- tx_offload
- packet_type
Adapted from an idea by Michael Qiu <qiud...@chinac.com>:
https://patchwork.ozlabs.org/patch/777570/
Signed-off-by: Mark Kavanagh <mark.b.kavan...@intel.com>
Acked-by: Michael Qiu <qiud...@chinac.com>
---
lib/dp-packet.h | 6 +++---
From: Michael Qiu <qiud...@chinac.com>
Currently, one packet is only copied to one segment
in function dpdk_do_tx_copy(), this could be an issue
when a jumbo frame comes, especially for multiple segments.
This patch calculate the segment number needed by the packet and
copy th
From: Michael Qiu <qiud...@chinac.com>
When doing packet clone, if packet source is from DPDK driver,
multi-segment must be considered, and copy the segment's
data one by one.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
---
lib/dp-packet.c | 27 ---
1
From: Michael Qiu <qiud...@chinac.com>
Currently, when doing packet copy, lots of DPDK mbuf's info
will be missed, like packet type, ol_flags, etc.
Those information is very important for DPDK to do
packets processing.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
---
lib/dp-pa
From: Michael Qiu <qiud...@chinac.com>
Currently, OVS only support DPDK single segment mbuf,
it could lead problems, like a large non-DPDK source
packet transmit to dpdk port.
Also, OVS doesn't copy enough info in mbuf when do
packet copy.
At the same time, vlan and tunnelling packet'
在 2017/5/31 18:59, Kavanagh, Mark B 写道:
From: Michael Qiu [mailto:qdy220091...@gmail.com]
Sent: Wednesday, May 31, 2017 9:58 AM
To: Ben Pfaff <b...@ovn.org>; Kavanagh, Mark B <mark.b.kavan...@intel.com>
Cc: d...@openvswitch.org; Lal, PrzemyslawX <przemyslawx....@intel.com>;
在 2017/5/5 23:44, Ben Pfaff 写道:
On Fri, May 05, 2017 at 08:57:27AM +, Kavanagh, Mark B wrote:
On Tue, May 02, 2017 at 02:10:43PM +0800, Michael Qiu wrote:
From: Michael Qiu <qiud...@chinac.com>
When a packet is from DPDK source, and it contains
multiple segments, data_len is not
Hi, Ben
在 2017/5/5 7:00, Darrell Ball 写道:
On 5/1/17, 11:10 PM, "ovs-dev-boun...@openvswitch.org on behalf of Michael Qiu"
<ovs-dev-boun...@openvswitch.org on behalf of qdy220091...@gmail.com> wrote:
From: Michael Qiu <qiud...@chinac.com>
When building with D
From: Michael Qiu <qiud...@chinac.com>
Currently, one packet is only copied to one segment
in function dpdk_do_tx_copy(), this could be an issue
when a jumbo frame comes, especially for multiple segments.
This patch calculate the segment number needed by the packet and
copy th
From: Michael Qiu <qiud...@chinac.com>
When doing packet clone, if packet source is from DPDK driver,
multi-segment must be considered, and copy the segment's
data one by one.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
---
lib/dp-packet.c | 27 ---
1
From: Michael Qiu <qiud...@chinac.com>
When a packet is from DPDK source, and it contains
multiple segments, data_len is not equal to the
packet size. This patch fix this issue.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Marcin Ksiadz <marcinx.ksi...@intel.
From: Michael Qiu <qiud...@chinac.com>
Currently, when doing packet copy, lots of DPDK mbuf's info
will be missed, like packet type, ol_flags, etc.
Those information is very important for DPDK to do
packets processing.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
---
lib/dp-pa
From: Michael Qiu <qiud...@chinac.com>
Currently, OVS only support DPDK single segment mbuf,
it could lead problems, like a large non-DPDK source
packet transmit to dpdk port.
Also, OVS doesn't copy enough info in mbuf when do
packet copy.
At the same time, vlan and tunnelling packet'
From: Michael Qiu <qiud...@chinac.com>
When building with DPDK, and using xmalloc() to get a new packet,
field mbuf of the packet will not be initialized, but it's very important for
DPDK port when copying the data to DPDK mbuf, because if ol_flags
and other info are random values, DPDK
From: Michael Qiu <qiud...@chinac.com>
Currently, one packet is only copied to one segment
in function dpdk_do_tx_copy(), this could be an issue
when a jumbo frame comes, especially for multiple segments.
