On Wed, Mar 15, 2017 at 07:10:49PM +, Kevin Traynor wrote:
> The virtio_net structure is used in both enqueue and dequeue datapaths.
> broadcast_rarp is checked with cmpset in the dequeue datapath regardless
> of whether descriptors are available or not.
>
> It is observed in some cases where
> From: Vincent JARDIN [mailto:vincent.jar...@6wind.com]
>
> Le 15/03/2017 à 11:55, Thomas Monjalon a écrit :
> >> I'd suggest that this is a good topic for the next Tech Board
> meeting.
> > I agree Tim.
> > CC'ing techboard to add this item to the agenda of the next meeting.
>
> Frankly, I disa
When setting up the VSIs, MAC filter is used for
receiving MAC broadcast packets.
We should follow it to implement the broadcast
promiscuous mode setting.
Fixes: 61fff9b4c68b (net/i40e: set VF broadcast mode from PF)
Signed-off-by: Wenzhuo Lu
---
drivers/net/i40e/i40e_ethdev.c | 16
Hi,
> On Mar 15, 2017, at 4:55 PM, Yongseok Koh wrote:
>
> This patchset is to add the Enhanced Multi-Packet Send feature which is
> newly introduced for ConnectX-5 families of adaptors.
>
> v3:
> * Fix indentation
> * Fix typo in comment
> * Change name of variables and functions
I forgot to
ConnectX-5 supports enhanced version of multi-packet send (MPS). An MPS Tx
descriptor can carry multiple packets either by including pointers of
packets or by inlining packets. Inlining packet data can be helpful to
better utilize PCIe bandwidth. In addition, Enhanced MPS supports hybrid
mode - mix
Enhanced multi-packet send mode is newly introduced for ConnectX-5 families
of adaptors.
Signed-off-by: Yongseok Koh
---
doc/guides/nics/mlx5.rst | 31 +++
1 file changed, 27 insertions(+), 4 deletions(-)
diff --git a/doc/guides/nics/mlx5.rst b/doc/guides/nics/mlx5.r
This patchset is to add the Enhanced Multi-Packet Send feature which is
newly introduced for ConnectX-5 families of adaptors.
v3:
* Fix indentation
* Fix typo in comment
* Change name of variables and functions
v2:
* Resolves conflicts with other patches in review.
* Improved performance by reloc
> On Mar 15, 2017, at 11:09 PM, Pascal Mazon wrote:
>
> Signed-off-by: Pascal Mazon
Acked-by: Keith Wiles
> ---
> drivers/net/tap/rte_eth_tap.c | 6 ++
> drivers/net/tap/rte_eth_tap.h | 1 +
> 2 files changed, 7 insertions(+)
Regards,
Keith
> On Mar 15, 2017, at 11:09 PM, Pascal Mazon wrote:
>
> Support for segmented packets (scatter/gather) is mandatory for most
> purposes, regardless of the MTU size. Tx packets are often the result of
> mbuf concatenation, and an mbuf is not necessarily large enough for Rx
> packets to fit in a s
Please let me know if I have missed and Ack for the TAP PMD. I believe I have
reviewed and agree with all of the changes to TAP at this time.
> On Mar 16, 2017, at 12:46 AM, Yigit, Ferruh wrote:
>
> On 3/15/2017 2:48 PM, Pascal Mazon wrote:
>> v2 change:
>> - use snprintf in tap_mtu set
>>
>>
> On Mar 15, 2017, at 10:48 PM, Pascal Mazon wrote:
>
> There is no reason not to support ARP on a tap netdevice.
> Focus on IFF_UP when a link status change is required.
