Memory interface (memif), provides high performance
packet transfer over shared memory.
Signed-off-by: Jakub Grajciar
---
MAINTAINERS |6 +
config/common_base |5 +
config/common_linux |1 +
doc/guides/n
Hi,
> -Original Message-
> From: Wei, Dan
> Sent: Thursday, May 30, 2019 22:59
> To: dev@dpdk.org
> Cc: Yigit, Ferruh ; Wei, Dan ;
> Xu, Rosen ; sta...@dpdk.org
> Subject: [DPDK v2] net/ipn3ke: modifications on AFU configurations
>
> Modify AFU configurations for new Blue Bitstream of A10
On 05/30, Stephen Hemminger wrote:
>On Thu, 30 May 2019 17:07:06 +0800
>Xiaolong Ye wrote:
>
>> This patch adds two parameters `start_queue` and `queue_count` to
>> specify the range of netdev queues used by AF_XDP pmd.
>>
>> Signed-off-by: Xiaolong Ye
>
>Why does this have to be a config option
On 05/30, Stephen Hemminger wrote:
>On Thu, 30 May 2019 17:07:05 +0800
>Xiaolong Ye wrote:
>
>> Implement zero copy of af_xdp pmd through mbuf's external memory
>> mechanism to achieve high performance.
>>
>> This patch also provides a new parameter "pmd_zero_copy" for user, so they
>> can choose
Add support for packets that consist of multiple segments.
Take into account that trailer bytes (padding, ESP tail, ICV)
can spawn across multiple segments.
Signed-off-by: Konstantin Ananyev
---
v1 -> v2:
merge with latest mainline
fix build problem for RTE_BUILD_SHARED_LIB=y
update programme
On Thu, 30 May 2019 22:25:20 +0100
Bruce Richardson wrote:
> +
> +#ifdef __cplusplus
> +extern "C" {
> +#endif
> +
This a driver private file, why the C++ guard here?
Hello,
(Sorry for late response, we should fix the automation issue before sending out
the result. This is test result of
http://git.dpdk.org/dpdk-stable/snapshot/dpdk-stable-18.11.2-rc1.tar.xz)
DPDK
https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fgit.dpdk.org%2Fdpdk-stable%2F
In the case where no drivers are registered with the system,
rte_pci_get_iommu_class should return RTE_IOVA_DC.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 91 -
1 file changed, 50 insertions(+), 41 deletions(-)
diff --git a/drivers/bus/pci/li
If the bus can't determine a preference for IOVA_PA vs.
IOVA_VA by looking at the devices and drivers, as a
last resort test if physical addresses are even accessible
in /proc/self/pagemap. If they are, use IOVA_PA. If they
are not, use IOVA_VA.
Signed-off-by: Ben Walker
---
lib/librte_eal/commo
All of the checks should respect the white and black lists.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/bus/pci/linux/pci.c b/drivers/bus/pci/linux/pci.c
index 6d311f4e0..d2464d2ae 100644
--- a/driver
If a device is found that is bound to the UIO driver,
only force IOVA_PA if there is a driver registered to use it.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --git a/drivers/bus/pci/linux/pci.c b/drivers/bus/pci/l
Take several independent if statements and convert to a
switch statement.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 21 -
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git a/drivers/bus/pci/linux/pci.c b/drivers/bus/pci/linux/pci.c
index 765c473e8
It's simpler to reverse the if statement here, especially
with an upcoming simplification.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/drivers/bus/pci/linux/pci.c b/drivers/bus/pci/linux/pci.c
index d2464d2ae
The three loops can now be easily combined into one.
This is slightly less efficient than before because it
doesn't break out early. But that can be addressed
later.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 19 +++
1 file changed, 3 insertions(+), 16 deletions
This performs a check for whether the device should be ignored
due to whitelist or blacklist. This check eventually needs
to apply to all of the other checks in rte_pci_get_iommu_class.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 44 +
1 file c
Make all of the loops first iterate over devices, then
drivers. This is in preparation for combining them
into a single loop.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/bus/pci/linux/pci.c b/d
Two of these loops easily collapse into a single loop.
