Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH] scsi: hisi_sas: Switch from msleep to usleep_range per
 timers-howto.txt
To:     Ding Xiang <dingxiang@cmss.chinamobile.com>,
        <jejb@linux.vnet.ibm.com>, <martin.petersen@oracle.com>,
        <linux-scsi@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
        Linuxarm <linuxarm@huawei.com>,
        chenxiang <chenxiang66@hisilicon.com>
References: <1535022539-10634-1-git-send-email-dingxiang@cmss.chinamobile.com>
From:   John Garry <john.garry@huawei.com>
Message-ID: <a89bf6bd-9789-5e6e-40ba-7e82780422d2@huawei.com>
Date:   Thu, 23 Aug 2018 18:01:10 +0100
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:45.0) Gecko/20100101
 Thunderbird/45.3.0
MIME-Version: 1.0
In-Reply-To: <1535022539-10634-1-git-send-email-dingxiang@cmss.chinamobile.com>
Content-Type: text/plain; charset="windows-1252"; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 23/08/2018 12:08, Ding Xiang wrote:
> use usleep_range to sleep for 10us - 20ms
>

Hi,

Thanks for the patch.

According to the document, msleep(1) will be ~20ms. As such, some driver 
behaviour may depend on this now.

At this stage, the drivers have been extensively tested (v2+v3), so I 
would rather not introduce possible instability. In addition, v1 is all 
but dead, so I can't test. Finally, none of the sleeps are in the 
fastpath, so the extra sleep does not cause much harm.

John

> Signed-off-by: Ding Xiang <dingxiang@cmss.chinamobile.com>
> ---
>  drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 8 ++++----
>  drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 6 +++---
>  drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 2 +-
>  3 files changed, 8 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
> index 8f60f0e..0584f711 100644
> --- a/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
> +++ b/drivers/scsi/hisi_sas/hisi_sas_v1_hw.c
> @@ -579,7 +579,7 @@ static int reset_hw_v1_hw(struct hisi_hba *hisi_hba)
>  		phy_ctrl |= PHY_CTRL_RESET_MSK;
>  		hisi_sas_phy_write32(hisi_hba, i, PHY_CTRL, phy_ctrl);
>  	}
> -	msleep(1); /* It is safe to wait for 50us */
> +	usleep_range(1000, 1100); /* It is safe to wait for 50us */
>
>  	/* Ensure DMA tx & rx idle */
>  	for (i = 0; i < hisi_hba->n_phy; i++) {
> @@ -632,7 +632,7 @@ static int reset_hw_v1_hw(struct hisi_hba *hisi_hba)
>  			     RESET_VALUE);
>  		regmap_write(hisi_hba->ctrl, hisi_hba->ctrl_clock_ena_reg + 4,
>  			     RESET_VALUE);
> -		msleep(1);
> +		usleep_range(1000, 1100);
>  		regmap_read(hisi_hba->ctrl, hisi_hba->ctrl_reset_sts_reg, &val);
>  		if (RESET_VALUE != (val & RESET_VALUE)) {
>  			dev_err(dev, "Reset failed\n");
> @@ -645,7 +645,7 @@ static int reset_hw_v1_hw(struct hisi_hba *hisi_hba)
>  			     RESET_VALUE);
>  		regmap_write(hisi_hba->ctrl, hisi_hba->ctrl_clock_ena_reg,
>  			     RESET_VALUE);
> -		msleep(1);
> +		usleep_range(1000, 1100);
>  		regmap_read(hisi_hba->ctrl, hisi_hba->ctrl_reset_sts_reg, &val);
>  		if (val & RESET_VALUE) {
>  			dev_err(dev, "De-reset failed\n");
> @@ -841,7 +841,7 @@ static void sl_notify_v1_hw(struct hisi_hba *hisi_hba, int phy_no)
>  	sl_control = hisi_sas_phy_read32(hisi_hba, phy_no, SL_CONTROL);
>  	sl_control |= SL_CONTROL_NOTIFY_EN_MSK;
>  	hisi_sas_phy_write32(hisi_hba, phy_no, SL_CONTROL, sl_control);
> -	msleep(1);
> +	usleep_range(1000, 1100);
>  	sl_control = hisi_sas_phy_read32(hisi_hba, phy_no, SL_CONTROL);
>  	sl_control &= ~SL_CONTROL_NOTIFY_EN_MSK;
>  	hisi_sas_phy_write32(hisi_hba, phy_no, SL_CONTROL, sl_control);
> diff --git a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
> index 9c5c5a6..4ca5f77 100644
> --- a/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
> +++ b/drivers/scsi/hisi_sas/hisi_sas_v2_hw.c
> @@ -1081,7 +1081,7 @@ static int reset_hw_v2_hw(struct hisi_hba *hisi_hba)
>  				reset_val);
>  		regmap_write(hisi_hba->ctrl, hisi_hba->ctrl_clock_ena_reg + 4,
>  				reset_val);
> -		msleep(1);
> +		usleep_range(1000, 1100);
>  		regmap_read(hisi_hba->ctrl, hisi_hba->ctrl_reset_sts_reg, &val);
>  		if (reset_val != (val & reset_val)) {
>  			dev_err(dev, "SAS reset fail.\n");
> @@ -1093,7 +1093,7 @@ static int reset_hw_v2_hw(struct hisi_hba *hisi_hba)
>  				reset_val);
>  		regmap_write(hisi_hba->ctrl, hisi_hba->ctrl_clock_ena_reg,
>  				reset_val);
> -		msleep(1);
> +		usleep_range(1000, 1100);
>  		regmap_read(hisi_hba->ctrl, hisi_hba->ctrl_reset_sts_reg,
>  				&val);
>  		if (val & reset_val) {
> @@ -1591,7 +1591,7 @@ static void sl_notify_v2_hw(struct hisi_hba *hisi_hba, int phy_no)
>  	sl_control = hisi_sas_phy_read32(hisi_hba, phy_no, SL_CONTROL);
>  	sl_control |= SL_CONTROL_NOTIFY_EN_MSK;
>  	hisi_sas_phy_write32(hisi_hba, phy_no, SL_CONTROL, sl_control);
> -	msleep(1);
> +	usleep_range(1000, 1100);
>  	sl_control = hisi_sas_phy_read32(hisi_hba, phy_no, SL_CONTROL);
>  	sl_control &= ~SL_CONTROL_NOTIFY_EN_MSK;
>  	hisi_sas_phy_write32(hisi_hba, phy_no, SL_CONTROL, sl_control);
> diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
> index 08b503e2..8f57737 100644
> --- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
> +++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
> @@ -834,7 +834,7 @@ static void sl_notify_v3_hw(struct hisi_hba *hisi_hba, int phy_no)
>  	sl_control = hisi_sas_phy_read32(hisi_hba, phy_no, SL_CONTROL);
>  	sl_control |= SL_CONTROL_NOTIFY_EN_MSK;
>  	hisi_sas_phy_write32(hisi_hba, phy_no, SL_CONTROL, sl_control);
> -	msleep(1);
> +	usleep_range(1000, 1100);
>  	sl_control = hisi_sas_phy_read32(hisi_hba, phy_no, SL_CONTROL);
>  	sl_control &= ~SL_CONTROL_NOTIFY_EN_MSK;
>  	hisi_sas_phy_write32(hisi_hba, phy_no, SL_CONTROL, sl_control);
>