Date: Tue, 28 Jul 2009 15:50:03 GMT From: Eygene Ryabinkin <rea-fbsd@codelabs.ru> To: freebsd-usb@FreeBSD.org Subject: Re: usb/137189: [usb][patch] create and use sysctl nodes for HID report descriptors Message-ID: <200907281550.n6SFo3va084603@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR usb/137189; it has been noted by GNATS.
From: Eygene Ryabinkin <rea-fbsd@codelabs.ru>
To: bug-followup@FreeBSD.org, freebsd-usb@FreeBSD.org
Cc: hselasky@c2i.net
Subject: Re: usb/137189: [usb][patch] create and use sysctl nodes for HID
report descriptors
Date: Tue, 28 Jul 2009 19:40:27 +0400
HPS wrote:
> > I can, but how to relate the device name under devfs to the descriptor
> > to be retrieved?
>
> By unit number and device address, like usbconfig.
Yes, but these numbers are the ugen ones; on my system, for example,
ums0 maps to ugen5.3.0.
>> I can enumerate all HID descriptors via libusb, but I
> > can't currently say which device uses them.
>
> Try:
>
> usbconfig show_ifdrv
Yes, that's better. But can I rely that the first item in the output of
the libusb20_dev_get_iface_desc() will be the device name? And still,
to get the report descriptor from one device, I should walk the entire
USB tree until I'll find the device I am interested in. May be there
are some other ways? I understand that enumeration is the principle of
the USB bus, but may be there are another ways to directly grab 'struct
libusb20_device *' or alike for the devfs node I am interested in?
> > Why do you dislike the sysctl approach? It is simple and reliable.
>
> It's duplicating access to data. There is not that much wrong about
> it, except it will not work if the device is of another kind. I.E. you
> would have to patch the HID sysctl node into every driver accessing
> HID descriptors?
Just now -- yes, I'll need it. But probably I can move this
functionality into the USB bus level -- it will automatically create
this sysctl node for all HID children and will dispose it on the detach.
usb_probe_and_attach() is a candidate for such functionality. Will it
be bad?
--
Eygene
_ ___ _.--. #
\`.|\..----...-'` `-._.-'_.-'` # Remember that it is hard
/ ' ` , __.--' # to read the on-line manual
)/' _/ \ `-_, / # while single-stepping the kernel.
`-'" `"\_ ,_.-;_.-\_ ', fsc/as #
_.-'_./ {_.' ; / # -- FreeBSD Developers handbook
{_.-``-' {_/ #
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200907281550.n6SFo3va084603>