[PATCH] i2c: mxs: Handle i2c DMA failure properly

[PATCH] i2c: mxs: Handle i2c DMA failure properly

[Marek Vasut]

Properly terminate the DMA transfer in case the DMA PIO transfer
or setup fails for any reason. While at it, enable support for
SMBUS_QUICK. This essentially fixes i2c-detect on this IP.

Signed-off-by: Marek Vasut <marex-ynQEQJNshbs@public.gmane.org>
Cc: Fabio Estevam <festevam-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: Tim Michals <tcmichals-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: Wolfram Sang <w.sang-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org>
---
 drivers/i2c/busses/i2c-mxs.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/i2c/busses/i2c-mxs.c b/drivers/i2c/busses/i2c-mxs.c
index 286ca19..6ed53da 100644
--- a/drivers/i2c/busses/i2c-mxs.c
+++ b/drivers/i2c/busses/i2c-mxs.c
@@ -287,12 +287,14 @@ read_init_dma_fail:
 select_init_dma_fail:
 	dma_unmap_sg(i2c->dev, &i2c->sg_io[0], 1, DMA_TO_DEVICE);
 select_init_pio_fail:
+	dmaengine_terminate_all(i2c->dmach);
 	return -EINVAL;
 
 /* Write failpath. */
 write_init_dma_fail:
 	dma_unmap_sg(i2c->dev, i2c->sg_io, 2, DMA_TO_DEVICE);
 write_init_pio_fail:
+	dmaengine_terminate_all(i2c->dmach);
 	return -EINVAL;
 }
 
@@ -357,7 +359,7 @@ static int mxs_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[],
 
 static u32 mxs_i2c_func(struct i2c_adapter *adap)
 {
-	return I2C_FUNC_I2C | (I2C_FUNC_SMBUS_EMUL & ~I2C_FUNC_SMBUS_QUICK);
+	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
 }
 
 static irqreturn_t mxs_i2c_isr(int this_irq, void *dev_id)
-- 
1.7.10.4

Re: [PATCH] i2c: mxs: Handle i2c DMA failure properly

[Fabio Estevam]

Hi Marek,

On Wed, Nov 14, 2012 at 12:31 PM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> Properly terminate the DMA transfer in case the DMA PIO transfer
> or setup fails for any reason. While at it, enable support for
> SMBUS_QUICK. This essentially fixes i2c-detect on this IP.

Shouldn't this be split into two patches?

Regards,

Fabio Estevam

Re: [PATCH] i2c: mxs: Handle i2c DMA failure properly

[Marek Vasut]

Dear Fabio Estevam,

> Hi Marek,
> 
> On Wed, Nov 14, 2012 at 12:31 PM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> > Properly terminate the DMA transfer in case the DMA PIO transfer
> > or setup fails for any reason. While at it, enable support for
> > SMBUS_QUICK. This essentially fixes i2c-detect on this IP.
> 
> Shouldn't this be split into two patches?

It should, but then it fixes one problem, so I stuck it into one ...

> Regards,
> 
> Fabio Estevam

Best regards,
Marek Vasut

Re: [PATCH] i2c: mxs: Handle i2c DMA failure properly

[Marek Vasut]

Dear Tim Michals,

> On Wed, Nov 14, 2012 at 9:22 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> > Dear Fabio Estevam,
> > 
> > > Hi Marek,
> > > 
> > > On Wed, Nov 14, 2012 at 12:31 PM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> > > > Properly terminate the DMA transfer in case the DMA PIO transfer
> > > > or setup fails for any reason. While at it, enable support for
> > > > SMBUS_QUICK. This essentially fixes i2c-detect on this IP.
> > > 
> > > Shouldn't this be split into two patches?
> > 
> > It should, but then it fixes one problem, so I stuck it into one ...
> > 
> > > Regards,
> > > 
> > > Fabio Estevam
> > 
> > Best regards,
> > Marek Vasut
>
> Dear Marek Vasut,
> 
> This patch
>  @@ -150,8 +150,6 @@ static void mxs_i2c_reset(struct mxs_i2c
>   writel(i2c->speed->timing2, i2c->regs + MXS_I2C_TIMING2);
> 
>   writel(MXS_I2C_IRQ_MASK << 8, i2c->regs + MXS_I2C_CTRL1_SET);
> -
> - dmaengine_terminate_all(i2c->dmach);
>  }

This patch doesn't apply, can you elaborate ?

Moreover ... read either RFC1855 properly [1] and/or read the mailing list 
etiquette please (!). Especially the part about top-posting. Do us a favor.

> This patch fixes current DMA requests which time out or error returned by
> the ISR.
> 
> Tim

[1] http://www.ietf.org/rfc/rfc1855.txt

Best regards,
Marek Vasut

Re: [PATCH] i2c: mxs: Handle i2c DMA failure properly

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 854 bytes --]

Tim,

On Wed, Nov 14, 2012 at 09:34:27PM -0600, Tim Michals wrote:
>    This patch
>    �@@ -150,8 +150,6 @@ static void mxs_i2c_reset(struct mxs_i2c
>    �       writel(i2c->speed->timing2, i2c->regs + MXS_I2C_TIMING2);
>    �
>    �       writel(MXS_I2C_IRQ_MASK << 8, i2c->regs + MXS_I2C_CTRL1_SET);
>    -
>    -       dmaengine_terminate_all(i2c->dmach);
>    �}��
>    This patch fixes current DMA requests which time out or error returned by
>    the ISR.�
>    Tim

Please elaborate because I am confused. Did you see problems after
applying Marek's patch and if so, please describe in what circumstances
so we could reproduce.

Thanks,

   Wolfram

-- 
Pengutronix e.K.                           | Wolfram Sang                |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 198 bytes --]

Re: [PATCH] i2c: mxs: Handle i2c DMA failure properly

[Fabio Estevam]

Hi Tim,

On Fri, Nov 16, 2012 at 1:24 PM, Tim Michals <tcmichals-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> My suggestion was to always clear the DMA in the reset i2c code.  This way
> if there is an error reported from the ISR or a timeout, it always clears
> it.  I've tested this approach with I2C devices connected and and doing a
> i2cdetect -r 0.

Thanks for the explanation.

Maybe you could post your proposed patch?

Thanks,

Fabio Estevam