[dpdk-dev] [PATCH] ci: enable vm env for aarch64 builds
Tests requiring hugepages do not work outside of VM environment because of security limitations. Enable aarch64 builds which run tests to run in a VM to avoid these limitations. Signed-off-by: Juraj Linkeš --- .travis.yml | 10 -- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/.travis.yml b/.travis.yml index d6eeab371..d73d3e4c5 100644 --- a/.travis.yml +++ b/.travis.yml @@ -88,7 +88,10 @@ jobs: arch: arm64 compiler: gcc - env: DEF_LIB="shared" RUN_TESTS=1 -arch: arm64 +dist: focal +arch: arm64-graviton2 +virt: vm +group: edge compiler: gcc - env: DEF_LIB="shared" BUILD_DOCS=1 arch: arm64 @@ -103,5 +106,8 @@ jobs: arch: arm64 compiler: clang - env: DEF_LIB="shared" RUN_TESTS=1 -arch: arm64 +dist: focal +arch: arm64-graviton2 +virt: vm +group: edge compiler: clang -- 2.20.1
[dpdk-dev] [PATCH] ci: enable vm env for aarch64 builds
Tests requiring hugepages do not work outside of VM environment because of security limitations. Enable aarch64 builds which run tests to run in a VM to avoid these limitations. Signed-off-by: Juraj Linkeš --- .travis.yml | 10 -- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/.travis.yml b/.travis.yml index d6eeab371..d73d3e4c5 100644 --- a/.travis.yml +++ b/.travis.yml @@ -88,7 +88,10 @@ jobs: arch: arm64 compiler: gcc - env: DEF_LIB="shared" RUN_TESTS=1 -arch: arm64 +dist: focal +arch: arm64-graviton2 +virt: vm +group: edge compiler: gcc - env: DEF_LIB="shared" BUILD_DOCS=1 arch: arm64 @@ -103,5 +106,8 @@ jobs: arch: arm64 compiler: clang - env: DEF_LIB="shared" RUN_TESTS=1 -arch: arm64 +dist: focal +arch: arm64-graviton2 +virt: vm +group: edge compiler: clang -- 2.20.1
Re: [dpdk-dev] [PATCH 2/2] net/ark: remove RTE_LIBRTE_ARK_PAD_TX configuration macro
On 8/24/2020 10:51 PM, Ed Czeck wrote: > On Mon, Aug 24, 2020 at 10:55 AM Ferruh Yigit wrote: >> >> On 8/24/2020 2:36 PM, Ed Czeck wrote: >>> Replace behavior with RTE_LIBRTE_ARK_MIN_TX_PKTLEN >>> with a default value of 0. >>> Update documentation as needed. >>> >>> Signed-off-by: Ed Czeck >>> --- >>> doc/guides/nics/ark.rst | 13 -- >>> drivers/net/ark/ark_ethdev_tx.c | 44 +++-- >>> drivers/net/ark/ark_logs.h | 8 -- >>> 3 files changed, 36 insertions(+), 29 deletions(-) >>> >>> diff --git a/doc/guides/nics/ark.rst b/doc/guides/nics/ark.rst >>> index c3ffcbbc2..f4d4539e5 100644 >>> --- a/doc/guides/nics/ark.rst >>> +++ b/doc/guides/nics/ark.rst >>> @@ -129,8 +129,10 @@ Configuration Information >>> * **CONFIG_RTE_LIBRTE_ARK_PMD** (default y): Enables or disables >>> inclusion >>> of the ARK PMD driver in the DPDK compilation. >>> >>> - * **CONFIG_RTE_LIBRTE_ARK_PAD_TX** (default y): When enabled TX >>> - packets are padded to 60 bytes to support downstream MACS. >>> + * **CONFIG_RTE_LIBRTE_ARK_MIN_TX_PKTLEN** (default 0): Sets the minimum >>> +packet length for tx packets to the FPGA. Packets less than this >>> +length are padded to meet the requirement. This allows padding to >>> +be offloaded or remain in host software. >>> >> >> There is no 'CONFIG_RTE_LIBRTE_ARK_MIN_TX_PKTLEN' option, I think it is >> confusing to add it to documentation. There is already some documentation in >> below for 'RTE_LIBRTE_ARK_MIN_TX_PKTLEN', what do you think removing this >> part? >> > Should I remove the all references to the CONFIG_RTE_LIBRTE_ARK from > the net/ark section of the guide? It appears that CONFIG_.* will be > removed along with the makefile build system. As you said it is removed by Ciara's patch, no need to touch it in this patch. > >> And can you please remove 'CONFIG_RTE_LIBRTE_ARK_PAD_TX' form >> 'config/common_base'? >> > I am hesitant to modify config/common_base, since it is slated for > removal and edits will cause merge conflicts. Are you OK with this? OK. > >> Another option is, do you think does it help your users to keep the config >> name >> same, like keep the 'CONFIG_RTE_LIBRTE_ARK_PAD_TX' as 'RTE_LIBRTE_ARK_PAD_TX' >> for the meson CFLAGs, instead of renaming it to >> 'RTE_LIBRTE_ARK_MIN_TX_PKTLEN'? > I will drop the CONFIG names and leave the compiler marco documented > in my next patch. > OK, thanks.
Re: [dpdk-dev] [PATCH 1/2] net/ark: remove compile time log macros in favor of run time log control
On 8/24/2020 10:40 PM, Ed Czeck wrote: > On Mon, Aug 24, 2020 at 11:09 AM Ferruh Yigit wrote: >> >> On 8/24/2020 3:40 PM, Bruce Richardson wrote: >>> On Mon, Aug 24, 2020 at 03:37:53PM +0100, Ferruh Yigit wrote: On 8/24/2020 2:36 PM, Ed Czeck wrote: > Use ARK_PMD_LOG in place of PMD_DRV_LOG, PMD_DEBUG_LOG, PMD_FUNC_LOG, > PMD_STATS_LOG, PMD_RX_LOG, and PMD_TX_LOG. > Review and adjust log levels and messages as needed. > > Signed-off-by: Ed Czeck > --- > doc/guides/nics/ark.rst | 13 -- > drivers/net/ark/ark_ddm.c | 12 ++--- > drivers/net/ark/ark_ethdev.c| 77 +++-- > drivers/net/ark/ark_ethdev_rx.c | 46 ++-- > drivers/net/ark/ark_ethdev_tx.c | 10 ++--- > drivers/net/ark/ark_logs.h | 63 +++ > drivers/net/ark/ark_mpu.c | 32 -- > drivers/net/ark/ark_pktchkr.c | 36 +++ > drivers/net/ark/ark_pktdir.c| 2 +- > drivers/net/ark/ark_pktgen.c| 22 +- > drivers/net/ark/ark_rqp.c | 4 +- > drivers/net/ark/ark_udm.c | 18 > 12 files changed, 125 insertions(+), 210 deletions(-) > > diff --git a/doc/guides/nics/ark.rst b/doc/guides/nics/ark.rst > index 06e8c3374..c3ffcbbc2 100644 > --- a/doc/guides/nics/ark.rst > +++ b/doc/guides/nics/ark.rst > @@ -132,19 +132,6 @@ Configuration Information > * **CONFIG_RTE_LIBRTE_ARK_PAD_TX** (default y): When enabled TX > packets are padded to 60 bytes to support downstream MACS. > > - * **CONFIG_RTE_LIBRTE_ARK_DEBUG_RX** (default n): Enables or disables > debug > - logging and internal checking of RX ingress logic within the ARK > PMD driver. > - > - * **CONFIG_RTE_LIBRTE_ARK_DEBUG_TX** (default n): Enables or disables > debug > - logging and internal checking of TX egress logic within the ARK PMD > driver. > - > - * **CONFIG_RTE_LIBRTE_ARK_DEBUG_STATS** (default n): Enables or > disables debug > - logging of detailed packet and performance statistics gathered in > - the PMD and FPGA. > - > - * **CONFIG_RTE_LIBRTE_ARK_DEBUG_TRACE** (default n): Enables or > disables debug > - logging of detailed PMD events and status. > - Can you please remove these config option from 'config/common_base'? >>> Given that the file is scheduled for deletion fairly soon, not sure if that >>> is worthwhile doing. >>> >> >> Since there is possible new version already because of comment in 2/2, I >> think >> better to update 'config/common_base' for the sake of completeness. >> >> But if there won't be a new version, this can be skipped pragmatically. > > I will avoid the update of config/common_base as changes there will > cause merge conflict. > OK
Re: [dpdk-dev] [PATCH 1/2] doc/linux_gsg: clarify instructions on running as non-root
On 8/24/2020 4:45 PM, Anatoly Burakov wrote: > The current instructions are slightly out of date when it comes to > providing information about setting up the system for using DPDK as > non-root, so update them. > > Cc: sta...@dpdk.org > > Signed-off-by: Anatoly Burakov Thanks for the doc update, it is useful. I did able to run testpmd as non-root using vfio-pci module. For series, Reviewed-by: Ferruh Yigit
Re: [dpdk-dev] [PATCH] cryptodev: revert ABI compatibility for ChaCha20-Poly1305
-Original Message-
From: Dybkowski, AdamX
Sent: piątek, 14 sierpnia 2020 12:00
To: dev@dpdk.org; Trahe, Fiona ; akhil.go...@nxp.com
Cc: Dybkowski, AdamX ; Kusztal, ArkadiuszX
Subject: [PATCH] cryptodev: revert ABI compatibility for ChaCha20-Poly1305
This reverts commit a0f0de06d457753c94688d551a6e8659b4d4e041 as the
rte_cryptodev_info_get function versioning was a temporary solution to maintain
ABI compatibility for ChaCha20-Poly1305 and is not needed in 20.11.
Fixes: a0f0de06d457 ("cryptodev: fix ABI compatibility for ChaCha20-Poly1305")
Signed-off-by: Adam Dybkowski
Reviewed-by: Arek Kusztal
---
lib/librte_cryptodev/meson.build | 1 -
lib/librte_cryptodev/rte_cryptodev.c | 147 +-
lib/librte_cryptodev/rte_cryptodev.h | 34 +---
.../rte_cryptodev_version.map | 6 -
4 files changed, 6 insertions(+), 182 deletions(-)
Acked-by: Arek Kusztal
Re: [dpdk-dev] [PATCH v4 3/3] ethdev: check if queue setupped when getting queue info
On 8/24/2020 12:01 PM, Wei Hu (Xavier) wrote: > From: "Wei Hu (Xavier)" > > This patch adds checking whether the related Tx or Rx queue has been > setuped in the rte_eth_rx_queue_info_get and rte_eth_tx_queue_info_get > API function to avoid illegal address access. > > Signed-off-by: Wei Hu (Xavier) Reviewed-by: Ferruh Yigit
Re: [dpdk-dev] [PATCH v4 0/3] minor updates for getting queue info
On 8/24/2020 12:01 PM, Wei Hu (Xavier) wrote: > This series include 3 patches. > One patch is a minor update to add checking whether the related Tx or Rx > queue has been setuped in the rte_eth_rx_queue_info_get and > rte_eth_tx_queue_info_get API function to avoid illegal address access. > > And the other two patches for hns3 PMD driver. > > Huisong Li (1): > net/hns3: support getting Tx and Rx queue information > > Wei Hu (Xavier) (2): > net/hns3: report Rx drop packets enable configuration > ethdev: check if queue setupped when getting queue info > Series applied to dpdk-next-net/main, thanks.
Re: [dpdk-dev] [PATCH 1/2] drivers: replace RTE CIO barriers with RTE IO barriers
Hi Honnappa,
> Hi Phil,
> Can you please add a statement in release notes indicating that these
> APIs are removed.
Sure. I will add it to the 2/2 patch in the new version.
Thanks,
Phil
>
> Thank you,
> Honnappa
>
> > -Original Message-
> > From: Phil Yang
> > Sent: Monday, August 24, 2020 3:41 AM
> > To: dev@dpdk.org
> > Cc: jer...@marvell.com; d...@linux.vnet.ibm.com; Ajit Khaparde
> > (ajit.khapa...@broadcom.com) ;
> > konstantin.anan...@intel.com; Honnappa Nagarahalli
> > ; Joyce Kong ;
> nd
> > ; Matan Azrad ; Shahaf Shuler
> > ; Viacheslav Ovsiienko
> > ; Ankur Dwivedi ;
> > Anoob Joseph ; Pavan Nikhilesh
> > ; Somnath Kotur
> > ; Wei Zhao ; Jeff
> > Guo ; Beilei Xing ; Ruifeng
> Wang
> > ; Harman Kalra ; Nithin
> > Dabilpuram ; Kiran Kumar K
> > ; Maxime Coquelin
> > ; Chenbo Xia ;
> > Zhihong Wang ; Mahipal Challa
> > ; Ori Kam
> > Subject: [PATCH 1/2] drivers: replace RTE CIO barriers with RTE IO barriers
> >
> > Replace the deprecated rte_cio_[rw]mb barriers with rte_io_[rw]mb
> barriers.
> >
> > Signed-off-by: Phil Yang
> > Signed-off-by: Joyce Kong
> > ---
> > drivers/common/mlx5/mlx5_common.h | 2 +-
> > drivers/crypto/octeontx2/otx2_cryptodev_ops.c | 2 +-
> > drivers/crypto/octeontx2/otx2_cryptodev_sec.c | 4 ++--
> > drivers/event/octeontx/ssovf_worker.c | 6 +++---
> > drivers/event/octeontx2/otx2_worker.h | 2 +-
> > drivers/net/bnxt/bnxt_hwrm.c | 2 +-
> > drivers/net/bnxt/bnxt_ring.h | 6 +++---
> > drivers/net/e1000/em_rxtx.c | 2 +-
> > drivers/net/i40e/i40e_rxtx.c | 2 +-
> > drivers/net/i40e/i40e_rxtx_vec_neon.c | 4 ++--
> > drivers/net/mlx5/mlx5_flow.c | 2 +-
> > drivers/net/mlx5/mlx5_flow_dv.c | 2 +-
> > drivers/net/mlx5/mlx5_rxq.c | 12 ++--
> > drivers/net/mlx5/mlx5_rxtx.c | 16
> > drivers/net/mlx5/mlx5_rxtx.h | 2 +-
> > drivers/net/mlx5/mlx5_rxtx_vec.h | 2 +-
> > drivers/net/mlx5/mlx5_rxtx_vec_altivec.h | 2 +-
> > drivers/net/mlx5/mlx5_rxtx_vec_neon.h | 4 ++--
> > drivers/net/mlx5/mlx5_rxtx_vec_sse.h | 2 +-
> > drivers/net/mlx5/mlx5_txq.c | 4 ++--
> > drivers/net/octeontx/octeontx_rxtx.h | 2 +-
> > drivers/net/octeontx2/otx2_ethdev_sec.c | 2 +-
> > drivers/net/octeontx2/otx2_ethdev_sec_tx.h| 2 +-
> > drivers/net/octeontx2/otx2_rx.c | 2 +-
> > drivers/net/octeontx2/otx2_tx.c | 6 +++---
> > drivers/net/virtio/virtio_rxtx.c | 2 +-
> > drivers/net/virtio/virtio_rxtx_simple_neon.c | 2 +-
> > drivers/net/virtio/virtqueue.h| 26
> > +-
> > drivers/raw/octeontx2_ep/otx2_ep_enqdeq.c | 4 ++--
> > drivers/regex/mlx5/mlx5_regex_fastpath.c | 4 ++--
> > 30 files changed, 66 insertions(+), 66 deletions(-)
> >
> > diff --git a/drivers/common/mlx5/mlx5_common.h
> > b/drivers/common/mlx5/mlx5_common.h
> > index 2cdb226..ed44a45 100644
> > --- a/drivers/common/mlx5/mlx5_common.h
> > +++ b/drivers/common/mlx5/mlx5_common.h
> > @@ -193,7 +193,7 @@ check_cqe(volatile struct mlx5_cqe *cqe, const
> > uint16_t cqes_n,
> >
> > if (unlikely((op_owner != (!!(idx))) || (op_code ==
> > MLX5_CQE_INVALID)))
> > return MLX5_CQE_STATUS_HW_OWN;
> > - rte_cio_rmb();
> > + rte_io_rmb();
> > if (unlikely(op_code == MLX5_CQE_RESP_ERR ||
> > op_code == MLX5_CQE_REQ_ERR))
> > return MLX5_CQE_STATUS_ERR;
> > diff --git a/drivers/crypto/octeontx2/otx2_cryptodev_ops.c
> > b/drivers/crypto/octeontx2/otx2_cryptodev_ops.c
> > index 9d51b17..df39cde 100644
> > --- a/drivers/crypto/octeontx2/otx2_cryptodev_ops.c
> > +++ b/drivers/crypto/octeontx2/otx2_cryptodev_ops.c
> > @@ -469,7 +469,7 @@ otx2_cpt_enqueue_req(const struct otx2_cpt_qp
> *qp,
> > * buffer immediately, a DMB is not required to push out
> > * LMTSTs.
> > */
> > - rte_cio_wmb();
> > + rte_io_wmb();
> > lmt_status = otx2_lmt_submit(qp->lf_nq_reg);
> > } while (lmt_status == 0);
> >
> > diff --git a/drivers/crypto/octeontx2/otx2_cryptodev_sec.c
> > b/drivers/crypto/octeontx2/otx2_cryptodev_sec.c
> > index 0741a59..72e6c41 100644
> > --- a/drivers/crypto/octeontx2/otx2_cryptodev_sec.c
> > +++ b/drivers/crypto/octeontx2/otx2_cryptodev_sec.c
> > @@ -107,7 +107,7 @@ otx2_cpt_enq_sa_write(struct
> > otx2_sec_session_ipsec_lp *lp,
> > inst.u64[3] = 0;
> > inst.res_addr = rte_mempool_virt2iova(res);
> >
> > - rte_cio_wmb();
> > + rte_io_wmb();
> >
> > do {
> > /* Copy CPT command to LMTLINE */
> > @@ -124,7 +124,7 @@ otx2_cpt_enq_sa_write(struct
> > otx2_sec_session_ipsec_lp *lp,
> > otx2_err("Request timed out");
> > return -ETIMEDOUT;
>
Re: [dpdk-dev] 19.11.4 patches review and test
Hello Luca, Testing with dpdk 19.11.4-rc1 from Red Hat looks good. We cover below 16 scenarios and and all get PASS on RHEL8 testing: (1)Guest with device assignment(PF) throughput testing(1G hugepage size): PASS (2)Guest with device assignment(PF) throughput testing(2M hugepage size) : PASS (3)Guest with device assignment(VF) throughput testing: PASS (4)PVP (host dpdk testpmd as vswitch) 1Q: throughput testing: PASS (5)PVP vhost-user 2Q throughput testing: PASS (6)PVP vhost-user 1Q - cross numa node throughput testing: PASS (7)Guest with vhost-user 2 queues throughput testing: PASS (8)vhost-user reconnect with dpdk-client, qemu-server: qemu reconnect: PASS (9)PVP 1Q live migration testing: PASS (10)PVP 1Q cross numa node live migration testing: PASS (11)Guest with ovs+dpdk+vhost-user 1Q live migration testing: PASS (12)Guest with ovs+dpdk+vhost-user 1Q live migration testing (2M): PASS (13)Guest with ovs+dpdk+vhost-user 2Q live migration testing: PASS (14) Allocate memory from the NUMA node which Virtio device locates: PASS (15)Host PF + DPDK testing: PASS (16)Host VF + DPDK testing: PASS Versions: kernel 4.18 qemu 5.1 dpdk: git://dpdk.org/dpdk-stable remotes/origin/19.11 # git log commit bb00cdf6cdcf690001ef4154a5aaad8c0dc5d9b9 (HEAD, tag: v19.11.4-rc1, origin/19.11) Author: Luca Boccassi Date: Tue Aug 18 19:00:36 2020 +0100 version: 19.11.4-rc1 Signed-off-by: Luca Boccassi NICs: X540-AT2 NIC(ixgbe, 10G) Best regards, Pei - Original Message - > From: "Luca Boccassi" > To: sta...@dpdk.org > Cc: dev@dpdk.org, "Abhishek Marathe" , "Akhil > Goyal" , "Ali > Alnubani" , "benjamin walker" > , "David Christensen" > , "Hemant Agrawal" , "Ian > Stokes" , "Jerin > Jacob" , "John McNamara" , > "Ju-Hyoung Lee" , > "Kevin Traynor" , "Pei Zhang" , > "pingx yu" , "qian q > xu" , "Raslan Darawsheh" , "Thomas > Monjalon" , "yuan > peng" , "zhaoyan chen" > Sent: Wednesday, August 19, 2020 2:12:22 AM > Subject: 19.11.4 patches review and test > > Hi all, > > Here is a list of patches targeted for stable release 19.11.4. > > The planned date for the final release is August 31st. > > Please help with testing and validation of your use cases and report > any issues/results with reply-all to this mail. For the final release > the fixes and reported validations will be added to the release notes. > > A release candidate tarball can be found at: > > https://dpdk.org/browse/dpdk-stable/tag/?id=v19.11.4-rc1 > > These patches are located at branch 19.11 of dpdk-stable repo: > https://dpdk.org/browse/dpdk-stable/ > > Thanks. > > Luca Boccassi > > --- > Adam Dybkowski (7): > test/crypto: fix asymmetric session mempool creation > crypto/qat: fix AES-XTS capabilities > crypto/qat: handle mixed hash-cipher requests on GEN3 > test/crypto: add mixed encypted-digest > common/qat: get firmware version > crypto/qat: handle mixed hash-cipher on GEN2 > common/qat: fix uninitialized variable > > Akhil Goyal (4): > crypto/dpaax_sec: fix 18-bit PDCP cases with HFN override > crypto/dpaax_sec: fix inline query for descriptors > crypto/dpaa2_sec: fix HFN override > common/dpaax: fix 12-bit null auth case > > Alan Dewar (1): > sched: fix port time rounding > > Alexander Kozyrev (3): > net/mlx5: fix vectorized Rx burst termination > net/mlx5: do not select legacy MPW implicitly > net/mlx5: fix descriptors number adjustment > > Ali Alnubani (2): > net/mlx5: fix typos in meter error messages > doc: fix a typo in mlx5 guide > > Alvin Zhang (1): > net/i40e: fix queue pairs configuration in VF > > Andrew Rybchenko (1): > net/sfc: do not enforce hash offload in RSS multi-queue > > Ankur Dwivedi (2): > common/cpt: fix encryption offset > test/crypto: change cipher offset for ESN vector > > Apeksha Gupta (2): > app/eventdev: fix capability check in pipeline ATQ test > event/dpaa2: add all-types queue capability flag > > Archana Muniganti (2): > examples/fips_validation: fix TDES interim callback > examples/fips_validation: fix count overwrite for TDES > > Archit Pandey (1): > sched: fix 64-bit rate > > Arek Kusztal (4): > drivers/qat: improve multi process on qat > drivers/qat: add multi process handling of driver id > drivers/qat: add handling of capabilities in multi process > crypto/qat: add minimum enq threshold > > Ayuj Verma (1): > examples/fips_validation: fix parsing of TDES vectors > > Bruce Richardson (12): > test: fix build with ring PMD but no bond PMD > mem: fix 32-bit init config with meson > rawdev: allow getting info for unknown device > rawdev: fill NUMA socket ID in info > rawdev: export dump function in map file > doc: rebuild with meson whenever a file changes > build: always link whole DPDK static libraries > devtools: test static li
Re: [dpdk-dev] [dpdk-stable] [PATCH] net/hinic: fix negative array index read
On 8/24/2020 1:54 PM, Ferruh Yigit wrote:
> On 8/24/2020 1:43 PM, wangyunjian wrote:
>> From: Yunjian Wang
>>
>> Negative array index read using variable 'i' as an index to array
>> 'filter_info->pkt_filters'. Fixed by add return value check.
>>
>> Coverity issue: 350364
>> Fixes: f4ca3fd54c4d ("net/hinic: create and destroy flow director filter")
>> Cc: sta...@dpdk.org
>>
>> Signed-off-by: Yunjian Wang
>
> Reviewed-by: Ferruh Yigit
>
Applied to dpdk-next-net/main, thanks.
[dpdk-dev] [v1] vdpa/mlx5: fix event channel setup
During vdap device setup, if some error happens, event channel release
stuck at polling event channel.
Event channel fd is set to nonblocking in cqe setup, so if any error
happens before this function and after event channel created, the
pooling before releasing resources will stuck.
This patch moves event channel to non-blocking mode right after
creation.
Fixes: 8395927cdf ("vdpa/mlx5: prepare HW queues")
Cc: ma...@nvidia.com
Signed-off-by: Xueming Li
---
drivers/vdpa/mlx5/mlx5_vdpa_event.c | 15 ---
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/drivers/vdpa/mlx5/mlx5_vdpa_event.c
b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
index 5a2d4fb1ec..bda547ffe0 100644
--- a/drivers/vdpa/mlx5/mlx5_vdpa_event.c
+++ b/drivers/vdpa/mlx5/mlx5_vdpa_event.c
@@ -51,6 +51,8 @@ mlx5_vdpa_event_qp_global_release(struct mlx5_vdpa_priv *priv)
static int
mlx5_vdpa_event_qp_global_prepare(struct mlx5_vdpa_priv *priv)
{
+ int flags, ret;
+
if (priv->eventc)
return 0;
if (mlx5_glue->devx_query_eqn(priv->ctx, 0, &priv->eqn)) {
@@ -66,6 +68,12 @@ mlx5_vdpa_event_qp_global_prepare(struct mlx5_vdpa_priv
*priv)
rte_errno);
goto error;
}
+ flags = fcntl(priv->eventc->fd, F_GETFL);
+ ret = fcntl(priv->eventc->fd, F_SETFL, flags | O_NONBLOCK);
+ if (ret) {
+ DRV_LOG(ERR, "Failed to change event channel FD.");
+ goto error;
+ }
priv->uar = mlx5_glue->devx_alloc_uar(priv->ctx, 0);
if (!priv->uar) {
rte_errno = errno;
@@ -376,7 +384,6 @@ mlx5_vdpa_interrupt_handler(void *cb_arg)
int
mlx5_vdpa_cqe_event_setup(struct mlx5_vdpa_priv *priv)
{
- int flags;
int ret;
if (!priv->eventc)
@@ -393,12 +400,6 @@ mlx5_vdpa_cqe_event_setup(struct mlx5_vdpa_priv *priv)
return -1;
}
}
- flags = fcntl(priv->eventc->fd, F_GETFL);
- ret = fcntl(priv->eventc->fd, F_SETFL, flags | O_NONBLOCK);
- if (ret) {
- DRV_LOG(ERR, "Failed to change event channel FD.");
- goto error;
- }
priv->intr_handle.fd = priv->eventc->fd;
priv->intr_handle.type = RTE_INTR_HANDLE_EXT;
if (rte_intr_callback_register(&priv->intr_handle,
--
2.17.1
Re: [dpdk-dev] [PATCH 1/1] vhost: fix iotlb mempool single-consumer flag
On 10/08/2020 15:11, Eugenio Pérez wrote: > Bugzilla bug: 523 > > Using testpmd as a vhost-user with iommu: > > /home/dpdk/build/app/dpdk-testpmd -l 1,3 \ > --vdev net_vhost0,iface=/tmp/vhost-user1,queues=1,iommu-support=1 \ > -- --auto-start --stats-period 5 --forward-mode=txonly > > And qemu with packed virtqueue: > > > > > > >function='0x0'/> > > ... > > > > > > > -- > > Is it possible to consume the iotlb's entries of the mempoo from different > threads. Thread sanitizer example output (after change rwlocks to POSIX ones): > > WARNING: ThreadSanitizer: data race (pid=76927) > Write of size 8 at 0x00017ffd5628 by thread T5: > #0 vhost_user_iotlb_cache_insert ../lib/librte_vhost/iotlb.c:181 > (dpdk-testpmd+0x769343) > #1 vhost_user_iotlb_msg ../lib/librte_vhost/vhost_user.c:2380 > (dpdk-testpmd+0x78e4bf) > #2 vhost_user_msg_handler ../lib/librte_vhost/vhost_user.c:2848 > (dpdk-testpmd+0x78fcf8) > #3 vhost_user_read_cb ../lib/librte_vhost/socket.c:311 > (dpdk-testpmd+0x770162) > #4 fdset_event_dispatch ../lib/librte_vhost/fd_man.c:286 > (dpdk-testpmd+0x7591c2) > #5 ctrl_thread_init ../lib/librte_eal/common/eal_common_thread.c:193 > (dpdk-testpmd+0xa2890b) > #6 (libtsan.so.0+0x2a68d) > > Previous read of size 8 at 0x00017ffd5628 by thread T3: > #0 vhost_user_iotlb_cache_find ../lib/librte_vhost/iotlb.c:252 > (dpdk-testpmd+0x76ee96) > #1 __vhost_iova_to_vva ../lib/librte_vhost/vhost.c:42 > (dpdk-testpmd+0x77488c) > #2 vhost_iova_to_vva ../lib/librte_vhost/vhost.h:753 > (dpdk-testpmd+0x7abeb3) > #3 map_one_desc ../lib/librte_vhost/virtio_net.c:497 > (dpdk-testpmd+0x7abeb3) > #4 fill_vec_buf_packed ../lib/librte_vhost/virtio_net.c:751 > (dpdk-testpmd+0x7abeb3) > #5 vhost_enqueue_single_packed ../lib/librte_vhost/virtio_net.c:1170 > (dpdk-testpmd+0x7abeb3) > #6 virtio_dev_rx_single_packed ../lib/librte_vhost/virtio_net.c:1346 > (dpdk-testpmd+0x7abeb3) > #7 virtio_dev_rx_packed ../lib/librte_vhost/virtio_net.c:1384 > (dpdk-testpmd+0x7abeb3) > #8 virtio_dev_rx ../lib/librte_vhost/virtio_net.c:1435 > (dpdk-testpmd+0x7b0654) > #9 rte_vhost_enqueue_burst ../lib/librte_vhost/virtio_net.c:1465 > (dpdk-testpmd+0x7b0654) > #10 eth_vhost_tx ../drivers/net/vhost/rte_eth_vhost.c:470 > (dpdk-testpmd+0x1ddfbd8) > #11 rte_eth_tx_burst ../lib/librte_ethdev/rte_ethdev.h:4800 > (dpdk-testpmd+0x505fdb) > #12 pkt_burst_transmit ../app/test-pmd/txonly.c:365 > (dpdk-testpmd+0x5106ad) > #13 run_pkt_fwd_on_lcore ../app/test-pmd/testpmd.c:2080 > (dpdk-testpmd+0x4f8951) > #14 start_pkt_forward_on_core ../app/test-pmd/testpmd.c:2106 > (dpdk-testpmd+0x4f89d7) > #15 eal_thread_loop ../lib/librte_eal/linux/eal_thread.c:127 > (dpdk-testpmd+0xa5b20a) > #16 (libtsan.so.0+0x2a68d) > > Location is global '' at 0x (rtemap_0+0x3ffd5628) > > Thread T5 'vhost-events' (tid=76933, running) created by main thread at: > #0 pthread_create (libtsan.so.0+0x2cd42) > #1 rte_ctrl_thread_create > ../lib/librte_eal/common/eal_common_thread.c:216 (dpdk-testpmd+0xa289e7) > #2 rte_vhost_driver_start ../lib/librte_vhost/socket.c:1190 > (dpdk-testpmd+0x7728ef) > #3 vhost_driver_setup ../drivers/net/vhost/rte_eth_vhost.c:1028 > (dpdk-testpmd+0x1de233d) > #4 eth_dev_configure ../drivers/net/vhost/rte_eth_vhost.c:1126 > (dpdk-testpmd+0x1de29cc) > #5 rte_eth_dev_configure ../lib/librte_ethdev/rte_ethdev.c:1439 > (dpdk-testpmd+0x991ce2) > #6 start_port ../app/test-pmd/testpmd.c:2450 (dpdk-testpmd+0x4f9b45) > #7 main ../app/test-pmd/testpmd.c:3777 (dpdk-testpmd+0x4fe1ac) > > Thread T3 'lcore-slave-3' (tid=76931, running) created by main thread at: > #0 pthread_create (libtsan.so.0+0x2cd42) > #1 rte_eal_init ../lib/librte_eal/linux/eal.c:1244 (dpdk-testpmd+0xa46e2b) > #2 main ../app/test-pmd/testpmd.c:3673 (dpdk-testpmd+0x4fdd75) > > -- > > Or: > WARNING: ThreadSanitizer: data race (pid=76927) > Write of size 1 at 0x00017ffd00f8 by thread T5: > #0 vhost_user_iotlb_cache_insert ../lib/librte_vhost/iotlb.c:182 > (dpdk-testpmd+0x769370) > #1 vhost_user_iotlb_msg ../lib/librte_vhost/vhost_user.c:2380 > (dpdk-testpmd+0x78e4bf) > #2 vhost_user_msg_handler ../lib/librte_vhost/vhost_user.c:2848 > (dpdk-testpmd+0x78fcf8) > #3 vhost_user_read_cb ../lib/librte_vhost/socket.c:311 > (dpdk-testpmd+0x770162) > #4 fdset_event_dispatch ../lib/librte_vhost/fd_man.c:286 > (dpdk-testpmd+0x7591c2) > #5 ctrl_thread_init ../lib/librte_eal/common/eal_common_thread.c:193 > (dpdk-testpmd+0xa2890b) > #6 (libtsan.so.0+0x2a68d) > > Previous write of size 1 at 0x00017ffd00f8 by thread T3: > #0 vhost_user_iotlb_pending_insert ../lib/librte_vhost/iotlb.c:86 > (dpdk-testpmd+0x75eb0c) > #1 __vhost_iova_to_vva ../lib/librte_vhost/vhost.c:58 > (dpdk-testpmd+0x774926) >
Re: [dpdk-dev] [PATCH 2/2] doc/linux_gsg: update information on using hugepages
On 24-Aug-20 6:13 PM, Bruce Richardson wrote: On Mon, Aug 24, 2020 at 04:45:01PM +0100, Anatoly Burakov wrote: Current information regarding hugepage usage is a little out of date. Update it to include information on in-memory mode, as well as on default mountpoints provided by systemd. Cc: sta...@dpdk.org Signed-off-by: Anatoly Burakov --- doc/guides/linux_gsg/sys_reqs.rst | 39 +++ 1 file changed, 24 insertions(+), 15 deletions(-) diff --git a/doc/guides/linux_gsg/sys_reqs.rst b/doc/guides/linux_gsg/sys_reqs.rst index a124656bcb..2ddd7ed667 100644 --- a/doc/guides/linux_gsg/sys_reqs.rst +++ b/doc/guides/linux_gsg/sys_reqs.rst @@ -155,8 +155,12 @@ Without hugepages, high TLB miss rates would occur with the standard 4k page siz Reserving Hugepages for DPDK Use -The allocation of hugepages should be done at boot time or as soon as possible after system boot -to prevent memory from being fragmented in physical memory. +The allocation of hugepages can be performed either at run time or at boot time. +In the general case, reserving hugepages at run time is perfectly fine, but in +use cases where having lots of physically contiguous memory is required, it is +preferable to reserve hugepages at boot time, as that will help in preventing +physical memory from becoming heavily fragmented. + Although we are removing the note about 1G pages requiring to be reserved at boot time, I think we should still mention here that some older kernel versions do not allow 1G reservations post-boot. Agreed, will fix in v2. To reserve hugepages at boot time, a parameter is passed to the Linux kernel on the kernel command line. For 2 MB pages, just pass the hugepages option to the kernel. For example, to reserve 1024 pages of 2 MB, use:: @@ -187,9 +191,9 @@ See the Documentation/admin-guide/kernel-parameters.txt file in your Linux sourc **Alternative:** -For 2 MB pages, there is also the option of allocating hugepages after the system has booted. +There is also the option of allocating hugepages after the system has booted. This is done by echoing the number of hugepages required to a nr_hugepages file in the ``/sys/devices/`` directory. -For a single-node system, the command to use is as follows (assuming that 1024 pages are required):: +For a single-node system, the command to use is as follows (assuming that 1024 of 2MB pages are required):: echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages @@ -198,22 +202,27 @@ On a NUMA machine, pages should be allocated explicitly on separate nodes:: echo 1024 > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages echo 1024 > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages -.. note:: - -For 1G pages, it is not possible to reserve the hugepage memory after the system has booted. - Using Hugepages with the DPDK ^ -Once the hugepage memory is reserved, to make the memory available for DPDK use, perform the following steps:: +If secondary process support is not required, DPDK is able to use hugepages +without any configuration by using "in-memory" mode. Please see +:ref:`linux_eal_parameters` for more details. + +If secondary process support is required, mount points for hugepages need to be +created. On modern Linux distributions, a default mount point for hugepages is provided +by the system and is located at ``/dev/hugepages``. This mount point will use the +default hugepage size set by the kernel parameters as described above. + +However, in order to use multiple hugepage sizes, it is necessary to manually Rather than multiple hugepage sizes, I'd suggest changing this to hugepage sizes other than the default. OK, will fix. Do we also want to add a line somewhere explaining that the default size can be set a boot using a kernel parameter? It's already there, right above this :) +create mount points for hugepage sizes that are not provided by the system +(e.g. 1GB pages). + +To make the hugepages of size 1GB available for DPDK use, perform the following steps:: mkdir /mnt/huge -mount -t hugetlbfs nodev /mnt/huge +mount -t hugetlbfs pagesize=1GB /mnt/huge The mount point can be made permanent across reboots, by adding the following line to the ``/etc/fstab`` file:: -nodev /mnt/huge hugetlbfs defaults 0 0 - -For 1GB pages, the page size must be specified as a mount option:: - -nodev /mnt/huge_1GB hugetlbfs pagesize=1GB 0 0 +nodev /mnt/huge hugetlbfs pagesize=1GB 0 0 -- 2.17.1 -- Thanks, Anatoly
Re: [dpdk-dev] [PATCH 1/2] doc/linux_gsg: clarify instructions on running as non-root
On 24-Aug-20 6:08 PM, Bruce Richardson wrote: On Mon, Aug 24, 2020 at 04:45:00PM +0100, Anatoly Burakov wrote: The current instructions are slightly out of date when it comes to providing information about setting up the system for using DPDK as non-root, so update them. Cc: sta...@dpdk.org Signed-off-by: Anatoly Burakov --- doc/guides/linux_gsg/enable_func.rst | 54 1 file changed, 39 insertions(+), 15 deletions(-) diff --git a/doc/guides/linux_gsg/enable_func.rst b/doc/guides/linux_gsg/enable_func.rst index b2bda80bb7..78b0f7c012 100644 --- a/doc/guides/linux_gsg/enable_func.rst +++ b/doc/guides/linux_gsg/enable_func.rst @@ -58,22 +58,34 @@ The application can then determine what action to take, if any, if the HPET is n if any, and on what is available on the system at runtime. Running DPDK Applications Without Root Privileges - +- -.. note:: +In order to run DPDK as non-root, the following Linux filesystem objects' +permissions should be adjusted to ensure that the Linux account being used to +run the DPDK application has access to them: -The instructions below will allow running DPDK as non-root with older -Linux kernel versions. However, since version 4.0, the kernel does not allow -unprivileged processes to read the physical address information from -the pagemaps file, making it impossible for those processes to use HW -devices which require physical addresses +* All directories which serve as hugepage mount points, for example, ``/dev/hugepages`` -Although applications using the DPDK use network ports and other hardware resources directly, -with a number of small permission adjustments it is possible to run these applications as a user other than "root". -To do so, the ownership, or permissions, on the following Linux file system objects should be adjusted to ensure that -the Linux user account being used to run the DPDK application has access to them: +* If the HPET is to be used, ``/dev/hpet`` -* All directories which serve as hugepage mount points, for example, ``/mnt/huge`` +When running as non-root user, there may be some additional resource limits +that are imposed by the system. Specifically, the following resource limits may +need to be adjusted in order to ensure normal DPDK operation: + +* RLIMIT_LOCKS (number of file locks that can be held by a process) + +* RLIMIT_NOFILE (number of open file descriptors that can be held open by a process) + +* RLIMIT_MEMLOCK (amount of pinned pages the process is allowed to have) + +The above limits can usually be adjusted by editing +``/etc/security/limits.conf`` file, and rebooting. + +Additionally, depending on which kernel driver is in use, the relevant +resources also should be accessible by the user running the DPDK application. + +For ``igb_uio`` or ``uio_pci_generic`` kernel drivers, the following Linux file +system objects' permissions should be adjusted: * The userspace-io device files in ``/dev``, for example, ``/dev/uio0``, ``/dev/uio1``, and so on @@ -82,11 +94,23 @@ the Linux user account being used to run the DPDK application has access to them /sys/class/uio/uio0/device/config /sys/class/uio/uio0/device/resource* -* If the HPET is to be used, ``/dev/hpet`` - .. note:: -On some Linux installations, ``/dev/hugepages`` is also a hugepage mount point created by default. +The instructions above will allow running DPDK with ``igb_uio`` driver as +non-root with older Linux kernel versions. However, since version 4.0, the +kernel does not allow unprivileged processes to read the physical address +information from the pagemaps file, making it impossible for those +processes to be used by non-privileged users. In such cases, using the VFIO +driver is recommended. + +For ``vfio-pci`` kernel driver, the following Linux file system objects' +permissions should be adjusted: + +* The VFIO device file , ``/dev/vfio/vfio`` + +* The directories under ``/dev/vfio`` that correspond to IOMMU group numbers of + devices intended to be used by DPDK, for example, ``/dev/vfio/50`` + Since we'd very much prefer in all cases people to use VFIO, I think the VFIO instructions should come first. Otherwise the text itself reads fine to me. OK, will fix in v2. /Bruce -- Thanks, Anatoly
[dpdk-dev] [PATCH v2 06/13] net/mlx5: remove Netlink dependency in shared code
From: Ophir Munk
This commit adds Linux implementation of routine mlx5_os_mac_addr_flush
as wrapper to Netlink API to avoid direct calls under non-Linux
operating systems.
