I'm hoping that someone (perhaps at Intel) can help us understand
an IXGBE RX packet loss issue we're able to reproduce with testpmd.
We run testpmd with various numbers of cores. We offer line-rate
traffic (~14.88 Mpps) to one ethernet port, and forward all received
packets via the second port.
Hi, all
Any comments on this?
Thanks,
Michael
On 2015/9/25 10:56, Qiu, Michael wrote:
> On 2015/9/7 22:46, Thomas Monjalon wrote:
>> 2015-08-24 17:22, Michael Qiu:
>>> For __SSE3__, the corresponding header file should be pmmintrin.h,
>>> tmmintrin.h works for __SSSE3__.
>> Please could you
This patch adds a workaround to drop flow control frames from being
transmitted from VSIs.
With this patch in place a malicious VF cannot send flow control or PFC
packets out on the wire.
Signed-off-by: Jingjing Wu
---
v2:
- reword comments
drivers/net/i40e/i40e_ethdev.c | 30
It supports both Admin queue based and directly writing registers
based RSS hash key and lookup table configuration, as X722 supports
AQ based configuration.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/i40e_ethdev.c | 229 ++---
In order to provide users early access of X722 and its A0 hardware,
new device IDs are added, and also compilation with those support
in base driver is enabled.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/Makefile | 1 +
lib/librte_eal/common/include/rte_pci_dev_ids.h
It supports both Admin queue based and directly writing registers
based RSS hash key and lookup table configuration, as X722 supports
AQ based configuration.
Signed-off-by: Helin Zhang
---
drivers/net/i40e/i40e_ethdev.h| 3 +
drivers/net/i40e/i40e_ethdev_vf.c | 230
Macro MAX_QUEUES was defined to 128, only allow 16 mem_pools in theory.
When running vmdq_app with more than 34 mem_pools,
it will cause the core_dump issue.
Change MAX_QUEUES to 1024 will solve this issue.
Signed-off-by: Xutao Sun
---
examples/vmdq/main.c | 2 +-
1 file changed, 1
New functions prototypes for bulk add/delete added to table API. New
functions allows adding/deleting multiple records with single function
call. For now those functions are implemented only for ACL table. For
other tables these function pointers are set to NULL.
Signed-off-by: Maciej Gajdzica
Added functions for adding/deleting multiple records to table owned by
pipeline.
Signed-off-by: Maciej Gajdzica
Signed-off-by: Marcin Kerlin
Acked-by: Cristian Dumitrescu
---
lib/librte_pipeline/rte_pipeline.c | 106 +
lib/librte_pipeline/rte_pipeline.h |
Added to acl table unit test check for bulk add and bulk delete.
Signed-off-by: Maciej Gajdzica
Acked-by: Cristian Dumitrescu
---
app/test/test_table_acl.c | 166 ++
1 file changed, 166 insertions(+)
diff --git a/app/test/test_table_acl.c
Added two new cli commands to firewall pipeline. Commands bulk add and
bulk delete takes as argument a file with rules to add/delete. The file
is parsed, and then rules are passed to backend functions which
add/delete records from pipeline tables.
Signed-off-by: Maciej Gajdzica
Acked-by:
The LIBABIVER number is incremented for table and pipeline libraries.
The release notes is updated and the deprecation announce is removed.
Signed-off-by: Maciej Gajdzica
Acked-by: Cristian Dumitrescu
---
doc/guides/rel_notes/deprecation.rst | 3 ---
Hi Avinash,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Yeddula, Avinash
> Sent: Monday, October 12, 2015 6:03 PM
> To: Dumitrescu, Cristian; dev at dpdk.org; Bly, Mike
> Subject: Re: [dpdk-dev] DPDK hash function related question
>
> Hi Cristian,
> I
On 2015/10/12 14:38, Wang Xiao W wrote:
> This patch enables fm10k TSO feature for both non-tunneling packet
> and tunneling packet.
>
> Signed-off-by: Wang Xiao W
> ---
Acked-by: Michael Qiu
On Mon, Aug 24, 2015 at 05:22:57PM +0800, Michael Qiu wrote:
> For __SSE3__, the corresponding header file should be pmmintrin.h,
> tmmintrin.h works for __SSSE3__.
>
> Signed-off-by: Michael Qiu
Since this is a bug-fix, it should probably have a fixes line in the commit.
