RE: [EXT] [dpdk-dev v1] lib/cryptodev: multi-process IPC request handler

Hi Fan, > Hi Akhil, > > > This is a library change you should cc all PMD owners while sending patch. > Kai is in holiday at the moment and will be back in a week. I will sync with > him > then. > > > > > This patch add in multi-process IPC request handler function in rte > > > cryptodev. This fun

[PATCH 00/18] Fixes and improvements in cnxk crypto PMDs

New features added: - AES-CCM with lookaside IPsec operations - DOCSIS algos with lookaside crypto operations - Redesigned datapath for improving performance Anoob Joseph (11): crypto/cnxk: remove zero IV crypto/cnxk: limit the meta buf cache to 128 crypto/cnxk: add separate path for pdcp

[PATCH 01/18] crypto/cnxk: add AES-CCM support

From: Archana Muniganti Add lookaside IPsec AES-CCM support in CN9K & CN10K PMDs. Signed-off-by: Archana Muniganti --- doc/guides/rel_notes/release_22_11.rst| 4 ++ drivers/common/cnxk/cnxk_security.c | 38 -- drivers/common/cnxk/roc_cpt.h | 13 ++

[PATCH 02/18] crypto/cnxk: add burst enqueue for event crypto

From: Volodymyr Fialko Added support for burst enqueue for cn10k event crypto adapter. Instruction will be grouped based on the queue pair and sent in a burst. Signed-off-by: Volodymyr Fialko --- drivers/crypto/cnxk/cn10k_cryptodev_ops.c | 141 +++--- drivers/crypto/cnxk/cn10k_

[PATCH 03/18] crypto/cnxk: remove zero IV

Zeroising AES-CMAC IV would be done in microcode. Extra clearing in DPDK is not required. Signed-off-by: Anoob Joseph Signed-off-by: Tejasree Kondoj --- drivers/crypto/cnxk/cnxk_se.h | 4 1 file changed, 4 deletions(-) diff --git a/drivers/crypto/cnxk/cnxk_se.h b/drivers/crypto/cnxk/cnxk_

[PATCH 04/18] crypto/cnxk: limit the meta buf cache to 128

Limit meta buf pool cache size as 128. Having 512 as the cache size would cause more time for refill. Signed-off-by: Anoob Joseph Signed-off-by: Volodymyr Fialko --- drivers/crypto/cnxk/cnxk_cryptodev_ops.c | 9 + 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/cry

[PATCH 05/18] crypto/cnxk: add separate path for pdcp chain opcode

Add separate datapath for PDCP chain opcode. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cn9k_cryptodev_ops.c | 7 +- drivers/crypto/cnxk/cnxk_se.h| 229 +++ drivers/crypto/cnxk/meson.build | 4 +- 3 files changed, 196 insertions(+), 44 deleti

[PATCH 06/18] crypto/cnxk: add separate datapath for pdcp cipher operation

Add separate datapath for PDCP opcode performing cipher operation. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cn10k_cryptodev_ops.c | 19 --- drivers/crypto/cnxk/cn9k_cryptodev_ops.c | 27 +--- drivers/crypto/cnxk/cnxk_se.h | 177 +++--- 3 files changed, 1

[PATCH 07/18] crypto/cnxk: remove MAC len check for AEAD

For both AES-GCM and ChaChaPoly, digest size must be non-zero. So remove the conditional handling in the datapath. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cnxk_se.h | 26 +- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/drivers/crypto/cnxk/cn

[PATCH 08/18] crypto/cnxk: fix endianness in anti-replay

From: Tejasree Kondoj Fixing anti-replay endianness issue in lookaside IPsec. Fixes: c1eac1b966c2 ("crypto/cnxk: add anti-replay as per new firmware") Cc: ktejas...@marvell.com Signed-off-by: Tejasree Kondoj --- drivers/crypto/cnxk/cn9k_cryptodev_ops.c | 21 +++-- 1 file chang

