Hi, Konstantin,
>
>
>> -Original Message-
>> From: outlook_739db8e1c4bc6fae at outlook.com
>> [mailto:outlook_739db8e1c4bc6fae at outlook.com] On Behalf Of Wang Dong
>> Sent: Thursday, April 16, 2015 12:36 PM
>> To: Ananyev, Konstantin; dev at dpdk.org
>> Subject: Re: [dpdk-dev] [PATCH] i
The current implementation of rte_wmb/rte_rmb for x86 is using processor memory
barrier. It's unnessary for IA processor, compiler memory barrier is enough.
But if dpdk runing on a AMD processor, maybe we should use processor memory
barrier.
I add a macro to distinguish them, if we compile DPDK
Hi Dong,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of WangDong
> Sent: Tuesday, May 05, 2015 4:38 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] librte_eal:Using compiler memory barrier for IA
> processor's rte_wmb/rte_rmb.
>
> The current implem
> -Original Message-
> From: Chilikin, Andrey
> Sent: Tuesday, May 05, 2015 4:43 PM
> To: Ananyev, Konstantin; Zhang, Helin; dev at dpdk.org
> Subject: RE: [dpdk-dev] [PATCH RFC 1/6] mbuf: update mbuf structure for QinQ
> support
>
> Hi Helin,
>
> I would agree with Konstantin about ne
Neil and John and anyone else, if I have been rude or ugly in anyway that was
not my intent. Please accept my apologies for being rude or condescending.
Sent from my iPhone
>> On May 5, 2015, at 12:07 PM, Neil Horman wrote:
>>
>> On Tue, May 05, 2015 at 04:43:08PM +, Wiles, Keith wrote:
>>
Hi list
I am doing something about compile DPDK OVS on AWS VM, while I met
some problem,
1. I can only set the hugepagesize to 2M, not to 1G, the reason is the
cpu only has
flag 'pse', which support only 2M huge page size, while not has
pdpe1gb, which support 1G hu
Sent from my iPhone
> On May 5, 2015, at 10:58 AM, John W. Linville
> wrote:
>
>> On Tue, May 05, 2015 at 04:43:08PM +, Wiles, Keith wrote:
>>
>> Please take a breath and relax as we all want the best for DPDK
Hi John
I am sorry you see it this way as I stated in the email parts you sni
Sent from my iPhone
> On May 5, 2015, at 6:56 AM, Neil Horman wrote:
>
>> On Mon, May 04, 2015 at 10:25:00PM -0500, Jim Thompson wrote:
>>
>>> On May 4, 2015, at 10:12 PM, Wiles, Keith wrote:
>>>
>>>
>>>
On 5/4/15, 10:48 AM, "Matthew Hall" wrote:
> On Mon, May 04, 2015 at
On Tue, 5 May 2015 13:39:45 +0800
Cunming Liang wrote:
> The patch does below for igb PF:
> - Setup NIC to generate MSI-X interrupts
> - Set the IVAR register to map interrupt causes to vectors
> - Implement interrupt enable/disable functions
>
> Signed-off-by: Danny Zhou
> Signed-off-by: Cunm
Hi Vladimir,
Why limit Toeplitz hash calculation to predefined tuples and length? Should it
be more general, something like
rte_softrss_be(void *input, uint32_t input_len, const uint8_t *rss_key) to
enable hash calculation for an input of any size? It would be useful for
distributing packets us
> -Original Message-
> From: Stephen Hemminger [mailto:stephen at networkplumber.org]
> Sent: Wednesday, April 29, 2015 6:05 PM
> To: Dumitrescu, Cristian
> Cc: dev at dpdk.org; Stephen Hemminger
> Subject: [PATCH 0/6] rte_sched: patches against 2.o
>
> This is a subset of earlier patche
From: De Lara Guarch, Pablo
Changed name to something more meaningful,
and mark rte_jhash2 as deprecated.
Signed-off-by: Pablo de Lara
---
app/test/test_func_reentrancy.c |2 +-
app/test/test_hash.c|4 ++--
lib/librte_hash/rte_jhash.h | 17 +++--
3 files c
rte_jhash is basically like _rte_jhash_2hashes but it returns only 1 hash,
instead of 2.
In order to remove duplicated code, rte_jhash calls _rte_jhash_2hashes,
passing 0 as the second seed and returning just the first hash value.
(performance penalty is negligible)
The same is done with rte_jhas
With the jhash update, two new functions were introduced:
- rte_jhash_2hashes: Same as rte_jhash, but takes two seeds
and return two hashes (uint32_ts)
- rte_jhash2_2hashes: Same as rte_jhash2, but takes two seeds
and return two hashes (uint32_ts)
Signed
Jenkins hash function was developed originally in 1996,
and was integrated in first versions of DPDK.