This sets the stage for future simplifications.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 31 ++-
1 file changed, 10 insertions(+), 21 deletions(-)
diff --git a/drivers/bus/pci/linux/pci.c b/dri
This is in preparation for future simplifications. The
functions are simply inlined for now.
Signed-off-by: Ben Walker
---
drivers/bus/pci/linux/pci.c | 176 +++-
1 file changed, 71 insertions(+), 105 deletions(-)
diff --git a/drivers/bus/pci/linux/pci.c b/driver
This function only returned the correct answer after
a call to initialize the memory subsystem. Make it work
prior to that.
Signed-off-by: Ben Walker
---
lib/librte_eal/linux/eal/eal_memory.c | 63 ---
1 file changed, 28 insertions(+), 35 deletions(-)
diff --git a/lib/li
Add local APIs to trigger data copies, and retrieve handle values once
those copies are completed. Included are unit tests to validate the data
is copies correctly.
Signed-off-by: Bruce Richardson
---
app/test/test_ioat_rawdev.c| 159 -
doc/guides/rawdevs/ioat
Allow initializing a driver instance.
Signed-off-by: Bruce Richardson
---
app/test/test_ioat_rawdev.c| 35 +-
doc/guides/rawdevs/ioat_rawdev.rst | 32 +
drivers/raw/ioat/ioat_rawdev.c | 75 ++
drivers/raw/ioat/rte_ioat_rawdev.h | 14
Add stats functions to track what is happening in the driver, and put
unit tests to check those.
Signed-off-by: Bruce Richardson
---
app/test/test_ioat_rawdev.c| 38 ++
doc/guides/rawdevs/ioat_rawdev.rst | 14 ++
drivers/raw/ioat/ioat_rawdev.c | 44 +++
Add in the "info_get" function to the driver, to allow us to query the
device. This allows us to have the unit test pick up the presence of
supported hardware or not.
Signed-off-by: Bruce Richardson
---
app/test/meson.build | 3 +++
app/test/test_ioat_rawdev.c| 23
Add the create/destroy driver functions so that we can actually allocate
a rawdev and destroy it when done. No rawdev API functions are actually
implemented at this point.
Signed-off-by: Bruce Richardson
---
doc/guides/rawdevs/ioat_rawdev.rst | 11
drivers/raw/ioat/ioat_rawdev.c | 93 ++
Add in the list of registers for the device. File is taken from the SPDK
project:
https://github.com/spdk/spdk/blob/master/include/spdk/ioat_spec.h
Signed-off-by: Bruce Richardson
---
drivers/raw/ioat/Makefile| 1 +
drivers/raw/ioat/meson.build | 3 +-
drivers/raw/ioat/rte_ioa
In order to allow binding/unbinding of devices for use by the
ioat_rawdev, we need to update the devbind script to add a new class
of device, and add device ids for the specific HW instances.
Signed-off-by: Bruce Richardson
---
doc/guides/rawdevs/ioat_rawdev.rst | 11 +++
usertools/dpdk-
Add stubs for ioat rawdev driver support in DPDK, specifically:
* makefile and meson build hooks
* initial public header file
* rawdev main C file, with probe and release functions
* release note update announcing the driver
* initial documentation for the new section in the rawdev doc
This patch series adds support for the Intel QuickData Technology
device, part of the Intel I/O Acceleration Technology (Intel I/OAT). It
is a raw device for allowing hardware DMA i.e. data copies in hardware.