[PATCH 09/18] crypto/cnxk: remove extra indirection for FC and Kasumi

Fill FC params routine now caters to only KASUMI & FC_GEN. Call appropriate functions directly. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cnxk_se.h | 51 +++ 1 file changed, 15 insertions(+), 36 deletions(-) diff --git a/drivers/crypto/cnxk/cnxk_se.h b/

[PATCH 10/18] crypto/cnxk: remove extra digest len check

When the flag ROC_SE_VALID_MAC_BUF is set, digest length will be non-zero. Remove extra check in datapath. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cnxk_se.h | 8 +++- 1 file changed, 3 insertions(+), 5 deletions(-) diff --git a/drivers/crypto/cnxk/cnxk_se.h b/drivers/crypto/cnxk

[PATCH 11/18] crypto/cnxk: avoid accessing se ctx in aes gcm path

Field op_minor is required only for digest encrypted cases with chained ops. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cnxk_se.h | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/crypto/cnxk/cnxk_se.h b/drivers/crypto/cnxk/cnxk_se.h index 64a66ef911..55f4

[PATCH 12/18] crypto/cnxk: remove auth iv from kasumi cipher

Auth IV is not applicable for kasumi cipher operation. Remove the same. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cnxk_se.h | 10 -- 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/drivers/crypto/cnxk/cnxk_se.h b/drivers/crypto/cnxk/cnxk_se.h index 55f411d50c..3de

[PATCH 13/18] crypto/cnxk: enable IE engine for Chacha-Poly

From: Tejasree Kondoj Change engine group of Chacha20-Poly1305 as it is supported on IE engines also. Signed-off-by: Tejasree Kondoj --- drivers/crypto/cnxk/cnxk_cryptodev_ops.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/crypto/cnxk/cnxk_cryptodev_ops.c b/driv

[PATCH 14/18] crypto/cnxk: use dedicated dp threads depending on operation

Identify the datapath thread to be used during session create. This can be used to call right function early on to avoid multiple session specific checks in datapath functions. Signed-off-by: Anoob Joseph --- drivers/crypto/cnxk/cnxk_cryptodev_ops.c | 29 + drivers/crypto/cnxk/cnxk_s

[PATCH 15/18] crypto/cnxk: remove unused ctx buf len

Ctx len buf field is unused. Remove it. Signed-off-by: Anoob Joseph --- drivers/common/cnxk/roc_se.h | 45 +++ drivers/crypto/cnxk/cnxk_se.h | 27 +++-- 2 files changed, 33 insertions(+), 39 deletions(-) diff --git a/drivers/common/cnxk/roc_se.h

[PATCH 16/18] drivers: change crypto adapter datapath error print to debug

From: Shijith Thotton Errors in crypto adapter datapath can be handled gracefully. So the error print can be moved under debug. Signed-off-by: Shijith Thotton --- drivers/common/cnxk/roc_platform.h| 2 ++ drivers/crypto/cnxk/cn10k_cryptodev_ops.c | 2 +- 2 files changed, 3 insertions(+

[PATCH 17/18] crypto/cnxk: update flow label copy capability

From: Vamsi Attunuru Update flow label copy capability in IPsec options. Signed-off-by: Vamsi Attunuru --- drivers/crypto/cnxk/cnxk_cryptodev_capabilities.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/crypto/cnxk/cnxk_cryptodev_capabilities.c b/drivers/crypto/cnxk/cnxk_cryptod

[PATCH 18/18] crypto/cnxk: add support for DOCSIS algorithm

From: Volodymyr Fialko Add support for offloading RTE_CRYPTO_CIPHER_AES_DOCSISBPI and RTE_CRYPTO_CIPHER_DES_DOCSISBPI algorithms to symmetric crypto session. Signed-off-by: Volodymyr Fialko --- doc/guides/cryptodevs/features/cn9k.ini | 2 + doc/guides/rel_notes/release_22_11.rst