>
> Fixes: f457b472b1f2 ("net/tap: add link up and down operations")
> Signed-off-by: Pascal Mazon
> Reviewed-by: Ferruh Yi
> -Original Message-
> From: Thomas Monjalon [mailto:thomas.monja...@6wind.com]
> Sent: Wednesday, March 15, 2017 8:16 PM
> To: Dumitrescu, Cristian
> Cc: De Lara Guarch, Pablo ; Singh,
> Jasvinder ; dev@dpdk.org; Doherty, Declan
>
> Subject: Re: [dpdk-dev] [PATCH v2 1/2] librte_net: ad
> On Mar 16, 2017, at 2:02 AM, Billy McFall wrote:
>
> See request from 11/21/2016:
> http://dpdk.org/ml/archives/dev/2016-November/050585.html
>
> Add a new API to free consumed buffers on TX ring. This addresses two
> scenarios:
> 1) Flooding a packet and want to reuse existing mbuf to avoid
> On Mar 15, 2017, at 10:02 PM, Vincent JARDIN wrote:
>
> Le 15/03/2017 à 05:10, O'Driscoll, Tim a écrit :
>>> so, still an nack because:
>>> - no performance data of avp vs virtio,
>> I don't think it should be a requirement for Allain to provide performance
>> data in order to justify getti
2017-03-15 19:03, Dumitrescu, Cristian:
> ...
>
> > > > > > I think it should be in librte_hash.
> > > > > >
> > > > > > Please check lib/librte_hash/rte_hash_crc.h
> > > > >
> > > > > Is it good to include payload crc calculation in hash library as I
> > > > > see all
> > hash
> > > > related f
The virtio_net structure is used in both enqueue and dequeue datapaths.
broadcast_rarp is checked with cmpset in the dequeue datapath regardless
of whether descriptors are available or not.
It is observed in some cases where dequeue and enqueue are performed by
different cores and no packets are a
> -Original Message-
> From: Dumitrescu, Cristian
> Sent: Wednesday, March 15, 2017 7:04 PM
> To: 'Thomas Monjalon' ; De Lara Guarch,
> Pablo
> Cc: Singh, Jasvinder ; dev@dpdk.org; Doherty,
> Declan
> Subject: RE: [dpdk-dev] [PATCH v2 1/2] librte_net: add crc init and compute
> APIs
>
...
> > > > > I think it should be in librte_hash.
> > > > >
> > > > > Please check lib/librte_hash/rte_hash_crc.h
> > > >
> > > > Is it good to include payload crc calculation in hash library as I see
> > > > all
> hash
> > > related functionality there?
> > >
> > > I think yes. Pablo?
> >
> >
On Wed, 15 Mar 2017, Yong Wang wrote:
> > -Original Message-
> > From: Charles (Chas) Williams [mailto:ciwil...@brocade.com]
> > Sent: Wednesday, March 15, 2017 5:35 AM
> > To: dev@dpdk.org
> > Cc: Yong Wang ; Charles (Chas) Williams
> >
> > Subject: [PATCH] net/vmxnet3: fix queue size
On Wed, Mar 15, 2017 at 1:35 PM, Charles (Chas) Williams
wrote:
> If the user reconfigures the queue size, then the previously allocated
> memzone may potentially be too small. Release the memzone when a queue
> is released and allocate a new one each time a queue is setup.
>
> While here convert
On 3/15/2017 2:08 PM, Vincent JARDIN wrote:
> Le 15/03/2017 à 12:29, Ferruh Yigit a écrit :
>> The scope of the patch is limited to PMD.
>> As long as it is maintained, it is good to have alternative approaches.
>
> From your logic, then, how many PMDs can be accepted?
>
> See my previous email:
Add support to the vHostdriver for the new API to force free consumed
buffers on Tx ring. vHost does not cache the mbufs so there is no work
to do.
Signed-off-by: Billy McFall
Acked-by: Maxime Coquelin
---
doc/guides/nics/features/vhost.ini | 1 +
drivers/net/vhost/rte_eth_vhost.c | 11 ++
Add support to the e1000 igb driver for the new API to force free
consumed buffers on Tx ring. This API is independent of the tx_rs_thresh
setting. With this API, buffers should be free even if tx_rs_thresh is
not met.
e1000 igb driver does not implement a tx_rs_thresh to free mbufs, it
frees a sl
Add a new API to force free consumed buffers on Tx ring. API will return
the number of packets freed (0-n) or error code if feature not supported
(-ENOTSUP) or input invalid (-ENODEV).
