[PATCH] i2c: xiic: print error code when xiic_start_xfer fails

[PATCH] i2c: xiic: print error code when xiic_start_xfer fails

[marc.ferland]

From: Marc Ferland <marc.ferland@sonatest.com>

xiic_start_xfer can fail for different reasons:

- EBUSY: bus is busy or i2c messages still in tx_msg or rx_msg
- ETIMEDOUT: timed-out trying to clear the RX fifo
- EINVAL: wrong clock settings

Printing the error code helps identifying the root cause.

Signed-off-by: Marc Ferland <marc.ferland@sonatest.com>
---
 drivers/i2c/busses/i2c-xiic.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c
index 71391b590ada..73729f0180a1 100644
--- a/drivers/i2c/busses/i2c-xiic.c
+++ b/drivers/i2c/busses/i2c-xiic.c
@@ -1165,7 +1165,7 @@ static int xiic_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 
 	err = xiic_start_xfer(i2c, msgs, num);
 	if (err < 0) {
-		dev_err(adap->dev.parent, "Error xiic_start_xfer\n");
+		dev_err(adap->dev.parent, "Error xiic_start_xfer: %d\n", err);
 		goto out;
 	}
 
-- 
2.34.1

Re: [PATCH] i2c: xiic: print error code when xiic_start_xfer fails

[Andi Shyti]

Hi Marc,

On Mon, May 06, 2024 at 11:45:20AM -0400, marc.ferland@gmail.com wrote:
> From: Marc Ferland <marc.ferland@sonatest.com>
> 
> xiic_start_xfer can fail for different reasons:
> 
> - EBUSY: bus is busy or i2c messages still in tx_msg or rx_msg
> - ETIMEDOUT: timed-out trying to clear the RX fifo

Wolfram has recently removed all the error printouts caused by
etimedout.

> - EINVAL: wrong clock settings
> 
> Printing the error code helps identifying the root cause.
> 
> Signed-off-by: Marc Ferland <marc.ferland@sonatest.com>
> ---
>  drivers/i2c/busses/i2c-xiic.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c
> index 71391b590ada..73729f0180a1 100644
> --- a/drivers/i2c/busses/i2c-xiic.c
> +++ b/drivers/i2c/busses/i2c-xiic.c
> @@ -1165,7 +1165,7 @@ static int xiic_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
>  
>  	err = xiic_start_xfer(i2c, msgs, num);
>  	if (err < 0) {
> -		dev_err(adap->dev.parent, "Error xiic_start_xfer\n");
> +		dev_err(adap->dev.parent, "Error xiic_start_xfer: %d\n", err);

I don't see the end user being interested in having the error
number printed in the dmesg. In fact, I doubt the end user is
interested in this message at all.

We could print the debug messages (or warnings) where the actual
failure occurs. For example, in the setclk case, a warning is
already being printed, and I don't see why we should also print
an error here.

Does that make sense?

Andi

>  		goto out;
>  	}
>  
> -- 
> 2.34.1
> 

Re: [PATCH] i2c: xiic: print error code when xiic_start_xfer fails

[Marc Ferland]

On Wed, May 8, 2024 at 8:51 AM Andi Shyti <andi.shyti@kernel.org> wrote:
>
> Hi Marc,
>
> On Mon, May 06, 2024 at 11:45:20AM -0400, marc.ferland@gmail.com wrote:
> > From: Marc Ferland <marc.ferland@sonatest.com>
> >
> > xiic_start_xfer can fail for different reasons:
> >
> > - EBUSY: bus is busy or i2c messages still in tx_msg or rx_msg
> > - ETIMEDOUT: timed-out trying to clear the RX fifo
>
> Wolfram has recently removed all the error printouts caused by
> etimedout.
>
> > - EINVAL: wrong clock settings
> >
> > Printing the error code helps identifying the root cause.
> >
> > Signed-off-by: Marc Ferland <marc.ferland@sonatest.com>
> > ---
> >  drivers/i2c/busses/i2c-xiic.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c
> > index 71391b590ada..73729f0180a1 100644
> > --- a/drivers/i2c/busses/i2c-xiic.c
> > +++ b/drivers/i2c/busses/i2c-xiic.c
> > @@ -1165,7 +1165,7 @@ static int xiic_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
> >
> >       err = xiic_start_xfer(i2c, msgs, num);
> >       if (err < 0) {
> > -             dev_err(adap->dev.parent, "Error xiic_start_xfer\n");
> > +             dev_err(adap->dev.parent, "Error xiic_start_xfer: %d\n", err);
>
> I don't see the end user being interested in having the error
> number printed in the dmesg. In fact, I doubt the end user is
> interested in this message at all.
>
> We could print the debug messages (or warnings) where the actual
> failure occurs. For example, in the setclk case, a warning is
> already being printed, and I don't see why we should also print
> an error here.
>
> Does that make sense?
>
Yeah, it makes sense. My goal here is to add a bit more context when
things go wrong.

I've been having these _very_ rare failures with this i2c adapter and
the only trace I am getting is this "Error xiic_start_xfer" message in
the kernel logs. So it seemed like a good idea to just print the error
code along with the message.

I'll resend a v2 with your suggestions.

Regards,

Marc