Re: [PATCH v3 2/2] net: have checksum routines accept unaligned data

On 2022-07-11 15:25, Olivier Matz wrote: > On Mon, Jul 11, 2022 at 02:11:32PM +0200, Mattias Rönnblom wrote: >> __rte_raw_cksum() (used by rte_raw_cksum() among others) accessed its >> data through an uint16_t pointer, which allowed the compiler to assume >> the data was 16-bit aligned. This in tur

[PATCH] net/tap: Allow jumbo frames

eth_dev_validate_mtu, introduced in 990912e676e, validates configured MTU plus overhead against max_rx_pktlen. Since TAP is a virtual device, it should support as big MTU as possible. ---  drivers/net/tap/rte_eth_tap.c | 2 +-  1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net

[PATCH 0/2] common/mlx5: fix multi-process mempool registration

Dmitry Kozlyuk (2): mempool: make event callbacks process-private common/mlx5: fix multi-process mempool registration drivers/common/mlx5/mlx5_common.c| 15 drivers/common/mlx5/mlx5_common_mr.c | 2 +- drivers/common/mlx5/mlx5_common_mr.h | 2 +- lib/mempool/rte_mempool.c

[PATCH 1/2] mempool: make event callbacks process-private

Callbacks for mempool events were registered in a process-shared tailq. This was inherently incorrect because the same function may be loaded to a different address in each process. Make the tailq process-private. Use the EAL tailq lock to reduce the number of different locks this module operates.

[PATCH 2/2] common/mlx5: fix multi-process mempool registration

The `mp_cb_registered` flag shared between all processes was used to ensure that for any IB device (MLX5 common device) mempool event callback was registered only once and mempools that had been existing before the device start were traversed only once to register them. Since mempool callback regis

RE: [PATCH] memif: memif driver does not crashes when there's different N of TX and RX queues

Hi Huzaifa, This patch looks good to me. And would you please help review my memif patches? https://patches.dpdk.org/project/dpdk/cover/20220701102815.1444223-1-joyce.k...@arm.com/ Thanks, Joyce > -Original Message- > From: huzaifa.rahman > Sent: Tuesday, July 26, 2022 6:16 PM > To: jgr

[PATCH 0/1] update abi maintainership

In anticipation of a role change. I wanted to make sure there is orderly handover of maintainership of the abi. I think the proposed change reflects the reality of what is happening in the project in anycase. Ray Kinsella (1): devtools: eol abi as a separate function MAINTAINERS | 27 +++

[PATCH 1/1] devtools: eol abi as a separate function

Developer tools associated with abi are maintained with as part of developer tooling, eal abi headers are maintained with eal, abi build scripts are maintained with the build system and abi policy and version documents along with rest of the documentation. Major change is that individual component

[PATCH v2] net/tap: Allow jumbo frames

eth_dev_validate_mtu, introduced in 990912e676e, validates configured MTU plus overhead against max_rx_pktlen. Since TAP is a virtual device, it should support as big MTU as possible. --- drivers/net/tap/rte_eth_tap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net

RE: [PATCH v5 07/12] net/nfp: add flower ctrl VNIC related logics

> On 8/5/22 09:32, Chaoyong He wrote: > > This commit adds the setup/start logic for the ctrl vNIC. This vNIC > > "This commit adds" -> "Add" > > > is used by the PMD and flower firmware as a communication channel > > between driver and firmware. In the case of OVS it is also used to > > communic

[PATCH v3] net/tap: Allow jumbo frames

eth_dev_validate_mtu, introduced in 990912e676e, validates configured MTU plus overhead against max_rx_pktlen. Since TAP is a virtual device, it should support as big MTU as possible. --- drivers/net/tap/rte_eth_tap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/t

RE: [PATCH v5 10/12] net/nfp: add flower representor framework

> On 8/5/22 09:32, Chaoyong He wrote: > > This commit adds the framework to support flower representors. > > "This commit adds" -> "Add" > > > The number of VF representors are parsed from the command line. For > > physical port representors the current logic aims to create a > > representor for