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/linux/mlx5_os.c | 18 ++
drivers/net/mlx5/mlx5.c | 7 ++-
drivers/net/mlx5/mlx5.h | 2 +-
3 files changed, 21 insertions(+), 6 deletions(-)
diff --git a/drivers/net/mlx5/linux/mlx5_os.c b/drivers/net/mlx5/linux/mlx5_os.c
index db955ae..bf1f82b 100644
--- a/drivers/net/mlx5/linux/mlx5_os.c
+++ b/drivers/net/mlx5/linux/mlx5_os.c
@@ -45,6 +45,7 @@
#include "mlx5_flow.h"
#include "rte_pmd_mlx5.h"
#include "mlx5_verbs.h"
+#include "mlx5_nl.h"
#define MLX5_TAGS_HLIST_ARRAY_SIZE 8192
@@ -2332,6 +2333,23 @@ mlx5_os_set_allmulti(struct rte_eth_dev *dev, int enable)
mlx5_ifindex(dev), !!enable);
}
+/**
+ * Flush device MAC addresses
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ *
+ */
+void
+mlx5_os_mac_addr_flush(struct rte_eth_dev *dev)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+
+ mlx5_nl_mac_addr_flush(priv->nl_socket_route, mlx5_ifindex(dev),
+ dev->data->mac_addrs,
+ MLX5_MAX_MAC_ADDRESSES, priv->mac_own);
+}
+
const struct eth_dev_ops mlx5_os_dev_ops = {
.dev_configure = mlx5_dev_configure,
.dev_start = mlx5_dev_start,
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index ca60926..fdda6ff 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -9,7 +9,6 @@
#include
#include
#include
-#include
#include
#include
@@ -1406,9 +1405,7 @@ mlx5_dev_close(struct rte_eth_dev *dev)
if (priv->reta_idx != NULL)
mlx5_free(priv->reta_idx);
if (priv->config.vf)
- mlx5_nl_mac_addr_flush(priv->nl_socket_route, mlx5_ifindex(dev),
- dev->data->mac_addrs,
- MLX5_MAX_MAC_ADDRESSES, priv->mac_own);
+ mlx5_os_mac_addr_flush(dev);
if (priv->nl_socket_route >= 0)
close(priv->nl_socket_route);
if (priv->nl_socket_rdma >= 0)
@@ -1446,7 +1443,7 @@ mlx5_dev_close(struct rte_eth_dev *dev)
/*
* Free the shared context in last turn, because the cleanup
* routines above may use some shared fields, like
-* mlx5_nl_mac_addr_flush() uses ibdev_path for retrieveing
+* mlx5_os_mac_addr_flush() uses ibdev_path for retrieveing
* ifindex if Netlink fails.
*/
mlx5_free_shared_dev_ctx(priv->sh);
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 3aea3b5..a45bd0b 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -24,7 +24,6 @@
#include
#include
#include
-#include
#include
#include
@@ -1018,6 +1017,7 @@ int mlx5_os_vf_mac_addr_modify(struct mlx5_priv *priv,
unsigned int iface_idx,
int mlx5_os_set_promisc(struct rte_eth_dev *dev, int enable);
int mlx5_os_set_allmulti(struct rte_eth_dev *dev, int enable);
int mlx5_os_set_nonblock_channel_fd(int fd);
+void mlx5_os_mac_addr_flush(struct rte_eth_dev *dev);
/* mlx5_txpp.c */
--
2.8.4
[dpdk-dev] [PATCH v2 02/13] common/mlx5: replace Linux __bexx types with rte
From: Ophir Munk
Replace Linux specific int types with their corresponding rte typedefs.
__be16 ==> rte_be16_t
__be32 ==> rte_be32_t
__be64 ==> rte_be64_t
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_prm.h | 14 +++---
1 file changed, 7 insertions(+), 7 deletions(-)
diff --git a/drivers/common/mlx5/mlx5_prm.h b/drivers/common/mlx5/mlx5_prm.h
index e0ebe12..69511bc 100644
--- a/drivers/common/mlx5/mlx5_prm.h
+++ b/drivers/common/mlx5/mlx5_prm.h
@@ -608,7 +608,7 @@ typedef uint8_t u8;
#define MLX5_SET(typ, p, fld, v) \
do { \
u32 _v = v; \
- *((__be32 *)(p) + __mlx5_dw_off(typ, fld)) = \
+ *((rte_be32_t *)(p) + __mlx5_dw_off(typ, fld)) = \
rte_cpu_to_be_32((rte_be_to_cpu_32(*((u32 *)(p) + \
__mlx5_dw_off(typ, fld))) & \
(~__mlx5_dw_mask(typ, fld))) | \
@@ -619,15 +619,15 @@ typedef uint8_t u8;
#define MLX5_SET64(typ, p, fld, v) \
do { \
MLX5_ASSERT(__mlx5_bit_sz(typ, fld) == 64); \
- *((__be64 *)(p) + __mlx5_64_off(typ, fld)) = \
+ *((rte_be64_t *)(p) + __mlx5_64_off(typ, fld)) = \
rte_cpu_to_be_64(v); \
} while (0)
#define MLX5_SET16(typ, p, fld, v) \
do { \
u16 _v = v; \
- *((__be16 *)(p) + __mlx5_16_off(typ, fld)) = \
- rte_cpu_to_be_16((rte_be_to_cpu_16(*((__be16 *)(p) + \
+ *((rte_be16_t *)(p) + __mlx5_16_off(typ, fld)) = \
+ rte_cpu_to_be_16((rte_be_to_cpu_16(*((rte_be16_t *)(p) + \
__mlx5_16_off(typ, fld))) & \
(~__mlx5_16_mask(typ, fld))) | \
(((_v) & __mlx5_mask16(typ, fld)) << \
@@ -639,14 +639,14 @@ typedef uint8_t u8;
__mlx5_dw_off(typ, fld))) >> __mlx5_dw_bit_off(typ, fld)) & \
__mlx5_mask(typ, fld))
#define MLX5_GET(typ, p, fld) \
- ((rte_be_to_cpu_32(*((__be32 *)(p) +\
+ ((rte_be_to_cpu_32(*((rte_be32_t *)(p) +\
__mlx5_dw_off(typ, fld))) >> __mlx5_dw_bit_off(typ, fld)) & \
__mlx5_mask(typ, fld))
#define MLX5_GET16(typ, p, fld) \
- ((rte_be_to_cpu_16(*((__be16 *)(p) + \
+ ((rte_be_to_cpu_16(*((rte_be16_t *)(p) + \
__mlx5_16_off(typ, fld))) >> __mlx5_16_bit_off(typ, fld)) & \
__mlx5_mask16(typ, fld))
-#define MLX5_GET64(typ, p, fld) rte_be_to_cpu_64(*((__be64 *)(p) + \
+#define MLX5_GET64(typ, p, fld) rte_be_to_cpu_64(*((rte_be64_t *)(p) + \
__mlx5_64_off(typ, fld)))
#define MLX5_FLD_SZ_BYTES(typ, fld) (__mlx5_bit_sz(typ, fld) / 8)
--
2.8.4
[dpdk-dev] [PATCH v2 03/13] net/mlx5: rename mlx5 enumeration REG_NONE
From: Ophir Munk
Enumerated variable REG_NONE (defined in mlx5_prm.h) is in conflict with
Windows definition (winnt.h): #define REG_NONE ( 0ul ) // No value type
To enable mlx5 PMD Windows compilation - rename REG_NONE as REG_NON.
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_prm.h | 2 +-
drivers/net/mlx5/mlx5_flow.c| 22 +++---
drivers/net/mlx5/mlx5_flow_dv.c | 8
3 files changed, 16 insertions(+), 16 deletions(-)
diff --git a/drivers/common/mlx5/mlx5_prm.h b/drivers/common/mlx5/mlx5_prm.h
index 69511bc..563e7c8 100644
--- a/drivers/common/mlx5/mlx5_prm.h
+++ b/drivers/common/mlx5/mlx5_prm.h
@@ -539,7 +539,7 @@ enum mlx5_modification_field {
#define MLX5_MREG_C_NUM (MLX5_MODI_META_REG_C_7 - MLX5_MODI_META_REG_C_0 + 1)
enum modify_reg {
- REG_NONE = 0,
+ REG_NON = 0,
REG_A,
REG_B,
REG_C_0,
diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
index 7150173..9aad24e 100644
--- a/drivers/net/mlx5/mlx5_flow.c
+++ b/drivers/net/mlx5/mlx5_flow.c
@@ -353,7 +353,7 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev,
case MLX5_METADATA_FDB:
switch (config->dv_xmeta_en) {
case MLX5_XMETA_MODE_LEGACY:
- return REG_NONE;
+ return REG_NON;
case MLX5_XMETA_MODE_META16:
return REG_C_0;
case MLX5_XMETA_MODE_META32:
@@ -363,7 +363,7 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev,
case MLX5_FLOW_MARK:
switch (config->dv_xmeta_en) {
case MLX5_XMETA_MODE_LEGACY:
- return REG_NONE;
+ return REG_NON;
case MLX5_XMETA_MODE_META16:
return REG_C_1;
case MLX5_XMETA_MODE_META32:
@@ -381,7 +381,7 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev,
return priv->mtr_color_reg != REG_C_2 ? REG_C_2 :
REG_C_3;
case MLX5_MTR_COLOR:
- MLX5_ASSERT(priv->mtr_color_reg != REG_NONE);
+ MLX5_ASSERT(priv->mtr_color_reg != REG_NON);
return priv->mtr_color_reg;
case MLX5_COPY_MARK:
/*
@@ -404,7 +404,7 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev,
return rte_flow_error_set(error, EINVAL,
RTE_FLOW_ERROR_TYPE_ITEM,
NULL, "invalid tag id");
- if (config->flow_mreg_c[id + start_reg - REG_C_0] == REG_NONE)
+ if (config->flow_mreg_c[id + start_reg - REG_C_0] == REG_NON)
return rte_flow_error_set(error, ENOTSUP,
RTE_FLOW_ERROR_TYPE_ITEM,
NULL, "unsupported tag id");
@@ -421,7 +421,7 @@ mlx5_flow_get_reg_id(struct rte_eth_dev *dev,
RTE_FLOW_ERROR_TYPE_ITEM,
NULL, "invalid tag id");
if (config->flow_mreg_c
- [id + 1 + start_reg - REG_C_0] != REG_NONE)
+ [id + 1 + start_reg - REG_C_0] != REG_NON)
return config->flow_mreg_c
[id + 1 + start_reg - REG_C_0];
return rte_flow_error_set(error, ENOTSUP,
@@ -459,7 +459,7 @@ mlx5_flow_ext_mreg_supported(struct rte_eth_dev *dev)
* - reg_c's are preserved across different domain (FDB and NIC) on
* packet loopback by flow lookup miss.
*/
- return config->flow_mreg_c[2] != REG_NONE;
+ return config->flow_mreg_c[2] != REG_NON;
}
/**
@@ -3011,7 +3011,7 @@ flow_mreg_add_copy_action(struct rte_eth_dev *dev,
uint32_t mark_id,
};
struct mlx5_flow_action_copy_mreg cp_mreg = {
.dst = REG_B,
- .src = REG_NONE,
+ .src = REG_NON,
};
struct rte_flow_action_jump jump = {
.group = MLX5_FLOW_MREG_ACT_TABLE_GROUP,
@@ -3499,7 +3499,7 @@ flow_hairpin_split(struct rte_eth_dev *dev,
actions_rx++;
set_tag = (void *)actions_rx;
set_tag->id = mlx5_flow_get_reg_id(dev, MLX5_HAIRPIN_RX, 0, NULL);
- MLX5_ASSERT(set_tag->id > REG_NONE);
+ MLX5_ASSERT(set_tag->id > REG_NON);
set_tag->data = *flow_id;
tag_action->conf = set_tag;
/* Create Tx item list. */
@@ -3511,7 +3511,7 @@ flow_hairpin_split(struct rte_eth_dev *dev,
tag_item = (void *)addr;
tag_item->data = *flow_id;
tag_item->id = mlx5_flow_get_reg_id(dev, MLX5_HAIRPIN_TX, 0, NULL);
- MLX5_ASSERT(set_tag->id > REG_NONE);
+ MLX5_ASSERT(set_tag->id > REG_NON);
item->spec = tag_i
[dpdk-dev] [PATCH v2 00/13] mlx5 PMD multi OS support - part #4
This patch series is part of preparing mlx5 PMD to compile and run under multiple OSs. Part #4 v1: initial version v2: checkpatch fixes Ophir Munk (13): common/mlx5: replace strsep with strtok_r common/mlx5: replace Linux __bexx types with rte net/mlx5: rename mlx5 enumeration REG_NONE net/mlx5: move mlx5_get_ifname prototype under Linux net/mlx5: fix removal of unused inclusion files net/mlx5: remove Netlink dependency in shared code net/mlx5: fix unused utility macros net/mlx5: call meter detach only if DR is supported net/mlx5: add ICMP protocol number definition net/mlx5: remove more DV dependencies net/mlx5: remove ibv_* dependency in Rx/Tx objects net/mlx5: separate VLAN strip modification linux/mlx5: refactor VLAN drivers/common/mlx5/linux/mlx5_common_os.h | 111 +++ drivers/common/mlx5/mlx5_common_pci.c | 14 +-- drivers/common/mlx5/mlx5_prm.h | 16 +-- drivers/net/mlx5/Makefile | 1 + drivers/net/mlx5/linux/meson.build | 1 + drivers/net/mlx5/linux/mlx5_os.c | 18 drivers/net/mlx5/linux/mlx5_os.h | 6 ++ drivers/net/mlx5/linux/mlx5_verbs.c| 28 + drivers/net/mlx5/linux/mlx5_vlan_os.c | 168 + drivers/net/mlx5/mlx5.c| 23 ++-- drivers/net/mlx5/mlx5.h| 30 +++--- drivers/net/mlx5/mlx5_devx.c | 48 + drivers/net/mlx5/mlx5_devx.h | 12 +++ drivers/net/mlx5/mlx5_flow.c | 29 +++-- drivers/net/mlx5/mlx5_flow_dv.c| 8 +- drivers/net/mlx5/mlx5_flow_meter.c | 4 + drivers/net/mlx5/mlx5_mac.c| 2 - drivers/net/mlx5/mlx5_rxq.c| 18 ++-- drivers/net/mlx5/mlx5_rxtx.h | 24 ++--- drivers/net/mlx5/mlx5_trigger.c| 3 +- drivers/net/mlx5/mlx5_txpp.c | 38 --- drivers/net/mlx5/mlx5_txq.c| 19 ++-- drivers/net/mlx5/mlx5_utils.h | 4 - drivers/net/mlx5/mlx5_vlan.c | 161 +-- 24 files changed, 527 insertions(+), 259 deletions(-) create mode 100644 drivers/net/mlx5/linux/mlx5_vlan_os.c create mode 100644 drivers/net/mlx5/mlx5_devx.c create mode 100644 drivers/net/mlx5/mlx5_devx.h -- 2.8.4
[dpdk-dev] [PATCH v2 05/13] net/mlx5: fix removal of unused inclusion files
From: Ophir Munk
Remove unused Linux included files:
, from file net/mlx5/mlx5_mac.c
from file net/mlx5/mlx5.c
Fixes: 771fa900b73a ("mlx5: introduce new driver for Mellanox ConnectX-4
adapters")
Cc: sta...@dpdk.org
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/mlx5.c | 1 -
drivers/net/mlx5/mlx5_mac.c | 2 --
2 files changed, 3 deletions(-)
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index b099b23..ca60926 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -9,7 +9,6 @@
#include
#include
#include
-#include
#include
#include
diff --git a/drivers/net/mlx5/mlx5_mac.c b/drivers/net/mlx5/mlx5_mac.c
index 88c52b2..bd786fd 100644
--- a/drivers/net/mlx5/mlx5_mac.c
+++ b/drivers/net/mlx5/mlx5_mac.c
@@ -9,8 +9,6 @@
#include
#include
#include
-#include
-#include
#include
#include
--
2.8.4
[dpdk-dev] [PATCH v2 04/13] net/mlx5: move mlx5_get_ifname prototype under Linux
From: Ophir Munk
mlx5_get_ifname() prototype includes 'IF_NAMESIZE' definition from Linux
file net/if.h. Since this API is only used under Linux and to enable
compilation under non-Linux OS - move this prototype from shared file
mlx5.h to file linux/mlx5_os.h.
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/linux/mlx5_os.h | 6 ++
drivers/net/mlx5/mlx5.c | 1 -
drivers/net/mlx5/mlx5.h | 2 --
3 files changed, 6 insertions(+), 3 deletions(-)
diff --git a/drivers/net/mlx5/linux/mlx5_os.h b/drivers/net/mlx5/linux/mlx5_os.h
index 31add39..759def2 100644
--- a/drivers/net/mlx5/linux/mlx5_os.h
+++ b/drivers/net/mlx5/linux/mlx5_os.h
@@ -6,6 +6,8 @@
#ifndef RTE_PMD_MLX5_OS_H_
#define RTE_PMD_MLX5_OS_H_
+#include
+
/* verb enumerations translations to local enums. */
enum {
DEV_SYSFS_NAME_MAX = IBV_SYSFS_NAME_MAX + 1,
@@ -15,4 +17,8 @@ enum {
#define PCI_DRV_FLAGS (RTE_PCI_DRV_INTR_LSC | \
RTE_PCI_DRV_INTR_RMV | \
RTE_PCI_DRV_PROBE_AGAIN)
+
+/* mlx5_ethdev_os.c */
+
+int mlx5_get_ifname(const struct rte_eth_dev *dev, char
(*ifname)[IF_NAMESIZE]);
#endif /* RTE_PMD_MLX5_OS_H_ */
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 1e4c695..b099b23 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -9,7 +9,6 @@
#include
#include
#include
-#include
#include
#include
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 1880a82..3aea3b5 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -10,7 +10,6 @@
#include
#include
#include
-#include
#include
#include
@@ -813,7 +812,6 @@ int mlx5_dev_configure_rss_reta(struct rte_eth_dev *dev);
/* mlx5_ethdev_os.c */
-int mlx5_get_ifname(const struct rte_eth_dev *dev, char
(*ifname)[IF_NAMESIZE]);
unsigned int mlx5_ifindex(const struct rte_eth_dev *dev);
int mlx5_get_mac(struct rte_eth_dev *dev, uint8_t (*mac)[RTE_ETHER_ADDR_LEN]);
int mlx5_get_mtu(struct rte_eth_dev *dev, uint16_t *mtu);
--
2.8.4
[dpdk-dev] [PATCH v2 01/13] common/mlx5: replace strsep with strtok_r
From: Ophir Munk
strsep() is a non-standardized API (by C or POSIX) and thus it is
non-portable between different operating systems. Replace it with
strtok_r() which is standardized by the C standard, and hence also by
POSIX.
The replacement occurs in the code that extracts individual PCI class
names (e.g. class=net:vdpa:foo:bar).
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/common/mlx5/mlx5_common_pci.c | 14 --
1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/common/mlx5/mlx5_common_pci.c
b/drivers/common/mlx5/mlx5_common_pci.c
index d4ff039..02417c6 100644
--- a/drivers/common/mlx5/mlx5_common_pci.c
+++ b/drivers/common/mlx5/mlx5_common_pci.c
@@ -72,6 +72,7 @@ bus_cmdline_options_handler(__rte_unused const char *key,
int class_val;
char *found;
char *nstr;
+ char *refstr = NULL;
*ret = 0;
nstr = strdup(class_names);
@@ -80,21 +81,22 @@ bus_cmdline_options_handler(__rte_unused const char *key,
return *ret;
}
nstr_org = nstr;
- while (nstr) {
+ found = strtok_r(nstr, ":", &refstr);
+ if (!found)
+ goto err;
+ do {
/* Extract each individual class name. Multiple
* class key,value is supplied as class=net:vdpa:foo:bar.
*/
- found = strsep(&nstr, ":");
- if (!found)
- continue;
- /* Check if its a valid class. */
class_val = class_name_to_value(found);
+ /* Check if its a valid class. */
if (class_val < 0) {
*ret = -EINVAL;
goto err;
}
*ret |= class_val;
- }
+ found = strtok_r(NULL, ":", &refstr);
+ } while (found);
err:
free(nstr_org);
if (*ret < 0)
--
2.8.4
[dpdk-dev] [PATCH v2 09/13] net/mlx5: add ICMP protocol number definition
From: Ophir Munk Some Windows distributions do not include ICMP protocol definitions IPPROTO_ICMP and IPPROTO_ICMPV6. This commit defines them if needed. Signed-off-by: Ophir Munk Acked-by: Matan Azrad --- drivers/net/mlx5/mlx5_flow.c | 7 +++ 1 file changed, 7 insertions(+) diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c index 9aad24e..4c29898 100644 --- a/drivers/net/mlx5/mlx5_flow.c +++ b/drivers/net/mlx5/mlx5_flow.c @@ -1244,6 +1244,10 @@ mlx5_flow_validate_attributes(struct rte_eth_dev *dev, * @return * 0 on success, a negative errno value otherwise and rte_errno is set. */ + +#ifndef IPPROTO_ICMPV6 +#define IPPROTO_ICMPV6 58 +#endif int mlx5_flow_validate_item_icmp6(const struct rte_flow_item *item, uint64_t item_flags, @@ -1296,6 +1300,9 @@ mlx5_flow_validate_item_icmp6(const struct rte_flow_item *item, * @return * 0 on success, a negative errno value otherwise and rte_errno is set. */ +#ifndef IPPROTO_ICMP +#define IPPROTO_ICMP 1 +#endif int mlx5_flow_validate_item_icmp(const struct rte_flow_item *item, uint64_t item_flags, -- 2.8.4
[dpdk-dev] [PATCH v2 08/13] net/mlx5: call meter detach only if DR is supported
From: Ophir Munk
Flow metering is supported only in direct rules (DR). Currently the APIs
of meter actions create and modify are under #ifdef
HAVE_MLX5_DR_CREATE_ACTION_FLOW_METER, while detaching the meter action
is executed unconditionally. This commit adds the same ifdef to API
mlx5_flow_meter_detach().
This commit avoids compilation failure of non-Linux operating systems
which do not support DR.
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/mlx5_flow_meter.c | 4
1 file changed, 4 insertions(+)
diff --git a/drivers/net/mlx5/mlx5_flow_meter.c
b/drivers/net/mlx5/mlx5_flow_meter.c
index bf34687..b36bc7b 100644
--- a/drivers/net/mlx5/mlx5_flow_meter.c
+++ b/drivers/net/mlx5/mlx5_flow_meter.c
@@ -1221,6 +1221,7 @@ mlx5_flow_meter_attach(struct mlx5_priv *priv, uint32_t
meter_id,
void
mlx5_flow_meter_detach(struct mlx5_flow_meter *fm)
{
+#ifdef HAVE_MLX5_DR_CREATE_ACTION_FLOW_METER
MLX5_ASSERT(fm->ref_cnt);
if (--fm->ref_cnt)
return;
@@ -1230,6 +1231,9 @@ mlx5_flow_meter_detach(struct mlx5_flow_meter *fm)
fm->ingress = 0;
fm->egress = 0;
fm->transfer = 0;
+#else
+ (void)fm;
+#endif
}
/**
--
2.8.4
[dpdk-dev] [PATCH v2 07/13] net/mlx5: fix unused utility macros
From: Ophir Munk
Remove utility macros INFO, WARN, ERROR. They are not in use and
conflict with identical definitions when compiled under Windows.
Fixes: 80f2d0ed7ff9 ("net/mlx5: add hardware flow debug dump")
Cc: sta...@dpdk.org
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/mlx5_utils.h | 4
1 file changed, 4 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_utils.h b/drivers/net/mlx5/mlx5_utils.h
index 97d931f..f078bdc 100644
--- a/drivers/net/mlx5/mlx5_utils.h
+++ b/drivers/net/mlx5/mlx5_utils.h
@@ -35,10 +35,6 @@ extern int mlx5_logtype;
__VA_ARGS__ PMD_DRV_LOG_STRIP PMD_DRV_LOG_OPAREN, \
PMD_DRV_LOG_CPAREN)
-#define INFO(...) DRV_LOG(INFO, __VA_ARGS__)
-#define WARN(...) DRV_LOG(WARNING, __VA_ARGS__)
-#define ERROR(...) DRV_LOG(ERR, __VA_ARGS__)
-
/* Convenience macros for accessing mbuf fields. */
#define NEXT(m) ((m)->next)
#define DATA_LEN(m) ((m)->data_len)
--
2.8.4
[dpdk-dev] [PATCH v2 10/13] net/mlx5: remove more DV dependencies
From: Ophir Munk
Several DV-based structs of type 'struct mlx5dv_devx_XXX' are replaced
with 'void *' to enable compilation under non-Linux operating systems.
New getter functions were added to retrieve the specific fields that
were previously accessed directly.
Replaced structs:
'struct mlx5dv_pp *'
'struct mlx5dv_devx_event_channel *'
'struct mlx5dv_devx_umem *'
'struct mlx5dv_devx_uar *'
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/common/mlx5/linux/mlx5_common_os.h | 111 +
drivers/net/mlx5/mlx5.c| 14 ++--
drivers/net/mlx5/mlx5.h| 12 ++--
drivers/net/mlx5/mlx5_rxq.c| 13 ++--
drivers/net/mlx5/mlx5_rxtx.h | 10 +--
drivers/net/mlx5/mlx5_txpp.c | 38 +-
drivers/net/mlx5/mlx5_txq.c| 17 +++--
7 files changed, 172 insertions(+), 43 deletions(-)
diff --git a/drivers/common/mlx5/linux/mlx5_common_os.h
b/drivers/common/mlx5/linux/mlx5_common_os.h
index 55c0902..3420937 100644
--- a/drivers/common/mlx5/linux/mlx5_common_os.h
+++ b/drivers/common/mlx5/linux/mlx5_common_os.h
@@ -90,4 +90,115 @@ mlx5_os_get_umem_id(void *umem)
return 0;
return ((struct mlx5dv_devx_umem *)umem)->umem_id;
}
+
+/**
+ * Get fd. Given a pointer to DevX channel object of type
+ * 'struct mlx5dv_devx_event_channel*' - return its fd.
+ *
+ * @param[in] channel
+ *Pointer to channel object.
+ *
+ * @return
+ *The fd if channel is valid, 0 otherwise.
+ */
+static inline int
+mlx5_os_get_devx_channel_fd(void *channel)
+{
+ if (!channel)
+ return 0;
+ return ((struct mlx5dv_devx_event_channel *)channel)->fd;
+}
+
+/**
+ * Get mmap offset. Given a pointer to an DevX UAR object of type
+ * 'struct mlx5dv_devx_uar *' - return its mmap offset.
+ *
+ * @param[in] uar
+ *Pointer to UAR object.
+ *
+ * @return
+ *The mmap offset if uar is valid, 0 otherwise.
+ */
+static inline off_t
+mlx5_os_get_devx_uar_mmap_offset(void *uar)
+{
+#ifdef HAVE_MLX5DV_DEVX_UAR_OFFSET
+ if (!uar)
+ return 0;
+ return ((struct mlx5dv_devx_uar *)uar)->mmap_off;
+#else
+ RTE_SET_USED(uar);
+ return 0;
+#endif
+}
+
+/**
+ * Get base addr pointer. Given a pointer to an UAR object of type
+ * 'struct mlx5dv_devx_uar *' - return its base address.
+ *
+ * @param[in] uar
+ *Pointer to an UAR object.
+ *
+ * @return
+ *The base address if UAR is valid, 0 otherwise.
+ */
+static inline void *
+mlx5_os_get_devx_uar_base_addr(void *uar)
+{
+#ifdef HAVE_MLX5DV_DEVX_UAR_OFFSET
+ if (!uar)
+ return NULL;
+ return ((struct mlx5dv_devx_uar *)uar)->base_addr;
+#else
+ RTE_SET_USED(uar);
+ return NULL;
+#endif
+}
+
+/**
+ * Get reg addr pointer. Given a pointer to an UAR object of type
+ * 'struct mlx5dv_devx_uar *' - return its reg address.
+ *
+ * @param[in] uar
+ *Pointer to an UAR object.
+ *
+ * @return
+ *The reg address if UAR is valid, 0 otherwise.
+ */
+static inline void *
+mlx5_os_get_devx_uar_reg_addr(void *uar)
+{
+#ifdef HAVE_MLX5DV_DEVX_UAR_OFFSET
+ if (!uar)
+ return NULL;
+ return ((struct mlx5dv_devx_uar *)uar)->reg_addr;
+#else
+ RTE_SET_USED(uar);
+ return NULL;
+#endif
+}
+
+/**
+ * Get page id. Given a pointer to an UAR object of type
+ * 'struct mlx5dv_devx_uar *' - return its page id.
+ *
+ * @param[in] uar
+ *Pointer to an UAR object.
+ *
+ * @return
+ *The page id if UAR is valid, 0 otherwise.
+ */
+static inline uint32_t
+mlx5_os_get_devx_uar_page_id(void *uar)
+{
+#ifdef HAVE_MLX5DV_DEVX_UAR_OFFSET
+ if (!uar)
+ return 0;
+ return ((struct mlx5dv_devx_uar *)uar)->page_id;
+#else
+ RTE_SET_USED(uar);
+ return 0;
+#endif
+}
+
#endif /* RTE_PMD_MLX5_COMMON_OS_H_ */
diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index fdda6ff..4a807fb 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -723,6 +723,7 @@ mlx5_alloc_rxtx_uars(struct mlx5_dev_ctx_shared *sh,
{
uint32_t uar_mapping, retry;
int err = 0;
+ void *base_addr;
for (retry = 0; retry < MLX5_ALLOC_UAR_RETRY; ++retry) {
#ifdef MLX5DV_UAR_ALLOC_TYPE_NC
@@ -781,7 +782,8 @@ mlx5_alloc_rxtx_uars(struct mlx5_dev_ctx_shared *sh,
err = ENOMEM;
goto exit;
}
- if (sh->tx_uar->base_addr)
+ base_addr = mlx5_os_get_devx_uar_base_addr(sh->tx_uar);
+ if (base_addr)
break;
/*
* The UARs are allocated by rdma_core within the
@@ -820,7 +822,8 @@ mlx5_alloc_rxtx_uars(struct mlx5_dev_ctx_shared *sh,
err = ENOMEM;
goto exit;
}
- if (sh->devx_rx_uar->base_addr)
+ base_addr = mlx5_os_get_devx_uar_base_addr(sh->d
[dpdk-dev] [PATCH v2 12/13] net/mlx5: separate VLAN strip modification
From: Ophir Munk
When updating a queue vlan stripping offload - either the WQ is modified
in Verbs or the RQ is modified in DevX. Add a vlan stripping modify
callback to 'struct mlx5_obj_ops' and assign it with the specific Verbs
and DevX implementations: 'rxq_obj_modify_wq_vlan_strip' and
'rxq_obj_modify_rq_vlan_strip' respectively.
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/linux/mlx5_verbs.c | 28 ++
drivers/net/mlx5/mlx5.h | 6 +
drivers/net/mlx5/mlx5_devx.c| 48 +
drivers/net/mlx5/mlx5_devx.h| 12 ++
drivers/net/mlx5/mlx5_vlan.c| 27 -
5 files changed, 98 insertions(+), 23 deletions(-)
create mode 100644 drivers/net/mlx5/mlx5_devx.c
create mode 100644 drivers/net/mlx5/mlx5_devx.h
diff --git a/drivers/net/mlx5/linux/mlx5_verbs.c
b/drivers/net/mlx5/linux/mlx5_verbs.c
index d41b0fe..6271f0f 100644
--- a/drivers/net/mlx5/linux/mlx5_verbs.c
+++ b/drivers/net/mlx5/linux/mlx5_verbs.c
@@ -19,6 +19,7 @@
#include
#include
#include
+#include
#include
/**
* Register mr. Given protection domain pointer, pointer to addr and length
@@ -61,3 +62,30 @@ const struct mlx5_verbs_ops mlx5_verbs_ops = {
.reg_mr = mlx5_reg_mr,
.dereg_mr = mlx5_dereg_mr,
};
+
+/**
+ * Modify Rx WQ vlan stripping offload
+ *
+ * @param rxq_obj
+ * Rx queue object.
+ *
+ * @return 0 on success, non-0 otherwise
+ */
+static int
+mlx5_rxq_obj_modify_wq_vlan_strip(struct mlx5_rxq_obj *rxq_obj, int on)
+{
+ uint16_t vlan_offloads =
+ (on ? IBV_WQ_FLAGS_CVLAN_STRIPPING : 0) |
+ 0;
+ struct ibv_wq_attr mod;
+ mod = (struct ibv_wq_attr){
+ .attr_mask = IBV_WQ_ATTR_FLAGS,
+ .flags_mask = IBV_WQ_FLAGS_CVLAN_STRIPPING,
+ .flags = vlan_offloads,
+ };
+ return mlx5_glue->modify_wq(rxq_obj->wq, &mod);
+}
+
+struct mlx5_obj_ops ibv_obj_ops = {
+ .rxq_obj_modify_vlan_strip = mlx5_rxq_obj_modify_wq_vlan_strip,
+};
diff --git a/drivers/net/mlx5/mlx5.h b/drivers/net/mlx5/mlx5.h
index 34d7a15..431f861 100644
--- a/drivers/net/mlx5/mlx5.h
+++ b/drivers/net/mlx5/mlx5.h
@@ -676,6 +676,11 @@ TAILQ_HEAD(mlx5_flow_meters, mlx5_flow_meter);
#define MLX5_PROC_PRIV(port_id) \
((struct mlx5_proc_priv *)rte_eth_devices[port_id].process_private)
+/* HW objects operations structure. */
+struct mlx5_obj_ops {
+ int (*rxq_obj_modify_vlan_strip)(struct mlx5_rxq_obj *rxq_obj, int on);
+};
+
struct mlx5_priv {
struct rte_eth_dev_data *dev_data; /* Pointer to device data. */
struct mlx5_dev_ctx_shared *sh; /* Shared device context. */
@@ -719,6 +724,7 @@ struct mlx5_priv {
void *rss_desc; /* Intermediate rss description resources. */
int flow_idx; /* Intermediate device flow index. */
int flow_nested_idx; /* Intermediate device flow index, nested. */
+ struct mlx5_obj_ops *obj_ops; /* HW objects operations. */
LIST_HEAD(rxq, mlx5_rxq_ctrl) rxqsctrl; /* DPDK Rx queues. */
LIST_HEAD(rxqobj, mlx5_rxq_obj) rxqsobj; /* Verbs/DevX Rx queues. */
uint32_t hrxqs; /* Verbs Hash Rx queues. */
diff --git a/drivers/net/mlx5/mlx5_devx.c b/drivers/net/mlx5/mlx5_devx.c
new file mode 100644
index 000..7340412
--- /dev/null
+++ b/drivers/net/mlx5/mlx5_devx.c
@@ -0,0 +1,48 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2020 Mellanox Technologies, Ltd
+ */
+
+#include
+#include
+#include
+#include
+#include
+
+#include
+#include
+#include
+
+#include
+#include
+#include
+
+#include "mlx5.h"
+#include "mlx5_common_os.h"
+#include "mlx5_rxtx.h"
+#include "mlx5_utils.h"
+#include "mlx5_devx.h"
+
+/**
+ * Modify RQ vlan stripping offload
+ *
+ * @param rxq_obj
+ * Rx queue object.
+ *
+ * @return 0 on success, non-0 otherwise
+ */
+static int
+mlx5_rxq_obj_modify_rq_vlan_strip(struct mlx5_rxq_obj *rxq_obj, int on)
+{
+ struct mlx5_devx_modify_rq_attr rq_attr;
+
+ memset(&rq_attr, 0, sizeof(rq_attr));
+ rq_attr.rq_state = MLX5_RQC_STATE_RDY;
+ rq_attr.state = MLX5_RQC_STATE_RDY;
+ rq_attr.vsd = (on ? 0 : 1);
+ rq_attr.modify_bitmask = MLX5_MODIFY_RQ_IN_MODIFY_BITMASK_VSD;
+ return mlx5_devx_cmd_modify_rq(rxq_obj->rq, &rq_attr);
+}
+
+struct mlx5_obj_ops devx_obj_ops = {
+ .rxq_obj_modify_vlan_strip = mlx5_rxq_obj_modify_rq_vlan_strip,
+};
diff --git a/drivers/net/mlx5/mlx5_devx.h b/drivers/net/mlx5/mlx5_devx.h
new file mode 100644
index 000..844985c
--- /dev/null
+++ b/drivers/net/mlx5/mlx5_devx.h
@@ -0,0 +1,12 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2020 Mellanox Technologies, Ltd
+ */
+
+#ifndef RTE_PMD_MLX5_DEVX_H_
+#define RTE_PMD_MLX5_DEVX_H_
+
+#include "mlx5.h"
+
+extern struct mlx5_obj_ops devx_obj_ops;
+
+#endif /* RTE_PMD_MLX5_DEVX_H_ */
diff --git a/drivers/net/mlx5/mlx5_vlan.c b/drivers/net/mlx5/mlx5_vlan.c
index 8
[dpdk-dev] [PATCH v2 13/13] linux/mlx5: refactor VLAN
From: Ophir Munk
File mlx5_vlan.c contains Netlink APIs (Linux dependent) as part of VM
workaround implementation. Move this implementation to file
linux/mlx5_vlan_os.c. To remove Netlink dependency in header files
change pointer of type 'struct mlx5_nl_vlan_vmwa_context *' to 'void *'.
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/Makefile | 1 +
drivers/net/mlx5/linux/meson.build| 1 +
drivers/net/mlx5/linux/mlx5_vlan_os.c | 168 ++
drivers/net/mlx5/mlx5.h | 8 +-
drivers/net/mlx5/mlx5_vlan.c | 134 ---
5 files changed, 175 insertions(+), 137 deletions(-)
create mode 100644 drivers/net/mlx5/linux/mlx5_vlan_os.c
diff --git a/drivers/net/mlx5/Makefile b/drivers/net/mlx5/Makefile
index 568c772..6097688 100644
--- a/drivers/net/mlx5/Makefile
+++ b/drivers/net/mlx5/Makefile
@@ -36,6 +36,7 @@ SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += linux/mlx5_os.c
SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += linux/mlx5_ethdev_os.c
SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += linux/mlx5_verbs.c
SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += linux/mlx5_mp_os.c
+SRCS-$(CONFIG_RTE_LIBRTE_MLX5_PMD) += linux/mlx5_vlan_os.c
# Basic CFLAGS.
CFLAGS += -O3
diff --git a/drivers/net/mlx5/linux/meson.build
b/drivers/net/mlx5/linux/meson.build
index 2def8e3..6c44021 100644
--- a/drivers/net/mlx5/linux/meson.build
+++ b/drivers/net/mlx5/linux/meson.build
@@ -8,5 +8,6 @@ sources += files(
'mlx5_ethdev_os.c',
'mlx5_verbs.c',
'mlx5_mp_os.c',
+ 'mlx5_vlan_os.c',
)
diff --git a/drivers/net/mlx5/linux/mlx5_vlan_os.c
b/drivers/net/mlx5/linux/mlx5_vlan_os.c
new file mode 100644
index 000..92fc17d
--- /dev/null
+++ b/drivers/net/mlx5/linux/mlx5_vlan_os.c
@@ -0,0 +1,168 @@
+/* SPDX-License-Identifier: BSD-3-Clause
+ * Copyright 2015 6WIND S.A.
