Hi Yinan,
On 7/24/20 6:55 AM, Wang, Yinan wrote:
> Hi Maxime,
>
> The performance drop issue can be fixed, thanks!
> The multi-queues interrupt issue still exist w/ this patch set.
Thanks for the test report, so that's only half good.
I'm setting up the multi-queues interrupt test case to furthe
hi, simei
On 7/24/2020 10:10 AM, Simei Su wrote:
When adding a RSS rule with GTPU_DWN/UP, it will search profile
table from the top index. If a RSS rule with GTPU_EH already exists,
then GTPU_DWN/UP packet will match GTPU_EH profile. This patch solves
this issue by removing existed GTPU_EH rule
This patch enable i40e outer vlan strip on and off in QinQ
mode with mask bit of DEV_RX_OFFLOAD_QINQ_STRIP, users can
use "vlan set qinq_strip on 0" to enable or "vlan set
qinq_strip on 0" to disable i40e outer vlan strip when
try with testpmd app.
Fixes: 4861cde46116 (i40e: new poll mode driver)
Add maintainers.
> -Original Message-
> From: Zhao1, Wei
> Sent: Friday, July 24, 2020 3:02 PM
> To: dev@dpdk.org
> Cc: sta...@dpdk.org; Zhang, Qi Z ; Zhao1, Wei
>
> Subject: [PATCH] net/i40e: enable i40e outer VLAN strip in QinQ
>
> This patch enable i40e outer vlan strip on and off i
The device ready state in vhost lib is now defined as the state
that first queue pair is ready. And kick/callfd may be updated
by QEMU when ifc device is configured.
Although now ifc driver only supports one queue pair, it still
has to update callfd when working with QEMU. This patch fixes
this vr
The hash look up table (LUT) is managed by global register but it is not
initialized when RSS is disabled. Once user wants to enable RSS during
runtime, the LUT will not be initialized.
This patch fixes the issue by initializing the LUT whatever RSS enabled
or not.
Fixes: feaae285b342 ("net/i40e:
> -Original Message-
> From: Guo, Jia
> Sent: Friday, July 24, 2020 10:22 AM
> To: Zhang, Qi Z ; Xing, Beilei ;
> Wu, Jingjing
> Cc: dev@dpdk.org; Guo, Junfeng ; Su, Simei
> ; Guo, Jia
> Subject: [dpdk-dev v1] net/iavf: add gtpu hash in default
>
> Add GTPU_IP and GTPU_EH hash in def
> -Original Message-
> From: dev On Behalf Of Wei Zhao
> Sent: Friday, July 24, 2020 15:02
> To: dev@dpdk.org
> Cc: sta...@dpdk.org; Zhang, Qi Z ; Zhao1, Wei
>
> Subject: [dpdk-dev] [PATCH] net/i40e: enable i40e outer VLAN strip in QinQ
>
> This patch enable i40e outer vlan strip on a
On 7/24/20 9:06 AM, Maxime Coquelin wrote:
> Hi Yinan,
>
> On 7/24/20 6:55 AM, Wang, Yinan wrote:
>> Hi Maxime,
>>
>> The performance drop issue can be fixed, thanks!
>> The multi-queues interrupt issue still exist w/ this patch set.
>
> Thanks for the test report, so that's only half good.
>
Tested-by: Zhang, XiX
Regards,
Xie Wei
-Original Message-
From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Shougang Wang
Sent: Friday, July 24, 2020 4:13 PM
To: dev@dpdk.org
Cc: Xing, Beilei ; Guo, Jia ; Yang,
Qiming ; Wang, ShougangX ;
sta...@dpdk.org
Subject: [dpdk-dev] [PATCH v5
> -Original Message-
> From: Guo, Jia
> Sent: Friday, July 24, 2020 12:07 PM
> To: Zhang, Qi Z ; Xing, Beilei ;
> Wu, Jingjing
> Cc: dev@dpdk.org; Guo, Junfeng ; Su, Simei
> ; Guo, Jia
> Subject: [dpdk-dev v1] net/iavf: delete unsupported rss types
>
> The combined hash type should b
Avoid race with unregister interrupt hanlder if interrupt
source has some active callbacks at the moment, use wrapper
around rte_intr_callback_unregister() to check for -EAGAIN
return value.