Otherwise the change
Hi Jiuling,
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Jiuling Bie
> Sent: Wednesday, October 07, 2015 5:54 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] [testpmd] enable lsc to avoid TX stall, TX stall happened
> in following sequence start show port
Hi David
Is there anything I can do to help getting this merged?
Thanks,
Chris
On Mon, Oct 5, 2015 at 12:44 PM, Dumitrescu, Cristian
wrote:
>
>
>> -Original Message-
>> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Christoph Gysin
>> Sent: Tuesday, September 29, 2015 7:53 AM
Hello Christoph,
On Tue, Oct 13, 2015 at 11:10 AM, Christoph Gysin wrote:
>
> Is there anything I can do to help getting this merged?
>
This is ok for me, cc-ing Thomas.
--
David Marchand
On Mon, Oct 12, 2015 at 10:58:17PM +0200, Steffen Bauch wrote:
> On 10/12/2015 10:39 AM, Yuanhan Liu wrote:
> >Hi,
> >
> >I just recognized that this dead loop is the same one that I have
> >experienced (see
> >http://dpdk.org/ml/archives/dev/2015-October/024737.html for
> >reference). Just
> -Original Message-
> From: Yeddula, Avinash [mailto:ayeddula at ciena.com]
> Sent: Monday, October 12, 2015 6:03 PM
> To: Dumitrescu, Cristian; dev at dpdk.org; Bly, Mike
> Subject: RE: DPDK hash function related question
>
> Hi Cristian,
> I have configured the hash function and it
Updating the firmware using the tool did solve the problem! Now link is up.
Thanks a lot for your time and help Stephen! :)
-Original Message-
From: Stephen Hemminger [mailto:step...@networkplumber.org]
Sent: Monday, October 12, 2015 8:14 PM
To: Shaham Fridenberg
Cc: dev at dpdk.org
Ping?
This is one of three outstanding DPDK patches I have which hasn't seen
any activitiy in a while. Is there a list of pending applies somewhere
to monitor activity?
// Simon
On Thu, 24 Sep 2015 09:43:28 +0200
Simon Kagstrom wrote:
> Otherwise building with KERNELCC="ccache gcc" will fail:
On Tue, Oct 13, 2015 at 11:13 AM, David Marchand
wrote:
> Hello Christoph,
>
> On Tue, Oct 13, 2015 at 11:10 AM, Christoph Gysin <
> christoph.gysin at gmail.com> wrote:
>
>>
>> Is there anything I can do to help getting this merged?
>>
>
> This is ok for me, cc-ing Thomas.
>
Thought I already
Hi Simon,
Sorry for not having answered before.
On 10/13/2015 02:10 PM, Simon Kagstrom wrote:
> Ping?
>
> This is one of three outstanding DPDK patches I have which hasn't seen
> any activitiy in a while. Is there a list of pending applies somewhere
> to monitor activity?
There is the
Hello,
I have a system using dpdk 1.8 with 82599ES ixgbe NICs. These are
provided to a virtual guest via pci passthrough. Our dpdk application
on the guest takes control of the NICs using igb_uio.
On certain systems, under conditions we have not yet figured out,
sending traffic causes the host
On 2015-10-13 14:26, Olivier MATZ wrote:
> Sorry for not having answered before.
Thanks for looking at it now though!
>> This is one of three outstanding DPDK patches I have which hasn't seen
>> any activitiy in a while. Is there a list of pending applies somewhere
>> to monitor activity?
>
>
2015-10-13 14:39, Simon K?gstr?m:
> On 2015-10-13 14:26, Olivier MATZ wrote:
> > There is the patchwork tool:
> > http://dpdk.org/dev/patchwork/project/dpdk/list/
>
> Thanks! I knew about it, but forgot to look there. It would be nice to
> have tags to signify e.g., for-v2.2, candidate-v2.2 etc.
Hi Chas,
On 10/05/2015 05:26 PM, Chas Williams wrote:
> If a system is using deterministic interface names, it may be easier in
> some cases to use the interface name to blacklist an interface.
>
> Signed-off-by: Chas Williams <3chas3 at gmail.com>
> ---
> app/test/test_devargs.c
Ping?
// Simon
On Mon, 7 Sep 2015 14:50:09 +0200
Simon Kagstrom wrote:
> Chaining/segmenting mbufs can be useful in many places, so make it
> global.
