From: Jerin Jacob <jerinjacobk@gmail.com>
To: Pavan Nikhilesh <pbhagavatula@marvell.com>
Cc: Jerin Jacob <jerinj@marvell.com>,
Nithin Dabilpuram <ndabilpuram@marvell.com>,
Kiran Kumar K <kirankumark@marvell.com>,
Sunil Kumar Kori <skori@marvell.com>,
Satha Rao <skoteshwar@marvell.com>, dpdk-dev <dev@dpdk.org>
Subject: Re: [PATCH] common/cnxk: add workaround for vWQE flush
Date: Mon, 24 Jan 2022 14:36:13 +0530 [thread overview]
Message-ID: <CALBAE1PLGdZUXc9okAQpKUQHJZ6qYCkoXx4x+e+r2Q-QGjcmLQ@mail.gmail.com> (raw)
In-Reply-To: <20211213205424.5588-1-pbhagavatula@marvell.com>
On Tue, Dec 14, 2021 at 2:24 AM <pbhagavatula@marvell.com> wrote:
>
> From: Pavan Nikhilesh <pbhagavatula@marvell.com>
>
> Due to an errata writing to vWQE flush register might hang NIX.
> Add workaround for vWQE flush hang by waiting for the max
> coalescing timeout to flush out any pending vWQEs.
Fixes: ee48f711f3b0 ("common/cnxk: support NIX inline inbound and
outbound setup")
Cc: stable@dpdk.org
>
> Signed-off-by: Pavan Nikhilesh <pbhagavatula@marvell.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
Applied to dpdk-next-net-eventdev/for-main. Thanks
> ---
> drivers/common/cnxk/roc_nix_inl.c | 3 +--
> drivers/common/cnxk/roc_nix_inl_dev.c | 12 ++++++++++++
> drivers/common/cnxk/roc_nix_inl_priv.h | 1 +
> drivers/common/cnxk/roc_nix_priv.h | 1 +
> drivers/common/cnxk/roc_nix_queue.c | 19 +++++++++++++++++--
> 5 files changed, 32 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/common/cnxk/roc_nix_inl.c b/drivers/common/cnxk/roc_nix_inl.c
> index f0fc690417..e8981c4aa4 100644
> --- a/drivers/common/cnxk/roc_nix_inl.c
> +++ b/drivers/common/cnxk/roc_nix_inl.c
> @@ -595,8 +595,7 @@ roc_nix_inl_dev_rq_put(struct roc_nix_rq *rq)
> plt_err("Failed to disable inline device rq, rc=%d", rc);
>
> /* Flush NIX LF for CN10K */
> - if (roc_model_is_cn10k())
> - plt_write64(0, inl_dev->nix_base + NIX_LF_OP_VWQE_FLUSH);
> + nix_rq_vwqe_flush(rq, inl_dev->vwqe_interval);
>
> return rc;
> }
> diff --git a/drivers/common/cnxk/roc_nix_inl_dev.c b/drivers/common/cnxk/roc_nix_inl_dev.c
> index a0fe6ecd82..10912a6c93 100644
> --- a/drivers/common/cnxk/roc_nix_inl_dev.c
> +++ b/drivers/common/cnxk/roc_nix_inl_dev.c
> @@ -346,6 +346,7 @@ nix_inl_nix_setup(struct nix_inl_dev *inl_dev)
> struct mbox *mbox = dev->mbox;
> struct nix_lf_alloc_rsp *rsp;
> struct nix_lf_alloc_req *req;
> + struct nix_hw_info *hw_info;
> size_t inb_sa_sz;
> int i, rc = -ENOSPC;
> void *sa;
> @@ -382,6 +383,17 @@ nix_inl_nix_setup(struct nix_inl_dev *inl_dev)
> inl_dev->qints = rsp->qints;
> inl_dev->cints = rsp->cints;
>
> + /* Get VWQE info if supported */
> + if (roc_model_is_cn10k()) {
> + mbox_alloc_msg_nix_get_hw_info(mbox);
> + rc = mbox_process_msg(mbox, (void *)&hw_info);
> + if (rc) {
> + plt_err("Failed to get HW info, rc=%d", rc);
> + goto lf_free;