Bruce Richardson (8):
raw/ioat: add initial support for ioat rawdev driver
usertools
On Thu, May 30, 2019 at 07:00:36PM +0200, David Marchand wrote:
>On Thu, May 30, 2019 at 3:39 PM Thomas Monjalon
><[1]tho...@monjalon.net> wrote:
>
> 30/05/2019 12:11, Bruce Richardson:
> > On Thu, May 30, 2019 at 09:40:08AM +0200, Thomas Monjalon wrote:
> > > 30/05/2019 09:
> -Original Message-
> From: Stephen Hemminger [mailto:step...@networkplumber.org]
> Sent: Thursday, May 30, 2019 10:57 AM
> To: Walker, Benjamin
> Cc: dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH 02/12] eal/pci: Inline several functions into
> rte_pci_get_iommu_class
>
> On Thu, 30 May
On Thu, 30 May 2019 10:48:09 -0700
Ben Walker wrote:
> This is in preparation for future simplifications. The
> functions are simply inlined for now.
>
> Signed-off-by: Ben Walker
> Change-Id: I129992c9b44f4575a28cc05b78297e15b6be4249
Please don't inline any functions that are not in the fast
The three loops can now be easily combined into one.
This is slightly less efficient than before because it
doesn't break out early. But that can be addressed
later.
Signed-off-by: Ben Walker
Change-Id: Ic97155bb478dddbcbeaa6d51947684ffef219a52
---
drivers/bus/pci/linux/pci.c | 19 +++--
If a device is found that is bound to the UIO driver,
only force IOVA_PA if there is a driver registered to use it.
Signed-off-by: Ben Walker
Change-Id: I8015f11a33ab1b7662bf374d6944eff8d7a74a07
---
drivers/bus/pci/linux/pci.c | 8 +++-
1 file changed, 7 insertions(+), 1 deletion(-)
diff --
If the bus can't determine a preference for IOVA_PA vs.
IOVA_VA by looking at the devices and drivers, as a
last resort test if physical addresses are even accessible
in /proc/self/pagemap. If they are, use IOVA_PA. If they
are not, use IOVA_VA.
Change-Id: If1eeb723283b80b24bd973987054fdad62f59cbd
Two of these loops easily collapse into a single loop.
This sets the stage for future simplifications.
Signed-off-by: Ben Walker
Change-Id: I3353f2e3585808cebff3f11805f96e4a1cc7fb3a
---
drivers/bus/pci/linux/pci.c | 31 ++-
1 file changed, 10 insertions(+), 21 deletio
Take several independent if statements and convert to a
switch statement.
Signed-off-by: Ben Walker
Change-Id: Ia77c88ea484b529e8b0c9e09e8ef22cf3210e669
---
drivers/bus/pci/linux/pci.c | 21 -
1 file changed, 12 insertions(+), 9 deletions(-)
diff --git a/drivers/bus/pci/linu
This function only returned the correct answer after
a call to initialize the memory subsystem. Make it work
prior to that.
Signed-off-by: Ben Walker
Change-Id: I8f3c5128fbf5da884a956bbcc72c5a13564825d5
---
lib/librte_eal/linux/eal/eal_memory.c | 63 ---
1 file changed, 2
All of the checks should respect the white and black lists.
Signed-off-by: Ben Walker
Change-Id: Ie66176bea49987d1fc0a03dbee2638d9dd6efbc5
---
drivers/bus/pci/linux/pci.c | 7 ++-
1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/bus/pci/linux/pci.c b/drivers/bus/pci/linux
In the case where no drivers are registered with the system,
rte_pci_get_iommu_class should return RTE_IOVA_DC.
Signed-off-by: Ben Walker
Change-Id: Ia5b0cae100cfcfe46a9e4996328f9746ce33cfd3
---
drivers/bus/pci/linux/pci.c | 79 ++---
1 file changed, 38 insertions
Make all of the loops first iterate over devices, then
drivers. This is in preparation for combining them
into a single loop.
Signed-off-by: Ben Walker
Change-Id: Ifb2bfcc60570a5d5a13481be3da0fc74bf00ef1f
---
drivers/bus/pci/linux/pci.c | 12 ++--
1 file changed, 6 insertions(+), 6 delet
It's simpler to reverse the if statement here, especially
with an upcoming simplification.
