This patch reverts:
commit 3a6f2eb8c5c5 ("net/mlx5: fix Memory Region registration")
Although granularity of chunks in a mempool is a cacheline, addresses are
extended to align to page boundary for performance reason in device when
registering a MR (Memory Region). This could make some reg
There're new metadata IPV4FRAG and IPV6FRAG in PPP
profile, this patch improves ptype parser to support
IPV4FRAG and IPV6FRAG.
Signed-off-by: Beilei Xing
---
drivers/net/i40e/i40e_ethdev.c | 44 +++---
1 file changed, 28 insertions(+), 16 deletions(-)
diff --
This patchset is mainly for fixing fail to update SW ptype
table and adding parser for IPV4FRAG and IPV6FRAG.
v4 changes:
- Slipt patchset and replace strncmp with strncasecmp.
v3 changes:
- Reorder IPv4 case.
v2 changes:
- Add parser for IPV4FRAG and IPV6FRAG.
Beilei Xing (3):
net/i40e: fix
Replace strncmp with strncasecmp in i40e_update_customized_ptype
function for compatibility.
Signed-off-by: Beilei Xing
---
drivers/net/i40e/i40e_ethdev.c | 55 +++---
1 file changed, 30 insertions(+), 25 deletions(-)
diff --git a/drivers/net/i40e/i40e_ethdev
Fail to update SW ptype mapping table when loading
PPP profile, though profile can be loaded successfully.
It will cause fail to parse SW ptype during receiving
packets. This patch fixes this issue.
Fixes: 11556c915a08 ("net/i40e: improve packet type parser")
Cc: sta...@dpdk.org
Signed-off-by: Be
On 01/19/2018 03:44 AM, Yongseok Koh wrote:
This commit introduces rte_dma_wmb() and rte_dma_rmb(), in order to
guarantee the ordering of coherent shared memory between the CPU and a DMA
capable device.
Signed-off-by: Yongseok Koh
Acked-by: Andrew Rybchenko
It is already really good. Many t
Hi Neil
From: Neil Horman, Friday, January 19, 2018 3:41 AM
> On Thu, Jan 18, 2018 at 08:21:34PM +, Matan Azrad wrote:
> > Hi Neil.
> >
> > From: Neil Horman, Thursday, January 18, 2018 8:42 PM
> > 1. What exactly do you want to improve?(in details) 2. Which API
> > specifically do you want
> On Jan 18, 2018, at 4:21 AM, Ananyev, Konstantin
> wrote:
>
> Hi,
>
>> -Original Message-
>> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Yongseok Koh
>> Sent: Wednesday, January 17, 2018 6:19 PM
>> To: Lu, Wenzhuo ; Wu, Jingjing
>> Cc: dev@dpdk.org; Yongseok Koh
>> Subject
The VIRTIO_F_ANY_LAYOUT feature indicates the device accepts arbitrary
descriptor layouts. The vhost-user lib already supports it, but the
feature declaration is missing. This patch fixes the mismatch.
Signed-off-by: Zhihong Wang
---
lib/librte_vhost/vhost.h | 1 +
1 file changed, 1 insertion(+
Hi,
It seems more logical to me to introduce tap_program (patch 3) before
its compiled version (patch 2).
Source code is indeed written down before compiling it.
The doc section is a good addition.
I'll be happy to see the upcoming utility for turning eBPF bytecode to C
arrays.
I'd have liked to
From: Andy Moreton
The type_data argument to ef10_rx_qcreate is only used
in builds with EFSYS_OPT_RX_PACKED_STREAM. note this as
an usued argument to avoid warnings in builds without
packed stream support.
Fixes: b749646dade4 ("net/sfc/base: add function to create packed stream RxQ")
Signed-of
align the config option name with config/common_base
Fixes: aaa4a221da26 ("event/sw: add new software-only eventdev driver")
Cc: sta...@dpdk.org
Signed-off-by: Jerin Jacob
---
drivers/event/sw/sw_evdev.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/event/sw/sw_evd
> On Jan 18, 2018, at 8:13 AM, Olivier Matz wrote:
>
> On Thu, Jan 18, 2018 at 05:04:27PM +0100, Nélio Laranjeiro wrote:
>> On Thu, Jan 18, 2018 at 02:00:42PM +0100, Olivier Matz wrote:
>>> On error, mlx5_dev_start() does not return a negative value
>>> as it is supposed to do. The consequence is
Yes, this is the target. I was in a 3-day meeting this week, and had limited
time to update the patch.