Signed-off-by: Renata Saiakhova
---
lib/librte_eal/freebsd/eal_interrupts.c | 16 ++--
lib/li
For pci_vfio, while removing the device by rte_dev_remove(),
pci_vfio_disable_notifier() will call rte_intr_callback_unregister(),
which may return -EAGAIN, if an interrupt source (kernel) has some active
callbacks right now. As a result, the callback (req notifier) can be never
unregistered,
and
From: Mike Baucom
- Add TCAM shadow tables for searching
- Add Search API to allow reuse of TCAM entries
Signed-off-by: Mike Baucom
Reviewed-by: Randy Schacher
Reviewed-by: Ajit Khaparde
---
drivers/net/bnxt/tf_core/tf_core.c| 73 ++
drivers/net/bnxt/tf_core/tf_core.h| 101 +
From: Mike Baucom
modified ulp mappper to use the new tf_search_tcam_entry API.
When search before allocation is requested, mapper calls
tc_search_tcam_entry with the alloc flag.
- On HIT, the result and tcam index is returned.
- On MISS, the tcam index is returned but the result is
created and
Some fixes, cleanups and changes to augment pre-existing
support in infrastructure
Please apply
v1->v2:
- Fixed some typos in patch [9/20].
v2->v3:
- Fixed coding style issues in patch [15/20] to use __rte_attribute.
- Updated commit messages.
- Added a patch to fix a logic error [21/22].
-
From: Kishore Padmanabha
PARIF is handler to a partition of the physical port.
For the offload miss rules, the parif miss path needs to be
considered. The higher parif are reserved for handling this.
Signed-off-by: Kishore Padmanabha
Reviewed-by: Michael Baucom
Reviewed-by: Ajit Khaparde
---
From: Kishore Padmanabha
Add support to enable or disable the NAT global registers.
The NAT feature is enabled in hardware during initialization
and disabled at deinitialization of the application.
Signed-off-by: Kishore Padmanabha
Reviewed-by: Michael Baucom
Reviewed-by: Ajit Khaparde
---
d
From: Kishore Padmanabha
The template is updated to support additional combinations
of NAT actions.
Signed-off-by: Kishore Padmanabha
Reviewed-by: Michael Baucom
Reviewed-by: Ajit Khaparde
---
drivers/net/bnxt/tf_ulp/ulp_template_db_act.c | 412 +++---
.../net/bnxt/tf_ulp/ulp_tem
From: Mike Baucom
Added tf_hash API for common hash uses across tf_core functions
Signed-off-by: Mike Baucom
Reviewed-by: Farah Smith
Reviewed-by: Kishore Padmanabha
---
drivers/net/bnxt/meson.build | 1 +
drivers/net/bnxt/tf_core/Makefile | 1 +
drivers/net/bnxt/tf_core/tf_hash.c
From: Venkat Duvvuru
When a packet is looped back from VF to VFR, it is marked to identify
the VFR interface. However, this mark_id shouldn't be percolated up to
the OVS as it is internal to pmd.
This patch fixes it by skipping mark injection into mbuf if the packet
is received on VFR interface.
From: Kishore Padmanabha
The parif for the egress rules need to be dynamically
configured based on the port type.
PARIF is handler to a partition of the physical port.
Signed-off-by: Kishore Padmanabha
Reviewed-by: Michael Baucom
---
drivers/net/bnxt/tf_ulp/ulp_def_rules.c | 11 ++--
dr
From: Kishore Padmanabha
Added egress template with VLAN tag match
Signed-off-by: Kishore Padmanabha
Reviewed-by: Shahaji Bhosle
---
.../net/bnxt/tf_ulp/ulp_template_db_class.c | 501 +-
.../net/bnxt/tf_ulp/ulp_template_db_enum.h| 28 +-
2 files changed, 509 insertions(
From: Kishore Padmanabha
This is a work around for the OVS setting offload rules that
are passing vlan priority mask as wild card and currently we
do not support it.