Signed-off-by: Billy McFall
---
doc/guides/conf.py | 7 +--
doc/guides/nics/features/
See request from 11/21/2016:
http://dpdk.org/ml/archives/dev/2016-November/050585.html
Add a new API to free consumed buffers on TX ring. This addresses two
scenarios:
1) Flooding a packet and want to reuse existing mbuf to avoid a packet
copy. Increment the reference count of the packet and pol
> -Original Message-
> From: Charles (Chas) Williams [mailto:ciwil...@brocade.com]
> Sent: Wednesday, March 15, 2017 5:35 AM
> To: dev@dpdk.org
> Cc: Yong Wang ; Charles (Chas) Williams
>
> Subject: [PATCH] net/vmxnet3: fix queue size changes
>
> If the user reconfigures the queue size, t
2017-03-14 19:17, Vladyslav Buslov:
> This patch extend next_hop field from 8-bits to 21-bits in LPM library
> for IPv6.
>
> Added versioning symbols to functions and updated
> library and applications that have a dependency on LPM library.
>
> Signed-off-by: Vladyslav Buslov
> Acked-by: Bruce R
2017-03-08 11:08, De Lara Guarch, Pablo:
> From: Thomas Monjalon [mailto:thomas.monja...@6wind.com]
> > 2017-03-02 13:03, Singh, Jasvinder:
> > > From: Thomas Monjalon [mailto:thomas.monja...@6wind.com]
> > > > 2017-02-28 12:08, Jasvinder Singh:
> > > > > lib/librte_net/rte_net_crc.c | 664
> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Subject: [dpdk-dev] [PATCH 02/39] app/test: fix eventdev reconfigure test
>
> Minimum value of nb_event_ports and/or nb_event_queues
> should be one before reconfiguring the event
> From: Jerin Jacob [mailto:jerin.ja...@caviumnetworks.com]
> Sent: Friday, March 3, 2017 5:28 PM
> To: dev@dpdk.org
> Subject: [dpdk-dev] [PATCH 01/39] eventdev: update PMD dequeue timeout
> conversion
> callback
>
> eventdev driver may return error on dequeue timeout tick conversion.
> Change t
This commit adds an entry in the programmers guide
explaining the eventdev library.
The rte_event struct, queues and ports are explained.
An API walktrough of a simple two stage atomic pipeline
provides the reader with a step by step overview of the
expected usage of the Eventdev API.
Signed-off-
2017-03-13 11:27, Pascal Mazon:
> Signed-off-by: Pascal Mazon
> ---
> app/test-pmd/cmdline.c | 12 +---
> doc/guides/testpmd_app_ug/testpmd_funcs.rst | 7 +++
> 2 files changed, 16 insertions(+), 3 deletions(-)
Applied, thanks
2017-03-15 06:19, David Hunt:
> +/**
> + * Number of packets to deal with in bursts. Needs to be 8 so as to
> + * fit in one cache line.
> + */
> +#define RTE_DIST_BURST_SIZE (sizeof(rte_xmm_t) / sizeof(uint16_t))
error: 'rte_xmm_t' undeclared here (arm compilation)
Can it be fixed by including r
On 3/15/2017 2:48 PM, Pascal Mazon wrote:
> v2 change:
> - use snprintf in tap_mtu set
>
> v3 change:
> - call tap_mac_set() only once in tap_setup_queue()
>
> v4 changes:
> - use new tap_ioctl function for shared behavior between ops
> - update supported packet types
> - remove IFF_NOA
2017-02-06 14:29, Olivier Matz:
> The objective of this RFC patchset is to introduce a framework to
> support dynamic log types in EAL. It also provides one example of use
> (in i40e).