+ * Copyright 2015 Mellanox Technologies, Ltd
+ */
+
+#include
+#include
+#include
+#include
+
+/*
+ * Not needed by this file; included to work around the lack of off_t
+ * definition for mlx5dv.h with unpatched rdma-core versions.
+ */
+#include
+
+#include
+#include
+#include
+#include
+
+#include
+#include
+#include
+
+/*
+ * Release VLAN network device, created for VM workaround.
+ *
+ * @param[in] dev
+ * Ethernet device object, Netlink context provider.
+ * @param[in] vlan
+ * Object representing the network device to release.
+ */
+void
+mlx5_vlan_vmwa_release(struct rte_eth_dev *dev,
+ struct mlx5_vf_vlan *vlan)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_nl_vlan_vmwa_context *vmwa = priv->vmwa_context;
+ struct mlx5_nl_vlan_dev *vlan_dev = &vmwa->vlan_dev[0];
+
+ MLX5_ASSERT(vlan->created);
+ MLX5_ASSERT(priv->vmwa_context);
+ if (!vlan->created || !vmwa)
+ return;
+ vlan->created = 0;
+ MLX5_ASSERT(vlan_dev[vlan->tag].refcnt);
+ if (--vlan_dev[vlan->tag].refcnt == 0 &&
+ vlan_dev[vlan->tag].ifindex) {
+ mlx5_nl_vlan_vmwa_delete(vmwa, vlan_dev[vlan->tag].ifindex);
+ vlan_dev[vlan->tag].ifindex = 0;
+ }
+}
+
+/**
+ * Acquire VLAN interface with specified tag for VM workaround.
+ *
+ * @param[in] dev
+ * Ethernet device object, Netlink context provider.
+ * @param[in] vlan
+ * Object representing the network device to acquire.
+ */
+void
+mlx5_vlan_vmwa_acquire(struct rte_eth_dev *dev,
+ struct mlx5_vf_vlan *vlan)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_nl_vlan_vmwa_context *vmwa = priv->vmwa_context;
+ struct mlx5_nl_vlan_dev *vlan_dev = &vmwa->vlan_dev[0];
+
+ MLX5_ASSERT(!vlan->created);
+ MLX5_ASSERT(priv->vmwa_context);
+ if (vlan->created || !vmwa)
+ return;
+ if (vlan_dev[vlan->tag].refcnt == 0) {
+ MLX5_ASSERT(!vlan_dev[vlan->tag].ifindex);
+ vlan_dev[vlan->tag].ifindex =
+ mlx5_nl_vlan_vmwa_create(vmwa, vmwa->vf_ifindex,
+vlan->tag);
+ }
+ if (vlan_dev[vlan->tag].ifindex) {
+ vlan_dev[vlan->tag].refcnt++;
+ vlan->created = 1;
+ }
+}
+
+/*
+ * Create per ethernet device VLAN VM workaround context
+ *
+ * @param dev
+ * Pointer to Ethernet device structure.
+ * @param ifindex
+ * Interface index.
+ *
+ * @Return
+ * Pointer to mlx5_nl_vlan_vmwa_context
+ */
+void *
+mlx5_vlan_vmwa_init(struct rte_eth_dev *dev, uint32_t ifindex)
+{
+ struct mlx5_priv *priv = dev->data->dev_private;
+ struct mlx5_dev_config *config = &priv->config;
+ struct mlx5_nl_vlan_vmwa_context *vmwa;
+ enum rte_hypervisor hv_type;
+
+ /* Do not engage workaround over PF. */
+ if (!config->vf)
+ return NULL;
+ /* Check whether there is desired virtual environment */
+ hv_type = rte_
[dpdk-dev] [PATCH v2 11/13] net/mlx5: remove ibv_* dependency in Rx/Tx objects
From: Ophir Munk
Replace pointers to ibv structs with pointers to void (file
mlx5_rxtx.h). Specifically the following pointers were replaced:
'struct ibv_cq *', 'struct ibv_wq *', 'struct ibv_comp_channel *',
'struct ibv_rwq_ind_table *a', 'struct ibv_qp *'.
Signed-off-by: Ophir Munk
Acked-by: Matan Azrad
---
drivers/net/mlx5/mlx5_rxq.c | 5 +++--
drivers/net/mlx5/mlx5_rxtx.h| 14 +++---
drivers/net/mlx5/mlx5_trigger.c | 3 ++-
drivers/net/mlx5/mlx5_txq.c | 2 +-
4 files changed, 13 insertions(+), 11 deletions(-)
diff --git a/drivers/net/mlx5/mlx5_rxq.c b/drivers/net/mlx5/mlx5_rxq.c
index 92d8876..946f745 100644
--- a/drivers/net/mlx5/mlx5_rxq.c
+++ b/drivers/net/mlx5/mlx5_rxq.c
@@ -1790,7 +1790,8 @@ mlx5_rxq_obj_new(struct rte_eth_dev *dev, uint16_t idx,
rte_errno = ENOMEM;
goto error;
}
- tmpl->fd = tmpl->ibv_channel->fd;
+ tmpl->fd = ((struct ibv_comp_channel *)
+ (tmpl->ibv_channel))->fd;
} else if (tmpl->type == MLX5_RXQ_OBJ_TYPE_DEVX_RQ) {
int devx_ev_flag =
MLX5DV_DEVX_CREATE_EVENT_CHANNEL_FLAGS_OMIT_EV_DATA;
@@ -3299,7 +3300,7 @@ mlx5_ind_table_obj_drop_new(struct rte_eth_dev *dev)
(priv->sh->ctx,
&(struct ibv_rwq_ind_table_init_attr){
.log_ind_tbl_size = 0,
- .ind_tbl = &rxq->wq,
+ .ind_tbl = (struct ibv_wq **)&rxq->wq,
.comp_mask = 0,
});
if (!tmpl.ind_table) {
diff --git a/drivers/net/mlx5/mlx5_rxtx.h b/drivers/net/mlx5/mlx5_rxtx.h
index 0fc7754..f3fe2e1 100644
--- a/drivers/net/mlx5/mlx5_rxtx.h
+++ b/drivers/net/mlx5/mlx5_rxtx.h
@@ -178,9 +178,9 @@ struct mlx5_rxq_obj {
RTE_STD_C11
union {
struct {
- struct ibv_wq *wq; /* Work Queue. */
- struct ibv_cq *ibv_cq; /* Completion Queue. */
- struct ibv_comp_channel *ibv_channel;
+ void *wq; /* Work Queue. */
+ void *ibv_cq; /* Completion Queue. */
+ void *ibv_channel;
};
struct {
struct mlx5_devx_obj *rq; /* DevX Rx Queue object. */
@@ -229,7 +229,7 @@ struct mlx5_ind_table_obj {
enum mlx5_ind_tbl_type type;
RTE_STD_C11
union {
- struct ibv_rwq_ind_table *ind_table; /**< Indirection table. */
+ void *ind_table; /**< Indirection table. */
struct mlx5_devx_obj *rqt; /* DevX RQT object. */
};
uint32_t queues_n; /**< Number of queues in the list. */
@@ -243,7 +243,7 @@ struct mlx5_hrxq {
struct mlx5_ind_table_obj *ind_table; /* Indirection table. */
RTE_STD_C11
union {
- struct ibv_qp *qp; /* Verbs queue pair. */
+ void *qp; /* Verbs queue pair. */
struct mlx5_devx_obj *tir; /* DevX TIR object. */
};
#ifdef HAVE_IBV_FLOW_DV_SUPPORT
@@ -350,8 +350,8 @@ struct mlx5_txq_obj {
RTE_STD_C11
union {
struct {
- struct ibv_cq *cq; /* Completion Queue. */
- struct ibv_qp *qp; /* Queue Pair. */
+ void *cq; /* Completion Queue. */
+ void *qp; /* Queue Pair. */
};
struct {
struct mlx5_devx_obj *sq;
diff --git a/drivers/net/mlx5/mlx5_trigger.c b/drivers/net/mlx5/mlx5_trigger.c
index 6f1e6d4..549af35 100644
--- a/drivers/net/mlx5/mlx5_trigger.c
+++ b/drivers/net/mlx5/mlx5_trigger.c
@@ -153,7 +153,8 @@ mlx5_rxq_start(struct rte_eth_dev *dev)
if (!rxq_ctrl->obj)
goto error;
if (obj_type == MLX5_RXQ_OBJ_TYPE_IBV)
- rxq_ctrl->wqn = rxq_ctrl->obj->wq->wq_num;
+ rxq_ctrl->wqn =
+ ((struct ibv_wq *)(rxq_ctrl->obj->wq))->wq_num;
else if (obj_type == MLX5_RXQ_OBJ_TYPE_DEVX_RQ)
rxq_ctrl->wqn = rxq_ctrl->obj->rq->id;
}
diff --git a/drivers/net/mlx5/mlx5_txq.c b/drivers/net/mlx5/mlx5_txq.c
index fed9d8a..66ad368 100644
--- a/drivers/net/mlx5/mlx5_txq.c
+++ b/drivers/net/mlx5/mlx5_txq.c
@@ -1317,7 +1317,7 @@ mlx5_txq_obj_new(struct rte_eth_dev *dev, uint16_t idx,
txq_data->cqe_n = log2above(cq_info.cqe_cnt);
txq_data->cqe_s = 1 << txq_data->cqe_n;
txq_data->cqe_m = txq_data->cqe_s - 1;
- txq_data->qp_num_8s = tmpl.qp->qp_num << 8;
+ txq_data->qp_num_8s = ((struct ibv_qp *)tmpl.qp)->qp_num << 8;
txq_data->wqes = qp.sq.buf;
txq_data->wqe_n = log2above(qp.sq.wqe_cnt);
txq_data->wqe_s = 1 << txq_
Re: [dpdk-dev] 19.11.4 patches review and test
On Tue, 2020-08-25 at 05:13 -0400, Pei Zhang wrote: > Hello Luca, > > Testing with dpdk 19.11.4-rc1 from Red Hat looks good. > > We cover below 16 scenarios and and all get PASS on RHEL8 testing: > > (1)Guest with device assignment(PF) throughput testing(1G hugepage size): PASS > (2)Guest with device assignment(PF) throughput testing(2M hugepage size) : > PASS > (3)Guest with device assignment(VF) throughput testing: PASS > (4)PVP (host dpdk testpmd as vswitch) 1Q: throughput testing: PASS > (5)PVP vhost-user 2Q throughput testing: PASS > (6)PVP vhost-user 1Q - cross numa node throughput testing: PASS > (7)Guest with vhost-user 2 queues throughput testing: PASS > (8)vhost-user reconnect with dpdk-client, qemu-server: qemu reconnect: PASS > (9)PVP 1Q live migration testing: PASS > (10)PVP 1Q cross numa node live migration testing: PASS > (11)Guest with ovs+dpdk+vhost-user 1Q live migration testing: PASS > (12)Guest with ovs+dpdk+vhost-user 1Q live migration testing (2M): PASS > (13)Guest with ovs+dpdk+vhost-user 2Q live migration testing: PASS > (14) Allocate memory from the NUMA node which Virtio device locates: PASS > (15)Host PF + DPDK testing: PASS > (16)Host VF + DPDK testing: PASS > > Versions: > > kernel 4.18 > qemu 5.1 > dpdk: git://dpdk.org/dpdk-stable remotes/origin/19.11 > # git log > commit bb00cdf6cdcf690001ef4154a5aaad8c0dc5d9b9 (HEAD, tag: v19.11.4-rc1, > origin/19.11) > Author: Luca Boccassi > Date: Tue Aug 18 19:00:36 2020 +0100 > > version: 19.11.4-rc1 > > Signed-off-by: Luca Boccassi > > > NICs: X540-AT2 NIC(ixgbe, 10G) > > Best regards, > > Pei Thank you! > > - Original Message - > > From: "Luca Boccassi" > > To: sta...@dpdk.org > > Cc: dev@dpdk.org, "Abhishek Marathe" , > > "Akhil Goyal" , "Ali > > Alnubani" , "benjamin walker" > > , "David Christensen" > > , "Hemant Agrawal" , "Ian > > Stokes" , "Jerin > > Jacob" , "John McNamara" , > > "Ju-Hyoung Lee" , > > "Kevin Traynor" , "Pei Zhang" , > > "pingx yu" , "qian q > > xu" , "Raslan Darawsheh" , > > "Thomas Monjalon" , "yuan > > peng" , "zhaoyan chen" > > Sent: Wednesday, August 19, 2020 2:12:22 AM > > Subject: 19.11.4 patches review and test > > > > Hi all, > > > > Here is a list of patches targeted for stable release 19.11.4. > > > > The planned date for the final release is August 31st. > > > > Please help with testing and validation of your use cases and report > > any issues/results with reply-all to this mail. For the final release > > the fixes and reported validations will be added to the release notes. > > > > A release candidate tarball can be found at: > > > > https://dpdk.org/browse/dpdk-stable/tag/?id=v19.11.4-rc1 > > > > These patches are located at branch 19.11 of dpdk-stable repo: > > https://dpdk.org/browse/dpdk-stable/ > > > > Thanks. > > > > Luca Boccassi > > > > --- > > Adam Dybkowski (7): > > test/crypto: fix asymmetric session mempool creation > > crypto/qat: fix AES-XTS capabilities > > crypto/qat: handle mixed hash-cipher requests on GEN3 > > test/crypto: add mixed encypted-digest > > common/qat: get firmware version > > crypto/qat: handle mixed hash-cipher on GEN2 > > common/qat: fix uninitialized variable > > > > Akhil Goyal (4): > > crypto/dpaax_sec: fix 18-bit PDCP cases with HFN override > > crypto/dpaax_sec: fix inline query for descriptors > > crypto/dpaa2_sec: fix HFN override > > common/dpaax: fix 12-bit null auth case > > > > Alan Dewar (1): > > sched: fix port time rounding > > > > Alexander Kozyrev (3): > > net/mlx5: fix vectorized Rx burst termination > > net/mlx5: do not select legacy MPW implicitly > > net/mlx5: fix descriptors number adjustment > > > > Ali Alnubani (2): > > net/mlx5: fix typos in meter error messages > > doc: fix a typo in mlx5 guide > > > > Alvin Zhang (1): > > net/i40e: fix queue pairs configuration in VF > > > > Andrew Rybchenko (1): > > net/sfc: do not enforce hash offload in RSS multi-queue > > > > Ankur Dwivedi (2): > > common/cpt: fix encryption offset > > test/crypto: change cipher offset for ESN vector > > > > Apeksha Gupta (2): > > app/eventdev: fix capability check in pipeline ATQ test > > event/dpaa2: add all-types queue capability flag > > > > Archana Muniganti (2): > > examples/fips_validation: fix TDES interim callback > > examples/fips_validation: fix count overwrite for TDES > > > > Archit Pandey (1): > > sched: fix 64-bit rate > > > > Arek Kusztal (4): > > drivers/qat: improve multi process on qat > > drivers/qat: add multi process handling of driver id > > drivers/qat: add handling of capabilities in multi process > > crypto/qat: add minimum enq threshold > > > > Ayuj Verma (1): > > examples/fips_validation: fix parsing of TDES vectors > > > > Bruce Richardson (12): > > test: fix build with ring
[dpdk-dev] [PATCH] bus/pci: support segment value as address domain on Windows
Set the domain value for rte_pci_addr probing on Windows to the value of the PCI segment returned by SPDRP_BUSNUMBER. Signed-off-by: Tal Shnaiderman --- drivers/bus/pci/windows/pci.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/bus/pci/windows/pci.c b/drivers/bus/pci/windows/pci.c index 489aa7902a..a40acec609 100644 --- a/drivers/bus/pci/windows/pci.c +++ b/drivers/bus/pci/windows/pci.c @@ -195,8 +195,8 @@ get_device_pci_address(HDEVINFO dev_info, return -1; } - addr->domain = 0; - addr->bus = bus_num; + addr->domain = bus_num >> 8; + addr->bus = bus_num & 0xff; addr->devid = dev_and_func >> 16; addr->function = dev_and_func & 0x; return 0; -- 2.16.1.windows.4
[dpdk-dev] [PATCH 1/2] config: remove explicit undefinition of unset values
Rather than explicitly clearing any setting of undefined values in our rte_config.h file, it's better to instead just add a comment that the value is not set. Using a comment allows the user to set the value using CFLAGS or similar mechanism without the config file clearing the value again. The text used " is not set" is modelled after the kernel approach of doing the same thing. Signed-off-by: Bruce Richardson --- config/rte_config.h | 12 ++-- 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/config/rte_config.h b/config/rte_config.h index 9bb915347c..1c5a86d6a7 100644 --- a/config/rte_config.h +++ b/config/rte_config.h @@ -85,17 +85,17 @@ /* ip_fragmentation defines */ #define RTE_LIBRTE_IP_FRAG_MAX_FRAG 4 -#undef RTE_LIBRTE_IP_FRAG_TBL_STAT +// RTE_LIBRTE_IP_FRAG_TBL_STAT is not set /* rte_power defines */ #define RTE_MAX_LCORE_FREQS 64 /* rte_sched defines */ -#undef RTE_SCHED_RED -#undef RTE_SCHED_COLLECT_STATS -#undef RTE_SCHED_SUBPORT_TC_OV +// RTE_SCHED_RED is not set +// RTE_SCHED_COLLECT_STATS is not set +// RTE_SCHED_SUBPORT_TC_OV is not set #define RTE_SCHED_PORT_N_GRINDERS 8 -#undef RTE_SCHED_VECTOR +// RTE_SCHED_VECTOR is not set /* KNI defines */ #define RTE_KNI_PREEMPT_DEFAULT 1 @@ -123,7 +123,7 @@ /* i40e defines */ #define RTE_LIBRTE_I40E_RX_ALLOW_BULK_ALLOC 1 -#undef RTE_LIBRTE_I40E_16BYTE_RX_DESC +// RTE_LIBRTE_I40E_16BYTE_RX_DESC is not set #define RTE_LIBRTE_I40E_QUEUE_NUM_PER_PF 64 #define RTE_LIBRTE_I40E_QUEUE_NUM_PER_VF 4 #define RTE_LIBRTE_I40E_QUEUE_NUM_PER_VM 4 -- 2.25.1
[dpdk-dev] [PATCH 2/2] config: allow overriding some build defaults
In case a developer uses CFLAGS to set different default values for the defines in the rte_config.h file, use #ifndef / #endif guards around the setting of those values. For those lines just "defining" a macro without assigning it a value to be used by code, drop the value argument (where possible) to make it clearer that that is what is happening, since those don't need the #ifdef guard. Signed-off-by: Bruce Richardson --- config/rte_config.h | 110 +++- 1 file changed, 99 insertions(+), 11 deletions(-) diff --git a/config/rte_config.h b/config/rte_config.h index 1c5a86d6a7..f39da76c13 100644 --- a/config/rte_config.h +++ b/config/rte_config.h @@ -1,5 +1,5 @@ /* SPDX-License-Identifier: BSD-3-Clause - * Copyright(c) 2017 Intel Corporation + * Copyright(c) 2017-2020 Intel Corporation */ /** @@ -20,10 +20,10 @@ /* legacy defines */ #ifdef RTE_EXEC_ENV_LINUX -#define RTE_EXEC_ENV_LINUXAPP 1 +#define RTE_EXEC_ENV_LINUXAPP #endif #ifdef RTE_EXEC_ENV_FREEBSD -#define RTE_EXEC_ENV_BSDAPP 1 +#define RTE_EXEC_ENV_BSDAPP #endif /* String that appears before the version number */ @@ -32,107 +32,195 @@ /** library defines / /* EAL defines */ +#define RTE_BACKTRACE +#ifndef RTE_MAX_HEAPS #define RTE_MAX_HEAPS 32 +#endif +#ifndef RTE_MAX_MEMSEG_LISTS #define RTE_MAX_MEMSEG_LISTS 128 +#endif +#ifndef RTE_MAX_MEMSEG_PER_LIST #define RTE_MAX_MEMSEG_PER_LIST 8192 +#endif +#ifndef RTE_MAX_MEM_MB_PER_LIST #define RTE_MAX_MEM_MB_PER_LIST 32768 +#endif +#ifndef RTE_MAX_MEMSEG_PER_TYPE #define RTE_MAX_MEMSEG_PER_TYPE 32768 +#endif +#ifndef RTE_MAX_MEM_MB_PER_TYPE #define RTE_MAX_MEM_MB_PER_TYPE 65536 +#endif +#ifndef RTE_MAX_MEMZONE #define RTE_MAX_MEMZONE 2560 +#endif +#ifndef RTE_MAX_TAILQ #define RTE_MAX_TAILQ 32 +#endif +#ifndef RTE_LOG_DP_LEVEL #define RTE_LOG_DP_LEVEL RTE_LOG_INFO -#define RTE_BACKTRACE 1 +#endif +#ifndef RTE_MAX_VFIO_CONTAINERS #define RTE_MAX_VFIO_CONTAINERS 64 +#endif /* bsd module defines */ +#ifndef RTE_CONTIGMEM_MAX_NUM_BUFS #define RTE_CONTIGMEM_MAX_NUM_BUFS 64 +#endif +#ifndef RTE_CONTIGMEM_DEFAULT_NUM_BUFS #define RTE_CONTIGMEM_DEFAULT_NUM_BUFS 1 +#endif +#ifndef RTE_CONTIGMEM_DEFAULT_BUF_SIZE #define RTE_CONTIGMEM_DEFAULT_BUF_SIZE (512*1024*1024) +#endif /* mempool defines */ +#ifndef RTE_MEMPOOL_CACHE_MAX_SIZE #define RTE_MEMPOOL_CACHE_MAX_SIZE 512 +#endif /* mbuf defines */ +#define RTE_MBUF_REFCNT_ATOMIC +#ifndef RTE_MBUF_DEFAULT_MEMPOOL_OPS #define RTE_MBUF_DEFAULT_MEMPOOL_OPS "ring_mp_mc" -#define RTE_MBUF_REFCNT_ATOMIC 1 +#endif +#ifndef RTE_PKTMBUF_HEADROOM #define RTE_PKTMBUF_HEADROOM 128 +#endif /* ether defines */ +#define RTE_ETHDEV_RXTX_CALLBACKS +#ifndef RTE_MAX_QUEUES_PER_PORT #define RTE_MAX_QUEUES_PER_PORT 1024 +#endif +#ifndef RTE_ETHDEV_QUEUE_STAT_CNTRS #define RTE_ETHDEV_QUEUE_STAT_CNTRS 16 -#define RTE_ETHDEV_RXTX_CALLBACKS 1 +#endif /* cryptodev defines */ +#ifndef RTE_CRYPTO_MAX_DEVS #define RTE_CRYPTO_MAX_DEVS 64 +#endif +#ifndef RTE_CRYPTODEV_NAME_LEN #define RTE_CRYPTODEV_NAME_LEN 64 +#endif /* compressdev defines */ +#ifndef RTE_COMPRESS_MAX_DEVS #define RTE_COMPRESS_MAX_DEVS 64 +#endif /* regexdev defines */ +#ifndef RTE_MAX_REGEXDEV_DEVS #define RTE_MAX_REGEXDEV_DEVS 32 +#endif /* eventdev defines */ +#ifndef RTE_EVENT_MAX_DEVS #define RTE_EVENT_MAX_DEVS 16 +#endif +#ifndef RTE_EVENT_MAX_QUEUES_PER_DEV #define RTE_EVENT_MAX_QUEUES_PER_DEV 64 +#endif +#ifndef RTE_EVENT_TIMER_ADAPTER_NUM_MAX #define RTE_EVENT_TIMER_ADAPTER_NUM_MAX 32 +#endif +#ifndef RTE_EVENT_ETH_INTR_RING_SIZE #define RTE_EVENT_ETH_INTR_RING_SIZE 1024 +#endif +#ifndef RTE_EVENT_CRYPTO_ADAPTER_MAX_INSTANCE #define RTE_EVENT_CRYPTO_ADAPTER_MAX_INSTANCE 32 +#endif +#ifndef RTE_EVENT_ETH_TX_ADAPTER_MAX_INSTANCE #define RTE_EVENT_ETH_TX_ADAPTER_MAX_INSTANCE 32 +#endif /* rawdev defines */ +#ifndef RTE_RAWDEV_MAX_DEVS #define RTE_RAWDEV_MAX_DEVS 64 +#endif /* ip_fragmentation defines */ +#ifndef RTE_LIBRTE_IP_FRAG_MAX_FRAG #define RTE_LIBRTE_IP_FRAG_MAX_FRAG 4 +#endif // RTE_LIBRTE_IP_FRAG_TBL_STAT is not set /* rte_power defines */ +#ifndef RTE_MAX_LCORE_FREQS #define RTE_MAX_LCORE_FREQS 64 +#endif /* rte_sched defines */ // RTE_SCHED_RED is not set // RTE_SCHED_COLLECT_STATS is not set // RTE_SCHED_SUBPORT_TC_OV is not set +#ifndef RTE_SCHED_PORT_N_GRINDERS #define RTE_SCHED_PORT_N_GRINDERS 8 +#endif // RTE_SCHED_VECTOR is not set /* KNI defines */ -#define RTE_KNI_PREEMPT_DEFAULT 1 +#define RTE_KNI_PREEMPT_DEFAULT /* rte_graph defines */ -#define RTE_GRAPH_BURST_SIZE 256 #define RTE_LIBRTE_GRAPH_STATS 1 +#ifndef RTE_GRAPH_BURST_SIZE +#define RTE_GRAPH_BURST_SIZE 256 +#endif /** driver defines / /* QuickAssist device */ /* Max. number of QuickAssist devices which can be attached */ +#ifndef RTE_PMD_QAT_MAX_PCI_DEVICES #define RTE_PMD_QAT_MAX_PCI_DEVICES 48 +#endif +#ifndef RTE_PMD_QAT_COMP_SGL_MAX_SEGMENTS #define RTE_PMD_QAT_COMP_
[dpdk-dev] [PATCH 02/11] net/hns3: get dev specifications from firmware
From: "Wei Hu (Xavier)"
This patch adds getting PF/VF device specifications from firmware.
Signed-off-by: Wei Hu (Xavier)
Signed-off-by: Hongbo Zheng
---
drivers/net/hns3/hns3_cmd.h | 15
drivers/net/hns3/hns3_dcb.c | 1 -
drivers/net/hns3/hns3_dcb.h | 2 +
drivers/net/hns3/hns3_ethdev.c| 61 ++-
drivers/net/hns3/hns3_ethdev.h| 6 +++
drivers/net/hns3/hns3_ethdev_vf.c | 59 +-
6 files changed, 141 insertions(+), 3 deletions(-)
diff --git a/drivers/net/hns3/hns3_cmd.h b/drivers/net/hns3/hns3_cmd.h
index a13b799f4..65aa8bad8 100644
--- a/drivers/net/hns3/hns3_cmd.h
+++ b/drivers/net/hns3/hns3_cmd.h
@@ -93,6 +93,8 @@ enum hns3_opcode_type {
HNS3_OPC_QUERY_32_BIT_REG = 0x0041,
HNS3_OPC_QUERY_64_BIT_REG = 0x0042,
+ HNS3_OPC_QUERY_DEV_SPECS= 0x0050,
+
/* MAC command */
HNS3_OPC_CONFIG_MAC_MODE= 0x0301,
HNS3_OPC_QUERY_LINK_STATUS = 0x0307,
@@ -805,6 +807,19 @@ struct hns3_reset_cmd {
uint8_t rsv[22];
};
+#define HNS3_QUERY_DEV_SPECS_BD_NUM4
+struct hns3_dev_specs_0_cmd {
+ uint32_t rsv0;
+ uint32_t mac_entry_num;
+ uint32_t mng_entry_num;
+ uint16_t rss_ind_tbl_size;
+ uint16_t rss_key_size;
+ uint16_t intr_ql_max;
+ uint8_t max_non_tso_bd_num;
+ uint8_t rsv1;
+ uint32_t max_tm_rate;
+};
+
#define HNS3_MAX_TQP_NUM_PER_FUNC 64
#define HNS3_DEFAULT_TX_BUF0x4000/* 16k bytes */
#define HNS3_TOTAL_PKT_BUF 0x108000 /* 1.03125M bytes */
diff --git a/drivers/net/hns3/hns3_dcb.c b/drivers/net/hns3/hns3_dcb.c
index 02628b6b6..c1be49e0f 100644
--- a/drivers/net/hns3/hns3_dcb.c
+++ b/drivers/net/hns3/hns3_dcb.c
@@ -19,7 +19,6 @@
#define HNS3_SHAPER_BS_U_DEF 5
#define HNS3_SHAPER_BS_S_DEF 20
#define BW_MAX_PERCENT 100
-#define HNS3_ETHER_MAX_RATE10
/*
* hns3_shaper_para_calc: calculate ir parameter for the shaper
diff --git a/drivers/net/hns3/hns3_dcb.h b/drivers/net/hns3/hns3_dcb.h
index 9c2c5f21c..1636c5ae8 100644
--- a/drivers/net/hns3/hns3_dcb.h
+++ b/drivers/net/hns3/hns3_dcb.h
@@ -5,6 +5,8 @@
#ifndef _HNS3_DCB_H_
#define _HNS3_DCB_H_
+#define HNS3_ETHER_MAX_RATE10
+
/* MAC Pause */
#define HNS3_TX_MAC_PAUSE_EN_MSK BIT(0)
#define HNS3_RX_MAC_PAUSE_EN_MSK BIT(1)
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index 44fd69fa1..951f26d42 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -2809,6 +2809,51 @@ hns3_parse_speed(int speed_cmd, uint32_t *speed)
return 0;
}
+static void
+hns3_set_default_dev_specifications(struct hns3_hw *hw)
+{
+ hw->max_non_tso_bd_num = HNS3_MAX_NON_TSO_BD_PER_PKT;
+ hw->rss_ind_tbl_size = HNS3_RSS_IND_TBL_SIZE;
+ hw->rss_key_size = HNS3_RSS_KEY_SIZE;
+ hw->max_tm_rate = HNS3_ETHER_MAX_RATE;
+}
+
+static void
+hns3_parse_dev_specifications(struct hns3_hw *hw, struct hns3_cmd_desc *desc)
+{
+ struct hns3_dev_specs_0_cmd *req0;
+
+ req0 = (struct hns3_dev_specs_0_cmd *)desc[0].data;
+
+ hw->max_non_tso_bd_num = req0->max_non_tso_bd_num;
+ hw->rss_ind_tbl_size = rte_le_to_cpu_16(req0->rss_ind_tbl_size);
+ hw->rss_key_size = rte_le_to_cpu_16(req0->rss_key_size);
+ hw->max_tm_rate = rte_le_to_cpu_32(req0->max_tm_rate);
+}
+
+static int
+hns3_query_dev_specifications(struct hns3_hw *hw)
+{
+ struct hns3_cmd_desc desc[HNS3_QUERY_DEV_SPECS_BD_NUM];
+ int ret;
+ int i;
+
+ for (i = 0; i < HNS3_QUERY_DEV_SPECS_BD_NUM - 1; i++) {
+ hns3_cmd_setup_basic_desc(&desc[i], HNS3_OPC_QUERY_DEV_SPECS,
+ true);
+ desc[i].flag |= rte_cpu_to_le_16(HNS3_CMD_FLAG_NEXT);
+ }
+ hns3_cmd_setup_basic_desc(&desc[i], HNS3_OPC_QUERY_DEV_SPECS, true);
+
+ ret = hns3_cmd_send(hw, desc, HNS3_QUERY_DEV_SPECS_BD_NUM);
+ if (ret)
+ return ret;
+
+ hns3_parse_dev_specifications(hw, desc);
+
+ return 0;
+}
+
static int
hns3_get_capability(struct hns3_hw *hw)
{
@@ -2832,11 +2877,25 @@ hns3_get_capability(struct hns3_hw *hw)
ret = rte_pci_read_config(pci_dev, &revision, HNS3_PCI_REVISION_ID_LEN,
HNS3_PCI_REVISION_ID);
if (ret != HNS3_PCI_REVISION_ID_LEN) {
- PMD_INIT_LOG(ERR, "failed to read pci revision id: %d", ret);
+ PMD_INIT_LOG(ERR, "failed to read pci revision id, ret = %d",
+ret);
return -EIO;
}
hw->revision = revision;
+ if (revision < PCI_REVISION_ID_HIP09_A) {
+ hns3_set_default_dev_specifications(hw);
+ return 0;
+ }
+
+ ret = hns3_query_dev_specifications(hw);
+ if (ret) {
+ PMD_INIT_LOG(ERR,
+
[dpdk-dev] [PATCH 01/11] net/hns3: get device capability from firmware
From: "Wei Hu (Xavier)"
This patch adds getting device capabilities from firmware, so driver can
supply differnet cpabilities and specifications to upper level
applications base on differnet versions of hardware network engine.
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_cmd.c| 36 --
drivers/net/hns3/hns3_cmd.h| 19 +-
drivers/net/hns3/hns3_ethdev.c | 3 ---
drivers/net/hns3/hns3_ethdev.h | 29 +++
4 files changed, 77 insertions(+), 10 deletions(-)
diff --git a/drivers/net/hns3/hns3_cmd.c b/drivers/net/hns3/hns3_cmd.c
index cbb09887c..0299072ef 100644
--- a/drivers/net/hns3/hns3_cmd.c
+++ b/drivers/net/hns3/hns3_cmd.c
@@ -426,8 +426,29 @@ hns3_cmd_send(struct hns3_hw *hw, struct hns3_cmd_desc
*desc, int num)
return retval;
}
+static void hns3_parse_capability(struct hns3_hw *hw,
+ struct hns3_query_version_cmd *cmd)
+{
+ uint32_t caps = rte_le_to_cpu_32(cmd->caps[0]);
+
+ if (hns3_get_bit(caps, HNS3_CAPS_UDP_GSO_B))
+ hns3_set_bit(hw->capability, HNS3_DEV_SUPPORT_UDP_GSO_B, 1);
+ if (hns3_get_bit(caps, HNS3_CAPS_ADQ_B))
+ hns3_set_bit(hw->capability, HNS3_DEV_SUPPORT_ADQ_B, 1);
+ if (hns3_get_bit(caps, HNS3_CAPS_PTP_B))
+ hns3_set_bit(hw->capability, HNS3_DEV_SUPPORT_PTP_B, 1);
+ if (hns3_get_bit(caps, HNS3_CAPS_TX_PUSH_B))
+ hns3_set_bit(hw->capability, HNS3_DEV_SUPPORT_TX_PUSH_B, 1);
+ if (hns3_get_bit(caps, HNS3_CAPS_PHY_IMP_B))
+ hns3_set_bit(hw->capability, HNS3_DEV_SUPPORT_COPPER_B, 1);
+ if (hns3_get_bit(caps, HNS3_CAPS_TQP_TXRX_INDEP_B))
+ hns3_set_bit(hw->capability, HNS3_CAPS_TQP_TXRX_INDEP_B, 1);
+ if (hns3_get_bit(caps, HNS3_CAPS_STASH_B))
+ hns3_set_bit(hw->capability, HNS3_DEV_SUPPORT_STASH_B, 1);
+}
+
static enum hns3_cmd_status
-hns3_cmd_query_firmware_version(struct hns3_hw *hw, uint32_t *version)
+hns3_cmd_query_firmware_version_and_capability(struct hns3_hw *hw)
{
struct hns3_query_version_cmd *resp;
struct hns3_cmd_desc desc;
@@ -438,10 +459,13 @@ hns3_cmd_query_firmware_version(struct hns3_hw *hw,
uint32_t *version)
/* Initialize the cmd function */
ret = hns3_cmd_send(hw, &desc, 1);
- if (ret == 0)
- *version = rte_le_to_cpu_32(resp->firmware);
+ if (ret)
+ return ret;
- return ret;
+ hw->fw_version = rte_le_to_cpu_32(resp->firmware);
+ hns3_parse_capability(hw, resp);
+
+ return 0;
}
int
@@ -519,13 +543,13 @@ hns3_cmd_init(struct hns3_hw *hw)
}
rte_atomic16_clear(&hw->reset.disable_cmd);
- ret = hns3_cmd_query_firmware_version(hw, &version);
+ ret = hns3_cmd_query_firmware_version_and_capability(hw);
if (ret) {
PMD_INIT_LOG(ERR, "firmware version query failed %d", ret);
goto err_cmd_init;
}
- hw->fw_version = version;
+ version = hw->fw_version;
PMD_INIT_LOG(INFO, "The firmware version is %lu.%lu.%lu.%lu",
hns3_get_field(version, HNS3_FW_VERSION_BYTE3_M,
HNS3_FW_VERSION_BYTE3_S),
diff --git a/drivers/net/hns3/hns3_cmd.h b/drivers/net/hns3/hns3_cmd.h
index d70f42e5d..a13b799f4 100644
--- a/drivers/net/hns3/hns3_cmd.h
+++ b/drivers/net/hns3/hns3_cmd.h
@@ -264,9 +264,26 @@ struct hns3_rx_priv_buff_cmd {
#define HNS3_FW_VERSION_BYTE1_MGENMASK(15, 8)
#define HNS3_FW_VERSION_BYTE0_S0
#define HNS3_FW_VERSION_BYTE0_MGENMASK(7, 0)
+
+enum HNS3_CAPS_BITS {
+ HNS3_CAPS_UDP_GSO_B,
+ HNS3_CAPS_ATR_B,
+ HNS3_CAPS_ADQ_B,
+ HNS3_CAPS_PTP_B,
+ HNS3_CAPS_INT_QL_B,
+ HNS3_CAPS_SIMPLE_BD_B,
+ HNS3_CAPS_TX_PUSH_B,
+ HNS3_CAPS_PHY_IMP_B,
+ HNS3_CAPS_TQP_TXRX_INDEP_B,
+ HNS3_CAPS_HW_PAD_B,
+ HNS3_CAPS_STASH_B,
+};
+#define HNS3_QUERY_CAP_LENGTH 3
struct hns3_query_version_cmd {
uint32_t firmware;
- uint32_t firmware_rsv[5];
+ uint32_t hardware;
+ uint32_t rsv;
+ uint32_t caps[HNS3_QUERY_CAP_LENGTH]; /* capabilities of device */
};
#define HNS3_RX_PRIV_EN_B 15
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index fab1914c3..44fd69fa1 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -2837,9 +2837,6 @@ hns3_get_capability(struct hns3_hw *hw)
}
hw->revision = revision;
- if (revision >= PCI_REVISION_ID_HIP09_A)
- hns3_set_bit(hw->capability, HNS3_DEV_SUPPORT_COPPER_B, 1);
-
return 0;
}
diff --git a/drivers/net/hns3/hns3_ethdev.h b/drivers/net/hns3/hns3_ethdev.h
index 0e665e59b..1914e588d 100644
--- a/drivers/net/hns3/hns3_ethdev.h
+++ b/drivers/net/hns3/hns3_ethdev.h
@@ -535,13 +535,42 @@ s
[dpdk-dev] [PATCH 00/11] updates for hns3 PMD driver
This series are features and fixes for hns3 PMD driver. Huisong Li (3): net/hns3: replace private macro with RTE MAX net/hns3: fix default MAC addr from firmware net/hns3: fix some incomplete command structures Wei Hu (Xavier) (8): net/hns3: get device capability from firmware net/hns3: get dev specifications from firmware net/hns3: compatibility issues about Rx interrupts net/hns3: compatibility issues about Tx padding short frame net/hns3: add more hardware error types net/hns3: support a maximun 256 FDIR counter net/hns3: change the log level to INFO net/hns3: fix Rx/Tx queue offload capability drivers/net/hns3/hns3_cmd.c | 36 +- drivers/net/hns3/hns3_cmd.h | 94 ++- drivers/net/hns3/hns3_dcb.c |1 - drivers/net/hns3/hns3_dcb.h | 14 +- drivers/net/hns3/hns3_ethdev.c| 187 - drivers/net/hns3/hns3_ethdev.h| 138 +++- drivers/net/hns3/hns3_ethdev_vf.c | 124 ++- drivers/net/hns3/hns3_fdir.c |5 + drivers/net/hns3/hns3_intr.c | 1236 - drivers/net/hns3/hns3_intr.h | 40 +- drivers/net/hns3/hns3_regs.h |7 + drivers/net/hns3/hns3_rxtx.c | 31 +- drivers/net/hns3/hns3_rxtx.h | 11 +- drivers/net/hns3/hns3_stats.c | 78 +- drivers/net/hns3/hns3_stats.h |2 + 15 files changed, 1666 insertions(+), 338 deletions(-) -- 2.27.0
[dpdk-dev] [PATCH 03/11] net/hns3: compatibility issues about Rx interrupts
From: "Wei Hu (Xavier)"
There are difference about queue's interrupt configurations for different
versions of hardware network engine, such as queue's interrupt mapping
mode, coalesce configuration, etc.