Signed-off-by: Kishore Padmanabha
Reviewed-by: Michael Baucom
---
drivers/net/bnxt/tf_ulp/ulp_rte_parser.c | 9 +
1 fil
From: Mike Baucom
Removed the hash calculation from tf_shadow_tcam in favor of using a
new common implementation.
Signed-off-by: Mike Baucom
Reviewed-by: Kishore Padmanabha
Reviewed-by: Farah Smith
Reviewed-by: Ajit Khaparde
---
drivers/net/bnxt/tf_core/tf_shadow_tcam.c | 77 +--
From: Mike Baucom
- Added Index Table shadow tables for searching
- Added Search API to allow reuse of Table entries
Signed-off-by: Mike Baucom
Reviewed-by: Farah Smith
---
drivers/net/bnxt/tf_core/tf_core.c| 66 +-
drivers/net/bnxt/tf_core/tf_core.h| 79 ++-
drivers/net/bnx
From: Venkat Duvvuru
Currently, the flow_ids of port_to_app/app_to_port & tx_cfa_action
for the first port are getting over-written by the second port because
these fields are stored in the ulp context which is common across the
ports.
This patch fixes the problem by having per port structure to
From: Mike Baucom
modified ulp mappper to use the new tf_search_tbl_entry API.
When search before allocation is requested, mapper calls
tc_search_tbl_entry with the alloc flag.
- On HIT, the result and table index is returned.
- On MISS, the table index is returned but the result is
created and
From: Venkat Duvvuru
Truflow stack adds VFR to VF and VF to VFR conduits when VF
representor is created. However, in the ingress direction the
VF's fw rules conflict with Truflow rules, resulting in not hitting
the Truflow VFR rules. To fix this, fw is going to remove it’s
VF rules when vf repres
From: Kishore Padmanabha
Added support for tcam table processing to enable the search
and allocate support. This also includes the tcam entry update
support.
Signed-off-by: Kishore Padmanabha
Reviewed-by: Michael Baucom
---
drivers/net/bnxt/tf_ulp/ulp_mapper.c | 317 -
From: Mike Baucom
Turn on shadow memory in the core to allow search before allocate.
This allows reuse of constrained resources.
Signed-off-by: Mike Baucom
Reviewed-by: Venkat Duvvuru
Reviewed-by: Ajit Khaparde
---
drivers/net/bnxt/tf_ulp/bnxt_ulp.c | 2 +-
1 file changed, 1 insertion(+), 1
From: Mike Baucom
Search before alloc allows reuse of constrained resources such as tcam,
encap, and source modifications. The new templates will search the
entry and alloc only if necessary.
Signed-off-by: Mike Baucom
Reviewed-by: Venkat Duvvuru
Reviewed-by: Ajit Khaparde
---
drivers/net/b
On 7/24/2020 5:07 AM, Jeff Guo wrote:
> The combined hash type should be bound with prefix protocol when
> configure it, so delete some useless and unsupported part for
> rss types mapping.
>
> Signed-off-by: Jeff Guo
Hi Jeff,
Can you please add the fixes line as reference? There is a refactori
From: Mike Baucom
Moved setting the refcnt for shadow tcam and table entries to the
allocation path only. The insert can be called multiple times for
updates and was resetting the refcnt to 1 each time. Now multiple
insertion/modifications will not change the reference count.
Signed-off-by: Mi
From: Somnath Kotur
No need to access rx_cfa_code, cfa_code_map from the VF-Rep functions
anymore.
Fixes: 322bd6e70272 ("net/bnxt: add port representor infrastructure")
Reviewed-by: Venkat Duvvuru
Signed-off-by: Somnath Kotur
Reviewed-by: Ajit Khaparde
---
drivers/net/bnxt/bnxt.h | 1
The if condition in bnxt_restore_mac_filters needs to check for
the result of logical AND. But it was not doing it resulting in
an incorrect check.