The function has been improved in 2006,
achieving up to 60% better performance, compared to the original one.
This patch integrates that code into the rte_jhash library.
Signed-off-by: Pablo de L
Cycles per hash calculation were measured per single operation.
It is much more accurate to run several iterations between measurements
and divide by number of iterations.
Signed-off-by: Pablo de Lara
---
app/test/test_hash_func_perf.c | 18 +-
1 files changed, 9 insertions(+),
This patch moves hash function performance tests to a separate file,
so user can check performance of the existing hash functions quicker,
without having to run all the other hash operation performance tests,
which takes some time.
Signed-off-by: Pablo de Lara
---
app/test/Makefile
Jenkins hash function was developed originally in 1996,
and was integrated in first versions of DPDK.
The function has been improved in 2006,
achieving up to 60% better performance, compared to the original one.
This patchset updates the current jhash in DPDK,
including two new functions that gene
Hi Helin,
I would agree with Konstantin about new naming for VLAN tags. I think we can
leave existing name for t vlan_tci and just name new VLAN tag differently. I
was thinking in the line of "vlan_tci_outer" or "stag_tci". So vlan_tci will
store single VLAN in case if only one L2 tag is presen
This patch depends on the Port Hotplug Framework.
It implements the rte_dev_uninit_t() function for the ring pmd.
Changes in V2:
Fix crash in the rte_pmd_ring_devuninit() function.
Signed-off-by: Bernard Iremonger
---
lib/librte_pmd_ring/rte_eth_ring.c | 92 +++---
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Michal Jastrzebski
> Sent: Thursday, April 30, 2015 1:16 PM
> To: dev at dpdk.org
> Cc: Wodkowski, PawelX
> Subject: [dpdk-dev] [PATCH v2] pipeline: add statistics for librte_pipeline
> ports and tables
>
> Fr
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Michal Jastrzebski
> Sent: Thursday, April 30, 2015 1:14 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v2 00/10] table: added table statistics
>
> From: Maciej Gajdzica
>
> Added statistics for every
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Michal Jastrzebski
> Sent: Thursday, April 30, 2015 1:07 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v2 00/13] port: added port statistics
>
> From: Maciej Gajdzica
>
> Added statistics for every t
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Michal Jastrzebski
> Sent: Thursday, April 30, 2015 1:03 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v2 0/3] port: added frag_ipv6 and ras_ipv6 ports
>
> From: Maciej Gajdzica
>
> Added ipv6 versio
Due to commit c0371da6 in kernel 3.19, which removed msg_iov
and msg_iovlen from struct msghdr, DPDK would not build.
Also, functions memcpy_toiovecend and memcpy_fromiovecend
were removed in commits ba7438ae and 57dd8a07, being substituted by
copy_from_iter and copy_to_iter.
This patch makes use
On Tue, May 05, 2015 at 04:43:08PM +, Wiles, Keith wrote:
>
>
> Sent from my iPhone
>
> > On May 5, 2015, at 6:56 AM, Neil Horman wrote:
> >
> >> On Mon, May 04, 2015 at 10:25:00PM -0500, Jim Thompson wrote:
> >>
> >>> On May 4, 2015, at 10:12 PM, Wiles, Keith
> >>> wrote:
> >>>
> >>>
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Michal Jastrzebski
> Sent: Thursday, April 30, 2015 12:58 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v2 0/3] port: added ethdev_writer_nodrop and
> ring_writer_nodrop ports
>
> From: Maciej Gajdzica
Hi Paul,
> -Original Message-
> From: Paul Emmerich [mailto:emmericp at net.in.tum.de]
> Sent: Tuesday, April 28, 2015 12:48 PM
> To: De Lara Guarch, Pablo
> Cc: Pavel Odintsov; dev at dpdk.org
> Subject: Re: [dpdk-dev] Performance regression in DPDK 1.8/2.0
>
> Hi,
>
>
> De Lara Guarch
On Thu, Apr 23, 2015 at 3:26 PM, Ravi Kerur wrote:
>
>
> On Thu, Apr 23, 2015 at 7:00 AM, Bruce Richardson <
> bruce.richardson at intel.com> wrote:
>
>> On Thu, Apr 23, 2015 at 06:53:44AM -0700, Ravi Kerur wrote:
>> > On Thu, Apr 23, 2015 at 2:23 AM, Ananyev, Konstantin <
>> > konstantin.ananyev
On Tue, May 05, 2015 at 06:30:48PM +, Wiles, Keith wrote:
>
> Sent from my iPhone
>
> > On May 5, 2015, at 10:58 AM, John W. Linville
> > wrote:
> >
> >> On Tue, May 05, 2015 at 04:43:08PM +, Wiles, Keith wrote:
> >>
> >> Please take a breath and relax as we all want the best for DPDK
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Pablo de Lara
> Sent: Tuesday, May 05, 2015 1:35 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH v2] kni: fix compilation issue in KNI vhost on
> kernel 3.19/4.0
>
> Due to commit c0371da6 in kernel 3.19
On Tue, May 05, 2015 at 04:43:08PM +, Wiles, Keith wrote:
> Please take a breath and relax as we all want the best for DPDK.