The following uses the configuration differences of the interrupt mapping
mode as an example.
1) For some versions of hardware network engine, such as kunpeng 920,
because of the hardware constraint, we need implement unmmapping
relationship configurations by binding all queues to the last interrupt
vector and reserving the last interrupt vector. This results in a
decrease of the maximum queues when upper applications call the
rte_eth_dev_configure API function to enable Rx interrupt.
2) And for another versions, sunch as kunpeng 930, hns3 PMD driver can
map/unmmap all interrupt vectors with queues when Rx interrupt is
enabled.
This patch resolves configuration differences about Rx interrupts based on
kunpeng 920 and kunpeng 930.
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_cmd.h | 14 +---
drivers/net/hns3/hns3_ethdev.c| 31 +
drivers/net/hns3/hns3_ethdev.h| 56 +++
drivers/net/hns3/hns3_ethdev_vf.c | 29 +---
drivers/net/hns3/hns3_regs.h | 4 +++
drivers/net/hns3/hns3_rxtx.c | 20 ++-
drivers/net/hns3/hns3_rxtx.h | 2 ++
7 files changed, 123 insertions(+), 33 deletions(-)
diff --git a/drivers/net/hns3/hns3_cmd.h b/drivers/net/hns3/hns3_cmd.h
index 65aa8bad8..c2b0361b6 100644
--- a/drivers/net/hns3/hns3_cmd.h
+++ b/drivers/net/hns3/hns3_cmd.h
@@ -368,21 +368,25 @@ struct hns3_func_status_cmd {
uint8_t rsv[2];
};
-#define HNS3_VEC_NUM_S 0
-#define HNS3_VEC_NUM_M GENMASK(7, 0)
+#define HNS3_PF_VEC_NUM_S 0
+#define HNS3_PF_VEC_NUM_M GENMASK(15, 0)
#define HNS3_MIN_VECTOR_NUM2 /* one for msi-x, another for IO */
struct hns3_pf_res_cmd {
uint16_t tqp_num;
uint16_t buf_size;
uint16_t msixcap_localid_ba_nic;
- uint16_t msixcap_localid_ba_rocee;
- uint16_t pf_intr_vector_number;
+ uint16_t nic_pf_intr_vector_number;
+ uint16_t roce_pf_intr_vector_number;
uint16_t pf_own_fun_number;
uint16_t tx_buf_size;
uint16_t dv_buf_size;
- uint32_t rsv[2];
+ uint16_t tqp_num_ext;
+ uint16_t roh_pf_intr_vector_number;
+ uint32_t rsv[1];
};
+#define HNS3_VF_VEC_NUM_S 0
+#define HNS3_VF_VEC_NUM_M GENMASK(7, 0)
struct hns3_vf_res_cmd {
uint16_t tqp_num;
uint16_t reserved;
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index 951f26d42..3cc1fbc32 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -2208,7 +2208,7 @@ hns3_bind_ring_with_vector(struct hns3_hw *hw, uint8_t
vector_id, bool mmap,
static int
hns3_init_ring_with_vector(struct hns3_hw *hw)
{
- uint8_t vec;
+ uint16_t vec;
int ret;
int i;
@@ -2219,27 +2219,23 @@ hns3_init_ring_with_vector(struct hns3_hw *hw)
* vector. In the initialization clearing the all hardware mapping
* relationship configurations between queues and interrupt vectors is
* needed, so some error caused by the residual configurations, such as
-* the unexpected Tx interrupt, can be avoid. Because of the hardware
-* constraints in hns3 hardware engine, we have to implement clearing
-* the mapping relationship configurations by binding all queues to the
-* last interrupt vector and reserving the last interrupt vector. This
-* method results in a decrease of the maximum queues when upper
-* applications call the rte_eth_dev_configure API function to enable
-* Rx interrupt.
+* the unexpected Tx interrupt, can be avoid.
*/
vec = hw->num_msi - 1; /* vector 0 for misc interrupt, not for queue */
- /* vec - 1: the last interrupt is reserved */
- hw->intr_tqps_num = vec > hw->tqps_num ? hw->tqps_num : vec - 1;
+ if (hw->intr.mapping_mode == HNS3_INTR_MAPPING_VEC_RSV_ONE)
+ vec = vec - 1; /* the last interrupt is reserved */
+ hw->intr_tqps_num = RTE_MIN(vec, hw->tqps_num);
for (i = 0; i < hw->intr_tqps_num; i++) {
/*
-* Set gap limiter and rate limiter configuration of queue's
-* interrupt.
+* Set gap limiter/rate limiter/quanity limiter algorithm
+* configuration for interrupt coalesce of queue's interrupt.
*/
hns3_set_queue_intr_gl(hw, i, HNS3_RING_GL_RX,
HNS3_TQP_INTR_GL_DEFAULT);
hns3_set_queue_intr_gl(hw, i, HNS3_RING_GL_TX,
HNS3_TQP_INTR_GL_DEFAULT);
hns3_set_queue_intr_rl(hw, i, HNS3_TQP_INTR_RL_DEFAULT);
+ hns3_set_queue_intr_
[dpdk-dev] [PATCH 05/11] net/hns3: add more hardware error types
From: "Wei Hu (Xavier)"
The new firmware adds the hardware error types reported by MSI-x mode.
These errors are defined as RAS errors in hardware and belong to a
different type from the MSI-x errors processed by the driver.
When hardware detects an hardware errors, which need to be handled with
the driver otherwise the device cannot run properly, it reports error
information through the MSI-x interrupt. After receiving the interrupt
reported by the hardware, the driver queries the error information and
identifies the error level, then rectifies the error. All errors will be
logged. In addition, the hardware may be reset at the function or global
level based on the error level. After the reset is complete, the hardware
will recovers to the normal status.
Signed-off-by: Hongbo Zheng
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_cmd.h| 46 +-
drivers/net/hns3/hns3_ethdev.c | 67 +-
drivers/net/hns3/hns3_ethdev.h | 34 +-
drivers/net/hns3/hns3_intr.c | 1236 ++--
drivers/net/hns3/hns3_intr.h | 40 +-
drivers/net/hns3/hns3_regs.h |3 +
drivers/net/hns3/hns3_stats.c | 78 +-
drivers/net/hns3/hns3_stats.h |2 +
8 files changed, 1249 insertions(+), 257 deletions(-)
diff --git a/drivers/net/hns3/hns3_cmd.h b/drivers/net/hns3/hns3_cmd.h
index c2b0361b6..87d60539d 100644
--- a/drivers/net/hns3/hns3_cmd.h
+++ b/drivers/net/hns3/hns3_cmd.h
@@ -100,7 +100,6 @@ enum hns3_opcode_type {
HNS3_OPC_QUERY_LINK_STATUS = 0x0307,
HNS3_OPC_CONFIG_MAX_FRM_SIZE= 0x0308,
HNS3_OPC_CONFIG_SPEED_DUP = 0x0309,
- HNS3_MAC_COMMON_INT_EN = 0x030E,
/* PFC/Pause commands */
HNS3_OPC_CFG_MAC_PAUSE_EN = 0x0701,
@@ -153,10 +152,6 @@ enum hns3_opcode_type {
HNS3_OPC_RX_COM_THRD_ALLOC = 0x0904,
HNS3_OPC_RX_COM_WL_ALLOC= 0x0905,
- /* SSU module INT commands */
- HNS3_SSU_ECC_INT_CMD= 0x0989,
- HNS3_SSU_COMMON_INT_CMD = 0x098C,
-
/* TQP management command */
HNS3_OPC_SET_TQP_MAP= 0x0A01,
@@ -166,11 +161,6 @@ enum hns3_opcode_type {
HNS3_OPC_CFG_COM_TQP_QUEUE = 0x0B20,
HNS3_OPC_RESET_TQP_QUEUE= 0x0B22,
- /* PPU module intr commands */
- HNS3_PPU_MPF_ECC_INT_CMD= 0x0B40,
- HNS3_PPU_MPF_OTHER_INT_CMD = 0x0B41,
- HNS3_PPU_PF_OTHER_INT_CMD = 0x0B42,
-
/* TSO command */
HNS3_OPC_TSO_GENERIC_CONFIG = 0x0C01,
HNS3_OPC_GRO_GENERIC_CONFIG = 0x0C10,
@@ -216,17 +206,30 @@ enum hns3_opcode_type {
HNS3_OPC_SFP_GET_SPEED = 0x7104,
/* Interrupts commands */
- HNS3_OPC_ADD_RING_TO_VECTOR = 0x1503,
- HNS3_OPC_DEL_RING_TO_VECTOR = 0x1504,
+ HNS3_OPC_ADD_RING_TO_VECTOR = 0x1503,
+ HNS3_OPC_DEL_RING_TO_VECTOR = 0x1504,
/* Error INT commands */
- HNS3_QUERY_MSIX_INT_STS_BD_NUM = 0x1513,
- HNS3_QUERY_CLEAR_ALL_MPF_MSIX_INT = 0x1514,
- HNS3_QUERY_CLEAR_ALL_PF_MSIX_INT= 0x1515,
-
- /* PPP module intr commands */
- HNS3_PPP_CMD0_INT_CMD = 0x2100,
- HNS3_PPP_CMD1_INT_CMD = 0x2101,
+ HNS3_OPC_MAC_COMMON_INT_EN = 0x030E,
+ HNS3_OPC_TM_SCH_ECC_INT_EN = 0x0829,
+ HNS3_OPC_SSU_ECC_INT_CMD= 0x0989,
+ HNS3_OPC_SSU_COMMON_INT_CMD = 0x098C,
+ HNS3_OPC_PPU_MPF_ECC_INT_CMD= 0x0B40,
+ HNS3_OPC_PPU_MPF_OTHER_INT_CMD = 0x0B41,
+ HNS3_OPC_PPU_PF_OTHER_INT_CMD = 0x0B42,
+ HNS3_OPC_COMMON_ECC_INT_CFG = 0x1505,
+ HNS3_OPC_QUERY_RAS_INT_STS_BD_NUM = 0x1510,
+ HNS3_OPC_QUERY_CLEAR_MPF_RAS_INT= 0x1511,
+ HNS3_OPC_QUERY_CLEAR_PF_RAS_INT = 0x1512,
+ HNS3_OPC_QUERY_MSIX_INT_STS_BD_NUM = 0x1513,
+ HNS3_OPC_QUERY_CLEAR_ALL_MPF_MSIX_INT = 0x1514,
+ HNS3_OPC_QUERY_CLEAR_ALL_PF_MSIX_INT= 0x1515,
+ HNS3_OPC_IGU_EGU_TNL_INT_EN = 0x1803,
+ HNS3_OPC_IGU_COMMON_INT_EN = 0x1806,
+ HNS3_OPC_TM_QCN_MEM_INT_CFG = 0x1A14,
+ HNS3_OPC_PPP_CMD0_INT_CMD = 0x2100,
+ HNS3_OPC_PPP_CMD1_INT_CMD = 0x2101,
+ HNS3_OPC_NCSI_INT_EN= 0x2401,
};
#define HNS3_CMD_FLAG_IN BIT(0)
@@ -236,6 +239,11 @@ enum hns3_opcode_type {
#define HNS3_CMD_FLAG_NO_INTR BIT(4)
#define HNS3_CMD_FLAG_ERR_INTR BIT(5)
+#define HNS3_MPF_RAS_INT_MIN_BD_NUM10
+#define HNS3_PF_RAS_INT_MIN_BD_NUM 4
+#define HNS3_MPF_MSIX_INT_MIN_BD_NUM 10
+#define HNS3_PF_MSIX_INT_MIN_BD_NUM4
+
#define HNS3_BUF_SIZE_UNIT 256
#define HNS3_BUF_MUL_BY2
#define HNS3_BUF_DIV_BY2
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index 4
[dpdk-dev] [PATCH 07/11] net/hns3: replace private macro with RTE MAX
From: Huisong Li
This patch uses RTE_MAX function in DPDK lib to replace the private
macro named max_t in hns3 PMD driver.
Signed-off-by: Huisong Li
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_ethdev.c | 5 ++---
drivers/net/hns3/hns3_ethdev.h | 5 -
2 files changed, 2 insertions(+), 8 deletions(-)
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index b9ee11413..fca035d4f 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -3285,7 +3285,7 @@ hns3_is_rx_buf_ok(struct hns3_hw *hw, struct
hns3_pkt_buf_alloc *buf_alloc,
+ pf->dv_buf_size;
shared_buf_tc = tc_num * aligned_mps + aligned_mps;
- shared_std = roundup(max_t(uint32_t, shared_buf_min, shared_buf_tc),
+ shared_std = roundup(RTE_MAX(shared_buf_min, shared_buf_tc),
HNS3_BUF_SIZE_UNIT);
rx_priv = hns3_get_rx_priv_buff_alloced(buf_alloc);
@@ -3315,8 +3315,7 @@ hns3_is_rx_buf_ok(struct hns3_hw *hw, struct
hns3_pkt_buf_alloc *buf_alloc,
if (tc_num)
hi_thrd = hi_thrd / tc_num;
- hi_thrd = max_t(uint32_t, hi_thrd,
- HNS3_BUF_MUL_BY * aligned_mps);
+ hi_thrd = RTE_MAX(hi_thrd, HNS3_BUF_MUL_BY * aligned_mps);
hi_thrd = rounddown(hi_thrd, HNS3_BUF_SIZE_UNIT);
lo_thrd = hi_thrd - aligned_mps / HNS3_BUF_DIV_BY;
} else {
diff --git a/drivers/net/hns3/hns3_ethdev.h b/drivers/net/hns3/hns3_ethdev.h
index ca4cade42..a42479d64 100644
--- a/drivers/net/hns3/hns3_ethdev.h
+++ b/drivers/net/hns3/hns3_ethdev.h
@@ -708,11 +708,6 @@ struct hns3_adapter {
#define DIV_ROUND_UP(n, d) (((n) + (d) - 1) / (d))
-#define max_t(type, x, y) ({\
- type __max1 = (x); \
- type __max2 = (y); \
- __max1 > __max2 ? __max1 : __max2; })
-
static inline void hns3_write_reg(void *base, uint32_t reg, uint32_t value)
{
rte_write32(value, (volatile void *)((char *)base + reg));
--
2.27.0
[dpdk-dev] [PATCH 04/11] net/hns3: compatibility issues about Tx padding short frame
From: "Wei Hu (Xavier)"
There are differece about padding ultra-short frame in Tx procession for
different versions of hardware network engine.
If packet length is less than minimum packet length supported by hardware
in Tx direction, driver need to pad it to avoid error. The minimum packet
length in Tx direction is 33 based on kunpeng 920, and 9 based on
kunpeng 930.
Signed-off-by: Wei Hu (Xavier)
Signed-off-by: Chengchang Tang
---
drivers/net/hns3/hns3_ethdev.c| 2 ++
drivers/net/hns3/hns3_ethdev.h| 8 +++-
drivers/net/hns3/hns3_ethdev_vf.c | 2 ++
drivers/net/hns3/hns3_rxtx.c | 11 +++
drivers/net/hns3/hns3_rxtx.h | 9 -
5 files changed, 26 insertions(+), 6 deletions(-)
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index 3cc1fbc32..4797cfb2f 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -2884,6 +2884,7 @@ hns3_get_capability(struct hns3_hw *hw)
hw->intr.mapping_mode = HNS3_INTR_MAPPING_VEC_RSV_ONE;
hw->intr.coalesce_mode = HNS3_INTR_COALESCE_NON_QL;
hw->intr.gl_unit = HNS3_INTR_COALESCE_GL_UINT_2US;
+ hw->min_tx_pkt_len = HNS3_HIP08_MIN_TX_PKT_LEN;
return 0;
}
@@ -2898,6 +2899,7 @@ hns3_get_capability(struct hns3_hw *hw)
hw->intr.mapping_mode = HNS3_INTR_MAPPING_VEC_ALL;
hw->intr.coalesce_mode = HNS3_INTR_COALESCE_QL;
hw->intr.gl_unit = HNS3_INTR_COALESCE_GL_UINT_1US;
+ hw->min_tx_pkt_len = HNS3_HIP09_MIN_TX_PKT_LEN;
return 0;
}
diff --git a/drivers/net/hns3/hns3_ethdev.h b/drivers/net/hns3/hns3_ethdev.h
index b8eb7ddc1..a5405147d 100644
--- a/drivers/net/hns3/hns3_ethdev.h
+++ b/drivers/net/hns3/hns3_ethdev.h
@@ -58,7 +58,8 @@
#define HNS3_MAX_MTU (HNS3_MAX_FRAME_LEN - HNS3_ETH_OVERHEAD)
#define HNS3_DEFAULT_MTU 1500UL
#define HNS3_DEFAULT_FRAME_LEN (HNS3_DEFAULT_MTU + HNS3_ETH_OVERHEAD)
-#define HNS3_MIN_PKT_SIZE 60
+#define HNS3_HIP08_MIN_TX_PKT_LEN 33
+#define HNS3_HIP09_MIN_TX_PKT_LEN 9
#define HNS3_4_TCS 4
#define HNS3_8_TCS 8
@@ -464,6 +465,11 @@ struct hns3_hw {
uint32_t capability;
uint32_t max_tm_rate;
+ /*
+* The minimun length of the packet supported by hardware in the Tx
+* direction.
+*/
+ uint32_t min_tx_pkt_len;
struct hns3_queue_intr intr;
diff --git a/drivers/net/hns3/hns3_ethdev_vf.c
b/drivers/net/hns3/hns3_ethdev_vf.c
index 44657d362..3b2ba69bb 100644
--- a/drivers/net/hns3/hns3_ethdev_vf.c
+++ b/drivers/net/hns3/hns3_ethdev_vf.c
@@ -1133,6 +1133,7 @@ hns3vf_get_capability(struct hns3_hw *hw)
hw->intr.mapping_mode = HNS3_INTR_MAPPING_VEC_RSV_ONE;
hw->intr.coalesce_mode = HNS3_INTR_COALESCE_NON_QL;
hw->intr.gl_unit = HNS3_INTR_COALESCE_GL_UINT_2US;
+ hw->min_tx_pkt_len = HNS3_HIP08_MIN_TX_PKT_LEN;
return 0;
}
@@ -1147,6 +1148,7 @@ hns3vf_get_capability(struct hns3_hw *hw)
hw->intr.mapping_mode = HNS3_INTR_MAPPING_VEC_ALL;
hw->intr.coalesce_mode = HNS3_INTR_COALESCE_QL;
hw->intr.gl_unit = HNS3_INTR_COALESCE_GL_UINT_1US;
+ hw->min_tx_pkt_len = HNS3_HIP09_MIN_TX_PKT_LEN;
return 0;
}
diff --git a/drivers/net/hns3/hns3_rxtx.c b/drivers/net/hns3/hns3_rxtx.c
index d39576621..308d0a671 100644
--- a/drivers/net/hns3/hns3_rxtx.c
+++ b/drivers/net/hns3/hns3_rxtx.c
@@ -1915,6 +1915,7 @@ hns3_tx_queue_setup(struct rte_eth_dev *dev, uint16_t
idx, uint16_t nb_desc,
txq->configured = true;
txq->io_base = (void *)((char *)hw->io_base + HNS3_TQP_REG_OFFSET +
idx * HNS3_TQP_REG_SIZE);
+ txq->min_tx_pkt_len = hw->min_tx_pkt_len;
txq->over_length_pkt_cnt = 0;
txq->exceed_limit_bd_pkt_cnt = 0;
txq->exceed_limit_bd_reassem_fail = 0;
@@ -2743,14 +2744,16 @@ hns3_xmit_pkts(void *tx_queue, struct rte_mbuf
**tx_pkts, uint16_t nb_pkts)
}
/*
-* If packet length is less than minimum packet size, driver
-* need to pad it.
+* If packet length is less than minimum packet length supported
+* by hardware in Tx direction, driver need to pad it to avoid
+* error.
*/
- if (unlikely(rte_pktmbuf_pkt_len(tx_pkt) < HNS3_MIN_PKT_SIZE)) {
+ if (unlikely(rte_pktmbuf_pkt_len(tx_pkt) <
+ txq->min_tx_pkt_len)) {
uint16_t add_len;
char *appended;
- add_len = HNS3_MIN_PKT_SIZE -
+ add_len = txq->min_tx_pkt_len -
rte_pktmbuf_pkt_len(tx_pkt);
appended = rte_pktmbuf_append(
[dpdk-dev] [PATCH 06/11] net/hns3: support a maximun 256 FDIR counter
From: "Wei Hu (Xavier)"
The FDIR counter was used to count the number of FDIR hit, the maximum
number of the counter is 128 based on kunpeng 920, and it was 256 based
on kunpeng 930.
The firmware is responsible to allocate counters for diffent PF devices,
so the available counter number of one PF may be bigger than 128.
Currently, there are two places using the counter in hns3 PMD driver:
1. Configure the counter. Driver uses the command whose opcode is
HNS3_OPC_FD_AD_OP, now we extend one bit to hold the high bit of
counter-id in the command format.
2. Query the statistic information of the counter. Driver uses the command
whose opcode is HNS3_OPC_FD_COUNTER_OP, now the command already support
16-bit counter-id.
Signed-off-by: Chengwen Feng
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_fdir.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/drivers/net/hns3/hns3_fdir.c b/drivers/net/hns3/hns3_fdir.c
index 6ab439d06..5c3dd05f2 100644
--- a/drivers/net/hns3/hns3_fdir.c
+++ b/drivers/net/hns3/hns3_fdir.c
@@ -41,6 +41,8 @@
#define HNS3_FD_AD_WR_RULE_ID_B0
#define HNS3_FD_AD_RULE_ID_S 1
#define HNS3_FD_AD_RULE_ID_M GENMASK(13, 1)
+#define HNS3_FD_AD_COUNTER_HIGH_BIT 7
+#define HNS3_FD_AD_COUNTER_HIGH_BIT_B 26
enum HNS3_PORT_TYPE {
HOST_PORT,
@@ -424,6 +426,9 @@ static int hns3_fd_ad_config(struct hns3_hw *hw, int loc,
action->write_rule_id_to_bd);
hns3_set_field(ad_data, HNS3_FD_AD_RULE_ID_M, HNS3_FD_AD_RULE_ID_S,
action->rule_id);
+ /* set extend bit if counter_id is in [128 ~ 255] */
+ if (action->counter_id & BIT(HNS3_FD_AD_COUNTER_HIGH_BIT))
+ hns3_set_bit(ad_data, HNS3_FD_AD_COUNTER_HIGH_BIT_B, 1);
ad_data <<= HNS3_FD_AD_DATA_S;
hns3_set_bit(ad_data, HNS3_FD_AD_DROP_B, action->drop_packet);
hns3_set_bit(ad_data, HNS3_FD_AD_DIRECT_QID_B,
--
2.27.0
[dpdk-dev] [PATCH 11/11] net/hns3: fix some incomplete command structures
From: Huisong Li
The descriptor of the command between firmware and hns3 PMD driver
consists of 8-byte header and 24-byte data field. The contents sent to
firmware are packaged into a command structure as the data field of
command descriptor.
There are some command structures in hns3_dcb.h file that are less than
24 byte. So this patch fixes these incomplete command structures.
Fixes: 62e3ccc2b94c6 ("net/hns3: support flow control")
Cc: sta...@dpdk.org
Signed-off-by: Huisong Li
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_dcb.h | 12 +++-
1 file changed, 11 insertions(+), 1 deletion(-)
diff --git a/drivers/net/hns3/hns3_dcb.h b/drivers/net/hns3/hns3_dcb.h
index 1636c5ae8..557d88bc1 100644
--- a/drivers/net/hns3/hns3_dcb.h
+++ b/drivers/net/hns3/hns3_dcb.h
@@ -26,16 +26,19 @@ enum hns3_shap_bucket {
struct hns3_priority_weight_cmd {
uint8_t pri_id;
uint8_t dwrr;
+ uint8_t rsvd[22];
};
struct hns3_qs_weight_cmd {
uint16_t qs_id;
uint8_t dwrr;
+ uint8_t rsvd[21];
};
struct hns3_pg_weight_cmd {
uint8_t pg_id;
uint8_t dwrr;
+ uint8_t rsvd[22];
};
struct hns3_ets_tc_weight_cmd {
@@ -50,6 +53,7 @@ struct hns3_qs_to_pri_link_cmd {
uint8_t priority;
#define HNS3_DCB_QS_PRI_LINK_VLD_MSK BIT(0)
uint8_t link_vld;
+ uint8_t rsvd1[18];
};
struct hns3_nq_to_qs_link_cmd {
@@ -57,6 +61,7 @@ struct hns3_nq_to_qs_link_cmd {
uint16_t rsvd;
#define HNS3_DCB_Q_QS_LINK_VLD_MSK BIT(10)
uint16_t qset_id;
+ uint8_t rsvd1[18];
};
#define HNS3_DCB_SHAP_IR_B_MSK GENMASK(7, 0)
@@ -74,12 +79,14 @@ struct hns3_pri_shapping_cmd {
uint8_t pri_id;
uint8_t rsvd[3];
uint32_t pri_shapping_para;
+ uint32_t rsvd1[4];
};
struct hns3_pg_shapping_cmd {
uint8_t pg_id;
uint8_t rsvd[3];
uint32_t pg_shapping_para;
+ uint32_t rsvd1[4];
};
#define HNS3_BP_GRP_NUM32
@@ -92,16 +99,18 @@ struct hns3_bp_to_qs_map_cmd {
uint8_t rsvd[2];
uint8_t qs_group_id;
uint32_t qs_bit_map;
- uint32_t rsvd1;
+ uint32_t rsvd1[4];
};
struct hns3_pfc_en_cmd {
uint8_t tx_rx_en_bitmap;
uint8_t pri_en_bitmap;
+ uint8_t rsvd[22];
};
struct hns3_port_shapping_cmd {
uint32_t port_shapping_para;
+ uint32_t rsvd[5];
};
struct hns3_cfg_pause_param_cmd {
@@ -119,6 +128,7 @@ struct hns3_pg_to_pri_link_cmd {
uint8_t pg_id;
uint8_t rsvd1[3];
uint8_t pri_bit_map;
+ uint8_t rsvd2[19];
};
enum hns3_shaper_level {
--
2.27.0
[dpdk-dev] [PATCH 10/11] net/hns3: fix Rx/Tx queue offload capability
From: "Wei Hu (Xavier)"
According to rte_eth_rx_queue_setup and rte_eth_tx_queue_setup API
function, rx_queue_offload_capa and rx_offload_capa, tx_queue_offload_capa
and tx_offload_capa must be mutually exclusive in the '.dev_infos_get' ops
implementation function. Otherwise, rte_eth_rx_queue_setup or
rte_eth_tx_queue_setup will fail, if user uses rx_offload_capa and
tx_offload_capa obtained by calling the rte_eth_dev_info_get API function.
Currently, offload capabilities are enabled for all Rx/Tx queues in hns3
PF and VF PMD driver, and offload capability only applied in a Rx/Tx
queue is not supported. This patch fixes Rx/Tx queue offload capability.
Fixes: 1f5ca0b460cd67 ("net/hns3: support some device operations")
Fixes: a5475d61fa34b8 ("net/hns3: support VF")
Cc: sta...@dpdk.org
Signed-off-by: Huisong Li
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_ethdev.c| 5 +++--
drivers/net/hns3/hns3_ethdev_vf.c | 5 +++--
2 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index 14e4b9e35..281d8b928 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -2459,6 +2459,7 @@ hns3_dev_infos_get(struct rte_eth_dev *eth_dev, struct
rte_eth_dev_info *info)
info->max_mac_addrs = HNS3_UC_MACADDR_NUM;
info->max_mtu = info->max_rx_pktlen - HNS3_ETH_OVERHEAD;
info->max_lro_pkt_size = HNS3_MAX_LRO_SIZE;
+ info->rx_queue_offload_capa = 0;
info->rx_offload_capa = (DEV_RX_OFFLOAD_IPV4_CKSUM |
DEV_RX_OFFLOAD_TCP_CKSUM |
DEV_RX_OFFLOAD_UDP_CKSUM |
@@ -2472,7 +2473,7 @@ hns3_dev_infos_get(struct rte_eth_dev *eth_dev, struct
rte_eth_dev_info *info)
DEV_RX_OFFLOAD_JUMBO_FRAME |
DEV_RX_OFFLOAD_RSS_HASH |
DEV_RX_OFFLOAD_TCP_LRO);
- info->tx_queue_offload_capa = DEV_TX_OFFLOAD_MBUF_FAST_FREE;
+ info->tx_queue_offload_capa = 0;
info->tx_offload_capa = (DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM |
DEV_TX_OFFLOAD_IPV4_CKSUM |
DEV_TX_OFFLOAD_TCP_CKSUM |
@@ -2483,7 +2484,7 @@ hns3_dev_infos_get(struct rte_eth_dev *eth_dev, struct
rte_eth_dev_info *info)
DEV_TX_OFFLOAD_VXLAN_TNL_TSO |
DEV_TX_OFFLOAD_GRE_TNL_TSO |
DEV_TX_OFFLOAD_GENEVE_TNL_TSO |
-info->tx_queue_offload_capa |
+DEV_TX_OFFLOAD_MBUF_FAST_FREE |
hns3_txvlan_cap_get(hw));
info->rx_desc_lim = (struct rte_eth_desc_lim) {
diff --git a/drivers/net/hns3/hns3_ethdev_vf.c
b/drivers/net/hns3/hns3_ethdev_vf.c
index 7fd0e6a43..2f7a96826 100644
--- a/drivers/net/hns3/hns3_ethdev_vf.c
+++ b/drivers/net/hns3/hns3_ethdev_vf.c
@@ -903,6 +903,7 @@ hns3vf_dev_infos_get(struct rte_eth_dev *eth_dev, struct
rte_eth_dev_info *info)
info->max_mtu = info->max_rx_pktlen - HNS3_ETH_OVERHEAD;
info->max_lro_pkt_size = HNS3_MAX_LRO_SIZE;
+ info->rx_queue_offload_capa = 0;
info->rx_offload_capa = (DEV_RX_OFFLOAD_IPV4_CKSUM |
DEV_RX_OFFLOAD_UDP_CKSUM |
DEV_RX_OFFLOAD_TCP_CKSUM |
@@ -915,7 +916,7 @@ hns3vf_dev_infos_get(struct rte_eth_dev *eth_dev, struct
rte_eth_dev_info *info)
DEV_RX_OFFLOAD_JUMBO_FRAME |
DEV_RX_OFFLOAD_RSS_HASH |
DEV_RX_OFFLOAD_TCP_LRO);
- info->tx_queue_offload_capa = DEV_TX_OFFLOAD_MBUF_FAST_FREE;
+ info->tx_queue_offload_capa = 0;
info->tx_offload_capa = (DEV_TX_OFFLOAD_OUTER_IPV4_CKSUM |
DEV_TX_OFFLOAD_IPV4_CKSUM |
DEV_TX_OFFLOAD_TCP_CKSUM |
@@ -926,7 +927,7 @@ hns3vf_dev_infos_get(struct rte_eth_dev *eth_dev, struct
rte_eth_dev_info *info)
DEV_TX_OFFLOAD_VXLAN_TNL_TSO |
DEV_TX_OFFLOAD_GRE_TNL_TSO |
DEV_TX_OFFLOAD_GENEVE_TNL_TSO |
-info->tx_queue_offload_capa |
+DEV_TX_OFFLOAD_MBUF_FAST_FREE |
hns3_txvlan_cap_get(hw));
info->rx_desc_lim = (struct rte_eth_desc_lim) {
--
2.27.0
[dpdk-dev] [PATCH 08/11] net/hns3: change the log level to INFO
From: "Wei Hu (Xavier)"
This patch changes the log level from NOTICE to INFO.
Signed-off-by: Wei Hu (Xavier)
---
drivers/net/hns3/hns3_ethdev.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index fca035d4f..3827d3277 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -5744,5 +5744,5 @@ static struct rte_pci_driver rte_hns3_pmd = {
RTE_PMD_REGISTER_PCI(net_hns3, rte_hns3_pmd);
RTE_PMD_REGISTER_PCI_TABLE(net_hns3, pci_id_hns3_map);
RTE_PMD_REGISTER_KMOD_DEP(net_hns3, "* igb_uio | vfio-pci");
-RTE_LOG_REGISTER(hns3_logtype_init, pmd.net.hns3.init, NOTICE);
-RTE_LOG_REGISTER(hns3_logtype_driver, pmd.net.hns3.driver, NOTICE);
+RTE_LOG_REGISTER(hns3_logtype_init, pmd.net.hns3.init, INFO);
+RTE_LOG_REGISTER(hns3_logtype_driver, pmd.net.hns3.driver, INFO);
--
2.27.0
[dpdk-dev] [PATCH 09/11] net/hns3: fix default MAC addr from firmware
From: Huisong Li
Currently, default MAC address obtained from firmware in hns3 PF PMD driver
is directly used by .mac_addr_set ops implementation function when the
rte_eth_dev_start API function is executed. At this moment, if the default
MAC addr isn't an unicast address, it will fail to set default MAC addr to
hardware.
So this patch adds the validity check of default MAC addr in hns3 PF PMD
driver. We will use a random unicast address, if the default MAC address
obtained from firmware is not a vailid unicast address.
In addition, this patch also adjusts the location of processing default MAC
addr in hns3 VF PMD driver so as to increase relevance and readability of
the code.
Fixes: eab21776717 ("net/hns3: support setting VF MAC address by PF driver")
Fixes: d51867db65c ("net/hns3: add initialization")
Cc: sta...@dpdk.org
Signed-off-by: Huisong Li
Signed-off-by: Wei Hu (Xavier)
Signed-off-by: Chengchang Tang
---
drivers/net/hns3/hns3_ethdev.c| 11 +++
drivers/net/hns3/hns3_ethdev_vf.c | 29 ++---
2 files changed, 25 insertions(+), 15 deletions(-)
diff --git a/drivers/net/hns3/hns3_ethdev.c b/drivers/net/hns3/hns3_ethdev.c
index 3827d3277..14e4b9e35 100644
--- a/drivers/net/hns3/hns3_ethdev.c
+++ b/drivers/net/hns3/hns3_ethdev.c
@@ -5574,6 +5574,8 @@ static int
hns3_dev_init(struct rte_eth_dev *eth_dev)
{
struct hns3_adapter *hns = eth_dev->data->dev_private;
+ char mac_str[RTE_ETHER_ADDR_FMT_SIZE];
+ struct rte_ether_addr *eth_addr;
struct hns3_hw *hw = &hns->hw;
int ret;
@@ -5646,6 +5648,15 @@ hns3_dev_init(struct rte_eth_dev *eth_dev)
goto err_rte_zmalloc;
}
+ eth_addr = (struct rte_ether_addr *)hw->mac.mac_addr;
+ if (!rte_is_valid_assigned_ether_addr(eth_addr)) {
+ rte_eth_random_addr(hw->mac.mac_addr);
+ rte_ether_format_addr(mac_str, RTE_ETHER_ADDR_FMT_SIZE,
+ (struct rte_ether_addr *)hw->mac.mac_addr);
+ hns3_warn(hw, "default mac_addr from firmware is an invalid "
+ "unicast address, using random MAC address %s",
+ mac_str);
+ }
rte_ether_addr_copy((struct rte_ether_addr *)hw->mac.mac_addr,
ð_dev->data->mac_addrs[0]);
diff --git a/drivers/net/hns3/hns3_ethdev_vf.c
b/drivers/net/hns3/hns3_ethdev_vf.c
index 3b2ba69bb..7fd0e6a43 100644
--- a/drivers/net/hns3/hns3_ethdev_vf.c
+++ b/drivers/net/hns3/hns3_ethdev_vf.c
@@ -1748,21 +1748,6 @@ hns3vf_init_vf(struct rte_eth_dev *eth_dev)
goto err_get_config;
}
- /*
-* The hns3 PF ethdev driver in kernel support setting VF MAC address
-* on the host by "ip link set ..." command. To avoid some incorrect
-* scenes, for example, hns3 VF PMD driver fails to receive and send
-* packets after user configure the MAC address by using the
-* "ip link set ..." command, hns3 VF PMD driver keep the same MAC
-* address strategy as the hns3 kernel ethdev driver in the
-* initialization. If user configure a MAC address by the ip command
-* for VF device, then hns3 VF PMD driver will start with it, otherwise
-* start with a random MAC address in the initialization.
-*/
- ret = rte_is_zero_ether_addr((struct rte_ether_addr *)hw->mac.mac_addr);
- if (ret)
- rte_eth_random_addr(hw->mac.mac_addr);
-
ret = hns3vf_clear_vport_list(hw);
if (ret) {
PMD_INIT_LOG(ERR, "Failed to clear tbl list: %d", ret);
@@ -2644,8 +2629,22 @@ hns3vf_dev_init(struct rte_eth_dev *eth_dev)
goto err_rte_zmalloc;
}
+ /*
+* The hns3 PF ethdev driver in kernel support setting VF MAC address
+* on the host by "ip link set ..." command. To avoid some incorrect
+* scenes, for example, hns3 VF PMD driver fails to receive and send
+* packets after user configure the MAC address by using the
+* "ip link set ..." command, hns3 VF PMD driver keep the same MAC
+* address strategy as the hns3 kernel ethdev driver in the
+* initialization. If user configure a MAC address by the ip command
+* for VF device, then hns3 VF PMD driver will start with it, otherwise
+* start with a random MAC address in the initialization.
+*/
+ if (rte_is_zero_ether_addr((struct rte_ether_addr *)hw->mac.mac_addr))
+ rte_eth_random_addr(hw->mac.mac_addr);
rte_ether_addr_copy((struct rte_ether_addr *)hw->mac.mac_addr,
ð_dev->data->mac_addrs[0]);
+
hw->adapter_state = HNS3_NIC_INITIALIZED;
/*
* Pass the information to the rte_eth_dev_close() that it should also
--
2.27.0
[dpdk-dev] [PATCH] lib/telemetry: fix passing full params string to command
Telemetry only passed the first param to the command handler if multiple
were entered by the user, separated by commas. Telemetry is required to
pass the full params string to the command, by splitting by a comma
delimiter only once to remove the command part of the string. This will
enable future commands to take multiple param values.