Fixes: b02f1573cd07 ("net/bnxt: restore MAC filters during reset recovery")
Signed-off-by: Ajit Khaparde
---
drivers/net/bnxt/bnxt_ethdev.c | 2 +-
When we compile PMD with CFLAGS set to -O -g, build fails because of
uninitialized error. This patch fixes it.
Fixes: 1e46b3962620 ("net/bnxt: fill cfa action in Tx descriptor")
Bugzilla ID: 509
Cc: step...@networkplumber.org
Signed-off-by: Ajit Khaparde
Reviewed-by: Somnath Kotur
Reviewed-by:
This patch fixes the uninitialized variable bug in QAT PMD.
Fixes: 9f27a860dc16 ("crypto/qat: move generic qp function to qp file")
Signed-off-by: Adam Dybkowski
---
drivers/common/qat/qat_qp.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/common/qat/qat_qp.c b/dri
The hash look up table (LUT) is managed by global register but it is not
initialized when RSS is disabled. Once user wants to enable RSS during
runtime, the LUT will not be initialized.
This patch fixes the issue by initializing the LUT whatever RSS enabled
or not.
Fixes: feaae285b342 ("net/i40e:
hi, ferruh
Fixes: 7be10c3004be ("net/iavf: add RSS configuration for VF")
Thanks for your info, the above fixed line should be added, and need I sent a
new version or you could help to add it when merge it?
Best regards,
Jeff Guo
-Original Message-
From: Ferruh Yigit
Sent: Friday, Ju
Acked-by: Jeff Guo
On 7/24/2020 5:38 PM, Shougang Wang wrote:
The hash look up table (LUT) is managed by global register but it is not
initialized when RSS is disabled. Once user wants to enable RSS during
runtime, the LUT will not be initialized.
This patch fixes the issue by initializing the
> -Original Message-
> From: Guo, Jia
> Sent: Friday, July 24, 2020 5:52 PM
> To: Yigit, Ferruh ; Zhang, Qi Z
> ;
> Xing, Beilei ; Wu, Jingjing
> Cc: dev@dpdk.org; Guo, Junfeng ; Su, Simei
>
> Subject: RE: [dpdk-dev] [dpdk-dev v1] net/iavf: delete unsupported rss types
>
> hi, ferruh
> -Original Message-
> From: dev On Behalf Of Jeff Guo
> Sent: Friday, July 24, 2020 5:55 PM
> To: Wang, ShougangX ; dev@dpdk.org
> Cc: Xing, Beilei ; Yang, Qiming
> ; sta...@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH v6] net/i40e: fix incorrect hash look up table
>
> Acked-by: Jeff Guo
Hi,
This series adds SR-IOV PF pmd driver support to have VF pmd
driver work over PF pmd driver instances in order to run the
adapter completely under DPDK environment for one of the use
cases like ovs-dpdk.
This is very initial bring-up with following testing covered -
* Enable/Disable SR-IOV V
This patch defines various PCI config space access APIs
in order to read and find IOV specific PCI capabilities.
With these definitions implemented, it enables the base
driver to do SR-IOV specific initialization and HW specific
configuration required from PF-PMD driver instance.
Signed-off-by: M
By adding generic API, this patch removes individual
functions/defines implemented by drivers to find PCI
extended capability.
Signed-off-by: Manish Chopra
Signed-off-by: Igor Russkikh
---
drivers/bus/pci/pci_common.c | 42 ++
drivers/bus/pci/rte_bus_pci.h
Based on number of VFs enabled at PCI, PF-PMD driver instance
enables/configures those VFs from hardware perspective, such
that in later patches they could get required HW access to
communicate with PFs for slowpath configuration and run the
fastpath themsleves.
This patch also add two new qede IO
This patch adds necessary infrastructure support (required to handle
messages from VF and sending ramrod on behalf of VF's configuration
request from alarm handler context) to start/load the VF-PMD driver
instance on top of PF-PMD driver instance.
Signed-off-by: Manish Chopra
Signed-off-by: Igor
This patch configures VFs with random mac if no MAC is
provided by the PF/bulletin. This also adds required bulletin
APIs by PF-PMD driver to communicate LINK properties/changes to
the VFs through bulletin update mechanism.