[PATCH v4] net/tap: Allow jumbo frames

eth_dev_validate_mtu, introduced in 990912e676e, validates configured MTU plus overhead against max_rx_pktlen. Since TAP is a virtual device, it should support as big MTU as possible. --- drivers/net/tap/rte_eth_tap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net

[PATCH v5] net/tap: Allow jumbo frames

eth_dev_validate_mtu, introduced in 990912e676e, validates configured MTU plus overhead against max_rx_pktlen. Since TAP is a virtual device, it should support as big MTU as possible. --- drivers/net/tap/rte_eth_tap.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/net/

[PATCH v6] net/tap: Allow jumbo frames

eth_dev_validate_mtu, introduced in 990912e676e, validates configured MTU plus overhead against max_rx_pktlen. Since TAP is a virtual device, it should support as big MTU as possible. Signed-off-by: Francesco Mancino --- drivers/net/tap/rte_eth_tap.c | 2 +- 1 file changed, 1 insertion(+), 1 del

Re: [PATCH] net/tap: Allow jumbo frames

On Mon, 8 Aug 2022 11:31:16 +0200 Francesco Mancino wrote: > eth_dev_validate_mtu, introduced in 990912e676e, validates configured > MTU plus overhead against max_rx_pktlen. > Since TAP is a virtual device, it should support as big MTU as possible. > --- >  drivers/net/tap/rte_eth_tap.c | 2 +- >

Re: [PATCH v5] net/tap: Allow jumbo frames

On Mon, 8 Aug 2022 14:29:25 +0200 Francesco Mancino wrote: > eth_dev_validate_mtu, introduced in 990912e676e, validates configured > MTU plus overhead against max_rx_pktlen. > Since TAP is a virtual device, it should support as big MTU as possible. In all your tries at the patch, you must not ha

Re: [PATCH v5 07/12] net/nfp: add flower ctrl VNIC related logics

On Mon, 8 Aug 2022 11:32:30 + Chaoyong He wrote: > > > + goto done; > > > + > > > + /* Allocate memory for the eth_dev of the vNIC */ > > > + hw->eth_dev = rte_zmalloc("ctrl_vnic_eth_dev", > > > > Why not rte_eth_dev_allocate()? Isn't an ethdev? > > Why do you bypsss ethdev layer i

[PATCH v7] net/tap: Allow jumbo frames

eth_dev_validate_mtu, introduced in 990912e676e, validates configured MTU plus overhead against max_rx_pktlen. Since TAP is a virtual device, it should support as big MTU as possible. Signed-off-by: Francesco Mancino --- drivers/net/tap/rte_eth_tap.c | 2 +- 1 file changed, 1 insertion(+), 1 del

Re: [PATCH v7] net/tap: Allow jumbo frames

On Mon, 8 Aug 2022 16:49:44 +0200 Francesco Mancino wrote: > eth_dev_validate_mtu, introduced in 990912e676e, validates configured > MTU plus overhead against max_rx_pktlen. > Since TAP is a virtual device, it should support as big MTU as possible. > > Signed-off-by: Francesco Mancino d

Re: [PATCH v7] net/tap: Allow jumbo frames

Thank you for the feedback. I am sorry for the email spam, it was a learning process. I am not sure what results we would get by querying the MTU before configuring it, since we can (and do) set the MTU using SIOCSIFMTU. On 2022-08-08 17:03, Stephen Hemminger wrote: > On Mon, 8 Aug 2022 16:49:4

Re: [PATCH v7] net/tap: Allow jumbo frames

On Mon, 8 Aug 2022 17:38:21 +0200 Francesco Mancino wrote: > Thank you for the feedback. > > I am sorry for the email spam, it was a learning process. > > I am not sure what results we would get by querying the MTU > before configuring it, since we can (and do) set the MTU using > SIOCSIFMTU.