A new patch v6 was sent out just now. It's actually sent twice -- forgot to
add version info and "--in-reply-to" in the first one. Please just ignore that
and I'm sorry for the disturbanc
This patch provides an option to do rte_memcpy() using 'restrict'
qualifier, which can induce GCC to do optimizations by using more
efficient instructions, providing some performance gain over memcpy()
on some ARM64 platforms/enviroments.
The memory copy performance differs between different ARM64
This patch provides an option to do rte_memcpy() using 'restrict'
qualifier, which can induce GCC to do optimizations by using more
efficient instructions, providing some performance gain over memcpy()
on some ARM64 platforms/enviroments.
The memory copy performance differs between different ARM64
FDIR Rx ring isn't initialized and Rx queue HW tail isn't updated
when there's error detected during programming FDIR flow. There'll
be some potential risk.
This patch updates FDIR Rx resource.
Fixes: a778a1fa2e4e ("i40e: set up and initialize flow director")
Fixes: 05999aab4ca6 ("i40e: add or del
> -Original Message-
> From: Xing, Beilei
> Sent: Friday, January 19, 2018 10:50 AM
> To: Zhang, Qi Z ; Wu, Jingjing
> Cc: dev@dpdk.org; sta...@dpdk.org
> Subject: [PATCH v2] net/i40e: fix fdir Rx resource defect
>
> FDIR Rx ring isn't initialized and Rx queue HW tail isn't updated
> wh
Commit second time to cc sta...@dpdk.org
https://dpdk.org/dev/patchwork/patch/34065/
> -Original Message-
> From: Zhao1, Wei
> Sent: Friday, January 19, 2018 11:15 AM
> To: 'Thomas Monjalon'
> Cc: Zhang, Helin ; dev@dpdk.org; Mcnamara, John
> ; Yigit, Ferruh ; Xing,
> Beilei
> Subject:
This patch add inforation about i40e queue region
realted to release notes, it has been missed before
in v17.11 release notes. This feature has been
implemented in v17.11.
Signed-off-by: Wei Zhao
---
v2:
-change this information to v18.02 release notes.
v3:
-rework it on dpdk-next-net-intel su
Hi, Thomas
Thank you! After discussion with Beilei, I have commit a new patch v4.
https://dpdk.org/dev/patchwork/patch/33956/
Beilei also commit another new patch, we are waiting for review by Mcnamara,
John.
Do you think this is ok?
> -Original Message-
> From: Thomas Monjalon [ma
Set the value of m->refcnt to 1, m->nb_segs to 1 and m->next to NULL
when the mbuf is initialized or stored inside the mempool (unused).
All of these are done in rte_pktmbuf_pool_create() and
rte_pktmbuf_prefree_seg().
So we remove the redundant code from i40e, ixgbe and avf module.
Fixes: 8f094a
On 1/19/2018 9:13 AM, Thomas Monjalon wrote:
18/01/2018 05:12, Jeff Guo:
+ * It registers the callback for the specific device.
+ * Multiple callbacks cal be registered at the same time.
+ *
+ * @param device_name
+ * The device name, that is the param name of the struct rte_device,
Why not
FDIR Rx ring isn't initialized and Rx queue HW tail isn't updated
when there's error detected during programming FDIR flow. There'll
be some potential risk.
This patch updates FDIR Rx resource.
Fixes: a778a1fa2e4e ("i40e: set up and initialize flow director")
Fixes: 05999aab4ca6 ("i40e: add or del
On 1/19/2018 7:52 AM, Thomas Monjalon Wrote:
17/01/2018 10:09, Thomas Monjalon:
17/01/2018 09:47, Jia He:
Hi Thomas
On 1/17/2018 4:24 PM, Thomas Monjalon Wrote:
17/01/2018 05:03, Jia He:
To support C11 memory model barrier, 2 options are suggested by Jerin:
1. use rte_smp_rmb
2. use load_a
fix one typo and a grammatical mistake.
