Re: [dpdk-dev] [PATCH] devtools: check stable tag in fixes

[Yuanhan Liu]

On Wed, Jan 18, 2017 at 05:25:10PM +, Ferruh Yigit wrote:
> On 1/18/2017 4:41 AM, Yuanhan Liu wrote:
> > On Tue, Jan 17, 2017 at 07:42:33PM +0100, Thomas Monjalon wrote:
> >> 2017-01-17 18:15, Ferruh Yigit:
> >>> On 1/17/2017 2:54 PM, Thomas Monjalon wrote:
>  The tag "Cc: sta...@dpdk.org" must be set when the commit must be
>  backported to a stable branch.
> 
>  It must be located just below the "Fixes:" tag (without blank line)
>  and followed by a blank line, separated from SoB and review tags below.
> >>>
> >>> I am OK to keep it if it will help stable tree maintenance, but I still
> >>> not clear about why we need this.
> >>>
> >>> If a patch is a fix, it should already have "Fixes:" line, so this can
> >>> be used to parse git history.
> > 
> > Same answer (as I have already replied to you in another email): not all fix
> > patches should be picked to stable branch. (I gave some examples below)
> 
> I was thinking all fixes will have "Cc: stable" tag, to be sure all
> fixes sent to stable mail list, and you will be the picking for stable tree.
> 
> Now you are suggesting some fixes may have "Fixes:" tag but not "Cc:
> stable" tag.
> 
> So this means now author/maintainer/committer will be the picking
> patches for stable tree, and to show this decision, will add "Cc:
> stable" to the commit log.
> 
> If author puts the "Cc: stable" tag, git send-email will ensure this
> patch will be sent to stable mail list too.
> But if author missed the "Cc: stable" tag, will it be enough for you if
> committer adds the tag into commit log? Or should patch sent to stable
> mail list too?

Yes, that'd be enough. I will write a short script to monitor the
official DPDK git tree regularly (say weekly), to see whether there
are any new candidates (those with 'cc: stable' tag) for stable
releases or not. If so, I will pick them to stable.

> 
> What is the relation between "Cc: stable" tag been in commit log and
> patch been sent to stable mail list?

Not too much. The stable mailing list provides a place for further
discussion on stable patches:

- A notification will be sent to the stable mailing list when a patch
  is chosen as a candidate for a stable release. People can shout out
  if he thinks some patches should not have been picked for a stable
  release.

- people can also jump out to say, "hey, seems you forgot commit xxx,
  which is a good have for stable release".

--yliu

Re: [dpdk-dev] [dpdk-stable] [PATCH] net/virtio-user: fix missing driver name

[Yuanhan Liu]

On Thu, Jan 19, 2017 at 09:16:58AM +0800, Tan, Jianfeng wrote:
> > >> On 1/18/2017 11:44 AM, Jianfeng Tan wrote:
> > >>> API rte_eth_dev_info_get() fills driver name according to drv_name
> > >>> of rte_eth_dev_data. But we have not fill such info in virtio_user.
> > >>>
> > >>> We do not use the same one with virtio device as some applications
> > >>> might depend on driver name to differetiate kinds of devices, such
> > >>> as VPP.
> > >>>
> > >>> Fixes: e9efa4d93821 ("net/virtio-user: add new virtual PCI driver")
> > >>> CC: sta...@dpdk.org
> > >>>
> > >>> Signed-off-by: Jianfeng Tan 
> > >>> ---
> > >>>  drivers/net/virtio/virtio_user_ethdev.c | 1 +
> > >>>  1 file changed, 1 insertion(+)
> > >>>
> > >>> diff --git a/drivers/net/virtio/virtio_user_ethdev.c
> > >> b/drivers/net/virtio/virtio_user_ethdev.c
> > >>> index c877968..110f6a9 100644
> > >>> --- a/drivers/net/virtio/virtio_user_ethdev.c
> > >>> +++ b/drivers/net/virtio/virtio_user_ethdev.c
> > >>> @@ -312,6 +312,7 @@ virtio_user_eth_dev_alloc(const char *name)
> > >>> hw->use_simple_rxtx = 0;
> > >>> hw->virtio_user_dev = dev;
> > >>> data->dev_private = hw;
> > >>> +   data->drv_name = "virtio_user PMD";
> > >>
> > >> How driver naming done changed in next-net.
> > >> Please check any other virtual driver in next-net for sample.
> > >
> > > Thank you for the info. And it seems that it's already fixed by that 
> > > patch.
> > 
> > That patch modified during merge, can you please confirm the latest
> > version in the repo?
> 
> Just checked latest master repo, this problem has been fixed by commit 
> 73db5badb04 ("net: align ethdev and eal driver names"). Thanks to David 
> Marchand.
> 
> Yuanhan, do you think this patch can be applied to the stable repo?

I think I would partially backport that commit to a stable release: just
pick the virtio changes, with a brand new commit log, stating the issue
need to be fixed and it's a partial backport from 73db5badb04 ("net: align
ethdev and eal driver names").

Okay to you?

--yliu

Re: [dpdk-dev] [dpdk-stable] [PATCH] net/virtio-user: fix missing driver name

[Tan, Jianfeng]

> -Original Message-
> From: Yuanhan Liu [mailto:yuanhan@linux.intel.com]
> Sent: Thursday, January 19, 2017 4:16 PM
> To: Tan, Jianfeng
> Cc: Yigit, Ferruh; dev@dpdk.org; sta...@dpdk.org
> Subject: Re: [dpdk-stable] [PATCH] net/virtio-user: fix missing driver name
> 
> On Thu, Jan 19, 2017 at 09:16:58AM +0800, Tan, Jianfeng wrote:
> > > >> On 1/18/2017 11:44 AM, Jianfeng Tan wrote:
> > > >>> API rte_eth_dev_info_get() fills driver name according to drv_name
> > > >>> of rte_eth_dev_data. But we have not fill such info in virtio_user.
> > > >>>
> > > >>> We do not use the same one with virtio device as some applications
> > > >>> might depend on driver name to differetiate kinds of devices, such
> > > >>> as VPP.
> > > >>>
> > > >>> Fixes: e9efa4d93821 ("net/virtio-user: add new virtual PCI driver")
> > > >>> CC: sta...@dpdk.org
> > > >>>
> > > >>> Signed-off-by: Jianfeng Tan 
> > > >>> ---
> > > >>>  drivers/net/virtio/virtio_user_ethdev.c | 1 +
> > > >>>  1 file changed, 1 insertion(+)
> > > >>>
> > > >>> diff --git a/drivers/net/virtio/virtio_user_ethdev.c
> > > >> b/drivers/net/virtio/virtio_user_ethdev.c
> > > >>> index c877968..110f6a9 100644
> > > >>> --- a/drivers/net/virtio/virtio_user_ethdev.c
> > > >>> +++ b/drivers/net/virtio/virtio_user_ethdev.c
> > > >>> @@ -312,6 +312,7 @@ virtio_user_eth_dev_alloc(const char *name)
> > > >>>   hw->use_simple_rxtx = 0;
> > > >>>   hw->virtio_user_dev = dev;
> > > >>>   data->dev_private = hw;
> > > >>> + data->drv_name = "virtio_user PMD";
> > > >>
> > > >> How driver naming done changed in next-net.
> > > >> Please check any other virtual driver in next-net for sample.
> > > >
> > > > Thank you for the info. And it seems that it's already fixed by that 
> > > > patch.
> > >
> > > That patch modified during merge, can you please confirm the latest
> > > version in the repo?
> >
> > Just checked latest master repo, this problem has been fixed by commit
> 73db5badb04 ("net: align ethdev and eal driver names"). Thanks to David
> Marchand.
> >
> > Yuanhan, do you think this patch can be applied to the stable repo?
> 
> I think I would partially backport that commit to a stable release: just
> pick the virtio changes, with a brand new commit log, stating the issue
> need to be fixed and it's a partial backport from 73db5badb04 ("net: align
> ethdev and eal driver names").
> 
> Okay to you?

That will be great. Actually, I'm wondering if possible to make your work 
easier by just adding another very simple fix only to stable branch.

Thanks,
Jianfeng

> 
>   --yliu

Re: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter

[Thomas Monjalon]

2017-01-19 13:34, Yong Liu:
> +* ABI changes are planned for 17.05: structure ``rte_eth_tunnel_filter_conf``
> +  will be extended with a new member ``vf_id`` in order to enable cloud 
> filter
> +  on VF device.

I think we should stop rely on this API, and migrate to rte_flow instead.
Adrien any thought?

[dpdk-dev] [PATCH] net/i40e: fix Rx checksum flag

[Xiao Wang]

When no error reported in Rx descriptor, we should set CKSUM_GOOD flag
before return.

Fixes: b704f9071b09 ("net/i40e: implement new Rx checksum flag")

Signed-off-by: Xiao Wang 
CC: sta...@dpdk.org
---
 drivers/net/i40e/i40e_rxtx.c | 5 -
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/net/i40e/i40e_rxtx.c b/drivers/net/i40e/i40e_rxtx.c
index 1c9a6c8..22809fa 100644
--- a/drivers/net/i40e/i40e_rxtx.c
+++ b/drivers/net/i40e/i40e_rxtx.c
@@ -151,8 +151,11 @@ static uint16_t i40e_xmit_pkts_simple(void *tx_queue,
uint64_t error_bits = (qword >> I40E_RXD_QW1_ERROR_SHIFT);
 
 #define I40E_RX_ERR_BITS 0x3f
-   if (likely((error_bits & I40E_RX_ERR_BITS) == 0))
+   if (likely((error_bits & I40E_RX_ERR_BITS) == 0)) {
+   flags |= (PKT_RX_IP_CKSUM_GOOD | PKT_RX_L4_CKSUM_GOOD);
return flags;
+   }
+
if (unlikely(error_bits & (1 << I40E_RX_DESC_ERROR_IPE_SHIFT)))
flags |= PKT_RX_IP_CKSUM_BAD;
else
-- 
1.8.3.1

Re: [dpdk-dev] [PATCH v6 0/8] Add crypto PMD optimized for ARMv8

[Zbigniew Bodek]

Thanks a lot!

Kind regards
Zbigniew

On 18.01.2017 22:14, De Lara Guarch, Pablo wrote:




-Original Message-
From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of
zbigniew.bo...@caviumnetworks.com
Sent: Wednesday, January 18, 2017 8:02 PM
To: dev@dpdk.org
Cc: De Lara Guarch, Pablo; Doherty, Declan;
jerin.ja...@caviumnetworks.com; jianbo@linaro.org;
hemant.agra...@nxp.com; Zbigniew Bodek
Subject: [dpdk-dev] [PATCH v6 0/8] Add crypto PMD optimized for ARMv8

From: Zbigniew Bodek 


Applied to dpdk-next-crypto.
Thanks,

Pablo

Re: [dpdk-dev] [PATCH v2] net/i40e: set no drop for traffic class

[Sexton, Rory]

Perhaps the best solution is as suggested to set
rte_eth_conf.dcb_capability_en = ETH_DCB_PFC_SUPPORT
rte_eth_conf.rxmode.mq_mode = ETH_MQ_RX_DCB_FLAG
and set rte_eth_dcb_rx_conf.nb_tcs to the number of tc's to apply
Using this port configuration will give the same behavior of the patch and it 
removes the need for an API change.

Rory

-Original Message-
From: Wu, Jingjing 
Sent: Tuesday, January 17, 2017 3:09 PM
To: Sexton, Rory 
Cc: dev@dpdk.org; Marjanovic, Nemanja 
Subject: RE: [PATCH v2] net/i40e: set no drop for traffic class



> -Original Message-
> From: Sexton, Rory
> Sent: Monday, January 16, 2017 11:52 PM
> To: Wu, Jingjing 
> Cc: dev@dpdk.org; Sexton, Rory ; Marjanovic, 
> Nemanja 
> Subject: [PATCH v2] net/i40e: set no drop for traffic class
> 
> From: Rory Sexton 
> 
> The default traffic class in i40e is set to drop versus on ixgbe it 
> isset to no drop. This means when packets build up in the RX SRAM on 
> the NIC, they are dropped, and they do this when the SW descriptor rings fill 
> up.
> 
> This patch changes this behaviour and our testing shows there are no 
> drops as a result.
> 
> Signed-off-by: Rory Sexton 
> Signed-off-by: Nemanja Marjanovic 
> ---
> v2:
> * Changed to use existing api to set priority register directly.
> 
>  drivers/net/i40e/i40e_ethdev.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/drivers/net/i40e/i40e_ethdev.c 
> b/drivers/net/i40e/i40e_ethdev.c index 67778ba..97339b5 100644
> --- a/drivers/net/i40e/i40e_ethdev.c
> +++ b/drivers/net/i40e/i40e_ethdev.c
> @@ -2985,8 +2985,11 @@ static int
>  i40e_priority_flow_ctrl_set(__rte_unused struct rte_eth_dev *dev,
>   __rte_unused struct rte_eth_pfc_conf *pfc_conf) {
> + struct i40e_hw *hw = I40E_DEV_PRIVATE_TO_HW(dev->data-
> >dev_private);
> +
>   PMD_INIT_FUNC_TRACE();
> 
> + I40E_WRITE_REG(hw, I40E_PRTDCB_TC2PFC, 0xff);

PRTDCB_TC2PFC  is the Bitmap who controls the use of Priority Flow Control 
(PFC) per each TC. Bit n set to 1b indicates TC n uses PFC in Rx and Tx. The TC 
is referred as a no-drop TC.

And if look the rte_eth_pfc_conf, there is a field called priority, which would 
map to a TC.
Currently, the TC and priority is 1:1 map when dcb is enabled.
So how about change it like:
Check dcb info, and map the priority to tc, then val = 0x1 << tc; 
I40E_WRITE_REG(hw, I40E_PRTDCB_TC2PFC, val);

Thanks
Jingjing



>   return -ENOSYS;
>  }
> 
> --
> 2.4.3

--
Intel Research and Development Ireland Limited
Registered in Ireland
Registered Office: Collinstown Industrial Park, Leixlip, County Kildare
Registered Number: 308263


This e-mail and any attachments may contain confidential material for the sole
use of the intended recipient(s). Any review or distribution by others is
strictly prohibited. If you are not the intended recipient, please contact the
sender and delete all copies.

[dpdk-dev] [PATCH] net/sfc: do not panic if alarms are not supported

[Andrew Rybchenko]

Alarms are not supported on the FreeBSD.
Application must poll link status periodically itself using
rte_eth_link_get_nowait() to avoid managment event queue overflow.

Fixes: 2de39f4e1310 ("net/sfc: periodic management EVQ polling using alarm")

Signed-off-by: Andrew Rybchenko 
Reviewed-by: Andrew Lee 
Reviewed-by: Andy Moreton 
---
 drivers/net/sfc/sfc_ev.c | 12 
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/drivers/net/sfc/sfc_ev.c b/drivers/net/sfc/sfc_ev.c
index c788986..fe6de6f 100644
--- a/drivers/net/sfc/sfc_ev.c
+++ b/drivers/net/sfc/sfc_ev.c
@@ -499,10 +499,14 @@
 
rc = rte_eal_alarm_set(SFC_MGMT_EV_QPOLL_PERIOD_US,
   sfc_ev_mgmt_periodic_qpoll, sa);
-   if (rc != 0)
-   sfc_panic(sa,
- "cannot rearm management EVQ polling alarm (rc=%d)",
- rc);
+   if (rc == -ENOTSUP) {
+   sfc_warn(sa, "alarms are not supported");
+   sfc_warn(sa, "management EVQ must be polled indirectly using 
no-wait link status update");
+   } else if (rc != 0) {
+   sfc_err(sa,
+   "cannot rearm management EVQ polling alarm (rc=%d)",
+   rc);
+   }
 }
 
 static void
-- 
1.8.2.3

[dpdk-dev] [PATCH] net/sfc: fix TSO limits imposed to the number of Tx queues

[Andrew Rybchenko]

From: Ivan Malov 

The number of Tx queues requested by the user must not be overridden;
instead, the limits imposed by TSO must be applied to the advertised
maximum

Fixes: fec33d5bb3eb ("net/sfc: support firmware-assisted TSO")

Signed-off-by: Ivan Malov 
Signed-off-by: Andrew Rybchenko 
Reviewed-by: Andrew Lee 
Reviewed-by: Andy Moreton 
---
 drivers/net/sfc/sfc.c| 26 +-
 drivers/net/sfc/sfc_tx.c |  5 -
 2 files changed, 17 insertions(+), 14 deletions(-)

diff --git a/drivers/net/sfc/sfc.c b/drivers/net/sfc/sfc.c
index 648ad8c..d35754c 100644
--- a/drivers/net/sfc/sfc.c
+++ b/drivers/net/sfc/sfc.c
@@ -203,6 +203,13 @@
limits.edl_max_txq_count =
MIN(encp->enc_txq_limit,
limits.edl_max_evq_count - 1 - limits.edl_max_rxq_count);
+
+   if (sa->tso)
+   limits.edl_max_txq_count =
+   MIN(limits.edl_max_txq_count,
+   encp->enc_fw_assisted_tso_v2_n_contexts /
+   encp->enc_hw_pf_count);
+
SFC_ASSERT(limits.edl_max_txq_count >= limits.edl_min_rxq_count);
 
/* Configure the minimum required resources needed for the
@@ -601,12 +608,21 @@
if (rc != 0)
goto fail_nic_reset;
 
+   encp = efx_nic_cfg_get(sa->nic);
+
+#ifdef RTE_LIBRTE_SFC_EFX_TSO
+   sa->tso = encp->enc_fw_assisted_tso_v2_enabled;
+   if (!sa->tso)
+   sfc_warn(sa, "TSO support isn't available on this adapter");
+#else /* !RTE_LIBRTE_SFC_EFX_TSO */
+   sa->tso = B_FALSE;
+#endif /* RTE_LIBRTE_SFC_EFX_TSO */
+
sfc_log_init(sa, "estimate resource limits");
rc = sfc_estimate_resource_limits(sa);
if (rc != 0)
goto fail_estimate_rsrc_limits;
 
-   encp = efx_nic_cfg_get(sa->nic);
sa->txq_max_entries = encp->enc_txq_max_ndescs;
SFC_ASSERT(rte_is_power_of_2(sa->txq_max_entries));
 
@@ -621,14 +637,6 @@
if (rc != 0)
goto fail_set_rss_defaults;
 
-#ifdef RTE_LIBRTE_SFC_EFX_TSO
-   sa->tso = efx_nic_cfg_get(sa->nic)->enc_fw_assisted_tso_v2_enabled;
-   if (!sa->tso)
-   sfc_warn(sa, "TSO support isn't available on this adapter");
-#else /* !RTE_LIBRTE_SFC_EFX_TSO */
-   sa->tso = B_FALSE;
-#endif /* RTE_LIBRTE_SFC_EFX_TSO */
-
sfc_log_init(sa, "fini nic");
efx_nic_fini(enp);
 
diff --git a/drivers/net/sfc/sfc_tx.c b/drivers/net/sfc/sfc_tx.c
index 3e64c0f..1292af3 100644
--- a/drivers/net/sfc/sfc_tx.c
+++ b/drivers/net/sfc/sfc_tx.c
@@ -312,11 +312,6 @@
 
sa->txq_count = sa->eth_dev->data->nb_tx_queues;
 
-   if (sa->tso)
-   sa->txq_count = MIN(sa->txq_count,
-  efx_nic_cfg_get(sa->nic)->enc_fw_assisted_tso_v2_n_contexts /
-  efx_nic_cfg_get(sa->nic)->enc_hw_pf_count);
-
sa->txq_info = rte_calloc_socket("sfc-txqs", sa->txq_count,
 sizeof(sa->txq_info[0]), 0,
 sa->socket_id);
-- 
1.8.2.3

[dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Jeff Guo]

From: "Guo, Jia" 

When bind with the "uio_pci_generic" module in X710/XL710/XXV710, the result
is failed. The "uio_pci_generic" module is not supported by X710/XL710/XXV710.

Signed-off-by: Guo, Jia 
Acked-by: Jingjing Wu 
---
v3: add related linux kernel commit info and fix some syntax issues
v2: add X710/XXV710 limitation case
---
 doc/guides/rel_notes/known_issues.rst | 26 ++
 1 file changed, 26 insertions(+)

diff --git a/doc/guides/rel_notes/known_issues.rst 
b/doc/guides/rel_notes/known_issues.rst
index 3cd4237..8b083a0 100644
--- a/doc/guides/rel_notes/known_issues.rst
+++ b/doc/guides/rel_notes/known_issues.rst
@@ -640,3 +640,29 @@ I40e VF may not receive packets in the promiscuous mode
 
 **Driver/Module**:
Poll Mode Driver (PMD).
+
+
+The uio_pci_generic module bind failed in X710/XL710/XXV710
+---
+
+**Description**:
+   The “uio_pci_generic” module is not supported by XL710, since the errata of 
XL710
+   states that the Interrupt Status bit is not implemented. The errata is the 
item #71
+   from the `xl710 controller spec
+   
`_.
+   The hw limitation is the same as other X710/XXV710 NICs.
+
+**Implication**:
+   When bind the “uio_pci_generic” module, it probe device and check the 
Interrupt Status
+   bit. Since it is not supported by X710/XL710/XXV710, it return a *failed* 
value. The statement
+   that these products don’t support INTx masking, is indicated in the related 
`linux kernel commit
+   < 
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/drivers/pci/quirks.c?id=8bcf4525c5d43306c5fd07e132bc8650e3491aec>`_.
+
+**Resolution/Workaround**:
+   Do not bind to use the “uio_pci_generic” module in X710/XL710/XXV710 NICs.
+
+**Affected Environment/Platform**:
+   All.
+
+**Driver/Module**:
+   Poll Mode Driver (PMD).
-- 
2.7.4

[dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Jeff Guo]

From: "Guo, Jia" 

When bind with the "uio_pci_generic" module in X710/XL710/XXV710,
the result is failed. The "uio_pci_generic" module is not supported
by X710/XL710/XXV710.

Signed-off-by: Guo, Jia 
Acked-by: Jingjing Wu 
---
v3: add related linux kernel commit info and fix some syntax issues
v2: add X710/XXV710 limitation case
---
 doc/guides/rel_notes/known_issues.rst | 26 ++
 1 file changed, 26 insertions(+)

diff --git a/doc/guides/rel_notes/known_issues.rst 
b/doc/guides/rel_notes/known_issues.rst
index 3cd4237..8b083a0 100644
--- a/doc/guides/rel_notes/known_issues.rst
+++ b/doc/guides/rel_notes/known_issues.rst
@@ -640,3 +640,29 @@ I40e VF may not receive packets in the promiscuous mode
 
 **Driver/Module**:
Poll Mode Driver (PMD).
+
+
+The uio_pci_generic module bind failed in X710/XL710/XXV710
+---
+
+**Description**:
+   The “uio_pci_generic” module is not supported by XL710, since the errata of 
XL710
+   states that the Interrupt Status bit is not implemented. The errata is the 
item #71
+   from the `xl710 controller spec
+   
`_.
+   The hw limitation is the same as other X710/XXV710 NICs.
+
+**Implication**:
+   When bind the “uio_pci_generic” module, it probe device and check the 
Interrupt Status
+   bit. Since it is not supported by X710/XL710/XXV710, it return a *failed* 
value. The statement
+   that these products don’t support INTx masking, is indicated in the related 
`linux kernel commit
+   < 
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/drivers/pci/quirks.c?id=8bcf4525c5d43306c5fd07e132bc8650e3491aec>`_.
+
+**Resolution/Workaround**:
+   Do not bind to use the “uio_pci_generic” module in X710/XL710/XXV710 NICs.
+
+**Affected Environment/Platform**:
+   All.
+
+**Driver/Module**:
+   Poll Mode Driver (PMD).
-- 
2.7.4

[dpdk-dev] [PATCH v2] net/sfc: do not panic if alarms are not supported

[Andrew Rybchenko]

Alarms are not supported on the FreeBSD.
Application must poll link status periodically itself using
rte_eth_link_get_nowait() to avoid management event queue overflow.

Fixes: 2de39f4e1310 ("net/sfc: periodic management EVQ polling using alarm")

Signed-off-by: Andrew Rybchenko 
Reviewed-by: Andrew Lee 
Reviewed-by: Andy Moreton 
---
v2:
 - fix spelling

 drivers/net/sfc/sfc_ev.c | 12 
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/drivers/net/sfc/sfc_ev.c b/drivers/net/sfc/sfc_ev.c
index c788986..fe6de6f 100644
--- a/drivers/net/sfc/sfc_ev.c
+++ b/drivers/net/sfc/sfc_ev.c
@@ -499,10 +499,14 @@
 
rc = rte_eal_alarm_set(SFC_MGMT_EV_QPOLL_PERIOD_US,
   sfc_ev_mgmt_periodic_qpoll, sa);
-   if (rc != 0)
-   sfc_panic(sa,
- "cannot rearm management EVQ polling alarm (rc=%d)",
- rc);
+   if (rc == -ENOTSUP) {
+   sfc_warn(sa, "alarms are not supported");
+   sfc_warn(sa, "management EVQ must be polled indirectly using 
no-wait link status update");
+   } else if (rc != 0) {
+   sfc_err(sa,
+   "cannot rearm management EVQ polling alarm (rc=%d)",
+   rc);
+   }
 }
 
 static void
-- 
1.8.2.3

Re: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Mcnamara, John]

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Jeff Guo
> Sent: Thursday, January 19, 2017 10:57 AM
> To: Zhang, Helin ; Wu, Jingjing
> 
> Cc: dev@dpdk.org; Guo, Jia ; g...@dpdk.org
> Subject: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for
> i40e
> 
> +
> +The uio_pci_generic module bind failed in X710/XL710/XXV710
> +---
> +
> +**Description**:
> +   The “uio_pci_generic” module is not supported by XL710, since the
> errata of XL710
> +   states that the Interrupt Status bit is not implemented. The errata is
> the item #71
> +   from the `xl710 controller spec
> +

>  -10-40-controller-spec-update.html>`_.
> +   The hw limitation is the same as other X710/XXV710 NICs.
> +
> +**Implication**:
> +   When bind the “uio_pci_generic” module, it probe device and check the
> Interrupt Status

Better as: When using ``bind uio_pci_generic``, it probes the device


> +   bit. Since it is not supported by X710/XL710/XXV710, it return a
> *failed* value. The statement
> +   that these products don’t support INTx masking, is indicated in the
> related `linux kernel commit
> +   < https://git.kernel.org/cgit/linux/kernel/git/stable/linux-
> stable.git/commit/drivers/pci/quirks.c?id=8bcf4525c5d43306c5fd07e132bc8650
> e3491aec>`_.
> +
> +**Resolution/Workaround**:
> +   Do not bind to use the “uio_pci_generic” module in X710/XL710/XXV710
> NICs.

Better as: Do not bind to ``uio_pci_generic`` in X710/XL710/XXV710 NICs.

Re: [dpdk-dev] [PATCH 0/2] Enable zero verdicts in ACLs

[Ananyev, Konstantin]

> -Original Message-
> From: Michał Mirosław [mailto:mirq-li...@rere.qmqm.pl]
> Sent: Wednesday, January 18, 2017 7:33 PM
> To: dev@dpdk.org
> Cc: Ananyev, Konstantin ; Thomas Monjalon 
> 
> Subject: [PATCH 0/2] Enable zero verdicts in ACLs
> 
> This set enables one to have ACL matches return 0 where the distinction
> from no-match case is not needed.
> 
> This is a resubmission of the patches as a series, rebased on net-next tree,
> no other changes vs v2.
> 
> v2: fixes to prog_guide and ACL tests
> 
> Michał Mirosław (2):
>   acl: remove invalid test
>   acl: allow zero verdict
> 
>  app/test/test_acl.c| 33 
> --
>  .../prog_guide/packet_classif_access_ctrl.rst  |  3 +-
>  lib/librte_acl/rte_acl.c   |  3 +-
>  lib/librte_acl/rte_acl.h   |  2 --
>  lib/librte_table/rte_table_acl.c   |  2 +-
>  5 files changed, 4 insertions(+), 39 deletions(-)
> 
> --

Acked-by: Konstantin Ananyev 

> 2.11.0

Re: [dpdk-dev] [PATCH] devtools: check stable tag in fixes

[Ferruh Yigit]

On 1/19/2017 8:05 AM, Yuanhan Liu wrote:
> On Wed, Jan 18, 2017 at 05:25:10PM +, Ferruh Yigit wrote:
>> On 1/18/2017 4:41 AM, Yuanhan Liu wrote:
>>> On Tue, Jan 17, 2017 at 07:42:33PM +0100, Thomas Monjalon wrote:
 2017-01-17 18:15, Ferruh Yigit:
> On 1/17/2017 2:54 PM, Thomas Monjalon wrote:
>> The tag "Cc: sta...@dpdk.org" must be set when the commit must be
>> backported to a stable branch.
>>
>> It must be located just below the "Fixes:" tag (without blank line)
>> and followed by a blank line, separated from SoB and review tags below.
>
> I am OK to keep it if it will help stable tree maintenance, but I still
> not clear about why we need this.
>
> If a patch is a fix, it should already have "Fixes:" line, so this can
> be used to parse git history.
>>>
>>> Same answer (as I have already replied to you in another email): not all fix
>>> patches should be picked to stable branch. (I gave some examples below)
>>
>> I was thinking all fixes will have "Cc: stable" tag, to be sure all
>> fixes sent to stable mail list, and you will be the picking for stable tree.
>>
>> Now you are suggesting some fixes may have "Fixes:" tag but not "Cc:
>> stable" tag.
>>
>> So this means now author/maintainer/committer will be the picking
>> patches for stable tree, and to show this decision, will add "Cc:
>> stable" to the commit log.
>>
>> If author puts the "Cc: stable" tag, git send-email will ensure this
>> patch will be sent to stable mail list too.
>> But if author missed the "Cc: stable" tag, will it be enough for you if
>> committer adds the tag into commit log? Or should patch sent to stable
>> mail list too?
> 
> Yes, that'd be enough. 

To highlight, and double check, the process will be updated as following:

- Author may or may not have "CC: sta...@dpdk.org" for fixes.

- Maintainer/Committer may add "CC: sta...@dpdk.org" tag to commit log,
but won't send a mail to the stable mail list.

- "CC: sta...@dpdk.org" tags will remain in commit logs.


> I will write a short script to monitor the
> official DPDK git tree regularly (say weekly), to see whether there
> are any new candidates (those with 'cc: stable' tag) for stable
> releases or not. If so, I will pick them to stable.
> 
>>
>> What is the relation between "Cc: stable" tag been in commit log and
>> patch been sent to stable mail list?
> 
> Not too much. The stable mailing list provides a place for further
> discussion on stable patches:
> 
> - A notification will be sent to the stable mailing list when a patch
>   is chosen as a candidate for a stable release. People can shout out
>   if he thinks some patches should not have been picked for a stable
>   release.
> 
> - people can also jump out to say, "hey, seems you forgot commit xxx,
>   which is a good have for stable release".
> 
>   --yliu
> 

Re: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Ferruh Yigit]

On 1/19/2017 10:56 AM, Jeff Guo wrote:
> From: "Guo, Jia" 
> 
> When bind with the "uio_pci_generic" module in X710/XL710/XXV710,
> the result is failed. The "uio_pci_generic" module is not supported
> by X710/XL710/XXV710.
> 
> Signed-off-by: Guo, Jia 

To be consistent on names, this should be "Jeff Guo "
both in sign-off and author field.

If you will send a new version of the patch, please take care of this
too, if there won't be a new version please confirm it is OK to fix
while applying.

Thanks,
ferruh

<...>

Re: [dpdk-dev] [PATCH v4 1/6] lib: distributor performance enhancements

[Hunt, David]

Thanks for the comments Bruce. Addressed below.


On 16/1/2017 4:36 PM, Bruce Richardson wrote:

On Mon, Jan 09, 2017 at 07:50:43AM +, David Hunt wrote:

Now sends bursts of up to 8 mbufs to each worker, and tracks
the in-flight flow-ids (atomic scheduling)

New file with a new api, similar to the old API except with _burst
at the end of the function names


Can you explain why this is necessary, and also how the new version
works compared to the old. I know this is explained in the cover letter,
but the cover letter does not make the git commit log.


Sure. I'll add extra comments into the git comment. The main reason is to
preserve the original API. This gives the user the choice to migrate to 
the new

API should they wish to.


Signed-off-by: David Hunt 
---


diff --git a/lib/librte_distributor/rte_distributor_burst.c 
b/lib/librte_distributor/rte_distributor_burst.c

new file mode 100644
index 000..ae7cf9d
--- /dev/null
+++ b/lib/librte_distributor/rte_distributor_burst.c
@@ -0,0 +1,558 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright(c) 2016 Intel Corporation. All rights reserved.

Update year since we aren't in 2016 any more.


+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above 
copyright
+ *   notice, this list of conditions and the following 
disclaimer in

+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of Intel Corporation nor the names of its
+ *   contributors may be used to endorse or promote products 
derived

+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND 
CONTRIBUTORS

+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND 
FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE 
COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, 
INCIDENTAL,

+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS 
OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND 
ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR 
TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF 
THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH 
DAMAGE.

+ */
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include "rte_distributor_priv.h"
+#include "rte_distributor_burst.h"
+
+TAILQ_HEAD(rte_dist_burst_list, rte_distributor_burst);
+
+static struct rte_tailq_elem rte_dist_burst_tailq = {
+.name = "RTE_DIST_BURST",
+};
+EAL_REGISTER_TAILQ(rte_dist_burst_tailq)
+
+/ APIs called by workers /
+
+/ Burst Packet APIs called by workers /
+
+/* This function should really be called return_pkt_burst() */

1) Why should it be?
2) Why isn't it called that?
Please explain the naming.


It seemed to me that the main use of this function was to return
the packets from the worker rather than requesting new packets,
whilst also toggling the bit to tell the distributor to send more packets.
So I guess it's OK as it is. I've removed the comment to remove this 
confusion.



+void
+rte_distributor_request_pkt_burst(struct rte_distributor_burst *d,
+unsigned int worker_id, struct rte_mbuf **oldpkt,
+unsigned int count)
+{
+struct rte_distributor_buffer_burst *buf = &(d->bufs[worker_id]);
+unsigned int i;
+
+volatile int64_t *retptr64;
+
+
+/* if we dont' have any packets to return, return. */
+if (count == 0)
+return;
+

So if we don't return anything we don't get any more packets, right?
What happens if we return fewer packets than we were previously given?
If that is allowed, why the restriction on returning at least one?


You are correct. We should be able to return 0, and still flip the 
handshake

bit to request more packets. This check will be removed.


