On Tue, Jul 12, 2016 at 04:32:12PM +0200, Maxime Coquelin wrote:
> Indirect descriptors are usually supported by virtio-net devices,
> allowing to dispatch a large number of large requests.
>
> When the virtio device sends a packet using indirect descriptors,
> only one slot is used in the ring, e
Hello.
Function rte_eth_dev_attach can return false positive result.
It happens because rte_eal_pci_probe_one returns zero if no driver is found for
the device:
ret = pci_probe_all_drivers(dev);
if (ret < 0)
goto err_return;
return 0;
(pci_probe_all_drivers returns 1 in that case)
For ex
Now that we have a new blank release notes, a new generated NIC features table,
and the obsolete code removed, we can start applying some patches for 16.11.
Version number 16.11-rc0 applied.
I wish you a happy start of release cycle :)
Signed-off-by: Thomas Monjalon
---
lib/librte_eal/common/include/rte_version.h | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/lib/librte_eal/common/include/rte_version.h
b/lib/librte_eal/common/include/rte_version.h
index 615deb7..2c20cb3 100644
--- a/lib/librte_eal/co
2016-08-01 15:04, Thomas Monjalon:
> The function __rte_mbuf_raw_alloc was reserved for internal use and
> has been deprecated in favor of the public function rte_mbuf_raw_alloc.
> It can be safely removed now.
>
> Signed-off-by: Thomas Monjalon
Applied
2016-07-29 16:50, David Marchand:
> On Fri, Jul 29, 2016 at 3:50 PM, Thomas Monjalon
> wrote:
> > The log history feature was deprecated in 16.07.
> > The remaining empty functions are removed in 16.11.
> >
> > Signed-off-by: Thomas Monjalon
> > ---
> > v2: fix LIBABIVER and compilation of test
>
2016-07-29 15:41, Thomas Monjalon:
> It was planned to remove some mempool functions which are deprecated
> since 16.07.
> As no other mempool ABI change is planned in 16.11, it is better
> to postpone and group every mempool ABI changes in 17.02.
>
> Signed-off-by: Thomas Monjalon
Applied
2016-08-03 15:32, Bruce Richardson:
> On Mon, Aug 01, 2016 at 11:37:03PM +0200, Thomas Monjalon wrote:
> > Hi John,
> >
> > 2016-07-29 12:59, John McNamara:
> > > Convert the NIC feature table in the overview doc into a set of ini
> > > files and add functions into the Sphinx conf.py file to auto-
Replying to everything at once, please see below.
On Tue, Jul 26, 2016 at 03:37:35PM +0530, Rahul Lakkireddy wrote:
> On Monday, July 07/25/16, 2016 at 09:40:02 -0700, John Fastabend wrote:
> > On 16-07-25 04:32 AM, Rahul Lakkireddy wrote:
> > > Hi Adrien,
> > >
> > > On Thursday, July 07/21/16,
On 3 August 2016 at 16:29, Ananyev, Konstantin
wrote:
>
> Hi Jianbo,
>
>> > Hi, Jianbo
>> >
>> > I have tested you patch on my X86 platform, the single core performance
>> > for Non-vector PMD will have about 1Mpps drop
>> > Non-vector PMD single core performance with patch : ~33.
Hi Thomas,
On 3 August 2016 at 15:58, Thomas Monjalon wrote:
> 2016-08-03 14:02, Jianbo Liu:
>> I think we can place a compiling condition before rte_rmb() to avoid
>> performance decrease on x86.
>> For example: #if defined(RTE_ARCH_ARM) || defined(RTE_ARCH_ARM64)
>
> Please could you explain w
RSS hash-key-size is retrieved from device configuration instead of
using a fixed size of 40 bytes.
Fixes: f79959ea1504 ("app/testpmd: allow to configure RSS hash key")
Signed-off-by: Mohammad Abdul Awal
---
app/test-pmd/cmdline.c | 24 +---
app/test-pmd/config.c | 17 +
Hi Kieran,
On Mon, Aug 01, 2016 at 04:08:51PM +0100, Kieran Mansley wrote:
> Apologies for coming a little late to this thread about the proposed new
> API for filtering etc.
>
> I've reviewed it based on Solarflare's needs and hardware capabilities,
> and think the proposal is likely to be a big
*Updates in v2*
- drop the #!/bin/echo now that it is no more executable
Quoting the first line of the script: "#! /bin/echo must be loaded with ."
Given that we should drop the .sh file ending as well as the executable
flag - both are not needed to source the file.
Signed-off-by: Christian Ehrha
On Wed, Aug 3, 2016 at 3:19 PM, Thomas Monjalon
wrote:
> > Do you want me to resubmit as-is but without RFC tag or would you accept
> it
> > as already submitted?
