On Thu, Jul 16, 2020 at 06:27:19PM -0700, Dmitry Torokhov wrote:
> Hi Jingle,
>
> On Thu, Jul 16, 2020 at 02:15:23PM +0800, jingle.wu wrote:
> > HI Dmitry:
> >
> > Just to confirm, the older devices (I assume that pattern 0 means older)
> > have version command that is numerically higher than the one for the
> > newer (pattern >= 1) devices?
> >
> > >> Yes, Pattern 1, 2 are newer devices.
> >
> > > @@ -324,7 +342,14 @@ static int elan_i2c_get_sm_version(struct i2c_client *client,
> > > return error;
> > > }
> > > *version = val[0];
> > > - *ic_type = val[1];
> > > +
> > > + error = elan_i2c_read_cmd(client, ETP_I2C_IAP_VERSION_CMD, val);
> > > + if (error) {
> > > + dev_err(&client->dev, "failed to get ic type: %d\n",
> > > + error);
> > > + return error;
> > > + }
> >
> > Could you please tell me why this chunk is needed?
> > >> Modify the old pattern IC firmware read the correct ic_type.
> >
> > In the elan_i2c_core.c, move this code to elan_i2c_i2c.c.
> > static int elan_query_device_info(struct elan_tp_data *data)
> > {
> > .....
> > if (data->pattern == 0x01)
> > ic_type = data->ic_type;
> > else
> > ic_type = data->iap_version;
> > .....
> > return 0;
> > }
>
> I am concerned that unconditionally substituting iap_version for ic_type
> for "pattern 0" devices will break check in
> elan_check_ASUS_special_fw() as it operates on the ic_type returned by
> ETP_I2C_OSM_VERSION_CMD and not iap_version.
I split the firmware handling code into a few patches and uploaded it
to a new elan-i2c branch:
https://git.kernel.org/pub/scm/linux/kernel/git/dtor/input.git elan-i2c
Please take a look and let me know if I messed it up or not. I will be
looking at the new packet format next.
Thanks.
--
Dmitry