+retptr64 = &(buf->retptr64[0]);



+
+int
+rte_distributor_get_pkt_burst(struct rte_distributor_burst *d,
+unsigned int worker_id, struct rte_mbuf **pkts,
+struct rte_mbuf **oldpkt, unsigned int return_count)
+{
+unsigned int count;
+uint64_t retries = 0;
+
+rte_distributor_request_pkt_burst(d, worker_id, oldpkt, 
return_count);

+
+count = rte_distributor_poll_pkt_burst(d, worker_id, pkts);
+while (count == 0) {
+rte_pause();
+retries++;
+if (retries > 1000)
+return 0;

This behaviour is differe

Re: [dpdk-dev] [RFC PATCH 00/11] generalise rte_ring to allow different datatypes

[Bruce Richardson]

On Tue, Jan 17, 2017 at 02:38:20PM +0100, Olivier Matz wrote:
> Hi Bruce,
> 
> Maybe it's worth checking the impact. The size check could be done only
> once per bulk, so it may not cost that much.
> 
> It's also possible to have a particular case for pointer size, and
> use a memcpy for other sizes.
> 
> 
 
> I think having a performance test showing storing the elt size in the
> ring structure has a deep impact would help to reach a consensus
> faster :)
> 
> 
Hi Olivier,

I did a quick prototype using a switch statement for three data element
sizes: 8, 16, and 32 bytes. The performance difference was neglible to
none. In most cases, with ring_perf_autotest on my system, there was a
small degradation, of less than 1 cycle per packet, and a few were
slightly faster, probably due to the natural variation in results
between runs. I did not test with any memcpy calls in the datapath, all
assignments were done using uint64_t's or vectors of the appropriate
sizes.

Therefore it looks like some kind of solution without macros and using a
stored element size is possible. However, I think there is a third
alternative too. It is outlined below as option 3.

1. Use macros as in original RFC

2. Update rte_ring like I did for tests described above so that
   create takes the size parameter, and the switch statment in
   enqueue and dequeue looks that up at runtime.
   This means that rte_ring becomes the type used for all
   transfers of all sizes. Also, enqueues/dequeue functions take
   void * or const void * obj_table parameters rather than void
   ** and void * const * obj_table. 
   Downside, this would change the ring API and ABI, and the
   ring maintains no type information

3. Update rte_ring as above but rename it to rte_common_ring,
   and have the element size parameter passed to enqueue and
   dequeue functions too - allowing the compiler to optimise the
   switch out. Then we update the existing rte_ring to use the
   rte_common_ring calls, passing in sizeof(void *) as parameter
   to each common call. An event-ring type, or any other ring
   types can similarly be written using common ring code, and
   present the appropriate type information on enqueue/dequeue
   to the apps using them.
   Downside: more code to maintain, and more specialised APIs.

Personally, because I like having type-specialised code, I prefer the
third option. It also gives us the ability to change the common code
without affecting the API/ABI of the rings [which could be updated later
after a proper deprecation period, if we want].

An example of a change I have in mind for this common code would be some
rework around the watermarks support. While the watermarks support is
useful, for the event_rings we actually need more information provided
from enqueue. To that end, I would see the common_rings code changed so
that the enqueue function returns an additional parameter of the
amount of space left in the ring. This information is computed by the
function anyway, and can therefore be efficiently returned by the calls.
For sp_enqueue, this extra parameter would allow the app to know the
minimum number of elements which can be successfully enqueued to the
ring in a subsequent call. The existing rte_ring code can use the return
value to calculate itself if the watermark is exceeded and return a
value as it does now. Other ring types can then decide themselves if
they want to provide watermark functionality, or what their API set
would be - though it's probably best to keep the APIs consistent.

Further thoughts?
/Bruce

Re: [dpdk-dev] [PATCH v4 2/6] lib: add distributor vector flow matching

[Hunt, David]

On 16/1/2017 4:40 PM, Bruce Richardson wrote:

On Mon, Jan 09, 2017 at 07:50:44AM +, David Hunt wrote:

--- a/lib/librte_distributor/rte_distributor_burst.c
+++ b/lib/librte_distributor/rte_distributor_burst.c
@@ -352,6 +352,9 @@ rte_distributor_process_burst(struct rte_distributor_burst 
*d,
}
  
  		switch (d->dist_match_fn) {

+   case RTE_DIST_MATCH_VECTOR:
+   find_match_vec(d, &flows[0], &matches[0]);
+   break;
default:
find_match_scalar(d, &flows[0], &matches[0]);
}

Will link not fail on non-x86 platforms due to find_match_vec not having
any implementation on those platforms?

/Bruce


I've added a fallback find_match_vec in rte_distributor_match_generic.c 
that calls find_match_scalar.


Rgds,
Dave.

Re: [dpdk-dev] [RFC PATCH 00/11] generalise rte_ring to allow different datatypes

[Ferruh Yigit]

On 1/19/2017 12:10 PM, Bruce Richardson wrote:
> On Tue, Jan 17, 2017 at 02:38:20PM +0100, Olivier Matz wrote:
>> Hi Bruce,
>>
>> Maybe it's worth checking the impact. The size check could be done only
>> once per bulk, so it may not cost that much.
>>
>> It's also possible to have a particular case for pointer size, and
>> use a memcpy for other sizes.
>>
>>
>  
>> I think having a performance test showing storing the elt size in the
>> ring structure has a deep impact would help to reach a consensus
>> faster :)
>>
>>
> Hi Olivier,
> 
> I did a quick prototype using a switch statement for three data element
> sizes: 8, 16, and 32 bytes. The performance difference was neglible to
> none. In most cases, with ring_perf_autotest on my system, there was a
> small degradation, of less than 1 cycle per packet, and a few were
> slightly faster, probably due to the natural variation in results
> between runs. I did not test with any memcpy calls in the datapath, all
> assignments were done using uint64_t's or vectors of the appropriate
> sizes.
> 
> Therefore it looks like some kind of solution without macros and using a
> stored element size is possible. However, I think there is a third
> alternative too. It is outlined below as option 3.
> 
>   1. Use macros as in original RFC
> 
>   2. Update rte_ring like I did for tests described above so that
>  create takes the size parameter, and the switch statment in
>  enqueue and dequeue looks that up at runtime.
>  This means that rte_ring becomes the type used for all
>  transfers of all sizes. Also, enqueues/dequeue functions take
>  void * or const void * obj_table parameters rather than void
>  ** and void * const * obj_table. 
>  Downside, this would change the ring API and ABI, and the
>  ring maintains no type information
> 
>   3. Update rte_ring as above but rename it to rte_common_ring,
>  and have the element size parameter passed to enqueue and
>  dequeue functions too - allowing the compiler to optimise the
>  switch out. Then we update the existing rte_ring to use the
>  rte_common_ring calls, passing in sizeof(void *) as parameter
>  to each common call. An event-ring type, or any other ring
>  types can similarly be written using common ring code, and
>  present the appropriate type information on enqueue/dequeue
>  to the apps using them.
>  Downside: more code to maintain, and more specialised APIs.
> 
> Personally, because I like having type-specialised code, I prefer the
> third option. It also gives us the ability to change the common code
> without affecting the API/ABI of the rings [which could be updated later
> after a proper deprecation period, if we want].

+1 for third option.

> 
> An example of a change I have in mind for this common code would be some
> rework around the watermarks support. While the watermarks support is
> useful, for the event_rings we actually need more information provided
> from enqueue. To that end, I would see the common_rings code changed so
> that the enqueue function returns an additional parameter of the
> amount of space left in the ring. This information is computed by the
> function anyway, and can therefore be efficiently returned by the calls.
> For sp_enqueue, this extra parameter would allow the app to know the
> minimum number of elements which can be successfully enqueued to the
> ring in a subsequent call. The existing rte_ring code can use the return
> value to calculate itself if the watermark is exceeded and return a
> value as it does now. Other ring types can then decide themselves if
> they want to provide watermark functionality, or what their API set
> would be - though it's probably best to keep the APIs consistent.
> 
> Further thoughts?
> /Bruce
> 

Re: [dpdk-dev] I/O device memory read/write implementation for sfc

[Ferruh Yigit]

On 1/19/2017 6:30 AM, Andrew Rybchenko wrote:
> On 01/19/2017 08:53 AM, Jerin Jacob wrote:
>> On Wed, Jan 18, 2017 at 07:31:39PM +, Ferruh Yigit wrote:
>>> Hi Andrew, Jerin,
>>>
>>> Existing "I/O device memory read/write" doesn't cover sfc driver because
>>> driver was in next-net tree.
>>>
>>> And both sfc and "I/O device memory read/write" patches will be
>>> integrated into main tree for rc1 release.
>>>
>>> If "I/O device memory read/write" implementation can be done for sfc
>>> before rc2, that changes can be merged into rc2 release.
>>>
>>> Who can work on this task?
>> sfc driver does not even support non x86 build for now. I can spend some 
>> time if
>> Andrew's does not have any bandwidth to work on this task.
> 
> Don't worry. I'll do.

Thank you Andrew.

> 
> Thanks,
> Andrew.

[dpdk-dev] [PATCHv5 00/33] NXP DPAA2 PMD

[Hemant Agrawal]

The patch series adds NXP’s QorIQ-Layerscape DPAA2 Architecture based
fsl-mc bus driver and network SoC PMD.  This version of the driver
supports NXP LS208xA, LS204xA and LS108x families Network SoCs.

DPAA2, or Data Path Acceleration Architecture, is a hardware architecture
designed for high-speed network packet processing. It uses a bus name
‘fsl-mc’, part of Linux Kernel Staging tree [3], for resource management.

A brief description of architecture is given below; detailed description
is part of the documentation in the patches itself.

DPAA2 contains hardware component called the Management Complex (or MC).
It manages the DPAA2 hardware resources.  The MC provides an object-based
abstraction for software drivers to use the DPAA2 hardware.

Some of the key objects are:
- DPNI, which refers to the network interface object.
- DPBP, which refers to HW based memory pool object
- DPIO, refers to processing context for accessing QBMAN

Besides the MC, DPAA2 also includes a Hardware based Queue and Buffer Manager
called QBMAN. Prime responsibility of QBMAN is to allow lockless access to
software/user-space to the queues and buffers implemented in the hardware.

The patch series could be logically structured into following sub-areas:
2. (Patch 0001) Enabling crc in armv8 core machine type
3. (Patch 0002) DPAA2 Architecture overview document
4. (Patch 0003) Common dpaa2 hw accelerator drivers for QBMAN.
5. (Patches 0004-0012) introduce fsl-mc bus
6. (Patches 0013-0017) introduce DPAA2 PMD, DPIO and mempool
7. (Patches 0018-0031) Support for DPAA2 Ethernet Device (ethdev)
7. (Patches 0032-0033) Additional functionality in DPAA2 ethdev.

The following design decisions are made during development:

1. DPAA2 implements a new bus called "fsl-mc" and some common accelerator 
drivers.
   These drivers will be shared with dpaa2 based crypto drivers.

2. DPAA2 implements the HW mempool offload with DPBP object.
 - The new pool is being configured using compile time option and pool name
   as "dpaa2".

3. It maintains per lcore DPIO objects and affine the DPIO instance to the
   processing threads accessing the QBMAN HW.

Prerequisites:
 - For running the PMD, NXP's SoC (board) and SDK (software/BSP) is required.
   Information about obtaining relevant software is available in the docs
   as part of the patch.
 - At present the series has limited support for Ethernet functions. But,
   more functionality would be made available in a phased manner.

Future Changes/Caveats:

1. VFIO code for fsl-mc bus is different than eal-vfio code for pci bus.
   This need to be re-worked to make possible re-use of the existing code.

2. shared lib support in case of parallel build fails due to dependency of
   drivers/common, pool, etc on other driver components. The dependency graph
   need to be improved to support it.

References:
[1] http://dpdk.org/dev/patchwork/patch/19557/
[2] https://www.kernel.org/doc/readme/drivers-staging-fsl-mc-README.txt
[3] http://dpdk.org/ml/archives/dev/2016-October/048949.html

---
v5:
* rebased over master (6818a7f4)

v4:
* rebased over master (1feda4d8) and patches from Shreyansh [1] for Bus Arch.

v3:
* rebased over master (eac901ce2) and patches from Shreyansh [1] for Bus Arch.
* Fixed comment from John on Patch-0003 for documentation
* Removed Patch-0001 for rte_device in rte_eth_dev; Already upstreamed through
  another series

v2:
* separated the "fsl-mc" bus from the dpaa2 pmd driver - introduced drivers/bus
* separated the "dpaa2" hw mempool from dpaa2 pmd driver - introduced 
drivers/pool
* removed documentation warnings and missing information.
* removed arm64 part specific code from driver
* changed rte_panic to errors
* reduced checkpatch warnings


Hemant Agrawal (33):
  mk/dpaa2: add the crc support to the machine type
  doc: add dpaa2 nic details
  drivers/common/dpaa2: adding qbman driver
  bus/fslmc: introducing fsl-mc bus driver
  bus/fslmc: introduce mc object functions
  bus/fslmc: add mc dpni object support
  bus/fslmc: add mc dpio object support
  bus/fslmc: add mc dpbp object support
  bus/fslmc: add mc dpseci object support
  eal/vfio: adding vfio utility functions in map file
  bus/fslmc: add vfio support
  bus/fslmc: scan for net and sec devices
  net/dpaa2: introducing NXP dpaa2 pmd driver
  bus/fslmc: add debug log message support
  drivers/common/dpaa2: dpio portal driver
  drivers/pool/dpaa2: adding hw offloaded mempool
  drivers/common/dpaa2: dpio routine to affine to crypto threads
  net/dpaa2: adding eth ops to dpaa2
  net/dpaa2: add rss flow distribution
  net/dpaa2: configure mac address at init
  net/dpaa2: attach the buffer pool to dpni
  net/dpaa2: add support for l3 and l4 checksum offload
  net/dpaa2: add support for promiscuous mode
  net/dpaa2: add mtu config support
  net/dpaa2: add packet rx and tx support
  net/dpaa2: rx packet parsing and packet type support
  net/dpaa2: link status update
  net/dpaa2: basic stats support
  net/dpaa2: enable st

[dpdk-dev] [PATCH] mbuf: use pktmbuf helper to create the pool

[Hemant Agrawal]

When possible, replace the uses of rte_mempool_create() with
the helper provided in librte_mbuf: rte_pktmbuf_pool_create().

This is the preferred way to create a mbuf pool.

This also updates the documentation.

Signed-off-by: Olivier Matz 
Signed-off-by: Hemant Agrawal 
---
This patch is derived from the RFC from Olivier:
http://dpdk.org/dev/patchwork/patch/15925/

 app/test/test_link_bonding_rssconf.c   | 11 
 doc/guides/sample_app_ug/ip_reassembly.rst | 13 +
 doc/guides/sample_app_ug/ipv4_multicast.rst| 12 
 doc/guides/sample_app_ug/l2_forward_job_stats.rst  | 33 --
 .../sample_app_ug/l2_forward_real_virtual.rst  | 26 +++--
 doc/guides/sample_app_ug/ptpclient.rst | 11 ++--
 doc/guides/sample_app_ug/quota_watermark.rst   | 26 ++---
 drivers/net/bonding/rte_eth_bond_8023ad.c  | 13 -
 examples/ip_pipeline/init.c| 18 ++--
 examples/ip_reassembly/main.c  | 16 +--
 examples/multi_process/l2fwd_fork/main.c   | 13 +++--
 examples/tep_termination/main.c| 16 +--
 lib/librte_mbuf/rte_mbuf.c |  7 +++--
 lib/librte_mbuf/rte_mbuf.h | 29 +++
 14 files changed, 106 insertions(+), 138 deletions(-)

diff --git a/app/test/test_link_bonding_rssconf.c 
b/app/test/test_link_bonding_rssconf.c
index 34f1c16..9034f62 100644
--- a/app/test/test_link_bonding_rssconf.c
+++ b/app/test/test_link_bonding_rssconf.c
@@ -67,7 +67,7 @@
 #define SLAVE_RXTX_QUEUE_FMT  ("rssconf_slave%d_q%d")
 
 #define NUM_MBUFS 8191
-#define MBUF_SIZE (1600 + sizeof(struct rte_mbuf) + RTE_PKTMBUF_HEADROOM)
+#define MBUF_SIZE (1600 + RTE_PKTMBUF_HEADROOM)
 #define MBUF_CACHE_SIZE 250
 #define BURST_SIZE 32
 
@@ -536,13 +536,12 @@ struct link_bonding_rssconf_unittest_params {
 
if (test_params.mbuf_pool == NULL) {
 
-   test_params.mbuf_pool = rte_mempool_create("RSS_MBUF_POOL", 
NUM_MBUFS *
-   SLAVE_COUNT, MBUF_SIZE, MBUF_CACHE_SIZE,
-   sizeof(struct rte_pktmbuf_pool_private), 
rte_pktmbuf_pool_init,
-   NULL, rte_pktmbuf_init, NULL, rte_socket_id(), 
0);
+   test_params.mbuf_pool = rte_pktmbuf_pool_create(
+   "RSS_MBUF_POOL", NUM_MBUFS * SLAVE_COUNT,
+   MBUF_CACHE_SIZE, 0, MBUF_SIZE, rte_socket_id());
 
TEST_ASSERT(test_params.mbuf_pool != NULL,
-   "rte_mempool_create failed\n");
+   "rte_pktmbuf_pool_create failed\n");
}
 
/* Create / initialize ring eth devs. */
diff --git a/doc/guides/sample_app_ug/ip_reassembly.rst 
b/doc/guides/sample_app_ug/ip_reassembly.rst
index 3c5cc70..d5097c6 100644
--- a/doc/guides/sample_app_ug/ip_reassembly.rst
+++ b/doc/guides/sample_app_ug/ip_reassembly.rst
@@ -223,11 +223,14 @@ each RX queue uses its own mempool.
 
 snprintf(buf, sizeof(buf), "mbuf_pool_%u_%u", lcore, queue);
 
-if ((rxq->pool = rte_mempool_create(buf, nb_mbuf, MBUF_SIZE, 0, 
sizeof(struct rte_pktmbuf_pool_private), rte_pktmbuf_pool_init, NULL,
-rte_pktmbuf_init, NULL, socket, MEMPOOL_F_SP_PUT | MEMPOOL_F_SC_GET)) 
== NULL) {
-
-RTE_LOG(ERR, IP_RSMBL, "mempool_create(%s) failed", buf);
-return -1;
+rxq->pool = rte_pktmbuf_pool_create(buf, nb_mbuf,
+   0, /* cache size */
+   0, /* priv size */
+   MBUF_DATA_SIZE, socket);
+if (rxq->pool == NULL) {
+   RTE_LOG(ERR, IP_RSMBL,
+   "rte_pktmbuf_pool_create(%s) failed", buf);
+   return -1;
 }
 
 Packet Reassembly and Forwarding
diff --git a/doc/guides/sample_app_ug/ipv4_multicast.rst 
b/doc/guides/sample_app_ug/ipv4_multicast.rst
index 72da8c4..d9ff249 100644
--- a/doc/guides/sample_app_ug/ipv4_multicast.rst
+++ b/doc/guides/sample_app_ug/ipv4_multicast.rst
@@ -145,12 +145,12 @@ Memory pools for indirect buffers are initialized 
differently from the memory po
 
 .. code-block:: c
 
-packet_pool = rte_mempool_create("packet_pool", NB_PKT_MBUF, 
PKT_MBUF_SIZE, 32, sizeof(struct rte_pktmbuf_pool_private),
- rte_pktmbuf_pool_init, NULL, 
rte_pktmbuf_init, NULL, rte_socket_id(), 0);
-
-header_pool = rte_mempool_create("header_pool", NB_HDR_MBUF, 
HDR_MBUF_SIZE, 32, 0, NULL, NULL, rte_pktmbuf_init, NULL, rte_socket_id(), 0);
-clone_pool = rte_mempool_create("clone_pool", NB_CLONE_MBUF,
-CLONE_MBUF_SIZE, 32, 0, NULL, NULL, rte_pktmbuf_init, NULL, 
rte_socket_id(), 0);
+packet_pool = rte_pktmbuf_pool_create("packet_pool", NB_PKT_MBUF, 32,
+   0, PKT_MBUF_DATA_SIZE, rte_socket_id());
+header_pool = rte_pktmbuf_pool_create("header_pool", NB_HDR_MBUF, 32,
+   0, HDR_MBUF_DATA_SIZE, rte_socket_id());
+clone_pool = rte_pktmbu

[dpdk-dev] [PATCHv5 02/33] doc: add dpaa2 nic details

[Hemant Agrawal]

This patch adds the NXP dpaa2 architecture and pmd details
in the Network interfaces section.

Signed-off-by: Hemant Agrawal 
---
 MAINTAINERS|   8 +
 doc/guides/nics/dpaa2.rst  | 594 +
 doc/guides/nics/features/dpaa2.ini |   8 +
 doc/guides/nics/index.rst  |   1 +
 doc/guides/rel_notes/release_17_02.rst |  11 +
 5 files changed, 622 insertions(+)
 create mode 100644 doc/guides/nics/dpaa2.rst
 create mode 100644 doc/guides/nics/features/dpaa2.ini

diff --git a/MAINTAINERS b/MAINTAINERS
index f071138..d79e1a5 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -354,6 +354,14 @@ M: Alejandro Lucero 
 F: drivers/net/nfp/
 F: doc/guides/nics/nfp.rst
 
+NXP DPAA2 PMD
+M: Hemant Agrawal 
+F: drivers/bus/fslmc/
+F: drivers/common/dpaa2/
+F: drivers/net/dpaa2/
+F: drivers/pool/dpaa2/
+F: doc/guides/nics/dpaa2.rst
+
 QLogic bnx2x
 M: Harish Patil 
 M: Rasesh Mody 
diff --git a/doc/guides/nics/dpaa2.rst b/doc/guides/nics/dpaa2.rst
new file mode 100644
index 000..0b30a6f
--- /dev/null
+++ b/doc/guides/nics/dpaa2.rst
@@ -0,0 +1,594 @@
+..  BSD LICENSE
+Copyright (C) NXP. 2016.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without
+modification, are permitted provided that the following conditions
+are met:
+
+* Redistributions of source code must retain the above copyright
+notice, this list of conditions and the following disclaimer.
+* Redistributions in binary form must reproduce the above copyright
+notice, this list of conditions and the following disclaimer in
+the documentation and/or other materials provided with the
+distribution.
+* Neither the name of NXP nor the names of its
+contributors may be used to endorse or promote products derived
+from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+DPAA2 Poll Mode Driver
+==
+
+The DPAA2 NIC PMD (**librte_pmd_dpaa2**) provides poll mode driver
+support for the inbuilt NIC found in the **NXP DPAA2** SoC family.
+
+More information can be found at `NXP Official Website
+`_.
+
+NXP DPAA2 (Data Path Acceleration Architecture Gen2)
+
+
+This section provides an overview of the NXP DPAA2 architecture
+and how it is integrated into the DPDK.
+
+Contents summary
+
+- DPAA2 overview
+- Overview of DPAA2 objects
+- DPAA2 driver architecture overview
+
+DPAA2 Overview
+~~
+
+Reference: `FSL MC BUS in Linux Kernel 
`_.
+
+DPAA2 is a hardware architecture designed for high-speed network
+packet processing.  DPAA2 consists of sophisticated mechanisms for
+processing Ethernet packets, queue management, buffer management,
+autonomous L2 switching, virtual Ethernet bridging, and accelerator
+(e.g. crypto) sharing.
+
+A DPAA2 hardware component called the Management Complex (or MC) manages the
+DPAA2 hardware resources.  The MC provides an object-based abstraction for
+software drivers to use the DPAA2 hardware.
+
+The MC uses DPAA2 hardware resources such as queues, buffer pools, and
+network ports to create functional objects/devices such as network
+interfaces, an L2 switch, or accelerator instances.
+
+The MC provides memory-mapped I/O command interfaces (MC portals)
+which DPAA2 software drivers use to operate on DPAA2 objects:
+
+The diagram below shows an overview of the DPAA2 resource management
+architecture:
+
+.. code-block:: console
+
+  +--+
+  |  OS  |
+  |DPAA2 drivers |
+  | ||
+  +-|+
+|
+| (create,discover,connect
+|  config,use,destroy)
+|
+  DPAA2 |
+  +| mc portal |-+
+  |   

[dpdk-dev] [PATCH] cryptodev: decouple from PCI device

[Hemant Agrawal]

This makes struct rte_cryptodev independent of struct rte_pci_device by
replacing it with a pointer to the generic struct rte_device.

This is inline with the recent changes in ethdev

Signed-off-by: Hemant Agrawal 
---
 drivers/crypto/qat/qat_qp.c| 12 +---
 drivers/crypto/qat/rte_qat_cryptodev.c |  6 +++---
 lib/librte_cryptodev/rte_cryptodev.c   |  6 +++---
 lib/librte_cryptodev/rte_cryptodev.h   |  4 ++--
 4 files changed, 17 insertions(+), 11 deletions(-)

diff --git a/drivers/crypto/qat/qat_qp.c b/drivers/crypto/qat/qat_qp.c
index 2e7188b..07e63fc 100644
--- a/drivers/crypto/qat/qat_qp.c
+++ b/drivers/crypto/qat/qat_qp.c
@@ -135,6 +135,7 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev *dev, 
uint16_t queue_pair_id,
int socket_id)
 {
struct qat_qp *qp;
+   struct rte_pci_device *pci_dev;
int ret;
 
PMD_INIT_FUNC_TRACE();
@@ -153,7 +154,9 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev *dev, 
uint16_t queue_pair_id,
return -EINVAL;
}
 
-   if (dev->pci_dev->mem_resource[0].addr == NULL) {
+   pci_dev = RTE_DEV_TO_PCI(dev->device);
+
+   if (pci_dev->mem_resource[0].addr == NULL) {
PMD_DRV_LOG(ERR, "Could not find VF config space "
"(UIO driver attached?).");
return -EINVAL;
@@ -174,7 +177,7 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev *dev, 
uint16_t queue_pair_id,
PMD_DRV_LOG(ERR, "Failed to alloc mem for qp struct");
return -ENOMEM;
}
-   qp->mmap_bar_addr = dev->pci_dev->mem_resource[0].addr;
+   qp->mmap_bar_addr = pci_dev->mem_resource[0].addr;
rte_atomic16_init(&qp->inflights16);
 
if (qat_tx_queue_create(dev, &(qp->tx_q),
@@ -289,6 +292,7 @@ static void qat_queue_delete(struct qat_queue *queue)
void *io_addr;
const struct rte_memzone *qp_mz;
uint32_t queue_size_bytes = nb_desc*desc_size;
+   struct rte_pci_device *pci_dev;
 
PMD_INIT_FUNC_TRACE();
if (desc_size > ADF_MSG_SIZE_TO_BYTES(ADF_MAX_MSG_SIZE)) {
@@ -349,7 +353,9 @@ static void qat_queue_delete(struct qat_queue *queue)
 
queue_base = BUILD_RING_BASE_ADDR(queue->base_phys_addr,
queue->queue_size);
-   io_addr = dev->pci_dev->mem_resource[0].addr;
+   pci_dev = RTE_DEV_TO_PCI(dev->device);
+
+   io_addr = pci_dev->mem_resource[0].addr;
 
WRITE_CSR_RING_BASE(io_addr, queue->hw_bundle_number,
queue->hw_queue_number, queue_base);
diff --git a/drivers/crypto/qat/rte_qat_cryptodev.c 
b/drivers/crypto/qat/rte_qat_cryptodev.c
index 1e7ee61..840d828 100644
--- a/drivers/crypto/qat/rte_qat_cryptodev.c
+++ b/drivers/crypto/qat/rte_qat_cryptodev.c
@@ -88,9 +88,9 @@
 
PMD_INIT_FUNC_TRACE();
PMD_DRV_LOG(DEBUG, "Found crypto device at %02x:%02x.%x",
-   cryptodev->pci_dev->addr.bus,
-   cryptodev->pci_dev->addr.devid,
-   cryptodev->pci_dev->addr.function);
+   RTE_DEV_TO_PCI(cryptodev->device)->addr.bus,
+   RTE_DEV_TO_PCI(cryptodev->device)->addr.devid,
+   RTE_DEV_TO_PCI(cryptodev->device)->addr.function);
 
cryptodev->dev_type = RTE_CRYPTODEV_QAT_SYM_PMD;
cryptodev->dev_ops = &crypto_qat_ops;
diff --git a/lib/librte_cryptodev/rte_cryptodev.c 
b/lib/librte_cryptodev/rte_cryptodev.c
index 127e8d0..8402b6a 100644
--- a/lib/librte_cryptodev/rte_cryptodev.c
+++ b/lib/librte_cryptodev/rte_cryptodev.c
@@ -446,7 +446,7 @@ struct rte_cryptodev *
"device data");
}
 
-   cryptodev->pci_dev = pci_dev;
+   cryptodev->device = &pci_dev->device;
cryptodev->driver = cryptodrv;
 
/* init user callbacks */
@@ -506,7 +506,7 @@ struct rte_cryptodev *
if (rte_eal_process_type() == RTE_PROC_PRIMARY)
rte_free(cryptodev->data->dev_private);
 
-   cryptodev->pci_dev = NULL;
+   cryptodev->device = NULL;
cryptodev->driver = NULL;
cryptodev->data = NULL;
 
@@ -867,7 +867,7 @@ struct rte_cryptodev *
RTE_FUNC_PTR_OR_RET(*dev->dev_ops->dev_infos_get);
(*dev->dev_ops->dev_infos_get)(dev, dev_info);
 
-   dev_info->pci_dev = dev->pci_dev;
+   dev_info->pci_dev = RTE_DEV_TO_PCI(dev->device);
if (dev->driver)
dev_info->driver_name = dev->driver->pci_drv.driver.name;
 }
diff --git a/lib/librte_cryptodev/rte_cryptodev.h 
b/lib/librte_cryptodev/rte_cryptodev.h
index 8f63e8f..ef072d2 100644
--- a/lib/librte_cryptodev/rte_cryptodev.h
+++ b/lib/librte_cryptodev/rte_cryptodev.h
@@ -621,8 +621,8 @@ struct rte_cryptodev {
/**< Functions exported by PMD */
uint64_t feature_flags;
/**< Supported features */
-   struct rte_pci_device *pci_dev;
-   /**< PCI info. supplied by probing */
+   struct rte_device *device;
+   /**< 

[dpdk-dev] [PATCHv5 01/33] mk/dpaa2: add the crc support to the machine type

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
Acked-by: Jerin Jacob 
---
 mk/machine/dpaa2/rte.vars.mk | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/mk/machine/dpaa2/rte.vars.mk b/mk/machine/dpaa2/rte.vars.mk
index 8541633..e4735c2 100644
--- a/mk/machine/dpaa2/rte.vars.mk
+++ b/mk/machine/dpaa2/rte.vars.mk
@@ -1,6 +1,7 @@
 #   BSD LICENSE
 #
-#   Copyright(c) 2016 Freescale Semiconductor, Inc. All rights reserved.
+#   Copyright (c) 2016 Freescale Semiconductor, Inc. All rights reserved.
+#   Copyright (c) 2016 NXP. All rights reserved.
 #
 #   Redistribution and use in source and binary forms, with or without
 #   modification, are permitted provided that the following conditions
@@ -53,7 +54,7 @@
 # CPU_CFLAGS =
 # CPU_LDFLAGS =
 # CPU_ASFLAGS =
-MACHINE_CFLAGS += -march=armv8-a
+MACHINE_CFLAGS += -march=armv8-a+crc
 
 ifdef CONFIG_RTE_ARCH_ARM_TUNE
 MACHINE_CFLAGS += -mcpu=$(CONFIG_RTE_ARCH_ARM_TUNE)
-- 
1.9.1

[dpdk-dev] [PATCHv5 06/33] bus/fslmc: add mc dpni object support

[Hemant Agrawal]

This patch add support for dpni object support in MC
driver.

DPNI represent a network interface object in DPAA2.

Signed-off-by: Alex Marginean 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/Makefile |1 +
 drivers/bus/fslmc/mc/dpni.c|  732 ++
 drivers/bus/fslmc/mc/fsl_dpkg.h|  177 
 drivers/bus/fslmc/mc/fsl_dpni.h| 1210 
 drivers/bus/fslmc/mc/fsl_dpni_cmd.h|  327 +++
 drivers/bus/fslmc/mc/fsl_net.h |  480 ++
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |   22 +
 7 files changed, 2949 insertions(+)
 create mode 100644 drivers/bus/fslmc/mc/dpni.c
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpkg.h
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpni.h
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpni_cmd.h
 create mode 100644 drivers/bus/fslmc/mc/fsl_net.h

diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index 21e9f17..c3616e6 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -49,6 +49,7 @@ EXPORT_MAP := rte_pmd_fslmcbus_version.map
 LIBABIVER := 1
 
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
+mc/dpni.c \
 mc/mc_sys.c
 
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_bus.c
diff --git a/drivers/bus/fslmc/mc/dpni.c b/drivers/bus/fslmc/mc/dpni.c
new file mode 100644
index 000..1e1415f
--- /dev/null
+++ b/drivers/bus/fslmc/mc/dpni.c
@@ -0,0 +1,732 @@
+/* Copyright 2013-2016 Freescale Semiconductor Inc.
+ * Copyright (c) 2016 NXP.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * * Neither the name of the above-listed copyright holders nor the
+ * names of any contributors may be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ *
+ * ALTERNATIVELY, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") as published by the Free Software
+ * Foundation, either version 2 of that License or (at your option) any
+ * later version.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+#include 
+#include 
+#include 
+#include 
+
+int dpni_prepare_key_cfg(const struct dpkg_profile_cfg *cfg,
+uint8_t *key_cfg_buf)
+{
+   int i, j;
+   int offset = 0;
+   int param = 1;
+   uint64_t *params = (uint64_t *)key_cfg_buf;
+
+   if (!key_cfg_buf || !cfg)
+   return -EINVAL;
+
+   params[0] |= mc_enc(0, 8, cfg->num_extracts);
+   params[0] = cpu_to_le64(params[0]);
+
+   if (cfg->num_extracts >= DPKG_MAX_NUM_OF_EXTRACTS)
+   return -EINVAL;
+
+   for (i = 0; i < cfg->num_extracts; i++) {
+   switch (cfg->extracts[i].type) {
+   case DPKG_EXTRACT_FROM_HDR:
+   params[param] |= mc_enc(0, 8,
+   cfg->extracts[i].extract.from_hdr.prot);
+   params[param] |= mc_enc(8, 4,
+   cfg->extracts[i].extract.from_hdr.type);
+   params[param] |= mc_enc(16, 8,
+   cfg->extracts[i].extract.from_hdr.size);
+   params[param] |= mc_enc(24, 8,
+   cfg->extracts[i].extract.
+   from_hdr.offset);
+   params[param] |= mc_enc(32, 32,
+   cfg->extracts[i].extract.
+   from_hdr.field);
+   params[param] = cpu_to_le64(params[param]);
+   param++;
+   params[param] |= mc_enc(0, 8,
+   cfg

[dpdk-dev] [PATCHv5 05/33] bus/fslmc: introduce mc object functions

[Hemant Agrawal]

This patch intoduces the DPAA2 MC(Management complex Driver).