With these changes, VF-PMD instance is able to run
fastpath over PF-PMD dr
This patch adds required bit to handle VF FLR
indication from Management FW (MFW) of the device
With that VFs were able to load in VM (VF attached as PCI
passthrough to the guest VM) followed by FLR successfully
Updated the docs/guides with the feature support
Signed-off-by: Manish Chopra
Signe
On Thu, Jul 23, 2020 at 11:09:02PM +0300, Parav Pandit wrote:
> From: Thomas Monjalon
>
> Define each sub-directory on its own line ended with a comma,
> and use a simple indent.
>
> Signed-off-by: Thomas Monjalon
> ---
> drivers/meson.build | 24 +---
> 1 file changed, 13
Replace NITROX PMD specific tests with generic test suite.
Signed-off-by: Nagadheeraj Rottela
---
app/test/test_cryptodev.c | 18 +-
1 file changed, 1 insertion(+), 17 deletions(-)
diff --git a/app/test/test_cryptodev.c b/app/test/test_cryptodev.c
index 70bf6fe2c..162134a5c 1006
This patch adds cipher only crypto operation support.
Signed-off-by: Nagadheeraj Rottela
---
doc/guides/cryptodevs/nitrox.rst | 2 -
drivers/crypto/nitrox/nitrox_sym.c| 3 +
drivers/crypto/nitrox/nitrox_sym_reqmgr.c | 189 --
3 files changed, 143 insertio
This patch adds AES-GCM AEAD algorithm.
Signed-off-by: Nagadheeraj Rottela
---
doc/guides/cryptodevs/features/nitrox.ini | 3 +
doc/guides/cryptodevs/nitrox.rst | 4 +
drivers/crypto/nitrox/nitrox_sym.c| 82 +++-
.../crypto/nitrox/nitrox_sym_capabilities.c
This patch set replaces the NITROX PMD specific test suite with generic
test suite and adds support for AES-GCM and cipher only offload.
Nagadheeraj Rottela (3):
test/crypto: replace NITROX PMD specific test suite
crypto/nitrox: support AES-GCM
crypto/nitrox: support cipher only crypto opera
On Thu, Jul 23, 2020 at 11:09:03PM +0300, Parav Pandit wrote:
> From: Thomas Monjalon
>
> Drivers dependencies are evaluated in the order defined per
> their parent directory (also called class).
> This strict ordering prevent from having 2 different drivers
> of the same class with different dep
This patch adds examples to the Telemetry HowTo guide, to demonstrate
commands that use parameters. The programmer's guide is also modified to
include details on writing a callback function for a new command.
Signed-off-by: Ciara Power
---
v3:
- Replaced direct code examples with generic examp
Hi Thomas,
Thanks for the review.
I have addressed your comments in the v3 patch, with one exception mentioned
below.
>[...]
>> + * List all commands.
>> +
>> + .. code-block:: console
>
>I think you can achieve the same result as code-block with adding double
>colons (::) at the end o
>-Original Message-
>From: Power, Ciara
>Sent: Friday 24 July 2020 12:21
>To: Laatz, Kevin
>Cc: dev@dpdk.org; Richardson, Bruce ;
>tho...@monjalon.net; Power, Ciara
>Subject: [PATCH v3] doc: add more detail to telemetry guides
>
>This patch adds examples to the Telemetry HowTo guide, t
> -Original Message-
> From: Dybkowski, AdamX
> Sent: Friday, July 24, 2020 10:40 AM
> To: dev@dpdk.org; Trahe, Fiona ; akhil.go...@nxp.com
> Cc: Dybkowski, AdamX
> Subject: [PATCH] common/qat: fix uninitialized variable bug
>
> This patch fixes the uninitialized variable bug in QAT P
> -Original Message-
> From: Trahe, Fiona
> Sent: Friday, 24 July, 2020 13:55
> To: Dybkowski, AdamX ; dev@dpdk.org;
> akhil.go...@nxp.com
> Cc: Trahe, Fiona
> Subject: RE: [PATCH] common/qat: fix uninitialized variable bug
>
>
>
> > -Original Message-
> > From: Dybkowski, Adam
There are a lot of per virtq operations in the live migration
handling.