>
> Signed-off-by: Simon Kagstrom
> Signed-off-by: Johan Faltstrom
> ---
> ChangeLog:
> v2:
> * Check for nb_segs byte overflow (Olivier
Ping?
// Simon
On Thu, 20 Aug 2015 08:51:06 +0200
Simon Kagstrom wrote:
> /proc/version_signature is the version for the host machine, but in
> e.g., chroots, this does not necessarily match that DPDK is built
> for. DPDK will then build for the wrong kernel version - that of the
> server, and
On 2015-10-13 14:45, Thomas Monjalon wrote:
> 2015-10-13 14:39, Simon K?gstr?m:
>> Two of the patches (this one included) I have outstanding are build fixes
>> for use in our build environment, so it would be nice to them upstreamed.
>
> Waiting for integration of your patches, maybe you have
Hi Simon,
I'm looking forward to this patch since we also build the DPDK for
kernel versions which differ from the host kernel.
IMHO, the check for "Ubuntu" also needs change since this is also a
"host system" check instead of "target system" check.
I.e. A user may want to build for Ubuntu on a
Hi Simon,
On 09/07/2015 02:50 PM, Simon Kagstrom wrote:
> Chaining/segmenting mbufs can be useful in many places, so make it
> global.
>
> Signed-off-by: Simon Kagstrom
> Signed-off-by: Johan Faltstrom
>
> [...]
>
> --- a/lib/librte_mbuf/rte_mbuf.h
> +++ b/lib/librte_mbuf/rte_mbuf.h
> @@
2015-10-13 14:50, Simon Kagstrom:
> Ping?
OK you apply the ping method we have just talked about :)
To make it really effective, you should have these headers:
To: Olivier Matz
Cc: dev at dpdk.org
Indeed, as the mbuf maintainer, he's the target of your ping.
And to make it clear,
Hi Francesco,
On 09/29/2015 06:04 PM, Francesco Montorsi wrote:
> From: Francesco Montorsi
>
> ---
> mk/rte.sdkbuild.mk | 6 ++
> 1 file changed, 6 insertions(+)
>
> diff --git a/mk/rte.sdkbuild.mk b/mk/rte.sdkbuild.mk
> index 38ec7bd..013aa89 100644
> --- a/mk/rte.sdkbuild.mk
> +++
From: Fan Zhang
This patchset links to ABI change announced for librte_table.
Key_mask parameters has been added to the hash table parameter
structure for 8-byte key and 16-byte key extendible bucket
and LRU tables.
v2:
*change in release note.
Acked-by: Cristian
From: Fan Zhang
Tests have been updated to work on added key_mask parameter for 8-byte
key extendible bucket and LRU tables.
Signed-off-by: Fan Zhang
---
app/test/test_table_combined.c | 4
app/test/test_table_tables.c | 6 --
2 files changed, 8
From: Fan Zhang
This patch relates to ABI change proposed for librte_table. key_mask
parameter is added for 8-byte key extendible bucket and LRU tables.
Signed-off-by: Fan Zhang
---
lib/librte_table/rte_table_hash.h | 6
From: Fan Zhang
This patch updates the flow_classification pipeline for added key_mask
parameter in 8/16-byte key hash parameters. The update provides user
optional key_mask configuration item applying to the packets.
Signed-off-by: Fan Zhang
---
From: Fan Zhang
The LIBABIVER number is incremented. The release notes is
updated and the deprecation announce is removed.
Signed-off-by: Fan Zhang
---
doc/guides/rel_notes/deprecation.rst | 3 ---
doc/guides/rel_notes/release_2_2.rst | 4 +++-
From: Fan Zhang
This patch relates to ABI change proposed for librte_table. key_mask
parameter is added for 16-byte key extendible bucket and LRU tables.
Signed-off-by: Fan Zhang
---
lib/librte_table/rte_table_hash.h | 6
From: Fan Zhang
This patch is to adding hash table operations for key signature
computed on lookup ("do-sig") for LRU hash tables and Extendible buckets.
Signed-off-by: Fan Zhang
---
lib/librte_table/rte_table_hash.h | 8 +
From: Fan Zhang
This patch adds parse_hex_string function to parse hex string to uint8_t
array.
Signed-off-by: Fan Zhang
---
examples/ip_pipeline/config_parse.c | 70 +
examples/ip_pipeline/pipeline.h | 4 +++
2 files changed,
Hi Maryam,
On 09/30/2015 10:20 AM, Maryam Tahhan wrote:
> Remove the deprecation tag and notice for imissed.