This is a minimal set of low level functions to send and
receive commands to the fsl-mc. It includes support for basic
management commands and commands to manipulate MC objects.

This is common to be used by various DPAA2 PMDs. e.g.net, crypto
and other drivers.

This is a low level library also used in kernel.

Signed-off-by: Cristian Sovaiala 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/Makefile|   7 ++
 drivers/bus/fslmc/mc/fsl_mc_cmd.h | 231 ++
 drivers/bus/fslmc/mc/fsl_mc_sys.h |  98 
 drivers/bus/fslmc/mc/mc_sys.c | 107 ++
 4 files changed, 443 insertions(+)
 create mode 100644 drivers/bus/fslmc/mc/fsl_mc_cmd.h
 create mode 100644 drivers/bus/fslmc/mc/fsl_mc_sys.h
 create mode 100644 drivers/bus/fslmc/mc/mc_sys.c

diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index c4f22ed..21e9f17 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -37,6 +37,10 @@ LIB = librte_pmd_fslmcbus.a
 
 CFLAGS += -O3
 CFLAGS += $(WERROR_FLAGS)
+CFLAGS += "-Wno-strict-aliasing"
+
+CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc
+CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/mc
 
 # versioning export map
 EXPORT_MAP := rte_pmd_fslmcbus_version.map
@@ -44,6 +48,9 @@ EXPORT_MAP := rte_pmd_fslmcbus_version.map
 # library version
 LIBABIVER := 1
 
+SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
+mc/mc_sys.c
+
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_bus.c
 
 # library dependencies
diff --git a/drivers/bus/fslmc/mc/fsl_mc_cmd.h 
b/drivers/bus/fslmc/mc/fsl_mc_cmd.h
new file mode 100644
index 000..cbd3995
--- /dev/null
+++ b/drivers/bus/fslmc/mc/fsl_mc_cmd.h
@@ -0,0 +1,231 @@
+/* Copyright 2013-2016 Freescale Semiconductor Inc.
+ * Copyright (c) 2016 NXP.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * * Neither the name of the above-listed copyright holders nor the
+ * names of any contributors may be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ *
+ * ALTERNATIVELY, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") as published by the Free Software
+ * Foundation, either version 2 of that License or (at your option) any
+ * later version.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+#ifndef __FSL_MC_CMD_H
+#define __FSL_MC_CMD_H
+
+#define MC_CMD_NUM_OF_PARAMS   7
+
+#define MAKE_UMASK64(_width) \
+   ((uint64_t)((_width) < 64 ? ((uint64_t)1 << (_width)) - 1 : \
+  (uint64_t)-1))
+
+static inline uint64_t mc_enc(int lsoffset, int width, uint64_t val)
+{
+   return (uint64_t)(((uint64_t)val & MAKE_UMASK64(width)) << lsoffset);
+}
+
+static inline uint64_t mc_dec(uint64_t val, int lsoffset, int width)
+{
+   return (uint64_t)((val >> lsoffset) & MAKE_UMASK64(width));
+}
+
+struct mc_command {
+   uint64_t header;
+   uint64_t params[MC_CMD_NUM_OF_PARAMS];
+};
+
+/**
+ * enum mc_cmd_status - indicates MC status at command response
+ * @MC_CMD_STATUS_OK: Completed successfully
+ * @MC_CMD_STATUS_READY: Ready to be processed
+ * @MC_CMD_STATUS_AUTH_ERR: Authentication error
+ * @MC_CMD_STATUS_NO_PRIVILEGE: No privilege
+ * @MC_CMD_STATUS_DMA_ERR: DMA or I/O error
+ * @MC_CMD_STATUS_CONFIG_ERR: Configuration error
+ * @MC_CMD_STATUS_TIMEOUT: Operation timed out
+ * @MC_CMD_STATUS_NO_RESOURCE: No resources
+ * @MC_CMD_STATUS_NO_MEMORY: No memory available
+ * @MC_CMD_STATUS_BUSY: Device is busy
+ * @MC_CMD_STATUS_UNSUPPORTED_OP: Unsupported operation
+ * @MC_CMD_STATUS_INVALID_STATE: Invalid state
+ */
+enum mc_cmd_status {
+   MC_CMD_STATUS_OK = 0x0,
+   MC_CMD_STATUS_READY = 0

[dpdk-dev] [PATCHv5 07/33] bus/fslmc: add mc dpio object support

[Hemant Agrawal]

This patch adds the DPIO object support in MC driver.

DPIO - Data Path Input Output represent the processing
context to access the QBMAN HW for packet I/O.

Signed-off-by: Alex Marginean 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/Makefile |   1 +
 drivers/bus/fslmc/mc/dpio.c| 272 
 drivers/bus/fslmc/mc/fsl_dpio.h| 275 +
 drivers/bus/fslmc/mc/fsl_dpio_cmd.h| 114 ++
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |   8 +
 5 files changed, 670 insertions(+)
 create mode 100644 drivers/bus/fslmc/mc/dpio.c
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpio.h
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpio_cmd.h

diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index c3616e6..fa1ce13 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -50,6 +50,7 @@ LIBABIVER := 1
 
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
 mc/dpni.c \
+mc/dpio.c \
 mc/mc_sys.c
 
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_bus.c
diff --git a/drivers/bus/fslmc/mc/dpio.c b/drivers/bus/fslmc/mc/dpio.c
new file mode 100644
index 000..35a06d6
--- /dev/null
+++ b/drivers/bus/fslmc/mc/dpio.c
@@ -0,0 +1,272 @@
+/* Copyright 2013-2016 Freescale Semiconductor Inc.
+ * Copyright (c) 2016 NXP.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * * Neither the name of the above-listed copyright holders nor the
+ * names of any contributors may be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ *
+ * ALTERNATIVELY, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") as published by the Free Software
+ * Foundation, either version 2 of that License or (at your option) any
+ * later version.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+#include 
+#include 
+#include 
+#include 
+
+int dpio_open(struct fsl_mc_io *mc_io,
+ uint32_t cmd_flags,
+ int dpio_id,
+ uint16_t *token)
+{
+   struct mc_command cmd = { 0 };
+   int err;
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPIO_CMDID_OPEN,
+ cmd_flags,
+ 0);
+   DPIO_CMD_OPEN(cmd, dpio_id);
+
+   /* send command to mc*/
+   err = mc_send_command(mc_io, &cmd);
+   if (err)
+   return err;
+
+   /* retrieve response parameters */
+   *token = MC_CMD_HDR_READ_TOKEN(cmd.header);
+
+   return 0;
+}
+
+int dpio_close(struct fsl_mc_io *mc_io,
+  uint32_t cmd_flags,
+  uint16_t token)
+{
+   struct mc_command cmd = { 0 };
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPIO_CMDID_CLOSE,
+ cmd_flags,
+ token);
+
+   /* send command to mc*/
+   return mc_send_command(mc_io, &cmd);
+}
+
+int dpio_create(struct fsl_mc_io   *mc_io,
+   uint16_tdprc_token,
+   uint32_tcmd_flags,
+   const struct dpio_cfg   *cfg,
+   uint32_t*obj_id)
+{
+   struct mc_command cmd = { 0 };
+   int err;
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPIO_CMDID_CREATE,
+ cmd_flags,
+ dprc_token);
+   DPIO_CMD_CREATE(cmd, cfg);
+
+   /* send command to mc*/
+   err = mc_send_command(mc_io, &cmd);
+   if (err)
+   return err;
+
+   /* retrieve response parameters */
+   CMD_CREATE_RSP

[dpdk-dev] [PATCHv5 04/33] bus/fslmc: introducing fsl-mc bus driver

[Hemant Agrawal]

The fslmc bus driver is a rte_bus driver which scans the fsl-mc bus
for NXP DPAA2 SoCs.

Signed-off-by: Hemant Agrawal 
---
 config/common_base |   6 +
 config/defconfig_arm64-dpaa2-linuxapp-gcc  |   5 +
 drivers/Makefile   |   1 +
 drivers/bus/Makefile   |  36 ++
 drivers/bus/fslmc/Makefile |  52 +
 drivers/bus/fslmc/fslmc_bus.c  | 125 +
 drivers/bus/fslmc/rte_fslmc.h  | 148 +
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |   7 ++
 8 files changed, 380 insertions(+)
 create mode 100644 drivers/bus/Makefile
 create mode 100644 drivers/bus/fslmc/Makefile
 create mode 100644 drivers/bus/fslmc/fslmc_bus.c
 create mode 100644 drivers/bus/fslmc/rte_fslmc.h
 create mode 100644 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map

diff --git a/config/common_base b/config/common_base
index a64fd8b..e6b4d60 100644
--- a/config/common_base
+++ b/config/common_base
@@ -291,6 +291,12 @@ CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_MBOX=n
 # Compile Support Libraries for NXP DPAA2
 #
 CONFIG_RTE_LIBRTE_DPAA2_COMMON=n
+
+#
+# Compile NXP DPAA2 FSL-MC Bus
+#
+CONFIG_RTE_LIBRTE_FSLMC_BUS=n
+
 # Compile burst-oriented VIRTIO PMD driver
 #
 CONFIG_RTE_LIBRTE_VIRTIO_PMD=y
diff --git a/config/defconfig_arm64-dpaa2-linuxapp-gcc 
b/config/defconfig_arm64-dpaa2-linuxapp-gcc
index c57c340..800e22b 100644
--- a/config/defconfig_arm64-dpaa2-linuxapp-gcc
+++ b/config/defconfig_arm64-dpaa2-linuxapp-gcc
@@ -46,3 +46,8 @@ CONFIG_RTE_MAX_NUMA_NODES=1
 # Compile Support Libraries for DPAA2
 #
 CONFIG_RTE_LIBRTE_DPAA2_COMMON=y
+
+#
+# Compile NXP DPAA2 FSL-MC Bus
+#
+CONFIG_RTE_LIBRTE_FSLMC_BUS=y
diff --git a/drivers/Makefile b/drivers/Makefile
index d5580f6..bdae63b 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -32,6 +32,7 @@
 include $(RTE_SDK)/mk/rte.vars.mk
 
 DIRS-y += common
+DIRS-y += bus
 DIRS-y += net
 DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += crypto
 
diff --git a/drivers/bus/Makefile b/drivers/bus/Makefile
new file mode 100644
index 000..60e9764
--- /dev/null
+++ b/drivers/bus/Makefile
@@ -0,0 +1,36 @@
+#   BSD LICENSE
+#
+#   Copyright(c) 2016 NXP. All rights reserved.
+#   All rights reserved.
+#
+#   Redistribution and use in source and binary forms, with or without
+#   modification, are permitted provided that the following conditions
+#   are met:
+#
+# * Redistributions of source code must retain the above copyright
+#   notice, this list of conditions and the following disclaimer.
+# * Redistributions in binary form must reproduce the above copyright
+#   notice, this list of conditions and the following disclaimer in
+#   the documentation and/or other materials provided with the
+#   distribution.
+# * Neither the name of NXP nor the names of its
+#   contributors may be used to endorse or promote products derived
+#   from this software without specific prior written permission.
+#
+#   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+#   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+#   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+#   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+#   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+#   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+#   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+#   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+#   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+#   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+#   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+DIRS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc
+
+include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
new file mode 100644
index 000..c4f22ed
--- /dev/null
+++ b/drivers/bus/fslmc/Makefile
@@ -0,0 +1,52 @@
+#   BSD LICENSE
+#
+#   Copyright (c) 2016 NXP. All rights reserved.
+#
+#   Redistribution and use in source and binary forms, with or without
+#   modification, are permitted provided that the following conditions
+#   are met:
+#
+# * Redistributions of source code must retain the above copyright
+#   notice, this list of conditions and the following disclaimer.
+# * Redistributions in binary form must reproduce the above copyright
+#   notice, this list of conditions and the following disclaimer in
+#   the documentation and/or other materials provided with the
+#   distribution.
+# * Neither the name of NXP nor the names of its
+#   contributors may be used to endorse or promote products derived
+#   from this software without specific prior written permission.
+#
+#   THIS SOFTWARE I

[dpdk-dev] [PATCHv5 08/33] bus/fslmc: add mc dpbp object support

[Hemant Agrawal]

DPBP object represent a hw based buffer pool instance
in the DPAA2 hardware.

Signed-off-by: Alex Marginean 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/Makefile |   1 +
 drivers/bus/fslmc/mc/dpbp.c| 230 +
 drivers/bus/fslmc/mc/fsl_dpbp.h| 220 +++
 drivers/bus/fslmc/mc/fsl_dpbp_cmd.h|  76 
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |   5 +
 5 files changed, 532 insertions(+)
 create mode 100644 drivers/bus/fslmc/mc/dpbp.c
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpbp.h
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpbp_cmd.h

diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index fa1ce13..412715d 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -50,6 +50,7 @@ LIBABIVER := 1
 
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
 mc/dpni.c \
+mc/dpbp.c \
 mc/dpio.c \
 mc/mc_sys.c
 
diff --git a/drivers/bus/fslmc/mc/dpbp.c b/drivers/bus/fslmc/mc/dpbp.c
new file mode 100644
index 000..2260d86
--- /dev/null
+++ b/drivers/bus/fslmc/mc/dpbp.c
@@ -0,0 +1,230 @@
+/* Copyright 2013-2016 Freescale Semiconductor Inc.
+ * Copyright (c) 2016 NXP.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * * Neither the name of the above-listed copyright holders nor the
+ * names of any contributors may be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ *
+ * ALTERNATIVELY, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") as published by the Free Software
+ * Foundation, either version 2 of that License or (at your option) any
+ * later version.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+#include 
+#include 
+#include 
+#include 
+
+int dpbp_open(struct fsl_mc_io *mc_io,
+ uint32_t cmd_flags,
+ int dpbp_id,
+ uint16_t *token)
+{
+   struct mc_command cmd = { 0 };
+   int err;
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPBP_CMDID_OPEN,
+ cmd_flags,
+ 0);
+   DPBP_CMD_OPEN(cmd, dpbp_id);
+
+   /* send command to mc*/
+   err = mc_send_command(mc_io, &cmd);
+   if (err)
+   return err;
+
+   /* retrieve response parameters */
+   *token = MC_CMD_HDR_READ_TOKEN(cmd.header);
+
+   return err;
+}
+
+int dpbp_close(struct fsl_mc_io *mc_io,
+  uint32_t cmd_flags,
+  uint16_t token)
+{
+   struct mc_command cmd = { 0 };
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPBP_CMDID_CLOSE, cmd_flags,
+ token);
+
+   /* send command to mc*/
+   return mc_send_command(mc_io, &cmd);
+}
+
+int dpbp_create(struct fsl_mc_io *mc_io,
+   uint16_t dprc_token,
+   uint32_t cmd_flags,
+   const struct dpbp_cfg *cfg,
+   uint32_t *obj_id)
+{
+   struct mc_command cmd = { 0 };
+   int err;
+
+   (void)(cfg); /* unused */
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPBP_CMDID_CREATE,
+ cmd_flags,
+ dprc_token);
+
+   /* send command to mc*/
+   err = mc_send_command(mc_io, &cmd);
+   if (err)
+   return err;
+
+   /* retrieve response parameters */
+   CMD_CREATE_RSP_GET_OBJ_ID_PARAM0(cmd, *obj_id);
+
+   return 0;
+}
+
+int dpbp_destroy(struct fsl_mc_io *mc_io,
+uint16_t dprc_token,
+   uint32_t cmd_flags,
+ 

[dpdk-dev] [PATCHv5 09/33] bus/fslmc: add mc dpseci object support

[Hemant Agrawal]

dpseci represent a instance of SEC HW in DPAA2.

Signed-off-by: Cristian Sovaiala 
Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/Makefile |   1 +
 drivers/bus/fslmc/mc/dpseci.c  | 527 
 drivers/bus/fslmc/mc/fsl_dpseci.h  | 661 +
 drivers/bus/fslmc/mc/fsl_dpseci_cmd.h  | 248 ++
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |  10 +
 5 files changed, 1447 insertions(+)
 create mode 100644 drivers/bus/fslmc/mc/dpseci.c
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpseci.h
 create mode 100644 drivers/bus/fslmc/mc/fsl_dpseci_cmd.h

diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index 412715d..e422861 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -50,6 +50,7 @@ LIBABIVER := 1
 
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
 mc/dpni.c \
+mc/dpseci.c \
 mc/dpbp.c \
 mc/dpio.c \
 mc/mc_sys.c
diff --git a/drivers/bus/fslmc/mc/dpseci.c b/drivers/bus/fslmc/mc/dpseci.c
new file mode 100644
index 000..173a40c
--- /dev/null
+++ b/drivers/bus/fslmc/mc/dpseci.c
@@ -0,0 +1,527 @@
+/* Copyright 2013-2016 Freescale Semiconductor Inc.
+ * Copyright (c) 2016 NXP.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ * * Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * * Neither the name of the above-listed copyright holders nor the
+ * names of any contributors may be used to endorse or promote products
+ * derived from this software without specific prior written permission.
+ *
+ *
+ * ALTERNATIVELY, this software may be distributed under the terms of the
+ * GNU General Public License ("GPL") as published by the Free Software
+ * Foundation, either version 2 of that License or (at your option) any
+ * later version.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+#include 
+#include 
+#include 
+#include 
+
+int dpseci_open(struct fsl_mc_io *mc_io,
+   uint32_t cmd_flags,
+   int dpseci_id,
+   uint16_t *token)
+{
+   struct mc_command cmd = { 0 };
+   int err;
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPSECI_CMDID_OPEN,
+ cmd_flags,
+ 0);
+   DPSECI_CMD_OPEN(cmd, dpseci_id);
+
+   /* send command to mc*/
+   err = mc_send_command(mc_io, &cmd);
+   if (err)
+   return err;
+
+   /* retrieve response parameters */
+   *token = MC_CMD_HDR_READ_TOKEN(cmd.header);
+
+   return 0;
+}
+
+int dpseci_close(struct fsl_mc_io *mc_io,
+uint32_t cmd_flags,
+uint16_t token)
+{
+   struct mc_command cmd = { 0 };
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPSECI_CMDID_CLOSE,
+ cmd_flags,
+ token);
+
+   /* send command to mc*/
+   return mc_send_command(mc_io, &cmd);
+}
+
+int dpseci_create(struct fsl_mc_io *mc_io,
+ uint16_t  dprc_token,
+ uint32_t  cmd_flags,
+ const struct dpseci_cfg   *cfg,
+ uint32_t  *obj_id)
+{
+   struct mc_command cmd = { 0 };
+   int err;
+
+   /* prepare command */
+   cmd.header = mc_encode_cmd_header(DPSECI_CMDID_CREATE,
+ cmd_flags,
+ dprc_token);
+   DPSECI_CMD_CREATE(cmd, cfg);
+
+   /* send command to mc*/
+   err = mc_send_command(mc_io, &cmd);
+   if (err)
+   return err;
+
+   /* retrieve response parameters */
+   CMD_CREATE_RSP_GET_OBJ_ID_PARAM0(cmd, *obj_id);
+
+   return 0;
+}
+
+int dpseci_

[dpdk-dev] [PATCHv5 10/33] eal/vfio: adding vfio utility functions in map file

[Hemant Agrawal]

adding extra vfio utility functions to map file.
They will be used by other vfio supported buses like fslmc bus
for NXP DPAA2 devices

Signed-off-by: Hemant Agrawal 
---
 lib/librte_eal/bsdapp/eal/rte_eal_version.map   | 3 +++
 lib/librte_eal/linuxapp/eal/rte_eal_version.map | 3 +++
 2 files changed, 6 insertions(+)

diff --git a/lib/librte_eal/bsdapp/eal/rte_eal_version.map 
b/lib/librte_eal/bsdapp/eal/rte_eal_version.map
index 2cf1ac8..e1e0532 100644
--- a/lib/librte_eal/bsdapp/eal/rte_eal_version.map
+++ b/lib/librte_eal/bsdapp/eal/rte_eal_version.map
@@ -183,5 +183,8 @@ DPDK_17.02 {
rte_bus_register;
rte_bus_scan;
rte_bus_unregister;
+vfio_get_container_fd;
+vfio_get_group_fd;
+vfio_get_group_no;
 
 } DPDK_16.11;
diff --git a/lib/librte_eal/linuxapp/eal/rte_eal_version.map 
b/lib/librte_eal/linuxapp/eal/rte_eal_version.map
index 3c68ff5..99d4446 100644
--- a/lib/librte_eal/linuxapp/eal/rte_eal_version.map
+++ b/lib/librte_eal/linuxapp/eal/rte_eal_version.map
@@ -187,5 +187,8 @@ DPDK_17.02 {
rte_bus_register;
rte_bus_scan;
rte_bus_unregister;
+   vfio_get_container_fd;
+   vfio_get_group_fd;
+   vfio_get_group_no;
 
 } DPDK_16.11;
-- 
1.9.1

[dpdk-dev] [PATCHv5 12/33] bus/fslmc: scan for net and sec devices

[Hemant Agrawal]

This patch will add support in fslmc vfio process to
scan and parse the dpni and dpseci object for net and crypto
devices. It will add the scanned devices to the fslmc bus.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/fslmc_vfio.c | 63 +-
 1 file changed, 62 insertions(+), 1 deletion(-)

diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index d51e33a..fd844e2 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -210,6 +210,48 @@ static int64_t vfio_map_mcp_obj(struct fslmc_vfio_group 
*group, char *mcp_obj)
return v_addr;
 }
 
+static inline int
+dpaa2_compare_dpaa2_dev(const struct rte_dpaa2_device *dev,
+const struct rte_dpaa2_device *dev2)
+{
+   /*not the same family device */
+   if (dev->dev_type != DPAA2_MC_DPNI_DEVID ||
+   dev->dev_type != DPAA2_MC_DPSECI_DEVID)
+   return -1;
+
+   if (dev->object_id == dev2->object_id)
+   return 0;
+   else
+   return 1;
+}
+
+static void
+fslmc_bus_add_device(struct rte_dpaa2_device *dev)
+{
+   struct rte_fslmc_device_list *dev_l;
+
+   dev_l = &rte_fslmc_bus.device_list;
+
+   /* device is valid, add in list (sorted) */
+   if (TAILQ_EMPTY(dev_l)) {
+   TAILQ_INSERT_TAIL(dev_l, dev, next);
+   } else {
+   struct rte_dpaa2_device *dev2;
+   int ret;
+
+   TAILQ_FOREACH(dev2, dev_l, next) {
+   ret = dpaa2_compare_dpaa2_dev(dev, dev2);
+   if (ret <= 0)
+   continue;
+
+   TAILQ_INSERT_BEFORE(dev2, dev, next);
+   return;
+   }
+
+   TAILQ_INSERT_TAIL(dev_l, dev, next);
+   }
+}
+
 /* Following function shall fetch total available list of MC devices
  * from VFIO container & populate private list of devices and other
  * data structures
@@ -218,7 +260,7 @@ int fslmc_vfio_process_group(void)
 {
struct fslmc_vfio_device *vdev;
struct vfio_device_info device_info = { .argsz = sizeof(device_info) };
-   char *temp_obj, *object_type __rte_unused, *mcp_obj, *dev_name;
+   char *temp_obj, *object_type, *mcp_obj, *dev_name;
int32_t object_id, i, dev_fd;
DIR *d;
struct dirent *dir;
@@ -348,6 +390,25 @@ int fslmc_vfio_process_group(void)
FSLMC_VFIO_LOG(ERR, "DPAA2 VFIO_DEVICE_GET_INFO fail");
goto FAILURE;
}
+   if (!strcmp(object_type, "dpni") ||
+   !strcmp(object_type, "dpseci")) {
+   struct rte_dpaa2_device *dev;
+
+   dev = malloc(sizeof(struct rte_dpaa2_device));
+   if (dev == NULL)
+   return -1;
+
+   memset(dev, 0, sizeof(*dev));
+   /* store hw_id of dpni/dpseci device */
+   dev->object_id = object_id;
+   dev->dev_type = (strcmp(object_type, "dpseci")) ?
+   DPAA2_MC_DPNI_DEVID : DPAA2_MC_DPSECI_DEVID;
+
+   FSLMC_VFIO_LOG(DEBUG, "DPAA2: Added [%s-%d]\n",
+ object_type, object_id);
+
+   fslmc_bus_add_device(dev);
+   }
}
closedir(d);
 
-- 
1.9.1

[dpdk-dev] [PATCHv5 11/33] bus/fslmc: add vfio support

[Hemant Agrawal]

Add support for using VFIO for dpaa2 based fsl-mc bus.

There are some differences in the way vfio used for fsl-mc bus
from the eal vfio.
 - The scanning of bus for individual objects on the basis of
   the DPRC container.
 - The use and mapping of MC portal for object access

With the evolution of bus model, they canbe further aligned with
eal vfio code.

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini |   1 +
 drivers/bus/fslmc/Makefile |   2 +
 drivers/bus/fslmc/fslmc_bus.c  |  10 +
 drivers/bus/fslmc/fslmc_vfio.c | 450 +
 drivers/bus/fslmc/fslmc_vfio.h |  74 
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |   2 +
 6 files changed, 539 insertions(+)
 create mode 100644 drivers/bus/fslmc/fslmc_vfio.c
 create mode 100644 drivers/bus/fslmc/fslmc_vfio.h

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index b2ad6ec..b176208 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -4,5 +4,6 @@
 ; Refer to default.ini for the full list of available PMD features.
 ;
 [Features]
+Linux VFIO   = Y
 ARMv8= Y
 Usage doc= Y
diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index e422861..a5a05de 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -41,6 +41,7 @@ CFLAGS += "-Wno-strict-aliasing"
 
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/mc
+CFLAGS += -I$(RTE_SDK)/lib/librte_eal/linuxapp/eal
 
 # versioning export map
 EXPORT_MAP := rte_pmd_fslmcbus_version.map
@@ -55,6 +56,7 @@ SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
 mc/dpio.c \
 mc/mc_sys.c
 
+SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_vfio.c
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_bus.c
 
 # library dependencies
diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index 8a4f519..ee794e7 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -42,6 +42,7 @@
 #include 
 
 #include "rte_fslmc.h"
+#include "fslmc_vfio.h"
 
 #define FSLMC_BUS_LOG(level, fmt, args...) \
RTE_LOG(level, EAL, "%s(): " fmt "\n", __func__, ##args)
@@ -51,6 +52,15 @@
 static int
 rte_fslmc_scan(void)
 {
+   if (fslmc_vfio_setup_group()) {
+   FSLMC_BUS_LOG(ERR, "fslmc: Unable to setup VFIO");
+   return -1;
+   }
+   if (fslmc_vfio_process_group()) {
+   FSLMC_BUS_LOG(ERR, "fslmc: Unable to setup devices");
+   return -1;
+   }
+   RTE_LOG(INFO, EAL, "fslmc: Bus scan completed\n");
return 0;
 }
 
diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
new file mode 100644
index 000..d51e33a
--- /dev/null
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -0,0 +1,450 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2015-2016 Freescale Semiconductor, Inc. All rights reserved.
+ *   Copyright (c) 2016 NXP. All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of Freescale Semiconductor, Inc nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include "rte_fslmc.

[dpdk-dev] [PATCHv5 14/33] bus/fslmc: add debug log message support

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 config/common_base|  7 +++
 config/defconfig_arm64-dpaa2-linuxapp-gcc |  5 ++
 drivers/bus/fslmc/Makefile|  5 ++
 drivers/bus/fslmc/fslmc_logs.h| 76 +++
 drivers/common/dpaa2/qbman/Makefile   |  5 ++
 drivers/net/dpaa2/Makefile|  5 ++
 drivers/net/dpaa2/dpaa2_ethdev.c  |  9 +++-
 7 files changed, 110 insertions(+), 2 deletions(-)
 create mode 100644 drivers/bus/fslmc/fslmc_logs.h

diff --git a/config/common_base b/config/common_base
index 6f513fe..9d35487 100644
--- a/config/common_base
+++ b/config/common_base
@@ -301,6 +301,13 @@ CONFIG_RTE_LIBRTE_FSLMC_BUS=n
 # Compile burst-oriented NXP DPAA2 PMD driver
 #
 CONFIG_RTE_LIBRTE_DPAA2_PMD=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_INIT=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_DRIVER=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_RX=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_TX=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_TX_FREE=n
+
+#
 # Compile burst-oriented VIRTIO PMD driver
 #
 CONFIG_RTE_LIBRTE_VIRTIO_PMD=y
diff --git a/config/defconfig_arm64-dpaa2-linuxapp-gcc 
b/config/defconfig_arm64-dpaa2-linuxapp-gcc
index 13c16c0..d3bc9d8 100644
--- a/config/defconfig_arm64-dpaa2-linuxapp-gcc
+++ b/config/defconfig_arm64-dpaa2-linuxapp-gcc
@@ -56,3 +56,8 @@ CONFIG_RTE_LIBRTE_FSLMC_BUS=y
 # Compile burst-oriented NXP DPAA2 PMD driver
 #
 CONFIG_RTE_LIBRTE_DPAA2_PMD=y
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_INIT=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_DRIVER=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_RX=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_TX=n
+CONFIG_RTE_LIBRTE_DPAA2_DEBUG_TX_FREE=n
diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index a5a05de..b74c333 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -35,8 +35,13 @@ include $(RTE_SDK)/mk/rte.vars.mk
 #
 LIB = librte_pmd_fslmcbus.a
 
+ifeq ($(CONFIG_RTE_LIBRTE_DPAA2_DEBUG_INIT),y)
+CFLAGS += -O0 -g
+CFLAGS += "-Wno-error"
+else
 CFLAGS += -O3
 CFLAGS += $(WERROR_FLAGS)
+endif
 CFLAGS += "-Wno-strict-aliasing"
 
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc
diff --git a/drivers/bus/fslmc/fslmc_logs.h b/drivers/bus/fslmc/fslmc_logs.h
new file mode 100644
index 000..a890e6c
--- /dev/null
+++ b/drivers/bus/fslmc/fslmc_logs.h
@@ -0,0 +1,76 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2016 NXP. All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of NXP nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#ifndef _FSLMC_LOGS_H_
+#define _FSLMC_LOGS_H_
+
+#define PMD_INIT_LOG(level, fmt, args...) \
+   RTE_LOG(level, PMD, "%s(): " fmt "\n", __func__, ##args)
+
+#ifdef RTE_LIBRTE_DPAA2_DEBUG_INIT
+#define PMD_INIT_FUNC_TRACE() PMD_INIT_LOG(DEBUG, " >>")
+#else
+#define PMD_INIT_FUNC_TRACE() do { } while (0)
+#endif
+
+#ifdef RTE_LIBRTE_DPAA2_DEBUG_RX
+#define PMD_RX_LOG(level, fmt, args...) \
+   RTE_LOG(level, PMD, "%s(): " fmt "\n", __func__, ## args)
+#else
+#define PMD_RX_LOG(level, fmt, args...) do { } while (0)
+#endif
+
+#ifdef RTE_LIBRTE_DPAA2_DEBUG_TX
+#define PMD_TX_LOG(level, fmt, args...) \
+   RTE_LOG(level, PMD, "%s(): " fmt "\n", __func__, ## args)
+#else
+#define PMD_TX_LOG(level, fmt, args...) do { } while (0)
+#endif
+
+#ifdef RTE_LIBRTE_DPAA2_DEBUG_TX_FREE
+#define PMD_TX_FREE_LOG(level, fmt, args...) \
+   RTE_LOG(level, PMD, "%s(): " fmt "\n", __func__, ## args)
+#else
+#define PMD_TX_FREE_LOG(level, fmt, args...) do { } while (0)
+#endif
+
+#ifdef RTE_LIBRTE_DPAA2_DEBUG_DRIVER
+#define PMD_DRV_LOG_RAW(

[dpdk-dev] [PATCHv5 13/33] net/dpaa2: introducing NXP dpaa2 pmd driver

[Hemant Agrawal]

add support for fsl-mc bus based dpaa2 pmd driver.

