USB_ENDPOINT_DESCRIPTOR oddness

From: Tom Udale (email_at_address.com)
Date: 06/15/04 

Date: Tue, 15 Jun 2004 14:58:01 GMT

I am trying to read the wMaxPacketSize from the USB_ENDPOINT_DESCRIPTOR
of a device so that I can set the MaximumTransferSize to an integral
multiple of the wMaxPacketSize when I set up the interface and I am
getting some odd values.

The device in question has 4 alternate interfaces, each with 2
endpoints - the default control endpoint and a high-speed ISO endpoint.
The USB_ENDPOINT_DESCRIPTOR values for the control endpoint all look
quite reasonable, but the ISO pipe values are kind of cooked. In
particular, for the interfaces that are supposed to be full and half
bandwidth, the reported value of wMaxPacketSize is 5120 (which I believe
should never be higher than 3072) while the default interface (which I
believe should be using no bandwidth) is 384.

Now, my question is, is the data in USB_ENDPOINT_DESCRIPTOR directly
copied from the hardware device (in which case I should be talking to
the hardware guys), or is it massaged alot by usbdi (in which case I
could be calling usbdi incorrectly or there could be some other bug)?

Any ideas?

As an additionaly bit of data, when I actually select the full bandwidth
interface, the packet size I get from the USBD_PIPE_INFORMATION is
indeed 3072.

If this is a hardware problem, I am kind of surprised that USBDI is not
whining more severely early on given that the hardware is reporting
invalid values.

Regards,

Tom Udale

Relevant Pages

  • Re: Newbie Modelling Interface Question
    ... A client's access to this subsystem is ... interface has a message identifier and a by-value data packet. ... The client then has a pointer to each interface. ... For example, in a hardware interface, clients usually want to manipulate ...
    (comp.object)
  • Re: OpenGL-based framebuffer concepts
    ... Agreed that kernel should only deal with necessary tasks as minimum as ... Designing the interface inevitably involves clear understanding of the ... hardware capabilities and closed hardware spec is an obvious obstacle. ... Open Graphics card would be a great thing ...
    (Linux-Kernel)
  • Re: Cubase SX or Pro-Tools M-Powered?
    ... electronic type composers, MIDI editing, sound stretching etc while ... besides a history lesson and marketing targets, they all employ a 'hyper-tape' type setup unifying audio, midi, and automation in one interface, but obviously going beyond the abilities of tape. ... ProTools would restrict your choice of hardware. ...
    (rec.audio.pro)
  • Re: [openib-general] Re: [PATCH][RFC][0/4] InfiniBand userspace verbs implementation
    ... The interface should be used ... > to map the hardware, there will always be opportunity for abuse. ... any faster than it would to invalidating a memory region. ... Instead it is allowed to block on such a request and only guarantees ...
    (Linux-Kernel)
  • Funded Ph.D. Position - Formalising the Hardware/Software Interface, TCD, Ireland
    ... Formalising the Interface between Software and Hardware ... Software Systems Laboratory, ... hardware and software, with particular emphasis on Flash Memory ...
    (comp.theory)