Before the driver support for queue update, when a virtq was not valid,
all the LM handling was terminated.
But now, when the driver supports queue update, the virtq can be invalid
as legal stage.
Skip invalid virtq in LM ha
> -Original Message-
> From: Dybkowski, AdamX
> Sent: Friday, July 24, 2020 12:58 PM
> To: Trahe, Fiona ; dev@dpdk.org; akhil.go...@nxp.com
> Subject: RE: [PATCH] common/qat: fix uninitialized variable bug
>
> > -Original Message-
> > From: Trahe, Fiona
> > Sent: Friday, 24 Ju
On Fri, Jul 24, 2020 at 9:06 AM Maxime Coquelin
wrote:
> On 7/24/20 6:55 AM, Wang, Yinan wrote:
> > Hi Maxime,
> >
> > The performance drop issue can be fixed, thanks!
> > The multi-queues interrupt issue still exist w/ this patch set.
>
> Thanks for the test report, so that's only half good.
> I'
This commit adds a new experimental API which allows the user
to retrieve the active state of an lcore. Knowing when the service
lcore is completed its polling loop can be useful to applications
to avoid race conditions when e.g. finalizing statistics.
The service thread itself now has a variable
This commit fixes a potential race condition in the tests
where the lcore running a service would increment a counter
that was already reset by the test-suite thread. The resulting
race-condition incremented value could cause CI failures, as
indicated by DPDK's CI.
This patch fixes the race-condit
On 23-Jul-20 3:48 PM, wangyunjian wrote:
From: Yunjian Wang
Currently, we will create new user mem map entry for the same memory
segment, but in fact it has already been added to the user mem maps.
It's not necessary to create it twice.
To resolve the issue, add support to remove the same entr
On Thu, Jul 23, 2020 at 10:10 PM Parav Pandit wrote:
>
> From: Thomas Monjalon
>
> Define each sub-directory on its own line ended with a comma,
> and use a simple indent.
>
> Signed-off-by: Thomas Monjalon
> ---
> drivers/meson.build | 24 +---
> 1 file changed, 13 insertio
23/07/2020 19:00, Stephen Hemminger:
> On Thu, 23 Jul 2020 17:13:17 +0500
> Sarosh Arif wrote:
>
> > Since this script has no maintainer who is responsible for providing ACK or
> > NAK on the patches related to this script?
> >
> > On Tue, Jun 2, 2020 at 12:39 PM Sarosh Arif wrote:
> >
> > > C
On Thu, Jul 23, 2020 at 10:10 PM Parav Pandit wrote:
>
> From: Thomas Monjalon
>
> Drivers dependencies are evaluated in the order defined per
> their parent directory (also called class).
> This strict ordering prevent from having 2 different drivers
> of the same class with different dependenci
Other PMDs may be using the mempool, so don't free it when destroying the
UMEM.
Fixes: d8a210774e1d ("net/af_xdp: support unaligned umem chunks")
Cc: sta...@dpdk.org
Signed-off-by: Ciara Loftus
---
drivers/net/af_xdp/rte_eth_af_xdp.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/ne
This commit adds a new experimental API which allows the user
to retrieve the active state of an lcore. Knowing when the service
lcore is completed its polling loop can be useful to applications
to avoid race conditions when e.g. finalizing statistics.
The service thread itself now has a variable
This commit fixes a potential race condition in the tests
where the lcore running a service would increment a counter
that was already reset by the test-suite thread. The resulting
race-condition incremented value could cause CI failures, as
indicated by DPDK's CI.
This patch fixes the race-condit
On Thu, Jul 23, 2020 at 10:10 PM Parav Pandit wrote:
>
> mlx5_common is shared library between mlx5 net, VDPA and regex PMD.
> It is better to use common initialization helper instead of using
> RTE_INIT_CLASS priority.
RTE_INIT_CLASS does not exist, I suppose you meant RTE_PRIORITY_CLASS.