Signed-off-by: Hemant Agrawal 
---
 config/common_base  |   4 +
 config/defconfig_arm64-dpaa2-linuxapp-gcc   |   5 +
 drivers/bus/Makefile|   2 +
 drivers/common/Makefile |   2 +
 drivers/net/Makefile|   2 +-
 drivers/net/dpaa2/Makefile  |  59 +++
 drivers/net/dpaa2/dpaa2_ethdev.c| 153 
 drivers/net/dpaa2/dpaa2_ethdev.h|  44 
 drivers/net/dpaa2/rte_pmd_dpaa2_version.map |   4 +
 mk/rte.app.mk   |   5 +
 10 files changed, 279 insertions(+), 1 deletion(-)
 create mode 100644 drivers/net/dpaa2/Makefile
 create mode 100644 drivers/net/dpaa2/dpaa2_ethdev.c
 create mode 100644 drivers/net/dpaa2/dpaa2_ethdev.h
 create mode 100644 drivers/net/dpaa2/rte_pmd_dpaa2_version.map

diff --git a/config/common_base b/config/common_base
index e6b4d60..6f513fe 100644
--- a/config/common_base
+++ b/config/common_base
@@ -297,6 +297,10 @@ CONFIG_RTE_LIBRTE_DPAA2_COMMON=n
 #
 CONFIG_RTE_LIBRTE_FSLMC_BUS=n
 
+#
+# Compile burst-oriented NXP DPAA2 PMD driver
+#
+CONFIG_RTE_LIBRTE_DPAA2_PMD=n
 # Compile burst-oriented VIRTIO PMD driver
 #
 CONFIG_RTE_LIBRTE_VIRTIO_PMD=y
diff --git a/config/defconfig_arm64-dpaa2-linuxapp-gcc 
b/config/defconfig_arm64-dpaa2-linuxapp-gcc
index 800e22b..13c16c0 100644
--- a/config/defconfig_arm64-dpaa2-linuxapp-gcc
+++ b/config/defconfig_arm64-dpaa2-linuxapp-gcc
@@ -51,3 +51,8 @@ CONFIG_RTE_LIBRTE_DPAA2_COMMON=y
 # Compile NXP DPAA2 FSL-MC Bus
 #
 CONFIG_RTE_LIBRTE_FSLMC_BUS=y
+
+#
+# Compile burst-oriented NXP DPAA2 PMD driver
+#
+CONFIG_RTE_LIBRTE_DPAA2_PMD=y
diff --git a/drivers/bus/Makefile b/drivers/bus/Makefile
index 60e9764..8f7864b 100644
--- a/drivers/bus/Makefile
+++ b/drivers/bus/Makefile
@@ -31,6 +31,8 @@
 
 include $(RTE_SDK)/mk/rte.vars.mk
 
+CONFIG_RTE_LIBRTE_FSLMC_BUS = $(CONFIG_RTE_LIBRTE_DPAA2_PMD)
+
 DIRS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc
 
 include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/common/Makefile b/drivers/common/Makefile
index e5bfecb..76ec2d1 100644
--- a/drivers/common/Makefile
+++ b/drivers/common/Makefile
@@ -31,6 +31,8 @@
 
 include $(RTE_SDK)/mk/rte.vars.mk
 
+CONFIG_RTE_LIBRTE_DPAA2_COMMON = $(CONFIG_RTE_LIBRTE_DPAA2_PMD)
+
 DIRS-$(CONFIG_RTE_LIBRTE_DPAA2_COMMON) += dpaa2
 
 include $(RTE_SDK)/mk/rte.subdir.mk
diff --git a/drivers/net/Makefile b/drivers/net/Makefile
index 40fc333..f716ca0 100644
--- a/drivers/net/Makefile
+++ b/drivers/net/Makefile
@@ -57,7 +57,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_THUNDERX_NICVF_PMD) += thunderx
 DIRS-$(CONFIG_RTE_LIBRTE_VIRTIO_PMD) += virtio
 DIRS-$(CONFIG_RTE_LIBRTE_VMXNET3_PMD) += vmxnet3
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_XENVIRT) += xenvirt
-
+DIRS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += dpaa2
 ifeq ($(CONFIG_RTE_LIBRTE_VHOST),y)
 DIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += vhost
 endif # $(CONFIG_RTE_LIBRTE_VHOST)
diff --git a/drivers/net/dpaa2/Makefile b/drivers/net/dpaa2/Makefile
new file mode 100644
index 000..9e7f958
--- /dev/null
+++ b/drivers/net/dpaa2/Makefile
@@ -0,0 +1,59 @@
+#   BSD LICENSE
+#
+#   Copyright (c) 2016 Freescale Semiconductor, Inc. All rights reserved.
+#   Copyright (c) 2016 NXP. All rights reserved.
+#
+#   Redistribution and use in source and binary forms, with or without
+#   modification, are permitted provided that the following conditions
+#   are met:
+#
+# * Redistributions of source code must retain the above copyright
+#   notice, this list of conditions and the following disclaimer.
+# * Redistributions in binary form must reproduce the above copyright
+#   notice, this list of conditions and the following disclaimer in
+#   the documentation and/or other materials provided with the
+#   distribution.
+# * Neither the name of Freescale Semiconductor, Inc nor the names of its
+#   contributors may be used to endorse or promote products derived
+#   from this software without specific prior written permission.
+#
+#   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+#   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+#   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+#   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+#   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+#   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+#   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+#   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+#   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+#   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+#   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+include $(RTE_SDK)/mk/rte.vars.mk
+
+#
+# library name
+#
+LIB = librte_pmd_dpaa2.a
+
+CFLAGS += -O3
+

[dpdk-dev] [PATCHv5 15/33] drivers/common/dpaa2: dpio portal driver

[Hemant Agrawal]

The portal driver is bound to DPIO objects discovered on the fsl-mc bus and
provides services that:
- allow other drivers, such as the Ethernet driver, to enqueue and dequeue
  frames for their respective objects

A system will typically allocate 1 DPIO object per CPU to allow queuing
operations to happen simultaneously across all CPUs.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/Makefile |   3 +
 drivers/bus/fslmc/fslmc_vfio.c |  17 +-
 drivers/bus/fslmc/fslmc_vfio.h |   5 +
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c   | 364 +
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.h   |  60 
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h|  68 +
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |   2 +
 drivers/common/Makefile|   4 +
 8 files changed, 522 insertions(+), 1 deletion(-)
 create mode 100644 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
 create mode 100644 drivers/bus/fslmc/portal/dpaa2_hw_dpio.h
 create mode 100644 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h

diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index b74c333..1b815dd 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -46,6 +46,7 @@ CFLAGS += "-Wno-strict-aliasing"
 
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/mc
+CFLAGS += -I$(RTE_SDK)/drivers/common/dpaa2/qbman/include
 CFLAGS += -I$(RTE_SDK)/lib/librte_eal/linuxapp/eal
 
 # versioning export map
@@ -61,10 +62,12 @@ SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
 mc/dpio.c \
 mc/mc_sys.c
 
+SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += portal/dpaa2_hw_dpio.c
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_vfio.c
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_bus.c
 
 # library dependencies
 DEPDIRS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += lib/librte_eal
+DEPDIRS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += lib/librte_pmd_dpaa2_qbman
 
 include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index fd844e2..8d24620 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -61,6 +61,9 @@
 #include "rte_fslmc.h"
 #include "fslmc_vfio.h"
 
+#include "portal/dpaa2_hw_pvt.h"
+#include "portal/dpaa2_hw_dpio.h"
+
 #define VFIO_MAX_CONTAINERS1
 
 #define FSLMC_VFIO_LOG(level, fmt, args...) \
@@ -261,12 +264,13 @@ int fslmc_vfio_process_group(void)
struct fslmc_vfio_device *vdev;
struct vfio_device_info device_info = { .argsz = sizeof(device_info) };
char *temp_obj, *object_type, *mcp_obj, *dev_name;
-   int32_t object_id, i, dev_fd;
+   int32_t object_id, i, dev_fd, ret;
DIR *d;
struct dirent *dir;
char path[PATH_MAX];
int64_t v_addr;
int ndev_count;
+   int dpio_count = 0;
struct fslmc_vfio_group *group = &vfio_groups[0];
static int process_once;
 
@@ -409,9 +413,20 @@ int fslmc_vfio_process_group(void)
 
fslmc_bus_add_device(dev);
}
+   if (!strcmp(object_type, "dpio")) {
+   ret = dpaa2_create_dpio_device(vdev,
+  &device_info,
+  object_id);
+   if (!ret)
+   dpio_count++;
+   }
}
closedir(d);
 
+   ret = dpaa2_affine_qbman_swp();
+   if (ret)
+   FSLMC_VFIO_LOG(DEBUG, "Error in affining qbman swp %d", ret);
+
return 0;
 
 FAILURE:
diff --git a/drivers/bus/fslmc/fslmc_vfio.h b/drivers/bus/fslmc/fslmc_vfio.h
index 5e58211..39994dd 100644
--- a/drivers/bus/fslmc/fslmc_vfio.h
+++ b/drivers/bus/fslmc/fslmc_vfio.h
@@ -71,4 +71,9 @@ int vfio_dmamap_mem_region(
 int fslmc_vfio_setup_group(void);
 int fslmc_vfio_process_group(void);
 
+/* create dpio device */
+int dpaa2_create_dpio_device(struct fslmc_vfio_device *vdev,
+struct vfio_device_info *obj_info,
+int object_id);
+
 #endif /* _FSLMC_VFIO_H_ */
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
new file mode 100644
index 000..011bd9f
--- /dev/null
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -0,0 +1,364 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2016 Freescale Semiconductor, Inc. All rights reserved.
+ *   Copyright (c) 2016 NXP. All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in

[dpdk-dev] [PATCHv5 17/33] drivers/common/dpaa2: dpio routine to affine to crypto threads

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c   | 45 ++
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.h   |  3 ++
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |  1 +
 3 files changed, 49 insertions(+)

diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index 011bd9f..d7de0d5 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -276,6 +276,51 @@ static inline struct dpaa2_dpio_dev 
*dpaa2_get_qbman_swp(void)
 }
 
 int
+dpaa2_affine_qbman_swp_sec(void)
+{
+   unsigned int lcore_id = rte_lcore_id();
+   uint64_t tid = syscall(SYS_gettid);
+
+   if (lcore_id == LCORE_ID_ANY)
+   lcore_id = rte_get_master_lcore();
+   /* if the core id is not supported */
+   else if (lcore_id >= RTE_MAX_LCORE)
+   return -1;
+
+   if (dpaa2_io_portal[lcore_id].sec_dpio_dev) {
+   PMD_DRV_LOG(INFO, "DPAA Portal=0x%x (%d) is being shared"
+   " between thread %lu and current  %lu",
+   dpaa2_io_portal[lcore_id].sec_dpio_dev,
+   dpaa2_io_portal[lcore_id].sec_dpio_dev->index,
+   dpaa2_io_portal[lcore_id].sec_tid,
+   tid);
+   RTE_PER_LCORE(_dpaa2_io).sec_dpio_dev
+   = dpaa2_io_portal[lcore_id].sec_dpio_dev;
+   rte_atomic16_inc(&dpaa2_io_portal
+[lcore_id].sec_dpio_dev->ref_count);
+   dpaa2_io_portal[lcore_id].sec_tid = tid;
+
+   PMD_DRV_LOG(DEBUG, "Old Portal=0x%x (%d) affined thread - %lu",
+   dpaa2_io_portal[lcore_id].sec_dpio_dev,
+   dpaa2_io_portal[lcore_id].sec_dpio_dev->index,
+   tid);
+   return 0;
+   }
+
+   /* Populate the dpaa2_io_portal structure */
+   dpaa2_io_portal[lcore_id].sec_dpio_dev = dpaa2_get_qbman_swp();
+
+   if (dpaa2_io_portal[lcore_id].sec_dpio_dev) {
+   RTE_PER_LCORE(_dpaa2_io).sec_dpio_dev
+   = dpaa2_io_portal[lcore_id].sec_dpio_dev;
+   dpaa2_io_portal[lcore_id].sec_tid = tid;
+   return 0;
+   } else {
+   return -1;
+   }
+}
+
+int
 dpaa2_create_dpio_device(struct fslmc_vfio_device *vdev,
 struct vfio_device_info *obj_info,
int object_id)
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h
index 682f3fa..b1a1b8f 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h
@@ -56,5 +56,8 @@ struct dpaa2_io_portal_t {
 /* Affine a DPIO portal to current processing thread */
 int dpaa2_affine_qbman_swp(void);
 
+/* Affine additional DPIO portal to current crypto processing thread */
+int dpaa2_affine_qbman_swp_sec(void);
+
 
 #endif /* _DPAA2_HW_DPIO_H_ */
diff --git a/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map 
b/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
index 76029b9..6937ad0 100644
--- a/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
+++ b/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
@@ -2,6 +2,7 @@ DPDK_17.02 {
global:
 
 dpaa2_affine_qbman_swp;
+dpaa2_affine_qbman_swp_sec;
 dpaa2_alloc_dpbp_dev;
 dpaa2_free_dpbp_dev;
 dpbp_disable;
-- 
1.9.1

[dpdk-dev] [PATCHv5 16/33] drivers/pool/dpaa2: adding hw offloaded mempool

[Hemant Agrawal]

Adding NXP DPAA2 architecture specific mempool support
Each mempool instance is represented by a DPBP object
from the FSL-MC bus.

This patch also registers a dpaa2 type MEMPOOL OPS

Signed-off-by: Hemant Agrawal 
---
 config/common_base|   1 +
 config/defconfig_arm64-dpaa2-linuxapp-gcc |   4 +
 drivers/Makefile  |   1 +
 drivers/bus/fslmc/Makefile|   2 +
 drivers/bus/fslmc/fslmc_vfio.c|   9 +-
 drivers/bus/fslmc/fslmc_vfio.h|   2 +
 drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c  | 137 +
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h   |  19 ++
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map|   2 +
 drivers/common/Makefile   |   3 +
 drivers/pool/Makefile |  38 +++
 drivers/pool/dpaa2/Makefile   |  67 +
 drivers/pool/dpaa2/dpaa2_hw_mempool.c | 339 ++
 drivers/pool/dpaa2/dpaa2_hw_mempool.h |  95 ++
 drivers/pool/dpaa2/rte_pmd_dpaa2_pool_version.map |   8 +
 mk/rte.app.mk |   1 +
 16 files changed, 727 insertions(+), 1 deletion(-)
 create mode 100644 drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c
 create mode 100644 drivers/pool/Makefile
 create mode 100644 drivers/pool/dpaa2/Makefile
 create mode 100644 drivers/pool/dpaa2/dpaa2_hw_mempool.c
 create mode 100644 drivers/pool/dpaa2/dpaa2_hw_mempool.h
 create mode 100644 drivers/pool/dpaa2/rte_pmd_dpaa2_pool_version.map

diff --git a/config/common_base b/config/common_base
index 9d35487..5a10e45 100644
--- a/config/common_base
+++ b/config/common_base
@@ -291,6 +291,7 @@ CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_MBOX=n
 # Compile Support Libraries for NXP DPAA2
 #
 CONFIG_RTE_LIBRTE_DPAA2_COMMON=n
+CONFIG_RTE_LIBRTE_DPAA2_POOL=n
 
 #
 # Compile NXP DPAA2 FSL-MC Bus
diff --git a/config/defconfig_arm64-dpaa2-linuxapp-gcc 
b/config/defconfig_arm64-dpaa2-linuxapp-gcc
index d3bc9d8..7665912 100644
--- a/config/defconfig_arm64-dpaa2-linuxapp-gcc
+++ b/config/defconfig_arm64-dpaa2-linuxapp-gcc
@@ -42,10 +42,14 @@ CONFIG_RTE_ARCH_ARM_TUNE="cortex-a57+fp+simd"
 CONFIG_RTE_MAX_LCORE=8
 CONFIG_RTE_MAX_NUMA_NODES=1
 
+CONFIG_RTE_PKTMBUF_HEADROOM=256
+
 #
 # Compile Support Libraries for DPAA2
 #
 CONFIG_RTE_LIBRTE_DPAA2_COMMON=y
+CONFIG_RTE_LIBRTE_DPAA2_POOL=n
+CONFIG_RTE_MBUF_DEFAULT_MEMPOOL_OPS="dpaa2"
 
 #
 # Compile NXP DPAA2 FSL-MC Bus
diff --git a/drivers/Makefile b/drivers/Makefile
index bdae63b..9fd268e 100644
--- a/drivers/Makefile
+++ b/drivers/Makefile
@@ -33,6 +33,7 @@ include $(RTE_SDK)/mk/rte.vars.mk
 
 DIRS-y += common
 DIRS-y += bus
+DIRS-y += pool
 DIRS-y += net
 DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += crypto
 
diff --git a/drivers/bus/fslmc/Makefile b/drivers/bus/fslmc/Makefile
index 1b815dd..35f30ad 100644
--- a/drivers/bus/fslmc/Makefile
+++ b/drivers/bus/fslmc/Makefile
@@ -47,6 +47,7 @@ CFLAGS += "-Wno-strict-aliasing"
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/mc
 CFLAGS += -I$(RTE_SDK)/drivers/common/dpaa2/qbman/include
+CFLAGS += -I$(RTE_SDK)/drivers/pool/dpaa2
 CFLAGS += -I$(RTE_SDK)/lib/librte_eal/linuxapp/eal
 
 # versioning export map
@@ -63,6 +64,7 @@ SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += \
 mc/mc_sys.c
 
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += portal/dpaa2_hw_dpio.c
+SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += portal/dpaa2_hw_dpbp.c
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_vfio.c
 SRCS-$(CONFIG_RTE_LIBRTE_FSLMC_BUS) += fslmc_bus.c
 
diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index 8d24620..2bfd7fd 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -270,7 +270,7 @@ int fslmc_vfio_process_group(void)
char path[PATH_MAX];
int64_t v_addr;
int ndev_count;
-   int dpio_count = 0;
+   int dpio_count = 0, dpbp_count = 0;
struct fslmc_vfio_group *group = &vfio_groups[0];
static int process_once;
 
@@ -420,6 +420,11 @@ int fslmc_vfio_process_group(void)
if (!ret)
dpio_count++;
}
+   if (!strcmp(object_type, "dpbp")) {
+   ret = dpaa2_create_dpbp_device(object_id);
+   if (!ret)
+   dpbp_count++;
+   }
}
closedir(d);
 
@@ -427,6 +432,8 @@ int fslmc_vfio_process_group(void)
if (ret)
FSLMC_VFIO_LOG(DEBUG, "Error in affining qbman swp %d", ret);
 
+   FSLMC_VFIO_LOG(DEBUG, "DPAA2: Added dpbp_count = %d dpio_count=%d\n",
+ dpbp_count, dpio_count);
return 0;
 
 FAILURE:
diff --git a/drivers/bus/fslmc/fslmc_vfio.h b/drivers/bus/fslmc/fslmc_vfio.h
index 39994dd..80c6869 100644
--- a/drivers/bus/fslmc/fslmc_vfio.h
+++ b/drivers/bus/fslmc/fslmc_vfio.h
@@ -76,4 +76,6 @@ int

[dpdk-dev] [PATCHv5 19/33] net/dpaa2: add rss flow distribution

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini |   1 +
 drivers/net/dpaa2/Makefile |   1 +
 drivers/net/dpaa2/base/dpaa2_hw_dpni.c | 287 +
 drivers/net/dpaa2/dpaa2_ethdev.c   |  31 +++-
 drivers/net/dpaa2/dpaa2_ethdev.h   |  12 ++
 5 files changed, 328 insertions(+), 4 deletions(-)
 create mode 100644 drivers/net/dpaa2/base/dpaa2_hw_dpni.c

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index 0b59725..20152a0 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -5,6 +5,7 @@
 ;
 [Features]
 Queue start/stop = Y
+RSS hash = Y
 Linux VFIO   = Y
 ARMv8= Y
 Usage doc= Y
diff --git a/drivers/net/dpaa2/Makefile b/drivers/net/dpaa2/Makefile
index 61831cc..657ee2a 100644
--- a/drivers/net/dpaa2/Makefile
+++ b/drivers/net/dpaa2/Makefile
@@ -57,6 +57,7 @@ EXPORT_MAP := rte_pmd_dpaa2_version.map
 # library version
 LIBABIVER := 1
 
+SRCS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += base/dpaa2_hw_dpni.c
 SRCS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += dpaa2_ethdev.c
 
 # library dependencies
diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c 
b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
new file mode 100644
index 000..c95c083
--- /dev/null
+++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
@@ -0,0 +1,287 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2016 Freescale Semiconductor, Inc. All rights reserved.
+ *   Copyright (c) 2016 NXP. All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of Freescale Semiconductor, Inc nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include 
+#include 
+
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+#include 
+
+#include 
+#include 
+
+#include "../dpaa2_ethdev.h"
+
+static void
+dpaa2_distset_to_dpkg_profile_cfg(
+   uint32_t req_dist_set,
+   struct dpkg_profile_cfg *kg_cfg);
+
+int
+dpaa2_setup_flow_dist(struct rte_eth_dev *eth_dev,
+ uint32_t req_dist_set)
+{
+   struct dpaa2_dev_priv *priv = eth_dev->data->dev_private;
+   struct fsl_mc_io *dpni = priv->hw;
+   struct dpni_rx_tc_dist_cfg tc_cfg;
+   struct dpkg_profile_cfg kg_cfg;
+   void *p_params;
+   int ret, tc_index = 0;
+
+   p_params = rte_malloc(
+   NULL, DIST_PARAM_IOVA_SIZE, RTE_CACHE_LINE_SIZE);
+   if (!p_params) {
+   RTE_LOG(ERR, PMD, "Memory unavaialble\n");
+   return -ENOMEM;
+   }
+   memset(p_params, 0, DIST_PARAM_IOVA_SIZE);
+   memset(&tc_cfg, 0, sizeof(struct dpni_rx_tc_dist_cfg));
+
+   dpaa2_distset_to_dpkg_profile_cfg(req_dist_set, &kg_cfg);
+   tc_cfg.key_cfg_iova = (uint64_t)(p_params);
+   tc_cfg.dist_size = eth_dev->data->nb_rx_queues;
+   tc_cfg.dist_mode = DPNI_DIST_MODE_HASH;
+
+   ret = dpni_prepare_key_cfg(&kg_cfg, p_params);
+   if (ret) {
+   RTE_LOG(ERR, PMD, "Unable to prepare extract parameters\n");
+   rte_free(p_params);
+   return ret;
+   }
+
+   ret = dpni_set_rx_tc_dist(dpni, CMD_PRI_LOW, priv->token, tc_index,
+ &tc_cfg);
+   rte_free(p_params);
+   if (ret) {
+   RTE_LOG(ERR, PMD, "Setting distribution for Rx failed with"
+   " err code: %d\n", ret);
+   return ret;
+   }
+
+   return 0;
+}
+
+int dpaa2

[dpdk-dev] [PATCHv5 18/33] net/dpaa2: adding eth ops to dpaa2

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini  |   1 +
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h |  20 ++
 drivers/net/dpaa2/Makefile  |   3 +
 drivers/net/dpaa2/dpaa2_ethdev.c| 412 +++-
 drivers/net/dpaa2/dpaa2_ethdev.h|  15 ++
 5 files changed, 450 insertions(+), 1 deletion(-)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index b176208..0b59725 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -4,6 +4,7 @@
 ; Refer to default.ini for the full list of available PMD features.
 ;
 [Features]
+Queue start/stop = Y
 Linux VFIO   = Y
 ARMv8= Y
 Usage doc= Y
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index 3b846a0..660537d 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -37,9 +37,12 @@
 #include 
 #include 
 
+#define DPAA2_DQRR_RING_SIZE   16
+   /** 
 #include 
+#include 
+
 #include "dpaa2_ethdev.h"
 
 /* Name of the DPAA2 Net PMD */
 static const char *drivername = "DPAA2 PMD";
 
+static void
+dpaa2_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
+{
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+
+   PMD_INIT_FUNC_TRACE();
+
+   dev_info->if_index = priv->hw_id;
+
+   dev_info->max_rx_queues = (uint16_t)priv->nb_rx_queues;
+   dev_info->max_tx_queues = (uint16_t)priv->nb_tx_queues;
+
+   dev_info->speed_capa = ETH_LINK_SPEED_1G |
+   ETH_LINK_SPEED_2_5G |
+   ETH_LINK_SPEED_10G;
+}
+
+static int
+dpaa2_alloc_rx_tx_queues(struct rte_eth_dev *dev)
+{
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   uint16_t dist_idx;
+   uint32_t vq_id;
+   struct dpaa2_queue *mc_q, *mcq;
+   uint32_t tot_queues;
+   int i;
+   struct dpaa2_queue *dpaa2_q;
+
+   PMD_INIT_FUNC_TRACE();
+
+   tot_queues = priv->nb_rx_queues + priv->nb_tx_queues;
+   mc_q = rte_malloc(NULL, sizeof(struct dpaa2_queue) * tot_queues,
+ RTE_CACHE_LINE_SIZE);
+   if (!mc_q) {
+   PMD_INIT_LOG(ERR, "malloc failed for rx/tx queues\n");
+   return -1;
+   }
+
+   for (i = 0; i < priv->nb_rx_queues; i++) {
+   mc_q->dev = dev;
+   priv->rx_vq[i] = mc_q++;
+   dpaa2_q = (struct dpaa2_queue *)priv->rx_vq[i];
+   dpaa2_q->q_storage = rte_malloc("dq_storage",
+   sizeof(struct queue_storage_info_t),
+   RTE_CACHE_LINE_SIZE);
+   if (!dpaa2_q->q_storage)
+   goto fail;
+
+   memset(dpaa2_q->q_storage, 0,
+  sizeof(struct queue_storage_info_t));
+   dpaa2_q->q_storage->dq_storage[0] = rte_malloc(NULL,
+   DPAA2_DQRR_RING_SIZE * sizeof(struct qbman_result),
+   RTE_CACHE_LINE_SIZE);
+   }
+
+   for (i = 0; i < priv->nb_tx_queues; i++) {
+   mc_q->dev = dev;
+   mc_q->flow_id = DPNI_NEW_FLOW_ID;
+   priv->tx_vq[i] = mc_q++;
+   }
+
+   vq_id = 0;
+   for (dist_idx = 0; dist_idx < priv->nb_rx_queues; dist_idx++) {
+   mcq = (struct dpaa2_queue *)priv->rx_vq[vq_id];
+   mcq->tc_index = DPAA2_DEF_TC;
+   mcq->flow_id = dist_idx;
+   vq_id++;
+   }
+
+   return 0;
+fail:
+   i -= 1;
+   mc_q = priv->rx_vq[0];
+   while (i >= 0) {
+   dpaa2_q = (struct dpaa2_queue *)priv->rx_vq[i];
+   rte_free(dpaa2_q->q_storage->dq_storage[0]);
+   rte_free(dpaa2_q->q_storage);
+   priv->rx_vq[i--] = NULL;
+   }
+   rte_free(mc_q);
+   return -1;
+}
+
+static int
+dpaa2_eth_dev_configure(struct rte_eth_dev *dev)
+{
+   struct rte_eth_dev_data *data = dev->data;
+   struct rte_eth_conf *eth_conf = &data->dev_conf;
+
+   PMD_INIT_FUNC_TRACE();
+
+   /* Check for correct configuration */
+   if (eth_conf->rxmode.mq_mode != ETH_MQ_RX_RSS &&
+   data->nb_rx_queues > 1) {
+   PMD_INIT_LOG(ERR, "Distribution is not enabled, "
+   "but Rx queues more than 1\n");
+   return -1;
+   }
+
+   return 0;
+}
+
+/* Function to setup RX flow information. It contains traffic class ID,
+ * flow ID, destination configuration etc.
+ */
+static int
+dpaa2_dev_rx_queue_setup(struct rte_eth_dev *dev,
+uint16_t rx_queue_id,
+uint16_t nb_rx_desc __rte_unused,
+unsigned int socket_id __rte_unused,
+const struct rte_eth_rxconf *rx_conf __rte_unused,
+struct rte_mempool *mb_po

[dpdk-dev] [PATCHv5 20/33] net/dpaa2: configure mac address at init

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 28 
 drivers/net/dpaa2/dpaa2_ethdev.h |  3 +++
 2 files changed, 31 insertions(+)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 1d7ca66..54f4498 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -63,6 +63,7 @@
 
dev_info->if_index = priv->hw_id;
 
+   dev_info->max_mac_addrs = priv->max_mac_filters;
dev_info->max_rx_queues = (uint16_t)priv->nb_rx_queues;
dev_info->max_tx_queues = (uint16_t)priv->nb_tx_queues;
 
@@ -447,6 +448,9 @@
 
priv->hw = dpni_dev;
priv->hw_id = hw_id;
+   priv->options = attr.options;
+   priv->max_mac_filters = attr.mac_filter_entries;
+   priv->max_vlan_filters = attr.vlan_filter_entries;
priv->flags = 0;
 
ret = dpaa2_alloc_rx_tx_queues(eth_dev);
@@ -455,6 +459,25 @@
return -ret;
}
 
+   /* Allocate memory for storing MAC addresses */
+   eth_dev->data->mac_addrs = rte_zmalloc("dpni",
+   ETHER_ADDR_LEN * attr.mac_filter_entries, 0);
+   if (eth_dev->data->mac_addrs == NULL) {
+   PMD_INIT_LOG(ERR, "Failed to allocate %d bytes needed to "
+   "store MAC addresses",
+   ETHER_ADDR_LEN * attr.mac_filter_entries);
+   return -ENOMEM;
+   }
+
+   ret = dpni_get_primary_mac_addr(dpni_dev, CMD_PRI_LOW,
+   priv->token,
+   (uint8_t *)(eth_dev->data->mac_addrs[0].addr_bytes));
+   if (ret) {
+   PMD_INIT_LOG(ERR, "DPNI get mac address failed:"
+   " Error Code = %d\n", ret);
+   return -ret;
+   }
+
eth_dev->dev_ops = &dpaa2_ethdev_ops;
eth_dev->data->drv_name = drivername;
 
@@ -493,6 +516,11 @@
priv->rx_vq[0] = NULL;
}
 
+   /* Allocate memory for storing MAC addresses */
+   if (eth_dev->data->mac_addrs) {
+   rte_free(eth_dev->data->mac_addrs);
+   eth_dev->data->mac_addrs = NULL;
+   }
 
/*Close the device at underlying layer*/
ret = dpni_close(dpni, CMD_PRI_LOW, priv->token);
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.h b/drivers/net/dpaa2/dpaa2_ethdev.h
index d24fcc6..2d13137 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.h
+++ b/drivers/net/dpaa2/dpaa2_ethdev.h
@@ -57,7 +57,10 @@ struct dpaa2_dev_priv {
void *rx_vq[MAX_RX_QUEUES];
void *tx_vq[MAX_TX_QUEUES];
 
+   uint32_t options;
uint16_t num_dist_per_tc[MAX_TCS];
+   uint8_t max_mac_filters;
+   uint8_t max_vlan_filters;
uint8_t num_tc;
uint8_t flags; /*dpaa2 config flags */
 };
-- 
1.9.1

[dpdk-dev] [PATCHv5 21/33] net/dpaa2: attach the buffer pool to dpni

[Hemant Agrawal]

This patch configures a MC-DPNI based DPAA2 PMD network
port with a DPBP based buffer pool.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h | 10 ++
 drivers/net/dpaa2/Makefile  |  3 ++
 drivers/net/dpaa2/base/dpaa2_hw_dpni.c  | 57 ++
 drivers/net/dpaa2/dpaa2_ethdev.c| 62 +
 drivers/net/dpaa2/dpaa2_ethdev.h|  6 
 5 files changed, 138 insertions(+)

diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index 660537d..b4f243b 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -50,6 +50,16 @@
 #define DPAA2_MBUF_MAX_ACQ_REL 7
 
 #define MAX_BPID 256
+#define DPAA2_MBUF_HW_ANNOTATION   64
+#define DPAA2_FD_PTA_SIZE  64
+
+#if (DPAA2_MBUF_HW_ANNOTATION + DPAA2_FD_PTA_SIZE) > RTE_PKTMBUF_HEADROOM
+#error "Annotation requirement is more than RTE_PKTMBUF_HEADROOM"
+#endif
+
+/* we will re-use the HEADROOM for annotation in RX */
+#define DPAA2_HW_BUF_RESERVE   0
+#define DPAA2_PACKET_LAYOUT_ALIGN  64 /*changing from 256 */
 
 struct dpaa2_dpio_dev {
TAILQ_ENTRY(dpaa2_dpio_dev) next;
diff --git a/drivers/net/dpaa2/Makefile b/drivers/net/dpaa2/Makefile
index 657ee2a..ca51402 100644
--- a/drivers/net/dpaa2/Makefile
+++ b/drivers/net/dpaa2/Makefile
@@ -49,6 +49,7 @@ CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/mc
 CFLAGS += -I$(RTE_SDK)/drivers/bus/fslmc/portal
 CFLAGS += -I$(RTE_SDK)/drivers/common/dpaa2/qbman/include
+CFLAGS += -I$(RTE_SDK)/drivers/pool/dpaa2
 CFLAGS += -I$(RTE_SDK)/lib/librte_eal/linuxapp/eal
 
 # versioning export map
@@ -62,7 +63,9 @@ SRCS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += dpaa2_ethdev.c
 
 # library dependencies
 DEPDIRS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += lib/librte_eal lib/librte_ether
+DEPDIRS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += lib/librte_mempool lib/librte_mbuf
 DEPDIRS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += lib/librte_pmd_dpaa2_qbman
 DEPDIRS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += lib/librte_pmd_fslmcbus
+DEPDIRS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += lib/librte_pmd_dpaa2_pool
 
 include $(RTE_SDK)/mk/rte.lib.mk
diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c 
b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
index c95c083..08f53b3 100644
--- a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
+++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
@@ -46,6 +46,7 @@
 
 #include 
 #include 
+#include 
 
 #include "../dpaa2_ethdev.h"
 
@@ -285,3 +286,59 @@ int dpaa2_remove_flow_dist(
}
kg_cfg->num_extracts = i;
 }
+
+int
+dpaa2_attach_bp_list(struct dpaa2_dev_priv *priv,
+void *blist)
+{
+   /* Function to attach a DPNI with a buffer pool list. Buffer pool list
+* handle is passed in blist.
+*/
+   int32_t retcode;
+   struct fsl_mc_io *dpni = priv->hw;
+   struct dpni_pools_cfg bpool_cfg;
+   struct dpaa2_bp_list *bp_list = (struct dpaa2_bp_list *)blist;
+   struct dpni_buffer_layout layout;
+   int tot_size;
+
+   /* ... rx buffer layout .
+* Check alignment for buffer layouts first
+*/
+
+   /* ... rx buffer layout ... */
+   tot_size = DPAA2_HW_BUF_RESERVE + RTE_PKTMBUF_HEADROOM;
+   tot_size = RTE_ALIGN_CEIL(tot_size,
+ DPAA2_PACKET_LAYOUT_ALIGN);
+
+   memset(&layout, 0, sizeof(struct dpni_buffer_layout));
+   layout.options = DPNI_BUF_LAYOUT_OPT_DATA_HEAD_ROOM;
+
+   layout.data_head_room =
+   tot_size - DPAA2_FD_PTA_SIZE - DPAA2_MBUF_HW_ANNOTATION;
+   retcode = dpni_set_buffer_layout(dpni, CMD_PRI_LOW, priv->token,
+DPNI_QUEUE_RX, &layout);
+   if (retcode) {
+   PMD_INIT_LOG(ERR, "Err(%d) in setting rx buffer layout\n",
+retcode);
+   return retcode;
+   }
+
+   /*Attach buffer pool to the network interface as described by the user*/
+   bpool_cfg.num_dpbp = 1;
+   bpool_cfg.pools[0].dpbp_id = bp_list->buf_pool.dpbp_node->dpbp_id;
+   bpool_cfg.pools[0].backup_pool = 0;
+   bpool_cfg.pools[0].buffer_size =
+   RTE_ALIGN_CEIL(bp_list->buf_pool.size,
+  256 /*DPAA2_PACKET_LAYOUT_ALIGN*/);
+
+   retcode = dpni_set_pools(dpni, CMD_PRI_LOW, priv->token, &bpool_cfg);
+   if (retcode != 0) {
+   PMD_INIT_LOG(ERR, "Error in attaching the buffer pool list"
+   " bpid = %d Error code = %d\n",
+   bpool_cfg.pools[0].dpbp_id, retcode);
+   return retcode;
+   }
+
+   priv->bp_list = bp_list;
+   return 0;
+}
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 54f4498..6bd18bc 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -48,6 +48,7 @@
 #include 
 #include 
 #i

[dpdk-dev] [PATCHv5 22/33] net/dpaa2: add support for l3 and l4 checksum offload

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini  |  2 +
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h |  6 +++
 drivers/net/dpaa2/dpaa2_ethdev.c| 72 +++--
 3 files changed, 76 insertions(+), 4 deletions(-)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index 20152a0..d50c62e 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -6,6 +6,8 @@
 [Features]
 Queue start/stop = Y
 RSS hash = Y
+L3 checksum offload  = Y
+L4 checksum offload  = Y
 Linux VFIO   = Y
 ARMv8= Y
 Usage doc= Y
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index b4f243b..71361a4 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -37,6 +37,12 @@
 #include 
 #include 
 