> + }
> + inl_dev->vwqe_interval = hw_info->vwqe_delay;
> + }
> +
> /* Register nix interrupts */
> rc = nix_inl_nix_register_irqs(inl_dev);
> if (rc) {
> diff --git a/drivers/common/cnxk/roc_nix_inl_priv.h b/drivers/common/cnxk/roc_nix_inl_priv.h
> index 3dc526f929..be53a3fa81 100644
> --- a/drivers/common/cnxk/roc_nix_inl_priv.h
> +++ b/drivers/common/cnxk/roc_nix_inl_priv.h
> @@ -35,6 +35,7 @@ struct nix_inl_dev {
> /* NIX data */
> uint8_t lf_tx_stats;
> uint8_t lf_rx_stats;
> + uint16_t vwqe_interval;
> uint16_t cints;
> uint16_t qints;
> struct roc_nix_rq rq;
> diff --git a/drivers/common/cnxk/roc_nix_priv.h b/drivers/common/cnxk/roc_nix_priv.h
> index 04575af295..deb2a6ba11 100644
> --- a/drivers/common/cnxk/roc_nix_priv.h
> +++ b/drivers/common/cnxk/roc_nix_priv.h
> @@ -377,6 +377,7 @@ int nix_rq_cfg(struct dev *dev, struct roc_nix_rq *rq, uint16_t qints, bool cfg,
> int nix_rq_ena_dis(struct dev *dev, struct roc_nix_rq *rq, bool enable);
> int nix_tm_bp_config_get(struct roc_nix *roc_nix, bool *is_enabled);
> int nix_tm_bp_config_set(struct roc_nix *roc_nix, bool enable);
> +void nix_rq_vwqe_flush(struct roc_nix_rq *rq, uint16_t vwqe_interval);
>
> /*
> * TM priv utils.
> diff --git a/drivers/common/cnxk/roc_nix_queue.c b/drivers/common/cnxk/roc_nix_queue.c
> index c8c8401d81..d5f6813e69 100644
> --- a/drivers/common/cnxk/roc_nix_queue.c
> +++ b/drivers/common/cnxk/roc_nix_queue.c
> @@ -28,6 +28,22 @@ nix_qsize_clampup(uint32_t val)
> return i;
> }
>
> +void
> +nix_rq_vwqe_flush(struct roc_nix_rq *rq, uint16_t vwqe_interval)
> +{
> + uint64_t wait_ns;
> +
> + if (!roc_model_is_cn10k())
> + return;
> + /* Due to HW errata writes to VWQE_FLUSH might hang, so instead
> + * wait for max vwqe timeout interval.
> + */
> + if (rq->vwqe_ena) {
> + wait_ns = rq->vwqe_wait_tmo * (vwqe_interval + 1) * 100;
> + plt_delay_us((wait_ns / 1E3) + 1);
> + }
> +}
> +
> int
> nix_rq_ena_dis(struct dev *dev, struct roc_nix_rq *rq, bool enable)
> {
> @@ -66,9 +82,8 @@ roc_nix_rq_ena_dis(struct roc_nix_rq *rq, bool enable)
> int rc;
>
> rc = nix_rq_ena_dis(&nix->dev, rq, enable);
> + nix_rq_vwqe_flush(rq, nix->vwqe_interval);
>
> - if (roc_model_is_cn10k())
> - plt_write64(rq->qid, nix->base + NIX_LF_OP_VWQE_FLUSH);
> return rc;
> }
>
> --
> 2.17.1
>
prev parent reply other threads:[~2022-01-24 9:06 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-13 20:54 pbhagavatula
2022-01-24 9:06 ` Jerin Jacob [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CALBAE1PLGdZUXc9okAQpKUQHJZ6qYCkoXx4x+e+r2Q-QGjcmLQ@mail.gmail.com \
--to=jerinjacobk@gmail.com \
--cc=dev@dpdk.org \
--cc=jerinj@marvell.com \
--cc=kirankumark@marvell.com \
--cc=ndabilpuram@marvell.com \
--cc=pbhagavatula@marvell.com \
--cc=skori@marvell.com \
--cc=skoteshwar@marvell.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).