From: Guenter Roeck <linux@roeck-us.net>
To: Hauke Mehrtens <hauke@hauke-m.de>
Cc: wim@linux-watchdog.org, linux-watchdog@vger.kernel.org,
john@phrozen.org, dev@kresin.me
Subject: Re: [v2,1/3] wdt: lantiq: update register names to better match spec
Date: Mon, 3 Sep 2018 16:09:15 -0700 [thread overview]
Message-ID: <20180903230915.GA30065@roeck-us.net> (raw)
In-Reply-To: <20180902113233.29113-2-hauke@hauke-m.de>
On Sun, Sep 02, 2018 at 01:32:31PM +0200, Hauke Mehrtens wrote:
> Some of the names of the bits were confusing to me.
> Now the bits share the same prefix as the register they are set on.
>
Plus you changed the description. Overall I find the new defines
more confusing than the old ones, especially the changed meaning.
Can you point me to a reference manual ?
> Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
> ---
> drivers/watchdog/lantiq_wdt.c | 32 ++++++++++++++++----------------
> 1 file changed, 16 insertions(+), 16 deletions(-)
>
> diff --git a/drivers/watchdog/lantiq_wdt.c b/drivers/watchdog/lantiq_wdt.c
> index 7f43cefa0eae..e5937be09bd4 100644
> --- a/drivers/watchdog/lantiq_wdt.c
> +++ b/drivers/watchdog/lantiq_wdt.c
> @@ -13,6 +13,7 @@
> #include <linux/module.h>
> #include <linux/fs.h>
> #include <linux/miscdevice.h>
> +#include <linux/bitops.h>
> #include <linux/watchdog.h>
> #include <linux/of_platform.h>
> #include <linux/uaccess.h>
> @@ -40,18 +41,17 @@
> * essentially the following two magic passwords need to be written to allow
> * IO access to the WDT core
> */
> -#define LTQ_WDT_PW1 0x00BE0000
> -#define LTQ_WDT_PW2 0x00DC0000
> +#define LTQ_WDT_CR_PW1 0x00BE0000
> +#define LTQ_WDT_CR_PW2 0x00DC0000
>
> -#define LTQ_WDT_CR 0x0 /* watchdog control register */
> -#define LTQ_WDT_SR 0x8 /* watchdog status register */
> +#define LTQ_WDT_CR 0x0 /* watchdog control register */
> +#define LTQ_WDT_CR_GEN BIT(31) /* enable bit */
> +#define LTQ_WDT_CR_PWL (0x3 << 26) /* Pre-warning limit set to 1/16 of max WDT period */
Old was "turn on power"
> +#define LTQ_WDT_CR_CLKDIV (0x3 << 24) /* set clock divider to 0x40000 */
Old was: "turn on clock and set divider to 0x40000". Note that the comment
itself does not really add much value, except that it suggests that one
of the bits may _not_ really set a clock mask.
So this doesn't just rename the defines, it also changes the meaning.
Also, while some may feel differently, I do think we should stick with the
80-column limit as long as it exists and generates checkpatch warnings.
> +#define LTQ_WDT_CR_PW_MASK GENMASK(23, 16) /* Password field */
> +#define LTQ_WDT_CR_RELOAD_MASK GENMASK(15, 0) /* Reload value */
>
The mixed use of GENMASK() for some defines and shift for others is
confusing.
> -#define LTQ_WDT_SR_EN (0x1 << 31) /* enable bit */
> -#define LTQ_WDT_SR_PWD (0x3 << 26) /* turn on power */
> -#define LTQ_WDT_SR_CLKDIV (0x3 << 24) /* turn on clock and set */
> - /* divider to 0x40000 */
> #define LTQ_WDT_DIVIDER 0x40000
> -#define LTQ_MAX_TIMEOUT ((1 << 16) - 1) /* the reload field is 16 bit */
>
I question if the change from LTQ_MAX_TIMEOUT to LTQ_WDT_CR_RELOAD_MASK
really adds value. Sure, the maximum value must fit into a mask, but it
is still a maximum value.
> static bool nowayout = WATCHDOG_NOWAYOUT;
>
> @@ -68,26 +68,26 @@ ltq_wdt_enable(void)
> {
> unsigned long int timeout = ltq_wdt_timeout *
> (ltq_io_region_clk_rate / LTQ_WDT_DIVIDER) + 0x1000;
> - if (timeout > LTQ_MAX_TIMEOUT)
> - timeout = LTQ_MAX_TIMEOUT;
> + if (timeout > LTQ_WDT_CR_RELOAD_MASK)
> + timeout = LTQ_WDT_CR_RELOAD_MASK;
And this is less confusing than before ? Not to me. I find the new code
more confusing.
>
> /* write the first password magic */
> - ltq_w32(LTQ_WDT_PW1, ltq_wdt_membase + LTQ_WDT_CR);
> + ltq_w32(LTQ_WDT_CR_PW1, ltq_wdt_membase + LTQ_WDT_CR);
> /* write the second magic plus the configuration and new timeout */
> - ltq_w32(LTQ_WDT_SR_EN | LTQ_WDT_SR_PWD | LTQ_WDT_SR_CLKDIV | -
> LTQ_WDT_PW2 | timeout, ltq_wdt_membase + LTQ_WDT_CR); +
> ltq_w32(LTQ_WDT_CR_GEN | LTQ_WDT_CR_PWL | LTQ_WDT_CR_CLKDIV | +
> LTQ_WDT_CR_PW2 | timeout, ltq_wdt_membase + LTQ_WDT_CR); }
>
> static void ltq_wdt_disable(void) { /* write the first password magic */ -
> ltq_w32(LTQ_WDT_PW1, ltq_wdt_membase + LTQ_WDT_CR); +
> ltq_w32(LTQ_WDT_CR_PW1, ltq_wdt_membase + LTQ_WDT_CR); /* * write the second
> password magic with no config * this turns the watchdog off */ -
> ltq_w32(LTQ_WDT_PW2, ltq_wdt_membase + LTQ_WDT_CR); +
> ltq_w32(LTQ_WDT_CR_PW2, ltq_wdt_membase + LTQ_WDT_CR); }
>
> static ssize_t
next prev parent reply other threads:[~2018-09-04 3:31 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-09-02 11:32 [PATCH v2 0/3] wdt: lantiq: Convert to watchdog_device Hauke Mehrtens
2018-09-02 11:32 ` [PATCH v2 1/3] wdt: lantiq: update register names to better match spec Hauke Mehrtens
2018-09-03 23:09 ` Guenter Roeck [this message]
2018-09-09 21:04 ` [v2,1/3] " Hauke Mehrtens
2018-09-02 11:32 ` [PATCH v2 2/3] wdt: lantiq: Convert to watchdog_device Hauke Mehrtens
2018-09-03 23:43 ` Guenter Roeck
2018-09-09 21:07 ` Hauke Mehrtens
2018-09-10 13:26 ` Guenter Roeck
2018-09-02 11:32 ` [PATCH v2 3/3] wdt: lantiq: add get_timeleft callback Hauke Mehrtens
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=20180903230915.GA30065@roeck-us.net \
--to=linux@roeck-us.net \
--cc=dev@kresin.me \
--cc=hauke@hauke-m.de \
--cc=john@phrozen.org \
--cc=linux-watchdog@vger.kernel.org \
--cc=wim@linux-watchdog.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).