public inbox for linux-usb@vger.kernel.org
 help / color / mirror / Atom feed
From: Dan Scally <dan.scally@ideasonboard.com>
To: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Cc: linux-usb@vger.kernel.org, gregkh@linuxfoundation.org,
	mgr@pengutronix.de, kieran.bingham@ideasonboard.com
Subject: Re: [PATCH 3/3] usb: gadget: uvc: Allow disabling of interrupt endpoint
Date: Sun, 1 Jan 2023 20:46:14 +0000	[thread overview]
Message-ID: <9fccd14e-a6b9-66ce-6318-9ac8031d1e44@ideasonboard.com> (raw)
In-Reply-To: <Y6ulV6idkEeVgdXr@pendragon.ideasonboard.com>

Hi Laurent

On 28/12/2022 02:09, Laurent Pinchart wrote:
> Hi Dan,
>
> Thank you for the patch.
>
> On Mon, Dec 05, 2022 at 02:37:58PM +0000, Daniel Scally wrote:
>> The f_uvc code includes an interrupt endpoint against the VideoControl
>> interface. According to section 2.4.2 of the UVC specification however
>> this endpoint is optional in at least some cases:
>>
>> "This endpoint is optional, but may be mandatory under certain
>> conditions"
>>
>> The conditions enumerated are whether...
>>
>> 1. The device supports hardware triggers
>> 2. The device implements any AutoUpdate controls
>> 3. The device implements any Asynchronous controls
>>
>> As all of those things are implementation dependent, this endpoint
>> might be unnecessary for some users. Check whether the user has
>> requested it be disable via configfs and don't proceed with its
> s/disable/disabled/
>
>> instantiation if so.
>>
>> Signed-off-by: Daniel Scally <dan.scally@ideasonboard.com>
>> ---
>>   drivers/usb/gadget/function/f_uvc.c | 42 ++++++++++++++++++-----------
>>   drivers/usb/gadget/function/uvc.h   |  1 +
>>   2 files changed, 27 insertions(+), 16 deletions(-)
>>
>> diff --git a/drivers/usb/gadget/function/f_uvc.c b/drivers/usb/gadget/function/f_uvc.c
>> index 49b7231742d6..76ec84d3a5fd 100644
>> --- a/drivers/usb/gadget/function/f_uvc.c
>> +++ b/drivers/usb/gadget/function/f_uvc.c
>> @@ -79,7 +79,7 @@ static struct usb_interface_descriptor uvc_control_intf = {
>>   	.bDescriptorType	= USB_DT_INTERFACE,
>>   	.bInterfaceNumber	= UVC_INTF_VIDEO_CONTROL,
>>   	.bAlternateSetting	= 0,
>> -	.bNumEndpoints		= 1,
>> +	.bNumEndpoints		= 0,
>>   	.bInterfaceClass	= USB_CLASS_VIDEO,
>>   	.bInterfaceSubClass	= UVC_SC_VIDEOCONTROL,
>>   	.bInterfaceProtocol	= 0x00,
>> @@ -290,14 +290,16 @@ uvc_function_set_alt(struct usb_function *f, unsigned interface, unsigned alt)
>>   		if (alt)
>>   			return -EINVAL;
>>   
>> -		uvcg_info(f, "reset UVC interrupt endpoint\n");
>> -		usb_ep_disable(uvc->interrupt_ep);
>> +		if (!uvc->disable_interrupt_ep) {
>> +			uvcg_info(f, "reset UVC interrupt endpoint\n");
>> +			usb_ep_disable(uvc->interrupt_ep);
>>   
>> -		if (!uvc->interrupt_ep->desc)
>> -			if (config_ep_by_speed(cdev->gadget, f, uvc->interrupt_ep))
>> -				return -EINVAL;
>> +			if (!uvc->interrupt_ep->desc)
>> +				if (config_ep_by_speed(cdev->gadget, f, uvc->interrupt_ep))
> Line wrap would be nice.
>
> 				if (config_ep_by_speed(cdev->gadget, f,
> 						       uvc->interrupt_ep))
>
>> +					return -EINVAL;
>>   
>> -		usb_ep_enable(uvc->interrupt_ep);
>> +			usb_ep_enable(uvc->interrupt_ep);
>> +		}
>>   
>>   		if (uvc->state == UVC_STATE_DISCONNECTED) {
>>   			memset(&v4l2_event, 0, sizeof(v4l2_event));
>> @@ -375,7 +377,8 @@ uvc_function_disable(struct usb_function *f)
>>   	uvc->state = UVC_STATE_DISCONNECTED;
>>   
>>   	usb_ep_disable(uvc->video.ep);
>> -	usb_ep_disable(uvc->interrupt_ep);
>> +	if (!uvc->disable_interrupt_ep)
>> +		usb_ep_disable(uvc->interrupt_ep);
>>   }
>>   
>>   /* --------------------------------------------------------------------------
>> @@ -523,8 +526,10 @@ uvc_copy_descriptors(struct uvc_device *uvc, enum usb_device_speed speed)
>>   	control_size = 0;
>>   	streaming_size = 0;
>>   	bytes = uvc_iad.bLength + uvc_control_intf.bLength
>> -	      + uvc_interrupt_ep.bLength + uvc_control_cs_ep.bLength
>> -	      + uvc_streaming_intf_alt0.bLength;
>> +	       + uvc_control_cs_ep.bLength + uvc_streaming_intf_alt0.bLength;
> Wrong indentation.
>
>> +
>> +	if (!uvc->disable_interrupt_ep)
>> +		bytes += uvc_interrupt_ep.bLength;
>>   
>>   	if (speed == USB_SPEED_SUPER) {
>>   		bytes += uvc_ss_control_comp.bLength;
>> @@ -569,7 +574,8 @@ uvc_copy_descriptors(struct uvc_device *uvc, enum usb_device_speed speed)
>>   	uvc_control_header->bInCollection = 1;
>>   	uvc_control_header->baInterfaceNr[0] = uvc->streaming_intf;
>>   
>> -	UVC_COPY_DESCRIPTOR(mem, dst, &uvc_interrupt_ep);
>> +	if (!uvc->disable_interrupt_ep)
>> +		UVC_COPY_DESCRIPTOR(mem, dst, &uvc_interrupt_ep);
>>   	if (speed == USB_SPEED_SUPER)
>>   		UVC_COPY_DESCRIPTOR(mem, dst, &uvc_ss_control_comp);
> This is the companion descriptor for the interrupt endpoint, it should
> only be included if the interrupt endpoint is enabled. Same for
> uvc_control_cs_ep.
>
>>   
>> @@ -656,12 +662,16 @@ uvc_function_bind(struct usb_configuration *c, struct usb_function *f)
>>   			    (opts->streaming_maxburst + 1));
>>   
>>   	/* Allocate endpoints. */
>> -	ep = usb_ep_autoconfig(cdev->gadget, &uvc_interrupt_ep);
>> -	if (!ep) {
>> -		uvcg_info(f, "Unable to allocate control EP\n");
>> -		goto error;
>> +	if (!opts->disable_interrupt_ep) {
>> +		ep = usb_ep_autoconfig(cdev->gadget, &uvc_interrupt_ep);
>> +		if (!ep) {
>> +			uvcg_info(f, "Unable to allocate interrupt EP\n");
>> +			goto error;
>> +		}
>> +		uvc->interrupt_ep = ep;
>> +		uvc_control_intf.bNumEndpoints = 1;
>>   	}
>> -	uvc->interrupt_ep = ep;
>> +	uvc->disable_interrupt_ep = opts->disable_interrupt_ep;
> Given that you always test for !uvc->disable_interrupt_ep, how about
> naming this enable_interrupt_ep and switching the tests ? I think
> positive logic is more readable that double negation. I would then move
> this line up, in order to test uvc->enable_interrupt_ep instead of
> !opts->disable_interrupt_ep above.


