On Fri, Sep 26, 2014 at 03:02:55PM -0700, Stephen Hemminger wrote:
> On Fri, 26 Sep 2014 10:45:49 -0400
> Neil Horman wrote:
>
> > On Fri, Sep 26, 2014 at 12:41:33PM +0200, Thomas Monjalon wrote:
> > > Hi Neil,
> > >
> > > 2014-09-24 14:19, Neil Horman:
> > > > Ping Thomas. I know you're busy,
> > As I remember the purpose of the patch was to fix the race condition inside
> > rte_alarm library.
> > I believe that the patch provided by Michal & Pawel fixes the issues you
> > discovered.
> > If you think, that is not the case, could you please provide a list of
> > remaining issues?
Signed-off-by: Huawei Xie
---
examples/vhost/Makefile | 52 +
1 file changed, 52 insertions(+)
create mode 100644 examples/vhost/Makefile
diff --git a/examples/vhost/Makefile b/examples/vhost/Makefile
new file mode 100644
index 000..a4d4fb0
The mbuf changes include:
1. flattened structure vlan_macip
2. removed rte_pktmbuf structure.
3. mbuf data pointer replaced by an offset
Other changes include:
1. fix sg mbuf xmit in virtio_tx_route
2. rename RTE_MBUF_SCATTER_GATHER to RTE_MBUF_REFCNT
3. add one TODO and FIXME
Signed-off-by:
This vhost example demonstrates how to integrate user space vhost
with DPDK accelerated ethernet vSwitch.
Signed-off-by: Huawei Xie
---
examples/vhost/main.c | 1455 +
examples/vhost/main.h | 47 +-
2 files changed, 431 insertions(+), 1071
copy old vhost example source files without any modification.
The subsequent patch will modify it to use new vhost lib API.
Signed-off-by: Huawei Xie
---
examples/vhost/main.c | 3722 +
examples/vhost/main.h | 86 ++
2 files changed, 3808
vhost lib is turned off by default as it requires fuse-devel package.
fuse-devel isn't installed in every linux distribution.
fuse-devel enables user space filesystem/char driver development.
vhost lib contains a user space char driver, which replies on this package.
Signed-off-by: Huawei Xie
fixes checkpatch reported issues.
Signed-off-by: Huawei Xie
---
lib/librte_vhost/rte_virtio_net.h | 55 +
lib/librte_vhost/vhost-net-cdev.c | 236 +++---
lib/librte_vhost/vhost-net-cdev.h | 33 +++---
lib/librte_vhost/vhost_rxtx.c | 57 +
There is no rte_pktmbuf structure in mbuf now. Its fields are merged to
rte_mbuf structure.
Signed-off-by: Huawei Xie
---
lib/librte_vhost/vhost_rxtx.c | 24
1 file changed, 12 insertions(+), 12 deletions(-)
diff --git a/lib/librte_vhost/vhost_rxtx.c
Merge vhost merge-able rx.
For vhost tx, previous vhost merge-able feature introduces virtio_dev_merge_tx,
and calls virtio_dev_tx and vritio_dev_merge_tx respectively depends on whether
the vhost device supports merge-able feature.
There is no so called merge-tx, it is actually fix for memcpy
This vhost lib consists of five APIs plus several other helper routines
for feature disable/enable.
1) rte_vhost_driver_register initialises vhost driver.
2) rte_vhost_driver_callback_register registers the callbacks.
Callbacks are called from vhost driver when virtio device is ready
for polling
This patch does simple split of the original vhost example source
files in vhost lib directory.
vhost rx/tx functions virtio_dev_rx/tx are copied from main.c to
new file vhost_rxtx.c and license header is added.
main.c and main.h are removed and will be copied to new vhost
example in subsequent
"git mv examples/vhost lib/librte_vhost"
This is a purely src file move, without any modification.
Subsequent patch will transform those src files to a vhost library.