>
> Features:
> - log types are identified by a string
> - at registration, a uniq identifier is associated to a
On 3/10/2017 3:39 PM, Jerin Jacob wrote:
> Fixes: e438796617dc ("net/thunderx: add PMD skeleton")
>
> Signed-off-by: Jerin Jacob
Still getting build error with 32bit when debug enabled:
In file included from .../drivers/net/thunderx/base/../nicvf_logs.h:36:0,
from .../drivers/n
Hi Nicolas,
On Wed, Mar 15, 2017 at 02:29:44PM +, Le Scouarnec Nicolas wrote:
> Hi Adrien,
>
> > > > > And about the tpid, ethertype. I have a thought that why we need it
> > > > > as it's
> > > > duplicate with the item type. I think the initial design is just
> > > > following the IEEE
>
Support for segmented packets (scatter/gather) is mandatory for most
purposes, regardless of the MTU size. Tx packets are often the result of
mbuf concatenation, and an mbuf is not necessarily large enough for Rx
packets to fit in a single one.
Signed-off-by: Pascal Mazon
---
drivers/net/tap/rte
Signed-off-by: Pascal Mazon
---
drivers/net/tap/rte_eth_tap.c | 6 ++
drivers/net/tap/rte_eth_tap.h | 1 +
2 files changed, 7 insertions(+)
diff --git a/drivers/net/tap/rte_eth_tap.c b/drivers/net/tap/rte_eth_tap.c
index 69fa282ca176..17496501e6ba 100644
--- a/drivers/net/tap/rte_eth_tap.c
+
By default, a tap netdevice is of no use when not fed by a separate
process. The ability to automatically feed it from another netdevice
allows applications to capture any kind of traffic normally destined to
the kernel stack.
This patch implements this ability through a new optional "remote"
para
This patchset adds the special "remote" feature to the tap PMD, that
actually enables capturing traffic from another netdevice. This is
especially useful to get packets into the DPDK app, when the remote
netdevice has no DPDK support.
The "remote" feature requires flow API support as flow rules wi
On Wed, Mar 15, 2017 at 6:29 AM, Olivier Matz
wrote:
> Hi Billy,
>
> On Thu, 9 Mar 2017 15:51:17 -0500, Billy McFall
> wrote:
> > Add a new API to force free consumed buffers on Tx ring. API will return
> > the number of packets freed (0-n) or error code if feature not supported
> > (-ENOTSUP)
Tap PMD is flexible, it supports any speed.
Signed-off-by: Pascal Mazon
Reviewed-by: Ferruh Yigit
---
doc/guides/nics/features/tap.ini | 1 +
drivers/net/tap/rte_eth_tap.c| 35 +++
2 files changed, 36 insertions(+)
diff --git a/doc/guides/nics/features/tap.
Each kernel netdevice may have queueing disciplines set for it, which
determine how to handle the packet (mostly on egress). That's part of
the TC (Traffic Control) mechanism.
Through TC, it is possible to set filter rules that match specific
packets, and act according to what is in the rule. This
In the next patch, access to struct pmd_internals will be necessary in
tap_flow.c to store the flows.
Signed-off-by: Pascal Mazon
Acked-by: Olga Shern
---
drivers/net/tap/Makefile | 1 +
drivers/net/tap/rte_eth_tap.c | 36 ++---
drivers/net/tap/rte_eth_tap.h | 75 +
Supported flow rules are now mapped to TC rules on the tap netdevice.
The netlink message used for creating the TC rule is stored in struct
rte_flow. That way, by simply changing a metadata in it, we can require
for the rule deletion without further parsing.
Supported items:
- eth: src and dst (wi
The flow API provides the ability to classify packets received by a tap
netdevice.
This patch only implements skeleton functions for flow API support, no
patterns are supported yet.
Signed-off-by: Pascal Mazon
Acked-by: Olga Shern
---
doc/guides/nics/features/tap.ini | 1 +
drivers/net/tap/M
This series add support for the flow API in tap PMD.