Fixes: b1ad0e124536 ("rawdev: add telemetry callbacks")
Fixes: c190daedb9b1 ("ethdev: add telemetry callbacks")
Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
Cc: bruce.richard...@intel.com
Cc: sta...@dpdk.org
Signed-off-by: Ciara Power
---
lib/librte_ethdev/rte_ethdev.c | 10 --
lib/librte_rawdev/rte_rawdev.c | 5 -
lib/librte_telemetry/telemetry.c | 2 +-
3 files changed, 13 insertions(+), 4 deletions(-)
diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
index 7858ad5f11..6ba09362db 100644
--- a/lib/librte_ethdev/rte_ethdev.c
+++ b/lib/librte_ethdev/rte_ethdev.c
@@ -5284,11 +5284,14 @@ handle_port_xstats(const char *cmd __rte_unused,
struct rte_eth_xstat_name *xstat_names;
int port_id, num_xstats;
int i, ret;
+ char *end_param;
if (params == NULL || strlen(params) == 0 || !isdigit(*params))
return -1;
- port_id = atoi(params);
+ port_id = strtoul(params, &end_param, 0);
+ if (*end_param != '\0')
+ RTE_ETHDEV_LOG(NOTICE, "Extra params passed to telemetry
command for ethdev xstats, first param in use");
if (!rte_eth_dev_is_valid_port(port_id))
return -1;
@@ -5330,11 +5333,14 @@ handle_port_link_status(const char *cmd __rte_unused,
static const char *status_str = "status";
int ret, port_id;
struct rte_eth_link link;
+ char *end_param;
if (params == NULL || strlen(params) == 0 || !isdigit(*params))
return -1;
- port_id = atoi(params);
+ port_id = strtoul(params, &end_param, 0);
+ if (*end_param != '\0')
+ RTE_ETHDEV_LOG(NOTICE, "Extra params passed to telemetry
command for ethdev link status, first param in use");
if (!rte_eth_dev_is_valid_port(port_id))
return -1;
diff --git a/lib/librte_rawdev/rte_rawdev.c b/lib/librte_rawdev/rte_rawdev.c
index 8f84d0b228..51c4470b0c 100644
--- a/lib/librte_rawdev/rte_rawdev.c
+++ b/lib/librte_rawdev/rte_rawdev.c
@@ -565,11 +565,14 @@ handle_dev_xstats(const char *cmd __rte_unused,
struct rte_rawdev_xstats_name *xstat_names;
int dev_id, num_xstats, i, ret;
unsigned int *ids;
+ char *end_param;
if (params == NULL || strlen(params) == 0 || !isdigit(*params))
return -1;
- dev_id = atoi(params);
+ dev_id = strtoul(params, &end_param, 0);
+ if (*end_param != '\0')
+ RTE_RDEV_LOG(NOTICE, "Extra params passed to telemetry command
for rawdev xstats, first param in use");
if (!rte_rawdev_pmd_is_valid_dev(dev_id))
return -1;
diff --git a/lib/librte_telemetry/telemetry.c b/lib/librte_telemetry/telemetry.c
index 0252282735..c6c2948709 100644
--- a/lib/librte_telemetry/telemetry.c
+++ b/lib/librte_telemetry/telemetry.c
@@ -248,7 +248,7 @@ client_handler(void *sock_id)
while (bytes > 0) {
buffer[bytes] = 0;
const char *cmd = strtok(buffer, ",");
- const char *param = strtok(NULL, ",");
+ const char *param = strtok(NULL, "\0");
telemetry_cb fn = unknown_command;
int i;
--
2.17.1
Re: [dpdk-dev] [PATCH RFC 0/1] net/af_xdp: shared UMEM support
On 8/11/2020 10:50 AM, Ciara Loftus wrote: > This RFC integrates shared UMEM support into the AF_XDP PMD. It is based on > the > WIP kernel series [1] by Magnus Karlsson. > > Detailed information on the shared UMEM feature can be found in the final > patch > in the aforementioned series. > > Support for the kernel feature can eventually be detected in DPDK by querying > the LINUX_KERNEL_VERSION. As of now the feature is not yet merged upstream, so > for this RFC it is assumed the user is using a patched version of v5.8. Hi Ciara, I don't see this feature in kernel in 5.9.0-rc2, is there any update in the kernel side? > > Shared UMEM is only available for zero copy mode. > > In order to share UMEM information between PMDs, the ethdevs wishing to share > must be tracked somehow. The method chosen to do so is similar to methods used > in the vHost [2] and vDPA drivers, where pointers to the ethdevs are > maintained > in an internal list. Proposals for alternate solutions are welcome. > > Performance data to follow with the v1. > > [1] > https://patchwork.ozlabs.org/project/netdev/cover/1595307848-20719-1-git-send-email-magnus.karls...@intel.com/ > [2] > https://git.dpdk.org/dpdk/commit/?id=ee584e9710b9abd60ee9faef664e106dcea10085 > > Ciara Loftus (1): > net/af_xdp: shared UMEM support > > doc/guides/nics/af_xdp.rst | 5 +- > drivers/net/af_xdp/rte_eth_af_xdp.c | 315 ++-- > 2 files changed, 252 insertions(+), 68 deletions(-) >
[dpdk-dev] [PATCH v2 1/2] doc/linux_gsg: clarify instructions on running as non-root
The current instructions are slightly out of date when it comes to providing information about setting up the system for using DPDK as non-root, so update them. Cc: sta...@dpdk.org Signed-off-by: Anatoly Burakov Reviewed-by: Ferruh Yigit --- Notes: v2: - Moved VFIO description to be first doc/guides/linux_gsg/enable_func.rst | 54 1 file changed, 39 insertions(+), 15 deletions(-) diff --git a/doc/guides/linux_gsg/enable_func.rst b/doc/guides/linux_gsg/enable_func.rst index b2bda80bb7..a000def6cc 100644 --- a/doc/guides/linux_gsg/enable_func.rst +++ b/doc/guides/linux_gsg/enable_func.rst @@ -58,22 +58,42 @@ The application can then determine what action to take, if any, if the HPET is n if any, and on what is available on the system at runtime. Running DPDK Applications Without Root Privileges - +- -.. note:: +In order to run DPDK as non-root, the following Linux filesystem objects' +permissions should be adjusted to ensure that the Linux account being used to +run the DPDK application has access to them: -The instructions below will allow running DPDK as non-root with older -Linux kernel versions. However, since version 4.0, the kernel does not allow -unprivileged processes to read the physical address information from -the pagemaps file, making it impossible for those processes to use HW -devices which require physical addresses +* All directories which serve as hugepage mount points, for example, ``/dev/hugepages`` -Although applications using the DPDK use network ports and other hardware resources directly, -with a number of small permission adjustments it is possible to run these applications as a user other than "root". -To do so, the ownership, or permissions, on the following Linux file system objects should be adjusted to ensure that -the Linux user account being used to run the DPDK application has access to them: +* If the HPET is to be used, ``/dev/hpet`` -* All directories which serve as hugepage mount points, for example, ``/mnt/huge`` +When running as non-root user, there may be some additional resource limits +that are imposed by the system. Specifically, the following resource limits may +need to be adjusted in order to ensure normal DPDK operation: + +* RLIMIT_LOCKS (number of file locks that can be held by a process) + +* RLIMIT_NOFILE (number of open file descriptors that can be held open by a process) + +* RLIMIT_MEMLOCK (amount of pinned pages the process is allowed to have) + +The above limits can usually be adjusted by editing +``/etc/security/limits.conf`` file, and rebooting. + +Additionally, depending on which kernel driver is in use, the relevant +resources also should be accessible by the user running the DPDK application. + +For ``vfio-pci`` kernel driver, the following Linux file system objects' +permissions should be adjusted: + +* The VFIO device file, ``/dev/vfio/vfio`` + +* The directories under ``/dev/vfio`` that correspond to IOMMU group numbers of + devices intended to be used by DPDK, for example, ``/dev/vfio/50`` + +For ``igb_uio`` or ``uio_pci_generic`` kernel drivers, the following Linux file +system objects' permissions should be adjusted: * The userspace-io device files in ``/dev``, for example, ``/dev/uio0``, ``/dev/uio1``, and so on @@ -82,11 +102,15 @@ the Linux user account being used to run the DPDK application has access to them /sys/class/uio/uio0/device/config /sys/class/uio/uio0/device/resource* -* If the HPET is to be used, ``/dev/hpet`` - .. note:: -On some Linux installations, ``/dev/hugepages`` is also a hugepage mount point created by default. +The instructions above will allow running DPDK with ``igb_uio`` driver as +non-root with older Linux kernel versions. However, since version 4.0, the +kernel does not allow unprivileged processes to read the physical address +information from the pagemaps file, making it impossible for those +processes to be used by non-privileged users. In such cases, using the VFIO +driver is recommended. + Power Management and Power Saving Functionality --- -- 2.17.1
[dpdk-dev] [PATCH v2 2/2] doc/linux_gsg: update information on using hugepages
Current information regarding hugepage usage is a little out of date. Update it to include information on in-memory mode, as well as on default mountpoints provided by systemd. Cc: sta...@dpdk.org Signed-off-by: Anatoly Burakov --- Notes: v2: - Reworked the description - Put runtime reservation first, and boot time as an alternative - Clarified wording and fixed typos - Mentioned that some kernel versions not supporting reserving 1G pages doc/guides/linux_gsg/sys_reqs.rst | 71 --- 1 file changed, 45 insertions(+), 26 deletions(-) diff --git a/doc/guides/linux_gsg/sys_reqs.rst b/doc/guides/linux_gsg/sys_reqs.rst index a124656bcb..8782d05579 100644 --- a/doc/guides/linux_gsg/sys_reqs.rst +++ b/doc/guides/linux_gsg/sys_reqs.rst @@ -155,8 +155,35 @@ Without hugepages, high TLB miss rates would occur with the standard 4k page siz Reserving Hugepages for DPDK Use -The allocation of hugepages should be done at boot time or as soon as possible after system boot -to prevent memory from being fragmented in physical memory. +The reservation of hugepages can be performed at run time. This is done by +echoing the number of hugepages required to a ``nr_hugepages`` file in the +``/sys/kernel/`` directory corresponding to a specific page size (in +Kilobytes). For a single-node system, the command to use is as follows +(assuming that 1024 of 2MB pages are required):: + +echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages + +On a NUMA machine, the above command will usually divide the number of hugepages +equally across all NUMA nodes (assuming there is enough memory on all NUMA +nodes). However, pages can also be reserved explicitly on individual NUMA +nodes using a ``nr_hugepages`` file in the ``/sys/devices/`` directory:: + +echo 1024 > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages +echo 1024 > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages + +.. note:: + +Some kernel versions may not allow reserving 1 GB hugepages at run time, so +reserving them at boot time may be the only option. Please see below for +instructions. + +**Alternative:** + +In the general case, reserving hugepages at run time is perfectly fine, but in +use cases where having lots of physically contiguous memory is required, it is +preferable to reserve hugepages at boot time, as that will help in preventing +physical memory from becoming heavily fragmented. + To reserve hugepages at boot time, a parameter is passed to the Linux kernel on the kernel command line. For 2 MB pages, just pass the hugepages option to the kernel. For example, to reserve 1024 pages of 2 MB, use:: @@ -185,35 +212,27 @@ the number of hugepages reserved at boot time is generally divided equally betwe See the Documentation/admin-guide/kernel-parameters.txt file in your Linux source tree for further details of these and other kernel options. -**Alternative:** - -For 2 MB pages, there is also the option of allocating hugepages after the system has booted. -This is done by echoing the number of hugepages required to a nr_hugepages file in the ``/sys/devices/`` directory. -For a single-node system, the command to use is as follows (assuming that 1024 pages are required):: - -echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages - -On a NUMA machine, pages should be allocated explicitly on separate nodes:: - -echo 1024 > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages -echo 1024 > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages - -.. note:: - -For 1G pages, it is not possible to reserve the hugepage memory after the system has booted. - Using Hugepages with the DPDK ^ -Once the hugepage memory is reserved, to make the memory available for DPDK use, perform the following steps:: +If secondary process support is not required, DPDK is able to use hugepages +without any configuration by using "in-memory" mode. Please see +:ref:`linux_eal_parameters` for more details. + +If secondary process support is required, mount points for hugepages need to be +created. On modern Linux distributions, a default mount point for hugepages is provided +by the system and is located at ``/dev/hugepages``. This mount point will use the +default hugepage size set by the kernel parameters as described above. + +However, in order to use hugepage sizes other than default, it is necessary to +manually create mount points for hugepage sizes that are not provided by the +system (e.g. 1GB pages). + +To make the hugepages of size 1GB available for DPDK use, perform the following steps:: mkdir /mnt/huge -mount -t hugetlbfs nodev /mnt/huge +mount -t hugetlbfs pagesize=1GB /mnt/huge The mount point can be made permanent across reboots, by adding the following line to the ``/etc/fstab`` file::
[dpdk-dev] [PATCH v8 1/5] net/bnxt: add support for aarch32
From: Ruifeng Wang
Expand vector PMD support to aarch32.
Signed-off-by: Ruifeng Wang
---
drivers/net/bnxt/bnxt_rxq.h | 2 +-
drivers/net/bnxt/bnxt_rxr.h | 2 +-
drivers/net/bnxt/bnxt_txr.h | 2 +-
3 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/bnxt/bnxt_rxq.h b/drivers/net/bnxt/bnxt_rxq.h
index d5ce3b6d5..1c4027711 100644
--- a/drivers/net/bnxt/bnxt_rxq.h
+++ b/drivers/net/bnxt/bnxt_rxq.h
@@ -22,7 +22,7 @@ struct bnxt_rx_queue {
uint16_tnb_rx_hold; /* num held free RX desc */
uint16_trx_free_thresh; /* max free RX desc to hold */
uint16_tqueue_id; /* RX queue index */
-#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
+#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) || defined(RTE_ARCH_ARM)
uint16_trxrearm_nb; /* number of descs to reinit. */
uint16_trxrearm_start; /* next desc index to reinit. */
#endif
diff --git a/drivers/net/bnxt/bnxt_rxr.h b/drivers/net/bnxt/bnxt_rxr.h
index 2bf46cd91..e2fba1647 100644
--- a/drivers/net/bnxt/bnxt_rxr.h
+++ b/drivers/net/bnxt/bnxt_rxr.h
@@ -221,7 +221,7 @@ int bnxt_init_one_rx_ring(struct bnxt_rx_queue *rxq);
int bnxt_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id);
int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id);
-#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
+#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) || defined(RTE_ARCH_ARM)
uint16_t bnxt_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
uint16_t nb_pkts);
int bnxt_rxq_vec_setup(struct bnxt_rx_queue *rxq);
diff --git a/drivers/net/bnxt/bnxt_txr.h b/drivers/net/bnxt/bnxt_txr.h
index 7715c11b8..38e5ac9df 100644
--- a/drivers/net/bnxt/bnxt_txr.h
+++ b/drivers/net/bnxt/bnxt_txr.h
@@ -59,7 +59,7 @@ uint16_t bnxt_xmit_pkts(void *tx_queue, struct rte_mbuf
**tx_pkts,
uint16_t nb_pkts);
uint16_t bnxt_dummy_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
uint16_t nb_pkts);
-#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
+#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) || defined(RTE_ARCH_ARM)
uint16_t bnxt_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts,
uint16_t nb_pkts);
#endif
--
2.20.1
[dpdk-dev] [PATCH v8 2/5] build: add aarch32 meson build flags
Add aarch32 extra build flags and aarch32 machine flags to generic
machine args.
Also modify how arm flags are updated in meson build - for 32-bit build,
update only if cross-compiling.
Signed-off-by: Juraj Linkeš
---
config/arm/meson.build | 20 ++--
1 file changed, 18 insertions(+), 2 deletions(-)
diff --git a/config/arm/meson.build b/config/arm/meson.build
index 8728051d5..b29f27097 100644
--- a/config/arm/meson.build
+++ b/config/arm/meson.build
@@ -63,6 +63,11 @@ flags_armada = [
['RTE_MAX_LCORE', 16]]
flags_default_extra = []
+flags_aarch32_extra = [
+['RTE_ARCH_ARM_NEON_MEMCPY', false],
+['RTE_ARCH_STRICT_ALIGN', true],
+['RTE_EAL_NUMA_AWARE_HUGEPAGES', false],
+['RTE_MAX_LCORE', 256]]
flags_n1sdp_extra = [
['RTE_MACHINE', '"n1sdp"'],
['RTE_MAX_NUMA_NODES', 1],
@@ -90,6 +95,7 @@ flags_octeontx2_extra = [
machine_args_generic = [
['default', ['-march=armv8-a+crc']],
['native', ['-march=native']],
+ ['aarch32', ['-march=armv8-a', '-mfpu=neon'], flags_aarch32_extra],
['0xd03', ['-mcpu=cortex-a53']],
['0xd04', ['-mcpu=cortex-a35']],
['0xd07', ['-mcpu=cortex-a57']],
@@ -129,15 +135,23 @@ impl_0x69 = ['Intel', flags_generic, machine_args_generic]
impl_dpaa = ['NXP DPAA', flags_dpaa, machine_args_generic]
dpdk_conf.set('RTE_FORCE_INTRINSICS', 1)
-
+update_flags = false
if not dpdk_conf.get('RTE_ARCH_64')
dpdk_conf.set('RTE_CACHE_LINE_SIZE', 64)
dpdk_conf.set('RTE_ARCH_ARM', 1)
dpdk_conf.set('RTE_ARCH_ARMv7', 1)
# the minimum architecture supported, armv7-a, needs the following,
# mk/machine/armv7a/rte.vars.mk sets it too
- machine_args += '-mfpu=neon'
+ if meson.is_cross_build()
+ update_flags = true
+ impl_id = meson.get_cross_property('implementor_id', 'aarch32')
+ impl_pn = meson.get_cross_property('implementor_pn', 'default')
+ machine = get_variable('impl_' + impl_id)
+ else
+ machine_args += '-mfpu=neon'
+ endif
else
+ update_flags = true
dpdk_conf.set('RTE_CACHE_LINE_SIZE', 128)
dpdk_conf.set('RTE_ARCH_ARM64', 1)
@@ -172,7 +186,9 @@ else
impl_pn = meson.get_cross_property('implementor_pn', 'default')
machine = get_variable('impl_' + impl_id)
endif
+endif
+if update_flags == true
# Apply Common Defaults. These settings may be overwritten by machine
# settings later.
foreach flag: flags_common_default
--
2.20.1
[dpdk-dev] [PATCH v8 0/5] aarch64 -> aarch32 cross compilation support
Add support for aarch32 cross build in meson and add aarch64 -> aarch32 cross build to Travis. Aarch32 is an execution state that allows execution of 32-bit code on armv8 machines. This execution state contains a superset of previous armv7 32-bit instructions and features. Thus the aarch32 build is distinct from arvm7 build. Tested here: [v6] https://travis-ci.com/github/jlinkes/dpdk/builds/179252628 [v7] https://travis-ci.com/github/jlinkes/dpdk/builds/179723067 [v8] https://travis-ci.com/github/jlinkes/dpdk/builds/181307330 v4: Removed disabled drivers which actually build on arm32. Also tested the patchset with series 9609 which fixes underlying failures. v5: Changed the condition for running test-null.sh in ci. Re-uploaded after underlying fixes have been committed. v6: Changed the condition for running test-null.sh again. Reworked the patch to do aarch32 build instead of arvm7-a build. Simplified meson build flags. Changed commit msgs. Added 32b qualifier to .travis.yml. Added a patch with fixes for bnxt. Added a patch with cross compilation docs. v7: Rebased the patchset. v8: Removed Makefile additions from net/bnxt patch. Juraj Linkeš (3): build: add aarch32 meson build flags build: add aarch32 to meson cross-compilation ci: add aarch64 -> aarch32 cross compiling jobs Phil Yang (1): doc: add aarch32 build guidance Ruifeng Wang (1): net/bnxt: add support for aarch32 .ci/linux-build.sh| 7 - .travis.yml | 19 config/arm/arm32_armv8a_linux_gcc | 17 +++ config/arm/meson.build| 20 +++-- .../linux_gsg/cross_build_dpdk_for_arm64.rst | 30 ++- drivers/net/bnxt/bnxt_rxq.h | 2 +- drivers/net/bnxt/bnxt_rxr.h | 2 +- drivers/net/bnxt/bnxt_txr.h | 2 +- 8 files changed, 86 insertions(+), 13 deletions(-) create mode 100644 config/arm/arm32_armv8a_linux_gcc -- 2.20.1
[dpdk-dev] [PATCH v8 3/5] build: add aarch32 to meson cross-compilation
Create meson cross file arm32_armv8a_linux_gcc. Use arm-linux-gnueabihf- toolset which comes with standard packages on most used systems, such as Ubuntu and CentOS. Signed-off-by: Juraj Linkeš --- config/arm/arm32_armv8a_linux_gcc | 17 + 1 file changed, 17 insertions(+) create mode 100644 config/arm/arm32_armv8a_linux_gcc diff --git a/config/arm/arm32_armv8a_linux_gcc b/config/arm/arm32_armv8a_linux_gcc new file mode 100644 index 0..d3c4bce6c --- /dev/null +++ b/config/arm/arm32_armv8a_linux_gcc @@ -0,0 +1,17 @@ +[binaries] +c = 'arm-linux-gnueabihf-gcc' +cpp = 'arm-linux-gnueabihf-cpp' +ar = 'arm-linux-gnueabihf-gcc-ar' +strip = 'arm-linux-gnueabihf-strip' +pkgconfig = 'arm-linux-gnueabihf-pkg-config' +pcap-config = '' + +[host_machine] +system = 'linux' +cpu_family = 'aarch64' +cpu = 'armv8-a' +endian = 'little' + +[properties] +implementor_id = '0x41' +implementor_pn = 'aarch32' -- 2.20.1
[dpdk-dev] [PATCH v8 4/5] ci: add aarch64 -> aarch32 cross compiling jobs
Add two jobs (static and shared libs), both building on aarch64 and producing 32-bit arm binaries executable on armv8-a, but not armv7. Do not run tests in these jobs. Signed-off-by: Juraj Linkeš --- .ci/linux-build.sh | 7 ++- .travis.yml| 19 +++ 2 files changed, 25 insertions(+), 1 deletion(-) diff --git a/.ci/linux-build.sh b/.ci/linux-build.sh index d079801d7..c042b065c 100755 --- a/.ci/linux-build.sh +++ b/.ci/linux-build.sh @@ -33,6 +33,11 @@ if [ "$AARCH64" = "1" ]; then OPTS="$OPTS --cross-file config/arm/arm64_armv8_linux_gcc" fi +if [ "$AARCH32" = "1" ]; then +# convert the arch specifier +OPTS="$OPTS --cross-file config/arm/arm32_armv8a_linux_gcc" +fi + if [ "$BUILD_DOCS" = "1" ]; then OPTS="$OPTS -Denable_docs=true" fi @@ -53,7 +58,7 @@ OPTS="$OPTS --buildtype=debugoptimized" meson build --werror $OPTS ninja -C build -if [ "$AARCH64" != "1" ]; then +if file build/app/dpdk-testpmd | sed 's/-/_/g' | grep -q "$(uname -m)"; then devtools/test-null.sh fi diff --git a/.travis.yml b/.travis.yml index d6eeab371..850faaac1 100644 --- a/.travis.yml +++ b/.travis.yml @@ -18,6 +18,10 @@ _aarch64_packages: &aarch64_packages - *required_packages - [gcc-aarch64-linux-gnu, libc6-dev-arm64-cross, pkg-config-aarch64-linux-gnu] +_arm_32b_packages: &arm_32b_packages + - *required_packages + - [gcc-arm-linux-gnueabihf, libc6-dev-armhf-cross, pkg-config-arm-linux-gnueabihf] + _build_32b_packages: &build_32b_packages - *required_packages - [gcc-multilib] @@ -98,6 +102,21 @@ jobs: packages: - *required_packages - *doc_packages + # aarch64 cross-compiling aarch32 jobs + - env: DEF_LIB="shared" AARCH32=1 +arch: arm64 +compiler: gcc +addons: + apt: +packages: + - *arm_32b_packages + - env: DEF_LIB="static" AARCH32=1 +arch: arm64 +compiler: gcc +addons: + apt: +packages: + - *arm_32b_packages # aarch64 clang jobs - env: DEF_LIB="static" arch: arm64 -- 2.20.1
[dpdk-dev] [PATCH v8 5/5] doc: add aarch32 build guidance
From: Phil Yang Add cross-compiling guidance for 32-bit aarch32 DPDK on aarch64 host. Signed-off-by: Phil Yang --- .../linux_gsg/cross_build_dpdk_for_arm64.rst | 30 ++- 1 file changed, 23 insertions(+), 7 deletions(-) diff --git a/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst b/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst index c5875a6d5..07085e3ea 100644 --- a/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst +++ b/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst @@ -1,15 +1,16 @@ .. SPDX-License-Identifier: BSD-3-Clause Copyright(c) 2018 ARM Corporation. -Cross compile DPDK for ARM64 - -This chapter describes how to cross compile DPDK for ARM64 from x86 build hosts. +Cross compile DPDK for aarch64 and aarch32 +== +This chapter describes how to cross compile DPDK for aarch64 from x86 build hosts and compile +32-bit aarch32 DPDK from aarch64 build hosts. .. note:: - Whilst it is recommended to natively build DPDK on ARM64 (just - like with x86), it is also possible to cross-build DPDK for ARM64. An - ARM64 cross compile GNU toolchain is used for this. + Whilst it is recommended to natively build DPDK on aarch64 (just + like with x86), it is also possible to cross-build DPDK for aarch64. + An aarch64 cross compile GNU toolchain is used for this. Obtain the cross tool chain --- @@ -88,18 +89,33 @@ To install it in Ubuntu:: sudo apt-get install pkg-config-aarch64-linux-gnu +.. note:: + +Some aarch64 platforms support EL0 aarch32 mode, which means the 32-bit aarch32 applications +can run on aarch64. The armhf architecture toolchain ``gcc-arm-linux-gnueabihf`` is required +for aarch32 on aarch64. To install it in Ubuntu:: + + sudo dpkg --add-architecture armhf + sudo apt-get update + sudo apt-get install -y gcc gcc-arm-linux-gnueabihf libc6:armhf make binutils + To cross-compile DPDK on a desired target machine we can use the following command:: meson cross-build --cross-file ninja -C cross-build -For example if the target machine is arm64 we can use the following +For example if the target machine is aarch64 we can use the following command:: meson arm64-build --cross-file config/arm/arm64_armv8_linux_gcc ninja -C arm64-build +If the target machine is aarch32 we can use the following command:: + + meson arm32-build --cross-file config/arm/arm32_armv8a_linux_gcc + ninja -C arm32-build + Configure and Cross Compile DPDK using Make --- To configure a build, choose one of the target configurations, like arm64-dpaa-linux-gcc and arm64-thunderx-linux-gcc. -- 2.20.1
Re: [dpdk-dev] [PATCH v2 1/2] doc/linux_gsg: clarify instructions on running as non-root
On Tue, Aug 25, 2020 at 01:17:48PM +0100, Anatoly Burakov wrote: > The current instructions are slightly out of date when it comes to > providing information about setting up the system for using DPDK as > non-root, so update them. > > Cc: sta...@dpdk.org > > Signed-off-by: Anatoly Burakov > Reviewed-by: Ferruh Yigit Acked-by: Bruce Richardson
Re: [dpdk-dev] [PATCH v2 2/2] doc/linux_gsg: update information on using hugepages
On Tue, Aug 25, 2020 at 01:17:49PM +0100, Anatoly Burakov wrote: > Current information regarding hugepage usage is a little out of date. > Update it to include information on in-memory mode, as well as on > default mountpoints provided by systemd. > > Cc: sta...@dpdk.org > > Signed-off-by: Anatoly Burakov > --- > > Notes: > v2: > - Reworked the description > - Put runtime reservation first, and boot time as an alternative > - Clarified wording and fixed typos > - Mentioned that some kernel versions not supporting reserving 1G pages > > doc/guides/linux_gsg/sys_reqs.rst | 71 --- > 1 file changed, 45 insertions(+), 26 deletions(-) > > diff --git a/doc/guides/linux_gsg/sys_reqs.rst > b/doc/guides/linux_gsg/sys_reqs.rst > index a124656bcb..8782d05579 100644 > --- a/doc/guides/linux_gsg/sys_reqs.rst > +++ b/doc/guides/linux_gsg/sys_reqs.rst > @@ -155,8 +155,35 @@ Without hugepages, high TLB miss rates would occur with > the standard 4k page siz > Reserving Hugepages for DPDK Use > > > -The allocation of hugepages should be done at boot time or as soon as > possible after system boot > -to prevent memory from being fragmented in physical memory. > +The reservation of hugepages can be performed at run time. This is done by > +echoing the number of hugepages required to a ``nr_hugepages`` file in the > +``/sys/kernel/`` directory corresponding to a specific page size (in > +Kilobytes). For a single-node system, the command to use is as follows > +(assuming that 1024 of 2MB pages are required):: > + > +echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages > + > +On a NUMA machine, the above command will usually divide the number of > hugepages > +equally across all NUMA nodes (assuming there is enough memory on all NUMA > +nodes). However, pages can also be reserved explicitly on individual NUMA > +nodes using a ``nr_hugepages`` file in the ``/sys/devices/`` directory:: > + > +echo 1024 > > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages > +echo 1024 > > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages > + > +.. note:: > + > +Some kernel versions may not allow reserving 1 GB hugepages at run time, > so > +reserving them at boot time may be the only option. Please see below for > +instructions. > + > +**Alternative:** > + > +In the general case, reserving hugepages at run time is perfectly fine, but > in > +use cases where having lots of physically contiguous memory is required, it > is > +preferable to reserve hugepages at boot time, as that will help in preventing > +physical memory from becoming heavily fragmented. > + > To reserve hugepages at boot time, a parameter is passed to the Linux kernel > on the kernel command line. > > For 2 MB pages, just pass the hugepages option to the kernel. For example, > to reserve 1024 pages of 2 MB, use:: > @@ -185,35 +212,27 @@ the number of hugepages reserved at boot time is > generally divided equally betwe > > See the Documentation/admin-guide/kernel-parameters.txt file in your Linux > source tree for further details of these and other kernel options. > > -**Alternative:** > - > -For 2 MB pages, there is also the option of allocating hugepages after the > system has booted. > -This is done by echoing the number of hugepages required to a nr_hugepages > file in the ``/sys/devices/`` directory. > -For a single-node system, the command to use is as follows (assuming that > 1024 pages are required):: > - > -echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages > - > -On a NUMA machine, pages should be allocated explicitly on separate nodes:: > - > -echo 1024 > > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages > -echo 1024 > > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages > - > -.. note:: > - > -For 1G pages, it is not possible to reserve the hugepage memory after > the system has booted. > - > Using Hugepages with the DPDK > ^ > > -Once the hugepage memory is reserved, to make the memory available for DPDK > use, perform the following steps:: > +If secondary process support is not required, DPDK is able to use hugepages > +without any configuration by using "in-memory" mode. Please see > +:ref:`linux_eal_parameters` for more details. > + > +If secondary process support is required, mount points for hugepages need to > be > +created. On modern Linux distributions, a default mount point for hugepages > is provided > +by the system and is located at ``/dev/hugepages``. This mount point will > use the > +default hugepage size set by the kernel parameters as described above. > + > +However, in order to use hugepage sizes other than default, it is necessary > to > +manually create mount points for hugepage sizes that are not provided by the > +system (e.g. 1GB pages). Th
Re: [dpdk-dev] [PATCH] lib/telemetry: fix passing full params string to command
On Tue, Aug 25, 2020 at 01:01:33PM +0100, Ciara Power wrote:
> Telemetry only passed the first param to the command handler if multiple
> were entered by the user, separated by commas. Telemetry is required to
> pass the full params string to the command, by splitting by a comma
> delimiter only once to remove the command part of the string. This will
> enable future commands to take multiple param values.
>
> Fixes: b1ad0e124536 ("rawdev: add telemetry callbacks")
> Fixes: c190daedb9b1 ("ethdev: add telemetry callbacks")
> Fixes: 6dd571fd07c3 ("telemetry: introduce new functionality")
> Cc: bruce.richard...@intel.com
> Cc: sta...@dpdk.org
>
> Signed-off-by: Ciara Power
> ---
> lib/librte_ethdev/rte_ethdev.c | 10 --
> lib/librte_rawdev/rte_rawdev.c | 5 -
> lib/librte_telemetry/telemetry.c | 2 +-
> 3 files changed, 13 insertions(+), 4 deletions(-)
>
> diff --git a/lib/librte_ethdev/rte_ethdev.c b/lib/librte_ethdev/rte_ethdev.c
> index 7858ad5f11..6ba09362db 100644
> --- a/lib/librte_ethdev/rte_ethdev.c
> +++ b/lib/librte_ethdev/rte_ethdev.c
> @@ -5284,11 +5284,14 @@ handle_port_xstats(const char *cmd __rte_unused,
> struct rte_eth_xstat_name *xstat_names;
> int port_id, num_xstats;
> int i, ret;
> + char *end_param;
>
> if (params == NULL || strlen(params) == 0 || !isdigit(*params))
> return -1;
>
> - port_id = atoi(params);
> + port_id = strtoul(params, &end_param, 0);
> + if (*end_param != '\0')
> + RTE_ETHDEV_LOG(NOTICE, "Extra params passed to telemetry
> command for ethdev xstats, first param in use");
The text is a bit long, so maybe you can put the string on it's own line.
Also, I'd change "params" to the full word "parameters", and remove "first
param in use". Maybe the word "ignoring" would substitute for the last
part, e.g.
"Extra parameters passed to ethdev telemetry command, ignoring."
[dpdk-dev] [PATCH v3 2/2] doc/linux_gsg: update information on using hugepages
Current information regarding hugepage usage is a little out of date. Update it to include information on in-memory mode, as well as on default mountpoints provided by systemd. Cc: sta...@dpdk.org Signed-off-by: Anatoly Burakov Acked-by: Bruce Richardson --- Notes: v3: - Clarified wording around non-default hugepage sizes v2: - Reworked the description - Put runtime reservation first, and boot time as an alternative - Clarified wording and fixed typos - Mentioned that some kernel versions not supporting reserving 1G pages doc/guides/linux_gsg/sys_reqs.rst | 70 +++ 1 file changed, 44 insertions(+), 26 deletions(-) diff --git a/doc/guides/linux_gsg/sys_reqs.rst b/doc/guides/linux_gsg/sys_reqs.rst index a124656bcb..587f9e85e5 100644 --- a/doc/guides/linux_gsg/sys_reqs.rst +++ b/doc/guides/linux_gsg/sys_reqs.rst @@ -155,8 +155,35 @@ Without hugepages, high TLB miss rates would occur with the standard 4k page siz Reserving Hugepages for DPDK Use -The allocation of hugepages should be done at boot time or as soon as possible after system boot -to prevent memory from being fragmented in physical memory. +The reservation of hugepages can be performed at run time. This is done by +echoing the number of hugepages required to a ``nr_hugepages`` file in the +``/sys/kernel/`` directory corresponding to a specific page size (in +Kilobytes). For a single-node system, the command to use is as follows +(assuming that 1024 of 2MB pages are required):: + +echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages + +On a NUMA machine, the above command will usually divide the number of hugepages +equally across all NUMA nodes (assuming there is enough memory on all NUMA +nodes). However, pages can also be reserved explicitly on individual NUMA +nodes using a ``nr_hugepages`` file in the ``/sys/devices/`` directory:: + +echo 1024 > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages +echo 1024 > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages + +.. note:: + +Some kernel versions may not allow reserving 1 GB hugepages at run time, so +reserving them at boot time may be the only option. Please see below for +instructions. + +**Alternative:** + +In the general case, reserving hugepages at run time is perfectly fine, but in +use cases where having lots of physically contiguous memory is required, it is +preferable to reserve hugepages at boot time, as that will help in preventing +physical memory from becoming heavily fragmented. + To reserve hugepages at boot time, a parameter is passed to the Linux kernel on the kernel command line. For 2 MB pages, just pass the hugepages option to the kernel. For example, to reserve 1024 pages of 2 MB, use:: @@ -185,35 +212,26 @@ the number of hugepages reserved at boot time is generally divided equally betwe See the Documentation/admin-guide/kernel-parameters.txt file in your Linux source tree for further details of these and other kernel options. -**Alternative:** - -For 2 MB pages, there is also the option of allocating hugepages after the system has booted. -This is done by echoing the number of hugepages required to a nr_hugepages file in the ``/sys/devices/`` directory. -For a single-node system, the command to use is as follows (assuming that 1024 pages are required):: - -echo 1024 > /sys/kernel/mm/hugepages/hugepages-2048kB/nr_hugepages - -On a NUMA machine, pages should be allocated explicitly on separate nodes:: - -echo 1024 > /sys/devices/system/node/node0/hugepages/hugepages-2048kB/nr_hugepages -echo 1024 > /sys/devices/system/node/node1/hugepages/hugepages-2048kB/nr_hugepages - -.. note:: - -For 1G pages, it is not possible to reserve the hugepage memory after the system has booted. - Using Hugepages with the DPDK ^ -Once the hugepage memory is reserved, to make the memory available for DPDK use, perform the following steps:: +If secondary process support is not required, DPDK is able to use hugepages +without any configuration by using "in-memory" mode. Please see +:ref:`linux_eal_parameters` for more details. + +If secondary process support is required, mount points for hugepages need to be +created. On modern Linux distributions, a default mount point for hugepages is provided +by the system and is located at ``/dev/hugepages``. This mount point will use the +default hugepage size set by the kernel parameters as described above. + +However, in order to use hugepage sizes other than the default, it is necessary +to manually create mount points for those hugepage sizes (e.g. 1GB pages). + +To make the hugepages of size 1GB available for DPDK use, perform the following steps:: mkdir /mnt/huge -mount -t hugetlbfs nodev /mnt/huge +mount -t hugetlbfs pagesize=1GB /mnt/huge The mount point can be made permanent ac
[dpdk-dev] [PATCH v3 1/2] doc/linux_gsg: clarify instructions on running as non-root
The current instructions are slightly out of date when it comes to providing information about setting up the system for using DPDK as non-root, so update them. Cc: sta...@dpdk.org Signed-off-by: Anatoly Burakov Reviewed-by: Ferruh Yigit Acked-by: Bruce Richardson --- Notes: v2: - Moved VFIO description to be first doc/guides/linux_gsg/enable_func.rst | 54 1 file changed, 39 insertions(+), 15 deletions(-) diff --git a/doc/guides/linux_gsg/enable_func.rst b/doc/guides/linux_gsg/enable_func.rst index b2bda80bb7..a000def6cc 100644 --- a/doc/guides/linux_gsg/enable_func.rst +++ b/doc/guides/linux_gsg/enable_func.rst @@ -58,22 +58,42 @@ The application can then determine what action to take, if any, if the HPET is n if any, and on what is available on the system at runtime. Running DPDK Applications Without Root Privileges - +- -.. note:: +In order to run DPDK as non-root, the following Linux filesystem objects' +permissions should be adjusted to ensure that the Linux account being used to +run the DPDK application has access to them: -The instructions below will allow running DPDK as non-root with older -Linux kernel versions. However, since version 4.0, the kernel does not allow -unprivileged processes to read the physical address information from -the pagemaps file, making it impossible for those processes to use HW -devices which require physical addresses +* All directories which serve as hugepage mount points, for example, ``/dev/hugepages`` -Although applications using the DPDK use network ports and other hardware resources directly, -with a number of small permission adjustments it is possible to run these applications as a user other than "root". -To do so, the ownership, or permissions, on the following Linux file system objects should be adjusted to ensure that -the Linux user account being used to run the DPDK application has access to them: +* If the HPET is to be used, ``/dev/hpet`` -* All directories which serve as hugepage mount points, for example, ``/mnt/huge`` +When running as non-root user, there may be some additional resource limits +that are imposed by the system. Specifically, the following resource limits may +need to be adjusted in order to ensure normal DPDK operation: + +* RLIMIT_LOCKS (number of file locks that can be held by a process) + +* RLIMIT_NOFILE (number of open file descriptors that can be held open by a process) + +* RLIMIT_MEMLOCK (amount of pinned pages the process is allowed to have) + +The above limits can usually be adjusted by editing +``/etc/security/limits.conf`` file, and rebooting. + +Additionally, depending on which kernel driver is in use, the relevant +resources also should be accessible by the user running the DPDK application. + +For ``vfio-pci`` kernel driver, the following Linux file system objects' +permissions should be adjusted: + +* The VFIO device file, ``/dev/vfio/vfio`` + +* The directories under ``/dev/vfio`` that correspond to IOMMU group numbers of + devices intended to be used by DPDK, for example, ``/dev/vfio/50`` + +For ``igb_uio`` or ``uio_pci_generic`` kernel drivers, the following Linux file +system objects' permissions should be adjusted: * The userspace-io device files in ``/dev``, for example, ``/dev/uio0``, ``/dev/uio1``, and so on @@ -82,11 +102,15 @@ the Linux user account being used to run the DPDK application has access to them /sys/class/uio/uio0/device/config /sys/class/uio/uio0/device/resource* -* If the HPET is to be used, ``/dev/hpet`` - .. note:: -On some Linux installations, ``/dev/hugepages`` is also a hugepage mount point created by default. +The instructions above will allow running DPDK with ``igb_uio`` driver as +non-root with older Linux kernel versions. However, since version 4.0, the +kernel does not allow unprivileged processes to read the physical address +information from the pagemaps file, making it impossible for those +processes to be used by non-privileged users. In such cases, using the VFIO +driver is recommended. + Power Management and Power Saving Functionality --- -- 2.17.1
Re: [dpdk-dev] [PATCH v8 5/5] doc: add aarch32 build guidance
Juraj Linkeš writes: > From: Phil Yang > > Add cross-compiling guidance for 32-bit aarch32 DPDK on aarch64 host. > > Signed-off-by: Phil Yang > --- > .../linux_gsg/cross_build_dpdk_for_arm64.rst | 30 ++- > 1 file changed, 23 insertions(+), 7 deletions(-) > > diff --git a/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > b/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > index c5875a6d5..07085e3ea 100644 > --- a/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > +++ b/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > @@ -1,15 +1,16 @@ > .. SPDX-License-Identifier: BSD-3-Clause > Copyright(c) 2018 ARM Corporation. > > -Cross compile DPDK for ARM64 > - > -This chapter describes how to cross compile DPDK for ARM64 from x86 build > hosts. > +Cross compile DPDK for aarch64 and aarch32 > +== > +This chapter describes how to cross compile DPDK for aarch64 from x86 build > hosts and compile > +32-bit aarch32 DPDK from aarch64 build hosts. > > .. note:: > > - Whilst it is recommended to natively build DPDK on ARM64 (just > - like with x86), it is also possible to cross-build DPDK for ARM64. An > - ARM64 cross compile GNU toolchain is used for this. > + Whilst it is recommended to natively build DPDK on aarch64 (just > + like with x86), it is also possible to cross-build DPDK for aarch64. > + An aarch64 cross compile GNU toolchain is used for this. > > Obtain the cross tool chain > --- > @@ -88,18 +89,33 @@ To install it in Ubuntu:: > > sudo apt-get install pkg-config-aarch64-linux-gnu > > +.. note:: > + > +Some aarch64 platforms support EL0 aarch32 mode, which means the 32-bit > aarch32 applications > +can run on aarch64. The armhf architecture toolchain > ``gcc-arm-linux-gnueabihf`` is required > +for aarch32 on aarch64. To install it in Ubuntu:: > + > + sudo dpkg --add-architecture armhf > + sudo apt-get update > + sudo apt-get install -y gcc gcc-arm-linux-gnueabihf libc6:armhf make > binutils > + Is 'make' still needed? I don't run Ubuntu, so I'm not sure. > To cross-compile DPDK on a desired target machine we can use the following > command:: > > meson cross-build --cross-file > ninja -C cross-build > > -For example if the target machine is arm64 we can use the following > +For example if the target machine is aarch64 we can use the following > command:: > > meson arm64-build --cross-file config/arm/arm64_armv8_linux_gcc > ninja -C arm64-build > > +If the target machine is aarch32 we can use the following command:: > + > + meson arm32-build --cross-file config/arm/arm32_armv8a_linux_gcc > + ninja -C arm32-build > + > Configure and Cross Compile DPDK using Make > --- > To configure a build, choose one of the target configurations, like > arm64-dpaa-linux-gcc and arm64-thunderx-linux-gcc.