+#ifndef false
+#define false  0
+#endif
+#ifndef true
+#define true   1
+#endif
 #define DPAA2_DQRR_RING_SIZE   16
/** min_rx_bufsize = DPAA2_MIN_RX_BUF_SIZE;
dev_info->max_rx_queues = (uint16_t)priv->nb_rx_queues;
dev_info->max_tx_queues = (uint16_t)priv->nb_tx_queues;
-
+   dev_info->rx_offload_capa =
+   DEV_RX_OFFLOAD_IPV4_CKSUM |
+   DEV_RX_OFFLOAD_UDP_CKSUM |
+   DEV_RX_OFFLOAD_TCP_CKSUM |
+   DEV_RX_OFFLOAD_OUTER_IPV4_CKSUM;
+   dev_info->tx_offload_capa =
+   DEV_TX_OFFLOAD_IPV4_CKSUM |
+   DEV_TX_OFFLOAD_UDP_CKSUM |
+   DEV_TX_OFFLOAD_TCP_CKSUM |
+   DEV_TX_OFFLOAD_SCTP_CKSUM |
+   DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM;
dev_info->speed_capa = ETH_LINK_SPEED_1G |
ETH_LINK_SPEED_2_5G |
ETH_LINK_SPEED_10G;
@@ -253,8 +263,13 @@
memset(&tx_conf_cfg, 0, sizeof(struct dpni_queue));
memset(&tx_flow_cfg, 0, sizeof(struct dpni_queue));
 
-   tc_id = 0;
-   flow_id = tx_queue_id;
+   if (priv->num_tc == 1) {
+   tc_id = 0;
+   flow_id = tx_queue_id % priv->num_dist_per_tc[tc_id];
+   } else {
+   tc_id = tx_queue_id;
+   flow_id = 0;
+   }
 
ret = dpni_set_queue(dpni, CMD_PRI_LOW, priv->token, DPNI_QUEUE_TX,
 tc_id, flow_id, options, &tx_flow_cfg);
@@ -303,6 +318,7 @@
struct dpaa2_dev_priv *priv = data->dev_private;
struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
struct dpni_queue cfg;
+   struct dpni_error_cfg   err_cfg;
uint16_t qdid;
struct dpni_queue_id qid;
struct dpaa2_queue *dpaa2_q;
@@ -338,6 +354,48 @@
dpaa2_q->fqid = qid.fqid;
}
 
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_RX_L3_CSUM, true);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to set RX l3 csum:Error = %d\n", ret);
+   return ret;
+   }
+
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_RX_L4_CSUM, true);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to get RX l4 csum:Error = %d\n", ret);
+   return ret;
+   }
+
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_TX_L3_CSUM, true);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to set TX l3 csum:Error = %d\n", ret);
+   return ret;
+   }
+
+   ret = dpni_set_offload(dpni, CMD_PRI_LOW, priv->token,
+  DPNI_OFF_TX_L4_CSUM, true);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to get TX l4 csum:Error = %d\n", ret);
+   return ret;
+   }
+
+   /*checksum errors, send them to normal path and set it in annotation */
+   err_cfg.errors = DPNI_ERROR_L3CE | DPNI_ERROR_L4CE;
+
+   err_cfg.error_action = DPNI_ERROR_ACTION_CONTINUE;
+   err_cfg.set_frame_annotation = true;
+
+   ret = dpni_set_errors_behavior(dpni, CMD_PRI_LOW,
+  priv->token, &err_cfg);
+   if (ret) {
+   PMD_INIT_LOG(ERR, "Error to dpni_set_errors_behavior:"
+"code = %d\n", ret);
+   return ret;
+   }
+
return 0;
 }
 
@@ -454,7 +512,13 @@
 */
priv->nb_rx_queues = priv->num_dist_per_tc[DPAA2_DEF_TC];
 
-   priv->nb_tx_queues = attr.num_queues;
+   if (attr.num_tcs == 1)
+   priv->nb_tx_queues = attr.num_queues;
+   else
+   priv->nb_tx_queues = attr.num_tcs;
+
+   PMD_INIT_LOG(DEBUG, "num_tc %d", priv->num_tc);
+   PMD_INIT_LOG(DEBUG, "nb_rx_queues %d", priv->nb_rx_queues);
 
eth_dev->data->nb_rx_queues = priv->nb_rx_queues;
eth_dev->data->nb_tx_queues = priv->nb_tx_queues;
-- 
1.9.1

[dpdk-dev] [PATCHv5 25/33] net/dpaa2: add packet rx and tx support

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h |  54 +++
 drivers/net/dpaa2/Makefile  |   1 +
 drivers/net/dpaa2/dpaa2_ethdev.c|   4 +
 drivers/net/dpaa2/dpaa2_ethdev.h|   3 +
 drivers/net/dpaa2/dpaa2_rxtx.c  | 260 
 5 files changed, 322 insertions(+)
 create mode 100644 drivers/net/dpaa2/dpaa2_rxtx.c

diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index 7c6cc7e..158dfef 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -43,10 +43,16 @@
 #ifndef true
 #define true   1
 #endif
+#define lower_32_bits(x) ((uint32_t)(x))
+#define upper_32_bits(x) ((uint32_t)(((x) >> 16) >> 16))
 
 #ifndef ETH_VLAN_HLEN
 #define ETH_VLAN_HLEN   4 /** < Vlan Header Length */
 #endif
+
+#define MAX_TX_RING_SLOTS  8
+   /** simple.addr_lo = lower_32_bits((uint64_t)(addr));   \
+   fd->simple.addr_hi = upper_32_bits((uint64_t)(addr));   \
+} while (0)
+#define DPAA2_SET_FD_LEN(fd, length)   (fd)->simple.len = length
+#define DPAA2_SET_FD_BPID(fd, bpid)((fd)->simple.bpid_offset |= bpid)
+#define DPAA2_SET_FD_OFFSET(fd, offset)\
+   ((fd->simple.bpid_offset |= (uint32_t)(offset) << 16))
+#define DPAA2_RESET_FD_CTRL(fd)(fd)->simple.ctrl = 0
+
+#defineDPAA2_SET_FD_ASAL(fd, asal) ((fd)->simple.ctrl |= (asal << 
16))
+#define DPAA2_SET_FD_FLC(fd, addr) do { \
+   fd->simple.flc_lo = lower_32_bits((uint64_t)(addr));\
+   fd->simple.flc_hi = upper_32_bits((uint64_t)(addr));\
+} while (0)
+#define DPAA2_GET_FD_ADDR(fd)  \
+((uint64_t)uint64_t)((fd)->simple.addr_hi)) << 32) + (fd)->simple.addr_lo))
+
+#define DPAA2_GET_FD_LEN(fd)   ((fd)->simple.len)
+#define DPAA2_GET_FD_BPID(fd)  (((fd)->simple.bpid_offset & 0x3FFF))
+#define DPAA2_GET_FD_OFFSET(fd)(((fd)->simple.bpid_offset & 
0x0FFF) >> 16)
+#define DPAA2_INLINE_MBUF_FROM_BUF(buf, meta_data_size) \
+   ((struct rte_mbuf *)((uint64_t)(buf) - (meta_data_size)))
+
+#define DPAA2_ASAL_VAL (DPAA2_MBUF_HW_ANNOTATION / 64)
+
+/* Only Enqueue Error responses will be
+ * pushed on FQID_ERR of Enqueue FQ
+ */
+#define DPAA2_EQ_RESP_ERR_FQ   0
+/* All Enqueue responses will be pushed on address
+ * set with qbman_eq_desc_set_response
+ */
+#define DPAA2_EQ_RESP_ALWAYS   1
+
 struct dpaa2_dpbp_dev *dpaa2_alloc_dpbp_dev(void);
 void dpaa2_free_dpbp_dev(struct dpaa2_dpbp_dev *dpbp);
 
diff --git a/drivers/net/dpaa2/Makefile b/drivers/net/dpaa2/Makefile
index ca51402..5e669df 100644
--- a/drivers/net/dpaa2/Makefile
+++ b/drivers/net/dpaa2/Makefile
@@ -59,6 +59,7 @@ EXPORT_MAP := rte_pmd_dpaa2_version.map
 LIBABIVER := 1
 
 SRCS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += base/dpaa2_hw_dpni.c
+SRCS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += dpaa2_rxtx.c
 SRCS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += dpaa2_ethdev.c
 
 # library dependencies
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index f98e7b8..448aac7 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -682,6 +682,8 @@
eth_dev->dev_ops = &dpaa2_ethdev_ops;
eth_dev->data->drv_name = drivername;
 
+   eth_dev->rx_pkt_burst = dpaa2_dev_rx;
+   eth_dev->tx_pkt_burst = dpaa2_dev_tx;
return 0;
 }
 
@@ -735,6 +737,8 @@
free(dpni);
 
eth_dev->dev_ops = NULL;
+   eth_dev->rx_pkt_burst = NULL;
+   eth_dev->tx_pkt_burst = NULL;
 
return 0;
 }
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.h b/drivers/net/dpaa2/dpaa2_ethdev.h
index a56b525..7196398 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.h
+++ b/drivers/net/dpaa2/dpaa2_ethdev.h
@@ -77,4 +77,7 @@ int dpaa2_remove_flow_dist(struct rte_eth_dev *eth_dev,
 
 int dpaa2_attach_bp_list(struct dpaa2_dev_priv *priv, void *blist);
 
+uint16_t dpaa2_dev_rx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts);
+uint16_t dpaa2_dev_tx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts);
+
 #endif /* _DPAA2_ETHDEV_H */
diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c
new file mode 100644
index 000..4b76be5
--- /dev/null
+++ b/drivers/net/dpaa2/dpaa2_rxtx.c
@@ -0,0 +1,260 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2016 Freescale Semiconductor, Inc. All rights reserved.
+ *   Copyright (c) 2016 NXP. All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+

[dpdk-dev] [PATCHv5 24/33] net/dpaa2: add mtu config support

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini  |  1 +
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h |  4 
 drivers/net/dpaa2/dpaa2_ethdev.c| 34 +
 3 files changed, 39 insertions(+)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index b7c274a..a6b7964 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -5,6 +5,7 @@
 ;
 [Features]
 Queue start/stop = Y
+MTU update   = Y
 Promiscuous mode = Y
 RSS hash = Y
 L3 checksum offload  = Y
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index 71361a4..7c6cc7e 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -43,6 +43,10 @@
 #ifndef true
 #define true   1
 #endif
+
+#ifndef ETH_VLAN_HLEN
+#define ETH_VLAN_HLEN   4 /** < Vlan Header Length */
+#endif
 #define DPAA2_DQRR_RING_SIZE   16
/** data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+   uint32_t frame_size = mtu + ETHER_HDR_LEN + ETHER_CRC_LEN;
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (dpni == NULL) {
+   RTE_LOG(ERR, PMD, "dpni is NULL");
+   return -EINVAL;
+   }
+
+   /* check that mtu is within the allowed range */
+   if ((mtu < ETHER_MIN_MTU) || (frame_size > DPAA2_MAX_RX_PKT_LEN))
+   return -EINVAL;
+
+   /* Set the Max Rx frame length as 'mtu' +
+* Maximum Ethernet header length
+*/
+   ret = dpni_set_max_frame_length(dpni, CMD_PRI_LOW, priv->token,
+   mtu + ETH_VLAN_HLEN);
+   if (ret) {
+   PMD_DRV_LOG(ERR, "setting the max frame length failed");
+   return -1;
+   }
+   PMD_DRV_LOG(INFO, "MTU is configured %d for the device\n", mtu);
+   return 0;
+}
+
 static struct eth_dev_ops dpaa2_ethdev_ops = {
.dev_configure= dpaa2_eth_dev_configure,
.dev_start= dpaa2_dev_start,
@@ -485,6 +518,7 @@
.promiscuous_enable   = dpaa2_dev_promiscuous_enable,
.promiscuous_disable  = dpaa2_dev_promiscuous_disable,
.dev_infos_get = dpaa2_dev_info_get,
+   .mtu_set   = dpaa2_dev_mtu_set,
.rx_queue_setup= dpaa2_dev_rx_queue_setup,
.rx_queue_release  = dpaa2_dev_rx_queue_release,
.tx_queue_setup= dpaa2_dev_tx_queue_setup,
-- 
1.9.1

[dpdk-dev] [PATCHv5 23/33] net/dpaa2: add support for promiscuous mode

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini |  1 +
 drivers/net/dpaa2/dpaa2_ethdev.c   | 41 ++
 2 files changed, 42 insertions(+)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index d50c62e..b7c274a 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -5,6 +5,7 @@
 ;
 [Features]
 Queue start/stop = Y
+Promiscuous mode = Y
 RSS hash = Y
 L3 checksum offload  = Y
 L4 checksum offload  = Y
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index b85e31f..76d8e6d 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -438,11 +438,52 @@
}
 }
 
+static void
+dpaa2_dev_promiscuous_enable(
+   struct rte_eth_dev *dev)
+{
+   int ret;
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (dpni == NULL) {
+   RTE_LOG(ERR, PMD, "dpni is NULL");
+   return;
+   }
+
+   ret = dpni_set_unicast_promisc(dpni, CMD_PRI_LOW, priv->token, true);
+   if (ret < 0)
+   RTE_LOG(ERR, PMD, "Unable to enable promiscuous mode %d", ret);
+}
+
+static void
+dpaa2_dev_promiscuous_disable(
+   struct rte_eth_dev *dev)
+{
+   int ret;
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (dpni == NULL) {
+   RTE_LOG(ERR, PMD, "dpni is NULL");
+   return;
+   }
+
+   ret = dpni_set_unicast_promisc(dpni, CMD_PRI_LOW, priv->token, false);
+   if (ret < 0)
+   RTE_LOG(ERR, PMD, "Unable to disable promiscuous mode %d", ret);
+}
 static struct eth_dev_ops dpaa2_ethdev_ops = {
.dev_configure= dpaa2_eth_dev_configure,
.dev_start= dpaa2_dev_start,
.dev_stop = dpaa2_dev_stop,
.dev_close= dpaa2_dev_close,
+   .promiscuous_enable   = dpaa2_dev_promiscuous_enable,
+   .promiscuous_disable  = dpaa2_dev_promiscuous_disable,
.dev_infos_get = dpaa2_dev_info_get,
.rx_queue_setup= dpaa2_dev_rx_queue_setup,
.rx_queue_release  = dpaa2_dev_rx_queue_release,
-- 
1.9.1

[dpdk-dev] [PATCHv5 26/33] net/dpaa2: rx packet parsing and packet type support

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini   |   1 +
 drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h | 257 +++
 drivers/net/dpaa2/dpaa2_ethdev.c |  23 +++
 drivers/net/dpaa2/dpaa2_rxtx.c   |  91 +-
 4 files changed, 371 insertions(+), 1 deletion(-)
 create mode 100644 drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index a6b7964..0746d4b 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -10,6 +10,7 @@ Promiscuous mode = Y
 RSS hash = Y
 L3 checksum offload  = Y
 L4 checksum offload  = Y
+Packet type parsing  = Y
 Linux VFIO   = Y
 ARMv8= Y
 Usage doc= Y
diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h 
b/drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h
new file mode 100644
index 000..9324c6a
--- /dev/null
+++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni_annot.h
@@ -0,0 +1,257 @@
+/*-
+ *   BSD LICENSE
+ *
+ *   Copyright (c) 2016 Freescale Semiconductor, Inc. All rights reserved.
+ *   Copyright (c) 2016 NXP. All rights reserved.
+ *
+ *   Redistribution and use in source and binary forms, with or without
+ *   modification, are permitted provided that the following conditions
+ *   are met:
+ *
+ * * Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * * Redistributions in binary form must reproduce the above copyright
+ *   notice, this list of conditions and the following disclaimer in
+ *   the documentation and/or other materials provided with the
+ *   distribution.
+ * * Neither the name of Freescale Semiconductor, Inc nor the names of its
+ *   contributors may be used to endorse or promote products derived
+ *   from this software without specific prior written permission.
+ *
+ *   THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ *   "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ *   LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ *   A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ *   OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ *   SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ *   LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ *   DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ *   THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ *   (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ *   OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/**
+ * @file
+ *
+ * DPNI packet parse results - implementation internal
+ */
+
+#ifndef _DPAA2_HW_DPNI_ANNOT_H_
+#define _DPAA2_HW_DPNI_ANNOT_H_
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* Annotation valid bits in FD FRC */
+#define DPAA2_FD_FRC_FASV  0x8000
+#define DPAA2_FD_FRC_FAEADV0x4000
+#define DPAA2_FD_FRC_FAPRV 0x2000
+#define DPAA2_FD_FRC_FAIADV0x1000
+#define DPAA2_FD_FRC_FASWOV0x0800
+#define DPAA2_FD_FRC_FAICFDV   0x0400
+
+/* Annotation bits in FD CTRL */
+#define DPAA2_FD_CTRL_ASAL 0x0002  /* ASAL = 128 */
+#define DPAA2_FD_CTRL_PTA  0x0080
+#define DPAA2_FD_CTRL_PTV1 0x0040
+
+/* Frame annotation status */
+struct dpaa2_fas {
+   uint8_t reserved;
+   uint8_t ppid;
+   __le16 ifpid;
+   __le32 status;
+} __packed;
+
+/**
+ * HW Packet Annotation  Register structures
+ */
+struct dpaa2_annot_hdr {
+   /**

[dpdk-dev] [PATCHv5 27/33] net/dpaa2: link status update

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini |   1 +
 drivers/net/dpaa2/dpaa2_ethdev.c   | 107 +
 2 files changed, 108 insertions(+)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index 0746d4b..0660cab 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -4,6 +4,7 @@
 ; Refer to default.ini for the full list of available PMD features.
 ;
 [Features]
+Link status  = Y
 Queue start/stop = Y
 MTU update   = Y
 Promiscuous mode = Y
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index cbcf806..70264b6 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -55,6 +55,58 @@
 /* Name of the DPAA2 Net PMD */
 static const char *drivername = "DPAA2 PMD";
 
+/**
+ * Atomically reads the link status information from global
+ * structure rte_eth_dev.
+ *
+ * @param dev
+ *   - Pointer to the structure rte_eth_dev to read from.
+ *   - Pointer to the buffer to be saved with the link status.
+ *
+ * @return
+ *   - On success, zero.
+ *   - On failure, negative value.
+ */
+static inline int
+dpaa2_dev_atomic_read_link_status(struct rte_eth_dev *dev,
+ struct rte_eth_link *link)
+{
+   struct rte_eth_link *dst = link;
+   struct rte_eth_link *src = &dev->data->dev_link;
+
+   if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
+   *(uint64_t *)src) == 0)
+   return -1;
+
+   return 0;
+}
+
+/**
+ * Atomically writes the link status information into global
+ * structure rte_eth_dev.
+ *
+ * @param dev
+ *   - Pointer to the structure rte_eth_dev to read from.
+ *   - Pointer to the buffer to be saved with the link status.
+ *
+ * @return
+ *   - On success, zero.
+ *   - On failure, negative value.
+ */
+static inline int
+dpaa2_dev_atomic_write_link_status(struct rte_eth_dev *dev,
+  struct rte_eth_link *link)
+{
+   struct rte_eth_link *dst = &dev->data->dev_link;
+   struct rte_eth_link *src = link;
+
+   if (rte_atomic64_cmpset((uint64_t *)dst, *(uint64_t *)dst,
+   *(uint64_t *)src) == 0)
+   return -1;
+
+   return 0;
+}
+
 static void
 dpaa2_dev_info_get(struct rte_eth_dev *dev, struct rte_eth_dev_info *dev_info)
 {
@@ -431,6 +483,7 @@
struct dpaa2_dev_priv *priv = dev->data->dev_private;
struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
int ret;
+   struct rte_eth_link link;
 
PMD_INIT_FUNC_TRACE();
 
@@ -440,6 +493,10 @@
 ret, priv->hw_id);
return;
}
+
+   /* clear the recorded link status */
+   memset(&link, 0, sizeof(link));
+   dpaa2_dev_atomic_write_link_status(dev, &link);
 }
 
 static void
@@ -532,6 +589,55 @@
return 0;
 }
 
+/* return 0 means link status changed, -1 means not changed */
+static int
+dpaa2_dev_link_update(struct rte_eth_dev *dev,
+   int wait_to_complete __rte_unused)
+{
+   int ret;
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+   struct rte_eth_link link, old;
+   struct dpni_link_state state = {0};
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (dpni == NULL) {
+   RTE_LOG(ERR, PMD, "error : dpni is NULL");
+   return 0;
+   }
+   memset(&old, 0, sizeof(old));
+   dpaa2_dev_atomic_read_link_status(dev, &old);
+
+   ret = dpni_get_link_state(dpni, CMD_PRI_LOW, priv->token, &state);
+   if (ret < 0) {
+   RTE_LOG(ERR, PMD, "error: dpni_get_link_state %d", ret);
+   return -1;
+   }
+
+   if ((old.link_status == state.up) && (old.link_speed == state.rate)) {
+   RTE_LOG(DEBUG, PMD, "No change in status\n");
+   return -1;
+   }
+
+   memset(&link, 0, sizeof(struct rte_eth_link));
+   link.link_status = state.up;
+   link.link_speed = state.rate;
+
+   if (state.options & DPNI_LINK_OPT_HALF_DUPLEX)
+   link.link_duplex = ETH_LINK_HALF_DUPLEX;
+   else
+   link.link_duplex = ETH_LINK_FULL_DUPLEX;
+
+   dpaa2_dev_atomic_write_link_status(dev, &link);
+
+   if (link.link_status)
+   PMD_DRV_LOG(INFO, "Port %d Link is Up\n", dev->data->port_id);
+   else
+   PMD_DRV_LOG(INFO, "Port %d Link is Down\n", dev->data->port_id);
+   return 0;
+}
+
 static struct eth_dev_ops dpaa2_ethdev_ops = {
.dev_configure= dpaa2_eth_dev_configure,
.dev_start= dpaa2_dev_start,
@@ -539,6 +645,7 @@
.dev_close= dpaa2_dev_close,
.promiscuous_enable   = dpaa2_dev_promiscuous_enable,
.promiscuous_disable  = dpaa2_dev_promiscuous_disable,
+

[dpdk-dev] [PATCHv5 29/33] net/dpaa2: enable stashing for LS2088A devices

[Hemant Agrawal]

As the hardware determines which core will process which packet,
performance is boosted by direct cache warming/stashing as well
as by providing biasing for core-to-flow affinity, which ensures
that flow-specific data structures can remain in the core’s cache.

This patch enables the one cache line data stashing for packet
annotation data and packet context

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_ethdev.c | 11 +++
 1 file changed, 11 insertions(+)

diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index a9d5f10..2cf395f 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -278,6 +278,17 @@
options = options | DPNI_QUEUE_OPT_USER_CTX;
cfg.user_context = (uint64_t)(dpaa2_q);
 
+   /*if ls2088 or rev2 device, enable the stashing */
+   if ((qbman_get_version() & 0x) > QMAN_REV_4000) {
+   options |= DPNI_QUEUE_OPT_FLC;
+   cfg.flc.stash_control = true;
+   cfg.flc.value &= 0xFFC0;
+   /* 00 00 00 - last 6 bit represent annotation, context stashing,
+* data stashing setting 01 01 00 (0x14) to enable
+* 1 line annotation, 1 line context
+*/
+   cfg.flc.value |= 0x14;
+   }
ret = dpni_set_queue(dpni, CMD_PRI_LOW, priv->token, DPNI_QUEUE_RX,
 dpaa2_q->tc_index, flow_id, options, &cfg);
if (ret) {
-- 
1.9.1

[dpdk-dev] [PATCHv5 30/33] net/dpaa2: add support for non hw buffer pool packet transmit

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/net/dpaa2/dpaa2_rxtx.c | 74 --
 1 file changed, 72 insertions(+), 2 deletions(-)

diff --git a/drivers/net/dpaa2/dpaa2_rxtx.c b/drivers/net/dpaa2/dpaa2_rxtx.c
index 7d73bde..55068e5 100644
--- a/drivers/net/dpaa2/dpaa2_rxtx.c
+++ b/drivers/net/dpaa2/dpaa2_rxtx.c
@@ -191,6 +191,54 @@ static void __attribute__ ((noinline)) __attribute__((hot))
DPAA2_GET_FD_BPID(fd), DPAA2_GET_FD_LEN(fd));
 }
 
+
+static inline int __attribute__((hot))
+eth_copy_mbuf_to_fd(struct rte_mbuf *mbuf,
+   struct qbman_fd *fd, uint16_t bpid)
+{
+   struct rte_mbuf *m;
+   void *mb = NULL;
+
+   if (hw_mbuf_alloc_bulk(bpid_info[bpid].bp_list->buf_pool.mp, &mb, 1)) {
+   PMD_TX_LOG(WARNING, "Unable to allocated DPAA2 buffer");
+   rte_pktmbuf_free(mbuf);
+   return -1;
+   }
+   m = (struct rte_mbuf *)mb;
+   memcpy((char *)m->buf_addr + mbuf->data_off,
+  (void *)((char *)mbuf->buf_addr + mbuf->data_off),
+   mbuf->pkt_len);
+
+   /* Copy required fields */
+   m->data_off = mbuf->data_off;
+   m->ol_flags = mbuf->ol_flags;
+   m->packet_type = mbuf->packet_type;
+   m->tx_offload = mbuf->tx_offload;
+
+   /*Resetting the buffer pool id and offset field*/
+   fd->simple.bpid_offset = 0;
+
+   DPAA2_SET_FD_ADDR(fd, (m->buf_addr));
+   DPAA2_SET_FD_LEN(fd, mbuf->data_len);
+   DPAA2_SET_FD_BPID(fd, bpid);
+   DPAA2_SET_FD_OFFSET(fd, mbuf->data_off);
+   DPAA2_SET_FD_ASAL(fd, DPAA2_ASAL_VAL);
+
+   PMD_TX_LOG(DEBUG, " mbuf %p BMAN buf addr %p",
+  (void *)mbuf, mbuf->buf_addr);
+
+   PMD_TX_LOG(DEBUG, " fdaddr =%lx bpid =%d meta =%d off =%d, len =%d",
+  DPAA2_GET_FD_ADDR(fd),
+   DPAA2_GET_FD_BPID(fd),
+   bpid_info[DPAA2_GET_FD_BPID(fd)].meta_data_size,
+   DPAA2_GET_FD_OFFSET(fd),
+   DPAA2_GET_FD_LEN(fd));
+   /*free the original packet */
+   rte_pktmbuf_free(mbuf);
+
+   return 0;
+}
+
 uint16_t
 dpaa2_dev_rx(void *queue, struct rte_mbuf **bufs, uint16_t nb_pkts)
 {
@@ -331,8 +379,29 @@ static void __attribute__ ((noinline)) __attribute__((hot))
DPAA2_RESET_FD_CTRL((&fd_arr[loop]));
DPAA2_SET_FD_FLC((&fd_arr[loop]), NULL);
mp = (*bufs)->pool;
-   bpid = mempool_to_bpid(mp);
-   eth_mbuf_to_fd(*bufs, &fd_arr[loop], bpid);
+   /* Not a hw_pkt pool allocated frame */
+   if (mp && !(mp->flags & MEMPOOL_F_HW_PKT_POOL)) {
+   PMD_TX_LOG(ERR, "non hw offload bufffer ");
+   /* alloc should be from the default buffer pool
+* attached to this interface
+*/
+   if (priv->bp_list) {
+   bpid = priv->bp_list->buf_pool.bpid;
+   } else {
+   PMD_TX_LOG(ERR, "errr: why no bpool"
+  " attached");
+   num_tx = 0;
+   goto skip_tx;
+   }
+   if (eth_copy_mbuf_to_fd(*bufs,
+   &fd_arr[loop], bpid)) {
+   bufs++;
+   continue;
+   }
+   } else {
+   bpid = mempool_to_bpid(mp);
+   eth_mbuf_to_fd(*bufs, &fd_arr[loop], bpid);
+   }
bufs++;
}
loop = 0;
@@ -345,5 +414,6 @@ static void __attribute__ ((noinline)) __attribute__((hot))
dpaa2_q->tx_pkts += frames_to_send;
nb_pkts -= frames_to_send;
}
+skip_tx:
return num_tx;
 }
-- 
1.9.1

[dpdk-dev] [PATCHv5 31/33] net/dpaa2: enabling the use of physical addresses

[Hemant Agrawal]

DPAA2 - ARM support both physical and virtual addressing.
This patch enables the compile time usages of physical
address instead of virtual address.

The current usages are also set to default as Physical
Address.

Signed-off-by: Hemant Agrawal 
---
 config/common_base|  1 +
 config/defconfig_arm64-dpaa2-linuxapp-gcc |  1 +
 drivers/bus/fslmc/portal/dpaa2_hw_pvt.h   | 66 +++
 drivers/net/dpaa2/base/dpaa2_hw_dpni.c|  4 +-
 drivers/net/dpaa2/dpaa2_rxtx.c| 16 
 drivers/pool/dpaa2/dpaa2_hw_mempool.c | 19 +++--
 6 files changed, 95 insertions(+), 12 deletions(-)

diff --git a/config/common_base b/config/common_base
index 5a10e45..ebd6281 100644
--- a/config/common_base
+++ b/config/common_base
@@ -292,6 +292,7 @@ CONFIG_RTE_LIBRTE_THUNDERX_NICVF_DEBUG_MBOX=n
 #
 CONFIG_RTE_LIBRTE_DPAA2_COMMON=n
 CONFIG_RTE_LIBRTE_DPAA2_POOL=n
+CONFIG_RTE_LIBRTE_DPAA2_USE_PHYS_IOVA=y
 
 #
 # Compile NXP DPAA2 FSL-MC Bus
diff --git a/config/defconfig_arm64-dpaa2-linuxapp-gcc 
b/config/defconfig_arm64-dpaa2-linuxapp-gcc
index 7665912..18c9589 100644
--- a/config/defconfig_arm64-dpaa2-linuxapp-gcc
+++ b/config/defconfig_arm64-dpaa2-linuxapp-gcc
@@ -50,6 +50,7 @@ CONFIG_RTE_PKTMBUF_HEADROOM=256
 CONFIG_RTE_LIBRTE_DPAA2_COMMON=y
 CONFIG_RTE_LIBRTE_DPAA2_POOL=n
 CONFIG_RTE_MBUF_DEFAULT_MEMPOOL_OPS="dpaa2"
+CONFIG_RTE_LIBRTE_DPAA2_USE_PHYS_IOVA=y
 
 #
 # Compile NXP DPAA2 FSL-MC Bus
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
index 158dfef..052a171 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_pvt.h
@@ -175,6 +175,72 @@ struct qbman_fle {
  */
 #define DPAA2_EQ_RESP_ALWAYS   1
 
+#ifdef RTE_LIBRTE_DPAA2_USE_PHYS_IOVA
+static void *dpaa2_mem_ptov(phys_addr_t paddr) __attribute__((unused));
+/* todo - this is costly, need to write a fast coversion routine */
+static void *dpaa2_mem_ptov(phys_addr_t paddr)
+{
+   const struct rte_memseg *memseg = rte_eal_get_physmem_layout();
+   int i;
+
+   for (i = 0; i < RTE_MAX_MEMSEG && memseg[i].addr_64 != 0; i++) {
+   if (paddr >= memseg[i].phys_addr &&
+  (char *)paddr < (char *)memseg[i].phys_addr + memseg[i].len)
+   return (void *)(memseg[i].addr_64
+   + (paddr - memseg[i].phys_addr));
+   }
+   return NULL;
+}
+
+static phys_addr_t dpaa2_mem_vtop(uint64_t vaddr) __attribute__((unused));
+static phys_addr_t dpaa2_mem_vtop(uint64_t vaddr)
+{
+   const struct rte_memseg *memseg = rte_eal_get_physmem_layout();
+   int i;
+
+   for (i = 0; i < RTE_MAX_MEMSEG && memseg[i].addr_64 != 0; i++) {
+   if (vaddr >= memseg[i].addr_64 &&
+   vaddr < memseg[i].addr_64 + memseg[i].len)
+   return memseg[i].phys_addr
+   + (vaddr - memseg[i].addr_64);
+   }
+   return (phys_addr_t)(NULL);
+}
+
+/**
+ * When we are using Physical addresses as IO Virtual Addresses,
+ * Need to call conversion routines dpaa2_mem_vtop & dpaa2_mem_ptov
+ * whereever required.
+ * These routines are called with help of below MACRO's
+ */
+
+#define DPAA2_MBUF_VADDR_TO_IOVA(mbuf) ((mbuf)->buf_physaddr)
+
+/**
+ * macro to convert Virtual address to IOVA
+ */
+#define DPAA2_VADDR_TO_IOVA(_vaddr) dpaa2_mem_vtop((uint64_t)(_vaddr))
+
+/**
+ * macro to convert IOVA to Virtual address
+ */
+#define DPAA2_IOVA_TO_VADDR(_iova) dpaa2_mem_ptov((phys_addr_t)(_iova))
+
+/**
+ * macro to convert modify the memory containing IOVA to Virtual address
+ */
+#define DPAA2_MODIFY_IOVA_TO_VADDR(_mem, _type) \
+   {_mem = (_type)(dpaa2_mem_ptov((phys_addr_t)(_mem))); }
+
+#else  /* RTE_LIBRTE_DPAA2_USE_PHYS_IOVA */
+
+#define DPAA2_MBUF_VADDR_TO_IOVA(mbuf) ((mbuf)->buf_addr)
+#define DPAA2_VADDR_TO_IOVA(_vaddr) (_vaddr)
+#define DPAA2_IOVA_TO_VADDR(_iova) (_iova)
+#define DPAA2_MODIFY_IOVA_TO_VADDR(_mem, _type)
+
+#endif /* RTE_LIBRTE_DPAA2_USE_PHYS_IOVA */
+
 struct dpaa2_dpbp_dev *dpaa2_alloc_dpbp_dev(void);
 void dpaa2_free_dpbp_dev(struct dpaa2_dpbp_dev *dpbp);
 
diff --git a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c 
b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
index 08f53b3..3dc60cc 100644
--- a/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
+++ b/drivers/net/dpaa2/base/dpaa2_hw_dpni.c
@@ -76,7 +76,7 @@
memset(&tc_cfg, 0, sizeof(struct dpni_rx_tc_dist_cfg));
 
dpaa2_distset_to_dpkg_profile_cfg(req_dist_set, &kg_cfg);
-   tc_cfg.key_cfg_iova = (uint64_t)(p_params);
+   tc_cfg.key_cfg_iova = (uint64_t)(DPAA2_VADDR_TO_IOVA(p_params));
tc_cfg.dist_size = eth_dev->data->nb_rx_queues;
tc_cfg.dist_mode = DPNI_DIST_MODE_HASH;
 
@@ -119,7 +119,7 @@ int dpaa2_remove_flow_dist(
memset(p_params, 0, DIST_PARAM_IOVA_SIZE);
memset(&tc_cfg, 0, sizeof(struct dpni_rx_tc_dist_cfg));
 
-   tc_cfg.key_cfg_iova = (uint64_t)(p_params);
+  

[dpdk-dev] [PATCHv5 33/33] drivers/common/dpaa2: frame queue based dq storage alloc

[Hemant Agrawal]

This patch adds generic functions for allowing dq storage
for the frame queues.
As the frame queues are common resource for different drivers
this is helpful.

