2022-08-16 07:38:20

by Alice Guo (OSS)

[permalink] [raw]
Subject: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue

From: Ye Li <[email protected]>

According to measure on i.MX7ULP and i.MX8ULP, the RCS done needs
about 3400us and 6700us respectively. So current 20us timeout is
not enough. When reconfiguring is on-going, unlock and configure CS
will lead to unknown result.

Increase the wait timeout value to 10ms and check the return value
of RCS wait to fix the issue

Signed-off-by: Ye Li <[email protected]>
Signed-off-by: Alice Guo <[email protected]>
Reviewed-by: Jacky Bai <[email protected]>
Acked-by: Jason Liu <[email protected]>
---
drivers/watchdog/imx7ulp_wdt.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/watchdog/imx7ulp_wdt.c b/drivers/watchdog/imx7ulp_wdt.c
index a0f6b8cea78f..12715c248688 100644
--- a/drivers/watchdog/imx7ulp_wdt.c
+++ b/drivers/watchdog/imx7ulp_wdt.c
@@ -39,7 +39,7 @@
#define DEFAULT_TIMEOUT 60
#define MAX_TIMEOUT 128
#define WDOG_CLOCK_RATE 1000
-#define WDOG_WAIT_TIMEOUT 20
+#define WDOG_WAIT_TIMEOUT 10000

static bool nowayout = WATCHDOG_NOWAYOUT;
module_param(nowayout, bool, 0000);
@@ -80,7 +80,7 @@ static int imx7ulp_wdt_enable(struct watchdog_device *wdog, bool enable)
writel(val | WDOG_CS_EN, wdt->base + WDOG_CS);
else
writel(val & ~WDOG_CS_EN, wdt->base + WDOG_CS);
- imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
+ ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);

enable_out:
local_irq_enable();
@@ -127,7 +127,9 @@ static int imx7ulp_wdt_set_timeout(struct watchdog_device *wdog,
if (ret)
goto timeout_out;
writel(val, wdt->base + WDOG_TOVAL);
- imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
+ ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
+ if (ret)
+ goto timeout_out;

wdog->timeout = timeout;

--
2.17.1


2022-08-22 14:26:53

by Guenter Roeck

[permalink] [raw]
Subject: Re: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue

On Tue, Aug 16, 2022 at 12:36:40PM +0800, Alice Guo (OSS) wrote:
> From: Ye Li <[email protected]>
>
> According to measure on i.MX7ULP and i.MX8ULP, the RCS done needs
> about 3400us and 6700us respectively. So current 20us timeout is
> not enough. When reconfiguring is on-going, unlock and configure CS
> will lead to unknown result.
>
> Increase the wait timeout value to 10ms and check the return value
> of RCS wait to fix the issue
>

You'll have to find a better solution. An active (non-sleep) wait of
10 ms is unacceptable.

Guenter

> Signed-off-by: Ye Li <[email protected]>
> Signed-off-by: Alice Guo <[email protected]>
> Reviewed-by: Jacky Bai <[email protected]>
> Acked-by: Jason Liu <[email protected]>
> ---
> drivers/watchdog/imx7ulp_wdt.c | 8 +++++---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/watchdog/imx7ulp_wdt.c b/drivers/watchdog/imx7ulp_wdt.c
> index a0f6b8cea78f..12715c248688 100644
> --- a/drivers/watchdog/imx7ulp_wdt.c
> +++ b/drivers/watchdog/imx7ulp_wdt.c
> @@ -39,7 +39,7 @@
> #define DEFAULT_TIMEOUT 60
> #define MAX_TIMEOUT 128
> #define WDOG_CLOCK_RATE 1000
> -#define WDOG_WAIT_TIMEOUT 20
> +#define WDOG_WAIT_TIMEOUT 10000
>
> static bool nowayout = WATCHDOG_NOWAYOUT;
> module_param(nowayout, bool, 0000);
> @@ -80,7 +80,7 @@ static int imx7ulp_wdt_enable(struct watchdog_device *wdog, bool enable)
> writel(val | WDOG_CS_EN, wdt->base + WDOG_CS);
> else
> writel(val & ~WDOG_CS_EN, wdt->base + WDOG_CS);
> - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
>
> enable_out:
> local_irq_enable();
> @@ -127,7 +127,9 @@ static int imx7ulp_wdt_set_timeout(struct watchdog_device *wdog,
> if (ret)
> goto timeout_out;
> writel(val, wdt->base + WDOG_TOVAL);
> - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> + if (ret)
> + goto timeout_out;
>
> wdog->timeout = timeout;
>
> --
> 2.17.1
>

