2018-05-11 13:51:16

by Charles Keepax

[permalink] [raw]
Subject: [PATCH] mfd: arizona: Don't use regmap_read_poll_timeout

Some Arizona CODECs have a small timing window where they will
NAK an I2C transaction if it happens before the boot done bit is
set. This can cause the read of the register containing the boot
done bit to fail until it is set. Since regmap_read_poll_timeout
will abort polling if a read fails it can't be reliably used to
poll the boot done bit over I2C.

Do a partial revert of ef84f885e037 ("mfd: arizona: Refactor
arizona_poll_reg"), removing the regmap_read_poll_timeout but
leaving the refactoring to make the arizona_poll_reg take more
sensible arguments.

Fixes: ef84f885e037 ("mfd: arizona: Refactor arizona_poll_reg")
Signed-off-by: Charles Keepax <[email protected]>
---
drivers/mfd/arizona-core.c | 29 +++++++++++++++++++++--------
1 file changed, 21 insertions(+), 8 deletions(-)

diff --git a/drivers/mfd/arizona-core.c b/drivers/mfd/arizona-core.c
index 83f1c5a516d9..2ed29ce08dac 100644
--- a/drivers/mfd/arizona-core.c
+++ b/drivers/mfd/arizona-core.c
@@ -240,18 +240,31 @@ static int arizona_poll_reg(struct arizona *arizona,
int timeout_ms, unsigned int reg,
unsigned int mask, unsigned int target)
{
+ ktime_t timeout = ktime_add_us(ktime_get(), timeout_ms * 1000);
unsigned int val = 0;
int ret;

- ret = regmap_read_poll_timeout(arizona->regmap,
- reg, val, ((val & mask) == target),
- ARIZONA_REG_POLL_DELAY_US,
- timeout_ms * 1000);
- if (ret)
- dev_err(arizona->dev, "Polling reg 0x%x timed out: %x\n",
- reg, val);
+ while (true) {
+ ret = regmap_read(arizona->regmap, reg, &val);

- return ret;
+ if ((val & mask) == target)
+ return 0;
+
+ if (ktime_compare(ktime_get(), timeout) > 0)
+ break;
+
+ usleep_range(ARIZONA_REG_POLL_DELAY_US / 2,
+ ARIZONA_REG_POLL_DELAY_US);
+ }
+
+ if (ret) {
+ dev_err(arizona->dev, "Failed polling reg 0x%x: %d\n",
+ reg, ret);
+ return ret;
+ }
+
+ dev_err(arizona->dev, "Polling reg 0x%x timed out: %x\n", reg, val);
+ return -ETIMEDOUT;
}

static int arizona_wait_for_boot(struct arizona *arizona)
--
2.11.0



2018-06-04 05:45:59

by Lee Jones

[permalink] [raw]
Subject: Re: [PATCH] mfd: arizona: Don't use regmap_read_poll_timeout

On Fri, 11 May 2018, Charles Keepax wrote:

> Some Arizona CODECs have a small timing window where they will
> NAK an I2C transaction if it happens before the boot done bit is
> set. This can cause the read of the register containing the boot
> done bit to fail until it is set. Since regmap_read_poll_timeout
> will abort polling if a read fails it can't be reliably used to
> poll the boot done bit over I2C.
>
> Do a partial revert of ef84f885e037 ("mfd: arizona: Refactor
> arizona_poll_reg"), removing the regmap_read_poll_timeout but
> leaving the refactoring to make the arizona_poll_reg take more
> sensible arguments.
>
> Fixes: ef84f885e037 ("mfd: arizona: Refactor arizona_poll_reg")
> Signed-off-by: Charles Keepax <[email protected]>
> ---
> drivers/mfd/arizona-core.c | 29 +++++++++++++++++++++--------
> 1 file changed, 21 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/mfd/arizona-core.c b/drivers/mfd/arizona-core.c
> index 83f1c5a516d9..2ed29ce08dac 100644
> --- a/drivers/mfd/arizona-core.c
> +++ b/drivers/mfd/arizona-core.c
> @@ -240,18 +240,31 @@ static int arizona_poll_reg(struct arizona *arizona,
> int timeout_ms, unsigned int reg,
> unsigned int mask, unsigned int target)
> {
> + ktime_t timeout = ktime_add_us(ktime_get(), timeout_ms * 1000);

USEC_PER_MSEC ?

