On Tue, Nov 02, 2021 at 09:58:06AM +0100, Uwe Kleine-König wrote:
> On Mon, Oct 04, 2021 at 07:32:16PM -0500, minyard@xxxxxxx wrote:
> > From: Corey Minyard <cminyard@xxxxxxxxxx>
> >
> > The timer is too slow and significantly reduces performance. Use an
> > hrtimer to get things working faster.
> >
> > Signed-off-by: Corey Minyard <minyard@xxxxxxx>
> > Tested-by: Andrew Manley <andrew.manley@xxxxxxxxxxxxxxx>
> > Reviewed-by: Andrew Manley <andrew.manley@xxxxxxxxxxxxxxx>
> > ---
> > drivers/i2c/busses/i2c-imx.c | 23 +++++++++++++++--------
> > 1 file changed, 15 insertions(+), 8 deletions(-)
> >
> > diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c
> > index 26a04dc0590b..4b0e9d1784dd 100644
> > --- a/drivers/i2c/busses/i2c-imx.c
> > +++ b/drivers/i2c/busses/i2c-imx.c
> > @@ -38,7 +38,7 @@
> > #include <linux/iopoll.h>
> > #include <linux/kernel.h>
> > #include <linux/spinlock.h>
> > -#include <linux/timer.h>
> > +#include <linux/hrtimer.h>
> > #include <linux/module.h>
> > #include <linux/of.h>
> > #include <linux/of_device.h>
> > @@ -53,6 +53,8 @@
> > /* This will be the driver name the kernel reports */
> > #define DRIVER_NAME "imx-i2c"
> >
> > +#define I2C_IMX_CHECK_DELAY 30000 /* Time to check for bus idle, in NS */
> > +
> > /*
> > * Enable DMA if transfer byte size is bigger than this threshold.
> > * As the hardware request, it must bigger than 4 bytes.\
> > @@ -214,8 +216,8 @@ struct imx_i2c_struct {
> > enum i2c_slave_event last_slave_event;
> >
> > /* For checking slave events. */
> > - spinlock_t slave_lock;
> > - struct timer_list slave_timer;
> > + spinlock_t slave_lock;
> > + struct hrtimer slave_timer;
>
> This is unrelated to this patch, moreover it was introduced only in
> patch 1.
The second line is important for this patch, of course. I assume you
mean the indention of the first line, which is just keeping things lined
up.
>
> > };
> >
> > static const struct imx_i2c_hwdata imx1_i2c_hwdata = {
> > @@ -783,13 +785,16 @@ static irqreturn_t i2c_imx_slave_handle(struct imx_i2c_struct *i2c_imx,
> > }
> >
> > out:
> > - mod_timer(&i2c_imx->slave_timer, jiffies + 1);
> > + hrtimer_try_to_cancel(&i2c_imx->slave_timer);
>
> Don't you need to check the return value here?
Not really. The possible return values are:
* * 0 when the timer was not active
* * 1 when the timer was active
* * -1 when the timer is currently executing the callback function and
* cannot be stopped
and if it returns 0 or 1, then everything is fine. If it returns -1,
then the code will still work, though it may be redone (or already have
been done) by the timer function. So it doesn't matter.
Maybe I should add a comment about this?
Thanks for reviewing.
-corey
>
> > + hrtimer_forward_now(&i2c_imx->slave_timer, I2C_IMX_CHECK_DELAY);
> > + hrtimer_restart(&i2c_imx->slave_timer);
> > return IRQ_HANDLED;
> > }
> >
>
> Best regards
> Uwe
>
> --
> Pengutronix e.K. | Uwe Kleine-König |
> Industrial Linux Solutions | https://www.pengutronix.de/ |