Sugge
> From: David Marchand
> Sent: Friday, July 24, 2020 7:15 PM
> To: Parav Pandit
> Cc: dev ; Gaetan Rivet ; Yigit, Ferruh
> ; Thomas Monjalon ;
> Raslan Darawsheh ; Ori Kam
> ; Matan Azrad ; Joyce Kong
>
> Subject: Re: [dpdk-dev] [PATCH v8 06/10] common/mlx5: avoid using class
> constructor pri
Hi Bruce,
> From: Bruce Richardson
> Sent: Friday, July 24, 2020 4:37 PM
>
> On Thu, Jul 23, 2020 at 11:09:03PM +0300, Parav Pandit wrote:
> > From: Thomas Monjalon
> >
> > Drivers dependencies are evaluated in the order defined per their
> > parent directory (also called class).
> > This stric
From: Patrick Fu
Async copy fails when single ring buffer vector is splited on multiple
physical pages. This happens because current hpa address translation
function doesn't handle multi-page buffers. A new gpa to hpa address
conversion function, which returns the hpa on the first hitting host
pa
24/07/2020 15:48, Parav Pandit:
> Hi Bruce,
>
> > From: Bruce Richardson
> > Sent: Friday, July 24, 2020 4:37 PM
> >
> > On Thu, Jul 23, 2020 at 11:09:03PM +0300, Parav Pandit wrote:
> > > From: Thomas Monjalon
> > >
> > > Drivers dependencies are evaluated in the order defined per their
> > >
On Fri, Jul 24, 2020 at 7:00 PM Thomas Monjalon wrote:
>
> 23/07/2020 19:00, Stephen Hemminger:
> > On Thu, 23 Jul 2020 17:13:17 +0500
> > Sarosh Arif wrote:
> >
> > > Since this script has no maintainer who is responsible for providing ACK
> > > or
> > > NAK on the patches related to this scrip
Hi, Guojia
> -Original Message-
> From: Guo, Jia
> Sent: Friday, July 24, 2020 3:13 PM
> To: Su, Simei ; Zhang, Qi Z ; Xing,
> Beilei
> Cc: dev@dpdk.org
> Subject: Re: [PATCH] net/ice: fix GTPU down/uplink and extension conflict
>
> hi, simei
>
> On 7/24/2020 10:10 AM, Simei Su wrote:
From: Thomas Monjalon
Define each sub-directory on its own line ended with a comma,
and use a simple indent.
Acked-by: Bruce Richardson
Reviewed-by: David Marchand
Signed-off-by: Thomas Monjalon
---
Changelog:
v7->8:
- new patch
---
drivers/meson.build | 24 +---
1 file
This series introduces mlx5 common driver layer to support multiple
class of devices for a single PCI device.
Motivation and example
--
mlx5 PCI device supports multiple class of devices such as net, vdpa
and regex devices.
Currently only one pmd (either net or vdpa) can bind
There are several drivers which duplicate bit generation macro.
Introduce a generic bit macros so that such drivers avoid redefining
same in multiple drivers.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
Acked-by: Morten Brørup
---
Changelog:
v4->v5:
- Addressed comments from Morten Brøru
From: Thomas Monjalon
Drivers dependencies are evaluated in the order defined per their parent
directory (also called class). This strict ordering prevents from
having pairs of drivers from two classes with different dependency
ordering. For example, if the mlx5 common code depends on the pci bus
Following two errors are reported when compiled with
gcc (GCC) 8.3.1 20191121 (Red Hat 8.3.1-5).
drivers/common/mlx5/linux/mlx5_glue.h:188:2:
error: function declaration isn't a prototype [-Werror=strict-prototypes]
drivers/common/mlx5/linux/mlx5_glue.h:188:2:
error: function declaration isn't a
When fail to initialize the device, avoid segmentation fault while
accessing unintialized priv.
Fixes: cfc672a90b74 ("regex/mlx5: support probing")
Signed-off-by: Parav Pandit
---
Changelog:
v7->v8:
- Rebased
- new patch
---
drivers/regex/mlx5/mlx5_regex.c | 7 ---
1 file changed, 4 inser
mlx5 PCI Device supports multiple classes of devices such as net, vdpa,
and/or regex.
