On Dec 19, 2018, at 21:09, Jiri Kosina [off-list ref] wrote:

On Fri, 14 Dec 2018, Kai-Heng Feng wrote:

quoted

While using Elan touchpads, the message floods:
[  136.138487] i2c_hid i2c-DELL08D6:00: i2c_hid_get_input: incomplete report (14/65535)

Though the message flood is annoying, the device it self works without
any issue. I suspect that the device in question takes too much time to
pull the IRQ back to high after I2C host has done reading its data.

Since the host receives all useful data, let's ignore the input report
when there's no data.

Interesting, never seen such a bug before.

quoted

Signed-off-by: Kai-Heng Feng <redacted>
---
drivers/hid/i2c-hid/i2c-hid-core.c | 9 +++++++++
1 file changed, 9 insertions(+)

diff --git a/drivers/hid/i2c-hid/i2c-hid-core.c b/drivers/hid/i2c-hid/i2c-hid-core.c
index 8555ce7e737b..1776afa6d69c 100644
--- a/drivers/hid/i2c-hid/i2c-hid-core.c
+++ b/drivers/hid/i2c-hid/i2c-hid-core.c

@@ -50,6 +50,7 @@

#define I2C_HID_QUIRK_NO_IRQ_AFTER_RESET	BIT(1)
#define I2C_HID_QUIRK_NO_RUNTIME_PM		BIT(2)
#define I2C_HID_QUIRK_DELAY_AFTER_SLEEP		BIT(3)
+#define I2C_HID_QUIRK_BOGUS_IRQ			BIT(4)

/* flags */
#define I2C_HID_STARTED		0

@@ -179,6 +180,8 @@ static const struct i2c_hid_quirks {

		I2C_HID_QUIRK_DELAY_AFTER_SLEEP },
	{ USB_VENDOR_ID_LG, I2C_DEVICE_ID_LG_8001,
		I2C_HID_QUIRK_NO_RUNTIME_PM },
+	{ USB_VENDOR_ID_ELAN, HID_ANY_ID,
+		 I2C_HID_QUIRK_BOGUS_IRQ },
	{ 0, 0 }
};

@@ -503,6 +506,12 @@ static void i2c_hid_get_input(struct i2c_hid *ihid)

		return;
	}

+	if (ihid->quirks & I2C_HID_QUIRK_BOGUS_IRQ && ret_size == 0xffff) {
+		dev_dbg(&ihid->client->dev,
+			"%s: IRQ triggered but there's no data\n", __func__);
+		return;
+	}

Would it perhaps make sense to make this some sort of printk_once(), so 
that it's immediately apparent from dmesg that the system/device is 
suffering from this particular problem? Might potentially be helpful piece 
of information.

Good advice, I’ll send a v2 patch to address this issue.

Kai-Heng

Thanks,

-- 
Jiri Kosina
SUSE Labs