On Wednesday 03 July 2002 23:00, David Brownell wrote:
> > ... my modem returns 6 bytes
> > (out of 16) when the ADSL line goes up or down.  With no transfer
> > flags set (transfer_flags = 0, the default) this is OK.  With
> > USB_DISABLE_SPD set I get an error.
>
> Sigh, you're right.  That's really not the default behavior
> I'd have expected ... that's what comes from bad naming.
>
> I'm going to prepare a 2.5 patch that updates documentation (the
> <linux/usb.h> doc says otherwise) and changes the name of that
> flag so it's not cryptic.  I'm thinking SHORT_NOT_OK would
> be short enough to be typable, and impossible to misinterpret.

How about NO_SHORT?

> p.s. So a "short packet detect" flag isn't really the same
>      as a "short packet disable" flag would have been.

Well, it was still unclear to me what "disable short packet detect"
was supposed to mean.  Interpretation (1): if there is to be an error
on short packets, the short packets need to be detected; thus
by turning off "short packet detect" there will be no errors on
short packets, thus "disable short packet detect" = SHORT_OK.
(2): by detecting that a packet is short, we can clear the error
the short packet generated; thus turning off "short packet detect"
means that we will get errors for short packets, i.e.
"disable short packet detect" = NO_SHORT.

It seems (2) is right, but who would have guessed?  (Answer: someone
who has read the uhci spec.)

Ciao,

Duncan.


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
No, I will not fix your computer.
http://thinkgeek.com/sf
_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to