>
> Signed-off-by: Maryam Tahhan
> ---
> doc/guides/rel_notes/deprecation.rst | 2 +-
> lib/librte_ether/rte_ethdev.h| 3 +--
> 2 files changed, 2 insertions(+), 3 deletions(-)
On Mon, Oct 12, 2015 at 10:18:30PM -0700, Stephen Hemminger wrote:
> On Tue, 13 Oct 2015 02:57:46 +
> "Sanford, Robert" wrote:
>
> > I'm hoping that someone (perhaps at Intel) can help us understand
> > an IXGBE RX packet loss issue we're able to reproduce with testpmd.
> >
> > We run
Hi Pablo,
On 10/02/2015 05:53 PM, Pablo de Lara wrote:
> From: "De Lara Guarch, Pablo"
>
> When creating a ring, a memzone is created to allocate it in memory,
> but the ring could not be freed, as memzones could not be.
>
> Since memzones can be freed now, then rings can be as well,
> taking
Hi Bruce,
On 09/30/2015 02:12 PM, Bruce Richardson wrote:
> Add a new field to the rte_ring structure to store the memzone pointer which
> contains the ring. For rings created using rte_ring_create(), the field will
> be set automatically.
>
> This new field will allow users of the ring to query
>>> [Robert:]
>>> 1. The 82599 device supports up to 128 queues. Why do we see trouble
>>> with as few as 5 queues? What could limit the system (and one port
>>> controlled by 5+ cores) from receiving at line-rate without loss?
>>>
>>> 2. As far as we can tell, the RX path only touches the device
On Tue, Oct 13, 2015 at 04:54:06PM +0800, Tiwei Bie wrote:
> The variable optind must be reinitialized to 1 in order to skip over
> argv[0] on FreeBSD. Because getopt() on FreeBSD will return -1 when
> it meets an argument which doesn't start with '-'.
>
> The variable optreset is provided on
Hi Pablo,
The issue is related to certain NIC(s). I observed this on Intel
82577LM(em). Basically show port info will read PHY registers to get link
status when lsc interrupt was disabled, which caused TX to stop. I don't
have other NICs so not sure it is a common issue or not.
Regards,
Jiuling
On 10/13/2015 7:47 AM, Sanford, Robert wrote:
[Robert:]
1. The 82599 device supports up to 128 queues. Why do we see trouble
with as few as 5 queues? What could limit the system (and one port
controlled by 5+ cores) from receiving at line-rate without loss?
2. As
Hi all,
The below will try to suggest a modification to the initialization of
Environment Abstraction Layer (AKA EAL) so it will be able to allocate memory
zones from same virtual memory addresses even if the primary process is not
similar to the secondary processes.
Problem:
The DPDK
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Nissim Nisimov
> Sent: Tuesday, October 13, 2015 4:40 PM
> To: 'dev at dpdk.org'
> Subject: [dpdk-dev] propose a solution for mapping same virtual address
> space to asymmetric processes
>
> Hi all,
>
> The
On Wed, 30 Sep 2015 09:20:56 +0100
Maryam Tahhan wrote:
> diff --git a/lib/librte_ether/rte_ethdev.h b/lib/librte_ether/rte_ethdev.h
> index fa06554..78bd94d 100644
> --- a/lib/librte_ether/rte_ethdev.h
> +++ b/lib/librte_ether/rte_ethdev.h
> @@ -194,8 +194,7 @@ struct rte_eth_stats {
>
> -Original Message-
> From: Stephen Hemminger [mailto:stephen at networkplumber.org]
> Sent: Friday, October 2, 2015 12:38 AM
> To: Richardson, Bruce; De Lara Guarch, Pablo
> Cc: dev at dpdk.org; Stephen Hemminger
> Subject: [PATCH] crc: deinline crc functions
>
> Because the CRC
> -Original Message-
> From: dev [mailto:dev-bounces at dpdk.org] On Behalf Of Thomas F Herbert
> Sent: Thursday, October 8, 2015 7:30 PM
> To: dev at dpdk.org
> Subject: [dpdk-dev] DPDK User Space: Session onUseability and Ease of Use
>
> All:
>
> Captured white board notes from Jim
Actually, this is a good opportunity to fix a bug that's been in this code
forever: it shouldn't be resetting optind to some arbitrary value: it should be
saving optind (and optarg and optopt) at the beginning, initializing optind to
1 before calling getopt_long(), then restoring all the values
Hi everybody, apologies for coming to this list with a tech support question.
