From: "Jernej Škrabec" <jernej.skrabec@gmail.com>
To: linux-sunxi@lists.linux.dev, linux-kernel@vger.kernel.org,
qianfanguijin@163.com
Cc: Maxime Ripard <mripard@kernel.org>, Chen-Yu Tsai <wens@csie.org>,
linux-arm-kernel@lists.infradead.org,
Evgeny Boger <boger@wirenboard.com>,
qianfan Zhao <qianfanguijin@163.com>
Subject: Re: [PATCH 1/2] drivers: net: mdio-sun4i: Speedup mdio read and write
Date: Tue, 06 Sep 2022 21:52:24 +0200 [thread overview]
Message-ID: <4737709.GXAFRqVoOG@kista> (raw)
In-Reply-To: <20220906075616.21347-1-qianfanguijin@163.com>
Dne torek, 06. september 2022 ob 09:56:15 CEST je qianfanguijin@163.com
napisal(a):
> From: qianfan Zhao <qianfanguijin@163.com>
>
> msleep(1) on my board takes about 30ms, and it is too long to accept.
> Use read_poll_timeout to speedup.
>
> Signed-off-by: qianfan Zhao <qianfanguijin@163.com>
> ---
> drivers/net/mdio/mdio-sun4i.c | 29 ++++++++++++-----------------
> 1 file changed, 12 insertions(+), 17 deletions(-)
>
> diff --git a/drivers/net/mdio/mdio-sun4i.c b/drivers/net/mdio/mdio-sun4i.c
> index f798de3276dc..168e2a375535 100644
> --- a/drivers/net/mdio/mdio-sun4i.c
> +++ b/drivers/net/mdio/mdio-sun4i.c
> @@ -26,8 +26,6 @@
> #define EMAC_MAC_MIND_REG (0x10)
> #define EMAC_MAC_SSRR_REG (0x14)
>
> -#define MDIO_TIMEOUT (msecs_to_jiffies(100))
> -
> struct sun4i_mdio_data {
> void __iomem *membase;
> struct regulator *regulator;
> @@ -36,8 +34,7 @@ struct sun4i_mdio_data {
> static int sun4i_mdio_read(struct mii_bus *bus, int mii_id, int regnum)
> {
> struct sun4i_mdio_data *data = bus->priv;
> - unsigned long timeout_jiffies;
> - int value;
> + int ret, tmp, value;
>
> /* issue the phy address and reg */
> writel((mii_id << 8) | regnum, data->membase + EMAC_MAC_MADR_REG);
> @@ -45,12 +42,11 @@ static int sun4i_mdio_read(struct mii_bus *bus, int
> mii_id, int regnum) writel(0x1, data->membase + EMAC_MAC_MCMD_REG);
>
> /* Wait read complete */
> - timeout_jiffies = jiffies + MDIO_TIMEOUT;
> - while (readl(data->membase + EMAC_MAC_MIND_REG) & 0x1) {
> - if (time_is_before_jiffies(timeout_jiffies))
> - return -ETIMEDOUT;
> - msleep(1);
> - }
> + ret = read_poll_timeout(readl, tmp, (tmp & 1) == 0,
> + 20, 10000, false,
> + data->membase +
EMAC_MAC_MIND_REG);
You should use readl_poll_timeout() instead, as instructed by documentation.
Additionally, are you sure about both delay parameters? They are both much
smaller than in original code.
Same comments apply to patch 2.
Best regards,
Jernej
> + if (ret < 0)
> + return ret;
>
> /* push down the phy io line */
> writel(0x0, data->membase + EMAC_MAC_MCMD_REG);
> @@ -64,7 +60,7 @@ static int sun4i_mdio_write(struct mii_bus *bus, int
> mii_id, int regnum, u16 value)
> {
> struct sun4i_mdio_data *data = bus->priv;
> - unsigned long timeout_jiffies;
> + int ret, tmp;
>
> /* issue the phy address and reg */
> writel((mii_id << 8) | regnum, data->membase + EMAC_MAC_MADR_REG);
> @@ -72,12 +68,11 @@ static int sun4i_mdio_write(struct mii_bus *bus, int
> mii_id, int regnum, writel(0x1, data->membase + EMAC_MAC_MCMD_REG);
>
> /* Wait read complete */
> - timeout_jiffies = jiffies + MDIO_TIMEOUT;
> - while (readl(data->membase + EMAC_MAC_MIND_REG) & 0x1) {
> - if (time_is_before_jiffies(timeout_jiffies))
> - return -ETIMEDOUT;
> - msleep(1);
> - }
> + ret = read_poll_timeout(readl, tmp, (tmp & 1) == 0,
> + 20, 10000, false,
> + data->membase +
EMAC_MAC_MIND_REG);
> + if (ret < 0)
> + return ret;
>
> /* push down the phy io line */
> writel(0x0, data->membase + EMAC_MAC_MCMD_REG);
> --
> 2.25.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
prev parent reply other threads:[~2022-09-06 19:53 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-09-06 7:56 [PATCH 1/2] drivers: net: mdio-sun4i: Speedup mdio read and write qianfanguijin
2022-09-06 7:56 ` [PATCH 2/2] drivers: net: mdio-sun4i: Wait mdio write done qianfanguijin
2022-09-06 19:52 ` Jernej Škrabec [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=4737709.GXAFRqVoOG@kista \
--to=jernej.skrabec@gmail.com \
--cc=boger@wirenboard.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-sunxi@lists.linux.dev \
--cc=mripard@kernel.org \
--cc=qianfanguijin@163.com \
--cc=wens@csie.org \
/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).