Fixes: b0152b1b40fe ("doc: update bonding")
Cc: sta...@dpdk.org
Signed-off-by: Zhiyong Yang
Acked-by: Marko Kovacevic
---
Changes in V3:
use "similar capabilities" instead of "the similar capabilities"
Changes in V2:
Fix the title and fixline format iss
Thanks, V3 will be coming soon.
> -Original Message-
> From: Kovacevic, Marko
> Sent: Monday, January 15, 2018 7:22 PM
> To: Yang, Zhiyong ; dev@dpdk.org
> Cc: Mcnamara, John
> Subject: RE: [dpdk-dev] [PATCH] doc: fix link bonding pmd typo in prog guide
>
> Actually just one more change
Hi Maxime,
> -Original Message-
> From: Maxime Coquelin [mailto:maxime.coque...@redhat.com]
> Sent: Thursday, January 18, 2018 5:04 PM
> To: Yang, Zhiyong ; dev@dpdk.org;
> y...@fridaylinux.org
> Cc: Wang, Wei W ; Tan, Jianfeng
>
> Subject: Re: [PATCH 00/11] net/vhostpci: A new vhostpci P
On Thu, Jan 18, 2018 at 08:21:34PM +, Matan Azrad wrote:
> Hi Neil.
>
> From: Neil Horman, Thursday, January 18, 2018 8:42 PM
>
>
> > > > But thats not all. The determination of success or failure in
> > > > claiming ownership is largely dependent on the behavior of other
> > > > threads ac
18/01/2018 05:12, Jeff Guo:
> + * It registers the callback for the specific device.
> + * Multiple callbacks cal be registered at the same time.
> + *
> + * @param device_name
> + * The device name, that is the param name of the struct rte_device,
Why not using rte_device pointer?
> + * null v
18/01/2018 05:23, Guo, Jia:
>
> On 1/18/2018 5:59 AM, Thomas Monjalon wrote:
> > 15/01/2018 11:48, Jeff Guo:
> >> + * It registers the callback for the specific event. Multiple
> >> + * callbacks cal be registered at the same time.
> >> + *
> >> + * @param device_name
> >> + * The device name.
>
11/01/2018 20:23, Amr Mokhtar:
> Hello,
> Latest update for Wireless Device abstraction library.
Applied with few changes:
- fixed title formats
- moved bbdev between ethdev and cryptodev in lists
- added in test-build.sh
- fixed compilation as shared library (must be better fixed in RC2)
Polling a new packet is basically sensing the generation bit in a
completion entry. For some processors not having strongly-ordered memory
model, there has to be an IO memory barrier between reading the generation
bit and other fields of the entry in order to guarantee data is not stale.
Fixes: 57
As rte_write64() has an IO barrier, there's no need to have a barrier
before the call.
Signed-off-by: Yongseok Koh
---
drivers/net/mlx5/mlx5_rxq.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 950472754..11438c86a 100644
--- a/
Cc: Thomas Speier
Signed-off-by: Yongseok Koh
Acked-by: Thomas Speier
Acked-by: Jianbo Liu
---
lib/librte_eal/common/include/arch/arm/rte_atomic_64.h | 4
1 file changed, 4 insertions(+)
diff --git a/lib/librte_eal/common/include/arch/arm/rte_atomic_64.h
b/lib/librte_eal/common/include
Signed-off-by: Yongseok Koh
---
drivers/net/mlx5/mlx5_rxtx.c | 4 ++--
drivers/net/mlx5/mlx5_rxtx.h | 2 +-
drivers/net/mlx5/mlx5_rxtx_vec.h | 2 +-
3 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_rxtx.c b/drivers/net/mlx5/mlx5_rxtx.c
index 3b8f71c28..