Signed-off-by: Ben Walker
Change-Id: I6cff80231032304f3f865fdf38157554fad7fd07
---
drivers/bus/pci/linux/pci.c | 6 ++
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/drivers/bus/pci/linu
This performs a check for whether the device should be ignored
due to whitelist or blacklist. This check eventually needs
to apply to all of the other checks in rte_pci_get_iommu_class.
Signed-off-by: Ben Walker
Change-Id: I8e63e4c2e4199f34561ea1d911e13d6d74a47322
---
drivers/bus/pci/linux/pci.c
This is in preparation for future simplifications. The
functions are simply inlined for now.
Signed-off-by: Ben Walker
Change-Id: I129992c9b44f4575a28cc05b78297e15b6be4249
---
drivers/bus/pci/linux/pci.c | 176 +++-
1 file changed, 71 insertions(+), 105 deletions(
In SPDK, not all drivers are registered with DPDK at start up time.
Previously, that meant DPDK always chose to set itself up in IOVA_PA
mode. Instead, when the correct iova choice is unclear based on the
devices and drivers known to DPDK at start up time, use other heuristics
(such as whether /pro
When a cryptodev is created in a primary process,
rte_cryptodev_data_alloc reserves a memzone.
However, this memzone was not released when the cryptodev
is uninitialized. After that, new cryptodev cannot be
created due to memzone name conflict.
This commit frees the memzone when a cryptodev is
uni
On Thu, May 30, 2019 at 3:39 PM Thomas Monjalon wrote:
> 30/05/2019 12:11, Bruce Richardson:
> > On Thu, May 30, 2019 at 09:40:08AM +0200, Thomas Monjalon wrote:
> > > 30/05/2019 09:31, David Marchand:
> > > > On Thu, May 30, 2019 at 12:51 AM Stephen Hemminger <
> > > > step...@networkplumber.org
Hi Lukasz,
> diff --git a/lib/librte_ipsec/esp_outb.c b/lib/librte_ipsec/esp_outb.c
> index c798bc4..ed5974b 100644
> --- a/lib/librte_ipsec/esp_outb.c
> +++ b/lib/librte_ipsec/esp_outb.c
> @@ -126,11 +126,11 @@ outb_tun_pkt_prepare(struct rte_ipsec_sa *sa,
> rte_be64_t sqc,
>
> /* pad len
From: Marcin Hajkowski
Send request to power manager for core id provided
by user to get related capabilities.
Signed-off-by: Marcin Hajkowski
---
.../guest_cli/vm_power_cli_guest.c| 119 +-
1 file changed, 117 insertions(+), 2 deletions(-)
diff --git a/examples/vm
From: Marcin Hajkowski
Send capabilities for requested cores.
Signed-off-by: Marcin Hajkowski
---
examples/vm_power_manager/channel_monitor.c | 67 +
1 file changed, 67 insertions(+)
diff --git a/examples/vm_power_manager/channel_monitor.c
b/examples/vm_power_manager/chan
From: Marcin Hajkowski
Add new packet type and commands for capabilities query.
Signed-off-by: Marcin Hajkowski
---
lib/librte_power/channel_commands.h | 14 ++
1 file changed, 14 insertions(+)
diff --git a/lib/librte_power/channel_commands.h
b/lib/librte_power/channel_commands.h
From: Marcin Hajkowski
Extend guest channel and sample apps to query CPU capabilities.
Please note that these changes depends on
(http://patchwork.dpdk.org/cover/52335/) and
(http://patchwork.dpdk.org/cover/52213/) which should be applied first.