It enables filtering specific packets incoming on the tap netdevice, to
process only desired ones. Under the hood, it uses kernel TC (traffic
control), which takes place very early in the stack, and supports most
common pattern items and actions
A tap netdevice actually receives every packet, without any filtering
whatsoever. There is no need for any multicast address registration
to receive multicast packets.
Signed-off-by: Pascal Mazon
Reviewed-by: Ferruh Yigit
---
doc/guides/nics/features/tap.ini | 1 +
drivers/net/tap/rte_eth_tap.
A tap netdevice does not support flow control; ensure nothing but
RTE_FC_NONE mode can be set.
Signed-off-by: Pascal Mazon
Reviewed-by: Ferruh Yigit
---
doc/guides/nics/features/tap.ini | 1 +
drivers/net/tap/rte_eth_tap.c| 19 +++
2 files changed, 20 insertions(+)
diff --
Advertize packet types supported by the librte_net.
Signed-off-by: Pascal Mazon
Reviewed-by: Ferruh Yigit
---
doc/guides/nics/features/tap.ini | 1 +
drivers/net/tap/Makefile | 1 +
drivers/net/tap/rte_eth_tap.c| 35 +++
3 files changed, 37 insertio
As soon as the netdevice is created, update pmd->mac_addr with its
actual MAC address.
Signed-off-by: Pascal Mazon
Reviewed-by: Ferruh Yigit
---
doc/guides/nics/features/tap.ini | 1 +
drivers/net/tap/rte_eth_tap.c| 39 +--
2 files changed, 34 insertions
The MTU is assigned to the tap netdevice according to the argument, but
packet transmission and reception just write/read on an fd with the
default limit being the socket buffer size.
As a new rte_eth_dev_data is allocated during tap device init, ensure it
is set again dev->data->mtu.
Once the act
There is no reason not to support ARP on a tap netdevice.
Focus on IFF_UP when a link status change is required.
Fixes: f457b472b1f2 ("net/tap: add link up and down operations")
Signed-off-by: Pascal Mazon
Reviewed-by: Ferruh Yigit
---
drivers/net/tap/rte_eth_tap.c | 4 ++--
1 file changed, 2 i
Create a socket for ioctl at tap device creation instead of opening it
and closing it every call to tap_link_set_flags().
Use a common tap_ioctl() function that can be extended for various uses
(such as MTU change, MAC address change, ...).
Signed-off-by: Pascal Mazon
Reviewed-by: Ferruh Yigit
v2 change:
- use snprintf in tap_mtu set
v3 change:
- call tap_mac_set() only once in tap_setup_queue()
v4 changes:
- use new tap_ioctl function for shared behavior between ops
- update supported packet types
- remove IFF_NOARP flag from link status change
- sync desired MTU to both r
On 3/15/2017 1:44 PM, Pascal Mazon wrote:
> On Wed, 15 Mar 2017 13:35:44 +
> Ferruh Yigit wrote:
>
>> On 3/14/2017 8:22 AM, Pascal Mazon wrote:
>>> Advertize packet types supported by the librte_net.
>>>
>>> Signed-off-by: Pascal Mazon
>>
>> <...>
>>
>>>
>>> #include
>>> #include
>>> @
Hi Adrien,
> > > > And about the tpid, ethertype. I have a thought that why we need it as
> > > > it's
> > > duplicate with the item type. I think the initial design is just
> > > following the IEEE
> > > spec to define the structures so we will not miss anything. But why not
> > > do some
> >
On 3/10/2017 11:13 PM, Yongseok Koh wrote:
> ConnectX-5 supports enhanced version of multi-packet send (MPS). An MPS Tx
> descriptor can carry multiple packets either by including pointers of
> packets or by inlining packets. Inlining packet data can be helpful to
> better utilize PCIe bandwidth. I
On Wed, Mar 15, 2017 at 12:15:56PM +0100, Thomas Monjalon wrote:
2017-03-15 03:28, Bruce Richardson:
On Tue, Mar 14, 2017 at 03:49:47PM +0100, Gaëtan Rivet wrote:
> - In the bonding, the init and configuration steps are still the
> responsibility of the application and no one else. The bonding
Le 15/03/2017 à 11:55, Thomas Monjalon a écrit :
I'd suggest that this is a good topic for the next Tech Board meeting.