This patch calculate the segment number needed by the packet and
copy th
From: Michael Qiu <qiud...@chinac.com>
When doing packet clone, if packet source is from DPDK driver,
multi-segment must be considered, and copy the segment's
data one by one.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
---
lib/dp-packet.c | 27 ---
1
From: Michael Qiu <qiud...@chinac.com>
When building with DPDK, and using xmalloc() to get a new packet,
field mbuf of the packet will not be initialized, but it's very important for
DPDK port when copying the data to DPDK mbuf, because if ol_flags
and other info are random values, DPDK
to different segments.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijiang Liu <liujiji...@chinac.com>
---
lib/netdev-dpdk.c | 55 ---
1 file changed, 48 insertions(+), 7 deletions(-)
diff --git a/lib/netdev-dpdk.c b/lib/n
to different segments.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijiang Liu <liujiji...@chinac.com>
---
lib/netdev-dpdk.c | 55 ---
1 file changed, 48 insertions(+), 7 deletions(-)
diff --git a/lib/netdev-dpdk.c b/lib/n
2016/10/28 18:50, Kavanagh, Mark B :
When doing packet clone, if packet source is from DPDK driver,
multi-segment must be considered, and copy the segment's
data one by one.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijiang Liu <liujiji...@chinac.com>
---
lib
may hang.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijiang Liu <liujiji...@chinac.com>
Thanks for the patch Michael; one minor nit, but otherwise looks good to me.
---
lib/dp-packet.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/lib/dp-packet.c b/lib
hang.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijiang Liu <liujiji...@chinac.com>
Thanks for the patch Michael; one minor nit, but otherwise looks good to me.
---
lib/dp-packet.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/lib/dp-packet.c b/lib/dp-pa
segments, data_len is not equal to the
Typo in the previous line - 'multipule'
Thanks for point it out, I'll fix it.
packet size. This patch fix this issue.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Marcin Ksiadz <marcinx.ksi...@intel.com>
Signed-off-by: M
2016/10/28 17:47, Kavanagh, Mark B :
Currently, when doing packet copy, lots of DPDK mbuf's info
will be missed, like packet type, ol_flags, etc.
Those information is very important for DPDK to do
packets processing.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijia
When doing packet clone, if packet source is from DPDK driver,
multi-segment must be considered, and copy the segment's
data one by one.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijiang Liu <liujiji...@chinac.com>
---
lib/dp-packet.c | 25 +
Currently, one packet is only copied to one segment
in function dpdk_do_tx_copy(), this could be an issue
when a jumboframe comes, especially for multipile segments.
This patch calculate the segment number needed by the packet and
copy the data to different segments.
Signed-off-by: Michael Qiu
When a packet is from DPDK source, and it contains
multipule segments, data_len is not equal to the
packet size. This patch fix this issue.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Marcin Ksiadz <marcinx.ksi...@intel.com>
Signed-off-by: Mark Kavanagh &
Currently, when doing packet copy, lots of DPDK mbuf's info
will be missed, like packet type, ol_flags, etc.
Those information is very important for DPDK to do
packets processing.
Signed-off-by: Michael Qiu <qiud...@chinac.com>
Signed-off-by: Jijiang Liu <liujiji...@chinac.com>
When building with DPDK, and using xmalloc() to get a new packet,
field mbuf of the packet will not be initialized, but it's very important for
DPDK port when copying the data to DPDK mbuf, because if ol_flags
and other info are random values, DPDK driver may hang.
Signed-off-by: Michael Qiu
-segment mbuf's
support.
This patchset enable DPDK multi-segment mbuf support in OVS.
Michael Qiu (5):
lib/dp-packet: init the mbuf to zero when build with DPDK
DPDK: copy additional packet info when do packet copy
lib/dp-packet: Fix data_len issue with multi-segments
lib/dp-packet: copy multi
在 2016/10/25 1:16, Daniele Di Proietto 写道:
On 24/10/2016 00:31, "Michael Qiu" <08005...@163.com> wrote:
Hi, all
When I self-compiled the OVS-2.6 from commit:
commit 7a0f907b2393626dac1387617355990eab69aef7
Author: Justin Pettit <jpet...@ovn.org>
Date: Tue Sep
Hi, all
When I self-compiled the OVS-2.6 from commit:
commit 7a0f907b2393626dac1387617355990eab69aef7
Author: Justin Pettit
Date: Tue Sep 27 10:13:42 2016 -0700
Set release date for 2.6.0.
Signed-off-by: Justin Pettit
Acked-by: Ben Pfaff
[+cc DPDK jianfeng]
For the gap, is it been fixed by below commit?
commit c59faf3fe8fe6dbd99e8cd9d703839728264c5ae
Author: Jianfeng Tan
Date: Mon Aug 1 03:56:54 2016 +
net/i40e: support TSO on tunneling packet
I just tested your vlan patch, the same thing
Hi, Mark
I have reviewed your RFC patch, I found one place could have issue, see
below:
+if (v > ETHER_MAX_LEN) {
+b->mbuf.data_len =
+(uint16_t)(b->mbuf.buf_len - b->mbuf.data_off);
+} else {
+b->mbuf.data_len = (uint16_t)v;
+}
+
Here, what will
Hi, Mark
OK, once it's ready, pls let me know and I'm glade to help to test it.
BTW, you mentioned the gap, is that the TSO and CSUM of tunnel offload?
I'm familiar with it, and I hope I could do something on it if possible.
2016/10/11 22:34, Kavanagh, Mark B :
Hi, all
This patch is
Hi, all
This patch is very important for users want to improve the performance
of the large packets.
But you know, in data center, lots of networks using vxlan, so if it
supports vxlan, then it will be very useful.
Would you guys has a plan to support it? I would like to help test it,
36 matches
Mail list logo