2022-08-23 06:16:03

by Alice Guo (OSS)

[permalink] [raw]
Subject: RE: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue



> -----Original Message-----
> From: Guenter Roeck <[email protected]> On Behalf Of Guenter Roeck
> Sent: Monday, August 22, 2022 10:10 PM
> To: Alice Guo (OSS) <[email protected]>
> Cc: [email protected]; [email protected];
> [email protected]; [email protected]; [email protected];
> dl-linux-imx <[email protected]>; [email protected];
> [email protected]; [email protected]
> Subject: Re: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue
>
> On Tue, Aug 16, 2022 at 12:36:40PM +0800, Alice Guo (OSS) wrote:
> > From: Ye Li <[email protected]>
> >
> > According to measure on i.MX7ULP and i.MX8ULP, the RCS done needs
> > about 3400us and 6700us respectively. So current 20us timeout is not
> > enough. When reconfiguring is on-going, unlock and configure CS will
> > lead to unknown result.
> >
> > Increase the wait timeout value to 10ms and check the return value of
> > RCS wait to fix the issue
> >
>
> You'll have to find a better solution. An active (non-sleep) wait of
> 10 ms is unacceptable.
>
> Guenter

Hi Guenter,

Sorry. I think this patch should be merged with " watchdog: imx7ulp_wdt: Handle wdog reconfigure failure", but I didn't merge them. I will send v2.

Best Regards,
Alice Guo

> > Signed-off-by: Ye Li <[email protected]>
> > Signed-off-by: Alice Guo <[email protected]>
> > Reviewed-by: Jacky Bai <[email protected]>
> > Acked-by: Jason Liu <[email protected]>
> > ---
> > drivers/watchdog/imx7ulp_wdt.c | 8 +++++---
> > 1 file changed, 5 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/watchdog/imx7ulp_wdt.c
> > b/drivers/watchdog/imx7ulp_wdt.c index a0f6b8cea78f..12715c248688
> > 100644
> > --- a/drivers/watchdog/imx7ulp_wdt.c
> > +++ b/drivers/watchdog/imx7ulp_wdt.c
> > @@ -39,7 +39,7 @@
> > #define DEFAULT_TIMEOUT 60
> > #define MAX_TIMEOUT 128
> > #define WDOG_CLOCK_RATE 1000
> > -#define WDOG_WAIT_TIMEOUT 20
> > +#define WDOG_WAIT_TIMEOUT 10000
> >
> > static bool nowayout = WATCHDOG_NOWAYOUT;
> module_param(nowayout,
> > bool, 0000); @@ -80,7 +80,7 @@ static int imx7ulp_wdt_enable(struct
> > watchdog_device *wdog, bool enable)
> > writel(val | WDOG_CS_EN, wdt->base + WDOG_CS);
> > else
> > writel(val & ~WDOG_CS_EN, wdt->base + WDOG_CS);
> > - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> >
> > enable_out:
> > local_irq_enable();
> > @@ -127,7 +127,9 @@ static int imx7ulp_wdt_set_timeout(struct
> watchdog_device *wdog,
> > if (ret)
> > goto timeout_out;
> > writel(val, wdt->base + WDOG_TOVAL);
> > - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > + if (ret)
> > + goto timeout_out;
> >
> > wdog->timeout = timeout;
> >
> > --
> > 2.17.1
> >

2022-08-23 15:04:33

by Guenter Roeck

[permalink] [raw]
Subject: Re: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue

On Tue, Aug 23, 2022 at 05:59:11AM +0000, Alice Guo (OSS) wrote:
>
>
> > -----Original Message-----
> > From: Guenter Roeck <[email protected]> On Behalf Of Guenter Roeck
> > Sent: Monday, August 22, 2022 10:10 PM
> > To: Alice Guo (OSS) <[email protected]>
> > Cc: [email protected]; [email protected];
> > [email protected]; [email protected]; [email protected];
> > dl-linux-imx <[email protected]>; [email protected];
> > [email protected]; [email protected]
> > Subject: Re: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue
> >
> > On Tue, Aug 16, 2022 at 12:36:40PM +0800, Alice Guo (OSS) wrote:
> > > From: Ye Li <[email protected]>
> > >
> > > According to measure on i.MX7ULP and i.MX8ULP, the RCS done needs
> > > about 3400us and 6700us respectively. So current 20us timeout is not
> > > enough. When reconfiguring is on-going, unlock and configure CS will
> > > lead to unknown result.
> > >
> > > Increase the wait timeout value to 10ms and check the return value of
> > > RCS wait to fix the issue
> > >
> >
> > You'll have to find a better solution. An active (non-sleep) wait of
> > 10 ms is unacceptable.
> >
> > Guenter
>
> Hi Guenter,
>
> Sorry. I think this patch should be merged with " watchdog: imx7ulp_wdt: Handle wdog reconfigure failure", but I didn't merge them. I will send v2.
>

That doesn't change the fact that a 10 ms hot wait is unacceptable.

Thanks,
Guenter

> Best Regards,
> Alice Guo
>
> > > Signed-off-by: Ye Li <[email protected]>
> > > Signed-off-by: Alice Guo <[email protected]>
> > > Reviewed-by: Jacky Bai <[email protected]>
> > > Acked-by: Jason Liu <[email protected]>
> > > ---
> > > drivers/watchdog/imx7ulp_wdt.c | 8 +++++---
> > > 1 file changed, 5 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/drivers/watchdog/imx7ulp_wdt.c
> > > b/drivers/watchdog/imx7ulp_wdt.c index a0f6b8cea78f..12715c248688
> > > 100644
> > > --- a/drivers/watchdog/imx7ulp_wdt.c
> > > +++ b/drivers/watchdog/imx7ulp_wdt.c
> > > @@ -39,7 +39,7 @@
> > > #define DEFAULT_TIMEOUT 60
> > > #define MAX_TIMEOUT 128
> > > #define WDOG_CLOCK_RATE 1000
> > > -#define WDOG_WAIT_TIMEOUT 20
> > > +#define WDOG_WAIT_TIMEOUT 10000
> > >
> > > static bool nowayout = WATCHDOG_NOWAYOUT;
> > module_param(nowayout,
> > > bool, 0000); @@ -80,7 +80,7 @@ static int imx7ulp_wdt_enable(struct
> > > watchdog_device *wdog, bool enable)
> > > writel(val | WDOG_CS_EN, wdt->base + WDOG_CS);
> > > else
> > > writel(val & ~WDOG_CS_EN, wdt->base + WDOG_CS);
> > > - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > > + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > >
> > > enable_out:
> > > local_irq_enable();
> > > @@ -127,7 +127,9 @@ static int imx7ulp_wdt_set_timeout(struct
> > watchdog_device *wdog,
> > > if (ret)
> > > goto timeout_out;
> > > writel(val, wdt->base + WDOG_TOVAL);
> > > - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > > + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > > + if (ret)
> > > + goto timeout_out;
> > >
> > > wdog->timeout = timeout;
> > >
> > > --
> > > 2.17.1
> > >

2022-08-24 07:05:56

by Alice Guo (OSS)

[permalink] [raw]
Subject: RE: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue

> -----Original Message-----
> From: Guenter Roeck <[email protected]> On Behalf Of Guenter Roeck
> Sent: Tuesday, August 23, 2022 8:07 PM
> To: Alice Guo (OSS) <[email protected]>
> Cc: [email protected]; [email protected];
> [email protected]; [email protected]; [email protected];
> dl-linux-imx <[email protected]>; [email protected];
> [email protected]; [email protected]
> Subject: Re: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout issue
>
> On Tue, Aug 23, 2022 at 05:59:11AM +0000, Alice Guo (OSS) wrote:
> >
> >
> > > -----Original Message-----
> > > From: Guenter Roeck <[email protected]> On Behalf Of Guenter Roeck
> > > Sent: Monday, August 22, 2022 10:10 PM
> > > To: Alice Guo (OSS) <[email protected]>
> > > Cc: [email protected]; [email protected];
> > > [email protected]; [email protected]; [email protected];
> > > dl-linux-imx <[email protected]>; [email protected];
> > > [email protected]; [email protected]
> > > Subject: Re: [PATCH 4/7] watchdog: imx7ulp_wdt: Fix RCS timeout
> > > issue
> > >
> > > On Tue, Aug 16, 2022 at 12:36:40PM +0800, Alice Guo (OSS) wrote:
> > > > From: Ye Li <[email protected]>
> > > >
> > > > According to measure on i.MX7ULP and i.MX8ULP, the RCS done needs
> > > > about 3400us and 6700us respectively. So current 20us timeout is
> > > > not enough. When reconfiguring is on-going, unlock and configure
> > > > CS will lead to unknown result.
> > > >
> > > > Increase the wait timeout value to 10ms and check the return value
> > > > of RCS wait to fix the issue
> > > >
> > >
> > > You'll have to find a better solution. An active (non-sleep) wait of
> > > 10 ms is unacceptable.
> > >
> > > Guenter
> >
> > Hi Guenter,
> >
> > Sorry. I think this patch should be merged with " watchdog: imx7ulp_wdt:
> Handle wdog reconfigure failure", but I didn't merge them. I will send v2.
> >
>
> That doesn't change the fact that a 10 ms hot wait is unacceptable.
>
> Thanks,
> Guenter

Hi Guenter,

I plan to use readl_poll_timeout to wait for RCS(Reconfiguration Success). I think the sleep_us of readl_poll_timeout can be used to avoid a 10 ms hot wait.

Best Regards,
Alice Guo

> > Best Regards,
> > Alice Guo
> >
> > > > Signed-off-by: Ye Li <[email protected]>
> > > > Signed-off-by: Alice Guo <[email protected]>
> > > > Reviewed-by: Jacky Bai <[email protected]>
> > > > Acked-by: Jason Liu <[email protected]>
> > > > ---
> > > > drivers/watchdog/imx7ulp_wdt.c | 8 +++++---
> > > > 1 file changed, 5 insertions(+), 3 deletions(-)
> > > >
> > > > diff --git a/drivers/watchdog/imx7ulp_wdt.c
> > > > b/drivers/watchdog/imx7ulp_wdt.c index a0f6b8cea78f..12715c248688
> > > > 100644
> > > > --- a/drivers/watchdog/imx7ulp_wdt.c
> > > > +++ b/drivers/watchdog/imx7ulp_wdt.c
> > > > @@ -39,7 +39,7 @@
> > > > #define DEFAULT_TIMEOUT 60
> > > > #define MAX_TIMEOUT 128
> > > > #define WDOG_CLOCK_RATE 1000
> > > > -#define WDOG_WAIT_TIMEOUT 20
> > > > +#define WDOG_WAIT_TIMEOUT 10000
> > > >
> > > > static bool nowayout = WATCHDOG_NOWAYOUT;
> > > module_param(nowayout,
> > > > bool, 0000); @@ -80,7 +80,7 @@ static int
> > > > imx7ulp_wdt_enable(struct watchdog_device *wdog, bool enable)
> > > > writel(val | WDOG_CS_EN, wdt->base + WDOG_CS);
> > > > else
> > > > writel(val & ~WDOG_CS_EN, wdt->base + WDOG_CS);
> > > > - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > > > + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > > >
> > > > enable_out:
> > > > local_irq_enable();
> > > > @@ -127,7 +127,9 @@ static int imx7ulp_wdt_set_timeout(struct
> > > watchdog_device *wdog,
> > > > if (ret)
> > > > goto timeout_out;
> > > > writel(val, wdt->base + WDOG_TOVAL);
> > > > - imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > > > + ret = imx7ulp_wdt_wait(wdt->base, WDOG_CS_RCS);
> > > > + if (ret)
> > > > + goto timeout_out;
> > > >
> > > > wdog->timeout = timeout;
> > > >
> > > > --
> > > > 2.17.1
> > > >