> unsigned int val = 0;
> int ret;
>
> - ret = regmap_read_poll_timeout(arizona->regmap,
> - reg, val, ((val & mask) == target),
> - ARIZONA_REG_POLL_DELAY_US,
> - timeout_ms * 1000);
> - if (ret)
> - dev_err(arizona->dev, "Polling reg 0x%x timed out: %x\n",
> - reg, val);
> + while (true) {
> + ret = regmap_read(arizona->regmap, reg, &val);
>
> - return ret;
> + if ((val & mask) == target)
> + return 0;
> +
> + if (ktime_compare(ktime_get(), timeout) > 0)
> + break;
> +
> + usleep_range(ARIZONA_REG_POLL_DELAY_US / 2,
> + ARIZONA_REG_POLL_DELAY_US);
> + }
> +
> + if (ret) {
> + dev_err(arizona->dev, "Failed polling reg 0x%x: %d\n",
> + reg, ret);
> + return ret;
> + }
> +
> + dev_err(arizona->dev, "Polling reg 0x%x timed out: %x\n", reg, val);
> + return -ETIMEDOUT;
> }
>
> static int arizona_wait_for_boot(struct arizona *arizona)

--
Lee Jones [李琼斯]
Linaro Services Technical Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog

2018-06-04 09:41:17

by Andy Shevchenko

[permalink] [raw]
Subject: Re: [PATCH] mfd: arizona: Don't use regmap_read_poll_timeout

On Mon, Jun 4, 2018 at 8:44 AM, Lee Jones <[email protected]> wrote:
> On Fri, 11 May 2018, Charles Keepax wrote:

>> + while (true) {
>> + ret = regmap_read(arizona->regmap, reg, &val);
>>
>> + if ((val & mask) == target)
>> + return 0;
>> +
>> + if (ktime_compare(ktime_get(), timeout) > 0)
>> + break;
>> +
>> + usleep_range(ARIZONA_REG_POLL_DELAY_US / 2,
>> + ARIZONA_REG_POLL_DELAY_US);
>> + }

From my point of view infinite loops not good for readability and maintenance.
Perhaps

do {
...
} while (ktime_compare(...));

?

--
With Best Regards,
Andy Shevchenko

2018-06-04 10:28:48

by Charles Keepax

[permalink] [raw]
Subject: Re: [PATCH] mfd: arizona: Don't use regmap_read_poll_timeout

On Mon, Jun 04, 2018 at 12:40:37PM +0300, Andy Shevchenko wrote:
> On Mon, Jun 4, 2018 at 8:44 AM, Lee Jones <[email protected]> wrote:
> > On Fri, 11 May 2018, Charles Keepax wrote:
>
> >> + while (true) {
> >> + ret = regmap_read(arizona->regmap, reg, &val);
> >>
> >> + if ((val & mask) == target)
> >> + return 0;
> >> +
> >> + if (ktime_compare(ktime_get(), timeout) > 0)
> >> + break;
> >> +
> >> + usleep_range(ARIZONA_REG_POLL_DELAY_US / 2,
> >> + ARIZONA_REG_POLL_DELAY_US);
> >> + }
>
> >From my point of view infinite loops not good for readability and maintenance.
> Perhaps
>
> do {
> ...
> } while (ktime_compare(...));
>
> ?

I would rather not do it that way since then it becomes
impossible to have a single poll with no delays happening. I can
refactor to remove the while(true) but it will make the code a
bit bigger, as I will probably need to duplicate some code
outside the loop.

Thanks,
Charles

2018-06-04 10:57:19

by Charles Keepax

[permalink] [raw]
Subject: Re: [PATCH] mfd: arizona: Don't use regmap_read_poll_timeout

On Mon, Jun 04, 2018 at 11:27:19AM +0100, Charles Keepax wrote:
> On Mon, Jun 04, 2018 at 12:40:37PM +0300, Andy Shevchenko wrote:
> > On Mon, Jun 4, 2018 at 8:44 AM, Lee Jones <[email protected]> wrote:
> > > On Fri, 11 May 2018, Charles Keepax wrote:
> >
> > >> + while (true) {
> > >> + ret = regmap_read(arizona->regmap, reg, &val);
> > >>
> > >> + if ((val & mask) == target)
> > >> + return 0;
> > >> +
> > >> + if (ktime_compare(ktime_get(), timeout) > 0)
> > >> + break;
> > >> +
> > >> + usleep_range(ARIZONA_REG_POLL_DELAY_US / 2,
> > >> + ARIZONA_REG_POLL_DELAY_US);
> > >> + }
> >
> > >From my point of view infinite loops not good for readability and maintenance.
> > Perhaps
> >
> > do {
> > ...
> > } while (ktime_compare(...));
> >
> > ?
>
> I would rather not do it that way since then it becomes
> impossible to have a single poll with no delays happening. I can
> refactor to remove the while(true) but it will make the code a
> bit bigger, as I will probably need to duplicate some code
> outside the loop.
>

Looking again I think it is fine if I do the ktime_compare at the top
of the loop though. Will respin.

Thanks,
Charles