>
> I will apply it as is.
As just discussed on #dpdk now that it is no more executable I'd even drop
the #!/bin/echo.
That is to av
Hi John,
I'm replying below to both messages.
On Tue, Aug 02, 2016 at 11:19:15AM -0700, John Fastabend wrote:
> On 16-07-23 02:10 PM, John Fastabend wrote:
> > On 16-07-21 12:20 PM, Adrien Mazarguil wrote:
> >> Hi Jerin,
> >>
> >> Sorry, looks like I missed your reply. Please see below.
> >>
> >
And add read memory barrier to avoid status inconsistency
between two RX descriptors readings.
Signed-off-by: Jianbo Liu
---
config/defconfig_arm64-armv8a-linuxapp-gcc | 2 +-
doc/guides/nics/overview.rst | 2 +-
drivers/net/i40e/i40e_rxtx.c | 2 ++
3 files changed, 4
On Mon, Aug 01, 2016 at 11:37:03PM +0200, Thomas Monjalon wrote:
> Hi John,
>
> 2016-07-29 12:59, John McNamara:
> > Convert the NIC feature table in the overview doc into a set of ini
> > files and add functions into the Sphinx conf.py file to auto-generate
> > them back into an RST table.
>
> I
On Fri, Jul 29, 2016 at 8:20 PM, Pablo de Lara <
pablo.de.lara.guarch at intel.com> wrote:
> dpdk-nic-bind.py script has been renamed to dpdk-devbind.py,
> but some references to the old script have remained.
> This commit completes the renaming.
>
> Fixes: a5d7a3f77ddc ("unify tools naming")
>
>
2016-08-03 11:42, Christian Ehrhardt:
> On Wed, Aug 3, 2016 at 11:26 AM, Bruce Richardson <
> bruce.richardson at intel.com> wrote:
>
> > Definitely no objection on the file mode change.
> >
> > For the dropping of the .sh extension, I don't think it matters much.
> > However,
> > given that .sh f
When a rule with depth > 24 is added into an existing
rule with depth <=24, a new tbl8 is allocated, the existing
rule first fulfill whole new tbl8, so the filed valid of
each entry in this tbl8 is always true and depth of each
entry is always <= 24 before adding the new rule with depth > 24.
Sign
As a bug-fix for lpm tlb8 recycle is introduced,
add a test case to verify tlb8 group is correctly
freed when it only includes a rule with depth=24.
Signed-off-by: Wei Dai
---
app/test/test_lpm.c | 80 -
1 file changed, 79 insertions(+), 1 dele
When all rules with depth > 24 are deleted in a same sub-table
(tlb8 group) and only a rule with depth <=24 is left in it,
this sub-table (tlb8 group) should be recycled.
Fixes: dc81ebbacaeb ("lpm: extend IPv4 next hop field")
Fixes: af75078fece3 ("first public release")
Signed-off-by: Wei Dai
-
On 3 August 2016 at 11:26, Yao, Lei A wrote:
> Hi, Jianbo
>
> I have tested you patch on my X86 platform, the single core performance for
> Non-vector PMD will have about 1Mpps drop
> Non-vector PMD single core performance with patch : ~33.9 Mpps
> Non-vector PMD single core perfo
googletest is a very nice test framework and we use it very
extensively in our company(Luminate Wireless), together with gmock.
I understand the resistance from the maintainers that are concerned
about introducing a C++ dependency to a pure C code base. The approach
we take doesn't require any cha
The operaton may have an undefined behavior or yield to an unexpected result.
A bit shift operation has a shift amount which is too large or has a negative
value.
Coverity issue: 30688
Fixes: ea977ff1cb0b ("examples/exception_path: fix shift operation in lcore
setup")
The previous patch forget t
[...]
>> The proposal looks very good. It satisfies most of the features
>> supported by Chelsio NICs. We are looking for suggestions on exposing
>> more additional features supported by Chelsio NICs via this API.
>>
>> Chelsio NICs have two regions in which filters can be pl
On Wed, Aug 3, 2016 at 11:26 AM, Bruce Richardson <
bruce.richardson at intel.com> wrote:
> Definitely no objection on the file mode change.
>
> For the dropping of the .sh extension, I don't think it matters much.
> However,
> given that .sh files are generally scripts to be executed, I think
> d
[...]
Considering that allowed pattern/actions combinations cannot be known in
advance and would result in an unpractically large number of capabilities
to
expose, a method is provided to validate a given rule from the current
device configuration state without actually a
On Wed, Aug 03, 2016 at 10:11:32AM +0200, Christian Ehrhardt wrote:
> sorry, I accidentally dropped dev list in one of my replies, readding.
