Hi Lothar,
it's been a while...
I've been working on the 1160 version of your 1362 driver for sometime
now. I'm getting very promising results :-)

Anyway, straight to the subject:

1) Maybe we should move DRIVER_AUTHOR etc to another file that is
implementation-specific (like ohci-xxx-regs.h), so that they inform of
the real maintainer of the driver. I don't like to move them out of the
main source file, but I can't think of something better...

2) You seem to initialize the values of RHDESCA and RHDESCB twice. Once
before calling hc_reset(ohci) but then the OHCI code does the same work
during hc_reset(ohci).

3) So, although it doesn't really matter, why do you init roothub port 1
as having "a device that is not removable?"

4) There is some kind of problem with setting the fminterval. I propose
to create an ohci_write_fminterval that also updates ohci->fminterval,
and call it before hc_reset(ohci), so that when hc_reset gets to state
"if(!ohci->fminterval){...}", it will not alter anything...

Enough with remarks/proposals. Let's get down to questions:

I) I'm able to communicate with the HC through it's regs, but get/set
descriptor functions seem to fail (silently). I'm not really getting
your plan here, so it's difficult to understand why. Could you tell me
where is the connection between ohci_urb_enqueue and your code? I mean
what is it that you do differently, and we end up translating ptds to
urbs? Or even better, can you give me a short list of function calls
when a get_config_descriptor happens?

II) Very general question, needing a very short reply: What is a pci
master device? how do I know if mine is one? an 'lspci -v' tells me in
the 'flags' field that it is a 'bus master' device...

-- 
Dimitris Lampridis <[EMAIL PROTECTED]>

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to