I cannot believe how rude, asinine, and condascending your attitude
in this thread has been. If this is the future of the DPDK community,
I'm surprised that anyone wo
v7 changes
- decouple epoll event and intr operation
- add condition check in the case intr vector is disabled
- renaming some APIs
v6 changes
- split rte_intr_wait_rx_pkt into two APIs 'wait' and 'set'.
- rewrite rte_intr_rx_wait/rte_intr_rx_set.
- using vector number instead of queue_id as
Fixes: 6065355a "pci: make device id tables const"
Following the above commit, compilation on FreeBSD with clang was broken,
giving the error message:
.../lib/librte_eal/bsdapp/eal/eal_pci.c:438:16: fatal error: assigning to
'struct rte_pci_id *' from 'const struct rte_pci_id *' discards qu
Demonstrate how to handle per rx queue interrupt in a NAPI-like
implementation in usersapce. PDK polling thread mainly works in
polling mode and switch to interrupt mode only if there is no
any packet received in recent polls.
Usersapce interrupt notification generally takes a lot more cycles
than
The patch does below for igb PF:
- Setup NIC to generate MSI-X interrupts
- Set the IVAR register to map interrupt causes to vectors
- Implement interrupt enable/disable functions
Signed-off-by: Danny Zhou
Signed-off-by: Cunming Liang
---
v7 changes
- add condition check when intr vector is not
The patch does below things for ixgbe PF and VF:
- Setup NIC to generate MSI-X interrupts
- Set the IVAR register to map interrupt causes to vectors
- Implement interrupt enable/disable functions
Signed-off-by: Danny Zhou
Signed-off-by: Yong Liu
Signed-off-by: Cunming Liang
---
v7 changes
- ad
The patch adds two dev_ops functions to enable and disable rx queue interrupts.
In addtion, it adds rte_eth_dev_rx_intr_ctl/rx_intr_q to support per port or
per queue rx intr event set.
Signed-off-by: Danny Zhou
Signed-off-by: Cunming Liang
---
v7 changes
- remove rx_intr_vec_get
- add rx_int
This patch does below:
- Create VFIO eventfds for each interrupt vector
- Assign per interrupt vector's eventfd to VFIO by ioctl
Signed-off-by: Danny Zhou
Signed-off-by: Cunming Liang
---
v7 changes
- cleanup unnecessary code change
- split event and intr operation to other patches
lib/lib
Signed-off-by: Danny Zhou
Signed-off-by: Cunming Liang
---
lib/librte_eal/linuxapp/eal/eal_interrupts.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/lib/librte_eal/linuxapp/eal/eal_interrupts.c
b/lib/librte_eal/linuxapp/eal/eal_interrupts.c
index 1090d7b..178a88e 1006
To make bsd compiling happy with new intr changes.
Signed-off-by: Cunming Liang
---
v7 changes
- remove stub 'linux only' function from source file
lib/librte_eal/bsdapp/eal/include/exec-env/rte_interrupts.h | 6 ++
1 file changed, 6 insertions(+)
diff --git a/lib/librte_eal/bsdapp/eal/in
The patch adds 'rte_intr_rx_ctl' to add or delete interrupt vector events
monitor on specified epoll instance.
Signed-off-by: Cunming Liang
---
v7 changes
- rename rte_intr_rx_set to rte_intr_rx_ctl.
- rte_intr_rx_ctl uses rte_epoll_ctl to register epoll event instance.
- the intr rx event in
The patch adds 'rte_epoll_wait' and 'rte_epoll_ctl' for async event wakeup.
It defines 'struct rte_epoll_event' as the event param.
The 'op' uses the same enum as epoll_wait/ctl does.
The epoll event support to carry a raw user data and to register a callback
which is exectuted during wakeup.
