i2c: core: ACPI: Properly set status byte to 0 for multi-byte writes
authorHans de Goede <hdegoede@redhat.com>
Sun, 12 Aug 2018 10:53:20 +0000 (12:53 +0200)
committerWolfram Sang <wsa@the-dreams.de>
Mon, 20 Aug 2018 12:08:59 +0000 (14:08 +0200)
acpi_gsb_i2c_write_bytes() returns i2c_transfer()'s return value, which
is the number of transfers executed on success, so 1.

The ACPI code expects us to store 0 in gsb->status for success, not 1.

Specifically this breaks the following code in the Thinkpad 8 DSDT:

            ECWR = I2CW = ECWR /* \_SB_.I2C1.BAT0.ECWR */
            If ((ECST == Zero))
            {
                ECRD = I2CR /* \_SB_.I2C1.I2CR */
            }

Before this commit we set ECST to 1, causing the read to never happen
breaking battery monitoring on the Thinkpad 8.

This commit makes acpi_gsb_i2c_write_bytes() return 0 when i2c_transfer()
returns 1, so the single write transfer completed successfully, and
makes it return -EIO on for other (unexpected) return values >= 0.

Cc: stable@vger.kernel.org
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
drivers/i2c/i2c-core-acpi.c

index 7c3b4740b94b644509ae3658fa2eda4885086bba..b8f303dea30588b3ee779fe0d4df0c828996a2a3 100644 (file)
@@ -482,11 +482,16 @@ static int acpi_gsb_i2c_write_bytes(struct i2c_client *client,
        msgs[0].buf = buffer;
 
        ret = i2c_transfer(client->adapter, msgs, ARRAY_SIZE(msgs));
-       if (ret < 0)
-               dev_err(&client->adapter->dev, "i2c write failed\n");
 
        kfree(buffer);
-       return ret;
+
+       if (ret < 0) {
+               dev_err(&client->adapter->dev, "i2c write failed: %d\n", ret);
+               return ret;
+       }
+
+       /* 1 transfer must have completed successfully */
+       return (ret == 1) ? 0 : -EIO;
 }
 
 static acpi_status