2018-05-23 09:59:05

by Esben Haabendal

Subject: [PATCH 2/4] i2c: imx: Fix race condition in dma read

From: Esben Haabendal <[email protected]>

This fixes a race condition, where the DMAEN bit ends up being set after
I2C slave has transmitted a byte following the dummy read. When that
happens, an interrupt is generated instead, and no DMA request is generated
to kickstart the DMA read, and a timeout happens after DMA_TIMEOUT (1 sec).

Fixed by setting the DMAEN bit before the dummy read.

Signed-off-by: Esben Haabendal <[email protected]>
---
drivers/i2c/busses/i2c-imx.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/i2c/busses/i2c-imx.c b/drivers/i2c/busses/i2c-imx.c
index 6fca5e64cffb..742b548437af 100644
--- a/drivers/i2c/busses/i2c-imx.c
+++ b/drivers/i2c/busses/i2c-imx.c
@@ -677,9 +677,6 @@ static int i2c_imx_dma_read(struct imx_i2c_struct *i2c_imx,
struct imx_i2c_dma *dma = i2c_imx->dma;
struct device *dev = &i2c_imx->adapter.dev;

- temp = imx_i2c_read_reg(i2c_imx, IMX_I2C_I2CR);
- temp |= I2CR_DMAEN;
- imx_i2c_write_reg(temp, i2c_imx, IMX_I2C_I2CR);

dma->chan_using = dma->chan_rx;
dma->dma_transfer_dir = DMA_DEV_TO_MEM;
@@ -819,6 +816,11 @@ static int i2c_imx_read(struct imx_i2c_struct *i2c_imx, struct i2c_msg *msgs, bo
if ((msgs->len - 1) || block_data)
temp &= ~I2CR_TXAK;
imx_i2c_write_reg(temp, i2c_imx, IMX_I2C_I2CR);
+ if (i2c_imx->dma && msgs->len >= DMA_THRESHOLD && !block_data) {
+ temp = imx_i2c_read_reg(i2c_imx, IMX_I2C_I2CR);
+ temp |= I2CR_DMAEN;
+ imx_i2c_write_reg(temp, i2c_imx, IMX_I2C_I2CR);
+ }
imx_i2c_read_reg(i2c_imx, IMX_I2C_I2DR); /* dummy read */

dev_dbg(&i2c_imx->adapter.dev, "<%s> read data\n", __func__);
--
2.17.0

2018-07-05 06:35:12

by Uwe Kleine-König

Subject: Re: [PATCH 2/4] i2c: imx: Fix race condition in dma read

On Wed, May 23, 2018 at 11:56:21AM +0200, Esben Haabendal wrote:
> From: Esben Haabendal <[email protected]>
>
> This fixes a race condition, where the DMAEN bit ends up being set after
> I2C slave has transmitted a byte following the dummy read. When that
> happens, an interrupt is generated instead, and no DMA request is generated
> to kickstart the DMA read, and a timeout happens after DMA_TIMEOUT (1 sec).
>
> Fixed by setting the DMAEN bit before the dummy read.

Does this fix the problem or just narrow the race window?

Best regards
Uwe

--
Pengutronix e.K. | Uwe Kleine-K?nig |
Industrial Linux Solutions | http://www.pengutronix.de/ |

2018-07-09 09:23:40

by Esben Haabendal

Subject: Re: [PATCH 2/4] i2c: imx: Fix race condition in dma read

Uwe Kleine-König <[email protected]> writes:

> On Wed, May 23, 2018 at 11:56:21AM +0200, Esben Haabendal wrote:
>> From: Esben Haabendal <[email protected]>
>>
>> This fixes a race condition, where the DMAEN bit ends up being set after
>> I2C slave has transmitted a byte following the dummy read. When that
>> happens, an interrupt is generated instead, and no DMA request is generated
>> to kickstart the DMA read, and a timeout happens after DMA_TIMEOUT (1 sec).
>>
>> Fixed by setting the DMAEN bit before the dummy read.
>
> Does this fix the problem or just narrow the race window?

It should fix the problem.

/Esben