I agree Tim.
CC'ing techboard to add this item to the agenda of the next meeting.
Frankly, I disagree, it is missing some discussions on the list.
Le 15/03/2017 à 12:29, Ferruh Yigit a écrit :
The scope of the patch is limited to PMD.
As long as it is maintained, it is good to have alternative approaches.
From your logic, then, how many PMDs can be accepted?
See my previous email: techboard should not bypass discussion of the
dev@ maili
Le 15/03/2017 à 05:10, O'Driscoll, Tim a écrit :
so, still an nack because:
- no performance data of avp vs virtio,
I don't think it should be a requirement for Allain to provide performance data
in order to justify getting this accepted into DPDK. Keith pointed out in a
previous comment on
Le 15/03/2017 à 05:10, O'Driscoll, Tim a écrit :
so, still an nack because:
- no performance data of avp vs virtio,
I don't think it should be a requirement for Allain to provide performance data
in order to justify getting this accepted into DPDK. Keith pointed out in a
previous comment on
clang error below:
2017-03-07 11:32, Bruce Richardson:
> + union {
> + uint32_t sp_enqueue; /**< True, if single producer. */
> + uint32_t sc_dequeue; /**< True, if single consumer. */
> + };
error: anonymous unions are a C11 extension
On Wed, 15 Mar 2017 13:35:44 +
Ferruh Yigit wrote:
> On 3/14/2017 8:22 AM, Pascal Mazon wrote:
> > Advertize packet types supported by the librte_net.
> >
> > Signed-off-by: Pascal Mazon
>
> <...>
>
> >
> > #include
> > #include
> > @@ -228,6 +229,8 @@ pmd_rx_burst(void *queue, stru
On 3/14/2017 8:22 AM, Pascal Mazon wrote:
> Add a few eth_dev ops to the tap PMD for completeness.
> We want it to behave as much as possible as a standard PMD.
>
> v2 change:
> - use snprintf in tap_mtu set
>
> v3 change:
> - call tap_mac_set() only once in tap_setup_queue()
>
> v4 changes:
On 3/14/2017 8:22 AM, Pascal Mazon wrote:
> Advertize packet types supported by the librte_net.
>
> Signed-off-by: Pascal Mazon
<...>
>
> +static const uint32_t*
> +tap_dev_supported_ptypes_get(struct rte_eth_dev *dev __rte_unused)
> +{
> + static const uint32_t ptypes[] = {
> +
On 3/14/2017 8:22 AM, Pascal Mazon wrote:
> Advertize packet types supported by the librte_net.
>
> Signed-off-by: Pascal Mazon
<...>
>
> #include
> #include
> @@ -228,6 +229,8 @@ pmd_rx_burst(void *queue, struct rte_mbuf **bufs,
> uint16_t nb_pkts)
> mbuf->data_len = len;
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
MAINTAINERS | 1 +
1 file changed, 1 insertion(+)
diff --git a/MAINTAINERS b/MAINTAINERS
index 39bc78e..0545911 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -494,6 +494,7 @@ F: doc/guides/sample_app_ug/ip_reassembly.rst
Distributor
M
Changes in the thread layout described, with an updated diagram.
Signed-off-by: David Hunt
---
doc/guides/sample_app_ug/dist_app.rst | 49 +++---
doc/guides/sample_app_ug/img/dist_app.svg | 276 +-
2 files changed, 179 insertions(+), 146 deletions(-)
diff --git
Approximately 10% performance increase due to these changes.
