From: Leon Romanovsky <leon@kernel.org>
To: liweihang <liweihang@huawei.com>
Cc: chenglang <chenglang@huawei.com>,
"dledford@redhat.com" <dledford@redhat.com>,
"jgg@nvidia.com" <jgg@nvidia.com>,
"linux-rdma@vger.kernel.org" <linux-rdma@vger.kernel.org>,
"linuxarm@openeuler.org" <linuxarm@openeuler.org>
Subject: Re: [PATCH for-next 2/6] RDMA/hns: Remove unsupported CMDQ mode
Date: Sun, 7 Feb 2021 10:53:59 +0200 [thread overview]
Message-ID: <20210207085359.GD4656@unreal> (raw)
In-Reply-To: <11e5e5424ad94e329c11e7a7a77f585e@huawei.com>
On Sun, Feb 07, 2021 at 08:15:09AM +0000, liweihang wrote:
> On 2021/2/7 15:52, chenglang wrote:
> >
> > On 2021/2/4 14:23, Weihang Li wrote:
> >> From: Lang Cheng <chenglang@huawei.com>
> >>
> >> HIP08/09 only supports CMDQ in non-interrupt mode, and the firmware always
> >> ignores the flag to indicate the mode. Therefore, remove the dead code.
> >>
> >> Fixes: a04ff739f2a9 ("RDMA/hns: Add command queue support for hip08 RoCE driver")
> >> Signed-off-by: Lang Cheng <chenglang@huawei.com>
> >> Signed-off-by: Weihang Li <liweihang@huawei.com>
> >> ---
> >> drivers/infiniband/hw/hns/hns_roce_hw_v2.c | 24 ++++++++----------------
> >> drivers/infiniband/hw/hns/hns_roce_hw_v2.h | 2 --
> >> 2 files changed, 8 insertions(+), 18 deletions(-)
> >>
> >> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> >> index 7a5a41d..260c17c 100644
> >> --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> >> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.c
> >> @@ -1197,8 +1197,7 @@ static void hns_roce_cmq_setup_basic_desc(struct hns_roce_cmq_desc *desc,
> >> {
> >> memset((void *)desc, 0, sizeof(struct hns_roce_cmq_desc));
> >> desc->opcode = cpu_to_le16(opcode);
> >> - desc->flag =
> >> - cpu_to_le16(HNS_ROCE_CMD_FLAG_NO_INTR | HNS_ROCE_CMD_FLAG_IN);
> >> + desc->flag = cpu_to_le16(HNS_ROCE_CMD_FLAG_IN);
> >> if (is_read)
> >> desc->flag |= cpu_to_le16(HNS_ROCE_CMD_FLAG_WR);
> >> else
> >> @@ -1275,18 +1274,12 @@ static int __hns_roce_cmq_send(struct hns_roce_dev *hr_dev,
> >> /* Write to hardware */
> >> roce_write(hr_dev, ROCEE_TX_CMQ_TAIL_REG, csq->next_to_use);
> >>
> >> - /*
> >> - * If the command is sync, wait for the firmware to write back,
> >> - * if multi descriptors to be sent, use the first one to check
> >> - */
> >> - if (le16_to_cpu(desc->flag) & HNS_ROCE_CMD_FLAG_NO_INTR) {
> >> - do {
> >> - if (hns_roce_cmq_csq_done(hr_dev))
> >> - break;
> >> - udelay(1);
> >> - timeout++;
> >> - } while (timeout < priv->cmq.tx_timeout);
> >> - }
> >> + do {
> >> + if (hns_roce_cmq_csq_done(hr_dev))
> >> + break;
> >> + udelay(1);
> >> + timeout++;
> >> + } while (timeout < priv->cmq.tx_timeout);
> >>
> >> if (hns_roce_cmq_csq_done(hr_dev)) {
> >> handle = 0;
> >> @@ -1626,8 +1619,7 @@ static int hns_roce_set_vf_switch_param(struct hns_roce_dev *hr_dev, int vf_id)
> >> if (ret)
> >> return ret;
> >>
> >> - desc.flag =
> >> - cpu_to_le16(HNS_ROCE_CMD_FLAG_NO_INTR | HNS_ROCE_CMD_FLAG_IN);
> >
> > The old firmware needs this redundant flag, it is best cleaned up after the
> > firmware version is released.
