Currently acpi_gsb_i2c_read_bytes() directly returns i2c_transfer's return
value. i2c_transfer returns a value < 0 on error and 2 (for 2 successfully
executed transfers) on success. But the ACPI code expects 0 on success, so
currently acpi_gsb_i2c_read_bytes()'s caller does:
if (status > 0)
status = 0;
This commit makes acpi_gsb_i2c_read_bytes() return a value which can be
directly consumed by the ACPI code, mirroring acpi_gsb_i2c_write_bytes(),
this commit also makes acpi_gsb_i2c_read_bytes() explitcly check that
i2c_transfer returns 2, rather then accepting any value > 0.
Signed-off-by: Hans de Goede <hdegoede@xxxxxxxxxx>
---
Changes in v2:
-New patch in v2 of this patch-set
---
drivers/i2c/i2c-core-acpi.c | 8 +++++---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/i2c/i2c-core-acpi.c b/drivers/i2c/i2c-core-acpi.c
index b8f303dea305..32affd3fa8bd 100644
--- a/drivers/i2c/i2c-core-acpi.c
+++ b/drivers/i2c/i2c-core-acpi.c
@@ -453,8 +453,12 @@ static int acpi_gsb_i2c_read_bytes(struct i2c_client *client,
else
dev_err(&client->adapter->dev, "i2c read %d bytes from client@%#x starting at reg %#x failed, error: %d\n",
data_len, client->addr, cmd, ret);
- } else {
+ /* 2 transfers must have completed successfully */
+ } else if (ret == 2) {
memcpy(data, buffer, data_len);
+ ret = 0;
+ } else {
+ ret = -EIO;
}
kfree(buffer);
@@ -595,8 +599,6 @@ i2c_acpi_space_handler(u32 function, acpi_physical_address command,
if (action == ACPI_READ) {
status = acpi_gsb_i2c_read_bytes(client, command,
gsb->data, info->access_length);
- if (status > 0)
- status = 0;
} else {
status = acpi_gsb_i2c_write_bytes(client, command,
gsb->data, info->access_length);
--
2.18.0