Marcin Hajkowski (3):
power: add new packet ty
On Thu, 30 May 2019 00:10:21 +
bugzi...@dpdk.org wrote:
> https://bugs.dpdk.org/show_bug.cgi?id=287
>
> Bug ID: 287
>Summary: netvsc PMD/dpdk/azure: Driver lockup with multi-queue
> configuration
>Product: DPDK
>Version: 18.1
Minutes 30 May 2019
---
Agenda:
* Release Dates
* Subtrees
* OvS
* Conferences
* Opens
Participants:
* Debian/Microsoft
* Intel
* Marvell
* Mellanox
* Red Hat
Release Dates
-
* v19.08 dates:
* Proposal/V1 Monday 03 June 2019
* Integration/Merg
https://bugs.dpdk.org/show_bug.cgi?id=288
Bug ID: 288
Summary: Target name recorded wrong when try to build dpdk with
x86_64-native-linux-gcc
Product: DPDK
Version: unspecified
Hardware: All
OS: All
HI Thomas,
Snipped
> > +
> > + To built other BPF examples, the compiler requires additional command-
> line options.
>
> "To built" -> "To build"
ok
>
> I think this note is vague. Don't you think it may confuse user if we don't
> explicit which kind of options are required?
Ok, the `v1` c
On Thu, 30 May 2019 17:07:06 +0800
Xiaolong Ye wrote:
> This patch adds two parameters `start_queue` and `queue_count` to
> specify the range of netdev queues used by AF_XDP pmd.
>
> Signed-off-by: Xiaolong Ye
Why does this have to be a config option, we already have max queues
and number of q
On Thu, 30 May 2019 17:07:05 +0800
Xiaolong Ye wrote:
> Implement zero copy of af_xdp pmd through mbuf's external memory
> mechanism to achieve high performance.
>
> This patch also provides a new parameter "pmd_zero_copy" for user, so they
> can choose to enable zero copy of af_xdp pmd or not.
30/05/2019 12:11, Bruce Richardson:
> On Thu, May 30, 2019 at 09:40:08AM +0200, Thomas Monjalon wrote:
> > 30/05/2019 09:31, David Marchand:
> > > On Thu, May 30, 2019 at 12:51 AM Stephen Hemminger <
> > > step...@networkplumber.org> wrote:
> > >
> > > > On Thu, 30 May 2019 00:46:30 +0200
> > > >
On Thu, 2019-05-30 at 15:38 +0300, Ilya Maximets wrote:
> Don't need to check dependencies if test apps will not be built
> anyway.
>
> Signed-off-by: Ilya Maximets <
> i.maxim...@samsung.com
> >
> ---
>
> Version 2:
> - 'get_option('tests')' check moved to the top.
>
> app/test/meson.build |
On Thu, 2019-05-30 at 14:59 +0300, Ilya Maximets wrote:
> On 30.05.2019 14:06, Luca Boccassi wrote:
> > On Thu, 2019-05-30 at 13:03 +0300, Ilya Maximets wrote:
> > > On 29.05.2019 23:37, Luca Boccassi wrote:
> > > > On Wed, 2019-05-29 at 19:39 +0300, Ilya Maximets wrote:
> > > > > The first thing m
According to API, 'rte_dev_probe()' and 'rte_dev_remove()' and their
'hotplug' equivalents must return 0 or negative error code. Bus code
returns positive values if device wasn't recognized by any driver, so
the result of 'bus->plug/unplug()' must be converted.
Positive on remove means that device
Replace the mbuf pointer array in the event eth Rx adapter
callback with an event array instead of an mbuf array. Using
an event array allows the application to change attributes
of the events enqueued by the SW adapter.
Signed-off-by: Nikhil Rao
---
lib/librte_eventdev/rte_event_eth_rx_adapter.
The application can install a callback invoked by
the Rx adapter. The callback can drop packets and populate
a callback argument with the number of dropped packets.
Add a Rx adapter stats field to keep track of the total
number of dropped packets.
Signed-off-by: Nikhil Rao
---
lib/librte_eventde
Ilya Maximets writes:
> Don't need to check dependencies if test apps will not be built anyway.