Signed-off-by: Huawei Xie
---
examples/vhost/Makefile | 60 -
examples/vhost/eventfd_link/Makefile
This set of patches transforms and refactors vhost example to a user
space vhost library and a new vhost example based on this library.
This library implements a user space vhost cuse driver, and provides
generic APIs for user space ethernet vSwitch to integrate us-vhost for
fast packet switching
Signed-off-by: Daniel Mrzyglod
---
app/test/test_link_bonding.c | 501 +++-
app/test/virtual_pmd.c |6 +-
app/test/virtual_pmd.h |7 +
lib/librte_pmd_bond/rte_eth_bond.h | 23 ++
It really depends on the devices offered in the VM. If direct device assignment
is not provided to a VM or if the node hypervisor doesn't have an optimized
para-virtual interface to a VM, I don't see any benefit using DPDK in VMs.
-Original Message-
From: dev
The function rte_snprintf() was deprecated in version 1.7.0
(commit 6f41fe75e2dd).
It's now totally removed.
Signed-off-by: Thomas Monjalon
---
app/test/Makefile | 7 --
app/test/test_string_fns.c | 136 +
On Fri, Sep 26, 2014 at 11:16:30AM -0400, Neil Horman wrote:
> On Fri, Sep 26, 2014 at 03:16:08PM +0100, Sergio Gonzalez Monroy wrote:
> > On Thu, Sep 25, 2014 at 02:52:32PM -0400, Neil Horman wrote:
> > > Add initial pass header files to support symbol versioning.
> > >
> > > ---
> > > Change
From: David Marchand
Signed-off-by: David Marchand
---
examples/cmdline/main.c| 3 +-
examples/cmdline/main.h| 45 -
examples/dpdk_qat/main.c | 3 +-
From: David Marchand
Signed-off-by: David Marchand
Signed-off-by: Thomas Monjalon
---
app/cmdline_test/cmdline_test.h | 7 --
app/dump_cfg/main.c | 4 +---
app/dump_cfg/main.h | 45 -
From: David Marchand
Signed-off-by: David Marchand
Signed-off-by: Thomas Monjalon
---
lib/Makefile| 1 -
lib/librte_eal/Makefile | 2 --
lib/librte_eal/common/Makefile | 3 ---
From: David Marchand
Signed-off-by: David Marchand
Signed-off-by: Thomas Monjalon
---
mk/exec-env/bsdapp/rte.vars.mk | 2 +-
mk/exec-env/linuxapp/rte.vars.mk | 2 +-
mk/rte.app.mk| 9 -
mk/rte.sdkroot.mk| 2 +-
From: David Marchand
Signed-off-by: David Marchand
Signed-off-by: Thomas Monjalon
---
config/common_bsdapp | 7 +--
config/common_linuxapp | 32 +---
2 files changed, 2 insertions(+), 37 deletions(-)
diff --git
There are some references to bare metal (i.e. without OS) support,
especially some options to build a libc with DPDK.
As there are currently no such support, it can be removed.
Some comments are cleaned in the same time.
Thanks to David for having done most of this effort:
config: no more bare
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of De Lara Guarch, Pablo
> Sent: Friday, September 26, 2014 4:12 PM
> To: Pattan, Reshma; dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2] distributor_app: new sample app
>
> Hi,
>
> > -Original
On Fri, Sep 26, 2014 at 2:31 PM, Thomas Monjalon
wrote:
> Since commit a155d430119 ("support link bonding device initialization"),
> rte_eal_pci_probe() is called in rte_eal_init().
> So it doesn't have to be called by application anymore.
> It has been fixed for testpmd in commit 2950a769315,
>
On Fri, Sep 26, 2014 at 06:07:14PM +, Ananyev, Konstantin wrote:
>
>
> > > As I remember the purpose of the patch was to fix the race condition
> > > inside rte_alarm library.
> > > I believe that the patch provided by Michal & Pawel fixes the issues you
> > > discovered.