Signed-off-by: Yongseok Koh
Acked-by: Jianbo Liu
---
lib/librte_eal/common/include/arch/arm/rte_atomic_32.h | 4
1 file changed, 4 insertions(+)
diff --git a/lib/librte_eal/common/include/arch/arm/rte_atomic_32.h
b/lib/librte_eal/common/include/arch/arm/rte_atomic_32.h
index 14c048640..a4
Signed-off-by: Yongseok Koh
---
lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h | 4
1 file changed, 4 insertions(+)
diff --git a/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
b/lib/librte_eal/common/include/arch/ppc_64/rte_atomic.h
index 150810cdb..46490f2b3 100644
--- a/l
Signed-off-by: Yongseok Koh
---
lib/librte_eal/common/include/arch/x86/rte_atomic.h | 4
1 file changed, 4 insertions(+)
diff --git a/lib/librte_eal/common/include/arch/x86/rte_atomic.h
b/lib/librte_eal/common/include/arch/x86/rte_atomic.h
index 36cfabc38..ae41e615f 100644
--- a/lib/librte
This commit introduces rte_dma_wmb() and rte_dma_rmb(), in order to
guarantee the ordering of coherent shared memory between the CPU and a DMA
capable device.
Signed-off-by: Yongseok Koh
---
lib/librte_eal/common/include/generic/rte_atomic.h | 52 ++
1 file changed, 52 insert
This patchset is to introduce DMA memory barriers, which could be more
efficient for coherent memory between I/O device and CPU, especially for
ARMv8.
v3:
* add more detailed comments about the new memory barriers.
v2:
* introduce DMA memory barriers.
Yongseok Koh (8):
eal: introduce DMA memor
11/01/2018 20:23, Amr Mokhtar:
> +export DIR_WIRELESS_SDK=/FlexRAN-1.3.0/SDK-R1.3.0/sdk/
This variable looks to be unused.
Adding this in test-build.sh (while applying):
test -z "$FLEXRAN_SDK" || \
sed -ri 's,(BBDEV_TURBO_SW=)n,\1y,' $1/.config
> -Original Message-
> From: Pavan Nikhilesh [mailto:pbhagavat...@caviumnetworks.com]
> Sent: Thursday, January 11, 2018 11:29 AM
> To: Carrillo, Erik G ;
> jerin.ja...@caviumnetworks.com; nipun.gu...@nxp.com;
> hemant.agra...@nxp.com
> Cc: dev@dpdk.org
> Subject: Re: [PATCH v6 08/23] eve
Ping
Are we targetting to integrate this optimization in DPDK 18.02?
I am expecting a v6, thanks.
15/01/2018 12:37, Thomas Monjalon:
> 15/01/2018 11:57, Herbert Guan:
> > Hi Thomas,
> >
> > From: Thomas Monjalon [mailto:tho...@monjalon.net]
> > > Hi,
> > >
> > > All the code is using ARM64, but
17/01/2018 10:09, Thomas Monjalon:
> 17/01/2018 09:47, Jia He:
> >
> > Hi Thomas
> >
> > On 1/17/2018 4:24 PM, Thomas Monjalon Wrote:
> > > 17/01/2018 05:03, Jia He:
> > >> To support C11 memory model barrier, 2 options are suggested by Jerin:
> > >> 1. use rte_smp_rmb
> > >> 2. use load_acquire/
04/12/2017 14:22, Pavan Nikhilesh:
> +/*
> + * BSD LICENSE
> + *
> + * Copyright (C) Cavium, Inc. 2017.
> + *
> + * Redistribution and use in source and binary forms, with or without
> + * modification, are permitted provided that the following conditions
> + * are met:
> + *
> + * *
11/12/2017 16:31, Bruce Richardson:
> On Mon, Nov 06, 2017 at 10:05:43AM -0800, Elza Mathew wrote:
> > Compile-time function selection can potentially lead to
> > lower performance on generic builds done by distros.
> > Replaced compile time flag checks with run-time
> > function selection.
> >
>
> >rte_member may have allocated a tailq entry or setum before failure,
> >so free them.
> >
> >Fixes: 857ed6c68cf2 ("member: implement main API")
Cc: sta...@dpdk.org
> >Cc: yipeng1.w...@intel.com
> >Signed-off-by: Anatoly Burakov
>
> Acked-by: Yipeng Wang
Applied, thanks
On Thu, Jan 18, 2018 at 11:22:51PM +0100, Thomas Monjalon wrote:
> 29/11/2017 20:17, Ferruh Yigit:
> > >>> On Thu, Oct 05, 2017 at 10:42:08PM +, Ophir Munk wrote:
> > This commit prevents control path operations from failing after a sub
> > device removal.