Sig
The patch adds interrupt vectors support in rte_intr_handle.
'vec_en' is set when interrupt vectors are detected and associated event fds
are set.
Those event fds are stored in efds[].
'intr_vec' is reserved for device driver to initialize the vector mapping table.
When the event fds add to a spec
v7 changes
- decouple epoll event and intr operation
- add condition check in the case intr vector is disabled
- renaming some APIs
v6 changes
- split rte_intr_wait_rx_pkt into two APIs 'wait' and 'set'.
- rewrite rte_intr_rx_wait/rte_intr_rx_set.
- using vector number instead of queue_id as
Due to commit c0371da6 in kernel 3.19, which removed msg_iov
and msg_iovlen from struct msghdr, DPDK would not build.
Also, functions memcpy_toiovecend and memcpy_fromiovecend
were removed in commits ba7438ae and 57dd8a07, being substituted by
copy_from_iter and copy_to_iter.
This patch makes use
Hi,
On 05/05/2015 11:21 AM, Gonzalez Monroy, Sergio wrote:
> On 04/05/2015 08:55, Olivier MATZ wrote:
>> Hi Sergio,
>>
>> On 04/15/2015 11:30 AM, Sergio Gonzalez Monroy wrote:
>>> This patch introduces a new LDLIBS variable to be set per library.
>>> Its purpose is to especify the library's depend
On Tue, 5 May 2015 13:39:42 +0800
Cunming Liang wrote:
> diff --git a/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c
> b/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c
> index aea1fb1..387f54c 100644
> --- a/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c
> +++ b/lib/librte_eal/linuxapp/eal/eal_pci_vfio.c
>
On Tue, 5 May 2015 13:39:44 +0800
Cunming Liang wrote:
>
> + /* set max interrupt vfio request */
> + if (pci_dev->intr_handle.vec_en) {
> + pci_dev->intr_handle.max_intr = hw->mac.max_rx_queues +
> + IXGBEVF_MAX_OTHER_INTR;
> + pci_dev->intr
On Tue, 5 May 2015 13:39:39 +0800
Cunming Liang wrote:
> static void
> +eal_intr_proc_rxtx_intr(int fd, struct rte_intr_handle *intr_handle)
> +{
Should be const intr_handle is not modified
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Pablo de Lara
> Sent: Monday, May 04, 2015 10:46 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH] kni: fix compilation issue on kernel 3.19
>
> Due to commit c0371da6 in kernel 3.19, which removed msg_io
On 2015-05-05 11:00, Olivier MATZ wrote:
> On 05/05/2015 11:00 AM, Olivier Matz wrote:
>> The script relpath.sh return the relative path of the first directory
>> from the second directory. It is used to generate relative symlinks,
>> which can be useful if the build directory is embedded in the dp
I fixed this issue with very dirty hacks.
I commented mentioned here lines in file
/usr/src/dpdk-2.0.0/x86_64-native-linuxapp-gcc/include/generic/rte_cpuflags.h:
//enum rte_cpu_flag_t;
//enum cpu_register_t;
// static const struct feature_entry cpu_feature_table[];
//static inline int
//rte_cpu_g
Hi Helin,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Helin Zhang
> Sent: Tuesday, May 05, 2015 3:32 AM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [PATCH RFC 1/6] mbuf: update mbuf structure for QinQ
> support
>
> To support QinQ, 'vlan_tci' should be
Hi Simon,
On 05/05/2015 11:00 AM, Olivier Matz wrote:
> The script relpath.sh return the relative path of the first directory
> from the second directory. It is used to generate relative symlinks,
> which can be useful if the build directory is embedded in the dpdk
> directory: the whole dpdk can
The script relpath.sh return the relative path of the first directory
from the second directory. It is used to generate relative symlinks,
which can be useful if the build directory is embedded in the dpdk
directory: the whole dpdk can be moved without breaking the links,
which is helpful for an in
Hello!
Could anybody help me with this issue? :(
In this file widely used enum forward declarations which completely
incompatible with C++ and need some rewrite.
On Wed, Apr 29, 2015 at 3:17 PM, Pavel Odintsov
wrote:
> Hello!
>
> I have C++ application compiles and works nice. But when I includ
As QinQ stripping and insertion have been supported, test commands
should be updated. In detail, "tx_vlan set vlan_id (port_id)" will
be changed to "tx_vlan set (port_id) vlan_id0[, vlan_id1]" to
support both single and double VLAN tag insertion; also VLAN tags
stripped from received packets will b
As hardware supports QinQ stripping and insertion, the offload
flags of them should be added in both PF and VF sides.