Re: [dpdk-dev] [PATCH RFC 1/1] net/af_xdp: shared UMEM support
On 8/11/2020 10:50 AM, Ciara Loftus wrote:
> A future kernel will introduce the ability to efficiently share a UMEM
> between AF_XDP sockets bound to different queue ids on the same or
> different devices. This patch integrates that functionality into the AF_XDP
> PMD.
>
> A PMD will attempt to share a UMEM with others if the shared_umem=1 vdev
> arg is set. UMEMs can only be shared across PMDs with the same mempool, up
> to a limited number of PMDs goverened by the size of the given mempool.
>
> The benefit of sharing UMEM across PMDs is a saving in memory due to not
> having to register the UMEM multiple times.
>
> Signed-off-by: Ciara Loftus
<...>
> @@ -1,5 +1,5 @@
> /* SPDX-License-Identifier: BSD-3-Clause
> - * Copyright(c) 2019 Intel Corporation.
> + * Copyright(c) 2020 Intel Corporation.
> */
> #include
> #include
> @@ -15,6 +15,7 @@
> #include
> #include
> #include
> +#include
> #include "af_xdp_deps.h"
> #include
>
> @@ -37,6 +38,11 @@
> #include
> #include
> #include
> +#include
> +
> +#if KERNEL_VERSION(5, 7, 0) < LINUX_VERSION_CODE
> +#define ETH_AF_XDP_SHARED_UMEM 1
> +#endif
I think better to separate these version checks from the actual code, what do
you think creating a compat.h under 'net/af_xdp' and move above logic there?
<...>
> @@ -888,9 +1048,15 @@ xsk_configure(struct pmd_internals *internals, struct
> pkt_rx_queue *rxq,
> cfg.bind_flags |= XDP_USE_NEED_WAKEUP;
> #endif
>
> - ret = xsk_socket__create(&rxq->xsk, internals->if_name,
> - rxq->xsk_queue_idx, rxq->umem->umem, &rxq->rx,
> - &txq->tx, &cfg);
> + if (!internals->shared_umem_configured) {
> + ret = xsk_socket__create(&rxq->xsk, internals->if_name,
> + rxq->xsk_queue_idx, rxq->umem->umem, &rxq->rx,
> + &txq->tx, &cfg);
> + } else {
> + ret = xsk_socket__create_shared(&rxq->xsk, internals->if_name,
> + rxq->xsk_queue_idx, rxq->umem->umem, &rxq->rx,
> + &txq->tx, &rxq->fq, &rxq->cq, &cfg);
> + }
Is the above dependency (ETH_AF_XDP_SHARED_UMEM) for the kernel 'af_xdp' code,
or for 'libbpf.so'?
The 'xsk_socket__create_shared()' API is not available in the latest
'libbpf.so', I wonder if the kernel version check is to align with the correct
'libbpf.so' version.
If not how the dependent version of the 'libbpf.so' managed for DPDK?
Re: [dpdk-dev] [PATCH v8 4/5] ci: add aarch64 -> aarch32 cross compiling jobs
Juraj Linkeš writes: > Add two jobs (static and shared libs), both building on aarch64 and > producing 32-bit arm binaries executable on armv8-a, but not armv7. > Do not run tests in these jobs. > > Signed-off-by: Juraj Linkeš > --- > .ci/linux-build.sh | 7 ++- > .travis.yml| 19 +++ > 2 files changed, 25 insertions(+), 1 deletion(-) > > diff --git a/.ci/linux-build.sh b/.ci/linux-build.sh > index d079801d7..c042b065c 100755 > --- a/.ci/linux-build.sh > +++ b/.ci/linux-build.sh > @@ -33,6 +33,11 @@ if [ "$AARCH64" = "1" ]; then > OPTS="$OPTS --cross-file config/arm/arm64_armv8_linux_gcc" > fi > > +if [ "$AARCH32" = "1" ]; then > +# convert the arch specifier > +OPTS="$OPTS --cross-file config/arm/arm32_armv8a_linux_gcc" > +fi > + > if [ "$BUILD_DOCS" = "1" ]; then > OPTS="$OPTS -Denable_docs=true" > fi > @@ -53,7 +58,7 @@ OPTS="$OPTS --buildtype=debugoptimized" > meson build --werror $OPTS > ninja -C build > > -if [ "$AARCH64" != "1" ]; then > +if file build/app/dpdk-testpmd | sed 's/-/_/g' | grep -q "$(uname -m)"; then I'm not sure about this. This will degrade the 32-bit x86 build, since it will no longer execute the test-null.sh script. Maybe we should re-use that flag for this series. BUILD_32BIT ?? > devtools/test-null.sh > fi > > diff --git a/.travis.yml b/.travis.yml > index d6eeab371..850faaac1 100644 > --- a/.travis.yml > +++ b/.travis.yml > @@ -18,6 +18,10 @@ _aarch64_packages: &aarch64_packages >- *required_packages >- [gcc-aarch64-linux-gnu, libc6-dev-arm64-cross, > pkg-config-aarch64-linux-gnu] > > +_arm_32b_packages: &arm_32b_packages > + - *required_packages > + - [gcc-arm-linux-gnueabihf, libc6-dev-armhf-cross, > pkg-config-arm-linux-gnueabihf] > + > _build_32b_packages: &build_32b_packages >- *required_packages >- [gcc-multilib] > @@ -98,6 +102,21 @@ jobs: > packages: >- *required_packages >- *doc_packages > + # aarch64 cross-compiling aarch32 jobs > + - env: DEF_LIB="shared" AARCH32=1 > +arch: arm64 > +compiler: gcc > +addons: > + apt: > +packages: > + - *arm_32b_packages > + - env: DEF_LIB="static" AARCH32=1 > +arch: arm64 > +compiler: gcc > +addons: > + apt: > +packages: > + - *arm_32b_packages ># aarch64 clang jobs >- env: DEF_LIB="static" > arch: arm64
Re: [dpdk-dev] [PATCH] ci: enable vm env for aarch64 builds
Juraj Linkeš writes: > Tests requiring hugepages do not work outside of VM environment because > of security limitations. Enable aarch64 builds which run tests to run in > a VM to avoid these limitations. > > Signed-off-by: Juraj Linkeš > --- Cool! Globally we have dist: set to 'bionic'. Does it make sense to upgrade all the distributions to 'focal' ? What is the rationale for not using 'bionic'? > .travis.yml | 10 -- > 1 file changed, 8 insertions(+), 2 deletions(-) > > diff --git a/.travis.yml b/.travis.yml > index d6eeab371..d73d3e4c5 100644 > --- a/.travis.yml > +++ b/.travis.yml > @@ -88,7 +88,10 @@ jobs: > arch: arm64 > compiler: gcc >- env: DEF_LIB="shared" RUN_TESTS=1 > -arch: arm64 > +dist: focal > +arch: arm64-graviton2 > +virt: vm > +group: edge > compiler: gcc >- env: DEF_LIB="shared" BUILD_DOCS=1 > arch: arm64 > @@ -103,5 +106,8 @@ jobs: > arch: arm64 > compiler: clang >- env: DEF_LIB="shared" RUN_TESTS=1 > -arch: arm64 > +dist: focal > +arch: arm64-graviton2 > +virt: vm > +group: edge > compiler: clang
Re: [dpdk-dev] [PATCH v2 02/18] raw/ioat: split header for readability
On 21/08/2020 17:29, Bruce Richardson wrote:
Rather than having a single long complicated header file for general use we
can split things so that there is one header with all the publically needed
information - data structs and function prototypes - while the rest of the
internal details are put separately. This makes it easier to read,
understand and use the APIs.
Signed-off-by: Bruce Richardson
---
There are a couple of checkpatch errors about spacing in this patch,
however, it appears that these are false positives.
---
drivers/raw/ioat/meson.build | 1 +
drivers/raw/ioat/rte_ioat_rawdev.h | 144 +-
drivers/raw/ioat/rte_ioat_rawdev_fns.h | 164 +
3 files changed, 171 insertions(+), 138 deletions(-)
create mode 100644 drivers/raw/ioat/rte_ioat_rawdev_fns.h
diff --git a/drivers/raw/ioat/rte_ioat_rawdev.h
b/drivers/raw/ioat/rte_ioat_rawdev.h
index 4bc6491d91..7ace5c085a 100644
--- a/drivers/raw/ioat/rte_ioat_rawdev.h
+++ b/drivers/raw/ioat/rte_ioat_rawdev.h
@@ -14,12 +14,7 @@
* @b EXPERIMENTAL: these structures and APIs may change without prior notice
*/
-#include
-#include
-#include
-#include
-#include
-#include "rte_ioat_spec.h"
+#include
/** Name of the device driver */
#define IOAT_PMD_RAWDEV_NAME rawdev_ioat
@@ -38,38 +33,6 @@ struct rte_ioat_rawdev_config {
bool hdls_disable;/**< if set, ignore user-supplied handle params */
};
-/**
- * @internal
- * Structure representing a device instance
- */
-struct rte_ioat_rawdev {
- struct rte_rawdev *rawdev;
- const struct rte_memzone *mz;
- const struct rte_memzone *desc_mz;
-
- volatile struct rte_ioat_registers *regs;
- phys_addr_t status_addr;
- phys_addr_t ring_addr;
-
- unsigned short ring_size;
- struct rte_ioat_generic_hw_desc *desc_ring;
- bool hdls_disable;
- __m128i *hdls; /* completion handles for returning to user */
-
-
- unsigned short next_read;
- unsigned short next_write;
-
- /* some statistics for tracking, if added/changed update xstats fns*/
- uint64_t enqueue_failed __rte_cache_aligned;
- uint64_t enqueued;
- uint64_t started;
- uint64_t completed;
-
- /* to report completions, the device will write status back here */
- volatile uint64_t status __rte_cache_aligned;
-};
-
/**
* Enqueue a copy operation onto the ioat device
*
@@ -104,38 +67,7 @@ struct rte_ioat_rawdev {
static inline int
rte_ioat_enqueue_copy(int dev_id, phys_addr_t src, phys_addr_t dst,
unsigned int length, uintptr_t src_hdl, uintptr_t dst_hdl,
- int fence)
-{
- struct rte_ioat_rawdev *ioat = rte_rawdevs[dev_id].dev_private;
This assignment needs to be type cast to "struct rte_ioat_rawdev *" for
C++ compilation compatibility.
There are a number of occurrences of this in this patch.
- unsigned short read = ioat->next_read;
- unsigned short write = ioat->next_write;
- unsigned short mask = ioat->ring_size - 1;
- unsigned short space = mask + read - write;
- struct rte_ioat_generic_hw_desc *desc;
-
- if (space == 0) {
- ioat->enqueue_failed++;
- return 0;
- }
-
- ioat->next_write = write + 1;
- write &= mask;
-
- desc = &ioat->desc_ring[write];
- desc->size = length;
- /* set descriptor write-back every 16th descriptor */
- desc->u.control_raw = (uint32_t)((!!fence << 4) | (!(write & 0xF)) <<
3);
- desc->src_addr = src;
- desc->dest_addr = dst;
- if (!ioat->hdls_disable)
- ioat->hdls[write] = _mm_set_epi64x((int64_t)dst_hdl,
- (int64_t)src_hdl);
-
- rte_prefetch0(&ioat->desc_ring[ioat->next_write & mask]);
-
- ioat->enqueued++;
- return 1;
-}
+ int fence);
+/**
+ * Returns details of copy operations that have been completed
+ */
+static inline int
+rte_ioat_completed_copies(int dev_id, uint8_t max_copies,
+ uintptr_t *src_hdls, uintptr_t *dst_hdls)
+{
+ struct rte_ioat_rawdev *ioat = rte_rawdevs[dev_id].dev_private;
+ unsigned short mask = (ioat->ring_size - 1);
+ unsigned short read = ioat->next_read;
+ unsigned short end_read, count;
+ int error;
+ int i = 0;
+
+ end_read = (rte_ioat_get_last_completed(ioat, &error) + 1) & mask;
+ count = (end_read - (read & mask)) & mask;
+
+ if (error) {
+ rte_errno = EIO;
+ return -1;
+ }
+
+ if (ioat->hdls_disable) {
+ read += count;
+ goto end;
+ }
+
+ if (count > max_copies)
+ count = max_copies;
+
+ for (; i < count - 1; i += 2, read += 2) {
+ __m128i hdls0 = _mm_load_si128(&ioat->hdls[read & mask]);
+ __m128i hdls1 = _mm_load_si128(&ioat->hdls
Re: [dpdk-dev] [PATCH v2 10/18] raw/ioat: add datapath data structures for idxd devices
On 21/08/2020 17:29, Bruce Richardson wrote:
Add in the relevant data structures for the data path for DSA devices. Also
include a device dump function to output the status of each device.
Signed-off-by: Bruce Richardson
---
drivers/raw/ioat/idxd_pci.c| 3 +-
drivers/raw/ioat/idxd_vdev.c | 1 +
drivers/raw/ioat/ioat_common.c | 34 +++
drivers/raw/ioat/ioat_private.h| 2 +
drivers/raw/ioat/ioat_rawdev_test.c| 3 +-
drivers/raw/ioat/rte_ioat_rawdev_fns.h | 80 ++
6 files changed, 121 insertions(+), 2 deletions(-)
diff --git a/drivers/raw/ioat/idxd_pci.c b/drivers/raw/ioat/idxd_pci.c
index 72f4ecebb7..ce238ae04c 100644
--- a/drivers/raw/ioat/idxd_pci.c
+++ b/drivers/raw/ioat/idxd_pci.c
@@ -52,7 +52,8 @@ idxd_is_wq_enabled(struct idxd_rawdev *idxd)
}
static const struct rte_rawdev_ops idxd_pci_ops = {
- .dev_selftest = idxd_rawdev_test
+ .dev_selftest = idxd_rawdev_test,
+ .dump = idxd_dev_dump,
};
/* each portal uses 4 x 4k pages */
diff --git a/drivers/raw/ioat/idxd_vdev.c b/drivers/raw/ioat/idxd_vdev.c
index 93c023a6e8..0f9aa48e84 100644
--- a/drivers/raw/ioat/idxd_vdev.c
+++ b/drivers/raw/ioat/idxd_vdev.c
@@ -32,6 +32,7 @@ struct idxd_vdev_args {
static const struct rte_rawdev_ops idxd_vdev_ops = {
.dev_selftest = idxd_rawdev_test,
+ .dump = idxd_dev_dump,
};
static void *
diff --git a/drivers/raw/ioat/ioat_common.c b/drivers/raw/ioat/ioat_common.c
index 37a14e514d..fb4f7055de 100644
--- a/drivers/raw/ioat/ioat_common.c
+++ b/drivers/raw/ioat/ioat_common.c
@@ -7,6 +7,36 @@
#include "ioat_private.h"
+int
+idxd_dev_dump(struct rte_rawdev *dev, FILE *f)
+{
+ struct idxd_rawdev *idxd = dev->dev_private;
+ struct rte_idxd_rawdev *rte_idxd = &idxd->public;
For C++ compatibility, we need to rename "public" since it is a reserved
word.
This will trigger renaming this wherever it is used in other patches, of
course :)
Thanks,
Kevin
Re: [dpdk-dev] [PATCH v2 08/18] raw/ioat: create rawdev instances on idxd PCI probe
On 21/08/2020 17:29, Bruce Richardson wrote:
When a matching device is found via PCI probe create a rawdev instance for
each queue on the hardware. Use empty self-test function for these devices
so that the overall rawdev_autotest does not report failures.
Signed-off-by: Bruce Richardson
---
drivers/raw/ioat/idxd_pci.c| 236 -
drivers/raw/ioat/ioat_common.c | 61 +++
drivers/raw/ioat/ioat_private.h| 31
drivers/raw/ioat/ioat_rawdev_test.c| 7 +
drivers/raw/ioat/ioat_spec.h | 64 +++
drivers/raw/ioat/meson.build | 1 +
drivers/raw/ioat/rte_ioat_rawdev_fns.h | 35 +++-
7 files changed, 432 insertions(+), 3 deletions(-)
create mode 100644 drivers/raw/ioat/ioat_common.c
diff --git a/drivers/raw/ioat/ioat_private.h b/drivers/raw/ioat/ioat_private.h
index d87d4d055e..32c824536d 100644
--- a/drivers/raw/ioat/ioat_private.h
+++ b/drivers/raw/ioat/ioat_private.h
@@ -14,6 +14,10 @@
* @b EXPERIMENTAL: these structures and APIs may change without prior notice
*/
+#include
+#include
+#include "rte_ioat_rawdev.h"
+
extern int ioat_pmd_logtype;
#define IOAT_PMD_LOG(level, fmt, args...) rte_log(RTE_LOG_ ## level, \
@@ -24,4 +28,31 @@ extern int ioat_pmd_logtype;
#define IOAT_PMD_ERR(fmt, args...)IOAT_PMD_LOG(ERR, fmt, ## args)
#define IOAT_PMD_WARN(fmt, args...) IOAT_PMD_LOG(WARNING, fmt, ## args)
+struct idxd_pci_common {
+ rte_spinlock_t lk;
+ volatile struct rte_idxd_bar0 *regs;
+ volatile struct rte_idxd_wqcfg *wq_regs;
+ volatile struct rte_idxd_grpcfg *grp_regs;
+ volatile void *portals;
+};
+
+struct idxd_rawdev {
+ struct rte_idxd_rawdev public; /* the public members, must be first */
For C++ compatibility, we cannot use "public" since it is a reserved
word in C++. Suggest renaming to "pub".
Thanks,
Kevin
Re: [dpdk-dev] [PATCH v2 13/18] raw/ioat: add data path for idxd devices
On 21/08/2020 17:29, Bruce Richardson wrote:
Add support for doing copies using DSA hardware. This is implemented by
just switching on the device type field at the start of the inline
functions. Since there is no hardware which will have both device types
present this branch will always be predictable after the first call,
meaning it has little to no perf penalty.
Signed-off-by: Bruce Richardson
---
drivers/raw/ioat/ioat_common.c | 1 +
drivers/raw/ioat/ioat_rawdev.c | 1 +
drivers/raw/ioat/rte_ioat_rawdev_fns.h | 164 +++--
3 files changed, 157 insertions(+), 9 deletions(-)
diff --git a/drivers/raw/ioat/rte_ioat_rawdev_fns.h
b/drivers/raw/ioat/rte_ioat_rawdev_fns.h
index 1939437d50..1950c8 100644
--- a/drivers/raw/ioat/rte_ioat_rawdev_fns.h
+++ b/drivers/raw/ioat/rte_ioat_rawdev_fns.h
@@ -194,8 +194,8 @@ struct rte_idxd_rawdev {
/**
* Enqueue a copy operation onto the ioat device
*/
-static inline int
-rte_ioat_enqueue_copy(int dev_id, phys_addr_t src, phys_addr_t dst,
+static __rte_always_inline int
+__ioat_enqueue_copy(int dev_id, phys_addr_t src, phys_addr_t dst,
unsigned int length, uintptr_t src_hdl, uintptr_t dst_hdl,
int fence)
{
@@ -233,8 +233,8 @@ rte_ioat_enqueue_copy(int dev_id, phys_addr_t src,
phys_addr_t dst,
/**
* Trigger hardware to begin performing enqueued copy operations
*/
-static inline void
-rte_ioat_do_copies(int dev_id)
+static __rte_always_inline void
+__ioat_perform_ops(int dev_id)
{
struct rte_ioat_rawdev *ioat = rte_rawdevs[dev_id].dev_private;
ioat->desc_ring[(ioat->next_write - 1) & (ioat->ring_size - 1)].u
@@ -248,8 +248,8 @@ rte_ioat_do_copies(int dev_id)
* @internal
* Returns the index of the last completed operation.
*/
-static inline int
-rte_ioat_get_last_completed(struct rte_ioat_rawdev *ioat, int *error)
+static __rte_always_inline int
+__ioat_get_last_completed(struct rte_ioat_rawdev *ioat, int *error)
{
uint64_t status = ioat->status;
@@ -263,8 +263,8 @@ rte_ioat_get_last_completed(struct rte_ioat_rawdev *ioat, int *error)
/**
* Returns details of copy operations that have been completed
*/
-static inline int
-rte_ioat_completed_copies(int dev_id, uint8_t max_copies,
+static __rte_always_inline int
+__ioat_completed_ops(int dev_id, uint8_t max_copies,
uintptr_t *src_hdls, uintptr_t *dst_hdls)
{
struct rte_ioat_rawdev *ioat = rte_rawdevs[dev_id].dev_private;
@@ -274,7 +274,7 @@ rte_ioat_completed_copies(int dev_id, uint8_t max_copies,
int error;
int i = 0;
- end_read = (rte_ioat_get_last_completed(ioat, &error) + 1) & mask;
+ end_read = (__ioat_get_last_completed(ioat, &error) + 1) & mask;
count = (end_read - (read & mask)) & mask;
if (error) {
@@ -311,4 +311,150 @@ rte_ioat_completed_copies(int dev_id, uint8_t max_copies,
return count;
}
+static __rte_always_inline int
+__idxd_enqueue_copy(int dev_id, rte_iova_t src, rte_iova_t dst,
+ unsigned int length, uintptr_t src_hdl, uintptr_t dst_hdl,
+ int fence __rte_unused)
+{
+ struct rte_idxd_rawdev *idxd = rte_rawdevs[dev_id].dev_private;
For C++ compatibility, "dev_private" needs to be type cast to "struct
rte_idxd_rawdev *" here.
+ struct rte_idxd_desc_batch *b = &idxd->batch_ring[idxd->next_batch];
+ uint32_t op_flags = (idxd_op_memmove << IDXD_CMD_OP_SHIFT) |
+ IDXD_FLAG_CACHE_CONTROL;
+
+ /* check for room in the handle ring */
+ if (((idxd->next_free_hdl + 1) & (idxd->hdl_ring_sz - 1)) ==
idxd->next_ret_hdl) {
+ rte_errno = ENOSPC;
+ return 0;
+ }
+ if (b->op_count >= BATCH_SIZE) {
+ rte_errno = ENOSPC;
+ return 0;
+ }
+ /* check that we can actually use the current batch */
+ if (b->submitted) {
+ rte_errno = ENOSPC;
+ return 0;
+ }
+
+ /* write the descriptor */
+ b->ops[b->op_count++] = (struct rte_idxd_hw_desc){
+ .op_flags = op_flags,
+ .src = src,
+ .dst = dst,
+ .size = length
+ };
+
+ /* store the completion details */
+ if (!idxd->hdls_disable)
+ idxd->hdl_ring[idxd->next_free_hdl] = (struct
rte_idxd_user_hdl) {
+ .src = src_hdl,
+ .dst = dst_hdl
+ };
+ if (++idxd->next_free_hdl == idxd->hdl_ring_sz)
+ idxd->next_free_hdl = 0;
+
+ return 1;
+}
+
+static __rte_always_inline void
+__idxd_movdir64b(volatile void *dst, const void *src)
+{
+ asm volatile (".byte 0x66, 0x0f, 0x38, 0xf8, 0x02"
+ :
+ : "a" (dst), "d" (src));
+}
+
+static __rte_always_inline void
+__idxd_perform_ops(int dev_id)
+{
+ struct rte_idxd_rawdev *idxd = rte_raw
Re: [dpdk-dev] [PATCH] app/testpmd: fix flow rules list after port stop
On 8/20/2020 9:40 AM, Gregory Etelson wrote:
> Hello,
>
> Is this patch scheduled for merge with dpdk.org ?
> Please update me.
>
> Regards,
> Gregory
>
>> -Original Message-
>> From: Gregory Etelson
>> Sent: Monday, August 10, 2020 19:15
>> To: dev@dpdk.org
>> Cc: Gregory Etelson ; Matan Azrad
>> ; Raslan Darawsheh ;
>> sta...@dpdk.org; Ori Kam ; Wenzhuo Lu
>> ; Beilei Xing ; Bernard
>> Iremonger
>> Subject: [PATCH] app/testpmd: fix flow rules list after port stop
>>
>> According to current RTE API, port flow rules must not be kept after port
>> stop.
Hi Gregory, Ori,
Can you please point where this is documented?
>>
>> Testpmd did not flush port flow rules after `port stop' command was called.
>> As the result, after the port was restarted, it showed bogus flow rules.
There are two issues,
1) According what I see in the rte_flow documentation, not sure if the "port
stop" should clear the rules:
"
PMDs, not applications, are responsible for maintaining flow rules configuration
when stopping and restarting a port or performing other actions which may affect
them. They can only be destroyed explicitly by applications.
"
As I tested with i40e, it keeps the rules after stop/start, cc'ing @Jeff,
@Beilei & @Qi if this is done intentionally.
2) From the perspective of the testers, users of the testpmd. If they are
testing a complex set of filter rules, stopping and starting the port flushing
all rules may be troublesome.
Since there is explicit command to remove a rte_flow rule or to remove them all,
user may prefer to call it when required to delete the rules, instead of this is
done implicitly in port stop.
Btw, this is based on PMD should handle the rules on stop/start, we need to
agree on it first, but even that is not the case, we are in the application
domain now and we can apply the rules back again in the 'start' if it serves
better to the user.
>>
>> Fixes: ce8d561418d4 ("app/testpmd: add port configuration settings")
>>
>> Cc: sta...@dpdk.org
>>
>> Signed-off-by: Gregory Etelson
>> Acked-by: Ori Kam
>> ---
>> app/test-pmd/testpmd.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/app/test-pmd/testpmd.c b/app/test-pmd/testpmd.c index
>> 7842c3b781..4ba5c41c6e 100644
>> --- a/app/test-pmd/testpmd.c
>> +++ b/app/test-pmd/testpmd.c
>> @@ -2627,6 +2627,9 @@ stop_port(portid_t pid)
>> RTE_PORT_HANDLING) == 0)
>> continue;
>>
>> +if (port->flow_list)
>> +port_flow_flush(pi);
>> +
>> rte_eth_dev_stop(pi);
>>
>> if (rte_atomic16_cmpset(&(port->port_status),
>> --
>> 2.25.1
>
Re: [dpdk-dev] [EXT] Re: [PATCH v4 1/4] ethdev: add tm support for shaper config in pkt mode
On 5/1/2020 2:16 PM, Nithin Dabilpuram wrote: > On Fri, May 01, 2020 at 11:27:02AM +0100, Ferruh Yigit wrote: >> External Email >> >> -- >> On 4/29/2020 10:03 AM, Bruce Richardson wrote: >>> On Wed, Apr 29, 2020 at 09:45:44AM +0100, Dumitrescu, Cristian wrote: > -Original Message- > From: Thomas Monjalon > Sent: Tuesday, April 28, 2020 4:54 PM > To: Jerin Jacob ; Dumitrescu, Cristian > > Cc: Richardson, Bruce ; Yigit, Ferruh > ; Luca Boccassi ; Nithin > Dabilpuram ; Singh, Jasvinder > ; Andrew Rybchenko > ; dev@dpdk.org; jer...@marvell.com; > kka...@marvell.com; Nithin Dabilpuram ; > Kinsella, Ray ; Neil Horman > ; Kevin Traynor ; David > Marchand > Subject: Re: [dpdk-dev] [PATCH v4 1/4] ethdev: add tm support for shaper > config in pkt mode > > 28/04/2020 17:04, Luca Boccassi: >> On Tue, 2020-04-28 at 15:45 +0100, Bruce Richardson wrote: >>> On Tue, Apr 28, 2020 at 03:06:20PM +0100, Ferruh Yigit wrote: On 4/27/2020 5:59 PM, Jerin Jacob wrote: > On Mon, Apr 27, 2020 at 10:19 PM Ferruh Yigit > wrote: >> On 4/27/2020 5:29 PM, Jerin Jacob wrote: >>> On Mon, Apr 27, 2020 at 9:42 PM Ferruh Yigit > wrote: On 4/27/2020 10:19 AM, Dumitrescu, Cristian wrote: > From: Yigit, Ferruh >> On 4/24/2020 11:28 AM, Dumitrescu, Cristian wrote: >>> From: Nithin Dabilpuram This patch also updates tm port/level/node capability > structures with exiting features of scheduler wfq packet mode, > scheduler wfq byte mode and private/shared shaper byte mode. SoftNIC PMD is also updated with new capabilities. > [...] >> Hi Nithin, >> >> It looks like patch is causing ABI break, I am getting following > warning [1], >> can you please check? >> >> [1] >> https://urldefense.proofpoint.com/v2/url?u=https-3A__pastebin.com_XYNFg14u&d=DwIDaQ&c=nKjWec2b6R0mOyPaz7xtfQ&r=FZ_tPCbgFOh18zwRPO9H0yDx8VW38vuapifdDfc8SFQ&m=ej5sP3-cEhEoCTZOia-QivXqgljtzBcMLtZGs-5c-Uc&s=B8z_5mQ2xO3C1izjmRe2zBApMrCUcW6KcAN-adglhJQ&e= >> > > Hi Ferruh, > > The RTE_TM API is marked as experimental, > but it looks that this was not correctly marked > when __rte_experimental ABI checker was introduced. > > It is marked as experimental at the top of the rte_tm.h, > similarly to other APIs introduced around same time, > but it was not correctly picked up by the ABI check procedure > when later introduced, so __rte_experimental was not added > to every function. > :( Is it time to mature them? As you said they are not marked as experimental both in header > file (function declarations) and .map file. The problem is, they are not marked as experimental in > DPDK_20.0 ABI (v19.11), so marking them as experimental now will break the ABI. Not > sure what to do, cc'ed a few ABI related names for comment. For me, we need to proceed as the experimental tag removed > and APIs become mature starting from v19.11, since this is what happened in > practice, and remove a few existing being experimental references in the doxygen > comments. >>> >>> I think, accidentally we can not make a library as NON- > experimental. >>> TM never went through experimental to mature transition(see git > log >>> lib/librte_ethdev/rte_tm.h) >>> It was a bug to not mark as experimental in each function in the > ABI process. >>> Some of the features like packet marking are not even > implemented by any HW. >>> I think, we can make API stable only all the features are > implemented >>> by one or two HW. > > Yes this is what was decided one or two years ago I think. > But rte_tm API was introduced 3 years ago and is implemented by 6 PMDs. > > > >> Fair enough, specially if the API is not ready yet. >> >> But they were part of stable ABI, and marking them as experimental > now will >> break the old applications using these APIs. > > it is still marked as EXPERIMENTAL everywhere and API is not ready > yet. > > rte_tm is implemented in 6 PMDs. > > Existing experimental marks are text only for human parsing. The compiler attribute and build time checks are missing, and the > symbol in the
Re: [dpdk-dev] [PATCH] net: rte_ether minor optimizations
On 7/27/2020 8:34 AM, Olivier Matz wrote: > On Fri, Jun 26, 2020 at 06:51:35PM +0200, Morten Brørup wrote: >> * rte_is_broadcast_ether_addr(): >> Use binary logic instead of comparisons and boolean logic, thus reducing >> the number of branches. >> It now resembles rte_is_zero_ether_addr(). >> >> * rte_ether_addr_copy(): >> The source code modifications were discussed on the mailing list: >> http://mails.dpdk.org/archives/dev/2020-June/171584.html >> Remove obsolete ICC-specific code and related comment. >> Restrict pointer aliasing (suggested by Jerin Jacob). >> Remove superfluous "Fast" from function description headline; all DPDK >> data plane functions are supposed to be fast. >> >> Signed-off-by: Morten Brørup > > Acked-by: Olivier Matz > Applied to dpdk-next-net/main, thanks.
[dpdk-dev] [PATCH v2] net/dpaa: fix public API port ID type
Updating the type for 'port' variable from 'uint8_t' to 'uint16_t'.
Fixes: 8c3495f5d2dd ("net/dpaa: support loopback API")
Cc: sta...@dpdk.org
Signed-off-by: Ferruh Yigit
---
v2:
* Remove deprecation notice
---
doc/guides/rel_notes/deprecation.rst | 3 ---
drivers/net/dpaa/dpaa_ethdev.c | 2 +-
drivers/net/dpaa/rte_pmd_dpaa.h | 2 +-
3 files changed, 2 insertions(+), 5 deletions(-)
diff --git a/doc/guides/rel_notes/deprecation.rst
b/doc/guides/rel_notes/deprecation.rst
index 345c38d5b6..249d4c50a1 100644
--- a/doc/guides/rel_notes/deprecation.rst
+++ b/doc/guides/rel_notes/deprecation.rst
@@ -254,9 +254,6 @@ Deprecation Notices
experimental in v20.11 DPDK release. For more details, please see `the thread
`_.
-* pmd_dpaa: The API ``rte_pmd_dpaa_set_tx_loopback`` will have extended
- ``port_id`` definition from ``uint8_t`` to ``uint16_t``.
-
* vhost: Vhost-user dequeue zero-copy support will be removed in 20.11.
The only known user is OVS where the feature is still experimental,
and has not received any update for 2.5 years.
diff --git a/drivers/net/dpaa/dpaa_ethdev.c b/drivers/net/dpaa/dpaa_ethdev.c
index c15e2b5462..e84f904071 100644
--- a/drivers/net/dpaa/dpaa_ethdev.c
+++ b/drivers/net/dpaa/dpaa_ethdev.c
@@ -1354,7 +1354,7 @@ is_dpaa_supported(struct rte_eth_dev *dev)
}
int
-rte_pmd_dpaa_set_tx_loopback(uint8_t port, uint8_t on)
+rte_pmd_dpaa_set_tx_loopback(uint16_t port, uint8_t on)
{
struct rte_eth_dev *dev;
diff --git a/drivers/net/dpaa/rte_pmd_dpaa.h b/drivers/net/dpaa/rte_pmd_dpaa.h
index 8d244bb491..ec45633ba2 100644
--- a/drivers/net/dpaa/rte_pmd_dpaa.h
+++ b/drivers/net/dpaa/rte_pmd_dpaa.h
@@ -29,6 +29,6 @@
* - (-EINVAL) if bad parameter.
*/
int
-rte_pmd_dpaa_set_tx_loopback(uint8_t port, uint8_t on);
+rte_pmd_dpaa_set_tx_loopback(uint16_t port, uint8_t on);
#endif /* _PMD_DPAA_H_ */
--
2.25.4
Re: [dpdk-dev] question regarding rx checksum offload flags
On Mon, Aug 24, 2020 at 4:11 PM Lance Richardson wrote: > The IP header checksum status for the outer IP header should determine > whether this flag is set in ol_flags: > PKT_RX_EIP_CKSUM_BAD A simpler question might be simply "how is PKT_RX_EIP_CKSUM_BAD" intended to be used? It seems to be intended to be set if the outer IP checksum is not correct.. do I have that right? Thanks, Lance
Re: [dpdk-dev] [PATCH] net/tap: avoid using SIGIO
On 8/17/2020 4:26 PM, Ferruh Yigit wrote: > On 7/28/2020 11:48 AM, Morten Brørup wrote: >>> From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Stephen Hemminger >>> Sent: Monday, July 27, 2020 9:44 PM >>> >>> On Mon, 27 Jul 2020 15:19:32 +0200 >>> Morten Brørup wrote: >>> > From: dev [mailto:dev-boun...@dpdk.org] On Behalf Of Stephen >>> Hemminger > Sent: Wednesday, July 15, 2020 1:58 AM > > SIGIO maybe used by application, instead choose another rt-signal. > Linux allows any signal to be used for signal based IO. > Search for an unused signal in the available rt-signal range. Just an observation. Feel free to ignore at your convenience: The problem is the same as for SIGIO if the application sets up its >>> own signal handler after this, and uses some hardcoded rt-signal that >>> happens to be the one found to be free. Unless the application doesn't use a hardcoded rt-signal, but also >>> searches for an unused one. So perhaps the "search for unused rt-signal" should be exposed as a >>> generic support function for the application (and this driver) to use. >>> >>> There is no safe way to use a signal deep inside DPDK in a driver. >>> >>> This is not the kind of thing that should be exposed to the >>> application. >>> >>> The algorithm for finding an RT signal conforms to the recommended >>> policy on the signal(7) >>> manual page. >>> >>>programs should never refer to real-time signals using hard- >>>coded numbers, but instead should always refer to real-time >>> signals >>>using the notation SIGRTMIN+n, and include suitable (run-time) >>> checks >>>that SIGRTMIN+n does not exceed SIGRTMAX. >>> >>> The application should be following the proscribed policy on the man >>> page. >>> If it doesn't it is broken. >> >> Great. That fully addresses my concern. >> >> Acked-by: Morten Brørup >> > > Reviewed-by: Ferruh Yigit > Applied to dpdk-next-net/main, thanks.
Re: [dpdk-dev] [RFC v2] lib/hash: integrate RCU QSBR
CI has reported some compilation issues for this patch. I will resolve these
issues once the RFC patch is approved.
Thank you!
> On Aug 18, 2020, at 11:05 PM, Dharmik Thakkar wrote:
>
> Integrate RCU QSBR to make it easier for the applications to use lock
> free algorithm.
>
> Resource reclamation implementation was split from the original
> series, and has already been part of RCU library. Rework the series
> to base hash integration on RCU reclamation APIs.
>
> Refer 'Resource reclamation framework for DPDK' available at [1]
> to understand various aspects of integrating RCU library
> into other libraries.
>
> [1] https://doc.dpdk.org/guides/prog_guide/rcu_lib.html
>
> Introduce a new API rte_hash_rcu_qsbr_add for application to
> register a RCU variable that hash library will use.
