Re: Redistribution of Hotfix KB943198

chris.aseltine@xxxxxxxxx wrote:

On Jul 11, 11:42 pm, Tim Roberts <t...@xxxxxxxxx> wrote:

When a single transfer is an exact multiple of the endpoint's packet size,
that is EXACTLY what the spec says. Without it, the transfer is considered
to be still in progress, hence the performance issue.

I'm unconvinced. The language in the spec refers to an "endpoint",
i.e. that means to me on the device side only: (section 5.8.3 of the
1.1 spec)

"A bulk transfer is complete when the *endpoint* (emphasis mine) does
one of the following:

- Has transferred exactly the amount of data expected
- Transfers a packet with a payload size less than wMaxPacketSize or
transfers a zero-length packet."

It's the same in both directions. Think about it logically. How else
would you ever signal to the device that a transfer is complete?
--
Tim Roberts, timr@xxxxxxxxx
Providenza & Boekelheide, Inc.
.

Relevant Pages

  • Re: USB GetDescriptor(Configuration) Error?
    ... And I'm not sure how I developed my false assumption about needing to send a zero length packet if the transfer length falls on the packet length boundary, since quite obviously you are correct about the USB 2 spec, and the USB 1.1 spec contains the same description of the process. ... I am still curious, however, when I watch bus traffic with an analyzer, the initial enumeration does involve a zero length data packet before the host begins the status stage, whereas the request my driver places does not. ...
    (microsoft.public.development.device.drivers)
  • Re: Ambiguities in TCP/IP - firewall bypassing
    ... > Identify what the packet should be, ... > the correct way to handle these packets, then these stacks are correct. ... As I said, the spec is ambiguous. ... The spec is silent on the use of ambiguous or contradictory flags ...
    (Bugtraq)
  • Re: Redistribution of Hotfix KB943198
    ... to be still in progress, ... The language in the spec refers to an "endpoint", ... transfers a zero-length packet." ...
    (microsoft.public.development.device.drivers)
  • Re: Bogon IPs traffic only seen by netflow, confined within a VLAN only
    ... things were caused by "jabbering" transceivers that would start transmitting ... their packet sooner than the spec allowed, resulting in the first few bytes ...
    (Incidents)
  • Ethernet packet size and QinQ standard?
    ... For 802.1Q tagging, the spec was changed to allow a max packet size ... I'm kind of surprized because some hardware does have the packet size ...
    (comp.dcom.lans.ethernet)