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
examples/distributor/main.c | 36 +++-
1 file changed, 23 insertions(+), 13 deletions(-)
diff --git a/examples/distributor/main.c b/examples/dis
Now that we're printing out a page of stats every second to the console,
we should give the stats it's own core so that we don't interfere with
the performance of the Rx core.
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
examples/distributor/main.c | 24
1 f
Signed-off-by: David Hunt
Acked-by: John McNamara
Acked-by: Bruce Richardson
---
doc/guides/prog_guide/packet_distrib_lib.rst | 3 +++
1 file changed, 3 insertions(+)
diff --git a/doc/guides/prog_guide/packet_distrib_lib.rst
b/doc/guides/prog_guide/packet_distrib_lib.rst
index b5bdabb..e0adca
On some machines, ports take several seconds to come up. This
patch causes the app to wait.
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
examples/distributor/main.c | 7 ---
1 file changed, 4 insertions(+), 3 deletions(-)
diff --git a/examples/distributor/main.c b/examples/dist
Give the distribution functionality it's own core for performance,
otherwise it's limited by the Rx core.
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
examples/distributor/main.c | 181 ++--
1 file changed, 123 insertions(+), 58 deletions(-)
This will allow us to see what's going on at various stages
throughout the sample app, with per-second visibility
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
examples/distributor/main.c | 140 +++-
1 file changed, 124 insertions(+), 16 deleti
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
test/test/test_distributor_perf.c | 75 ++-
1 file changed, 51 insertions(+), 24 deletions(-)
diff --git a/test/test/test_distributor_perf.c
b/test/test/test_distributor_perf.c
index 1dd326b..732d86d 1
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
test/test/test_distributor.c | 116 ++-
1 file changed, 82 insertions(+), 34 deletions(-)
diff --git a/test/test/test_distributor.c b/test/test/test_distributor.c
index 7a30513..890a852 100644
--- a
Also bumped up the ABI version number in the Makefile
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
lib/librte_distributor/Makefile| 2 +-
lib/librte_distributor/rte_distributor.c | 57 +++---
lib/librte_distributor/rte_distributor_v1705.h |
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
lib/librte_distributor/Makefile | 1 -
1 file changed, 1 deletion(-)
diff --git a/lib/librte_distributor/Makefile b/lib/librte_distributor/Makefile
index a812fe4..2b28eff 100644
--- a/lib/librte_distributor/Makefile
+++ b/lib/librte_distr
This is the main switch over between the legacy API and the new
burst API. We rename all the functions in rte_distributor.c to remove
the _v1705, and we add in _v20 in the rte_distributor_v20.c
We also rename the rte_distributor_next.h as rte_distributor.h, as
this is now the public header.
At th
In the next few patches, we'll want to test old and new API,
so here we're allowing different parameters to be passed to
the tests, instead of just a distributor struct.
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
test/test/test_distributor.c | 64 +-
Add an optimised version of the in-flight flow matching algorithm
using SIMD instructions. This should give up to 1.5x over the scalar
versions performance.
Falls back to scalar version if SSE4.2 not available
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
lib/librte_distributor/Make
This patch includes the code for new burst-capable distributor library.
It also includes the rte_distributor_next.h file which will
be used as the public header once we add in the symbol versioning
for v20 and v1705 APIs, at which stage we will rename it to
rte_distributor.h.
The new distributor
We'll be adding internal implementation definitions in here
that are common to both burst and legacy APIs.
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
lib/librte_distributor/rte_distributor_private.h | 136 +++
lib/librte_distributor/rte_distributor_v20.c |
Move files out of the way so that we can replace with new
versions of the distributor libtrary. Files are named in
such a way as to match the symbol versioning that we will
apply for backward ABI compatibility.
Signed-off-by: David Hunt
Acked-by: Bruce Richardson
---
lib/librte_distributor/Make
This patch aims to improve the throughput of the distributor library.