I'm completely stumped about using non-Intel transceivers with DPDK. testpmd is
bailing here: PMD: eth_ixgbe_dev_init(): Unsupported SFP+ Module / PMD:
eth_ixgbe_dev_init(): Hardware Initialization Failure: -19
My
Hi Bruce,
Using "--base-virtaddr" requires knowledge on the huge pages wanted address
going to be used and might vary on different uses of the application.
We suggest a more generic solution which wont require any previous knowledge
and will be "bullet proof" as much as possible.
Regards,
I believe I've discovered my problem:
https://gist.github.com/AlexForster/0fb4699bcdf196cf5462
As mentioned previously, I have two X520-Q1 cards installed. It appears that
initialization of the first card obeys allow_unsupported_sfp=1, but
initialization of the second card does not.
Is this a
Agreed, this handles the off by one error on wrap around and should be
faster.
-Roger
On 10/12/15 11:41 AM, Alexander Duyck wrote:
> On 10/12/2015 06:33 AM, Harry van Haaren wrote:
>> Fix a misinterpretation of VF stats in ixgbe
>>
>> Signed-off-by: Harry van Haaren
>> ---
>>
ack
On 10/12/15 12:45 PM, Harry van Haaren wrote:
> Fix a misinterpreatation of VF statistic macro in e1000/igb.
>
> Signed-off-by: Harry van Haaren
> ---
> drivers/net/e1000/igb_ethdev.c | 3 +--
> 1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/net/e1000/igb_ethdev.c
Thanks John for the summary.
2015-10-13 16:36, Mcnamara, John:
> - Move the EAL to the kernel.
Please explain what you mean here.
It's difficult to imagine.
> * Latest version of the docs. - Needs support from 6Wind.
OK
It's as simple as a git hook.
Anybody to write and test one locally?
The
On 10/13/2015 07:47 AM, Sanford, Robert wrote:
[Robert:]
1. The 82599 device supports up to 128 queues. Why do we see trouble
with as few as 5 queues? What could limit the system (and one port
controlled by 5+ cores) from receiving at line-rate without loss?
2. As far
On 10/13/2015 11:57 AM, Alex Forster wrote:
> I believe I've discovered my problem:
> https://gist.github.com/AlexForster/0fb4699bcdf196cf5462
>
> As mentioned previously, I have two X520-Q1 cards installed. It appears that
> initialization of the first card obeys allow_unsupported_sfp=1, but
>
On Thu, 1 Oct 2015 07:25:45 -0400
Neil Horman wrote:
> On Wed, Sep 30, 2015 at 05:37:05PM +0200, Thomas Monjalon wrote:
> > 2015-09-30 10:52, Neil Horman:
> > > On Wed, Sep 30, 2015 at 10:28:53AM +0200, David Marchand wrote:
> > > > On Tue, Sep 29, 2015 at 9:25 PM, Stephen Hemminger <
> > > >
On Mon, 20 Apr 2015 13:33:29 -0700
Ravi Kerur wrote:
> Remove unnecessary type casting in functions.
>
> Tested on Ubuntu (14.04 x86_64) with "make test".
> "make test" results match the results with baseline.
> "Memcpy perf" results match the results with baseline.
>
> Signed-off-by: Ravi
On Tue, 14 Jul 2015 09:57:04 +0100
Sergio Gonzalez Monroy wrote:
> Current implemetation allows reserving/creating memzones but not the opposite
> (unreserve/free). This affects mempools and other memzone based objects.
>
> From my point of view, implementing free functionality for memzones
On Mon, 18 May 2015 13:01:43 -0700
Ravi Kerur wrote:
> This patch implements memcmp and use librte_hash as the first candidate
> to use rte_memcmp which is implemented using AVX/SSE intrinsics.
>
> Tested with GCC(4.8.2) and Clang(3.4-1) compilers and both tests show better
> performance on
On Fri, 5 Jun 2015 10:46:36 +0800
Wen-Chi Yang wrote:
> Due to eal_alarm_callback() and rte_eal_alarm_set() use gettimeofday()
> to get the current time, and gettimeofday() is affected by jumps.
>
> For example, set up a rte_alarm which will be triggerd next second (
> current time + 1 second)
70 matches
Mail list logo