> >
> > Following
18/01/2018 17:54, santosh:
> On Thursday 18 January 2018 07:14 PM, Pavan Nikhilesh wrote:
> > The mbox messages are naturally aligned and in some cases compiler
> > optimization might disregard natural alignment.
> > Use volatile key word to force compiler to disable optimizing and
> > maintain ali
18/01/2018 16:18, Jasvinder Singh:
> Move all static variables defined in header file to c file.
>
> error log: /usr/bin/ld: Warning: size of symbol 'count' changed
> from 8 in test_pmd_perf.o to 24 in test_flow_classify.o
>
> Fixes: 50bdac5916d9 ("flow_classify: remove table id parameter from
17/01/2018 09:51, Hemant Agrawal:
> If the allocation is from a single memzone, optimize
> the phy-virt address conversions.
>
> Signed-off-by: Hemant Agrawal
> ---
> v2: use register memory area instead of new flag in mempool
Applied, thanks
08/12/2017 16:46, Olivier Matz:
> When RTE_MBUF_REFCNT_ATOMIC=n, the decrement of the mbuf reference
> counter uses an atomic operation. This is not necessary and impacts
> the performance (seen with TRex traffic generator).
>
> We cannot replace rte_atomic16_add_return() by rte_mbuf_refcnt_update
> -Original Message-
> From: Pavan Nikhilesh [mailto:pbhagavat...@caviumnetworks.com]
> Sent: Thursday, January 11, 2018 6:19 AM
> To: Carrillo, Erik G ;
> jerin.ja...@caviumnetworks.com; nipun.gu...@nxp.com;
> hemant.agra...@nxp.com
> Cc: dev@dpdk.org
> Subject: Re: [PATCH v6 15/23] eventt
18/01/2018 14:51, Matan Azrad:
> Previous fail-safe code didn't support probed sub-devices capture and
> failed when it tried to probe them.
>
> Skip fail-safe sub-device probing when it already was probed.
>
> Signed-off-by: Matan Azrad
> Cc: Gaetan Rivet
Was acked by Gaetan in v5.
29/11/2017 20:17, Ferruh Yigit:
> >>> On Thu, Oct 05, 2017 at 10:42:08PM +, Ophir Munk wrote:
> This commit prevents control path operations from failing after a sub
> device removal.
>
> Following are the failure steps:
> 1. The physical device is removed due to change
18/01/2018 21:52, Matan Azrad:
> From: Thomas Monjalon, Thursday, January 18, 2018 10:44 PM
> > 18/01/2018 17:35, Matan Azrad:
> > > rte_eth_dev_allocate(const char *name) {
> > > uint16_t port_id;
> > > - struct rte_eth_dev *eth_dev;
> > > + struct rte_eth_dev *eth_dev = NULL
18/01/2018 17:35, Matan Azrad:
> The ownership of a port is implicit in DPDK.
> Making it explicit is better from the next reasons:
> 1. It will define well who is in charge of the port usage synchronization.
> 2. A library could work on top of a port.
> 3. A port can work on top of another port.
>
From: Thomas Monjalon, Thursday, January 18, 2018 10:44 PM
> 18/01/2018 17:35, Matan Azrad:
> > rte_eth_dev_allocate(const char *name) {
> > uint16_t port_id;
> > - struct rte_eth_dev *eth_dev;
> > + struct rte_eth_dev *eth_dev = NULL;
> > +
> > + /* Synchronize share
18/01/2018 17:35, Matan Azrad:
> rte_eth_dev_allocate(const char *name)
> {
> uint16_t port_id;
> - struct rte_eth_dev *eth_dev;
> + struct rte_eth_dev *eth_dev = NULL;
> +
> + /* Synchronize share data one time allocation between local threads.
> */
I don't understand
Hi Neil.
From: Neil Horman, Thursday, January 18, 2018 8:42 PM
> > > But thats not all. The determination of success or failure in
> > > claiming ownership is largely dependent on the behavior of other
> > > threads actions, not a function of the state of the system at the moment
> ownership is
18/01/2018 19:26, Stephen Hemminger:
> On Thu, 18 Jan 2018 08:43:44 +
> Matan Azrad wrote:
>
> > NetVSC netdevices which are already routed should not be probed because
> > they are used for management purposes by the HyperV.
