When doing i2cdetect quick write mode, we would get transfer
error ENOMEM, and i2cdetect shows there's no device at the address.
Quoting from mt8183 datasheet, the number of transfers to be
transferred in one transaction should be set to bigger than 1,
so we should forbid zero-length transfer and update functionality.
Incorrect return:
localhost ~ # i2cdetect -q -y 0
0 1 2 3 4 5 6 7 8 9 a b c d e f
00: -- -- -- -- -- -- -- -- -- -- -- -- --
10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
70: -- -- -- -- -- -- -- --
After this patch:
localhost ~ # i2cdetect -q -y 0
Error: Can't use SMBus Quick Write command on this bus
localhost ~ # i2cdetect -y 0
Warning: Can't use SMBus Quick Write command, will skip some addresses
0 1 2 3 4 5 6 7 8 9 a b c d e f
00:
10:
20:
30: -- -- -- -- -- -- -- --
40:
50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
60:
70:
Reported-by: Alexandru M Stan <[email protected]>
Signed-off-by: Hsin-Yi Wang <[email protected]>
---
Previous patch and discussion:
http://patchwork.ozlabs.org/patch/1042684/
---
drivers/i2c/busses/i2c-mt65xx.c | 13 +++++++++++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/drivers/i2c/busses/i2c-mt65xx.c b/drivers/i2c/busses/i2c-mt65xx.c
index 252edb433fdf..2842ca4b8c3b 100644
--- a/drivers/i2c/busses/i2c-mt65xx.c
+++ b/drivers/i2c/busses/i2c-mt65xx.c
@@ -234,6 +234,10 @@ static const struct i2c_adapter_quirks mt7622_i2c_quirks = {
.max_num_msgs = 255,
};
+static const struct i2c_adapter_quirks mt8183_i2c_quirks = {
+ .flags = I2C_AQ_NO_ZERO_LEN,
+};
+
static const struct mtk_i2c_compatible mt2712_compat = {
.regs = mt_i2c_regs_v1,
.pmic_i2c = 0,
@@ -298,6 +302,7 @@ static const struct mtk_i2c_compatible mt8173_compat = {
};
static const struct mtk_i2c_compatible mt8183_compat = {
+ .quirks = &mt8183_i2c_quirks,
.regs = mt_i2c_regs_v2,
.pmic_i2c = 0,
.dcm = 0,
@@ -870,7 +875,11 @@ static irqreturn_t mtk_i2c_irq(int irqno, void *dev_id)
static u32 mtk_i2c_functionality(struct i2c_adapter *adap)
{
- return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
+ if (adap->quirks->flags & I2C_AQ_NO_ZERO_LEN)
+ return I2C_FUNC_I2C |
+ (I2C_FUNC_SMBUS_EMUL & ~I2C_FUNC_SMBUS_QUICK);
+ else
+ return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
}
static const struct i2c_algorithm mtk_i2c_algorithm = {
@@ -933,8 +942,8 @@ static int mtk_i2c_probe(struct platform_device *pdev)
i2c->dev = &pdev->dev;
i2c->adap.dev.parent = &pdev->dev;
i2c->adap.owner = THIS_MODULE;
- i2c->adap.algo = &mtk_i2c_algorithm;
i2c->adap.quirks = i2c->dev_comp->quirks;
+ i2c->adap.algo = &mtk_i2c_algorithm;
i2c->adap.timeout = 2 * HZ;
i2c->adap.retries = 1;
--
2.20.1
On Thu, 2019-08-22 at 13:57 +0800, Hsin-Yi Wang wrote:
> When doing i2cdetect quick write mode, we would get transfer
> error ENOMEM, and i2cdetect shows there's no device at the address.
> Quoting from mt8183 datasheet, the number of transfers to be
> transferred in one transaction should be set to bigger than 1,
> so we should forbid zero-length transfer and update functionality.
<...>
> @@ -933,8 +942,8 @@ static int mtk_i2c_probe(struct platform_device *pdev)
> i2c->dev = &pdev->dev;
> i2c->adap.dev.parent = &pdev->dev;
> i2c->adap.owner = THIS_MODULE;
> - i2c->adap.algo = &mtk_i2c_algorithm;
> i2c->adap.quirks = i2c->dev_comp->quirks;
> + i2c->adap.algo = &mtk_i2c_algorithm;
> i2c->adap.timeout = 2 * HZ;
> i2c->adap.retries = 1;
>
Why do you need to change this part?
Joe.C
On Thu, Aug 22, 2019 at 5:24 PM Yingjoe Chen <[email protected]> wrote:
>
> Why do you need to change this part?
>
> Joe.C
>
You're right. We don't need to change this. I'll revert this back.