> > > If you
On Fri, Sep 26, 2014 at 12:22:56PM -0400, Neil Horman wrote:
> On Fri, Sep 26, 2014 at 04:33:04PM +0100, Sergio Gonzalez Monroy wrote:
> > On Fri, Sep 26, 2014 at 11:16:30AM -0400, Neil Horman wrote:
> > > On Fri, Sep 26, 2014 at 03:16:08PM +0100, Sergio Gonzalez Monroy wrote:
> > > > On Thu, Sep
For apps that were using default rte_eth_rxconf and rte_eth_txconf
structures, these have been removed and now they are obtained by
calling rte_eth_rxconf_defaults and rte_eth_txconf_defaults, just
before setting up RX/TX queues.
Signed-off-by: Pablo de Lara
---
examples/dpdk_qat/main.c
Many sample apps use duplicated code to set rte_eth_txconf and rte_eth_rxconf
structures. This patch allows the user to get a default optimal RX/TX
configuration
through these two functions, and still any parameters may be tweaked as wished,
before setting up queues.
Signed-off-by: Pablo de Lara
These patches add two new API functions to get an optimal values
for the RX/TX configuration structures (rte_eth_rxconf and rte_eth_txconf),
so users can get these configurations and modify or use them directly,
to set up RX/TX queues. Besides, most of the apps that were modifying little
or none
On Thu, Sep 25, 2014 at 02:52:32PM -0400, Neil Horman wrote:
> Add initial pass header files to support symbol versioning.
>
> ---
> Change notes
> v2)
> * Fixed ifdef in rte_compat.h to test for RTE_BUILD_SHARED_LIB instead of the
> non-existant RTE_SYMBOL_VERSIONING
>
> * Fixed VERSION_SYMBOL
Hi,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of reshmapa
> Sent: Wednesday, September 24, 2014 3:17 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v2] distributor_app: new sample app
>
> From: Reshma Pattan
>
> A new sample app that shows the
On Fri, 26 Sep 2014 10:45:49 -0400
Neil Horman wrote:
> On Fri, Sep 26, 2014 at 12:41:33PM +0200, Thomas Monjalon wrote:
> > Hi Neil,
> >
> > 2014-09-24 14:19, Neil Horman:
> > > Ping Thomas. I know you're busy, but I would like this to not fall off
> > > anyones
> > > radar. You alluded to
Hi Shimamoto-san,
There are a lot of sighting relate to "DMAR:[fault reason 06] PTE Read access
is not set"
https://www.mail-archive.com/kvm at vger.kernel.org/msg106573.html
This might be related to IOMMU, and kernel code.
Here is what we know :-
1) Disabling VT-d in bios also removed the
No need to restrict usage of non Intel SFP.
If (hw->phy.type == ixgbe_phy_sfp_intel) is false,
a warning will be logged.
It was disabled for ixgbe and enabled but unused for i40e.
Signed-off-by: Thomas Monjalon
---
config/common_bsdapp| 2 --
config/common_linuxapp
Since commit a155d430119 ("support link bonding device initialization"),
rte_eal_pci_probe() is called in rte_eal_init().
So it doesn't have to be called by application anymore.
It has been fixed for testpmd in commit 2950a769315,
and this patch remove it from other applications.