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.c   | 32 ++
 drivers/bus/fslmc/portal/dpaa2_hw_dpio.h   |  7 ++
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |  2 ++
 drivers/net/dpaa2/dpaa2_ethdev.c   |  8 +++
 4 files changed, 45 insertions(+), 4 deletions(-)

diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
index d7de0d5..55b5ad7 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.c
@@ -407,3 +407,35 @@ static inline struct dpaa2_dpio_dev 
*dpaa2_get_qbman_swp(void)
 
return 0;
 }
+
+void
+dpaa2_free_dq_storage(struct queue_storage_info_t *q_storage)
+{
+   int i = 0;
+
+   for (i = 0; i < NUM_DQS_PER_QUEUE; i++) {
+   if (q_storage->dq_storage[i])
+   rte_free(q_storage->dq_storage[i]);
+   }
+}
+
+int
+dpaa2_alloc_dq_storage(struct queue_storage_info_t *q_storage)
+{
+   int i = 0;
+
+   for (i = 0; i < NUM_DQS_PER_QUEUE; i++) {
+   q_storage->dq_storage[i] = rte_malloc(NULL,
+   DPAA2_DQRR_RING_SIZE * sizeof(struct qbman_result),
+   RTE_CACHE_LINE_SIZE);
+   if (!q_storage->dq_storage[i])
+   goto fail;
+   }
+   return 0;
+fail:
+   i -= 1;
+   while (i >= 0)
+   rte_free(q_storage->dq_storage[i]);
+
+   return -1;
+}
diff --git a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h 
b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h
index b1a1b8f..f2e1168 100644
--- a/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h
+++ b/drivers/bus/fslmc/portal/dpaa2_hw_dpio.h
@@ -59,5 +59,12 @@ struct dpaa2_io_portal_t {
 /* Affine additional DPIO portal to current crypto processing thread */
 int dpaa2_affine_qbman_swp_sec(void);
 
+/* allocate memory for FQ - dq storage */
+int
+dpaa2_alloc_dq_storage(struct queue_storage_info_t *q_storage);
+
+/* free memory for FQ- dq storage */
+void
+dpaa2_free_dq_storage(struct queue_storage_info_t *q_storage);
 
 #endif /* _DPAA2_HW_DPIO_H_ */
diff --git a/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map 
b/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
index 17befc7..bccdc75 100644
--- a/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
+++ b/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
@@ -4,7 +4,9 @@ DPDK_17.02 {
 dpaa2_affine_qbman_swp;
 dpaa2_affine_qbman_swp_sec;
 dpaa2_alloc_dpbp_dev;
+dpaa2_alloc_dq_storage;
 dpaa2_free_dpbp_dev;
+dpaa2_free_dq_storage;
 dpbp_disable;
 dpbp_enable;
 dpbp_get_attributes;
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 183b5b1..695ee61 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -49,6 +49,7 @@
 #include 
 #include 
 #include 
+#include 
 
 #include "dpaa2_ethdev.h"
 
@@ -170,9 +171,8 @@
 
memset(dpaa2_q->q_storage, 0,
   sizeof(struct queue_storage_info_t));
-   dpaa2_q->q_storage->dq_storage[0] = rte_malloc(NULL,
-   DPAA2_DQRR_RING_SIZE * sizeof(struct qbman_result),
-   RTE_CACHE_LINE_SIZE);
+   if (dpaa2_alloc_dq_storage(dpaa2_q->q_storage))
+   goto fail;
}
 
for (i = 0; i < priv->nb_tx_queues; i++) {
@@ -196,7 +196,7 @@
mc_q = priv->rx_vq[0];
while (i >= 0) {
dpaa2_q = (struct dpaa2_queue *)priv->rx_vq[i];
-   rte_free(dpaa2_q->q_storage->dq_storage[0]);
+   dpaa2_free_dq_storage(dpaa2_q->q_storage);
rte_free(dpaa2_q->q_storage);
priv->rx_vq[i--] = NULL;
}
-- 
1.9.1

[dpdk-dev] [PATCHv5 28/33] net/dpaa2: basic stats support

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 doc/guides/nics/features/dpaa2.ini |  1 +
 drivers/net/dpaa2/dpaa2_ethdev.c   | 86 ++
 2 files changed, 87 insertions(+)

diff --git a/doc/guides/nics/features/dpaa2.ini 
b/doc/guides/nics/features/dpaa2.ini
index 0660cab..d43f404 100644
--- a/doc/guides/nics/features/dpaa2.ini
+++ b/doc/guides/nics/features/dpaa2.ini
@@ -12,6 +12,7 @@ RSS hash = Y
 L3 checksum offload  = Y
 L4 checksum offload  = Y
 Packet type parsing  = Y
+Basic stats  = Y
 Linux VFIO   = Y
 ARMv8= Y
 Usage doc= Y
diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c b/drivers/net/dpaa2/dpaa2_ethdev.c
index 70264b6..a9d5f10 100644
--- a/drivers/net/dpaa2/dpaa2_ethdev.c
+++ b/drivers/net/dpaa2/dpaa2_ethdev.c
@@ -589,6 +589,90 @@
return 0;
 }
 
+static
+void dpaa2_dev_stats_get(struct rte_eth_dev *dev,
+struct rte_eth_stats *stats)
+{
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+   int32_t  retcode;
+   uint8_t page0 = 0, page1 = 1, page2 = 2;
+   union dpni_statistics value;
+
+   memset(&value, 0, sizeof(union dpni_statistics));
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (!dpni) {
+   RTE_LOG(ERR, PMD, "dpni is NULL");
+   return;
+   }
+
+   if (!stats) {
+   RTE_LOG(ERR, PMD, "stats is NULL");
+   return;
+   }
+
+   /*Get Counters from page_0*/
+   retcode = dpni_get_statistics(dpni, CMD_PRI_LOW, priv->token,
+ page0, &value);
+   if (retcode)
+   goto err;
+
+   stats->ipackets = value.page_0.ingress_all_frames;
+   stats->ibytes = value.page_0.ingress_all_bytes;
+
+   /*Get Counters from page_1*/
+   retcode = dpni_get_statistics(dpni, CMD_PRI_LOW, priv->token,
+ page1, &value);
+   if (retcode)
+   goto err;
+
+   stats->opackets = value.page_1.egress_all_frames;
+   stats->obytes = value.page_1.egress_all_bytes;
+
+   /*Get Counters from page_2*/
+   retcode = dpni_get_statistics(dpni, CMD_PRI_LOW, priv->token,
+ page2, &value);
+   if (retcode)
+   goto err;
+
+   stats->ierrors = value.page_2.ingress_discarded_frames;
+   stats->oerrors = value.page_2.egress_discarded_frames;
+   stats->imissed = value.page_2.ingress_nobuffer_discards;
+
+   return;
+
+err:
+   RTE_LOG(ERR, PMD, "Operation not completed:Error Code = %d\n", retcode);
+   return;
+};
+
+static
+void dpaa2_dev_stats_reset(struct rte_eth_dev *dev)
+{
+   struct dpaa2_dev_priv *priv = dev->data->dev_private;
+   struct fsl_mc_io *dpni = (struct fsl_mc_io *)priv->hw;
+   int32_t  retcode;
+
+   PMD_INIT_FUNC_TRACE();
+
+   if (dpni == NULL) {
+   RTE_LOG(ERR, PMD, "dpni is NULL");
+   return;
+   }
+
+   retcode =  dpni_reset_statistics(dpni, CMD_PRI_LOW, priv->token);
+   if (retcode)
+   goto error;
+
+   return;
+
+error:
+   RTE_LOG(ERR, PMD, "Operation not completed:Error Code = %d\n", retcode);
+   return;
+};
+
 /* return 0 means link status changed, -1 means not changed */
 static int
 dpaa2_dev_link_update(struct rte_eth_dev *dev,
@@ -646,6 +730,8 @@
.promiscuous_enable   = dpaa2_dev_promiscuous_enable,
.promiscuous_disable  = dpaa2_dev_promiscuous_disable,
.link_update   = dpaa2_dev_link_update,
+   .stats_get = dpaa2_dev_stats_get,
+   .stats_reset   = dpaa2_dev_stats_reset,
.dev_infos_get = dpaa2_dev_info_get,
.dev_supported_ptypes_get = dpaa2_supported_ptypes_get,
.mtu_set   = dpaa2_dev_mtu_set,
-- 
1.9.1

[dpdk-dev] [PATCHv5 32/33] bus/fslmc: add support for dmamap to ARM SMMU

[Hemant Agrawal]

Signed-off-by: Hemant Agrawal 
---
 drivers/bus/fslmc/fslmc_vfio.c | 96 ++
 drivers/bus/fslmc/fslmc_vfio.h |  1 +
 drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |  1 +
 drivers/net/dpaa2/dpaa2_ethdev.c   |  2 +
 4 files changed, 100 insertions(+)

diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index 2bfd7fd..bea32b5 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -76,8 +76,10 @@
 static struct fslmc_vfio_group vfio_groups[VFIO_MAX_GRP];
 static struct fslmc_vfio_container vfio_containers[VFIO_MAX_CONTAINERS];
 static int container_device_fd;
+static uint32_t *msi_intr_vaddr;
 void *(*mcp_ptr_list);
 static uint32_t mcp_id;
+static int is_dma_done;
 
 static int vfio_connect_container(struct fslmc_vfio_group *vfio_group)
 {
@@ -147,6 +149,35 @@ static int vfio_connect_container(struct fslmc_vfio_group 
*vfio_group)
return 0;
 }
 
+static int vfio_map_irq_region(struct fslmc_vfio_group *group)
+{
+   int ret;
+   unsigned long *vaddr = NULL;
+   struct vfio_iommu_type1_dma_map map = {
+   .argsz = sizeof(map),
+   .flags = VFIO_DMA_MAP_FLAG_READ | VFIO_DMA_MAP_FLAG_WRITE,
+   .vaddr = 0x603,
+   .iova = 0x603,
+   .size = 0x1000,
+   };
+
+   vaddr = (unsigned long *)mmap(NULL, 0x1000, PROT_WRITE |
+   PROT_READ, MAP_SHARED, container_device_fd, 0x603);
+   if (vaddr == MAP_FAILED) {
+   FSLMC_VFIO_LOG(ERR, "Unable to map region (errno = %d)", errno);
+   return -errno;
+   }
+
+   msi_intr_vaddr = (uint32_t *)((char *)(vaddr) + 64);
+   map.vaddr = (unsigned long)vaddr;
+   ret = ioctl(group->container->fd, VFIO_IOMMU_MAP_DMA, &map);
+   if (ret == 0)
+   return 0;
+
+   FSLMC_VFIO_LOG(ERR, "VFIO_IOMMU_MAP_DMA fails (errno = %d)", errno);
+   return -errno;
+}
+
 int vfio_dmamap_mem_region(uint64_t vaddr,
   uint64_t iova,
   uint64_t size)
@@ -170,6 +201,71 @@ int vfio_dmamap_mem_region(uint64_t vaddr,
return 0;
 }
 
+int fslmc_vfio_dmamap(void)
+{
+   int ret;
+   struct fslmc_vfio_group *group;
+   struct vfio_iommu_type1_dma_map dma_map = {
+   .argsz = sizeof(struct vfio_iommu_type1_dma_map),
+   .flags = VFIO_DMA_MAP_FLAG_READ | VFIO_DMA_MAP_FLAG_WRITE,
+   };
+
+   int i;
+   const struct rte_memseg *memseg;
+
+   if (is_dma_done)
+   return 0;
+   is_dma_done = 1;
+
+   for (i = 0; i < RTE_MAX_MEMSEG; i++) {
+   memseg = rte_eal_get_physmem_layout();
+   if (memseg == NULL) {
+   FSLMC_VFIO_LOG(ERR, "Cannot get physical layout.");
+   return -ENODEV;
+   }
+
+   if (memseg[i].addr == NULL && memseg[i].len == 0)
+   break;
+
+   dma_map.size = memseg[i].len;
+   dma_map.vaddr = memseg[i].addr_64;
+#ifdef RTE_LIBRTE_DPAA2_USE_PHYS_IOVA
+   dma_map.iova = memseg[i].phys_addr;
+#else
+   dma_map.iova = dma_map.vaddr;
+#endif
+
+   /* SET DMA MAP for IOMMU */
+   group = &vfio_groups[0];
+
+   if (!group->container) {
+   FSLMC_VFIO_LOG(ERR, "Container is not connected ");
+   return -1;
+   }
+
+   FSLMC_VFIO_LOG(DEBUG, "-->Initial SHM Virtual ADDR %llX",
+dma_map.vaddr);
+   FSLMC_VFIO_LOG(DEBUG, "-> DMA size 0x%llX\n", dma_map.size);
+   ret = ioctl(group->container->fd, VFIO_IOMMU_MAP_DMA,
+   &dma_map);
+   if (ret) {
+   FSLMC_VFIO_LOG(ERR, "VFIO_IOMMU_MAP_DMA API"
+  "(errno = %d)", errno);
+   return ret;
+   }
+   FSLMC_VFIO_LOG(DEBUG, "-> dma_map.vaddr = 0x%llX",
+dma_map.vaddr);
+   }
+
+   /* TODO - This is a W.A. as VFIO currently does not add the mapping of
+* the interrupt region to SMMU. This should be removed once the
+* support is added in the Kernel.
+*/
+   vfio_map_irq_region(group);
+
+   return 0;
+}
+
 static int64_t vfio_map_mcp_obj(struct fslmc_vfio_group *group, char *mcp_obj)
 {
int64_t v_addr = (int64_t)MAP_FAILED;
diff --git a/drivers/bus/fslmc/fslmc_vfio.h b/drivers/bus/fslmc/fslmc_vfio.h
index 80c6869..7d562d3 100644
--- a/drivers/bus/fslmc/fslmc_vfio.h
+++ b/drivers/bus/fslmc/fslmc_vfio.h
@@ -70,6 +70,7 @@ int vfio_dmamap_mem_region(
 
 int fslmc_vfio_setup_group(void);
 int fslmc_vfio_process_group(void);
+int fslmc_vfio_dmamap(void);
 
 /* create dpio device */
 int dpaa2_create_dpio_device(struct fslmc_vfio_device *v

Re: [dpdk-dev] [PATCH] mbuf: use pktmbuf helper to create the pool

[Hemant Agrawal]

Please ignore.

Apologies for repeated sent. This patch was  posted earlier.

- Hemant

> -Original Message-
> From: Hemant Agrawal [mailto:hemant.agra...@nxp.com]
> Sent: Thursday, January 19, 2017 6:53 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; bruce.richard...@intel.com; Shreyansh Jain
> ; john.mcnam...@intel.com;
> ferruh.yi...@intel.com; jerin.ja...@caviumnetworks.com; Olivier Matz
> ; Hemant Agrawal 
> Subject: [PATCH] mbuf: use pktmbuf helper to create the pool
> 
> When possible, replace the uses of rte_mempool_create() with the helper
> provided in librte_mbuf: rte_pktmbuf_pool_create().
> 
> This is the preferred way to create a mbuf pool.
> 
> This also updates the documentation.
> 
> Signed-off-by: Olivier Matz 
> Signed-off-by: Hemant Agrawal 
> ---
> This patch is derived from the RFC from Olivier:
> http://dpdk.org/dev/patchwork/patch/15925/
> 
>  app/test/test_link_bonding_rssconf.c   | 11 
>  doc/guides/sample_app_ug/ip_reassembly.rst | 13 +
>  doc/guides/sample_app_ug/ipv4_multicast.rst| 12 
>  doc/guides/sample_app_ug/l2_forward_job_stats.rst  | 33 -
> -
>  .../sample_app_ug/l2_forward_real_virtual.rst  | 26 +++--
>  doc/guides/sample_app_ug/ptpclient.rst | 11 ++--
>  doc/guides/sample_app_ug/quota_watermark.rst   | 26 ++---
>  drivers/net/bonding/rte_eth_bond_8023ad.c  | 13 -
>  examples/ip_pipeline/init.c| 18 ++--
>  examples/ip_reassembly/main.c  | 16 +--
>  examples/multi_process/l2fwd_fork/main.c   | 13 +++--
>  examples/tep_termination/main.c| 16 +--
>  lib/librte_mbuf/rte_mbuf.c |  7 +++--
>  lib/librte_mbuf/rte_mbuf.h | 29 +++
>  14 files changed, 106 insertions(+), 138 deletions(-)
> 
> diff --git a/app/test/test_link_bonding_rssconf.c
> b/app/test/test_link_bonding_rssconf.c
> index 34f1c16..9034f62 100644
> --- a/app/test/test_link_bonding_rssconf.c
> +++ b/app/test/test_link_bonding_rssconf.c
> @@ -67,7 +67,7 @@
>  #define SLAVE_RXTX_QUEUE_FMT  ("rssconf_slave%d_q%d")
> 
>  #define NUM_MBUFS 8191
> -#define MBUF_SIZE (1600 + sizeof(struct rte_mbuf) +
> RTE_PKTMBUF_HEADROOM)
> +#define MBUF_SIZE (1600 + RTE_PKTMBUF_HEADROOM)
>  #define MBUF_CACHE_SIZE 250
>  #define BURST_SIZE 32
> 
> @@ -536,13 +536,12 @@ struct link_bonding_rssconf_unittest_params {
> 
>   if (test_params.mbuf_pool == NULL) {
> 
> - test_params.mbuf_pool =
> rte_mempool_create("RSS_MBUF_POOL", NUM_MBUFS *
> - SLAVE_COUNT, MBUF_SIZE,
> MBUF_CACHE_SIZE,
> - sizeof(struct rte_pktmbuf_pool_private),
> rte_pktmbuf_pool_init,
> - NULL, rte_pktmbuf_init, NULL, rte_socket_id(),
> 0);
> + test_params.mbuf_pool = rte_pktmbuf_pool_create(
> + "RSS_MBUF_POOL", NUM_MBUFS * SLAVE_COUNT,
> + MBUF_CACHE_SIZE, 0, MBUF_SIZE, rte_socket_id());
> 
>   TEST_ASSERT(test_params.mbuf_pool != NULL,
> - "rte_mempool_create failed\n");
> + "rte_pktmbuf_pool_create failed\n");
>   }
> 
>   /* Create / initialize ring eth devs. */ diff --git
> a/doc/guides/sample_app_ug/ip_reassembly.rst
> b/doc/guides/sample_app_ug/ip_reassembly.rst
> index 3c5cc70..d5097c6 100644
> --- a/doc/guides/sample_app_ug/ip_reassembly.rst
> +++ b/doc/guides/sample_app_ug/ip_reassembly.rst
> @@ -223,11 +223,14 @@ each RX queue uses its own mempool.
> 
>  snprintf(buf, sizeof(buf), "mbuf_pool_%u_%u", lcore, queue);
> 
> -if ((rxq->pool = rte_mempool_create(buf, nb_mbuf, MBUF_SIZE, 0,
> sizeof(struct rte_pktmbuf_pool_private), rte_pktmbuf_pool_init, NULL,
> -rte_pktmbuf_init, NULL, socket, MEMPOOL_F_SP_PUT |
> MEMPOOL_F_SC_GET)) == NULL) {
> -
> -RTE_LOG(ERR, IP_RSMBL, "mempool_create(%s) failed", buf);
> -return -1;
> +rxq->pool = rte_pktmbuf_pool_create(buf, nb_mbuf,
> + 0, /* cache size */
> + 0, /* priv size */
> + MBUF_DATA_SIZE, socket);
> +if (rxq->pool == NULL) {
> + RTE_LOG(ERR, IP_RSMBL,
> + "rte_pktmbuf_pool_create(%s) failed", buf);
> + return -1;
>  }
> 
>  Packet Reassembly and Forwarding
> diff --git a/doc/guides/sample_app_ug/ipv4_multicast.rst
> b/doc/guides/sample_app_ug/ipv4_multicast.rst
> index 72da8c4..d9ff249 100644
> --- a/doc/guides/sample_app_ug/ipv4_multicast.rst
> +++ b/doc/guides/sample_app_ug/ipv4_multicast.rst
> @@ -145,12 +145,12 @@ Memory pools for indirect buffers are initialized
> differently from the memory po
> 
>  .. code-block:: c
> 
> -packet_pool = rte_mempool_create("packet_pool", NB_PKT_MBUF,
> PKT_MBUF_SIZE, 32, sizeof(struct rte_pktmbuf_pool_private),
> - 

Re: [dpdk-dev] [PATCH] cryptodev: decouple from PCI device

[Hemant Agrawal]

Please ignore.

Apologies for repeated sent. This patch was  posted earlier.

- Hemant

> -Original Message-
> From: Hemant Agrawal [mailto:hemant.agra...@nxp.com]
> Sent: Thursday, January 19, 2017 6:53 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; bruce.richard...@intel.com; Shreyansh Jain
> ; john.mcnam...@intel.com;
> ferruh.yi...@intel.com; jerin.ja...@caviumnetworks.com; Hemant Agrawal
> 
> Subject: [PATCH] cryptodev: decouple from PCI device
> 
> This makes struct rte_cryptodev independent of struct rte_pci_device by
> replacing it with a pointer to the generic struct rte_device.
> 
> This is inline with the recent changes in ethdev
> 
> Signed-off-by: Hemant Agrawal 
> ---
>  drivers/crypto/qat/qat_qp.c| 12 +---
>  drivers/crypto/qat/rte_qat_cryptodev.c |  6 +++---
>  lib/librte_cryptodev/rte_cryptodev.c   |  6 +++---
>  lib/librte_cryptodev/rte_cryptodev.h   |  4 ++--
>  4 files changed, 17 insertions(+), 11 deletions(-)
> 
> diff --git a/drivers/crypto/qat/qat_qp.c b/drivers/crypto/qat/qat_qp.c index
> 2e7188b..07e63fc 100644
> --- a/drivers/crypto/qat/qat_qp.c
> +++ b/drivers/crypto/qat/qat_qp.c
> @@ -135,6 +135,7 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev
> *dev, uint16_t queue_pair_id,
>   int socket_id)
>  {
>   struct qat_qp *qp;
> + struct rte_pci_device *pci_dev;
>   int ret;
> 
>   PMD_INIT_FUNC_TRACE();
> @@ -153,7 +154,9 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev
> *dev, uint16_t queue_pair_id,
>   return -EINVAL;
>   }
> 
> - if (dev->pci_dev->mem_resource[0].addr == NULL) {
> + pci_dev = RTE_DEV_TO_PCI(dev->device);
> +
> + if (pci_dev->mem_resource[0].addr == NULL) {
>   PMD_DRV_LOG(ERR, "Could not find VF config space "
>   "(UIO driver attached?).");
>   return -EINVAL;
> @@ -174,7 +177,7 @@ int qat_crypto_sym_qp_setup(struct rte_cryptodev
> *dev, uint16_t queue_pair_id,
>   PMD_DRV_LOG(ERR, "Failed to alloc mem for qp struct");
>   return -ENOMEM;
>   }
> - qp->mmap_bar_addr = dev->pci_dev->mem_resource[0].addr;
> + qp->mmap_bar_addr = pci_dev->mem_resource[0].addr;
>   rte_atomic16_init(&qp->inflights16);
> 
>   if (qat_tx_queue_create(dev, &(qp->tx_q), @@ -289,6 +292,7 @@
> static void qat_queue_delete(struct qat_queue *queue)
>   void *io_addr;
>   const struct rte_memzone *qp_mz;
>   uint32_t queue_size_bytes = nb_desc*desc_size;
> + struct rte_pci_device *pci_dev;
> 
>   PMD_INIT_FUNC_TRACE();
>   if (desc_size > ADF_MSG_SIZE_TO_BYTES(ADF_MAX_MSG_SIZE)) { @@ -
> 349,7 +353,9 @@ static void qat_queue_delete(struct qat_queue *queue)
> 
>   queue_base = BUILD_RING_BASE_ADDR(queue->base_phys_addr,
>   queue->queue_size);
> - io_addr = dev->pci_dev->mem_resource[0].addr;
> + pci_dev = RTE_DEV_TO_PCI(dev->device);
> +
> + io_addr = pci_dev->mem_resource[0].addr;
> 
>   WRITE_CSR_RING_BASE(io_addr, queue->hw_bundle_number,
>   queue->hw_queue_number, queue_base); diff --git
> a/drivers/crypto/qat/rte_qat_cryptodev.c
> b/drivers/crypto/qat/rte_qat_cryptodev.c
> index 1e7ee61..840d828 100644
> --- a/drivers/crypto/qat/rte_qat_cryptodev.c
> +++ b/drivers/crypto/qat/rte_qat_cryptodev.c
> @@ -88,9 +88,9 @@
> 
>   PMD_INIT_FUNC_TRACE();
>   PMD_DRV_LOG(DEBUG, "Found crypto device at %02x:%02x.%x",
> - cryptodev->pci_dev->addr.bus,
> - cryptodev->pci_dev->addr.devid,
> - cryptodev->pci_dev->addr.function);
> + RTE_DEV_TO_PCI(cryptodev->device)->addr.bus,
> + RTE_DEV_TO_PCI(cryptodev->device)->addr.devid,
> + RTE_DEV_TO_PCI(cryptodev->device)->addr.function);
> 
>   cryptodev->dev_type = RTE_CRYPTODEV_QAT_SYM_PMD;
>   cryptodev->dev_ops = &crypto_qat_ops;
> diff --git a/lib/librte_cryptodev/rte_cryptodev.c
> b/lib/librte_cryptodev/rte_cryptodev.c
> index 127e8d0..8402b6a 100644
> --- a/lib/librte_cryptodev/rte_cryptodev.c
> +++ b/lib/librte_cryptodev/rte_cryptodev.c
> @@ -446,7 +446,7 @@ struct rte_cryptodev *
>   "device data");
>   }
> 
> - cryptodev->pci_dev = pci_dev;
> + cryptodev->device = &pci_dev->device;
>   cryptodev->driver = cryptodrv;
> 
>   /* init user callbacks */
> @@ -506,7 +506,7 @@ struct rte_cryptodev *
>   if (rte_eal_process_type() == RTE_PROC_PRIMARY)
>   rte_free(cryptodev->data->dev_private);
> 
> - cryptodev->pci_dev = NULL;
> + cryptodev->device = NULL;
>   cryptodev->driver = NULL;
>   cryptodev->data = NULL;
> 
> @@ -867,7 +867,7 @@ struct rte_cryptodev *
>   RTE_FUNC_PTR_OR_RET(*dev->dev_ops->dev_infos_get);
>   (*dev->dev_ops->dev_infos_get)(dev, dev_info);
> 
> - dev_info->pci_dev = dev->pci_dev;
> + dev_info->pci_dev = RTE_DEV_TO_PCI(dev->device);
>   if (dev->driver)
>  

Re: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Guo, Jia]

Hi,john
  The "bind uio_pci_generic`` is not the definitely command, but very 
appreciate for your correcting the other syntax issue. I will involve them in 
v4, Thanks very much.

Best regards,
Jeff Guo


-Original Message-
From: Mcnamara, John 
Sent: Thursday, January 19, 2017 7:17 PM
To: Guo, Jia ; Zhang, Helin ; Wu, 
Jingjing 
Cc: dev@dpdk.org; Guo, Jia ; g...@dpdk.org
Subject: RE: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e



> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Jeff Guo
> Sent: Thursday, January 19, 2017 10:57 AM
> To: Zhang, Helin ; Wu, Jingjing 
> 
> Cc: dev@dpdk.org; Guo, Jia ; g...@dpdk.org
> Subject: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue 
> for i40e
> 
> +
> +The uio_pci_generic module bind failed in X710/XL710/XXV710
> +---
> +
> +**Description**:
> +   The “uio_pci_generic” module is not supported by XL710, since the
> errata of XL710
> +   states that the Interrupt Status bit is not implemented. The 
> + errata is
> the item #71
> +   from the `xl710 controller spec
> +

>  l710 -10-40-controller-spec-update.html>`_.
> +   The hw limitation is the same as other X710/XXV710 NICs.
> +
> +**Implication**:
> +   When bind the “uio_pci_generic” module, it probe device and check 
> +the
> Interrupt Status

Better as: When using ``bind uio_pci_generic``, it probes the device


> +   bit. Since it is not supported by X710/XL710/XXV710, it return a
> *failed* value. The statement
> +   that these products don’t support INTx masking, is indicated in 
> + the
> related `linux kernel commit
> +   < https://git.kernel.org/cgit/linux/kernel/git/stable/linux-
> stable.git/commit/drivers/pci/quirks.c?id=8bcf4525c5d43306c5fd07e132bc
> 8650
> e3491aec>`_.
> +
> +**Resolution/Workaround**:
> +   Do not bind to use the “uio_pci_generic” module in 
> +X710/XL710/XXV710
> NICs.

Better as: Do not bind to ``uio_pci_generic`` in X710/XL710/XXV710 NICs.

Re: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Guo, Jia]

Yes, rerruh, I will recheck the gitconfig and consistent the sign-off name in 
v4. thanks for your suggestion. 

Best regards,
Jeff Guo


-Original Message-
From: Yigit, Ferruh 
Sent: Thursday, January 19, 2017 8:06 PM
To: Guo, Jia ; Zhang, Helin ; Wu, 
Jingjing 
Cc: dev@dpdk.org; g...@dpdk.org
Subject: Re: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

On 1/19/2017 10:56 AM, Jeff Guo wrote:
> From: "Guo, Jia" 
> 
> When bind with the "uio_pci_generic" module in X710/XL710/XXV710, the 
> result is failed. The "uio_pci_generic" module is not supported by 
> X710/XL710/XXV710.
> 
> Signed-off-by: Guo, Jia 

To be consistent on names, this should be "Jeff Guo "
both in sign-off and author field.

If you will send a new version of the patch, please take care of this too, if 
there won't be a new version please confirm it is OK to fix while applying.

Thanks,
ferruh

<...>

[dpdk-dev] [PATCH v4] doc: add known uio_pci_generic issue for i40e

[Jeff Guo]

When bind the "uio_pci_generic" module in X710/XL710/XXV710,
the result is failed. The "uio_pci_generic" module is not
supported by X710/XL710/XXV710.

Signed-off-by: Jeff Guo 
Acked-by: Jingjing Wu 
---
v4: modify the signed-off name to be consistent.
v3: add related linux kernel commit info and fix some syntax issues
v2: add X710/XXV710 limitation case
---
 doc/guides/rel_notes/known_issues.rst | 26 ++
 1 file changed, 26 insertions(+)

diff --git a/doc/guides/rel_notes/known_issues.rst 
b/doc/guides/rel_notes/known_issues.rst
index 3cd4237..c1e06fc 100644
--- a/doc/guides/rel_notes/known_issues.rst
+++ b/doc/guides/rel_notes/known_issues.rst
@@ -640,3 +640,29 @@ I40e VF may not receive packets in the promiscuous mode
 
 **Driver/Module**:
Poll Mode Driver (PMD).
+
+
+The uio_pci_generic module bind failed in X710/XL710/XXV710
+---
+
+**Description**:
+   The “uio_pci_generic” module is not supported by XL710, since the errata of 
XL710
+   states that the Interrupt Status bit is not implemented. The errata is the 
item #71
+   from the `xl710 controller spec
+   
`_.
+   The hw limitation is the same as other X710/XXV710 NICs.
+
+**Implication**:
+   When bind the “uio_pci_generic” module, it probes device and check the 
Interrupt Status
+   bit. Since it is not supported by X710/XL710/XXV710, it return a *failed* 
value. The statement
+   that these products don’t support INTx masking, is indicated in the related 
`linux kernel commit
+   < 
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git/commit/drivers/pci/quirks.c?id=8bcf4525c5d43306c5fd07e132bc8650e3491aec>`_.
+
+**Resolution/Workaround**:
+   Do not bind the “uio_pci_generic” module in X710/XL710/XXV710 NICs.
+
+**Affected Environment/Platform**:
+   All.
+
+**Driver/Module**:
+   Poll Mode Driver (PMD).
-- 
2.7.4

Re: [dpdk-dev] Performance issue in DPDK setup

[Wiles, Keith]

> On Jan 19, 2017, at 1:42 AM, Arnon Warshavsky  wrote:
> 
>> Is there a application available in DPDK where we can specify the sending
>> rate i.e. ask the application to send 1 Gbps traffic etc, so that it would
>> send 1 Gb traffic uniformly in the second.
>> Please guide us on this.
>> 
> 
> Hi
> The DPDK based Pkt-gen is doing a very good job for such uniform packet
> blasting

Thanks Arnon,

Just to be picky and not to be confused with the Lunix kernel pktgen the name 
for the DPDK version is call Pktgen-DPDK no dash in pktgen :-)

> http://dpdk.org/download
> http://pktgen.readthedocs.io/en/latest/
> 
> /Arnon

Regards,
Keith

[dpdk-dev] [PATCH] eal: fix FreeBSD build

[Pablo de Lara]

rte_bus_scan() and rte_bus_probe() have been introduced
in eal.c, but it is missing the rte_bus.h header file,
for BSD systems.

Fixes: f44abbc12fa0 ("bus: add scanning")
Fixes: c3cec1d80708 ("bus: add probing")

Signed-off-by: Pablo de Lara 
---
 lib/librte_eal/bsdapp/eal/eal.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/lib/librte_eal/bsdapp/eal/eal.c b/lib/librte_eal/bsdapp/eal/eal.c
index 534aeea..ee7c9de 100644
--- a/lib/librte_eal/bsdapp/eal/eal.c
+++ b/lib/librte_eal/bsdapp/eal/eal.c
@@ -64,6 +64,7 @@
 #include 
 #include 
 #include 
+#include 
 #include 
 #include 
 #include 
-- 
2.7.4

Re: [dpdk-dev] [PATCH] eal: fix FreeBSD build

[Thomas Monjalon]

2017-01-19 14:21, Pablo de Lara:
> rte_bus_scan() and rte_bus_probe() have been introduced
> in eal.c, but it is missing the rte_bus.h header file,
> for BSD systems.
> 
> Fixes: f44abbc12fa0 ("bus: add scanning")
> Fixes: c3cec1d80708 ("bus: add probing")
> 
> Signed-off-by: Pablo de Lara 

Oh :/ I missed a last test when working on these patches.

Applied, thanks

Re: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Mcnamara, John]

> -Original Message-
> From: Guo, Jia
> Sent: Thursday, January 19, 2017 1:50 PM
> To: Mcnamara, John ; Zhang, Helin
> ; Wu, Jingjing 
> Cc: dev@dpdk.org; g...@dpdk.org
> Subject: RE: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue
> for i40e
> 
> Hi,john
>   The "bind uio_pci_generic`` is not the definitely command, but very
> appreciate for your correcting the other syntax issue. I will involve them
> in v4, Thanks very much.