> >
> > prevent routed netvsc devices probing.
> >
> > Signed-off
On Thu, Jan 18, 2018 at 05:20:31PM +, Matan Azrad wrote:
> Hi Neil
>
> From: Neil Horman, Thursday, January 18, 2018 6:55 PM
> > On Thu, Jan 18, 2018 at 02:00:23PM +, Matan Azrad wrote:
> > > Hi Neil
> > >
> > > From: Neil Horman, Thursday, January 18, 2018 3:10 PM
> > > > On Wed, Jan 17,
From: Stephen Hemminger, Thursday, January 18, 2018 8:28 PM
> On Thu, 18 Jan 2018 08:43:45 +
> Matan Azrad wrote:
>
> > This parameter allows specifying any non-NetVSC interface or routed
> > NetVSC interfaces to use with tap sub-devices for development purposes.
> >
> > Signed-off-by: Adri
From: Stephen Hemminger, Thursday, January 18, 2018 8:26 PM
> On Thu, 18 Jan 2018 08:43:43 +
> Matan Azrad wrote:
>
> > +
> > +/**
> > + * Alarm callback that regularly probes system network interfaces.
> > + *
> > + * This callback runs at a frequency determined by
> > +VDEV_NETVSC_PROBE_M
On Thu, 18 Jan 2018 08:43:45 +
Matan Azrad wrote:
> This parameter allows specifying any non-NetVSC interface or routed
> NetVSC interfaces to use with tap sub-devices for development purposes.
>
> Signed-off-by: Adrien Mazarguil
> Signed-off-by: Matan Azrad
Might whitelist work for this?
On Thu, 18 Jan 2018 08:43:44 +
Matan Azrad wrote:
> NetVSC netdevices which are already routed should not be probed because
> they are used for management purposes by the HyperV.
>
> prevent routed netvsc devices probing.
>
> Signed-off-by: Raslan Darawsheh
> Signed-off-by: Matan Azrad
J
On Thu, 18 Jan 2018 08:43:43 +
Matan Azrad wrote:
> +
> +/**
> + * Alarm callback that regularly probes system network interfaces.
> + *
> + * This callback runs at a frequency determined by VDEV_NETVSC_PROBE_MS as
> + * long as an vdev_netvsc context instance exists.
> + *
> + * @param arg
>
> On Jan 18, 2018, at 3:56 AM, Andrew Rybchenko
> wrote:
>
> On 01/17/2018 09:39 PM, Yongseok Koh wrote:
>>> On Jan 17, 2018, at 5:46 AM, Thomas Monjalon wrote:
>>>
>>> 16/01/2018 10:10, Jianbo Liu:
The 01/16/2018 10:49, Andrew Rybchenko wrote:
> On 01/16/2018 04:10 AM, Yongseok Koh
Hi Ferruh
From: Ferruh Yigit, Thursday, January 18, 2018 7:31 PM
> On 1/18/2018 11:27 AM, Matan Azrad wrote:
> > rte_eth_dev_is_removed API was added to detect a device removal
> > synchronously.
> >
> > When a device removal occurs during control command execution, many
> > different errors can b
Hi Ferruh
From: Ferruh Yigit, Thursday, January 18, 2018 7:18 PM
> On 1/18/2018 11:27 AM, Matan Azrad wrote:
> > There is time between the physical removal of the device until PMDs
> > get a RMV interrupt. At this time DPDK PMDs and applications still
> > don't know about the removal.
> >
> > Curr
On Thu, Jan 18, 2018 at 05:18:22PM +, Ferruh Yigit wrote:
> On 1/18/2018 11:27 AM, Matan Azrad wrote:
> > There is time between the physical removal of the device until PMDs get
> > a RMV interrupt. At this time DPDK PMDs and applications still don't
> > know about the removal.
> >
> > Current
On 1/18/2018 11:27 AM, Matan Azrad wrote:
> rte_eth_dev_is_removed API was added to detect a device removal
> synchronously.
>
> When a device removal occurs during control command execution, many
> different errors can be reported to the user.
