11 +00:00
|
|
|
* The deprecated ring PMD functions are removed:
|
|
|
|
rte_eth_ring_pair_create() and rte_eth_ring_pair_attach().
|
|
|
|
|
2015-09-29 06:53:01 +00:00
|
|
|
* The devargs union field virtual is renamed to virt for C++ compatibility.
|
|
|
|
|
2015-09-02 07:59:58 +00:00
|
|
|
|
|
|
|
ABI Changes
|
|
|
|
-----------
|
|
|
|
|
2015-08-26 09:18:53 +00:00
|
|
|
* The EAL and ethdev structures rte_intr_handle and rte_eth_conf were changed
|
|
|
|
to support Rx interrupt. It was already done in 2.1 for CONFIG_RTE_NEXT_ABI.
|
|
|
|
|
2015-08-31 15:18:22 +00:00
|
|
|
* The ethdev flow director entries for SCTP were changed.
|
|
|
|
It was already done in 2.1 for CONFIG_RTE_NEXT_ABI.
|
|
|
|
|
2015-10-27 12:51:43 +00:00
|
|
|
* The new fields rx_desc_lim and tx_desc_lim are added into rte_eth_dev_info
|
|
|
|
structure.
|
|
|
|
|
2015-08-31 14:51:21 +00:00
|
|
|
* The mbuf structure was changed to support unified packet type.
|
|
|
|
It was already done in 2.1 for CONFIG_RTE_NEXT_ABI.
|
|
|
|
|
2015-09-01 13:29:37 +00:00
|
|
|
* The dummy malloc library is removed. The content was moved into EAL in 2.1.
|
|
|
|
|
2015-09-01 13:41:05 +00:00
|
|
|
* The LPM structure is changed. The deprecated field mem_location is removed.
|
|
|
|
|
2015-09-17 16:03:19 +00:00
|
|
|
* librte_table LPM: A new parameter to hold the table name will be added to
|
|
|
|
the LPM table parameter structure.
|
2015-09-02 07:59:58 +00:00
|
|
|
|
2015-09-11 13:35:45 +00:00
|
|
|
* librte_port: Macros to access the packet meta-data stored within the packet
|
|
|
|
buffer has been adjusted to cover the packet mbuf structure.
|
|
|
|
|
2015-10-22 14:03:03 +00:00
|
|
|
* librte_cfgfile: Allow longer names and values by increasing the constants
|
|
|
|
CFG_NAME_LEN and CFG_VALUE_LEN to 64 and 256 respectively.
|
|
|
|
|
2015-09-11 13:35:45 +00:00
|
|
|
|
2015-09-02 07:59:58 +00:00
|
|
|
Shared Library Versions
|
|
|
|
-----------------------
|
|
|
|
|
|
|
|
The libraries prepended with a plus sign were incremented in this version.
|
|
|
|
|
|
|
|
.. code-block:: diff
|
|
|
|
|
2015-08-26 09:18:53 +00:00
|
|
|
+ libethdev.so.2
|
2015-09-01 14:57:31 +00:00
|
|
|
+ librte_acl.so.2
|
2015-10-22 14:03:03 +00:00
|
|
|
+ librte_cfgfile.so.2
|
2015-09-02 07:59:58 +00:00
|
|
|
librte_cmdline.so.1
|
|
|
|
librte_distributor.so.1
|
2015-08-26 09:18:53 +00:00
|
|
|
+ librte_eal.so.2
|
2015-09-04 09:05:40 +00:00
|
|
|
+ librte_hash.so.2
|
2015-09-02 07:59:58 +00:00
|
|
|
librte_ip_frag.so.1
|
|
|
|
librte_ivshmem.so.1
|
|
|
|
librte_jobstats.so.1
|
2015-06-15 16:51:12 +00:00
|
|
|
+ librte_kni.so.2
|
2015-09-02 07:59:58 +00:00
|
|
|
librte_kvargs.so.1
|
2015-09-01 13:41:05 +00:00
|
|
|
+ librte_lpm.so.2
|
2015-08-31 14:51:21 +00:00
|
|
|
+ librte_mbuf.so.2
|
2015-09-02 07:59:58 +00:00
|
|
|
librte_mempool.so.1
|
|
|
|
librte_meter.so.1
|
|
|
|
librte_pipeline.so.1
|
|
|
|
librte_pmd_bond.so.1
|
2015-06-15 16:51:11 +00:00
|
|
|
+ librte_pmd_ring.so.2
|
2015-09-11 13:35:45 +00:00
|
|
|
+ librte_port.so.2
|
2015-09-02 07:59:58 +00:00
|
|
|
librte_power.so.1
|
|
|
|
librte_reorder.so.1
|
|
|
|
librte_ring.so.1
|
|
|
|
librte_sched.so.1
|
2015-09-17 16:03:19 +00:00
|
|
|
+ librte_table.so.2
|
2015-09-02 07:59:58 +00:00
|
|
|
librte_timer.so.1
|
|
|
|
librte_vhost.so.1
|