Hi Jeff, 

Variables, functions, program names, file names and any other literal should 
be in backtick quotes, as outlined in the DPDK Documentation: 

http://dpdk.org/doc/guides/contributing/documentation.html#code-and-literal-block-sections

So uio_pci_generic should be quoted with backticks. If you want to add bind 
then do:
``--bind= uio_pci_generic`` or similar.

Either way it shouldn't be quoted using the "smart" Unicode or non-Unicode 
quotes as in the
patch. All quotes should be straight ASCII double quotes (""). The Html/PDF 
renderers
will convert to the correct quotes as required.



> > related `linux kernel commit
> > +   < https://git.kernel.org/cgit/linux/kernel/git/stable/linux-
> > stable.git/commit/drivers/pci/quirks.c?id=8bcf4525c5d43306c5fd07e132bc
> > 8650
> > e3491aec>`_.

Also, this link generates a doc build error. I think because of the space
between the < and the https:.

John

Re: [dpdk-dev] [PATCH v4 1/6] lib: distributor performance enhancements

[Hunt, David]

On 13/1/2017 3:19 PM, Bruce Richardson wrote:

On Mon, Jan 09, 2017 at 07:50:43AM +, David Hunt wrote:

Now sends bursts of up to 8 mbufs to each worker, and tracks
the in-flight flow-ids (atomic scheduling)

New file with a new api, similar to the old API except with _burst
at the end of the function names

Signed-off-by: David Hunt 
---
  lib/librte_distributor/Makefile|   2 +
  lib/librte_distributor/rte_distributor.c   |  72 +--
  lib/librte_distributor/rte_distributor_burst.c | 558 +
  lib/librte_distributor/rte_distributor_burst.h | 255 ++
  lib/librte_distributor/rte_distributor_priv.h  | 189 +++
  lib/librte_distributor/rte_distributor_version.map |   9 +
  6 files changed, 1014 insertions(+), 71 deletions(-)
  create mode 100644 lib/librte_distributor/rte_distributor_burst.c
  create mode 100644 lib/librte_distributor/rte_distributor_burst.h
  create mode 100644 lib/librte_distributor/rte_distributor_priv.h


Run a documentation sanity check after this patch throws up a few
warnings:

--- /dev/null   2017-01-10 10:26:01.206201474 +
+++ /tmp/doc-check/doc.txt  2017-01-13 15:19:50.717102848 +
@@ -0,0 +1,6 @@
+/home/bruce/dpdk-clean/lib/librte_distributor/rte_distributor_burst.h:187:
warning: argument 'mbuf' of command @param is not found in the argument
list of rte_distributor_return_pkt_burst(struct rte_distributor_burst
*d, unsigned int worker_id, struct rte_mbuf **oldpkt, int num)
+/home/bruce/dpdk-clean/lib/librte_distributor/rte_distributor_burst.h:199:
warning: The following parameters of
rte_distributor_return_pkt_burst(struct rte_distributor_burst *d,
unsigned int worker_id, struct rte_mbuf **oldpkt, int num) are not
documented:
+  parameter 'oldpkt'
+  parameter 'num'
+/home/bruce/dpdk-clean/lib/librte_distributor/rte_distributor_priv.h:73:
warning: Found unknown command `\in_flight_bitmask'
+/home/bruce/dpdk-clean/lib/librte_distributor/rte_distributor_priv.h:73:
warning: Found unknown command `\rte_distributor_process'

Regards,
/Bruce


Will be cleaned up in next revsion.

Thanks,
Dave.

Re: [dpdk-dev] [PATCH v4 2/6] lib: add distributor vector flow matching

[Hunt, David]

On 13/1/2017 3:26 PM, Bruce Richardson wrote:

On Mon, Jan 09, 2017 at 07:50:44AM +, David Hunt wrote:

Signed-off-by: David Hunt 
---
  lib/librte_distributor/Makefile|   4 +
  lib/librte_distributor/rte_distributor_burst.c |  11 +-
  lib/librte_distributor/rte_distributor_match_sse.c | 113 +
  lib/librte_distributor/rte_distributor_priv.h  |   6 ++
  4 files changed, 133 insertions(+), 1 deletion(-)
  create mode 100644 lib/librte_distributor/rte_distributor_match_sse.c

diff --git a/lib/librte_distributor/Makefile b/lib/librte_distributor/Makefile
index 2acc54d..a725aaf 100644
--- a/lib/librte_distributor/Makefile
+++ b/lib/librte_distributor/Makefile
@@ -44,6 +44,10 @@ LIBABIVER := 1
  # all source are stored in SRCS-y
  SRCS-$(CONFIG_RTE_LIBRTE_DISTRIBUTOR) := rte_distributor.c
  SRCS-$(CONFIG_RTE_LIBRTE_DISTRIBUTOR) += rte_distributor_burst.c
+ifeq ($(CONFIG_RTE_ARCH_X86),y)
+SRCS-$(CONFIG_RTE_LIBRTE_DISTRIBUTOR) += rte_distributor_match_sse.c
+endif
+
  

I believe some of the intrinsics used in the vector code are SSE4.2
instructions, so you need to pass that flag for the compilation for e.g.
the "default" target for packaging into distros.


  # install this header file
  SYMLINK-$(CONFIG_RTE_LIBRTE_DISTRIBUTOR)-include := rte_distributor.h
diff --git a/lib/librte_distributor/rte_distributor_burst.c 
b/lib/librte_distributor/rte_distributor_burst.c
index ae7cf9d..35044c4 100644
--- a/lib/librte_distributor/rte_distributor_burst.c
+++ b/lib/librte_distributor/rte_distributor_burst.c
@@ -352,6 +352,9 @@ rte_distributor_process_burst(struct rte_distributor_burst 
*d,
}
  
  		switch (d->dist_match_fn) {

+   case RTE_DIST_MATCH_VECTOR:
+   find_match_vec(d, &flows[0], &matches[0]);
+   break;
default:
find_match_scalar(d, &flows[0], &matches[0]);
}
@@ -538,7 +541,13 @@ rte_distributor_create_burst(const char *name,
snprintf(d->name, sizeof(d->name), "%s", name);
d->num_workers = num_workers;
  
-	d->dist_match_fn = RTE_DIST_MATCH_SCALAR;

+#if defined(RTE_ARCH_X86)
+   if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_SSE2)) {
+   d->dist_match_fn = RTE_DIST_MATCH_VECTOR;
+   } else {
+#endif
+   d->dist_match_fn = RTE_DIST_MATCH_SCALAR;
+   }
  

Two issues here:
1) the check needs to be for SSE4.2, not SSE2 [minimum for DPDK on x86
is SSE3 anyway, so no need for any checks for SSE2]
2) The closing brace should be ifdefed out to fix compilation on non-x86
platforms. A simpler/better solution might actually be to remove the
braces since only a single line is involved in each branch.

Regards,
/Bruce



Will be resolved up in next revision.

Thanks,
Dave.

[dpdk-dev] [PATCH] net/sfc: enable TSO by default

[Andrew Rybchenko]

Signed-off-by: Andrew Rybchenko 
---
It is a mistake that TSO support is compiled out by default.
We would be happy to enable it by deafult, but strictly speaking
it is not a bug fix.

Arguments to enable are:
 - be more feature-rich (and user-friendly) in default config
 - the most of internal testing is done with TSO enabled

 config/common_base | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/config/common_base b/config/common_base
index b9fb8e2..134e9b9 100644
--- a/config/common_base
+++ b/config/common_base
@@ -261,7 +261,7 @@ CONFIG_RTE_LIBRTE_BNXT_PMD=y
 #
 CONFIG_RTE_LIBRTE_SFC_EFX_PMD=y
 CONFIG_RTE_LIBRTE_SFC_EFX_DEBUG=n
-CONFIG_RTE_LIBRTE_SFC_EFX_TSO=n
+CONFIG_RTE_LIBRTE_SFC_EFX_TSO=y
 
 #
 # Compile software PMD backed by SZEDATA2 device
-- 
1.8.2.3

Re: [dpdk-dev] [dpdk-users] IGB_UIO: PCI Resources Management

[Ferruh Yigit]

On 1/13/2017 5:33 AM, Tan, Jianfeng wrote:
> 
> 
>> -Original Message-
>> From: Yigit, Ferruh
>> Sent: Friday, January 13, 2017 10:05 AM
>> To: Tan, Jianfeng; Alejandro Lucero
>> Cc: Gregory Etelson; dev; us...@dpdk.org
>> Subject: Re: [dpdk-users] [dpdk-dev] IGB_UIO: PCI Resources Management
>>
>> On 1/13/2017 1:51 AM, Tan, Jianfeng wrote:
>>>
>>>
 -Original Message-
 From: users [mailto:users-boun...@dpdk.org] On Behalf Of Ferruh Yigit
 Sent: Thursday, January 12, 2017 8:22 PM
 To: Alejandro Lucero
 Cc: Gregory Etelson; dev; us...@dpdk.org
 Subject: Re: [dpdk-users] [dpdk-dev] IGB_UIO: PCI Resources
>> Management

 On 1/12/2017 12:12 PM, Alejandro Lucero wrote:
>
>
> On Thu, Jan 12, 2017 at 11:55 AM, Ferruh Yigit  > wrote:
>
> On 12/9/2016 8:54 AM, Gregory Etelson wrote:
> > Hello,
> >
> > IGB_UIO driver does not close port PCI activities after DPDK process
 exits.
> > DPDK API provides rte_eth_dev_close() to manage port PCI,
> > but it can be skipped if process receives SIGKILL signal
>
> I guess I understand the problem.
>
>
> This is a known problem, but it is not just a UIO problem, and this
> patch does not solve it, maybe it just solves part of it.
>
> In fact, a DPDK program crashing could imply the NIC DMAing after that
> and after that memory was assigned to another program.

 Yes.
 Can there be a way to stop NIC DMA, (or prevent it access to mem
 anymore) when app crashes?
 I think that is what this patch is looking for.
>>>
>>> If I understand it correctly, you are looking for this patch?
>>> http://dpdk.org/dev/patchwork/patch/17495/
>>>
>>
>> That is good, thanks Jianfeng, I will check it.
>>
>> btw, patch's current state is rejected, which is by mistake, it seems I
>> confused it with "iomem and ioport mapping" patch, sorry about it, I
>> will update its status immediately.
> 
> No problem at all. This patch is rejected as it's based on "iomem and ioport 
> mapping" patch. As "iomem and ioport mapping" patch has backward 
> compatibility issue, we need to figure out a way to resubmit this patch 
> without changing the original "iomem and ioport mapping" in igb_uio.

I thinks implementing uio_info->release and uio_info.open is good idea,
but I have a few questions:

1- What is the the dependency to "iomem and ioport mapping" patch?

2- If we keep pci_enable_device() in probe() can this prevent moving
registering/freeing interrupts in open()/release()

3- And is pci_disable_device() done in release is enough to stop NIC DMA
to access memory?


I did a simple test, implemented simple uio_info->release and
uio_info.open, which only does pci_disable_device() and
pci_enable_device(),
but this prevent app receiving packets in its second run, independent
from app terminated gracefully or not. Any idea why this is not working?


btw, I can produce the problematic case, as George Prekas described in:
http://dpdk.org/ml/archives/users/2016-September/001026.html

CC'ed George, since he also seems interested in issue.

> 
> Thanks,
> Jianfeng
> 

Re: [dpdk-dev] [dpdk-users] IGB_UIO: PCI Resources Management

[Ferruh Yigit]

On 1/13/2017 11:10 AM, Alejandro Lucero wrote:
> I completely misread the patch, and I wrongly thought that code was
> linked to module removal, but I see this is not about that, but about
> releasing the /dev/uio file calling release function, what is done by
> the kernel when the process exits.
> 
> So yes, the patch avoids the problem I talked about.
> 
> However, calling that specific ixgbe driver function will break other
> devices relying on igb_uio. What about implementing a notifier chain for
> this? The igb_uio code would be agnostic and each interested driver,
> like ixgbe or nfp_net, could execute the specific port close code when
> the notifier chain triggers.

Can you please elaborate, how can this be done?
This is kernel code, and interested drivers are in userspace. This patch
benefit from Linux kernel driver of same device.

> 
> 
> On Fri, Jan 13, 2017 at 5:33 AM, Tan, Jianfeng  > wrote:
> 
> 
> 
> > -Original Message-
> > From: Yigit, Ferruh
> > Sent: Friday, January 13, 2017 10:05 AM
> > To: Tan, Jianfeng; Alejandro Lucero
> > Cc: Gregory Etelson; dev; us...@dpdk.org 
> > Subject: Re: [dpdk-users] [dpdk-dev] IGB_UIO: PCI Resources Management
> >
> > On 1/13/2017 1:51 AM, Tan, Jianfeng wrote:
> > >
> > >
> > >> -Original Message-
> > >> From: users [mailto:users-boun...@dpdk.org
> ] On Behalf Of Ferruh Yigit
> > >> Sent: Thursday, January 12, 2017 8:22 PM
> > >> To: Alejandro Lucero
> > >> Cc: Gregory Etelson; dev; us...@dpdk.org 
> > >> Subject: Re: [dpdk-users] [dpdk-dev] IGB_UIO: PCI Resources
> > Management
> > >>
> > >> On 1/12/2017 12:12 PM, Alejandro Lucero wrote:
> > >>>
> > >>>
> > >>> On Thu, Jan 12, 2017 at 11:55 AM, Ferruh Yigit
> mailto:ferruh.yi...@intel.com>
> > >>>  >> wrote:
> > >>>
> > >>> On 12/9/2016 8:54 AM, Gregory Etelson wrote:
> > >>> > Hello,
> > >>> >
> > >>> > IGB_UIO driver does not close port PCI activities after
> DPDK process
> > >> exits.
> > >>> > DPDK API provides rte_eth_dev_close() to manage port PCI,
> > >>> > but it can be skipped if process receives SIGKILL signal
> > >>>
> > >>> I guess I understand the problem.
> > >>>
> > >>>
> > >>> This is a known problem, but it is not just a UIO problem, and
> this
> > >>> patch does not solve it, maybe it just solves part of it.
> > >>>
> > >>> In fact, a DPDK program crashing could imply the NIC DMAing
> after that
> > >>> and after that memory was assigned to another program.
> > >>
> > >> Yes.
> > >> Can there be a way to stop NIC DMA, (or prevent it access to mem
> > >> anymore) when app crashes?
> > >> I think that is what this patch is looking for.
> > >
> > > If I understand it correctly, you are looking for this patch?
> > > http://dpdk.org/dev/patchwork/patch/17495/
> 
> > >
> >
> > That is good, thanks Jianfeng, I will check it.
> >
> > btw, patch's current state is rejected, which is by mistake, it
> seems I
> > confused it with "iomem and ioport mapping" patch, sorry about it, I
> > will update its status immediately.
> 
> No problem at all. This patch is rejected as it's based on "iomem
> and ioport mapping" patch. As "iomem and ioport mapping" patch has
> backward compatibility issue, we need to figure out a way to
> resubmit this patch without changing the original "iomem and ioport
> mapping" in igb_uio.
> 
> Thanks,
> Jianfeng
> 
> 

Re: [dpdk-dev] [PATCHv5 02/33] doc: add dpaa2 nic details

[Thomas Monjalon]

2017-01-19 18:53, Hemant Agrawal:
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> +NXP DPAA2 PMD
> +M: Hemant Agrawal 
> +F: drivers/bus/fslmc/
> +F: drivers/common/dpaa2/
> +F: drivers/net/dpaa2/
> +F: drivers/pool/dpaa2/
> +F: doc/guides/nics/dpaa2.rst

Please introduce line by line in the patches
adding these files.

> +++ b/doc/guides/nics/features/dpaa2.ini
> @@ -0,0 +1,8 @@
> +;
> +; Supported features of the 'dpaa2' network poll mode driver.
> +;
> +; Refer to default.ini for the full list of available PMD features.
> +;
> +[Features]
> +ARMv8= Y
> +Usage doc= Y

You should introduce this file with the first PMD patch.

> +* **Added a new driver for NXP DPAA2 - FSLMC bus.**
> +
> +  Added the new bus "fslmc" driver for NXP DPAA2 devices. See the
> +  "Network Interface Controller Drivers" document for more details on this 
> new
> +  driver.
> +
> +* **Added a new driver for NXP DPAA2 Network PMD.**
> +
> +  Added the new "dpaa2" net driver for NXP DPAA2 devices. See the
> +  "Network Interface Controller Drivers" document for more details on this 
> new
> +  driver.

You can use sphinx references to create a link in the generated HTML.

Please move doc patch after (or within) the code.

Re: [dpdk-dev] [PATCHv5 04/33] bus/fslmc: introducing fsl-mc bus driver

[Thomas Monjalon]

2017-01-19 18:53, Hemant Agrawal:
> The fslmc bus driver is a rte_bus driver which scans the fsl-mc bus
> for NXP DPAA2 SoCs.
> 
> Signed-off-by: Hemant Agrawal 
> ---
>  config/common_base |   6 +
>  config/defconfig_arm64-dpaa2-linuxapp-gcc  |   5 +
>  drivers/Makefile   |   1 +
>  drivers/bus/Makefile   |  36 ++
>  drivers/bus/fslmc/Makefile |  52 +
>  drivers/bus/fslmc/fslmc_bus.c  | 125 +
>  drivers/bus/fslmc/rte_fslmc.h  | 148 
> +
>  drivers/bus/fslmc/rte_pmd_fslmcbus_version.map |   7 ++

Good to see a first bus driver.

Re: [dpdk-dev] [PATCHv5 06/33] bus/fslmc: add mc dpni object support

[Thomas Monjalon]

2017-01-19 18:53, Hemant Agrawal:
> This patch add support for dpni object support in MC
> driver.
> 
> DPNI represent a network interface object in DPAA2.

I really think you need to provide a design doc for this bus,
event if it a collection of links to some web resources.

Re: [dpdk-dev] [PATCHv5 10/33] eal/vfio: adding vfio utility functions in map file

[Thomas Monjalon]

2017-01-19 18:53, Hemant Agrawal:
> --- a/lib/librte_eal/bsdapp/eal/rte_eal_version.map
> +++ b/lib/librte_eal/bsdapp/eal/rte_eal_version.map
> @@ -183,5 +183,8 @@ DPDK_17.02 {
> rte_bus_register;
> rte_bus_scan;
> rte_bus_unregister;
> +vfio_get_container_fd;
> +vfio_get_group_fd;
> +vfio_get_group_no;

Indentation looks wrong here.

You cannot export some symbols without adding them a rte_ prefix.

Re: [dpdk-dev] [PATCHv5 11/33] bus/fslmc: add vfio support

[Thomas Monjalon]

2017-01-19 18:53, Hemant Agrawal:
> @@ -45,8 +45,10 @@ DPDK_17.02 {
>  dpseci_open;
>  dpseci_reset;
>  dpseci_set_rx_queue;
> +mcp_ptr_list;
>  rte_fslmc_driver_register;
>  rte_fslmc_driver_unregister;
> +vfio_dmamap_mem_region;

These symbols do not comply with the DPDK namespace (rte_*).

Re: [dpdk-dev] [PATCHv5 02/33] doc: add dpaa2 nic details

[Mcnamara, John]

> -Original Message-
> From: Hemant Agrawal [mailto:hemant.agra...@nxp.com]
> Sent: Thursday, January 19, 2017 1:23 PM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Richardson, Bruce
> ; shreyansh.j...@nxp.com; Mcnamara, John
> ; Yigit, Ferruh ;
> jerin.ja...@caviumnetworks.com; Hemant Agrawal 
> Subject: [PATCHv5 02/33] doc: add dpaa2 nic details
> 
>...
> 
> +* **Added a new driver for NXP DPAA2 - FSLMC bus.**
> +
> +  Added the new bus "fslmc" driver for NXP DPAA2 devices. See the
> + "Network Interface Controller Drivers" document for more details on
> + this new  driver.
> +
> +* **Added a new driver for NXP DPAA2 Network PMD.**
> +
> +  Added the new "dpaa2" net driver for NXP DPAA2 devices. See the
> + "Network Interface Controller Drivers" document for more details on
> + this new  driver.
> 
>  New Features
>  

These additions should be after "New Features" and not before, if you do
another revision you can fix it. If not it could be fixed on commit or I
can fix it in the final Release note review. 

But overall the NIC guide is very good.

Acked-by: John McNamara 

Re: [dpdk-dev] [PATCH] doc: update release notes for I/O device memory access API

[Mcnamara, John]

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Jerin Jacob
> Sent: Thursday, January 19, 2017 5:43 AM
> To: dev@dpdk.org
> Cc: thomas.monja...@6wind.com; Jerin Jacob
> 
> Subject: [dpdk-dev] [PATCH] doc: update release notes for I/O device
> memory access API
> 
> Signed-off-by: Jerin Jacob 


Acked-by: John McNamara 

Re: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter

[Mcnamara, John]

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Yong Liu
> Sent: Thursday, January 19, 2017 5:34 AM
> To: dev@dpdk.org
> Cc: Liu, Yong 
> Subject: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter
> 
> Signed-off-by: Yong Liu 

Acked-by: John McNamara 

Re: [dpdk-dev] [PATCH v11 13/13] doc: remove deprecation notice for rte_bus

[Mcnamara, John]

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Thomas Monjalon
> Sent: Thursday, January 19, 2017 4:45 AM
> To: Shreyansh Jain 
> Cc: dev@dpdk.org
> Subject: [dpdk-dev] [PATCH v11 13/13] doc: remove deprecation notice for
> rte_bus
> 
> From: Shreyansh Jain 
> 
> Signed-off-by: Shreyansh Jain 

Acked-by: John McNamara 

Re: [dpdk-dev] [PATCH v4] ethdev: fix port data mismatched in multiple process model

[Ferruh Yigit]

On 1/9/2017 7:50 AM, Yuanhan Liu wrote:
> Assume we have two virtio ports, 00:03.0 and 00:04.0. The first one is
> managed by the kernel driver, while the later one is managed by DPDK.
> 
> Now we start the primary process. 00:03.0 will be skipped by DPDK virtio
> PMD driver (since it's being used by the kernel). 00:04.0 would be
> successfully initiated by DPDK virtio PMD (if nothing abnormal happens).
> After that, we would get a port id 0, and all the related info needed
> by virtio (virtio_hw) is stored at rte_eth_dev_data[0].
> 
> Then we start the secondary process. As usual, 00:03.0 will be firstly
> probed. It firstly tries to get a local eth_dev structure for it (by
> rte_eth_dev_allocate):
> 
> port_id = rte_eth_dev_find_free_port();
> ...
> 
> eth_dev = &rte_eth_devices[port_id];
> eth_dev->data = &rte_eth_dev_data[port_id];
> ...
> 
> return eth_dev;
> 
> Since it's a first PCI device, port_id will be 0. eth_dev->data would
> then point to rte_eth_dev_data[0]. And here things start going wrong,
> as rte_eth_dev_data[0] actually stores the virtio_hw for 00:04.0.
> 
> That said, in the secondary process, DPDK will continue to drive PCI
> device 00.03.0 (despite the fact it's been managed by kernel), with
> the info from PCI device 00:04.0. Which is wrong.
> 
> The fix is to attach the port already registered by the primary process:
> iterate the rte_eth_dev_data[], and get the port id who's PCI ID matches
> the current PCI device.
> 
> This would let us maintain same port ID for the same PCI device, keeping
> the chance of referencing to wrong data minimal.
> 
> Fixes: af75078fece3 ("first public release")
> 
> Cc: sta...@dpdk.org
> Cc: Thomas Monjalon 
> Cc: Bruce Richardson 
> Cc: Ferruh Yigit 
> Signed-off-by: Yuanhan Liu 
> ---
> 
> v4: - assign eth_dev in the common eth_dev init help function
>   it also renamed to eth_dev_get, to not confuse with the
>   eth_dev_init callback.
> - move primoary process specific assignments to rte_eth_dev_allocate
> - drop the virtio example in comment
> - combine two code block for primary into one
> 
> v3: - do not move rte_eth_dev_data_alloc to pci_probe
> - rename eth_dev_attach to eth_dev_attach_secondary
> - introduce eth_dev_init() for common eth_dev struct initiation
> - move comment block inside the "if" block
> ---
>  lib/librte_ether/rte_ethdev.c | 71 
> +--
>  1 file changed, 62 insertions(+), 9 deletions(-)
> 
> diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
> index fde8112..1453df1 100644
> --- a/lib/librte_ether/rte_ethdev.c
> +++ b/lib/librte_ether/rte_ethdev.c
> @@ -189,6 +189,19 @@ struct rte_eth_dev *
>   return RTE_MAX_ETHPORTS;
>  }
>  
> +static struct rte_eth_dev *
> +eth_dev_get(uint8_t port_id)
> +{
> + struct rte_eth_dev *eth_dev = &rte_eth_devices[port_id];
> +
> + eth_dev->data = &rte_eth_dev_data[port_id];
> + eth_dev->attached = DEV_ATTACHED;
> + eth_dev_last_created_port = port_id;
> + nb_ports++;
> +
> + return eth_dev;
> +}
> +
>  struct rte_eth_dev *
>  rte_eth_dev_allocate(const char *name)
>  {
> @@ -210,13 +223,41 @@ struct rte_eth_dev *
>   return NULL;
>   }
>  
> - eth_dev = &rte_eth_devices[port_id];
> - eth_dev->data = &rte_eth_dev_data[port_id];
> + eth_dev = eth_dev_get(port_id);

There can be a merge issue here, please help me understand.

In repo, different from seen here, this patch does this here:
  -   eth_dev = &rte_eth_devices[port_id];
  -   eth_dev->data = &rte_eth_dev_data[port_id];
  -   memset(eth_dev->data, 0, sizeof(*eth_dev->data));
  +   memset(&rte_eth_devices[port_id], 0, sizeof(*eth_dev->data));
  +   eth_dev = eth_dev_get(port_id);

Which no more resets the eth_dev->data, but rte_eth_devices[port_id]
(with sizeof(*eth_dev->data))

memset(eth_dev->data) added by Jan Blunck on comment:
7f95f78a8aea ("ethdev: clear data when allocating device")

most probably it should stay as "memset(eth_dev->data)", but if not,
please aware that commit 7f95f78a8aea removed some assignment from
drivers relying this memset, they needs to be added back.

>   snprintf(eth_dev->data->name, sizeof(eth_dev->data->name), "%s", name);
>   eth_dev->data->port_id = port_id;
> - eth_dev->attached = DEV_ATTACHED;
> - eth_dev_last_created_port = port_id;
> - nb_ports++;
> +
> + return eth_dev;
> +}
> +
> +/*
> + * Attach to a port already registered by the primary process, which
> + * makes sure that the same device would have the same port id both
> + * in the primary and secondary process.
> + */
> +static struct rte_eth_dev *
> +eth_dev_attach_secondary(const char *name)
> +{
> + uint8_t i;
> + struct rte_eth_dev *eth_dev;
> +
> + if (rte_eth_dev_data == NULL)
> + rte_eth_dev_data_alloc();
> +
> + for (i = 0; i < RTE_MAX_ETHPORTS; i++) {
> +

Re: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter

[Adrien Mazarguil]

On Thu, Jan 19, 2017 at 10:06:34AM +0100, Thomas Monjalon wrote:
> 2017-01-19 13:34, Yong Liu:
> > +* ABI changes are planned for 17.05: structure 
> > ``rte_eth_tunnel_filter_conf``
> > +  will be extended with a new member ``vf_id`` in order to enable cloud 
> > filter
> > +  on VF device.
> 
> I think we should stop rely on this API, and migrate to rte_flow instead.
> Adrien any thought?

I'm all for using rte_flow in any case. I've already documented an approach
to convert TUNNEL filter rules to rte_flow rules [1], although it may be
incomplete due to my limited experience with this filter type. We already
know several tunnel item types must be added (currently only VXLAN is
defined).

I understand ixgbe/i40e currently map rte_flow on top of the legacy
framework, therefore extending this structure might still be needed in the
meantime. Not sure we should prevent this change as long as such rules can
be configured through rte_flow as well.

[1] 
http://dpdk.org/doc/guides/prog_guide/rte_flow.html#tunnel-to-eth-ipv4-ipv6-vxlan-or-other-queue

-- 
Adrien Mazarguil
6WIND

[dpdk-dev] [PATCH] lib/librte_ether: error handling on MAC address replay

[Steve Shin]

This patch fixes a bug in replaying MAC address to the hardware
in rte_eth_dev_config_restore() routine.

Signed-off-by: Steve Shin 
---
 lib/librte_ether/rte_ethdev.c | 10 ++
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 4790faf..7e01f10 100644
--- a/lib/librte_ether/rte_ethdev.c
+++ b/lib/librte_ether/rte_ethdev.c
@@ -951,10 +951,12 @@ rte_eth_dev_config_restore(uint8_t port_id)
continue;
 
/* add address to the hardware */
-   if  (*dev->dev_ops->mac_addr_add &&
-   (dev->data->mac_pool_sel[i] & (1ULL << pool)))
-   (*dev->dev_ops->mac_addr_add)(dev, &addr, i, pool);
-   else {
+   if  (*dev->dev_ops->mac_addr_add) {
+   if (dev->data->mac_pool_sel[i] & (1ULL << pool))
+   (*dev->dev_ops->mac_addr_add)(dev, &addr, i, 
pool);
+   else
+   continue;
+   } else {
RTE_PMD_DEBUG_TRACE("port %d: MAC address array not 
supported\n",
port_id);
/* exit the loop but not return an error */
-- 
2.9.3

Re: [dpdk-dev] [PATCHv5 03/33] drivers/common/dpaa2: adding qbman driver

[Ferruh Yigit]

On 1/19/2017 1:23 PM, Hemant Agrawal wrote:
> QBMAN, is a hardware block which interfaces with the other
> accelerating hardware blocks (For e.g., WRIOP) on NXP's DPAA2
> SoC for queue, buffer and packet scheduling.
> 
> This patch introduces a userspace driver for interfacing with
> the QBMAN hw block.
> 
> The qbman-portal component provides APIs to do the low level
> hardware bit twiddling for operations such as:
>   -initializing Qman software portals
>   -building and sending portal commands
>   -portal interrupt configuration and processing
> 
> This same/similar code is used in kernel and compat file is used
> to make it working in user space.
> 
> Signed-off-by: Geoff Thorpe 
> Signed-off-by: Roy Pledge 
> Signed-off-by: Hemant Agrawal 
> ---

<...>

> diff --git a/drivers/common/dpaa2/qbman/include/compat.h 
> b/drivers/common/dpaa2/qbman/include/compat.h
> new file mode 100644
> index 000..d321cc6
> --- /dev/null
> +++ b/drivers/common/dpaa2/qbman/include/compat.h
> @@ -0,0 +1,403 @@
> +/* Copyright (c) 2008-2016 Freescale Semiconductor, Inc.
> + * All rights reserved.

Should this say "#   BSD LICENSE" ? As in other file comments..
There are multiple this kind of usage.

> + *
> + * Redistribution and use in source and binary forms, with or without
> + * modification, are permitted provided that the following conditions are 
> met:
> + * * Redistributions of source code must retain the above copyright
> + *notice, this list of conditions and the following disclaimer.
> + * * Redistributions in binary form must reproduce the above copyright
> + *notice, this list of conditions and the following disclaimer in the
> + *documentation and/or other materials provided with the distribution.
> + * * Neither the name of Freescale Semiconductor nor the
> + *names of its contributors may be used to endorse or promote products
> + *derived from this software without specific prior written permission.
> + *

<...>

Re: [dpdk-dev] [PATCHv5 04/33] bus/fslmc: introducing fsl-mc bus driver

[Ferruh Yigit]

On 1/19/2017 1:23 PM, Hemant Agrawal wrote:
> The fslmc bus driver is a rte_bus driver which scans the fsl-mc bus
> for NXP DPAA2 SoCs.
> 
> Signed-off-by: Hemant Agrawal 
> ---

<...>

> +#
> +# library name
> +#
> +LIB = librte_pmd_fslmcbus.a

Since now there is a bus folder/driver, what do you think nameming
library with librte_bus_ prefix, like: librte_bus_fslmc.a

<...>

> +
> +static int
> +rte_fslmc_probe(void)
> +{
> + int ret = -1;

If any bus->probe() fails, rte_bus_probe() breaks and returns error,
which cause app to exit.
Here if there is no device or driver in the bus, function is returning
error, I guess it should be returning zero for this case.

> + struct rte_dpaa2_device *dev;
> + struct rte_dpaa2_driver *drv;
> +
> + TAILQ_FOREACH(dev, &rte_fslmc_bus.device_list, next) {
> + TAILQ_FOREACH(drv, &rte_fslmc_bus.driver_list, next) {
> + ret = rte_fslmc_match(drv, dev);
> + if (ret)
> + continue;
> +
> + if (!drv->probe)
> + continue;
> +
> + ret = drv->probe(drv, dev);
> + if (ret)
> + FSLMC_BUS_LOG(ERR, "Unable to probe.\n");
> + break;
> + }
> + }
> + return ret;
> +}

<...>

Re: [dpdk-dev] [PATCHv5 05/33] bus/fslmc: introduce mc object functions

[Ferruh Yigit]

On 1/19/2017 1:23 PM, Hemant Agrawal wrote:
> This patch intoduces the DPAA2 MC(Management complex Driver).
> 
> This is a minimal set of low level functions to send and
> receive commands to the fsl-mc. It includes support for basic
> management commands and commands to manipulate MC objects.
> 
> This is common to be used by various DPAA2 PMDs. e.g.net, crypto
> and other drivers.
> 
> This is a low level library also used in kernel.
> 
> Signed-off-by: Cristian Sovaiala 
> Signed-off-by: Hemant Agrawal 

<...>

> diff --git a/drivers/bus/fslmc/mc/fsl_mc_cmd.h 
> b/drivers/bus/fslmc/mc/fsl_mc_cmd.h
> new file mode 100644
> index 000..cbd3995
> --- /dev/null
> +++ b/drivers/bus/fslmc/mc/fsl_mc_cmd.h
> @@ -0,0 +1,231 @@
> +/* Copyright 2013-2016 Freescale Semiconductor Inc.
> + * Copyright (c) 2016 NXP.
> + *
> + * Redistribution and use in source and binary forms, with or without
> + * modification, are permitted provided that the following conditions are 
> met:
> + * * Redistributions of source code must retain the above copyright
> + * notice, this list of conditions and the following disclaimer.
> + * * Redistributions in binary form must reproduce the above copyright
> + * notice, this list of conditions and the following disclaimer in the
> + * documentation and/or other materials provided with the distribution.
> + * * Neither the name of the above-listed copyright holders nor the
> + * names of any contributors may be used to endorse or promote products
> + * derived from this software without specific prior written permission.
> + *
> + *
> + * ALTERNATIVELY, this software may be distributed under the terms of the
> + * GNU General Public License ("GPL") as published by the Free Software
> + * Foundation, either version 2 of that License or (at your option) any
> + * later version.

