The patch ensures that uhci.c doesn't use urb->status after the
completion callback if it doesn't need to.

fix the "small packet" problem and debug messages cleanup

removed CONFIG_USB_HIDDEV #ifdefs in the driver.

This patch against 2.5.7 fixes an oops and a memleak in the konicawc driver
and also adds an option to set the FPS.

updated the version number due the previous changes

  
  - Nitpickey bugfix to root hub config descriptors ... can't use
    the same one for high and full speed, since the encoding
    is different (255 ms FS == 0xff, 256 ms HS == 0x12).
  - Related, force period to 1/4 second rather than doing
    any sanity checking for the roothub timer (from Georg)
  - Don't "giveback" urbs on submit path errors (from Georg)
    ... means they don't get completion callbacks
  - Additional error checks on URB data (from Georg)
  - Uses <linux/completion.h> for unlink synchronization
  - The "already unlinking" error case is reported like other
    unlinking errors (not as success)
  - Ripped out urb->next handling ... it wasn't compatible
    with the ISO loop model, and at this point I believe it
    should be completely replaced with queuing urbs inside
    of the HCDs.  (Every HCD handles it for ISO, UHCI needs
    a magic flag to enable it for bulk ...)

  
- Makes "audio" set the urb->interval in its periodic URBs,
  so it can submit through the hcd framework. 

Module count of a serial converter driver is currently not
decremented if a disconnect happens while the tty is held open.
The fix is to close the device in usb_serial_disconnect() so that
module refcounts are properly updated.

fixes to let io_edgeport work properly as a usb serial console

use proper GFP flag setting for submitting a urb.

- make serial console only selectable if the code is not a module.
- removed lots of CONFIG_EXPERMENTAL dependancies as the drivers have been stable for some time now.

  
  Periodically folk have run into problems where usb-ohci oopses
  due to device refcount bugs ...
    
  This is a minor patch to move the sanity check out of usb-ohci
  into the generic bits of usbcore.    There are comments that
  suggest a path for a more comprehensive approach too.
    
  Applies cleanly against 2.5.7 and I've been testing with it
  for a while.  I can't think of any reason it shouldn't also go
  into 2.4, beyond the patch not applying cleanly there ... :)

  
      - bugfix: control endpoints can't stall
      - bugfix: remove bogus intr unlink optimization,
          by sharing intr/iso code
      - bugfix: iso submit uses urb->interval
      - removed iso urb->next ring logic
          (belongs in hcd layer if anywhere)
      - simplify/shorten/correct completion handling
      - in debug, labels setup packets as such
      - bring CVS ids back up to date

The issue was that the poisoning would posing the data *before* it gave
it to us as well, but I didn't clear out a pointer (qh->urbp) and later
dereferenced it.

Thanks for the report again and for the pci_pool patch so I could
troubleshoot it.

JE

semaphore cleanup and proper link detection

It updates the "ohci-hcd" driver to address two issues, one            
of which could potentially have caused oopses.  First, it takes
out calls to usb_dec_dev_use() on the submit error path; that's
already done elsewhere.  (Noted by Georg Acher.)  Second,
it removes some pointless diagnostics.  One is for schedule
overrun interrupts, the other is for accessing the frame counter.

Don't know if you saw this on the list or not. Here's a patch which
fixes zero packet support for uhci.c

- Setup the TD correctly, we weren't shifting the null data size
- Only add the zero packet if the transfer was a multiple of the
  endpoint of the size

JE

  
added support for the Palm i705 device.
thanks to Thomas Riemer for the information.

I took some time to clean up the code a little, and to add the new
calls to the documentation.  See patch below.

--
Paul

added rtl8150 usb ethernet driver

fix problem which cause hotplug/unplug crash the kernel

added optus@home uep1045a driver to the list of supported devices.

changes due to open_count being handled by the usb-serial core code.

- Moved all manipulation of port->open_count into the core.  Now the
  individual driver's open and close functions are called only when the
  first open() and last close() is called.  Making the drivers a bit
  smaller and simpler.
- Fixed a bug if a driver didn't have the owner field set.

Added support for the Palm m130 device, thanks to Udo Eisenbarth
<udo.eisenbarth@web.de> for the information.

into kroah.com:/home/greg/linux/BK/work

changed the interval for checking if the device is connected yet or not.
Thanks to Itai Nahshon <nahshon@actcom.co.il> for the information.

Port changes from 2.4:
  We do request_sense ourselves to comply with
  the scanner command set

documented the fact that the interval is not always reported in ms.

This is an updated version of a patch I sent around a
while back.  It's against 2.5.7-pre1 (so presumably is
fine on 2.5.7), and addresses feedback against that
earlier patch.
  
It's bugfixes, mostly for highspeed support, to what
/proc/bus/usb/devices shows:
  
- Shows isochronous periods correctly (logarithmic
  encoding, possibly 1/2/4 microframes if highspeed)
- Likewise for high-speed interrupt periods (similar)
- Makes high bandwidth endpoints look like they
  just do bigger packets (up to 3 KBytes/uframe)
- Shows highspeed bandwidth correctlly (80% reserved,
  vs 90% reserved for full/low speed).

Don't submit urbs while holding spinlocks. Not strictly required in
2.5.x, but it's always better to do less while holding a spinlock.
Also a good idea to keep 2.{4,5}.x drivers in sync.

Unfortunately, I left out one line from my spinlock cleanup patch
recently.

As a result, using interrupt URB's could cause a deadlock on SMP
kernels.

This should fix the deadlock Greg reported.

JE

uhci.c would call the completion callback when the call to submit_urb
failed. This is a rare situation.

This patch only calls the completion handler if the URB successfully
completed immediately (as in the case of talking to the virtual root
hub).

JE

My previous patch which cleaned up some of the spinlocks, moved one of
the spinlocks around a call to kmem_cache_alloc. It would sometimes
erroneously call it with GFP_KERNEL.

This patch fixes the problem by always calling it with GFP_ATOMIC.
Thanks to Greg for pointing this out to me.

JE

    - refer to "usbfs"
    - describe the /proc/bus/usb/BBB/DDD files
    - more info about the .../drivers and .../devices
    - ... generally, gives more information.
  
This is ever so slightly forward looking in how it describes
bandwidth requirements for high speed periodic transfers,
it's expecting a bugfix patch that's in my queue.  (That info            
is currently broken/meaningless.)

  
This updates the Philips tweak so that it also applies
to late-model ICH chips from Intel.  (Or so I'm told ... :)
That's at least three EHCI implementations known to
behave on Linux.  (And one hopes VIA soon too...)

It also cleans up a few comments relating to 64bit DMA;
recent API spec updates make it look like no games are
needed with the PCI DMA mask -- it doesn't change the
segment used by pci_pool allocations, so it just needs
to get turned on if the overall system does 64bit DMA.

I've a query in to see if those Philips/Intel tweaks are
chip quirks, or just something the EHCI 1.0 spec isn't
at all clear about.

inline cleanup to save a chunk of memory in usb-ohci

Comment and documentation cleanups

This patch completely punts on passing SLAB_POISON, I've
gotten burnt by it once too many.  Seems like the slab code
changed somewhere.  I've got a separate patch to make the
pci_pool code use CONFIG_DEBUG_SLAB, which I'll send
around.
    
Meanwhile, I needed this to get at least EHCI to intialize on
a 2.5.7-pre1 system.  Please merge.