Signed-off-by:
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Neil Horman
> Sent: Friday, September 26, 2014 2:40 PM
> To: Wodkowski, PawelX
> Cc: dev at dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v2] Change alarm cancel function to
> thread-safe:
>
> On Fri, Sep 26, 2014 at 12:37:54PM +,
add test command to configure flexible masks for each flow type
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
app/test-pmd/cmdline.c | 173 +
1 file changed, 173 insertions(+)
diff --git
implement operation to flexible masks for each flow type in i40e pmd driver
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_fdir.c | 121
1 file changed, 121 insertions(+)
diff --git
define structures for configuring flexible masks
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_ether/rte_eth_ctrl.h | 24
1 file changed, 24 insertions(+)
diff --git a/lib/librte_ether/rte_eth_ctrl.h
add test command to configure flexible payload
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
app/test-pmd/cmdline.c | 144 +
1 file changed, 144 insertions(+)
diff --git a/app/test-pmd/cmdline.c
define structures for configuring flexible payload
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_ether/rte_eth_ctrl.h | 42 +
1 file changed, 42 insertions(+)
diff --git a/lib/librte_ether/rte_eth_ctrl.h
add test command to flush flow director table
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
app/test-pmd/cmdline.c | 49 +
1 file changed, 49 insertions(+)
diff --git a/app/test-pmd/cmdline.c
implement operation to flush flow director table
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_fdir.c | 43 +
1 file changed, 43 insertions(+)
diff --git a/lib/librte_pmd_i40e/i40e_fdir.c
display flow director's statistics information
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
app/test-pmd/config.c | 40 +++-
1 file changed, 31 insertions(+), 9 deletions(-)
diff --git a/app/test-pmd/config.c
implement operation to get flow director information in i40e pmd driver
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_fdir.c | 34 ++
1 file changed, 34 insertions(+)
diff --git
define structures for getting flow director information
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_ether/rte_eth_ctrl.h | 40
lib/librte_ether/rte_ethdev.h | 23 ---
2 files
support to set the FDIR flag and report FD_ID in mbuf if match
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_rxtx.c | 48 +
1 file changed, 48 insertions(+)
diff --git
support to get the fdir_match counter
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_ethdev.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/lib/librte_pmd_i40e/i40e_ethdev.c
b/lib/librte_pmd_i40e/i40e_ethdev.c
index
add commands which can be used to test adding or deleting 8 flow types of the
flow director filters: ipv4, tcpv4, udpv4, sctpv4, ipv6, tcpv6, udpv6, sctpv6
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
app/test-pmd/cmdline.c | 447
deal with two operations for flow director
- RTE_ETH_FILTER_OP_ADD
- RTE_ETH_FILTER_OP_DELETE
encode the flow inputs to programming packet
sent the packet to filter programming queue and check status on the status
report queue
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by:
define structures to add or delete flow director filter
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_ether/rte_eth_ctrl.h | 159
1 file changed, 159 insertions(+)
diff --git
Define new APIs to support configure multi-kind filters using same APIs
- rte_eth_dev_filter_supported
- rte_eth_dev_filter_ctrl
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_ether/Makefile | 1 +
lib/librte_ether/rte_eth_ctrl.h | 78
set flexible payload related registers to default value at initialization time.
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_ethdev.c | 33 ++
lib/librte_pmd_i40e/i40e_fdir.c | 49
release fortville resources on flow director, includes
- queue 0 pair release
- release vsi
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_ethdev.c | 4 +++-
lib/librte_pmd_i40e/i40e_ethdev.h | 1 +
lib/librte_pmd_i40e/i40e_fdir.c
set up fortville resources to support flow director, includes
- queue 0 pair allocated and set up for flow director
- create vsi
- reserve memzone for flow director programming packet
Signed-off-by: Jingjing Wu
Acked-by: Chen Jing D(Mark)
Acked-by: Helin Zhang
---
The patch set supports flow director on fortville.
It includes:
- set up/tear down fortville resources to support flow director, such as queue
and vsi.
- define new APIs to support multi-kind filters and their operations.
- support operation to add or delete 8 flow types of the flow director
> > > Maybe I don't see something obvious? :)
>
> I think you're missing the fact that your patch doesn't do what you assert
> above
> either :)
Issue is not in setting alarms but canceling it. If you look closer to my patch
you
see that it address this issue (look at added *do { lock(); ;
Hi, all,
I found ixgbe in couple with rte_mbuf (and probably other PMD drivers)
don't support GSO, I reverse engineered the linux kernel's ixgbe's gso
support and got it working in 1.6. Could it be useful to provide the patch?
Regards,
Vadim.
V3 change:
- Rename field name to rx_deferred_start/tx_deferred_start in
both ixgbe and i40e PMD.
- Move the doxygen comments for rx_deferred_start after it is declared.
- Simplify/split the long description and move some to doxygen comments of
rte_eth_dev_rx_queue_start and
A customer plan to buy some global Amazon VMs to run their DPDK 1.3(will
upgrade to DPDK1.6 or 1.7) based VPN applications on global sites.
Thanks a lot;
Hi Neil,
2014-09-24 14:19, Neil Horman:
> Ping Thomas. I know you're busy, but I would like this to not fall off anyones
> radar. You alluded to concerns regarding what, for lack of a better term,
> ABI/API lockin. I had asked you to enuumerate/elaborate on specifics, but
> never
> heard back.
> So basically cancel() just set ALARM_CANCELLED and leaves actual alarm
> deletion to the callback()?
> That was the thought, yes.
>
> > I think it is doable - but I don't see any real advantage with that
> > approach.
> > Yes, code will become a bit simpler, as we'll have one point when we
On Fri, Sep 26, 2014 at 03:41:58PM +, Ananyev, Konstantin wrote:
>
>
> > -Original Message-
> > From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Neil Horman
> > Sent: Friday, September 26, 2014 4:02 PM
> > To: Wodkowski, PawelX
> > Cc: dev at dpdk.org
> > Subject: Re:
2014-09-16 16:16, Neil Horman:
> On Fri, Sep 12, 2014 at 02:05:23PM -0400, John W. Linville wrote:
> > Ping? Are there objections to this patch from mid-July?
>
> Thomas, Where are you on this? It seems like if you don't have any objections
> to this patch, it should go in, in ilght of the lack
2014-09-25 23:29, Ananyev, Konstantin:
> From: Thomas Monjalon
> > 2014-09-25 13:07, Cao, Waterman:
> > > I will work with team to see if we can improve test report.
> > > Because intel validation team will continue to upgrade test cases to
> > > verify feature,
> > > I think that it's still
On Fri, Sep 26, 2014 at 03:16:08PM +0100, Sergio Gonzalez Monroy wrote:
> On Thu, Sep 25, 2014 at 02:52:32PM -0400, Neil Horman wrote:
> > Add initial pass header files to support symbol versioning.
> >
> > ---
> > Change notes
> > v2)
> > * Fixed ifdef in rte_compat.h to test for
On Fri, Sep 26, 2014 at 04:39:44PM +0200, Thomas Monjalon wrote:
> The function rte_snprintf() was deprecated in version 1.7.0
> (commit 6f41fe75e2dd).
> It's now totally removed.
>
> Signed-off-by: Thomas Monjalon
Acked-by: Neil Horman
>
On Fri, Sep 26, 2014 at 02:01:05PM +, Wodkowski, PawelX wrote:
> > > > Maybe I don't see something obvious? :)
> >
> > I think you're missing the fact that your patch doesn't do what you assert
> > above
> > either :)
>
> Issue is not in setting alarms but canceling it. If you look closer
On Fri, Sep 26, 2014 at 12:41:33PM +0200, Thomas Monjalon wrote:
> Hi Neil,
>
> 2014-09-24 14:19, Neil Horman:
> > Ping Thomas. I know you're busy, but I would like this to not fall off
> > anyones
> > radar. You alluded to concerns regarding what, for lack of a better term,
> > ABI/API lockin.
On Fri, Sep 26, 2014 at 04:03:57PM +0200, Thomas Monjalon wrote:
> There are some references to bare metal (i.e. without OS) support,
> especially some options to build a libc with DPDK.
> As there are currently no such support, it can be removed.
> Some comments are cleaned in the same time.
>
>
On Fri, Sep 26, 2014 at 02:31:24PM +0200, Thomas Monjalon wrote:
> Since commit a155d430119 ("support link bonding device initialization"),
> rte_eal_pci_probe() is called in rte_eal_init().
> So it doesn't have to be called by application anymore.
> It has been fixed for testpmd in commit
On Fri, Sep 26, 2014 at 11:28:05AM +0200, Thomas Monjalon wrote:
> 2014-09-16 16:16, Neil Horman:
> > On Fri, Sep 12, 2014 at 02:05:23PM -0400, John W. Linville wrote:
> > > Ping? Are there objections to this patch from mid-July?
> >
> > Thomas, Where are you on this? It seems like if you don't
Add test cases in testpmd to test VxLAN Tx Checksum offload, which include
- IPv4 tunnel and IPv6 tunnel
- outer L3, inner L3 and L4 checksum offload for Tx side.
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by: Jingjing Wu
Acked-by: Jing Chen
---
app/test-pmd/config.c |6
Support VxLAN Tx checksum offload, which include
- outer L3(IP) checksum offload
- inner L3(IP) checksum offload
- inner L4(UDP, TCP and SCTP) checksum offload
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by: Jingjing Wu
Acked-by: Jing Chen
---
lib/librte_mbuf/rte_mbuf.h
Add tunnel_filter command in testpmd to test VxLAN packet filter API.
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by: Jingjing Wu
Acked-by: Jing Chen
---
app/test-pmd/cmdline.c | 152
1 files changed, 152 insertions(+), 0
The implementation of VxLAN tunnel filter in librte_pmd_i40e, which include
- add the i40e_dev_filter_ctrl() function.
- add the i40e_dev_tunnel_filter_set() function.
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by: Jingjing Wu
Acked-by: Jing Chen
---
Introduce a new filter framewok in librte_ether. As to the implemetation
discussion, please refer to
http://dpdk.org/ml/archives/dev/2014-September/005179.html, and VxLAN tunnel
filter implementation is based on
it.
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by: Jingjing Wu
---
Add commands to test VxLAN packet identification, which include
- use commands to add/delete VxLAN UDP port.
- use rxonly mode to receive VxLAN packet.
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by: Jingjing Wu
Acked-by: Jing Chen
---
app/test-pmd/cmdline.c| 78
Support tunneling UDP port configuration on i40e in librte_pmd_i40e.
Currently, only VxLAN is implemented, which include
- VxLAN UDP port initialization
- Implement the APIs to configure VxLAN UDP port in librte_pmd_i40e.
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by: Jingjing
Add data structures and APIs in librte_ether for supporting tunneling UDP port
configuration on i40e,
Currently, only VxLAN is implemented, which include
- VxLAN UDP port initialization
- Add APIs to configure VxLAN UDP port
Signed-off-by: Jijiang Liu
Acked-by: Helin Zhang
Acked-by:
The patch set supports VxLAN on Fortville based on current mbuf structure. When
Bruce's Mbuf Structure Rework(part 3) is applied, there will be minor changes
later.
It includes:
- Support VxLAN packet identification by configuring tunneling UDP port.
- Support VxLAN packet filters. It uses
Since the introduction of vector PMD, a bug in ixgbe_rxq_rearm could
cause a crash. As long as the memory pool allocated to the RX queue
has mbufs available, there is no problem. After allocation of _all_
mbufs from the memory pool, previously returned mbufs by
rte_eth_rx_burst could be accessed
> >
> > diff --git a/lib/librte_eal/linuxapp/eal/eal_alarm.c
> > b/lib/librte_eal/linuxapp/eal/eal_alarm.c
> > index 480f0cb..73b6dc5 100644
> > --- a/lib/librte_eal/linuxapp/eal/eal_alarm.c
> > +++ b/lib/librte_eal/linuxapp/eal/eal_alarm.c
> > @@ -64,6 +64,9 @@
> > #define MS_PER_S 1000
> >
On Fri, Sep 26, 2014 at 06:33:12AM +, Wodkowski, PawelX wrote:
> > Given what you said above, I agree, at least in the current implementation.
> > It
> > still seems like theres a simpler solution that doesn't require all the
> > comparative gymnastics.
>
> Yes, there is simpler solution,
On Fri, Sep 26, 2014 at 12:37:54PM +, Wodkowski, PawelX wrote:
> > So basically cancel() just set ALARM_CANCELLED and leaves actual alarm
> > deletion to the callback()?