It uses a similar handshake mechanism to the previous version of
the library, in that bits are used to indicate when packets are ready
to be sent to a worker and ready to be returned from a worker. One main
difference is that in
2017-03-14 16:59, Ferruh Yigit:
> On 3/14/2017 4:33 PM, Allain Legacy wrote:
> > From: Matt Peters
> >
> > Allow the BAR setup to succeed if a device has at least 1 BAR region
> > defined. Previously, the device probe would only succeed if at least one
> > memory BAR existed, but there are devic
2017-03-14 10:14, Olivier Matz:
> From: Hemant Agrawal
>
> When possible, replace the uses of rte_mempool_create() with
> the helper provided in librte_mbuf: rte_pktmbuf_pool_create().
>
> This is the preferred way to create a mbuf pool.
>
> This also updates the documentation.
>
> Signed-off-
2017-03-10 18:37, Dumitrescu, Cristian:
> From: O'Driscoll, Tim
> > > > OK I better understand now.
> > > > You should add this level of explanation in your patch.
> > > >
> > > > However I am reluctant to add an API if there is no user.
> > > > I think we should wait to have at least one existing
If the user reconfigures the queue size, then the previously allocated
memzone may potentially be too small. Release the memzone when a queue
is released and allocate a new one each time a queue is setup.
While here convert to rte_eth_dma_zone_reserve() which does basically
the same things as the
On 03/15/2017 06:05 AM, Jan Blunck wrote:
On Wed, Mar 15, 2017 at 10:45 AM, Charles (Chas) Williams
wrote:
On 03/15/2017 04:18 AM, Jan Blunck wrote:
On Tue, Mar 14, 2017 at 5:38 PM, Charles (Chas) Williams
wrote:
On 03/14/2017 12:11 PM, Jan Blunck wrote:
On Mon, Mar 13, 2017 at 11
Signed-off-by: Nelio Laranjeiro
Acked-by: Adrien Mazarguil
---
drivers/net/mlx5/mlx5_flow.c | 10 ++
1 file changed, 10 insertions(+)
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index e789c4c..5468642 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/n
> -Original Message-
> From: Dai, Wei
> Sent: Tuesday, March 14, 2017 2:48 PM
> To: dev@dpdk.org
> Cc: Ananyev, Konstantin ; Dai, Wei
> ; sta...@dpdk.org
> Subject: [PATCH v2] examples/ip_fragmentation: fix check of packet type
>
> The packet_type in mbuf is not correctly filled by ixgb
On 3/15/2017 4:10 AM, O'Driscoll, Tim wrote:
> I've included a couple of specific comments inline below, and a general
> comment here.
>
> We have somebody proposing to add a new driver to DPDK. It's standalone and
> doesn't affect any of the core libraries.
> They're willing to maintain the dri
2017-03-15 03:28, Bruce Richardson:
> On Tue, Mar 14, 2017 at 03:49:47PM +0100, Gaëtan Rivet wrote:
> > > > 2. Bonding and link availability
> > > >
> > > > The hot-plug functionality is not a core function of the bonding PMD.
> > > > It is only interested in knowing if the link is active or not.
On 3/9/2017 5:23 PM, Andrew Rybchenko wrote:
> From: Ivan Malov
>
> The patch is to add support for MCDI proxy which comes in
> useful, particularly, while running over VF: few commands
> will normally fail with EPERM, but in some cases the host
> driver (i.e. running over the corresponding PF, t
On 3/9/2017 5:22 PM, Andrew Rybchenko wrote:
> Running Solarflare PMD over PCI virtual functions requires few fixes:
> - periodic MAC stats DMA is not supported for VFs
> - promiscious and all-multicast modes may be denied by FW/PF
> - MCDI proxy authorization made by FW should be properly handl
2017-03-15 04:10, O'Driscoll, Tim:
> I've included a couple of specific comments inline below, and a general
> comment here.
>
> We have somebody proposing to add a new driver to DPDK. It's standalone and
> doesn't affect any of the core libraries. They're willing to maintain the
> driver and h
1 - 100 of 120 matches
Mail list logo