>
> On Tue, Aug 2, 2016 at 11:29 PM, Thomas Monjalon 6wind.com>
> wrote:
>
> > > > > Given that we should drop the .sh file ending as well as the
> > execu
On 02/08/2016 22:52, Thomas Monjalon wrote:
> 2016-08-02 21:37, Declan Doherty:
[..]
> You are not advocating but the unit test must be written in C++.
> I don't think it is a good idea to force people to write and maintain the
> tests
> in a different language than the code it tests.
In princip
On Tue, Aug 02, 2016 at 11:36:41PM +0200, Thomas Monjalon wrote:
> 2016-08-02 17:04, Bruce Richardson:
> > Having to make this change twice shows up the fact that we are still
> > carrying
> > around some version changes for older releases. Given that we are now past
> > the
> > 16.07 release, th
sorry, I accidentally dropped dev list in one of my replies, readding.
On Tue, Aug 2, 2016 at 11:29 PM, Thomas Monjalon
wrote:
> > > > Given that we should drop the .sh file ending as well as the
> executable
> > > > flag - both are not needed to source the file.
> > >
> > > Hmmm, it is still a
On 8/3/2016 8:13 AM, Dai, Wei wrote:
> This patch is same as patch 15063 which has been acked by Bruce Richardson.
> Only title and message body of git log has been revised according to Bruce's
> advice.
For future, you can carry Ac with new version of the patch for these
kind of cases.
2016-08-03 14:02, Jianbo Liu:
> I think we can place a compiling condition before rte_rmb() to avoid
> performance decrease on x86.
> For example: #if defined(RTE_ARCH_ARM) || defined(RTE_ARCH_ARM64)
Please could you explain why a memory barrier would be needed on ARM but
not on x86? What about o
> -Original Message-
...
> You are not advocating but the unit test must be written in C++.
> I don't think it is a good idea to force people to write and maintain the
> tests
> in a different language than the code it tests.
I know where you are coming from on this point, and I general
Hi,
2016-08-02 15:31, txcy uio:
> Hello
>
> I am getting a segmentation fault on freebsd 10 due to the shared config
> mismatch between the primary and secondary process -
>
> The following commit seems to have fixed this issue long time back on Linux
> but was never ported to freebsd.
> http://
On Wed, Aug 03, 2016 at 09:57:39AM +, Doherty, Declan wrote:
>
>
> > -Original Message-
> ...
> > You are not advocating but the unit test must be written in C++.
> > I don't think it is a good idea to force people to write and maintain the
> > tests
> > in a different language than
Hi Jianbo,
> > Hi, Jianbo
> >
> > I have tested you patch on my X86 platform, the single core performance
> > for Non-vector PMD will have about 1Mpps drop
> > Non-vector PMD single core performance with patch : ~33.9
> > Mpps
> > Non-vector PMD single core performance without p
On Tue, Aug 02, 2016 at 11:52:24PM +0200, Thomas Monjalon wrote:
> 2016-08-02 21:37, Declan Doherty:
> > I've been trying out using google test as a possible replacement for our
> > unit
> > test framework and have put to together this series of patches with help
> > from
> > Anatoly as RFC to ge
On Tue, Aug 02, 2016 at 09:37:48PM +0100, Declan Doherty wrote:
> Adds new command line options which allows the user to stop
> application echoing log output to stdout, logs are still
> written to syslog.
>
> Signed-off-by: Declan Doherty
> ---
> lib/librte_eal/common/eal_common_log.c | 14
This patch is same as patch 15094 which has been acked by Bruce Richardson.
Only minor typo is corrected according to Bruce's feedback.
> -Original Message-
> From: Dai, Wei
> Sent: Wednesday, August 3, 2016 3:04 PM
> To: dev at dpdk.org
> Cc: Dai, Wei
> Subject: [PATCH v3 3/3] lpm: remov
This patch is same as patch 15064 which has been acked by Bruce Richardson.
Resending it as 2/3 of a series of patches for same rte_lpm.c .
> -Original Message-
> From: Dai, Wei
> Sent: Wednesday, August 3, 2016 3:04 PM
> To: dev at dpdk.org
> Cc: Dai, Wei
> Subject: [PATCH v3 3/3] lpm: r
This patch is same as patch 15063 which has been acked by Bruce Richardson.
Only title and message body of git log has been revised according to Bruce's
advice.
> -Original Message-
> From: Dai, Wei
> Sent: Wednesday, August 3, 2016 3:02 PM
> To: dev at dpdk.org
> Cc: Dai, Wei
> Subject:
Hi, Jianbo
I have tested you patch on my X86 platform, the single core performance for
Non-vector PMD will have about 1Mpps drop
Non-vector PMD single core performance with patch : ~33.9 Mpps
Non-vector PMD single core performance without patch: ~35.1 Mpps
Is there any wa
45 matches
Mail list logo