> > That was the thought, yes.
> >
> > > I think it is doable - but I don't see any real advantage with that
> > >
During initialization of rte_table_hash_ext and rte_table_hash_lru, a
contiguous region of memory is allocated to store meta data, buckets,
extended buckets, keys, stack of keys, stack of extended buckets and
data entries. The size of each region depends on the hash table
configuration.
The
During initialization of rte_hash_table_ext and rte_hash_table_lru,
t->data_size_shl is calculated. This member contains the number of
bits to shift left during calculation of the location of entries in
the hash table. To determine the number of bits to shift left, the
size of the entry (as
If a key is not found in a bucket and the bucket has been extended,
the extended buckets also have to checked for potentially matching
keys. The extended buckets are checked at the end of the lookup. In
most cases, this logic is skipped as it is uncommon to have buckets in
an extended state.
In
When an entry is deleted from an extensible rte_table_hash, the bucket
that stored the entry can become empty. If this is the case, the
bucket needs to be removed from the chain of buckets.
During removal of the bucket, the chain should be updated first. If
the bucket that will be removed is
This set of patches fixes bugs in the packet framework. Some of the
bugs occur in corner cases (i.e. when a lookup is performed on a few
packets or when buckets are in extended states) while others can cause
memory to be accessed beyond what is reserved during initialization
time.
Balazs Nemeth
I encountered an issue that DPDK doesn't work with "iommu=pt intel_iommu=on"
on HP ProLiant DL380p Gen8 server. I'm using the following environment;
HW: ProLiant DL380p Gen8
CPU: E5-2697 v2
OS: RHEL7
kernel: kernel-3.10.0-123 and the latest kernel 3.17-rc6+
DPDK: v1.7.1-53-gce5abac
On Thu, Sep 25, 2014 at 11:24:30PM +, Ananyev, Konstantin wrote:
> > From: Neil Horman [mailto:nhorman at tuxdriver.com]
> > Sent: Thursday, September 25, 2014 6:24 PM
> > To: Ananyev, Konstantin
> > Cc: Jastrzebski, MichalX K; dev at dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH v2] Change
> Given what you said above, I agree, at least in the current implementation.
> It
> still seems like theres a simpler solution that doesn't require all the
> comparative gymnastics.
Yes, there is simpler solution, but this solution involve recursive locking.
DPDK recursive spinlocks are no an
IBM Power architecture has different memory architecture with x86. When
the physical memory address is in ascending order, the mmaped virtual
address is in descending order. This patch modified the memory segment
detection code to make it work for Power.
Signed-off-by: Chao Zhu
---
IBM Power architecture has different huge page sizes (16MB, 16GB) than
x86.This patch inserts RTE_PGSIZE_16M and RTE_PGSIZE_16G to the
rte_page_sizes enum variable and adds huge page size support of DPDK for
IBM Power architecture.
Signed-off-by: Chao Zhu
---
app/test/test_memzone.c
IBM Power architecture has different cache line size (128 bytes) than
x86 (64 bytes). This patch defines CACHE_LINE_SIZE to 128 bytes to
override the default value 64 bytes to support IBM Power Architecture.
Signed-off-by: Chao Zhu
---
app/test/test_malloc.c |8
iopl() call is mostly for the i386 architecture. In Power architecture.
It doesn't exist. This patch modified rte_eal_iopl_init() and make it
return -1 on Power. This means rte_config.flags will not contain
EAL_FLG_HIGH_IOPL flag on IBM Power architecture.
Signed-off-by: Chao Zhu
---
Intel processors contain registers to identify CPU flags, DPDK reads the
registers to get the CPU flags. IBM Power processor doesn't have such
registers. This patch uses aux vector software register to get CPU flags
and add CPU flag checking support for IBM Power architecture.
Signed-off-by: Chao
1 - 100 of 119 matches
Mail list logo