>
> Suggested-by: Honnappa Nagarahalli
> Signed-off-by: Dharmik Thakkar
> Reviewed-by: Ruifeng Wang
> ---
> v2:
> - Remove defer queue related functions and use resource reclamation
> APIs from the RCU QSBR library instead
>
> - Remove patch (net/ixgbe: avoid multpile definitions of 'bool')
> from the series as it is already accepted
>
> ---
> lib/Makefile | 2 +-
> lib/librte_hash/Makefile | 2 +-
> lib/librte_hash/meson.build | 1 +
> lib/librte_hash/rte_cuckoo_hash.c| 291 +--
> lib/librte_hash/rte_cuckoo_hash.h| 8 +
> lib/librte_hash/rte_hash.h | 75 ++-
> lib/librte_hash/rte_hash_version.map | 2 +-
> 7 files changed, 308 insertions(+), 73 deletions(-)
>
> diff --git a/lib/Makefile b/lib/Makefile
> index 8f5b68a2d469..dffe31c829f0 100644
> --- a/lib/Makefile
> +++ b/lib/Makefile
> @@ -51,7 +51,7 @@ DIRS-$(CONFIG_RTE_LIBRTE_VHOST) += librte_vhost
> DEPDIRS-librte_vhost := librte_eal librte_mempool librte_mbuf librte_ethdev \
> librte_net librte_hash librte_cryptodev
> DIRS-$(CONFIG_RTE_LIBRTE_HASH) += librte_hash
> -DEPDIRS-librte_hash := librte_eal librte_ring
> +DEPDIRS-librte_hash := librte_eal librte_ring librte_rcu
> DIRS-$(CONFIG_RTE_LIBRTE_EFD) += librte_efd
> DEPDIRS-librte_efd := librte_eal librte_ring librte_hash
> DIRS-$(CONFIG_RTE_LIBRTE_RIB) += librte_rib
> diff --git a/lib/librte_hash/Makefile b/lib/librte_hash/Makefile
> index ec9f86499262..10e697f48652 100644
> --- a/lib/librte_hash/Makefile
> +++ b/lib/librte_hash/Makefile
> @@ -8,7 +8,7 @@ LIB = librte_hash.a
>
> CFLAGS += -O3
> CFLAGS += $(WERROR_FLAGS) -I$(SRCDIR)
> -LDLIBS += -lrte_eal -lrte_ring
> +LDLIBS += -lrte_eal -lrte_ring -lrte_rcu
>
> EXPORT_MAP := rte_hash_version.map
>
> diff --git a/lib/librte_hash/meson.build b/lib/librte_hash/meson.build
> index 6ab46ae9d768..0977a63fd279 100644
> --- a/lib/librte_hash/meson.build
> +++ b/lib/librte_hash/meson.build
> @@ -10,3 +10,4 @@ headers = files('rte_crc_arm64.h',
>
> sources = files('rte_cuckoo_hash.c', 'rte_fbk_hash.c')
> deps += ['ring']
> +deps += ['rcu']
> diff --git a/lib/librte_hash/rte_cuckoo_hash.c
> b/lib/librte_hash/rte_cuckoo_hash.c
> index 0a6d474713a2..01c2cbe0e38b 100644
> --- a/lib/librte_hash/rte_cuckoo_hash.c
> +++ b/lib/librte_hash/rte_cuckoo_hash.c
> @@ -52,6 +52,11 @@ static struct rte_tailq_elem rte_hash_tailq = {
> };
> EAL_REGISTER_TAILQ(rte_hash_tailq)
>
> +struct __rte_hash_rcu_dq_entry {
> + uint32_t key_idx;
> + uint32_t ext_bkt_idx; /**< Extended bkt index */
> +};
> +
> struct rte_hash *
> rte_hash_find_existing(const char *name)
> {
> @@ -210,7 +215,10 @@ rte_hash_create(const struct rte_hash_parameters *params)
>
> if (params->extra_flag & RTE_HASH_EXTRA_FLAGS_RW_CONCURRENCY_LF) {
> readwrite_concur_lf_support = 1;
> - /* Enable not freeing internal memory/index on delete */
> + /* Enable not freeing internal memory/index on delete.
> + * If internal RCU is enabled, freeing of internal memory/index
> + * is done on delete
> + */
> no_free_on_del = 1;
> }
>
> @@ -505,6 +513,10 @@ rte_hash_free(struct rte_hash *h)
>
> rte_mcfg_tailq_write_unlock();
>
> + /* RCU clean up. */
> + if (h->dq)
> + rte_rcu_qsbr_dq_delete(h->dq);
> +
> if (h->use_local_cache)
> rte_free(h->local_free_slots);
> if (h->writer_takes_lock)
> @@ -612,6 +624,16 @@ rte_hash_reset(struct rte_hash *h)
> return;
>
> __hash_rw_writer_lock(h);
> +
> + /* RCU clean up. */
> + if (h->hash_rcu_cfg->v) {
> + rte_rcu_qsbr_dq_delete(h->dq);
> + h->dq = NULL;
> + if (rte_hash_rcu_qsbr_add(h, h->hash_rcu_cfg) < 0) {
> + RTE_LOG(ERR, HASH, "RCU init failed\n");
> + return;
> + }
> + }
> memset(h->buckets, 0, h->num_buckets * sizeof(struct rte_hash_bucket));
> memset(h->key_store, 0, h->key_entry_size * (h->entries + 1));
> *h->tbl_chng_cnt = 0;
> @@ -952,6 +974,37 @@ rte_h
[dpdk-dev] [PATCH 2/2] build: find max lcore programmatically
For Arm, RTE_MAX_LCORE is hard-coded into the config. It leads to incorrect
RTE_MAX_LCORE when machines have same Implemener and part number
but different number of CPUs.
For x86, RTE_MAX_LCORE is always set to 128 (using the value
set in meson_options.txt)
Use python script to find max lcore when using native build to
correctly set RTE_MAX_LCORE.
Signed-off-by: Dharmik Thakkar
Reviewed-by: Ruifeng Wang
---
config/get_max_lcores.py | 13 +
config/meson.build | 13 -
2 files changed, 25 insertions(+), 1 deletion(-)
create mode 100755 config/get_max_lcores.py
diff --git a/config/get_max_lcores.py b/config/get_max_lcores.py
new file mode 100755
index ..ebf1c7efdadd
--- /dev/null
+++ b/config/get_max_lcores.py
@@ -0,0 +1,13 @@
+#!/usr/bin/python3
+# SPDX-License-Identifier: BSD-3-Clause
+# Copyright(c) 2020 Arm Limited
+
+import os
+
+max_lcores = []
+
+nCPU = os.cpu_count()
+
+max_lcores.append(str(nCPU & 0xFFF)) # Number of CPUs
+
+print(' '.join(max_lcores))
diff --git a/config/meson.build b/config/meson.build
index 6996e5cbeaa5..80c05bc15d2f 100644
--- a/config/meson.build
+++ b/config/meson.build
@@ -237,11 +237,22 @@ else # for 32-bit we need smaller reserved memory areas
dpdk_conf.set('RTE_MAX_MEM_MB', 2048)
endif
-
compile_time_cpuflags = []
subdir(arch_subdir)
dpdk_conf.set('RTE_COMPILE_TIME_CPUFLAGS', ','.join(compile_time_cpuflags))
+# set max lcores
+if machine != 'default' and not meson.is_cross_build()
+ # The script returns max lcores
+ params = files('get_max_lcores.py')
+ cmd_out = run_command(params)
+ if cmd_out.returncode() == 0
+ cmd_lcore = cmd_out.stdout().to_lower().strip().split(' ')
+ endif
+ max_lcore = cmd_lcore[0].to_int()
+ dpdk_conf.set('RTE_MAX_LCORE', max_lcore)
+endif
+
# set the install path for the drivers
dpdk_conf.set_quoted('RTE_EAL_PMD_PATH', eal_pmd_path)
--
2.17.1
[dpdk-dev] [PATCH 1/2] config/arm: avoid variable reuse
Rename 'machine' to 'machine_properties' in config/arm/meson.build since
'machine' is previously being used in config/meson.build
Signed-off-by: Dharmik Thakkar
Reviewed-by: Ruifeng Wang
---
config/arm/meson.build | 18 +-
1 file changed, 9 insertions(+), 9 deletions(-)
diff --git a/config/arm/meson.build b/config/arm/meson.build
index 8728051d5e38..e89ecdc4ccd2 100644
--- a/config/arm/meson.build
+++ b/config/arm/meson.build
@@ -141,12 +141,12 @@ else
dpdk_conf.set('RTE_CACHE_LINE_SIZE', 128)
dpdk_conf.set('RTE_ARCH_ARM64', 1)
- machine = []
+ machine_properties = []
cmd_generic = ['generic', '', '', 'default', '']
cmd_output = cmd_generic # Set generic by default
machine_args = [] # Clear previous machine args
if arm_force_default_march and not meson.is_cross_build()
- machine = impl_generic
+ machine_properties = impl_generic
impl_pn = 'default'
elif not meson.is_cross_build()
# The script returns ['Implementer', 'Variant', 'Architecture',
@@ -158,9 +158,9 @@ else
cmd_output = cmd.stdout().to_lower().strip().split(' ')
endif
# Set to generic if variable is not found
- machine = get_variable('impl_' + cmd_output[0], ['generic'])
- if machine[0] == 'generic'
- machine = impl_generic
+ machine_properties = get_variable('impl_' + cmd_output[0],
['generic'])
+ if machine_properties[0] == 'generic'
+ machine_properties = impl_generic
cmd_output = cmd_generic
endif
impl_pn = cmd_output[3]
@@ -170,7 +170,7 @@ else
else
impl_id = meson.get_cross_property('implementor_id', 'generic')
impl_pn = meson.get_cross_property('implementor_pn', 'default')
- machine = get_variable('impl_' + impl_id)
+ machine_properties = get_variable('impl_' + impl_id)
endif
# Apply Common Defaults. These settings may be overwritten by machine
@@ -181,14 +181,14 @@ else
endif
endforeach
- message('Implementer : ' + machine[0])
- foreach flag: machine[1]
+ message('Implementer : ' + machine_properties[0])
+ foreach flag: machine_properties[1]
if flag.length() > 0
dpdk_conf.set(flag[0], flag[1])
endif
endforeach
- foreach marg: machine[2]
+ foreach marg: machine_properties[2]
if marg[0] == impl_pn
foreach flag: marg[1]
if cc.has_argument(flag)
--
2.17.1
Re: [dpdk-dev] [PATCH v8 1/5] net/bnxt: add support for aarch32
On Tue, Aug 25, 2020 at 6:04 AM Juraj Linkeš
wrote:
> From: Ruifeng Wang
>
> Expand vector PMD support to aarch32.
>
> Signed-off-by: Ruifeng Wang
>
Acked-by: Ajit Khaparde
> ---
> drivers/net/bnxt/bnxt_rxq.h | 2 +-
> drivers/net/bnxt/bnxt_rxr.h | 2 +-
> drivers/net/bnxt/bnxt_txr.h | 2 +-
> 3 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/net/bnxt/bnxt_rxq.h b/drivers/net/bnxt/bnxt_rxq.h
> index d5ce3b6d5..1c4027711 100644
> --- a/drivers/net/bnxt/bnxt_rxq.h
> +++ b/drivers/net/bnxt/bnxt_rxq.h
> @@ -22,7 +22,7 @@ struct bnxt_rx_queue {
> uint16_tnb_rx_hold; /* num held free RX desc */
> uint16_trx_free_thresh; /* max free RX desc to
> hold */
> uint16_tqueue_id; /* RX queue index */
> -#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
> +#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) ||
> defined(RTE_ARCH_ARM)
> uint16_trxrearm_nb; /* number of descs to reinit.
> */
> uint16_trxrearm_start; /* next desc index to
> reinit. */
> #endif
> diff --git a/drivers/net/bnxt/bnxt_rxr.h b/drivers/net/bnxt/bnxt_rxr.h
> index 2bf46cd91..e2fba1647 100644
> --- a/drivers/net/bnxt/bnxt_rxr.h
> +++ b/drivers/net/bnxt/bnxt_rxr.h
> @@ -221,7 +221,7 @@ int bnxt_init_one_rx_ring(struct bnxt_rx_queue *rxq);
> int bnxt_rx_queue_start(struct rte_eth_dev *dev, uint16_t rx_queue_id);
> int bnxt_rx_queue_stop(struct rte_eth_dev *dev, uint16_t rx_queue_id);
>
> -#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
> +#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) ||
> defined(RTE_ARCH_ARM)
> uint16_t bnxt_recv_pkts_vec(void *rx_queue, struct rte_mbuf **rx_pkts,
> uint16_t nb_pkts);
> int bnxt_rxq_vec_setup(struct bnxt_rx_queue *rxq);
> diff --git a/drivers/net/bnxt/bnxt_txr.h b/drivers/net/bnxt/bnxt_txr.h
> index 7715c11b8..38e5ac9df 100644
> --- a/drivers/net/bnxt/bnxt_txr.h
> +++ b/drivers/net/bnxt/bnxt_txr.h
> @@ -59,7 +59,7 @@ uint16_t bnxt_xmit_pkts(void *tx_queue, struct rte_mbuf
> **tx_pkts,
>uint16_t nb_pkts);
> uint16_t bnxt_dummy_xmit_pkts(void *tx_queue, struct rte_mbuf **tx_pkts,
> uint16_t nb_pkts);
> -#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64)
> +#if defined(RTE_ARCH_X86) || defined(RTE_ARCH_ARM64) ||
> defined(RTE_ARCH_ARM)
> uint16_t bnxt_xmit_pkts_vec(void *tx_queue, struct rte_mbuf **tx_pkts,
> uint16_t nb_pkts);
> #endif
> --
> 2.20.1
>
>
[dpdk-dev] [PATCH 2/2] trace: add size_t as a generic trace point
From: Pavan Nikhilesh
Add size_t as a generic trace point. Also, update
test_generic_trace_point() to validate size_t emitter.
Signed-off-by: Pavan Nikhilesh
---
app/test/test_trace.c | 1 +
lib/librte_eal/common/eal_common_trace_points.c | 2 ++
lib/librte_eal/include/rte_eal_trace.h | 6 ++
lib/librte_eal/rte_eal_version.map | 3 +++
4 files changed, 12 insertions(+)
diff --git a/app/test/test_trace.c b/app/test/test_trace.c
index ddcf416b2..0f9df83c4 100644
--- a/app/test/test_trace.c
+++ b/app/test/test_trace.c
@@ -166,6 +166,7 @@ test_generic_trace_points(void)
rte_eal_trace_generic_double(2.504);
rte_eal_trace_generic_ptr(&tmp);
rte_eal_trace_generic_str("my string");
+ rte_eal_trace_generic_size_t(sizeof(void *));
RTE_EAL_TRACE_GENERIC_FUNC;
return TEST_SUCCESS;
diff --git a/lib/librte_eal/common/eal_common_trace_points.c
b/lib/librte_eal/common/eal_common_trace_points.c
index 292ec91be..0b0b25461 100644
--- a/lib/librte_eal/common/eal_common_trace_points.c
+++ b/lib/librte_eal/common/eal_common_trace_points.c
@@ -36,6 +36,8 @@ RTE_TRACE_POINT_REGISTER(rte_eal_trace_generic_ptr,
lib.eal.generic.ptr)
RTE_TRACE_POINT_REGISTER(rte_eal_trace_generic_str,
lib.eal.generic.string)
+RTE_TRACE_POINT_REGISTER(rte_eal_trace_generic_size_t,
+ lib.eal.generic.size_t)
RTE_TRACE_POINT_REGISTER(rte_eal_trace_generic_func,
lib.eal.generic.func)
diff --git a/lib/librte_eal/include/rte_eal_trace.h
b/lib/librte_eal/include/rte_eal_trace.h
index 6b1a813c7..19df549d2 100644
--- a/lib/librte_eal/include/rte_eal_trace.h
+++ b/lib/librte_eal/include/rte_eal_trace.h
@@ -129,6 +129,12 @@ RTE_TRACE_POINT(
rte_trace_point_emit_string(str);
)
+RTE_TRACE_POINT(
+ rte_eal_trace_generic_size_t,
+ RTE_TRACE_POINT_ARGS(size_t sz),
+ rte_trace_point_emit_size_t(sz);
+)
+
RTE_TRACE_POINT(
rte_eal_trace_generic_func,
RTE_TRACE_POINT_ARGS(const char *func),
diff --git a/lib/librte_eal/rte_eal_version.map
b/lib/librte_eal/rte_eal_version.map
index 0b18e2ef8..4ccfa593f 100644
--- a/lib/librte_eal/rte_eal_version.map
+++ b/lib/librte_eal/rte_eal_version.map
@@ -397,6 +397,9 @@ EXPERIMENTAL {
rte_mp_disable;
rte_thread_register;
rte_thread_unregister;
+
+ #added in 20.11
+ __rte_eal_trace_generic_size_t;
};
INTERNAL {
--
2.17.1
[dpdk-dev] [PATCH 1/2] trace: fix size_t field emitter
From: Pavan Nikhilesh
Add size_t CTF format metadata, this is needed by CTF analyzers to
parse the emitted CTF trace.
Fixes: 262c4ee791c6 ("trace: add size_t field emitter")
Signed-off-by: Pavan Nikhilesh
---
lib/librte_eal/common/eal_common_trace_ctf.c | 5 +
1 file changed, 5 insertions(+)
diff --git a/lib/librte_eal/common/eal_common_trace_ctf.c
b/lib/librte_eal/common/eal_common_trace_ctf.c
index 302e2bb74..9dc91df0f 100644
--- a/lib/librte_eal/common/eal_common_trace_ctf.c
+++ b/lib/librte_eal/common/eal_common_trace_ctf.c
@@ -83,6 +83,11 @@ meta_data_type_emit(char **meta, int *offset)
"typealias integer {size = 32; base = x;} := long;\n"
#endif
"typealias integer {size = 8; signed = false; encoding = ASCII;
} := string_bounded_t;\n\n"
+#ifdef RTE_ARCH_64
+ "typealias integer {size = 64; base = x;} := size_t;\n"
+#else
+ "typealias integer {size = 32; base = x;} := size_t;\n"
+#endif
"typealias floating_point {\n"
"exp_dig = 8;\n"
"mant_dig = 24;\n"
--
2.17.1
[dpdk-dev] [PATCH v1] doc: update ice production information
The E810 (ice) has been shipping to customers, not just for evaluation any more. Signed-off-by: Haiyue Wang --- doc/guides/nics/ice.rst | 8 1 file changed, 8 deletions(-) diff --git a/doc/guides/nics/ice.rst b/doc/guides/nics/ice.rst index 9a9f4a6bb..6ac657c13 100644 --- a/doc/guides/nics/ice.rst +++ b/doc/guides/nics/ice.rst @@ -12,9 +12,6 @@ the Intel Ethernet Controller E810. Prerequisites - -- The E810 is currently in sampling state only. To obtain early samples and/or get further information - about kernel drivers, firmware and DDP support, please speak to your Intel representative. - - Follow the DPDK :ref:`Getting Started Guide for Linux ` to setup the basic DPDK environment. - To get better performance on Intel platforms, please follow the "How to get best performance with NICs on Intel platforms" @@ -310,8 +307,3 @@ is stored in ``ice_adapter->active_pkg_type``. A symbolic link to the DDP package file is also ok. The same package file is used by both the kernel driver and the DPDK PMD. - -limitation -~~ - -Ice code released is for evaluation only currently. -- 2.28.0
[dpdk-dev] [PATCH 1/3] ethdev: add a field for rte_eth_rxq_info
Add a field named rx_buf_size in rte_eth_rxq_info to indicate the buffer
size used in receiving packets for HW.
In this way, upper-layer users can get this information by calling
rte_eth_rx_queue_info_get.
Signed-off-by: Chengchang Tang
Acked-by: Andrew Rybchenko
---
lib/librte_ethdev/rte_ethdev.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/lib/librte_ethdev/rte_ethdev.h b/lib/librte_ethdev/rte_ethdev.h
index 70295d7..9fed5cb 100644
--- a/lib/librte_ethdev/rte_ethdev.h
+++ b/lib/librte_ethdev/rte_ethdev.h
@@ -1420,6 +1420,8 @@ struct rte_eth_rxq_info {
struct rte_eth_rxconf conf; /**< queue config parameters. */
uint8_t scattered_rx; /**< scattered packets RX supported. */
uint16_t nb_desc; /**< configured number of RXDs. */
+ /**< buffer size used for hardware when receive packets. */
+ uint16_t rx_buf_size;
} __rte_cache_min_aligned;
/**
--
2.7.4
[dpdk-dev] [PATCH 2/3] app/testpmd: Add RX buffer size display in queue info querry
Add RX buffer size to queue info querry cmd so that the user can get the
buffer length used by HW queue for receiving packets.
Signed-off-by: Chengchang Tang
---
app/test-pmd/config.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/app/test-pmd/config.c b/app/test-pmd/config.c
index 30bee33..b432ac6 100644
--- a/app/test-pmd/config.c
+++ b/app/test-pmd/config.c
@@ -452,6 +452,7 @@ rx_queue_infos_display(portid_t port_id, uint16_t queue_id)
(qinfo.conf.rx_deferred_start != 0) ? "on" : "off");
printf("\nRX scattered packets: %s",
(qinfo.scattered_rx != 0) ? "on" : "off");
+ printf("\nRX buffer size: %hu", qinfo.rx_buf_size);
printf("\nNumber of RXDs: %hu", qinfo.nb_desc);
if (rte_eth_rx_burst_mode_get(port_id, queue_id, &mode) == 0)
--
2.7.4
[dpdk-dev] [PATCH 0/3] add RX buffer size for rte_eth_rxq_info
In common practice, PMD configure the RX buffer size which indicate the buffer length could be used for HW in receiving packets according to the data room size of the object in mempool. But in fact, the final value is related to the specifications of HW, and its values will affect the number of fragments in receiving packets when scatter is enabled. By the way, some PMDs may force to enable scatter when the MTU is bigger than the HW RX buffer size. At present, we have no way to expose relevant information to upper layer users. So, add a field named rx_buf_size in rte_eth_rxq_info to indicate the buffer size used in receiving packets for HW. And this field is optional, so there is no need to forcibly update all PMDs. This patchset also add hns3 PMD implementation and update the testpmd. Chengchang Tang (3): ethdev: add a field for rte_eth_rxq_info app/testpmd: Add RX buffer size display in queue info querry net/hns3: add rx buffer size to rx qinfo querry app/test-pmd/config.c | 1 + drivers/net/hns3/hns3_rxtx.c | 2 ++ lib/librte_ethdev/rte_ethdev.h | 2 ++ 3 files changed, 5 insertions(+) -- 2.7.4
[dpdk-dev] [PATCH 3/3] net/hns3: add RX buffer size to rx qinfo querry
Report hns3 PMD configured RX buffer size in RX qinfo querry. Signed-off-by: Chengchang Tang --- drivers/net/hns3/hns3_rxtx.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/net/hns3/hns3_rxtx.c b/drivers/net/hns3/hns3_rxtx.c index fc1a256..50881d4 100644 --- a/drivers/net/hns3/hns3_rxtx.c +++ b/drivers/net/hns3/hns3_rxtx.c @@ -2830,6 +2830,8 @@ hns3_rxq_info_get(struct rte_eth_dev *dev, uint16_t queue_id, qinfo->mp = rxq->mb_pool; qinfo->nb_desc = rxq->nb_rx_desc; qinfo->scattered_rx = dev->data->scattered_rx; + /* Report the HW RX buffer length to user */ + qinfo->rx_buf_size = rxq->rx_buf_len; /* * If there are no available Rx buffer descriptors, incoming packets -- 2.7.4
Re: [dpdk-dev] 19.11.3 patches review and test
Hi Luca, Update LTS 19.11.4 test result for Intel part. No new issue is found except known issues. * Intel(R) Testing # Basic Intel(R) NIC testing * PF(i40e):Passed - Exception message when starting testpmd for testing external memory. It has been fixed on 20.05.patch link: http://patches.dpdk.org/patch/66041/ * PF(ixgbe):Passed * PF(ice):Passed - The MD5 is not same between kernel ethtool and dpdk ethtool when testing userspace_ethtool/retrieve_eeprom. It has been fixed on 20.08.patch link: http://patches.dpdk.org/patch/72354/ - With latest ice driver and firmware package exception found when port reset vf of testpmd. It is also found on latest dpdk version, dev is debugging. * VF(i40e):Passed * VF(ixgbe):Passed * VF(ice):Passed * Build or compile: Passed * Intel NIC single core/NIC performance: Passed #Basic cryptodev and virtio testing * vhost/virtio basic loopback, PVP and performance test: Passed. - udp-fragmentation-offload can't be setup on vm Ubuntu1910.it is kernel issue, tracked on: https://bugzilla.kernel.org/show_bug.cgi?id=207075 - l3fwd-power can wake up lcore but can not sleep again. It is also found on 20.08, dev is debugging. * cryptodev Function: Passed. - fips_cryptodev test failed for TDES. It is also found on 20.08, dev is debugging. https://bugs.dpdk.org/show_bug.cgi?id=512 * cryptodev Performance: Passed. - known unstable issue of test case 1c1t 3CPM. not effect LTS release. Thanks. Regards, Chen Bo > -Original Message- > From: Yu, PingX > Sent: June 10, 2020 15:20 > To: luca.bocca...@gmail.com; sta...@dpdk.org > Cc: dev@dpdk.org; Abhishek Marathe ; > Akhil Goyal ; Ali Alnubani ; > Walker, Benjamin ; David Christensen > ; Hemant Agrawal ; > Stokes, Ian ; Jerin Jacob ; > Mcnamara, John ; Ju-Hyoung Lee > ; Kevin Traynor ; Pei Zhang > ; Xu, Qian Q ; Raslan > Darawsheh ; Thomas Monjalon > ; Peng, Yuan ; Chen, > Zhaoyan > Subject: RE: 19.11.3 patches review and test > > Hi Luca, > Update LTS 19.11.3 test result for Intel part. All passed and no new issue is > found. > > * Intel(R) Testing > > # Basic Intel(R) NIC testing > * PF(i40e):Passed > * PF(ixgbe):Passed > * PF(ice):Passed > * VF(i40e):Passed > * VF(ixgbe):Passed > * VF(ice):Passed > * Build or compile: Passed > * Intel NIC single core/NIC performance: Passed > > #Basic cryptodev and virtio testing > * vhost/virtio basic loopback, PVP and performance test: Passed. > known issue: https://bugzilla.kernel.org/show_bug.cgi?id=207075 > * cryptodev Function: Passed. > * cryptodev Performance: Passed. > known unstable issue of test case 1c1t 3CPM. not effect LTS release. > > Thanks. > Regards, > Yu Ping > > > -Original Message- > > From: luca.bocca...@gmail.com > > Sent: Thursday, June 4, 2020 3:44 AM > > To: sta...@dpdk.org > > Cc: dev@dpdk.org; Abhishek Marathe > > ; Akhil Goyal ; > > Ali Alnubani ; Walker, Benjamin > > ; David Christensen > > ; Hemant Agrawal ; > > Stokes, Ian ; Jerin Jacob ; > > Mcnamara, John ; Ju-Hyoung Lee > > ; Kevin Traynor ; Pei > Zhang > > ; Yu, PingX ; Xu, Qian Q > > ; Raslan Darawsheh ; > Thomas > > Monjalon ; Peng, Yuan ; > > Chen, Zhaoyan > > Subject: 19.11.3 patches review and test > > > > Hi all, > > > > Here is a list of patches targeted for stable release 19.11.3. > > > > The planned date for the final release is the 17th of June. > > > > Please help with testing and validation of your use cases and report > > any issues/results with reply-all to this mail. For the final release > > the fixes and reported validations will be added to the release notes. > > > > A release candidate tarball can be found at: > > > > https://dpdk.org/browse/dpdk-stable/tag/?id=v19.11.3-rc1 > > > > These patches are located at branch 19.11 of dpdk-stable repo: > > https://dpdk.org/browse/dpdk-stable/ > > > > Thanks. > > > > Luca Boccassi > > > > --- > > Adam Dybkowski (5): > > cryptodev: fix missing device id range checking > > common/qat: fix GEN3 marketing name > > app/crypto-perf: fix display of sample test vector > > crypto/qat: support plain SHA1..SHA512 hashes > > cryptodev: fix SHA-1 digest enum comment > > > > Ajit Khaparde (3): > > net/bnxt: fix FW version query > > net/bnxt: fix error log for command timeout > > net/bnxt: fix using RSS config struct > > > > Akhil Goyal (1): > > ipsec: fix build dependency on hash lib > > > > Alex Kiselev (1): > > lpm6: fix size of tbl8 group > > > > Alex Marginean (1): > > net/enetc: fix Rx lock-up > > > > Alexander Kozyrev (8): > > net/mlx5: reduce Tx completion index memory loads > > net/mlx5: add device parameter for MPRQ stride size > > net/mlx5: enable MPRQ multi-stride operations > > net/mlx5: add multi-segment packets in MPRQ mode > > net/mlx5: set dynamic flow metadata in Rx queues > > net/mlx5: improve logging of MPRQ selection > > net/mlx5: fix assert in d
[dpdk-dev] Recall: 19.11.3 patches review and test
Chen, BoX C would like to recall the message, "19.11.3 patches review and test".
Re: [dpdk-dev] 19.11.4 patches review and test
Hi Luca, Update LTS 19.11.4 test result for Intel part. No new issue is found except known issues. * Intel(R) Testing # Basic Intel(R) NIC testing * PF(i40e):Passed - Exception message when starting testpmd for testing external memory. It has been fixed on 20.05.patch link: http://patches.dpdk.org/patch/66041/ * PF(ixgbe):Passed * PF(ice):Passed - The MD5 is not same between kernel ethtool and dpdk ethtool when testing userspace_ethtool/retrieve_eeprom. It has been fixed on 20.08.patch link: http://patches.dpdk.org/patch/72354/ - With latest ice driver and firmware package exception found when port reset vf of testpmd. It is also found on latest dpdk version, dev is debugging. * VF(i40e):Passed * VF(ixgbe):Passed * VF(ice):Passed * Build or compile: Passed * Intel NIC single core/NIC performance: Passed #Basic cryptodev and virtio testing * vhost/virtio basic loopback, PVP and performance test: Passed. - udp-fragmentation-offload can't be setup on vm Ubuntu1910.it is kernel issue, tracked on: https://bugzilla.kernel.org/show_bug.cgi?id=207075 - l3fwd-power can wake up lcore but can not sleep again. It is also found on 20.08, dev is debugging. * cryptodev Function: Passed. - fips_cryptodev test failed for TDES. It is also found on 20.08, dev is debugging. https://bugs.dpdk.org/show_bug.cgi?id=512 * cryptodev Performance: Passed. - known unstable issue of test case 1c1t 3CPM. not effect LTS release. Thanks. Regards, Chen Bo > -Original Message- > From: dev On Behalf Of Luca Boccassi > Sent: August 19, 2020 2:12 > To: sta...@dpdk.org > Cc: dev@dpdk.org; Abhishek Marathe ; > Akhil Goyal ; Ali Alnubani ; > Walker, Benjamin ; David Christensen > ; Hemant Agrawal ; > Stokes, Ian ; Jerin Jacob ; > Mcnamara, John ; Ju-Hyoung Lee > ; Kevin Traynor ; Pei Zhang > ; Yu, PingX ; Xu, Qian Q > ; Raslan Darawsheh ; > Thomas Monjalon ; Peng, Yuan > ; Chen, Zhaoyan > Subject: [dpdk-dev] 19.11.4 patches review and test > > Hi all, > > Here is a list of patches targeted for stable release 19.11.4. > > The planned date for the final release is August 31st. > > Please help with testing and validation of your use cases and report any > issues/results with reply-all to this mail. For the final release the fixes > and > reported validations will be added to the release notes. > > A release candidate tarball can be found at: > > https://dpdk.org/browse/dpdk-stable/tag/?id=v19.11.4-rc1 > > These patches are located at branch 19.11 of dpdk-stable repo: > https://dpdk.org/browse/dpdk-stable/ > > Thanks. > > Luca Boccassi > > --- > Adam Dybkowski (7): > test/crypto: fix asymmetric session mempool creation > crypto/qat: fix AES-XTS capabilities > crypto/qat: handle mixed hash-cipher requests on GEN3 > test/crypto: add mixed encypted-digest > common/qat: get firmware version > crypto/qat: handle mixed hash-cipher on GEN2 > common/qat: fix uninitialized variable > > Akhil Goyal (4): > crypto/dpaax_sec: fix 18-bit PDCP cases with HFN override > crypto/dpaax_sec: fix inline query for descriptors > crypto/dpaa2_sec: fix HFN override > common/dpaax: fix 12-bit null auth case > > Alan Dewar (1): > sched: fix port time rounding > > Alexander Kozyrev (3): > net/mlx5: fix vectorized Rx burst termination > net/mlx5: do not select legacy MPW implicitly > net/mlx5: fix descriptors number adjustment > > Ali Alnubani (2): > net/mlx5: fix typos in meter error messages > doc: fix a typo in mlx5 guide > > Alvin Zhang (1): > net/i40e: fix queue pairs configuration in VF > > Andrew Rybchenko (1): > net/sfc: do not enforce hash offload in RSS multi-queue > > Ankur Dwivedi (2): > common/cpt: fix encryption offset > test/crypto: change cipher offset for ESN vector > > Apeksha Gupta (2): > app/eventdev: fix capability check in pipeline ATQ test > event/dpaa2: add all-types queue capability flag > > Archana Muniganti (2): > examples/fips_validation: fix TDES interim callback > examples/fips_validation: fix count overwrite for TDES > > Archit Pandey (1): > sched: fix 64-bit rate > > Arek Kusztal (4): > drivers/qat: improve multi process on qat > drivers/qat: add multi process handling of driver id > drivers/qat: add handling of capabilities in multi process > crypto/qat: add minimum enq threshold > > Ayuj Verma (1): > examples/fips_validation: fix parsing of TDES vectors > > Bruce Richardson (12): > test: fix build with ring PMD but no bond PMD > mem: fix 32-bit init config with meson > rawdev: allow getting info for unknown device > rawdev: fill NUMA socket ID in info > rawdev: export dump function in map file > doc: rebuild with meson whenever a file changes > build: always link whole DPDK static libraries > devtools: test static lin
[dpdk-dev] [PATCH] gso: fix pyld_unit_size issue for udp gso
From: Yi Yang
IP fragment offset in IP header must be multiple of
8, if pyld_unit_size isn't multiple of 8, IP total
length will be wrong after reassemble, IP payload
also will be wrong because some data is lost on
reassembling.
This has been verified in OVS DPDK by VLAN TSO iperf3
UDP case.
Fixes: b166d4f30b66 ("gso: support UDP/IPv4 fragmentation")
Signed-off-by: Yi Yang
---
lib/librte_gso/gso_udp4.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/lib/librte_gso/gso_udp4.c b/lib/librte_gso/gso_udp4.c
index 21fea09..6fa68f2 100644
--- a/lib/librte_gso/gso_udp4.c
+++ b/lib/librte_gso/gso_udp4.c
@@ -69,7 +69,10 @@
return 1;
}
- pyld_unit_size = gso_size - hdr_offset;
+ /* pyld_unit_size must be a multiple of 8 because frag_off
+* uses 8 bytes as unit.
+*/
+ pyld_unit_size = (gso_size - hdr_offset) & ~7U;
/* Segment the payload */
ret = gso_do_segment(pkt, hdr_offset, pyld_unit_size, direct_pool,
--
1.8.3.1
Re: [dpdk-dev] [PATCH v8 5/5] doc: add aarch32 build guidance
Aaron Conole writes: > Subject: Re: [PATCH v8 5/5] doc: add aarch32 build guidance > > Juraj Linkeš writes: > > > From: Phil Yang > > > > Add cross-compiling guidance for 32-bit aarch32 DPDK on aarch64 host. > > > > Signed-off-by: Phil Yang > > --- > > .../linux_gsg/cross_build_dpdk_for_arm64.rst | 30 ++- > > 1 file changed, 23 insertions(+), 7 deletions(-) > > > > diff --git a/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > b/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > > index c5875a6d5..07085e3ea 100644 > > --- a/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > > +++ b/doc/guides/linux_gsg/cross_build_dpdk_for_arm64.rst > > @@ -1,15 +1,16 @@ > > .. SPDX-License-Identifier: BSD-3-Clause > > Copyright(c) 2018 ARM Corporation. > > > > -Cross compile DPDK for ARM64 > > - > > -This chapter describes how to cross compile DPDK for ARM64 from x86 > build hosts. > > +Cross compile DPDK for aarch64 and aarch32 > > +== > > +This chapter describes how to cross compile DPDK for aarch64 from x86 > build hosts and compile > > +32-bit aarch32 DPDK from aarch64 build hosts. > > > > .. note:: > > > > - Whilst it is recommended to natively build DPDK on ARM64 (just > > - like with x86), it is also possible to cross-build DPDK for ARM64. An > > - ARM64 cross compile GNU toolchain is used for this. > > + Whilst it is recommended to natively build DPDK on aarch64 (just > > + like with x86), it is also possible to cross-build DPDK for aarch64. > > + An aarch64 cross compile GNU toolchain is used for this. > > > > Obtain the cross tool chain > > --- > > @@ -88,18 +89,33 @@ To install it in Ubuntu:: > > > > sudo apt-get install pkg-config-aarch64-linux-gnu > > > > +.. note:: > > + > > +Some aarch64 platforms support EL0 aarch32 mode, which means the > 32-bit aarch32 applications > > +can run on aarch64. The armhf architecture toolchain ``gcc-arm-linux- > gnueabihf`` is required > > +for aarch32 on aarch64. To install it in Ubuntu:: > > + > > + sudo dpkg --add-architecture armhf > > + sudo apt-get update > > + sudo apt-get install -y gcc gcc-arm-linux-gnueabihf libc6:armhf make > binutils > > + > > Is 'make' still needed? I don't run Ubuntu, so I'm not sure. No, both 'GCC' and 'make' are not mandatory in this case. Will update it in the next version. > > > To cross-compile DPDK on a desired target machine we can use the > following > > command:: > > > > meson cross-build --cross-file > > ninja -C cross-build > > > > -For example if the target machine is arm64 we can use the following ^^^ Will use 'users' instead > > +For example if the target machine is aarch64 we can use the following ^^^ Will use 'users' instead > > command:: > > > > meson arm64-build --cross-file config/arm/arm64_armv8_linux_gcc > > ninja -C arm64-build > > > > +If the target machine is aarch32 we can use the following command:: ^^^ Will use 'users' instead > > + > > + meson arm32-build --cross-file config/arm/arm32_armv8a_linux_gcc > > + ninja -C arm32-build > > + > > Configure and Cross Compile DPDK using Make > > --- > > To configure a build, choose one of the target configurations, like arm64- > dpaa-linux-gcc and arm64-thunderx-linux-gcc.
[dpdk-dev] [PATCH] net/ice: add support for Auxiliary IP Offset Rx descriptor
Add RXDID #25 to support Auxiliary IP Offset Rx descriptor, including
FlexiMD.4: Outer/Single IPv4 Header offset
FlexiMD.5: Outer/Single IPv6 Header offset
And parse the valid IP Offset into mbuf by flexible descriptor
section via devargs "proto_xtr" with "proto_xtr=ip_offset".
This patch depends on:
commit e0c8ed904083 (net/ice: optimize the FlexiMD hardware check)
Signed-off-by: Junfeng Guo
---
doc/guides/nics/ice.rst | 14 -
drivers/net/ice/base/ice_lan_tx_rx.h| 1 +
drivers/net/ice/ice_ethdev.c| 10 +++-
drivers/net/ice/ice_ethdev.h| 1 +
drivers/net/ice/ice_rxtx.c | 77 +
drivers/net/ice/rte_pmd_ice.h | 13 +
drivers/net/ice/rte_pmd_ice_version.map | 1 +
7 files changed, 90 insertions(+), 27 deletions(-)
diff --git a/doc/guides/nics/ice.rst b/doc/guides/nics/ice.rst
index 9a9f4a6bb..ffaaad4e4 100644
--- a/doc/guides/nics/ice.rst
+++ b/doc/guides/nics/ice.rst
@@ -102,7 +102,7 @@ Runtime Config Options
The grouping ``()`` can be omitted for single element group. If no queues are
specified, PMD will use this protocol extraction type for all queues.