>
> Signed-off-by: Ilya Maximets
> ---
>
> Version 2:
> - 'get_option('tests')' check moved to the top.
>
> app/test/meson.build | 141 ++-
> 1 file changed,
Replace the mbuf pointer array in the event eth Rx adapter
callback with an event array instead of an mbuf array. Using
an event array allows the application to change attributes
of the events enqueued by the SW adapter.
Signed-off-by: Nikhil Rao
---
lib/librte_eventdev/rte_event_eth_rx_adapter.
On Thu, May 30, 2019 at 03:38:36PM +0300, Ilya Maximets wrote:
> Don't need to check dependencies if test apps will not be built anyway.
>
> Signed-off-by: Ilya Maximets
> ---
>
> Version 2:
> - 'get_option('tests')' check moved to the top.
>
> app/test/meson.build | 141
Don't need to check dependencies if test apps will not be built anyway.
Signed-off-by: Ilya Maximets
---
Version 2:
- 'get_option('tests')' check moved to the top.
app/test/meson.build | 141 ++-
1 file changed, 72 insertions(+), 69 deletions(-)
diff
> -Original Message-
> From: Ferruh Yigit
> Sent: Wednesday, May 29, 2019 7:29 PM
> To: Jakub Grajciar -X (jgrajcia - PANTHEON TECHNOLOGIES at Cisco)
> ; dev@dpdk.org
> Subject: Re: [dpdk-dev] [RFC v9] /net: memory interface (memif)
> > +
> > +.. csv-table:: **Memif configuration option
On Thu, May 30, 2019 at 03:06:17PM +0300, Ilya Maximets wrote:
> On 30.05.2019 14:55, Bruce Richardson wrote:
> > On Wed, May 29, 2019 at 07:39:57PM +0300, Ilya Maximets wrote:
> >> Don't need to check dependencies if test apps will not be built anyway.
> >>
> >> Signed-off-by: Ilya Maximets
> >>
On 30.05.2019 14:55, Bruce Richardson wrote:
> On Wed, May 29, 2019 at 07:39:57PM +0300, Ilya Maximets wrote:
>> Don't need to check dependencies if test apps will not be built anyway.
>>
>> Signed-off-by: Ilya Maximets
>> ---
>> app/test/meson.build | 38 +++---
>>
> -Original Message-
> From: Ian Stokes
> Sent: Thursday, May 30, 2019 11:16 AM
> To: Kevin Traynor ; dpdk stable
> Cc: dev@dpdk.org; Sitong Liu ; Pei Zhang
> ; Raslan Darawsheh ;
> qian.q...@intel.com; Ju-Hyoung Lee ; Ali Alnubani
> ; David Christensen ;
> benjamin.wal...@intel.com; Thom
On 30.05.2019 14:06, Luca Boccassi wrote:
> On Thu, 2019-05-30 at 13:03 +0300, Ilya Maximets wrote:
>> On 29.05.2019 23:37, Luca Boccassi wrote:
>>> On Wed, 2019-05-29 at 19:39 +0300, Ilya Maximets wrote:
The first thing many developers do before start building DPDK is
disabling all the n
On Wed, May 29, 2019 at 07:39:57PM +0300, Ilya Maximets wrote:
> Don't need to check dependencies if test apps will not be built anyway.
>
> Signed-off-by: Ilya Maximets
> ---
> app/test/meson.build | 38 +++---
> 1 file changed, 19 insertions(+), 19 deletions(-)
On 30.05.2019 13:22, Bruce Richardson wrote:
> On Wed, May 29, 2019 at 09:37:20PM +0100, Luca Boccassi wrote:
>> On Wed, 2019-05-29 at 19:39 +0300, Ilya Maximets wrote:
>>> The first thing many developers do before start building DPDK is
>>> disabling all the not needed divers and libraries. This h
On Thu, 2019-05-30 at 13:03 +0300, Ilya Maximets wrote:
> On 29.05.2019 23:37, Luca Boccassi wrote:
> > On Wed, 2019-05-29 at 19:39 +0300, Ilya Maximets wrote:
> > > The first thing many developers do before start building DPDK is
> > > disabling all the not needed divers and libraries. This happen
When WQEs are posted to the HW to send packets, the PMD may get a
completion report with error from the HW, aka error CQE which is
associated to a bad WQE.