> >
> > Thanks.
>
> Got it, I will drop this one from the series.
And how will it work new kernel with old FW? Or isn't it possible and all
users must upgrade their kernel/FW at the same time?
Thanks
>
> Thanks
> Weihang
>
> >
> >> + desc.flag = cpu_to_le16(HNS_ROCE_CMD_FLAG_IN);
> >> desc.flag &= cpu_to_le16(~HNS_ROCE_CMD_FLAG_WR);
> >> roce_set_bit(swt->cfg, VF_SWITCH_DATA_CFG_ALW_LPBK_S, 1);
> >> roce_set_bit(swt->cfg, VF_SWITCH_DATA_CFG_ALW_LCL_LPBK_S, 0);
> >> diff --git a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
> >> index 9f97e32..986a287 100644
> >> --- a/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
> >> +++ b/drivers/infiniband/hw/hns/hns_roce_hw_v2.h
> >> @@ -128,14 +128,12 @@
> >> #define HNS_ROCE_CMD_FLAG_OUT_VALID_SHIFT 1
> >> #define HNS_ROCE_CMD_FLAG_NEXT_SHIFT 2
> >> #define HNS_ROCE_CMD_FLAG_WR_OR_RD_SHIFT 3
> >> -#define HNS_ROCE_CMD_FLAG_NO_INTR_SHIFT 4
> >> #define HNS_ROCE_CMD_FLAG_ERR_INTR_SHIFT 5
> >>
> >> #define HNS_ROCE_CMD_FLAG_IN BIT(HNS_ROCE_CMD_FLAG_IN_VALID_SHIFT)
> >> #define HNS_ROCE_CMD_FLAG_OUT BIT(HNS_ROCE_CMD_FLAG_OUT_VALID_SHIFT)
> >> #define HNS_ROCE_CMD_FLAG_NEXT BIT(HNS_ROCE_CMD_FLAG_NEXT_SHIFT)
> >> #define HNS_ROCE_CMD_FLAG_WR BIT(HNS_ROCE_CMD_FLAG_WR_OR_RD_SHIFT)
> >> -#define HNS_ROCE_CMD_FLAG_NO_INTR BIT(HNS_ROCE_CMD_FLAG_NO_INTR_SHIFT)
> >> #define HNS_ROCE_CMD_FLAG_ERR_INTR BIT(HNS_ROCE_CMD_FLAG_ERR_INTR_SHIFT)
> >>
> >> #define HNS_ROCE_CMQ_DESC_NUM_S 3
>
next prev parent reply other threads:[~2021-02-07 8:54 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-02-04 6:23 [PATCH for-next 0/6] RDMA/hns: Fix and refactor CMDQ related code Weihang Li
2021-02-04 6:23 ` [PATCH for-next 1/6] RDMA/hns: Remove unused member and variable of CMDQ Weihang Li
2021-02-04 6:23 ` [PATCH for-next 2/6] RDMA/hns: Remove unsupported CMDQ mode Weihang Li
[not found] ` <aabcf1a1-1cdf-5d05-cc11-daa36f9f10fa@huawei.com>
2021-02-07 8:15 ` liweihang
2021-02-07 8:53 ` Leon Romanovsky [this message]
2021-02-04 6:23 ` [PATCH for-next 3/6] RDMA/hns: Fixes missing error code of CMDQ Weihang Li
2021-02-04 6:23 ` [PATCH for-next 4/6] RDMA/hns: Remove redundant operations on CMDQ Weihang Li
2021-02-04 6:23 ` [PATCH for-next 5/6] RDMA/hns: Adjust fields and variables about CMDQ tail/head Weihang Li
2021-02-04 6:23 ` [PATCH for-next 6/6] RDMA/hns: Refactor process of posting CMDQ Weihang Li
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=20210207085359.GD4656@unreal \
--to=leon@kernel.org \
--cc=chenglang@huawei.com \
--cc=dledford@redhat.com \
--cc=jgg@nvidia.com \
--cc=linux-rdma@vger.kernel.org \
--cc=linuxarm@openeuler.org \
--cc=liweihang@huawei.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