>
> Adjust all ethdev APIs error reports to return -
Hi Neil
From: Neil Horman, Thursday, January 18, 2018 6:55 PM
> On Thu, Jan 18, 2018 at 02:00:23PM +, Matan Azrad wrote:
> > Hi Neil
> >
> > From: Neil Horman, Thursday, January 18, 2018 3:10 PM
> > > On Wed, Jan 17, 2018 at 05:01:10PM +, Ananyev, Konstantin wrote:
> > > >
> > > >
> > > >
On 1/18/2018 11:27 AM, Matan Azrad wrote:
> There is time between the physical removal of the device until PMDs get
> a RMV interrupt. At this time DPDK PMDs and applications still don't
> know about the removal.
>
> Current removal detection is achieved only by registration to device RMV
> event
18/01/2018 17:35, Matan Azrad:
> rte_eth_dev_find_free_port() found a free port by state checking.
> The state field are in local process memory, so other DPDK processes
> may get the same port ID because their local states may be different.
>
> Replace the state checking by the ethdev port name c
18/01/2018 17:35, Matan Azrad:
> rte_eth_dev_data structure is allocated per ethdev port and can be
> used to get a data of the port internally.
>
> rte_eth_dev_attach_secondary tries to find the port identifier using
> rte_eth_dev_data name field comparison and may get an identifier of
> invalid
On Thu, Jan 18, 2018 at 11:27:11AM +, Matan Azrad wrote:
> Add support to get removal status of mlx5 device.
> It is not supported in secondary process.
>
> Signed-off-by: Matan Azrad
Acked-by: Adrien Mazarguil
--
Adrien Mazarguil
6WIND
On Thu, Jan 18, 2018 at 11:27:10AM +, Matan Azrad wrote:
> Add support to get removal status of mlx4 device.
>
> Signed-off-by: Matan Azrad
Acked-by: Adrien Mazarguil
--
Adrien Mazarguil
6WIND
On Thu, Jan 18, 2018 at 02:00:23PM +, Matan Azrad wrote:
> Hi Neil
>
> From: Neil Horman, Thursday, January 18, 2018 3:10 PM
> > On Wed, Jan 17, 2018 at 05:01:10PM +, Ananyev, Konstantin wrote:
> > >
> > >
> > > > -Original Message-
> > > > From: Neil Horman [mailto:nhor...@tuxdriv
On Thursday 18 January 2018 07:14 PM, Pavan Nikhilesh wrote:
> The mbox messages are naturally aligned and in some cases compiler
> optimization might disregard natural alignment.
> Use volatile key word to force compiler to disable optimizing and
> maintain alignment.
>
> Fixes: aecb8e093b52 ("ev
On 1/18/2018 3:18 PM, Jasvinder Singh wrote:
> Move all static variables defined in header file to c file.
>
> error log: /usr/bin/ld: Warning: size of symbol 'count' changed
> from 8 in test_pmd_perf.o to 24 in test_flow_classify.o
>
> Fixes: 50bdac5916d9 ("flow_classify: remove table id param
On 1/18/2018 1:06 PM, maciej.cze...@caviumnetworks.com wrote:
> From: Maciej Czekaj
Author fixed as "Maciej Czekaj "
s/caviunetworks/caviumnetworks
>
> This patch removes all references to old-style offload API
> replacing them with new offload flags.
>
> Signed-off-by: Maciej Czekaj
Applie
Fail-safe PMD sub devices management is based on ethdev port mechanism.
So, the sub-devices management structures are exposed to other DPDK
entities which may use them in parallel to fail-safe PMD.
Use the new port ownership mechanism to avoid multiple managments of
fail-safe PMD sub-devices.
Sig
Testpmd should not use ethdev ports which are managed by other DPDK
entities.
Set Testpmd ownership to each port which is not used by other entity and
prevent any usage of ethdev ports which are not owned by Testpmd.
Signed-off-by: Matan Azrad
---
app/test-pmd/cmdline.c | 89 ++
Call dedicated ethdev API to free port in remove time as was done in
other fail-safe places.
Signed-off-by: Matan Azrad
Acked-by: Gaetan Rivet
---
drivers/net/failsafe/failsafe_ether.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/failsafe/failsafe_ether.c
b/d
Ethernet port allocation was not thread safe, means 2 threads which tried
to allocate a new port at the same time might get an identical port
identifier and caused to memory overwrite.