The error reason may be bad address, wrong lkey, bad sizes, etc.
that can wrongly be configured by the PMD or by the user.
Checking all the o
The SQ errors recovery mechanism in the PMD invokes a Verbs
functions to modify the RQ states in order to reset the SQ and to
reactivate it.
These Verbs functions are not allowed to be invoked from a secondary
process, hence the PMD skips the recovery when the error is captured
by secondary proces
On Wed, May 29, 2019 at 09:37:20PM +0100, Luca Boccassi wrote:
> On Wed, 2019-05-29 at 19:39 +0300, Ilya Maximets wrote:
> > The first thing many developers do before start building DPDK is
> > disabling all the not needed divers and libraries. This happens
> > just because more than a half of DPDK
When WQEs are posted to the HW to receive packets, the PMD may receive
a completion report with error from the HW, aka error CQE which is
associated to a bad WQE.
The error reason may be bad address, wrong lkey, small buffer size,
etc. that can wrongly be configured by the PMD or by the user.
Che
The RQ errors recovery mechanism in the PMD invokes a Verbs functions to
modify the RQ states in order to reset the RQ and to reactivate it.
These Verbs functions are not allowed to be invoked from a secondary
process, hence the PMD skips the recovery when the error is captured by
secondary proces
Move the RQ WQEs initialization code to separate function as an
arrangement to CQE error recovering for code reuse.
CC: sta...@dpdk.org
Signed-off-by: Matan Azrad
---
drivers/net/mlx5/mlx5_rxq.c | 43 ++-
drivers/net/mlx5/mlx5_rxtx.c | 53 +++
When bad device arguments are added to the DPDK command line, the PMD
ignores all the command line arguments specified by the user and uses
the default values instead.
This behavior doesn't make sense because the user intension is to force
some device parameters and expects to get an error in case
There is a full correlation between the CQE indexes to the WQE indexes
in the vectorized Rx queues management.
When the RQ is inserted to the reset state, the correlation may break
because the HW starts the RQ polling from index 0 while the CQ polling
continues regularly.
As an arrangement to CQE
The RQ WQEs must be written in the memory before the HW gets the RQ
doorbell, hence a memory barrier should be triggered after the WQEs
writing and before the doorbell writing.
The current code used rte_wmb barrier which ensures that all the memory
stores were done while it is enough to use rte_ci
Add a global function in the PMD which dumps debug information to
specific file.
The data can be printed in hexadecimal format or as regular string.
The number of debug files per PMD entity should be limited by a new PMD
probe parameter called max_dump_files_num.
The files will be created in the
Add support for data-path Rx and Tx completions with error handling:
1. Detect the error.
2. Do not crash.
3. Report it in statistics counters.
4. Dump debug information to system log file.
5. Recover the error under the hood.
6. Add support for secondary process recovery.
No performance impact w
On Thu, May 30, 2019 at 09:07:44AM +0100, Burakov, Anatoly wrote:
> On 29-May-19 9:11 PM, David Marchand wrote:
> > On Wed, May 29, 2019 at 6:31 PM Anatoly Burakov
> > wrote:
> >
> > > This patchset removes the shared memory config from public
> > > API, and replaces all usages of said config wit
On Thu, May 30, 2019 at 09:40:08AM +0200, Thomas Monjalon wrote:
> 30/05/2019 09:31, David Marchand:
> > On Thu, May 30, 2019 at 12:51 AM Stephen Hemminger <
> > step...@networkplumber.org> wrote:
> >
> > > On Thu, 30 May 2019 00:46:30 +0200
> > > Thomas Monjalon wrote:
> > >
> > > > 23/05/2019 1
Hi All,
A reminder that there is no longer a default in practice of having short
term stable branches/releases for xx.02/05/08 DPDK master releases.