Re: [dpdk-kmods] windows/netuio: fix bar parsing

Hi Dmitry, On 8/6/2022 2:01 PM, Dmitry Kozlyuk wrote: 2022-08-02 12:10 (UTC-0700), Pallavi Kadam: For certain PCIe devices, BAR values are not continuous. This patch maps all the BARs and avoids skipping the next BAR addresses. Fixes: e28aabd88279 ("windows/netuio: introduce NetUIO kernel driv

[PATCH 0/3] cleanup bsf and fls inline function return types

The cleanup resulted from request to review [1] the following functions where there appeared to be inconsistency in return type or parameter type selections for the following inline functions. Since the original request some instances have been fixed so this series completes the outstanding return

[PATCH 3/3] test: fix sign compare warning for rte_bsf64 return type change

From: Tyler Retzlaff rte_bsf64 return type has been changed insert cast to suppress sign-comapre warning with int. Signed-off-by: Tyler Retzlaff --- app/test/test_mbuf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/app/test/test_mbuf.c b/app/test/test_mbuf.c index e09b25

[PATCH 2/3] eal: change rte_fls and rte_bsf to return uint32_t

From: Tyler Retzlaff return fixed width uint32_t to be consistent with what appears to be the original authors intent. it doesn't make much sense to return signed integers for these functions. Signed-off-by: Tyler Retzlaff --- lib/eal/include/rte_common.h | 6 +++--- 1 file changed, 3 insertio

[PATCH 1/3] doc: announce cleanup of rte_{bsf, fls} inline functions type use

From: Tyler Retzlaff The cleanup resulted from request to review [1] the following functions where there appeared to be inconsistency in return type or parameter type selections for the following inline functions. rte_bsf32() rte_bsf32_safe() rte_bsf64() rte_bsf64_safe() rte_fls_u32() rte_fls_u6

Re: [PATCH 0/3] cleanup bsf and fls inline function return types

On Mon, 8 Aug 2022 14:21:29 -0700 Tyler Retzlaff wrote: > The cleanup resulted from request to review [1] the following > functions where there appeared to be inconsistency in return type > or parameter type selections for the following inline functions. > > Since the original request some inst

Re: [PATCH v7] ip_frag: add IPv4 fragment copy packet API

08/08/2022 02:48, Huichao Cai пишет: Some NIC drivers support MBUF_FAST_FREE(Device supports optimization for fast release of mbufs. When set application must guarantee that per-queue all mbufs comes from the same mempool,has refcnt = 1,direct and non-segmented.)offload. In order to adapt to this

Re: [dpdk-kmods] windows/netuio: fix bar parsing

Hi Dmitry, On 8/6/2022 2:01 PM, Dmitry Kozlyuk wrote: 2022-08-02 12:10 (UTC-0700), Pallavi Kadam: For certain PCIe devices, BAR values are not continuous. This patch maps all the BARs and avoids skipping the next BAR addresses. Fixes: e28aabd88279 ("windows/netuio: introduce NetUIO kernel driv

[PATCH 0/2] support flow subscription

Add support AVF can be able to subscribe a flow from PF. Signed-off-by: Jie Wang Jie Wang (2): common/iavf: support flow subscription net/iavf: enable flow subscription rule support for AVF doc/guides/rel_notes/release_22_11.rst | 4 + drivers/common/iavf/virtchnl.h | 104 +++- d

[PATCH 1/2] common/iavf: support flow subscription

VF is able to subscribe a flow from PF by VIRTCHNL_FLOW_SUBSCRIBE. PF is expected to offload a rule to hardware which will redirect the packet that matching the required pattern to this VF. Only a flow with dst mac address as PF's mac address can be subscribed. VIRTCHNL_VF_OFFLOAD_FSUB_PF is use

[PATCH 2/2] net/iavf: enable flow subscription rule support for AVF

Add support AVF can be able to subscribe a flow from PF. The supported patterns are listed below: eth/vlan/ipv4 eth/ipv4(6) eth/ipv4(6)/udp eth/ipv4(6)/tcp Signed-off-by: Jie Wang --- doc/guides/rel_notes/release_22_11.rst | 4 + drivers/net/iavf/iavf.h| 10 + drivers/net/iav