Actually, all the port configurations were not thread safe from ethdev
point of view.
The port ownership mechanis
The ownership of a port is implicit in DPDK.
Making it explicit is better from the next reasons:
1. It will define well who is in charge of the port usage synchronization.
2. A library could work on top of a port.
3. A port can work on top of another port.
Also in the fail-safe case, an issue has
rte_eth_dev_data structure is allocated per ethdev port and can be
used to get a data of the port internally.
rte_eth_dev_attach_secondary tries to find the port identifier using
rte_eth_dev_data name field comparison and may get an identifier of
invalid port in case of this port was released by t
Add ownership mechanism to DPDK Ethernet devices to avoid multiple management
of a device by different DPDK entities.
The port ownership mechanism is a good point to redefine the synchronization
rules in ethdev:
1. The port allocation and port release synchronization will be managed
by
rte_eth_dev_find_free_port() found a free port by state checking.
The state field are in local process memory, so other DPDK processes
may get the same port ID because their local states may be different.
Replace the state checking by the ethdev port name checking,
so, if the name is an empty stri
On Wed, Jan 17, 2018 at 05:01:10PM +, Ananyev, Konstantin wrote:
>
>
> > -Original Message-
> > From: Neil Horman [mailto:nhor...@tuxdriver.com]
> > Sent: Wednesday, January 17, 2018 2:00 PM
> > To: Matan Azrad
> > Cc: Ananyev, Konstantin ; Thomas Monjalon
> > ; Gaetan Rivet
> > ; W
On 1/18/2018 10:57 AM, Tomasz Duszynski wrote:
> This patch-set contains a couple of changes which let the application
> using pmd to set mtu and mac addresses before port is initialized.
>
> Tomasz Duszynski (2):
> net/mrvl: allow changing MTU before port init
> net/mrvl: allow adding ether a
On Thu, Jan 18, 2018 at 02:00:43PM +0100, Olivier Matz wrote:
> If there is no memory available on the same numa node than the
> device, it is preferable to fallback on another socket instead
> of failing.
>
> Fixes: 1e3a39f72d5d ("net/mlx5: allocate verbs object into shared memory")
> Cc: sta...@
After trying RSS and Queue also try End for flow action
to allow for HW that don't support flow features with
inline crypto.
Signed-off-by: Radu Nicolau
---
examples/ipsec-secgw/ipsec.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/examples/ipsec-secgw/ipsec.c b/examples/ipsec-secgw/
On Thu, Jan 18, 2018 at 05:04:27PM +0100, Nélio Laranjeiro wrote:
> On Thu, Jan 18, 2018 at 02:00:42PM +0100, Olivier Matz wrote:
> > On error, mlx5_dev_start() does not return a negative value
> > as it is supposed to do. The consequence is that the application
> > (ex: testpmd) does not notice th
On Thu, Jan 18, 2018 at 02:00:42PM +0100, Olivier Matz wrote:
> On error, mlx5_dev_start() does not return a negative value
> as it is supposed to do. The consequence is that the application
> (ex: testpmd) does not notice that the port is not started
> and begins the rxtx on an uninitialized port,
On Thu, Jan 18, 2018 at 11:48:09PM +0800, Tiwei Bie wrote:
> On Thu, Jan 18, 2018 at 03:55:53PM +0100, Olivier Matz wrote:
> > On Thu, Jan 18, 2018 at 10:05:49PM +0800, Tiwei Bie wrote:
> > > On Thu, Jan 18, 2018 at 10:07:32AM +0100, Olivier Matz wrote:
> > > > Rationalize the function virtio_dev_f
Also moved offloaded packets from cryptodev queues
Signed-off-by: Radu Nicolau
Reviewed-by: Akhil Goyal
---
examples/ipsec-secgw/ipsec-secgw.c | 42 +-
examples/ipsec-secgw/ipsec.c | 12 ---
examples/ipsec-secgw/ipsec.h | 4 ++--
3 files
Add function to get security session size.
Signed-off-by: Radu Nicolau
---
lib/librte_security/rte_security.c | 7 +++
lib/librte_security/rte_security.h | 12
lib/librte_security/rte_security_driver.h| 14 ++
lib/librte_security/rte_security
1 - 100 of 259 matches
Mail list logo