To support these multiple classes, change mlx5_class to a
bitmap values so that if users asks to enable multiple of them, all
supported classes can be parsed.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
-
Add generic mlx5 PCI PMD layer as part of existing common_mlx5
module. This enables multiple classes (net, regex, vdpa) PMDs
to be supported at same time.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
---
Changelog:
v8->v9:
- Added missing LDFLAG for pci bus
- Fixed white spaces at start o
Migrate mlx5 net, vdpa and regex PMD to start using mlx5 common class
driver.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
---
Changelog:
v8->v9:
- Removed leftover inclusion of mlx5_pci bus include directory
---
drivers/common/mlx5/mlx5_common_pci.c | 6 ++
drivers/net/mlx5/Makefile
mlx5_common is shared library between mlx5 net, VDPA and regex PMD.
It is better to use common initialization helper instead of using
RTE_PRIORITY_CLASS priority.
Reviewed-by: David Marchand
Suggested-by: David Marchand
Signed-off-by: Parav Pandit
---
Changelog:
v7->v8:
- new patch
---
driver
Now that mlx5_pci PMD checks for enabled classes and performs
probe(), remove() of associated classes, individual class driver
does not need to check if other driver is enabled.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_common.c | 37
> From: Parav Pandit
> Sent: Friday, July 24, 2020 7:54 PM
>
> This series introduces mlx5 common driver layer to support multiple class of
> devices for a single PCI device.
>
[..]
>
> Changelog:
> v8->v9:
> - Updated commit message
> - Fixed LD_FLAGS
> - Fixed white space to tab
> - Fixe
From: Thomas Monjalon
Define each sub-directory on its own line ended with a comma,
and use a simple indent.
Acked-by: Bruce Richardson
Reviewed-by: David Marchand
Signed-off-by: Thomas Monjalon
---
Changelog:
v7->8:
- new patch
---
drivers/meson.build | 24 +---
1 file
This series introduces mlx5 common driver layer to support multiple
class of devices for a single PCI device.
Motivation and example
--
mlx5 PCI device supports multiple class of devices such as net, vdpa
and regex devices.
Currently only one pmd (either net or vdpa) can bind
There are several drivers which duplicate bit generation macro.
Introduce a generic bit macros so that such drivers avoid redefining
same in multiple drivers.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
Acked-by: Morten Brørup
---
Changelog:
v4->v5:
- Addressed comments from Morten Brøru
From: Thomas Monjalon
Drivers dependencies are evaluated in the order defined per their parent
directory (also called class). This strict ordering prevents from
having pairs of drivers from two classes with different dependency
ordering. For example, if the mlx5 common code depends on the pci bus
When fail to initialize the device, avoid segmentation fault while
accessing unintialized priv.
Fixes: cfc672a90b74 ("regex/mlx5: support probing")
Signed-off-by: Parav Pandit
---
Changelog:
v9->v10:
- Corrected type in commit log
v7->v8:
- Rebased
- new patch
---
drivers/regex/mlx5/mlx5_reg
Following two errors are reported when compiled with
gcc (GCC) 8.3.1 20191121 (Red Hat 8.3.1-5).
drivers/common/mlx5/linux/mlx5_glue.h:188:2:
error: function declaration isn't a prototype [-Werror=strict-prototypes]
drivers/common/mlx5/linux/mlx5_glue.h:188:2:
error: function declaration isn't a
mlx5 PCI Device supports multiple classes of devices such as net, vdpa,
and/or regex.
To support these multiple classes, change mlx5_class to a
bitmap values so that if users asks to enable multiple of them, all
supported classes can be parsed.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
-
Add generic mlx5 PCI PMD layer as part of existing common_mlx5
module. This enables multiple classes (net, regex, vdpa) PMDs
to be supported at same time.
Signed-off-by: Parav Pandit
Acked-by: Matan Azrad
---
Changelog:
v9->v10:
- Fixed alignment
v8->v9:
- Added missing LDFLAG for pci bus
- F
1 - 100 of 132 matches
Mail list logo
