Re: [PATCH 1/2] HID: uclogic: Store firmware name in params

["José Expósito" <jose.exposito89@gmail.com>]

Hi Benjamin,

Thanks a lot for the quick review :)

On Thu, Mar 21, 2024 at 04:01:06PM +0100, Benjamin Tissoires wrote:
> 
> Hi José,
> 
> On Mar 21 2024, José Expósito wrote:
> > A future patch will need to access the firmware name to expose it to
> > userspace via sysfs.
> > 
> > Store it in `uclogic_params->fw_name`.
> > 
> > Signed-off-by: José Expósito <jose.exposito89@gmail.com>
> > ---
> >  drivers/hid/hid-uclogic-params.c | 14 +++++++-------
> >  drivers/hid/hid-uclogic-params.h |  5 +++++
> >  2 files changed, 12 insertions(+), 7 deletions(-)
> > 
> > diff --git a/drivers/hid/hid-uclogic-params.c b/drivers/hid/hid-uclogic-params.c
> > index 9859dad36495..79ec3eb80f84 100644
> > --- a/drivers/hid/hid-uclogic-params.c
> > +++ b/drivers/hid/hid-uclogic-params.c
> > @@ -121,6 +121,7 @@ void uclogic_params_hid_dbg(const struct hid_device *hdev,
> >  		params->invalid ? "true" : "false");
> >  	hid_dbg(hdev, ".desc_ptr = %p\n", params->desc_ptr);
> >  	hid_dbg(hdev, ".desc_size = %u\n", params->desc_size);
> > +	hid_dbg(hdev, ".fw_name = %s\n", params->fw_name);
> >  	hid_dbg(hdev, ".pen = {\n");
> >  	uclogic_params_pen_hid_dbg(hdev, &params->pen);
> >  	hid_dbg(hdev, "\t}\n");
> > @@ -652,6 +653,7 @@ void uclogic_params_cleanup(struct uclogic_params *params)
> >  	if (!params->invalid) {
> >  		size_t i;
> >  		kfree(params->desc_ptr);
> > +		kfree(params->fw_name);
> >  		uclogic_params_pen_cleanup(&params->pen);
> >  		for (i = 0; i < ARRAY_SIZE(params->frame_list); i++)
> >  			uclogic_params_frame_cleanup(&params->frame_list[i]);
> > @@ -837,7 +839,6 @@ static int uclogic_params_huion_init(struct uclogic_params *params,
> >  	/* The resulting parameters (noop) */
> >  	struct uclogic_params p = {0, };
> >  	static const char transition_ver[] = "HUION_T153_160607";
> > -	char *ver_ptr = NULL;
> >  	const size_t ver_len = sizeof(transition_ver) + 1;
> >  	__u8 *params_ptr = NULL;
> >  	size_t params_len = 0;
> > @@ -870,14 +871,14 @@ static int uclogic_params_huion_init(struct uclogic_params *params,
> >  	}
> >  
> >  	/* Try to get firmware version */
> > -	ver_ptr = kzalloc(ver_len, GFP_KERNEL);
> > -	if (ver_ptr == NULL) {
> > +	p.fw_name = kzalloc(ver_len, GFP_KERNEL);
> > +	if (!p.fw_name) {
> >  		rc = -ENOMEM;
> >  		goto cleanup;
> >  	}
> > -	rc = usb_string(udev, 201, ver_ptr, ver_len);
> > +	rc = usb_string(udev, 201, p.fw_name, ver_len);
> >  	if (rc == -EPIPE) {
> > -		*ver_ptr = '\0';
> > +		*p.fw_name = '\0';
> >  	} else if (rc < 0) {
> >  		hid_err(hdev,
> >  			"failed retrieving Huion firmware version: %d\n", rc);
> > @@ -885,7 +886,7 @@ static int uclogic_params_huion_init(struct uclogic_params *params,
> >  	}
> >  
> >  	/* If this is a transition firmware */
> > -	if (strcmp(ver_ptr, transition_ver) == 0) {
> > +	if (strcmp(p.fw_name, transition_ver) == 0) {
> >  		hid_dbg(hdev,
> >  			"transition firmware detected, not probing pen v2 parameters\n");
> >  	} else {
> > @@ -1028,7 +1029,6 @@ static int uclogic_params_huion_init(struct uclogic_params *params,
> >  	rc = 0;
> >  cleanup:
> >  	kfree(params_ptr);
> > -	kfree(ver_ptr);
> >  	uclogic_params_cleanup(&p);
> >  	return rc;
> >  }
> > diff --git a/drivers/hid/hid-uclogic-params.h b/drivers/hid/hid-uclogic-params.h
> > index d6ffadb2f601..412c916770f5 100644
> > --- a/drivers/hid/hid-uclogic-params.h
> > +++ b/drivers/hid/hid-uclogic-params.h
> > @@ -232,6 +232,11 @@ struct uclogic_params {
> >  	 * List of event hooks.
> >  	 */
> >  	struct uclogic_raw_event_hook *event_hooks;
> > +	/*
> > +	 * Firmware name, exposed to userspace via sysfs as it is used to
> > +	 * identify the tablet.
> > +	 */
> > +	char *fw_name;
> 
> I can't remember if this was on the table or not, but any reasons to not
> use hid->uniq[64] field instead of a custom sysfs?
> If there is already a value, we could just append the firmware version
> with a colon (:) separator.
> 
> The main reason would be to not export a new sysfs specifically for this
> driver, but it would also allow HID-BPF to have access to the value and
> thus allow to detect if the device works with the given bpf program.

That's a very good point.

I tested with some tablets and for HUION, the uniq field is empty.
Other vendors just put garbage there. In fact, we override it in
tablets with battery to avoid a crash [1].

So, it'll be fine to just override the garbage in that field with
an actual unique ID. Sent v2 with the sugested change [2].

Thanks,
Jose

[1] https://github.com/JoseExposito/linux/commit/9845a1304a06805c3193c58283ca428c32c14fa7#diff-b3e9836cea703e50e8febb1f09b86f92833200e7bda4c9c1ea5a1d8b569b01dbR1257-R1262
[2] https://lore.kernel.org/linux-input/20240322100210.107152-1-jose.exposito89@gmail.com/T/

> Cheers,
> Benjamin
> 
> >  };
> >  
> >  /* Driver data */
> > -- 
> > 2.44.0
> >