Dual licensed, GPL and BSD ?

> + *
> + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS 
> IS"
> + * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
> + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
> + * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE
> + * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
> + * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
> + * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
> + * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
> + * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
> + * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
> + * POSSIBILITY OF SUCH DAMAGE.
> + */
<...>

Re: [dpdk-dev] [PATCHv5 11/33] bus/fslmc: add vfio support

[Ferruh Yigit]

On 1/19/2017 1:23 PM, Hemant Agrawal wrote:
> Add support for using VFIO for dpaa2 based fsl-mc bus.
> 
> There are some differences in the way vfio used for fsl-mc bus
> from the eal vfio.
>  - The scanning of bus for individual objects on the basis of
>the DPRC container.
>  - The use and mapping of MC portal for object access
> 
> With the evolution of bus model, they canbe further aligned with
> eal vfio code.
> 
> Signed-off-by: Hemant Agrawal 

<...>

> diff --git a/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map 
> b/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
> index c4b3408..411200c 100644
> --- a/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
> +++ b/drivers/bus/fslmc/rte_pmd_fslmcbus_version.map
> @@ -45,8 +45,10 @@ DPDK_17.02 {
>  dpseci_open;
>  dpseci_reset;
>  dpseci_set_rx_queue;
> +mcp_ptr_list;
>  rte_fslmc_driver_register;
>  rte_fslmc_driver_unregister;
> +vfio_dmamap_mem_region;

This function seems called by files within bus folder, does it need to
be exported? Not tested, just asking..

>  
>   local: *;
>  };
> 

Re: [dpdk-dev] [PATCHv5 13/33] net/dpaa2: introducing NXP dpaa2 pmd driver

[Ferruh Yigit]

On 1/19/2017 1:23 PM, Hemant Agrawal wrote:
> add support for fsl-mc bus based dpaa2 pmd driver.
> 
> Signed-off-by: Hemant Agrawal 

<...>

> diff --git a/drivers/common/Makefile b/drivers/common/Makefile
> index e5bfecb..76ec2d1 100644
> --- a/drivers/common/Makefile
> +++ b/drivers/common/Makefile
> @@ -31,6 +31,8 @@
>  
>  include $(RTE_SDK)/mk/rte.vars.mk
>  
> +CONFIG_RTE_LIBRTE_DPAA2_COMMON = $(CONFIG_RTE_LIBRTE_DPAA2_PMD)

This logic make sense, is there any reason DPAA2_COMMON to be a config
option, it doesn't look like something a user would like to change on
its own.

Instead, as done here, if there is a user of common folder it is enabled
in Makefile. For this DPAA2_COMMON doesn't need to be a config option
itself I think.

> +
>  DIRS-$(CONFIG_RTE_LIBRTE_DPAA2_COMMON) += dpaa2
>  
>  include $(RTE_SDK)/mk/rte.subdir.mk
> diff --git a/drivers/net/Makefile b/drivers/net/Makefile
> index 40fc333..f716ca0 100644
> --- a/drivers/net/Makefile
> +++ b/drivers/net/Makefile
> @@ -57,7 +57,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_THUNDERX_NICVF_PMD) += thunderx
>  DIRS-$(CONFIG_RTE_LIBRTE_VIRTIO_PMD) += virtio
>  DIRS-$(CONFIG_RTE_LIBRTE_VMXNET3_PMD) += vmxnet3
>  DIRS-$(CONFIG_RTE_LIBRTE_PMD_XENVIRT) += xenvirt
> -
> +DIRS-$(CONFIG_RTE_LIBRTE_DPAA2_PMD) += dpaa2

Add alphabetically please.

>  ifeq ($(CONFIG_RTE_LIBRTE_VHOST),y)
>  DIRS-$(CONFIG_RTE_LIBRTE_PMD_VHOST) += vhost
>  endif # $(CONFIG_RTE_LIBRTE_VHOST)

<...>

> diff --git a/drivers/net/dpaa2/dpaa2_ethdev.c 
> b/drivers/net/dpaa2/dpaa2_ethdev.c
> new file mode 100644
> index 000..2295f82
> --- /dev/null
> +++ b/drivers/net/dpaa2/dpaa2_ethdev.c
<...>
> +/* Name of the DPAA2 Net PMD */
> +static const char *drivername = "DPAA2 PMD";

Custom names not preferred, please check any other PMD to be consistent.

<...>

> +static int
> +rte_dpaa2_probe(struct rte_dpaa2_driver *dpaa2_drv,
> + struct rte_dpaa2_device *dpaa2_dev)
> +{
> + struct eth_driver*eth_drv;

whitespace error.

> + struct rte_eth_dev *eth_dev;
> + char ethdev_name[RTE_ETH_NAME_MAX_LEN];
> +
> + int diag;
> +
> + eth_drv = (struct eth_driver *)dpaa2_drv;

How this suppose to work?

struct eth_driver {

struct rte_pci_driver pci_drv;
eth_dev_init_t eth_dev_init;
eth_dev_uninit_t eth_dev_uninit;
unsigned int dev_private_size;
};

struct rte_dpaa2_driver {
TAILQ_ENTRY(rte_dpaa2_driver) next;
struct rte_driver driver;
struct rte_fslmc_bus *fslmc_bus;
uint32_t drv_flags;
uint16_t drv_type;
rte_dpaa2_probe_t probe
rte_dpaa2_remove_t remove;
};

> +
> + sprintf(ethdev_name, "dpni-%d", dpaa2_dev->object_id);
> +
> + eth_dev = rte_eth_dev_allocate(ethdev_name);
> + if (eth_dev == NULL)
> + return -ENOMEM;
> +
> + if (rte_eal_process_type() == RTE_PROC_PRIMARY) {
> + eth_dev->data->dev_private = rte_zmalloc(
> + "ethdev private structure",
> + sizeof(struct dpaa2_dev_priv),
> + RTE_CACHE_LINE_SIZE);
> + if (eth_dev->data->dev_private == NULL) {
> + RTE_LOG(CRIT, PMD, "Cannot allocate memzone for"
> + " private port data\n");
> + return -ENOMEM;

release allocated port?

> + }
> + }
> + eth_dev->device = &dpaa2_dev->device;
> + dpaa2_dev->eth_dev = eth_dev;
> + eth_dev->driver = eth_drv;
> + eth_dev->data->rx_mbuf_alloc_failed = 0;
> +
> + /* init user callbacks */
> + TAILQ_INIT(ð_dev->link_intr_cbs);

This is no more required, since done by rte_eth_dev_allocate()

> +
> + /*
> +  * Set the default MTU.
> +  */
> + eth_dev->data->mtu = ETHER_MTU;

Same, no more required to set in pmd.

> +
> + /* Invoke PMD device initialization function */
> + diag = dpaa2_dev_init(eth_dev);
> + if (diag == 0)
> + return 0;
> +
> + if (rte_eal_process_type() == RTE_PROC_PRIMARY)
> + rte_free(eth_dev->data->dev_private);
> + rte_eth_dev_release_port(eth_dev);
> + return diag;
> +}

<...>

> +static struct rte_dpaa2_driver rte_dpaa2_pmd = {
> + .drv_type = DPAA2_MC_DPNI_DEVID,
> + .driver = {
> + .name = "DPAA2 PMD",

This is not required, RTE_PMD_REGISTER_DPAA2 will set it.

> + },
> + .probe = rte_dpaa2_probe,
> + .remove = rte_dpaa2_remove,

These are PMD specific APIs, PCI equivalent of these are eth_dev_init()
and eth_dev_uninit() functions. Instead of rte_eth_dev_pci_probe() like
function.
Here it seems used as how it is used for virtual devices. But even for
virtual devices, it is desired to have more proper virtual bus structure.

I understand this part is a little problematic now, because of the
eth_driver <-> pci_driver relation, this is not generic.
What do you think first solve eth_driver <-> pci_driv

Re: [dpdk-dev] [PATCH] lib/librte_ether: error handling on MAC address replay

[Steve Shin (jonshin)]

Dear maintainer, 

Sorry that I forgot to add “Fixes:” line as follows:

   Fixes: 4bdefaade6d1 ("ethdev: VMDQ enhancements")

Can you please add the above line as part of comment?

Thanks,
Steve

On 1/19/17, 10:47 AM, "Steve Shin (jonshin)"  wrote:

This patch fixes a bug in replaying MAC address to the hardware
in rte_eth_dev_config_restore() routine.

Signed-off-by: Steve Shin 
---
 lib/librte_ether/rte_ethdev.c | 10 ++
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 4790faf..7e01f10 100644
--- a/lib/librte_ether/rte_ethdev.c
+++ b/lib/librte_ether/rte_ethdev.c
@@ -951,10 +951,12 @@ rte_eth_dev_config_restore(uint8_t port_id)
continue;
 
/* add address to the hardware */
-   if  (*dev->dev_ops->mac_addr_add &&
-   (dev->data->mac_pool_sel[i] & (1ULL << pool)))
-   (*dev->dev_ops->mac_addr_add)(dev, &addr, i, pool);
-   else {
+   if  (*dev->dev_ops->mac_addr_add) {
+   if (dev->data->mac_pool_sel[i] & (1ULL << pool))
+   (*dev->dev_ops->mac_addr_add)(dev, &addr, i, 
pool);
+   else
+   continue;
+   } else {
RTE_PMD_DEBUG_TRACE("port %d: MAC address array not 
supported\n",
port_id);
/* exit the loop but not return an error */
-- 
2.9.3

Re: [dpdk-dev] [dpdk-users] IGB_UIO: PCI Resources Management

[Alejandro Lucero]

On Thu, Jan 19, 2017 at 4:36 PM, Ferruh Yigit 
wrote:

> On 1/13/2017 11:10 AM, Alejandro Lucero wrote:
> > I completely misread the patch, and I wrongly thought that code was
> > linked to module removal, but I see this is not about that, but about
> > releasing the /dev/uio file calling release function, what is done by
> > the kernel when the process exits.
> >
> > So yes, the patch avoids the problem I talked about.
> >
> > However, calling that specific ixgbe driver function will break other
> > devices relying on igb_uio. What about implementing a notifier chain for
> > this? The igb_uio code would be agnostic and each interested driver,
> > like ixgbe or nfp_net, could execute the specific port close code when
> > the notifier chain triggers.
>
> Can you please elaborate, how can this be done?
> This is kernel code, and interested drivers are in userspace. This patch
> benefit from Linux kernel driver of same device.
>
>
The patch seems to add  a function to the ixgve_vf kernel driver which will
be called from igb_uio. The patch does not show the usual path to the
kernel ethernet drivers because, I guess, it is a patch for a specific
software build. So the interested drivers are not in user space, and a
driver, igb_uio, calling a userspace function? No way.

Any netdev driver interested to close ports, those left working by a
crashed DPDK app, could register a notifier like this:


int device_notifier(struct notifier_block *nb, unsigned long action, void
*data) {

...
if (action == NETDEV_GOING_DOWN) {

...
}


struct notifier_block device_nb = {
   .notifier_call =device_notifier,
};

...

register_netdev_notifer(&device_nb);

...


Inside igb_uio, something like this:

raw_notifier_call_chain(&netdev_chain, val, dev);

which would trigger the invocation of all registered notifier functions.

Please, note that this is just a possibility, and the notifier chain to use
could not be the best one, same for the event NETDEV_GOING_DOWN.

Also, how can the drivers know if that event is for them?, This is
something to work on.

Because UIO is part of the kernel, maybe it could be pushed a specific
chain for this purpose if we can not use any currently available.


> >
> >
> > On Fri, Jan 13, 2017 at 5:33 AM, Tan, Jianfeng  > > wrote:
> >
> >
> >
> > > -Original Message-
> > > From: Yigit, Ferruh
> > > Sent: Friday, January 13, 2017 10:05 AM
> > > To: Tan, Jianfeng; Alejandro Lucero
> > > Cc: Gregory Etelson; dev; us...@dpdk.org 
> > > Subject: Re: [dpdk-users] [dpdk-dev] IGB_UIO: PCI Resources
> Management
> > >
> > > On 1/13/2017 1:51 AM, Tan, Jianfeng wrote:
> > > >
> > > >
> > > >> -Original Message-
> > > >> From: users [mailto:users-boun...@dpdk.org
> > ] On Behalf Of Ferruh Yigit
> > > >> Sent: Thursday, January 12, 2017 8:22 PM
> > > >> To: Alejandro Lucero
> > > >> Cc: Gregory Etelson; dev; us...@dpdk.org  >
> > > >> Subject: Re: [dpdk-users] [dpdk-dev] IGB_UIO: PCI Resources
> > > Management
> > > >>
> > > >> On 1/12/2017 12:12 PM, Alejandro Lucero wrote:
> > > >>>
> > > >>>
> > > >>> On Thu, Jan 12, 2017 at 11:55 AM, Ferruh Yigit
> > mailto:ferruh.yi...@intel.com>
> > > >>>  > >> wrote:
> > > >>>
> > > >>> On 12/9/2016 8:54 AM, Gregory Etelson wrote:
> > > >>> > Hello,
> > > >>> >
> > > >>> > IGB_UIO driver does not close port PCI activities after
> > DPDK process
> > > >> exits.
> > > >>> > DPDK API provides rte_eth_dev_close() to manage port PCI,
> > > >>> > but it can be skipped if process receives SIGKILL signal
> > > >>>
> > > >>> I guess I understand the problem.
> > > >>>
> > > >>>
> > > >>> This is a known problem, but it is not just a UIO problem, and
> > this
> > > >>> patch does not solve it, maybe it just solves part of it.
> > > >>>
> > > >>> In fact, a DPDK program crashing could imply the NIC DMAing
> > after that
> > > >>> and after that memory was assigned to another program.
> > > >>
> > > >> Yes.
> > > >> Can there be a way to stop NIC DMA, (or prevent it access to mem
> > > >> anymore) when app crashes?
> > > >> I think that is what this patch is looking for.
> > > >
> > > > If I understand it correctly, you are looking for this patch?
> > > > http://dpdk.org/dev/patchwork/patch/17495/
> > 
> > > >
> > >
> > > That is good, thanks Jianfeng, I will check it.
> > >
> > > btw, patch's current state is rejected, which is by mistake, it
> > seems I
> > > confused it with "iomem and ioport mapping" patch, sorry about it,
> I
> > > will update its status immediately.
> >
> >   

[dpdk-dev] A question about the macro "RTE_ETH_IS_IPV4_HDR" used in the example l3fwd-acl

[Wu, Xiaoban]

Dear DPDK users,


I am working on the l3fwd-acl example 
http://dpdk.org/doc/guides/sample_app_ug/l3_forward_access_ctrl.html.


I just find that the macro "RTE_ETH_IS_IPV4_HDR" used on line 699 and 752 of 
the main.c rejected all the normal UDP packets.


Hence, I want to know if the PMD driver is required to parse the packet and 
update the "packet_type" field in the "struct rte_mbuf" whenever the packet is 
received?


If the answer is yes, then the PMD "librte_pmd_mlx5" seems to fail to do so, 
since all the packet_type I got are all zero which are not the required 
0x0010.


If the answer is no, then there would be better some warning on the webpage.


Thank you very much!


Best wishes,

Xiaoban

Re: [dpdk-dev] [PATCH] lib/librte_ether: error handling on MAC address replay

[Igor Ryzhov]

Hello Steve,

Thank you for the patch.

I think a couple of improvements can be done:
1. Function existence check – if (*dev->dev_ops->mac_addr_add) – can be
taken out of the loop. We don't need to check it on each iteration.
2. I'm not completely sure, but I think loop can be started from 1, not
from 0, because mac_pool_sel[0] is always zero. Am I right?

Best regards,
Igor

On Thu, Jan 19, 2017 at 10:35 PM, Steve Shin (jonshin) 
wrote:

> Dear maintainer,
>
> Sorry that I forgot to add “Fixes:” line as follows:
>
>Fixes: 4bdefaade6d1 ("ethdev: VMDQ enhancements")
>
> Can you please add the above line as part of comment?
>
> Thanks,
> Steve
>
> On 1/19/17, 10:47 AM, "Steve Shin (jonshin)"  wrote:
>
> This patch fixes a bug in replaying MAC address to the hardware
> in rte_eth_dev_config_restore() routine.
>
> Signed-off-by: Steve Shin 
> ---
>  lib/librte_ether/rte_ethdev.c | 10 ++
>  1 file changed, 6 insertions(+), 4 deletions(-)
>
> diff --git a/lib/librte_ether/rte_ethdev.c
> b/lib/librte_ether/rte_ethdev.c
> index 4790faf..7e01f10 100644
> --- a/lib/librte_ether/rte_ethdev.c
> +++ b/lib/librte_ether/rte_ethdev.c
> @@ -951,10 +951,12 @@ rte_eth_dev_config_restore(uint8_t port_id)
> continue;
>
> /* add address to the hardware */
> -   if  (*dev->dev_ops->mac_addr_add &&
> -   (dev->data->mac_pool_sel[i] & (1ULL << pool)))
> -   (*dev->dev_ops->mac_addr_add)(dev, &addr, i,
> pool);
> -   else {
> +   if  (*dev->dev_ops->mac_addr_add) {
> +   if (dev->data->mac_pool_sel[i] & (1ULL << pool))
> +   (*dev->dev_ops->mac_addr_add)(dev, &addr,
> i, pool);
> +   else
> +   continue;
> +   } else {
> RTE_PMD_DEBUG_TRACE("port %d: MAC address array
> not supported\n",
> port_id);
> /* exit the loop but not return an error */
> --
> 2.9.3
>
>
>
>

[dpdk-dev] [PATCH] mbuf: remove redundant line in rte_pktmbuf_attach

[Ilya Matveychikov]

mi->next will be assigned to NULL few lines later, trivial patch

Signed-off-by: Ilya V. Matveychikov 
---
 lib/librte_mbuf/rte_mbuf.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/lib/librte_mbuf/rte_mbuf.h b/lib/librte_mbuf/rte_mbuf.h
index ead7c6e..5589d54 100644
--- a/lib/librte_mbuf/rte_mbuf.h
+++ b/lib/librte_mbuf/rte_mbuf.h
@@ -1139,7 +1139,6 @@ static inline void rte_pktmbuf_attach(struct rte_mbuf 
*mi, struct rte_mbuf *m)
mi->buf_addr = m->buf_addr;
mi->buf_len = m->buf_len;
 
-   mi->next = m->next;
mi->data_off = m->data_off;
mi->data_len = m->data_len;
mi->port = m->port;
-- 
2.7.4

[dpdk-dev] FW: [PATCH v7] app/testpmd: supported offload capabilities query

[Yang, Qiming]

Hi, Thomas
Does this patch be forget? Or still have something need to amend?

> -Original Message-
> From: De Lara Guarch, Pablo
> Sent: Monday, January 16, 2017 6:10 PM
> To: Yang, Qiming ; dev@dpdk.org
> Cc: Yang, Qiming 
> Subject: RE: [dpdk-dev] [PATCH v7] app/testpmd: supported offload
> capabilities query
> 
> 
> 
> > -Original Message-
> > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Qiming Yang
> > Sent: Monday, January 16, 2017 2:31 AM
> > To: dev@dpdk.org
> > Cc: Yang, Qiming
> > Subject: [dpdk-dev] [PATCH v7] app/testpmd: supported offload
> > capabilities query
> >
> > Add two new commands "show port cap " and "show port cap all"to
> > diaplay what offload capabilities supported in ports. It will not only
> > display all the capabilities of the port, but also the enabling
> > condition for each capability in the running time.
> >
> > Signed-off-by: Qiming Yang 
> > Acked-by: Jingjing Wu 
> > Acked-by: Beilei Xing 
> 
> Acked-by: Pablo de Lara 

Re: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter

[Lu, Wenzhuo]

Hi Adrien, Thomas, Yong,

> -Original Message-
> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Adrien Mazarguil
> Sent: Friday, January 20, 2017 2:46 AM
> To: Thomas Monjalon
> Cc: Liu, Yong; dev@dpdk.org
> Subject: Re: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter
> 
> On Thu, Jan 19, 2017 at 10:06:34AM +0100, Thomas Monjalon wrote:
> > 2017-01-19 13:34, Yong Liu:
> > > +* ABI changes are planned for 17.05: structure
> > > +``rte_eth_tunnel_filter_conf``
> > > +  will be extended with a new member ``vf_id`` in order to enable
> > > +cloud filter
> > > +  on VF device.
> >
> > I think we should stop rely on this API, and migrate to rte_flow instead.
> > Adrien any thought?
> 
> I'm all for using rte_flow in any case. I've already documented an approach to
> convert TUNNEL filter rules to rte_flow rules [1], although it may be
> incomplete due to my limited experience with this filter type. We already
> know several tunnel item types must be added (currently only VXLAN is
> defined).
> 
> I understand ixgbe/i40e currently map rte_flow on top of the legacy
> framework, therefore extending this structure might still be needed in the
> meantime. Not sure we should prevent this change as long as such rules can be
> configured through rte_flow as well.
> 
> [1] http://dpdk.org/doc/guides/prog_guide/rte_flow.html#tunnel-to-eth-ipv4-
> ipv6-vxlan-or-other-queue
The problem is we haven't finished transferring all the functions from the 
regular filters to the generic filters. 
For example, igb, fm10k and enic haven't support generic filters yet. Ixgbe and 
i40e have supported the basic functions, but some advance features are not 
transferred to generic filters yet.
Seems it's not the time to remove the regular filters. Yong, I suggest to 
support both generic filter and regular filter in parallel.
So, we need to announce ABI change for the regular filter, until someday we 
remove the regular filter API. 

> 
> --
> Adrien Mazarguil
> 6WIND

Re: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter

[Liu, Yong]

Wenzhou, thanks for note.

> -Original Message-
> From: Lu, Wenzhuo
> Sent: Friday, January 20, 2017 10:15 AM
> To: Adrien Mazarguil ; Thomas Monjalon
> 
> Cc: Liu, Yong ; dev@dpdk.org
> Subject: RE: [dpdk-dev] [PATCH] doc: announce ABI change for cloud filter
> 
> Hi Adrien, Thomas, Yong,
> 
> > -Original Message-
> > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Adrien Mazarguil
> > Sent: Friday, January 20, 2017 2:46 AM
> > To: Thomas Monjalon
> > Cc: Liu, Yong; dev@dpdk.org
> > Subject: Re: [dpdk-dev] [PATCH] doc: announce ABI change for cloud
> > filter
> >
> > On Thu, Jan 19, 2017 at 10:06:34AM +0100, Thomas Monjalon wrote:
> > > 2017-01-19 13:34, Yong Liu:
> > > > +* ABI changes are planned for 17.05: structure
> > > > +``rte_eth_tunnel_filter_conf``
> > > > +  will be extended with a new member ``vf_id`` in order to enable
> > > > +cloud filter
> > > > +  on VF device.
> > >
> > > I think we should stop rely on this API, and migrate to rte_flow instead.
> > > Adrien any thought?
> >
> > I'm all for using rte_flow in any case. I've already documented an
> > approach to convert TUNNEL filter rules to rte_flow rules [1],
> > although it may be incomplete due to my limited experience with this
> > filter type. We already know several tunnel item types must be added
> > (currently only VXLAN is defined).
> >
> > I understand ixgbe/i40e currently map rte_flow on top of the legacy
> > framework, therefore extending this structure might still be needed in
> > the meantime. Not sure we should prevent this change as long as such
> > rules can be configured through rte_flow as well.
> >
> > [1]
> > http://dpdk.org/doc/guides/prog_guide/rte_flow.html#tunnel-to-eth-ipv4
> > -
> > ipv6-vxlan-or-other-queue
> The problem is we haven't finished transferring all the functions from the 
> regular
> filters to the generic filters.
> For example, igb, fm10k and enic haven't support generic filters yet. Ixgbe 
> and
> i40e have supported the basic functions, but some advance features are not
> transferred to generic filters yet.
> Seems it's not the time to remove the regular filters. Yong, I suggest to 
> support
> both generic filter and regular filter in parallel.
> So, we need to announce ABI change for the regular filter, until someday we
> remove the regular filter API.


I will enable VF support both in generic filter and original cloud filter API.
So I think we still need ABI announcement for structure modification.

> 
> >
> > --
> > Adrien Mazarguil
> > 6WIND

Re: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

[Guo, Jia]

Ok, that make sense, I will re-correct it then. Thanks.

Best regards,
Jeff Guo

-Original Message-
From: Mcnamara, John 
Sent: Thursday, January 19, 2017 10:54 PM
To: Guo, Jia ; Zhang, Helin ; Wu, 
Jingjing 
Cc: dev@dpdk.org; g...@dpdk.org
Subject: RE: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic issue for i40e

> -Original Message-
> From: Guo, Jia
> Sent: Thursday, January 19, 2017 1:50 PM
> To: Mcnamara, John ; Zhang, Helin 
> ; Wu, Jingjing 
> Cc: dev@dpdk.org; g...@dpdk.org
> Subject: RE: [dpdk-dev] [PATCH v3] doc: add known uio_pci_generic 
> issue for i40e
> 
> Hi,john
>   The "bind uio_pci_generic`` is not the definitely command, but 
> very appreciate for your correcting the other syntax issue. I will 
> involve them in v4, Thanks very much.

Hi Jeff, 

Variables, functions, program names, file names and any other literal should be 
in backtick quotes, as outlined in the DPDK Documentation: 

http://dpdk.org/doc/guides/contributing/documentation.html#code-and-literal-block-sections

So uio_pci_generic should be quoted with backticks. If you want to add bind 
then do:
``--bind= uio_pci_generic`` or similar.

Either way it shouldn't be quoted using the "smart" Unicode or non-Unicode 
quotes as in the patch. All quotes should be straight ASCII double quotes (""). 
The Html/PDF renderers will convert to the correct quotes as required.



> > related `linux kernel commit
> > +   < https://git.kernel.org/cgit/linux/kernel/git/stable/linux-
> > stable.git/commit/drivers/pci/quirks.c?id=8bcf4525c5d43306c5fd07e132
> > bc
> > 8650
> > e3491aec>`_.

Also, this link generates a doc build error. I think because of the space 
between the < and the https:.

John

Re: [dpdk-dev] [PATCH] lib/librte_ether: error handling on MAC address replay

[Steve Shin (jonshin)]

Thanks Igor for your comments!

A good point on the first item.
For the second question, I walked through device initialization routines to see 
any other potential issue.
I found one case where index 0 is still required to replay MAC address; ex) MAC 
address change on the PHY port – index 0
No code path was found to restore a changed MAC value with index 0. Therefore, 
we may have to use the existing rte_eth_dev_config_restore() function to replay 
0 indexed MAC.
Currently rte_eth_dev_default_mac_addr_set() is called to program PHY MAC 
without setting mac_pool_sel for 0 index.
So the following code is also required inside 
rte_eth_dev_default_mac_addr_set() routine as a complete solution:
@@ -2237,6 +2234,9 @@ struct rte_eth_dev *

(*dev->dev_ops->mac_addr_set)(dev, addr);

+   /* Update pool bitmap in NIC data structure */
+   dev->data->mac_pool_sel[0] = 1;
+
return 0;

Any thoughts would be appreciated.

Regards,
Steve

From: Igor Ryzhov 
Date: Thursday, January 19, 2017 at 2:39 PM
To: Steve Shin 
Cc: "dev@dpdk.org" , "ferruh.yi...@intel.com" 

Subject: Re: [dpdk-dev] [PATCH] lib/librte_ether: error handling on MAC address 
replay

Hello Steve,

Thank you for the patch.

I think a couple of improvements can be done:
1. Function existence check – if (*dev->dev_ops->mac_addr_add) – can be taken 
out of the loop. We don't need to check it on each iteration.
2. I'm not completely sure, but I think loop can be started from 1, not from 0, 
because mac_pool_sel[0] is always zero. Am I right?

Best regards,
Igor

On Thu, Jan 19, 2017 at 10:35 PM, Steve Shin (jonshin) 
mailto:jons...@cisco.com>> wrote:
Dear maintainer,

Sorry that I forgot to add “Fixes:” line as follows:

   Fixes: 4bdefaade6d1 ("ethdev: VMDQ enhancements")

Can you please add the above line as part of comment?

Thanks,
Steve

On 1/19/17, 10:47 AM, "Steve Shin (jonshin)" 
mailto:jons...@cisco.com>> wrote:

This patch fixes a bug in replaying MAC address to the hardware
in rte_eth_dev_config_restore() routine.

Signed-off-by: Steve Shin mailto:jons...@cisco.com>>
---
 lib/librte_ether/rte_ethdev.c | 10 ++
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/lib/librte_ether/rte_ethdev.c b/lib/librte_ether/rte_ethdev.c
index 4790faf..7e01f10 100644
--- a/lib/librte_ether/rte_ethdev.c
+++ b/lib/librte_ether/rte_ethdev.c
@@ -951,10 +951,12 @@ rte_eth_dev_config_restore(uint8_t port_id)
continue;

/* add address to the hardware */
-   if  (*dev->dev_ops->mac_addr_add &&
-   (dev->data->mac_pool_sel[i] & (1ULL << pool)))
-   (*dev->dev_ops->mac_addr_add)(dev, &addr, i, pool);
-   else {
+   if  (*dev->dev_ops->mac_addr_add) {
+   if (dev->data->mac_pool_sel[i] & (1ULL << pool))
+   (*dev->dev_ops->mac_addr_add)(dev, &addr, i, 
pool);
+   else
+   continue;
+   } else {
RTE_PMD_DEBUG_TRACE("port %d: MAC address array not 
supported\n",
port_id);
/* exit the loop but not return an error */
--
2.9.3

[dpdk-dev] [PATCH v4] doc: add known uio_pci_generic issue for i40e

[Jeff Guo]

When bind the "uio_pci_generic" module in X710/XL710/XXV710,
the result is failed. The "uio_pci_generic" module is not
supported by X710/XL710/XXV710.

Signed-off-by: Jeff Guo 
Acked-by: Jingjing Wu 
---
v4: modify the signed-off name to be consistent and fix make doc issue.
v3: add related linux kernel commit info and fix some syntax issues
v2: add X710/XXV710 limitation case
---
 doc/guides/rel_notes/known_issues.rst | 26 ++
 1 file changed, 26 insertions(+)

diff --git a/doc/guides/rel_notes/known_issues.rst 
b/doc/guides/rel_notes/known_issues.rst
index 3cd4237..b1e581b 100644
--- a/doc/guides/rel_notes/known_issues.rst
+++ b/doc/guides/rel_notes/known_issues.rst
@@ -640,3 +640,29 @@ I40e VF may not receive packets in the promiscuous mode
 
 **Driver/Module**:
Poll Mode Driver (PMD).
+
+
+The uio_pci_generic module bind failed in X710/XL710/XXV710
+---
+
+**Description**:
+   The “uio_pci_generic” module is not supported by XL710, since the errata of 
XL710
+   states that the Interrupt Status bit is not implemented. The errata is the 
item #71
+   from the `xl710 controller spec
+   
`_.
+   The hw limitation is the same as other X710/XXV710 NICs.
+
+**Implication**:
+   When use "--bind=uio_pci_generic”, the "uio_pci_generic" module probes 
device and check the Interrupt
+   Status bit. Since it is not supported by X710/XL710/XXV710, it return a 
*failed* value. The statement
+   that these products don’t support INTx masking, is indicated in the related 
`linux kernel commit
+   
`_.
+
+**Resolution/Workaround**:
+   Do not bind the “uio_pci_generic” module in X710/XL710/XXV710 NICs.
+
+**Affected Environment/Platform**:
+   All.
+
+**Driver/Module**:
+   Poll Mode Driver (PMD).
-- 
2.7.4

Re: [dpdk-dev] [PATCHv5 02/33] doc: add dpaa2 nic details

[Hemant Agrawal]

On 1/19/2017 10:38 PM, Thomas Monjalon wrote:

2017-01-19 18:53, Hemant Agrawal:

--- a/MAINTAINERS
+++ b/MAINTAINERS
+NXP DPAA2 PMD
+M: Hemant Agrawal 
+F: drivers/bus/fslmc/
+F: drivers/common/dpaa2/
+F: drivers/net/dpaa2/
+F: drivers/pool/dpaa2/
+F: doc/guides/nics/dpaa2.rst


Please introduce line by line in the patches
adding these files.



Ok


+++ b/doc/guides/nics/features/dpaa2.ini
@@ -0,0 +1,8 @@
+;
+; Supported features of the 'dpaa2' network poll mode driver.
+;
+; Refer to default.ini for the full list of available PMD features.
+;
+[Features]
+ARMv8= Y
+Usage doc= Y


You should introduce this file with the first PMD patch.



Ok


+* **Added a new driver for NXP DPAA2 - FSLMC bus.**
+
+  Added the new bus "fslmc" driver for NXP DPAA2 devices. See the
+  "Network Interface Controller Drivers" document for more details on this new
+  driver.
+
+* **Added a new driver for NXP DPAA2 Network PMD.**
+
+  Added the new "dpaa2" net driver for NXP DPAA2 devices. See the
+  "Network Interface Controller Drivers" document for more details on this new
+  driver.


You can use sphinx references to create a link in the generated HTML.

Please move doc patch after (or within) the code.



ok

Re: [dpdk-dev] [PATCHv5 02/33] doc: add dpaa2 nic details

[Hemant Agrawal]

On 1/19/2017 11:04 PM, Mcnamara, John wrote:




-Original Message-
From: Hemant Agrawal [mailto:hemant.agra...@nxp.com]
Sent: Thursday, January 19, 2017 1:23 PM
To: dev@dpdk.org
Cc: thomas.monja...@6wind.com; Richardson, Bruce
; shreyansh.j...@nxp.com; Mcnamara, John
; Yigit, Ferruh ;
jerin.ja...@caviumnetworks.com; Hemant Agrawal 
Subject: [PATCHv5 02/33] doc: add dpaa2 nic details

...

+* **Added a new driver for NXP DPAA2 - FSLMC bus.**
+
+  Added the new bus "fslmc" driver for NXP DPAA2 devices. See the
+ "Network Interface Controller Drivers" document for more details on
+ this new  driver.
+
+* **Added a new driver for NXP DPAA2 Network PMD.**
+
+  Added the new "dpaa2" net driver for NXP DPAA2 devices. See the
+ "Network Interface Controller Drivers" document for more details on
+ this new  driver.

 New Features
 


These additions should be after "New Features" and not before, if you do
another revision you can fix it. If not it could be fixed on commit or I
can fix it in the final Release note review.

But overall the NIC guide is very good.

Acked-by: John McNamara 



I am doing next rev, so I will fix it.