- remove items that are fixed (usually in the distant past)
- usbnet is (almost) done for non wifi
- non trivial usbdi documentation exists now

G/C

switch urndis(4) to usbnet. thanks to maya@ for testing and helping
fix the few issues in the conversion.

update ethernet driver notes: mark udav as done, and every thing
except umb(4) as patch available, with urndis probably working
(need to test the latest change before being commited.)

still need testers for: aue, cue, kue, mue, upl and url.
no umb patch yet.

Note that smsc(4) is done

add a list of done/todo items to the shared ethernet driver code.

branches: 1.46.2;
Remove an item that doesn't need to happen

branches: 1.45.2; 1.45.4; 1.45.10; 1.45.28;
Remove a couple of items

branches: 1.44.2;
Remove a couple of item that have been done.

branches: 1.43.2;
Remove a couple of items.

branches: 1.42.126; 1.42.138;
Remove an item - uhidev(4) now exists.

OK'd by Lennart.

branches: 1.41.16; 1.41.24; 1.41.26;
Add an item

branches: 1.40.2; 1.40.4;
Implement what in Intel-speech is known as "bandwidth reclamation".
It means that we continously poll USB devices that have a pending transfer
instead of polling just once every ms. This speeds up some transfers
at the expense of using more PCI bandwidth.

Take away some, add some.

Remove item.

Add an item.

Add one, remove some.

Add more items and reorganize.

Add some, remove some.

Take off an item.

The list is growing...

Remove one, add one.

Rearrange to indicate what should be done before 1.5

Add another item.

Remove some items!

Remove an item.

Add another item.

Sync with reality.

Add another item.

Cosmetics and a couple of diagnostic messages.

A number of stylistic changes to increase readability (many suggested
by Nick Hibma):
use NULL not 0
declare all local definitions static
rename s/usbd_request/usbd_xfer/ s/reqh/xfer/
rename s/r/err/
use implicit test for no err
KNF

Update.

branches: 1.20.2; 1.20.4; 1.20.6;
Make sure the HC deactivation is propagated.

Remove one, add none!

Make sure timeouts count as interrupt context too.

Rearrange the code a little so we can decide if we are in process
or interrupt context in a reliable way. Mainly used for DIAGNOSTIC.

Add a flag in the request to determine if the data copying is done by the
driver or the usbdi layer.

* Move DMA buffer allocation to HC independent code.
* Remove (almost) unused definitions USBD_XFER_OUT and USBD_XFER_IN.

Remove some and add some.

Do a brain dump of items to do.

One item off this list done!

Add to the growing list.

Make sure to mark the device as dying already in the (de)activate routine.
This avoids access to it before the detach routine has blown it away.

Merge the soft_{td,ed} with the real {td,ed}. This saves time and space.
(And fix typo in corresponding uhci change.)

More things to do...

Add some more items.

Remove some, add some.

Some new items.

Remove two items that are done.

More things to do.

branches: 1.2.4;
More things to do.

Add a (partial) list of what's left to do.