Note, this is relevant for xx.02/05/08 based short term (~3 month)
stables only. DPDK LTS based off xx.11 is *not* changing.
This is to allow more
On 29.05.2019 23:37, Luca Boccassi wrote:
> On Wed, 2019-05-29 at 19:39 +0300, Ilya Maximets wrote:
>> The first thing many developers do before start building DPDK is
>> disabling all the not needed divers and libraries. This happens
>> just because more than a half of DPDK dirvers and libraries a
On 29.05.2019 23:15, Michael Santana Francisco wrote:
> On 5/29/19 12:39 PM, Ilya Maximets wrote:
>> The first thing many developers do before start building DPDK is
>> disabling all the not needed divers and libraries. This happens
>> just because more than a half of DPDK dirvers and libraries are
Fixes: f1debd77efaf ("net/af_xdp: introduce AF_XDP PMD")
Cc: sta...@dpdk.org
Signed-off-by: Xiaolong Ye
---
drivers/net/af_xdp/rte_eth_af_xdp.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/net/af_xdp/rte_eth_af_xdp.c
b/drivers/net/af_xdp/rte_eth_af_xdp.c
index f56aabcae..fc25d245b
This patch series mainly includes 3 new features for AF_XDP pmd. They
are separated independent features, the reason I take them in one
patchset is that they have code dependency.
1. zero copy
This patch enables `zero copy` between af_xdp umem and mbuf by using
external mbuf mechanism.
2. multi-
Implement zero copy of af_xdp pmd through mbuf's external memory
mechanism to achieve high performance.
This patch also provides a new parameter "pmd_zero_copy" for user, so they
can choose to enable zero copy of af_xdp pmd or not.
To be clear, "zero copy" here is different from the "zero copy mo
This patch adds two parameters `start_queue` and `queue_count` to
specify the range of netdev queues used by AF_XDP pmd.
Signed-off-by: Xiaolong Ye
---
doc/guides/nics/af_xdp.rst | 3 +-
drivers/net/af_xdp/rte_eth_af_xdp.c | 194
2 files changed, 141 inser
On 5/21/2019 3:01 PM, Kevin Traynor wrote:
Hi all,
Here is a list of patches targeted for LTS release 18.11.2.
The planned date for the final release is 11th June.
Please help with testing and validation of your use cases and report
any issues/results. For the final release I will update the r
Added release note entry for test-compress-perf application
Signed-off-by: Tomasz Jozwiak
---
doc/guides/rel_notes/release_19_08.rst | 3 +++
1 file changed, 3 insertions(+)
diff --git a/doc/guides/rel_notes/release_19_08.rst
b/doc/guides/rel_notes/release_19_08.rst
index b9510f9..543e7d3 1006
This patch adds a possibility to force controlled process termination
as a result of two signals: SIGTERM and SIGINT
Signed-off-by: Tomasz Jozwiak
---
app/test-compress-perf/comp_perf_options.h| 1 +
app/test-compress-perf/comp_perf_test_benchmark.c | 13
app/test-compress-
On 29-May-19 9:11 PM, David Marchand wrote:
On Wed, May 29, 2019 at 6:31 PM Anatoly Burakov
wrote:
This patchset removes the shared memory config from public
API, and replaces all usages of said config with new API
calls.
The patchset is mostly a search-and-replace job and should
be pretty ea
This patch adds a benchmark part to
compression-perf-tool as a separate test case, which can be
executed multi-threaded.
Signed-off-by: Tomasz Jozwiak
---
app/test-compress-perf/Makefile | 1 +
app/test-compress-perf/comp_perf_test_benchmark.c | 139 --
ap
1 - 100 of 108 matches
Mail list logo