Signed-off-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_ethdev.c| 2 ++
lib/librte_pmd_i40e/i40e_ethdev_vf.c | 13 +
2 files changed, 15 insertions(+)
diff --gi
As offload capabilities of QinQ stripping and insertion are
supported by some of the supported hardware, the offload capability
flags should be added accordingly.
Signed-off-by: Helin Zhang
---
lib/librte_ether/rte_ethdev.h | 28 +++-
1 file changed, 15 insertions(+), 13
To support QinQ stripping and insertion, QinQ L2 tags should be
extracted from RX descriptors and stored in mbuf for RX stripping,
and should be read from mbuf and set correspondingly in TX
descriptors.
Signed-off-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_rxtx.c | 83 +
Reconfiguration is needed to support QinQ stripping and insertion,
as hardware does not support them by default.
Signed-off-by: Helin Zhang
---
lib/librte_pmd_i40e/i40e_ethdev.c | 48 +++
1 file changed, 48 insertions(+)
diff --git a/lib/librte_pmd_i40e/i40e_
To support QinQ, 'vlan_tci' should be replaced by 'vlan_tci0' and
'vlan_tci1'. Also new offload flags of 'PKT_RX_QINQ_PKT' and
'PKT_TX_QINQ_PKT' should be added.
Signed-off-by: Helin Zhang
---
app/test-pmd/flowgen.c| 2 +-
app/test-pmd/macfwd.c | 2 +-
app/test-
As i40e hardware can be reconfigured to support QinQ stripping and
insertion, this patch set is to enable that with the update of
'struct rte_mbuf', and testpmd commands.
Note that the Vector-PMD will be updated soon later.
Helin Zhang (6):
mbuf: update mbuf structure for QinQ support
i40e: re
On 04/05/2015 08:55, Olivier MATZ wrote:
> Hi Sergio,
>
> On 04/15/2015 11:30 AM, Sergio Gonzalez Monroy wrote:
>> This patch introduces a new LDLIBS variable to be set per library.
>> Its purpose is to especify the library's dependent libraries to
>> be explicitly linked against.
>>
>> Given the c
On Mon, May 04, 2015 at 10:25:00PM -0500, Jim Thompson wrote:
>
> > On May 4, 2015, at 10:12 PM, Wiles, Keith wrote:
> >
> >
> >
> > On 5/4/15, 10:48 AM, "Matthew Hall" wrote:
> >
> >> On Mon, May 04, 2015 at 12:43:48PM +, Qiu, Michael wrote:
> >>> What mail client do you use? I think m
Hi Keith,
On 05/04/2015 06:50 PM, Wiles, Keith wrote:
>
>
> Sent from my iPhone
>
>> On May 4, 2015, at 8:27 AM, Olivier MATZ wrote:
>>
>>
>>
>>> On 05/04/2015 04:36 PM, Wiles, Keith wrote:
>>>
>>>
On 5/4/15, 2:18 AM, "Olivier MATZ" wrote:
Hi Keith,
> On 05/01/2015 04:22
Software implementation of the Toeplitz hash function used by RSS.
Can be used either for packet distribution on single queue NIC
or for simulating of RSS computation on specific NIC (for example
after GRE header decapsulating).
v2 changes
- Add ipv6 support
- Various style fixes
Signed-off-by: V
On 5/4/15, 10:48 AM, "Matthew Hall" wrote:
>On Mon, May 04, 2015 at 12:43:48PM +, Qiu, Michael wrote:
>> What mail client do you use? I think mail client supporting thread mode
>> is important for patch review.
>
>Like many UNIX people, I use mutt.
>
>My concern is that, if we're making th
Hi Marc
On 5/4/15, 2:08 PM, "Marc Sune" wrote:
>
>
>On 02/05/15 15:59, Wiles, Keith wrote:
>>
>> On 5/2/15, 6:40 AM, "Neil Horman" wrote:
>>
>>> On Fri, May 01, 2015 at 01:36:58PM -0700, Matthew Hall wrote:
On Fri, May 01, 2015 at 10:59:32PM +0300, Aaro Koskinen wrote:
> Projects like
On 5/4/15, 1:34 PM, "Marc Sune" wrote:
>
>
>On 01/05/15 20:17, Wiles, Keith wrote:
>>
>> On 5/1/15, 1:09 PM, "Stephen Hemminger"
>>wrote:
>>
>>> On Fri, 1 May 2015 15:56:32 +
>>> "Wiles, Keith" wrote:
>>>
Hi Everyone,
I believe the DPDK community would benefit from moving t
71 matches
Mail list logo