- Protocol is : ``vlan, ipv4, ipv6, ipv6_flow, tcp``.
+ Protocol is : ``vlan, ipv4, ipv6, ipv6_flow, tcp, ip_offset``.
.. code-block:: console
@@ -183,6 +183,18 @@ Runtime Config Options
TCPHDR2 - Reserved
+ .. table:: Protocol extraction : ``ip_offset``
+
+ +++
+ | IPHDR2 | IPHDR1 |
+ +++
+ | IPv6 HDR Offset | IPv4 HDR Offset |
+ +++
+
+ IPHDR1 - Outer/Single IPv4 Header offset.
+
+ IPHDR2 - Outer/Single IPv6 Header offset.
+
Use ``rte_net_ice_dynf_proto_xtr_metadata_get`` to access the protocol
extraction metadata, and use ``RTE_PKT_RX_DYNF_PROTO_XTR_*`` to get the
metadata type of ``struct rte_mbuf::ol_flags``.
diff --git a/drivers/net/ice/base/ice_lan_tx_rx.h
b/drivers/net/ice/base/ice_lan_tx_rx.h
index c47114d16..ec0c9f3ab 100644
--- a/drivers/net/ice/base/ice_lan_tx_rx.h
+++ b/drivers/net/ice/base/ice_lan_tx_rx.h
@@ -715,6 +715,7 @@ enum ice_rxdid {
ICE_RXDID_COMMS_AUX_IPV6= 19,
ICE_RXDID_COMMS_AUX_IPV6_FLOW = 20,
ICE_RXDID_COMMS_AUX_TCP = 21,
+ ICE_RXDID_COMMS_AUX_IP_OFFSET = 25,
ICE_RXDID_LAST = 63,
};
diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c
index d67390f29..4a161ae8a 100644
--- a/drivers/net/ice/ice_ethdev.c
+++ b/drivers/net/ice/ice_ethdev.c
@@ -65,6 +65,9 @@ static struct proto_xtr_ol_flag
ice_proto_xtr_ol_flag_params[] = {
[PROTO_XTR_TCP] = {
.param = { .name = "ice_dynflag_proto_xtr_tcp" },
.ol_flag = &rte_net_ice_dynflag_proto_xtr_tcp_mask },
+ [PROTO_XTR_IP_OFFSET] = {
+ .param = { .name = "ice_dynflag_proto_xtr_ip_offset" },
+ .ol_flag = &rte_net_ice_dynflag_proto_xtr_ip_offset_mask },
};
#define ICE_DFLT_OUTER_TAG_TYPE ICE_AQ_VSI_OUTER_TAG_VLAN_9100
@@ -323,6 +326,7 @@ lookup_proto_xtr_type(const char *xtr_name)
{ "ipv6", PROTO_XTR_IPV6 },
{ "ipv6_flow", PROTO_XTR_IPV6_FLOW },
{ "tcp", PROTO_XTR_TCP },
+ { "ip_offset", PROTO_XTR_IP_OFFSET },
};
uint32_t i;
@@ -568,6 +572,10 @@ ice_check_proto_xtr_support(struct ice_hw *hw)
[PROTO_XTR_TCP] = { ICE_RXDID_COMMS_AUX_TCP,
ICE_RX_OPC_EXTRACT,
ICE_PROT_TCP_IL, ICE_PROT_ID_INVAL },
+ [PROTO_XTR_IP_OFFSET] = { ICE_RXDID_COMMS_AUX_IP_OFFSET,
+ ICE_RX_OPC_PROTID,
+ ICE_PROT_IPV4_OF_OR_S,
+ ICE_PROT_IPV6_OF_OR_S },
};
uint32_t i;
@@ -5145,7 +5153,7 @@ RTE_PMD_REGISTER_PCI(net_ice, rte_ice_pmd);
RTE_PMD_REGISTER_PCI_TABLE(net_ice, pci_id_ice_map);
RTE_PMD_REGISTER_KMOD_DEP(net_ice, "* igb_uio | uio_pci_generic | vfio-pci");
RTE_PMD_REGISTER_PARAM_STRING(net_ice,
- ICE_PROTO_XTR_ARG
"=[queue:]"
+ ICE_PROTO_XTR_ARG
"=[queue:]"
ICE_SAFE_MODE_SUPPORT_ARG "=<0|1>"
ICE_PIPELINE_MODE_SUPPORT_ARG "=<0|1>"
ICE_FLOW_MARK_SUPPORT_ARG "=<0|1>");
diff --git a/drivers/net/ice/ice_ethdev.h b/drivers/net/ice/ice_ethdev.h
index 06b9b8683..758caa8b4 100644
--- a/drivers/net/ice/ice_ethdev.h
+++ b/drivers/net/ice/ice_ethdev.h
@@ -262,6 +262,7 @@ enum proto_xtr_type {
PROTO_XTR_IPV6,
PROTO_XTR_IPV6_FLOW,
PROTO_XTR_TCP,
+ PROTO_XTR_IP
Re: [dpdk-dev] [PATCH 2/2] build: find max lcore programmatically
On Wed, Aug 26, 2020 at 2:44 AM Dharmik Thakkar wrote:
>
> For Arm, RTE_MAX_LCORE is hard-coded into the config. It leads to incorrect
> RTE_MAX_LCORE when machines have same Implemener and part number
> but different number of CPUs.
> For x86, RTE_MAX_LCORE is always set to 128 (using the value
> set in meson_options.txt)
>
> Use python script to find max lcore when using native build to
> correctly set RTE_MAX_LCORE.
We may need to build on the native arm64 machine and use it on another
arm64 machine(Just like x86).
So I think, at least for default config(which will be used by
distribution) to support max
lcores as fixed. I am not sure this patch changes those aspects or
not? Please check.
>
> Signed-off-by: Dharmik Thakkar
> Reviewed-by: Ruifeng Wang
> ---
> config/get_max_lcores.py | 13 +
> config/meson.build | 13 -
> 2 files changed, 25 insertions(+), 1 deletion(-)
> create mode 100755 config/get_max_lcores.py
>
> diff --git a/config/get_max_lcores.py b/config/get_max_lcores.py
> new file mode 100755
> index ..ebf1c7efdadd
> --- /dev/null
> +++ b/config/get_max_lcores.py
> @@ -0,0 +1,13 @@
> +#!/usr/bin/python3
> +# SPDX-License-Identifier: BSD-3-Clause
> +# Copyright(c) 2020 Arm Limited
> +
> +import os
> +
> +max_lcores = []
> +
> +nCPU = os.cpu_count()
> +
> +max_lcores.append(str(nCPU & 0xFFF)) # Number of CPUs
> +
> +print(' '.join(max_lcores))
> diff --git a/config/meson.build b/config/meson.build
> index 6996e5cbeaa5..80c05bc15d2f 100644
> --- a/config/meson.build
> +++ b/config/meson.build
> @@ -237,11 +237,22 @@ else # for 32-bit we need smaller reserved memory areas
> dpdk_conf.set('RTE_MAX_MEM_MB', 2048)
> endif
>
> -
> compile_time_cpuflags = []
> subdir(arch_subdir)
> dpdk_conf.set('RTE_COMPILE_TIME_CPUFLAGS', ','.join(compile_time_cpuflags))
>
> +# set max lcores
> +if machine != 'default' and not meson.is_cross_build()
> + # The script returns max lcores
> + params = files('get_max_lcores.py')
> + cmd_out = run_command(params)
> + if cmd_out.returncode() == 0
> + cmd_lcore = cmd_out.stdout().to_lower().strip().split(' ')
> + endif
> + max_lcore = cmd_lcore[0].to_int()
> + dpdk_conf.set('RTE_MAX_LCORE', max_lcore)
> +endif
> +
> # set the install path for the drivers
> dpdk_conf.set_quoted('RTE_EAL_PMD_PATH', eal_pmd_path)
>
> --
> 2.17.1
>
Re: [dpdk-dev] [PATCH 2/2] build: find max lcore programmatically
> On Aug 25, 2020, at 11:47 PM, Jerin Jacob wrote:
>
> On Wed, Aug 26, 2020 at 2:44 AM Dharmik Thakkar
> wrote:
>>
>> For Arm, RTE_MAX_LCORE is hard-coded into the config. It leads to incorrect
>> RTE_MAX_LCORE when machines have same Implemener and part number
>> but different number of CPUs.
>> For x86, RTE_MAX_LCORE is always set to 128 (using the value
>> set in meson_options.txt)
>>
>> Use python script to find max lcore when using native build to
>> correctly set RTE_MAX_LCORE.
>
> We may need to build on the native arm64 machine and use it on another
> arm64 machine(Just like x86).
> So I think, at least for default config(which will be used by
> distribution) to support max
> lcores as fixed. I am not sure this patch changes those aspects or
> not? Please check.
This patch does *not* affect ‘default’ build type and cross-compilation.
>
>>
>> Signed-off-by: Dharmik Thakkar
>> Reviewed-by: Ruifeng Wang
>> ---
>> config/get_max_lcores.py | 13 +
>> config/meson.build | 13 -
>> 2 files changed, 25 insertions(+), 1 deletion(-)
>> create mode 100755 config/get_max_lcores.py
>>
>> diff --git a/config/get_max_lcores.py b/config/get_max_lcores.py
>> new file mode 100755
>> index ..ebf1c7efdadd
>> --- /dev/null
>> +++ b/config/get_max_lcores.py
>> @@ -0,0 +1,13 @@
>> +#!/usr/bin/python3
>> +# SPDX-License-Identifier: BSD-3-Clause
>> +# Copyright(c) 2020 Arm Limited
>> +
>> +import os
>> +
>> +max_lcores = []
>> +
>> +nCPU = os.cpu_count()
>> +
>> +max_lcores.append(str(nCPU & 0xFFF)) # Number of CPUs
>> +
>> +print(' '.join(max_lcores))
>> diff --git a/config/meson.build b/config/meson.build
>> index 6996e5cbeaa5..80c05bc15d2f 100644
>> --- a/config/meson.build
>> +++ b/config/meson.build
>> @@ -237,11 +237,22 @@ else # for 32-bit we need smaller reserved memory areas
>>dpdk_conf.set('RTE_MAX_MEM_MB', 2048)
>> endif
>>
>> -
>> compile_time_cpuflags = []
>> subdir(arch_subdir)
>> dpdk_conf.set('RTE_COMPILE_TIME_CPUFLAGS', ','.join(compile_time_cpuflags))
>>
>> +# set max lcores
>> +if machine != 'default' and not meson.is_cross_build()
>> + # The script returns max lcores
>> + params = files('get_max_lcores.py')
>> + cmd_out = run_command(params)
>> + if cmd_out.returncode() == 0
>> + cmd_lcore = cmd_out.stdout().to_lower().strip().split(' ')
>> + endif
>> + max_lcore = cmd_lcore[0].to_int()
>> + dpdk_conf.set('RTE_MAX_LCORE', max_lcore)
>> +endif
>> +
>> # set the install path for the drivers
>> dpdk_conf.set_quoted('RTE_EAL_PMD_PATH', eal_pmd_path)
>>
>> --
>> 2.17.1
>>
Re: [dpdk-dev] [PATCH v2] net/dpaa: fix public API port ID type
Acked-by: Hemant Agrawal
> -Original Message-
> From: Ferruh Yigit
> Sent: Tuesday, August 25, 2020 11:21 PM
> To: dev@dpdk.org; Ray Kinsella ; Neil Horman
> ; John McNamara ;
> Marko Kovacevic ; Hemant Agrawal
> ; Sachin Saxena
> Cc: Ferruh Yigit ; sta...@dpdk.org
> Subject: [PATCH v2] net/dpaa: fix public API port ID type
>
> Updating the type for 'port' variable from 'uint8_t' to 'uint16_t'.
>
> Fixes: 8c3495f5d2dd ("net/dpaa: support loopback API")
> Cc: sta...@dpdk.org
>
> Signed-off-by: Ferruh Yigit
> ---
> v2:
> * Remove deprecation notice
> ---
> doc/guides/rel_notes/deprecation.rst | 3 ---
> drivers/net/dpaa/dpaa_ethdev.c | 2 +-
> drivers/net/dpaa/rte_pmd_dpaa.h | 2 +-
> 3 files changed, 2 insertions(+), 5 deletions(-)
>
> diff --git a/doc/guides/rel_notes/deprecation.rst
> b/doc/guides/rel_notes/deprecation.rst
> index 345c38d5b6..249d4c50a1 100644
> --- a/doc/guides/rel_notes/deprecation.rst
> +++ b/doc/guides/rel_notes/deprecation.rst
> @@ -254,9 +254,6 @@ Deprecation Notices
>experimental in v20.11 DPDK release. For more details, please see `the
> thread
>
> s.dpdk.org%2Farchives%2Fdev%2F2020-
> April%2F164970.html&data=02%7C01%7Chemant.agrawal%40nxp.com
> %7C7277661bfed84b6efd8e08d8491f754a%7C686ea1d3bc2b4c6fa92cd99c5c3
> 01635%7C0%7C0%7C637339746748499018&sdata=zhi6GXeFItNrWhst3L1
> TTOPqTG47kthyyYHBe8Sn7Mw%3D&reserved=0>`_.
>
> -* pmd_dpaa: The API ``rte_pmd_dpaa_set_tx_loopback`` will have
> extended
> - ``port_id`` definition from ``uint8_t`` to ``uint16_t``.
> -
> * vhost: Vhost-user dequeue zero-copy support will be removed in 20.11.
>The only known user is OVS where the feature is still experimental,
>and has not received any update for 2.5 years.
> diff --git a/drivers/net/dpaa/dpaa_ethdev.c
> b/drivers/net/dpaa/dpaa_ethdev.c index c15e2b5462..e84f904071 100644
> --- a/drivers/net/dpaa/dpaa_ethdev.c
> +++ b/drivers/net/dpaa/dpaa_ethdev.c
> @@ -1354,7 +1354,7 @@ is_dpaa_supported(struct rte_eth_dev *dev) }
>
> int
> -rte_pmd_dpaa_set_tx_loopback(uint8_t port, uint8_t on)
> +rte_pmd_dpaa_set_tx_loopback(uint16_t port, uint8_t on)
> {
> struct rte_eth_dev *dev;
>
> diff --git a/drivers/net/dpaa/rte_pmd_dpaa.h
> b/drivers/net/dpaa/rte_pmd_dpaa.h index 8d244bb491..ec45633ba2 100644
> --- a/drivers/net/dpaa/rte_pmd_dpaa.h
> +++ b/drivers/net/dpaa/rte_pmd_dpaa.h
> @@ -29,6 +29,6 @@
> * - (-EINVAL) if bad parameter.
> */
> int
> -rte_pmd_dpaa_set_tx_loopback(uint8_t port, uint8_t on);
> +rte_pmd_dpaa_set_tx_loopback(uint16_t port, uint8_t on);
>
> #endif /* _PMD_DPAA_H_ */
> --
> 2.25.4
[dpdk-dev] [PATCH v3 2/3] examples/multi_process: cleanup bus objects while terminating app
From: Rohit Raj
Certain bus objects may need to be closed and re-acquired
while terminating and rerunning the client application.
Hence a signal handler is required to catch the termination
of the App and hence closing the bus objects.
This patch adds the missing signal handler in the client
app and closes the Bus objects in both client and server
applications when the signal Handler is called.
Signed-off-by: Rohit Raj
---
.../multi_process/client_server_mp/mp_client/client.c | 11 +++
.../multi_process/client_server_mp/mp_server/main.c | 4 +++-
2 files changed, 14 insertions(+), 1 deletion(-)
diff --git a/examples/multi_process/client_server_mp/mp_client/client.c
b/examples/multi_process/client_server_mp/mp_client/client.c
index 361d90b54..c37516b4c 100644
--- a/examples/multi_process/client_server_mp/mp_client/client.c
+++ b/examples/multi_process/client_server_mp/mp_client/client.c
@@ -11,6 +11,7 @@
#include
#include
#include
+#include
#include
#include
@@ -196,6 +197,14 @@ handle_packet(struct rte_mbuf *buf)
}
+static void
+signal_handler(int signal)
+{
+ if (signal == SIGINT)
+ rte_eal_cleanup();
+ exit(0);
+}
+
/*
* Application main function - loops through
* receiving and processing packets. Never returns
@@ -217,6 +226,8 @@ main(int argc, char *argv[])
argc -= retval;
argv += retval;
+ signal(SIGINT, signal_handler);
+
if (parse_app_args(argc, argv) < 0)
rte_exit(EXIT_FAILURE, "Invalid command-line arguments\n");
diff --git a/examples/multi_process/client_server_mp/mp_server/main.c
b/examples/multi_process/client_server_mp/mp_server/main.c
index 280dab867..b0241cc20 100644
--- a/examples/multi_process/client_server_mp/mp_server/main.c
+++ b/examples/multi_process/client_server_mp/mp_server/main.c
@@ -275,11 +275,13 @@ signal_handler(int signal)
{
uint16_t port_id;
- if (signal == SIGINT)
+ if (signal == SIGINT) {
RTE_ETH_FOREACH_DEV(port_id) {
rte_eth_dev_stop(port_id);
rte_eth_dev_close(port_id);
}
+ rte_eal_cleanup();
+ }
exit(0);
}
--
2.17.1
[dpdk-dev] [PATCH v3 3/3] bus/fslmc: support bus close API
From: Rohit Raj
This patch add support for closing the bus objects which
were acquired In the bus probe.
Some devices need to be cleaned while in both primary and
secondary process and while some need to be cleaned only in
case of primary process.
The devices are closed as per the white list used while
creating the objects in a particular process.
Signed-off-by: Rohit Raj
Acked-by: Sachin Saxena
---
drivers/bus/fslmc/fslmc_bus.c | 15 +++-
drivers/bus/fslmc/fslmc_vfio.c | 89 -
drivers/bus/fslmc/fslmc_vfio.h | 3 +-
drivers/bus/fslmc/mc/fsl_dpcon.h| 1 +
drivers/bus/fslmc/portal/dpaa2_hw_dpbp.c| 31 ++-
drivers/bus/fslmc/portal/dpaa2_hw_dpci.c| 32 +++-
drivers/bus/fslmc/portal/dpaa2_hw_dpio.c| 34
drivers/bus/fslmc/rte_bus_fslmc_version.map | 1 +
drivers/bus/fslmc/rte_fslmc.h | 5 +-
drivers/event/dpaa2/dpaa2_hw_dpcon.c| 32 +++-
drivers/net/dpaa2/dpaa2_mux.c | 18 -
11 files changed, 253 insertions(+), 8 deletions(-)
diff --git a/drivers/bus/fslmc/fslmc_bus.c b/drivers/bus/fslmc/fslmc_bus.c
index beb3dd008..9d3da6be3 100644
--- a/drivers/bus/fslmc/fslmc_bus.c
+++ b/drivers/bus/fslmc/fslmc_bus.c
@@ -1,6 +1,6 @@
/* SPDX-License-Identifier: BSD-3-Clause
*
- * Copyright 2016,2018-2019 NXP
+ * Copyright 2016,2018-2020 NXP
*
*/
@@ -365,6 +365,18 @@ rte_fslmc_match(struct rte_dpaa2_driver *dpaa2_drv,
return 1;
}
+static int
+rte_fslmc_close(void)
+{
+ int ret = 0;
+
+ ret = fslmc_vfio_close_group();
+ if (ret)
+ DPAA2_BUS_ERR("Unable to close devices %d", ret);
+
+ return 0;
+}
+
static int
rte_fslmc_probe(void)
{
@@ -637,6 +649,7 @@ struct rte_fslmc_bus rte_fslmc_bus = {
.bus = {
.scan = rte_fslmc_scan,
.probe = rte_fslmc_probe,
+ .close = rte_fslmc_close,
.parse = rte_fslmc_parse,
.find_device = rte_fslmc_find_device,
.get_iommu_class = rte_dpaa2_get_iommu_class,
diff --git a/drivers/bus/fslmc/fslmc_vfio.c b/drivers/bus/fslmc/fslmc_vfio.c
index 9134ec552..34ccc9db3 100644
--- a/drivers/bus/fslmc/fslmc_vfio.c
+++ b/drivers/bus/fslmc/fslmc_vfio.c
@@ -1,7 +1,7 @@
/* SPDX-License-Identifier: BSD-3-Clause
*
* Copyright (c) 2015-2016 Freescale Semiconductor, Inc. All rights reserved.
- * Copyright 2016-2019 NXP
+ * Copyright 2016-2020 NXP
*
*/
@@ -692,6 +692,54 @@ rte_dpaa2_vfio_setup_intr(struct rte_intr_handle
*intr_handle,
return -1;
}
+static void
+fslmc_close_iodevices(struct rte_dpaa2_device *dev)
+{
+ struct rte_dpaa2_object *object = NULL;
+ struct rte_dpaa2_driver *drv;
+ int ret, probe_all;
+
+ switch (dev->dev_type) {
+ case DPAA2_IO:
+ case DPAA2_CON:
+ case DPAA2_CI:
+ case DPAA2_BPOOL:
+ case DPAA2_MUX:
+ TAILQ_FOREACH(object, &dpaa2_obj_list, next) {
+ if (dev->dev_type == object->dev_type)
+ object->close(dev->object_id);
+ else
+ continue;
+ }
+ break;
+ case DPAA2_ETH:
+ case DPAA2_CRYPTO:
+ case DPAA2_QDMA:
+ probe_all = rte_fslmc_bus.bus.conf.scan_mode !=
+ RTE_BUS_SCAN_WHITELIST;
+ TAILQ_FOREACH(drv, &rte_fslmc_bus.driver_list, next) {
+ if (drv->drv_type != dev->dev_type)
+ continue;
+ if (rte_dev_is_probed(&dev->device))
+ continue;
+ if (probe_all ||
+ (dev->device.devargs &&
+dev->device.devargs->policy ==
+RTE_DEV_WHITELISTED)) {
+ ret = drv->remove(dev);
+ if (ret)
+ DPAA2_BUS_ERR("Unable to remove");
+ }
+ }
+ break;
+ default:
+ break;
+ }
+
+ DPAA2_BUS_LOG(DEBUG, "Device (%s) Closed",
+ dev->device.name);
+}
+
/*
* fslmc_process_iodevices for processing only IO (ETH, CRYPTO, and possibly
* EVENT) devices.
@@ -796,6 +844,45 @@ fslmc_process_mcp(struct rte_dpaa2_device *dev)
return ret;
}
+int
+fslmc_vfio_close_group(void)
+{
+ struct rte_dpaa2_device *dev, *dev_temp;
+
+ TAILQ_FOREACH_SAFE(dev, &rte_fslmc_bus.device_list, next, dev_temp) {
+ if (dev->device.devargs &&
+ dev->device.devargs->policy == RTE_DEV_BLACKLISTED) {
+ DPAA2_BUS_LOG(DEBUG, "%s Blacklisted, skipping",
+ dev->device.name);
+ TAILQ_REMOVE(&rte_fslmc_bus.device_list
[dpdk-dev] [PATCH v3 1/3] eal: add API for bus close
From: Rohit Raj
As per the current code we have API for bus probe, but the
bus close API is missing. This breaks the multi process
scenarios as objects are not cleaned while terminating the
secondary processes.
This patch adds a new API rte_bus_close() for cleanup of
bus objects which were acquired during probe.
Signed-off-by: Rohit Raj
---
v3:
* nit: combined nested if statements
v2:
* Moved rte_bus_close call to rte_eal_cleanup path.
lib/librte_eal/common/eal_common_bus.c | 32 +-
lib/librte_eal/include/rte_bus.h | 25 +++-
lib/librte_eal/linux/eal.c | 1 +
lib/librte_eal/rte_eal_version.map | 3 +++
4 files changed, 59 insertions(+), 2 deletions(-)
diff --git a/lib/librte_eal/common/eal_common_bus.c
b/lib/librte_eal/common/eal_common_bus.c
index baa5b532a..5fd7cf6c5 100644
--- a/lib/librte_eal/common/eal_common_bus.c
+++ b/lib/librte_eal/common/eal_common_bus.c
@@ -1,5 +1,5 @@
/* SPDX-License-Identifier: BSD-3-Clause
- * Copyright 2016 NXP
+ * Copyright 2016,2020 NXP
*/
#include
@@ -56,6 +56,36 @@ rte_bus_scan(void)
return 0;
}
+int
+rte_bus_close(void)
+{
+ int ret;
+ struct rte_bus *bus, *vbus = NULL;
+
+ TAILQ_FOREACH(bus, &rte_bus_list, next) {
+ if (!strcmp(bus->name, "vdev")) {
+ vbus = bus;
+ continue;
+ }
+
+ if (bus->close) {
+ ret = bus->close();
+ if (ret)
+ RTE_LOG(ERR, EAL, "Bus (%s) close failed.\n",
+ bus->name);
+ }
+ }
+
+ if (vbus && vbus->close) {
+ ret = vbus->close();
+ if (ret)
+ RTE_LOG(ERR, EAL, "Bus (%s) close failed.\n",
+ vbus->name);
+ }
+
+ return 0;
+}
+
/* Probe all devices of all buses */
int
rte_bus_probe(void)
diff --git a/lib/librte_eal/include/rte_bus.h b/lib/librte_eal/include/rte_bus.h
index d3034d0ed..af4787b18 100644
--- a/lib/librte_eal/include/rte_bus.h
+++ b/lib/librte_eal/include/rte_bus.h
@@ -1,5 +1,5 @@
/* SPDX-License-Identifier: BSD-3-Clause
- * Copyright 2016 NXP
+ * Copyright 2016,2020 NXP
*/
#ifndef _RTE_BUS_H_
@@ -67,6 +67,18 @@ typedef int (*rte_bus_scan_t)(void);
*/
typedef int (*rte_bus_probe_t)(void);
+/**
+ * Implementation specific close function which is responsible for closing
+ * devices on that bus.
+ *
+ * This is called while iterating over each registered bus.
+ *
+ * @return
+ * 0 for successful close
+ * !0 for any error while closing
+ */
+typedef int (*rte_bus_close_t)(void);
+
/**
* Device iterator to find a device on a bus.
*
@@ -248,6 +260,7 @@ struct rte_bus {
const char *name;/**< Name of the bus */
rte_bus_scan_t scan; /**< Scan for devices attached to bus */
rte_bus_probe_t probe; /**< Probe devices on bus */
+ rte_bus_close_t close; /**< Close devices on bus */
rte_bus_find_device_t find_device; /**< Find a device on the bus */
rte_bus_plug_t plug; /**< Probe single device for drivers */
rte_bus_unplug_t unplug; /**< Remove single device from driver */
@@ -301,6 +314,16 @@ int rte_bus_scan(void);
*/
int rte_bus_probe(void);
+/**
+ * For each device on the buses, call the device specific close.
+ *
+ * @return
+ * 0 for successful close
+ * !0 otherwise
+ */
+__rte_experimental
+int rte_bus_close(void);
+
/**
* Dump information of all the buses registered with EAL.
*
diff --git a/lib/librte_eal/linux/eal.c b/lib/librte_eal/linux/eal.c
index 0960f01d0..a8a63d8ca 100644
--- a/lib/librte_eal/linux/eal.c
+++ b/lib/librte_eal/linux/eal.c
@@ -1357,6 +1357,7 @@ rte_eal_cleanup(void)
if (rte_eal_process_type() == RTE_PROC_PRIMARY)
rte_memseg_walk(mark_freeable, NULL);
+ rte_bus_close();
rte_service_finalize();
rte_mp_channel_cleanup();
rte_trace_save();
diff --git a/lib/librte_eal/rte_eal_version.map
b/lib/librte_eal/rte_eal_version.map
index 0b18e2ef8..dc5266f22 100644
--- a/lib/librte_eal/rte_eal_version.map
+++ b/lib/librte_eal/rte_eal_version.map
@@ -397,6 +397,9 @@ EXPERIMENTAL {
rte_mp_disable;
rte_thread_register;
rte_thread_unregister;
+
+ # added in 20.11
+ rte_bus_close;
};
INTERNAL {
--
2.17.1
Re: [dpdk-dev] [PATCH] ci: enable vm env for aarch64 builds
> -Original Message- > From: Aaron Conole > Sent: Tuesday, August 25, 2020 4:18 PM > To: Juraj Linkeš > Cc: tho...@monjalon.net; david.march...@redhat.com; > maicolgabr...@hotmail.com; dev@dpdk.org; Kevin Laatz > > Subject: Re: [PATCH] ci: enable vm env for aarch64 builds > > Juraj Linkeš writes: > > > Tests requiring hugepages do not work outside of VM environment > > because of security limitations. Enable aarch64 builds which run tests > > to run in a VM to avoid these limitations. > > > > Signed-off-by: Juraj Linkeš > > --- > > Cool! > > Globally we have dist: set to 'bionic'. Does it make sense to upgrade all the > distributions to 'focal' ? What is the rationale for not using 'bionic'? > I tried bionic, but it ran with focal anyway, so the rationale is 'not supported for arm vm environments'. This was released for early adopters, so they may expand the support to bionic later. > > .travis.yml | 10 -- > > 1 file changed, 8 insertions(+), 2 deletions(-) > > > > diff --git a/.travis.yml b/.travis.yml index d6eeab371..d73d3e4c5 > > 100644 > > --- a/.travis.yml > > +++ b/.travis.yml > > @@ -88,7 +88,10 @@ jobs: > > arch: arm64 > > compiler: gcc > >- env: DEF_LIB="shared" RUN_TESTS=1 > > -arch: arm64 > > +dist: focal > > +arch: arm64-graviton2 > > +virt: vm > > +group: edge > > compiler: gcc > >- env: DEF_LIB="shared" BUILD_DOCS=1 > > arch: arm64 > > @@ -103,5 +106,8 @@ jobs: > > arch: arm64 > > compiler: clang > >- env: DEF_LIB="shared" RUN_TESTS=1 > > -arch: arm64 > > +dist: focal > > +arch: arm64-graviton2 > > +virt: vm > > +group: edge > > compiler: clang >
Re: [dpdk-dev] [PATCH v8 4/5] ci: add aarch64 -> aarch32 cross compiling jobs
> -Original Message- > From: Aaron Conole > Sent: Tuesday, August 25, 2020 4:16 PM > To: Juraj Linkeš > Cc: bruce.richard...@intel.com; maicolgabr...@hotmail.com; dev@dpdk.org > Subject: Re: [PATCH v8 4/5] ci: add aarch64 -> aarch32 cross compiling jobs > > Juraj Linkeš writes: > > > Add two jobs (static and shared libs), both building on aarch64 and > > producing 32-bit arm binaries executable on armv8-a, but not armv7. > > Do not run tests in these jobs. > > > > Signed-off-by: Juraj Linkeš > > --- > > .ci/linux-build.sh | 7 ++- > > .travis.yml| 19 +++ > > 2 files changed, 25 insertions(+), 1 deletion(-) > > > > diff --git a/.ci/linux-build.sh b/.ci/linux-build.sh index > > d079801d7..c042b065c 100755 > > --- a/.ci/linux-build.sh > > +++ b/.ci/linux-build.sh > > @@ -33,6 +33,11 @@ if [ "$AARCH64" = "1" ]; then > > OPTS="$OPTS --cross-file config/arm/arm64_armv8_linux_gcc" > > fi > > > > +if [ "$AARCH32" = "1" ]; then > > +# convert the arch specifier > > +OPTS="$OPTS --cross-file config/arm/arm32_armv8a_linux_gcc" > > +fi > > + > > if [ "$BUILD_DOCS" = "1" ]; then > > OPTS="$OPTS -Denable_docs=true" > > fi > > @@ -53,7 +58,7 @@ OPTS="$OPTS --buildtype=debugoptimized" > > meson build --werror $OPTS > > ninja -C build > > > > -if [ "$AARCH64" != "1" ]; then > > +if file build/app/dpdk-testpmd | sed 's/-/_/g' | grep -q "$(uname > > +-m)"; then > > I'm not sure about this. This will degrade the 32-bit x86 build, since it > will no > longer execute the test-null.sh script. > > Maybe we should re-use that flag for this series. BUILD_32BIT ?? > So this omits running test-null.sh with 32-bit builds on 64bit systems and we want to run that? Then it's probably the easiest to just disable the check for x86_64 -> aarch64 and x86_64 -> aarch32 builds. I wanted something more elegant here but it's a bit more complicated. > > devtools/test-null.sh > > fi > > > > diff --git a/.travis.yml b/.travis.yml index d6eeab371..850faaac1 > > 100644 > > --- a/.travis.yml > > +++ b/.travis.yml > > @@ -18,6 +18,10 @@ _aarch64_packages: &aarch64_packages > >- *required_packages > >- [gcc-aarch64-linux-gnu, libc6-dev-arm64-cross, > > pkg-config-aarch64-linux-gnu] > > > > +_arm_32b_packages: &arm_32b_packages > > + - *required_packages > > + - [gcc-arm-linux-gnueabihf, libc6-dev-armhf-cross, > > +pkg-config-arm-linux-gnueabihf] > > + > > _build_32b_packages: &build_32b_packages > >- *required_packages > >- [gcc-multilib] > > @@ -98,6 +102,21 @@ jobs: > > packages: > >- *required_packages > >- *doc_packages > > + # aarch64 cross-compiling aarch32 jobs > > + - env: DEF_LIB="shared" AARCH32=1 > > +arch: arm64 > > +compiler: gcc > > +addons: > > + apt: > > +packages: > > + - *arm_32b_packages > > + - env: DEF_LIB="static" AARCH32=1 > > +arch: arm64 > > +compiler: gcc > > +addons: > > + apt: > > +packages: > > + - *arm_32b_packages > ># aarch64 clang jobs > >- env: DEF_LIB="static" > > arch: arm64 >
Re: [dpdk-dev] [PATCH 1/1] vhost: fix iotlb mempool single-consumer flag
Hi Eugenio, > -Original Message- > From: Eugenio Pérez > Sent: Monday, August 10, 2020 10:11 PM > To: dev@dpdk.org > Cc: Adrian Moreno Zapata ; Maxime Coquelin > ; sta...@dpdk.org; Wang, Zhihong > ; Xia, Chenbo > Subject: [PATCH 1/1] vhost: fix iotlb mempool single-consumer flag > > Bugzilla bug: 523 > > Using testpmd as a vhost-user with iommu: > > /home/dpdk/build/app/dpdk-testpmd -l 1,3 \ > --vdev net_vhost0,iface=/tmp/vhost-user1,queues=1,iommu-support=1 > \ > -- --auto-start --stats-period 5 --forward-mode=txonly > > And qemu with packed virtqueue: > > > > > > >function='0x0'/> > > ... > > > > > > The fix looks fine to me. But the commit message is a little bit complicated to me (also, some lines too long). Since this bug is clear and could be described by something like 'control thread which handles iotlb msg and forwarding thread which uses iotlb to translate address may modify same entry of mempool and may cause a loop in iotlb_pending_entries list'. Do you think it makes sense? Thanks for the fix! Chenbo > -- > > Is it possible to consume the iotlb's entries of the mempoo from different > threads. Thread sanitizer example output (after change rwlocks to POSIX > ones): > > WARNING: ThreadSanitizer: data race (pid=76927) > Write of size 8 at 0x00017ffd5628 by thread T5: > #0 vhost_user_iotlb_cache_insert ../lib/librte_vhost/iotlb.c:181 > (dpdk-testpmd+0x769343) > #1 vhost_user_iotlb_msg ../lib/librte_vhost/vhost_user.c:2380 (dpdk- > testpmd+0x78e4bf) > #2 vhost_user_msg_handler ../lib/librte_vhost/vhost_user.c:2848 (dpdk- > testpmd+0x78fcf8) > #3 vhost_user_read_cb ../lib/librte_vhost/socket.c:311 (dpdk- > testpmd+0x770162) > #4 fdset_event_dispatch ../lib/librte_vhost/fd_man.c:286 (dpdk- > testpmd+0x7591c2) > #5 ctrl_thread_init ../lib/librte_eal/common/eal_common_thread.c:193 > (dpdk-testpmd+0xa2890b) > #6 (libtsan.so.0+0x2a68d) > > Previous read of size 8 at 0x00017ffd5628 by thread T3: > #0 vhost_user_iotlb_cache_find ../lib/librte_vhost/iotlb.c:252 (dpdk- > testpmd+0x76ee96) > #1 __vhost_iova_to_vva ../lib/librte_vhost/vhost.c:42 (dpdk- > testpmd+0x77488c) > #2 vhost_iova_to_vva ../lib/librte_vhost/vhost.h:753 (dpdk- > testpmd+0x7abeb3) > #3 map_one_desc ../lib/librte_vhost/virtio_net.c:497 (dpdk- > testpmd+0x7abeb3) > #4 fill_vec_buf_packed ../lib/librte_vhost/virtio_net.c:751 (dpdk- > testpmd+0x7abeb3) > #5 vhost_enqueue_single_packed ../lib/librte_vhost/virtio_net.c:1170 > (dpdk-testpmd+0x7abeb3) > #6 virtio_dev_rx_single_packed ../lib/librte_vhost/virtio_net.c:1346 > (dpdk-testpmd+0x7abeb3) > #7 virtio_dev_rx_packed ../lib/librte_vhost/virtio_net.c:1384 (dpdk- > testpmd+0x7abeb3) > #8 virtio_dev_rx ../lib/librte_vhost/virtio_net.c:1435 (dpdk- > testpmd+0x7b0654) > #9 rte_vhost_enqueue_burst ../lib/librte_vhost/virtio_net.c:1465 > (dpdk-testpmd+0x7b0654) > #10 eth_vhost_tx ../drivers/net/vhost/rte_eth_vhost.c:470 (dpdk- > testpmd+0x1ddfbd8) > #11 rte_eth_tx_burst ../lib/librte_ethdev/rte_ethdev.h:4800 (dpdk- > testpmd+0x505fdb) > #12 pkt_burst_transmit ../app/test-pmd/txonly.c:365 (dpdk- > testpmd+0x5106ad) > #13 run_pkt_fwd_on_lcore ../app/test-pmd/testpmd.c:2080 (dpdk- > testpmd+0x4f8951) > #14 start_pkt_forward_on_core ../app/test-pmd/testpmd.c:2106 (dpdk- > testpmd+0x4f89d7) > #15 eal_thread_loop ../lib/librte_eal/linux/eal_thread.c:127 (dpdk- > testpmd+0xa5b20a) > #16 (libtsan.so.0+0x2a68d) > > Location is global '' at 0x (rtemap_0+0x3ffd5628) > > Thread T5 'vhost-events' (tid=76933, running) created by main thread at: > #0 pthread_create (libtsan.so.0+0x2cd42) > #1 > rte_ctrl_thread_create ../lib/librte_eal/common/eal_common_thread.c:216 > (dpdk-testpmd+0xa289e7) > #2 rte_vhost_driver_start ../lib/librte_vhost/socket.c:1190 (dpdk- > testpmd+0x7728ef) > #3 vhost_driver_setup ../drivers/net/vhost/rte_eth_vhost.c:1028 (dpdk- > testpmd+0x1de233d) > #4 eth_dev_configure ../drivers/net/vhost/rte_eth_vhost.c:1126 (dpdk- > testpmd+0x1de29cc) > #5 rte_eth_dev_configure ../lib/librte_ethdev/rte_ethdev.c:1439 (dpdk- > testpmd+0x991ce2) > #6 start_port ../app/test-pmd/testpmd.c:2450 (dpdk-testpmd+0x4f9b45) > #7 main ../app/test-pmd/testpmd.c:3777 (dpdk-testpmd+0x4fe1ac) > > Thread T3 'lcore-slave-3' (tid=76931, running) created by main thread at: > #0 pthread_create (libtsan.so.0+0x2cd42) > #1 rte_eal_init ../lib/librte_eal/linux/eal.c:1244 (dpdk- > testpmd+0xa46e2b) > #2 main ../app/test-pmd/testpmd.c:3673 (dpdk-testpmd+0x4fdd75) > > -- > > Or: > WARNING: ThreadSanitizer: data race (pid=76927) > Write of size 1 at 0x00017ffd00f8 by thread T5: > #0 vhost_user_iotlb_cache_insert ../lib/librte_vhost/iotlb.c:182 > (dpdk-testpmd+0x769370) > #1 vhost_user_iotlb_msg ../lib/librte_vhost/vhost_user.c:2380 (dpdk- > testpm