Sure, that's fine. Should we still keep the default functionality (I.E. 
set enable_interrupt_ep=true by default)? The thing is...that endpoint 
doesn't actually work. There's no mechanism in the f_uvc module that 
allows userspace to send a packet through it. Adding one is in the 
pipeline now, but given it's unuseable anyway retaining backwards 
compatability doesn't seem valuable.

>>   
>>   	if (gadget_is_superspeed(c->cdev->gadget))
>>   		ep = usb_ep_autoconfig_ss(cdev->gadget, &uvc_ss_streaming_ep,
>> diff --git a/drivers/usb/gadget/function/uvc.h b/drivers/usb/gadget/function/uvc.h
>> index 48b71e04c2b1..0d0ef9b90b1a 100644
>> --- a/drivers/usb/gadget/function/uvc.h
>> +++ b/drivers/usb/gadget/function/uvc.h
>> @@ -149,6 +149,7 @@ struct uvc_device {
>>   	struct usb_ep *interrupt_ep;
>>   	struct usb_request *control_req;
>>   	void *control_buf;
>> +	bool disable_interrupt_ep;
>>   
>>   	unsigned int streaming_intf;
>>   

  reply	other threads:[~2023-01-01 20:46 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-12-05 14:37 [PATCH 0/3] Add ability to disable UVC Gadget's interrupt endpoint Daniel Scally
2022-12-05 14:37 ` [PATCH 1/3] usb: gadget: uvc: Rename uvc_control_ep Daniel Scally
2022-12-08 15:33   ` Greg KH
2022-12-08 15:39     ` Dan Scally
2022-12-28  1:59   ` Laurent Pinchart
2023-01-01 20:25     ` Dan Scally
2022-12-05 14:37 ` [PATCH 2/3] usb: gadget: uvc: Add new disable_interrupt_ep attribute Daniel Scally
2022-12-08 15:32   ` Greg KH
2022-12-28  2:02   ` Laurent Pinchart
2022-12-05 14:37 ` [PATCH 3/3] usb: gadget: uvc: Allow disabling of interrupt endpoint Daniel Scally
2022-12-28  2:09   ` Laurent Pinchart
2023-01-01 20:46     ` Dan Scally [this message]
2023-01-02 10:24       ` Laurent Pinchart

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=9fccd14e-a6b9-66ce-6318-9ac8031d1e44@ideasonboard.com \
    --to=dan.scally@ideasonboard.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=kieran.bingham@ideasonboard.com \
    --cc=laurent.pinchart@ideasonboard.com \
    --cc=linux-usb@vger.kernel.org \
    --cc=mgr@pengutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox