plunky@rya-online.net said:
> Do you have any other USB hardware running through the ehci driver?

I've tried a mouse and an external harddisk, both still work.

> I'm a
> bit concerned that it looks as if this would affect all ehci transfers, so
> surely there would have been other failures reported..?

The change does only affect "short transfers", this is where the
length of the read transfer isn't known in advance. I didn't check
all the drivers in detail, but usually the length is known -- eg
the report size for HIDs or the transfer size for mass storage.
In addition, the bug causing the hangs did only strike if the
difference between the attempted and the real transfer crossed
a blocksize boundary, which at least requires that the attempted
transfer is larger than the blocksize.
In addition, full/low speed devices do only touch ehci if there
is a high-speed hub in between.
This all makes it prety plausible for me that there were no
bug reports before.

I'll just commit that change; it is small enough to have it
backed out if serious problems show up.

best regards
Matthias