[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: USB-backport 2.3.x for 2.2.14 in potato

On Thu, Apr 06, 2000 at 07:49:04PM -0500, Adam Keys wrote:
> URL?
> How stable is this code?  Are there any major issues known?  What percentage
> of the hardware supported by 2.3.x is supported by the backport?  What all
> resides in userspace?  I think these are the questions relevant to getting
> such a major subsystem included into a deeply frozen distribution.
  Stability depends on the device.  Keyboards/mice/joysticks(HID), boring
stuff, has been stable for quite some time.  All three host controler
drivers are stable enough for basic use.
 As for specific drivers, here's my take on them.
 * acm(modems, ISDN TA's, etc): should be useable by most but still has a few
minor issues
 * audio: works but does not support the full USB audio class. Fails to
recognize some devices as USB audio.  Doesn't attempt to implement the midi
end, or probably anything but basic PCM audio.
 * cpia: For CPiA based USB video cameras.  Was pretty stable but code was
recently merged with driver for parallel port models, so it's a little
rough still.
 * dabusb: Digital audio broadcast. I assume it's function, but as far as I
know only the developers of the driver have the hardware.
 * dc2xx: Kodac dc2xx still camera.  Works with many cameras and I haven't
noticed any "it's broken" messages in a while. I believe there is talk of
moving this to user space.
 * dsbr100: D-Link USB radio control.  Appears function, actual audio is
handled by the audio driver.  In kernel so it can register itself as a
video4linux tuner.
 * hub: USB hubs.  There may be a few issues, but by and large functional.
 * hid: Human Interface Devices, ie anything you can touch.  Keybaords,
mice, joysticks, control buttons and knobs.  Works great. All special
button might not have appropriate maps yet.
 * ibmcam: IBM USB videocameras.  A little twitchy, but it might be the
hardware.  It works with some RGB-BGR issues, I'm not sure if the driver or
my video is at fault.
 * kaweth: KLSI based USB ethernet.  Hasn't been merged with the kernel tree
yet, but functions on OHCI controlers with ~11% packet loss.
 * mdc8000: Mustek MDC800 Digital Camera.  Kernel part of the driver,
requires gPhoto support.  Can probably be moved to user space later.
 * ov511: OV511 based USB webcams: reported good, with minor issues.
 * pegasus: Pegasus based USB 10/100 ethernet & PNA: reported to work.(No,
use can't handle 100Mbit, but this won't force a 100Mbit hub down to 10)
 * plusb: Prolific pl-2302 based USB ethernet: Also reported to work.
 * printer: Handles USB printers, shouldn't be in the backports, however,
because of the scale of the parport changes in 2.3
 * rio500: Diamond Rio500: communicates with Rio MP3 players.  There is talk
of moving this into the userspace utility once it stablizes a bit.
 * serial: USB->serial converters, including Handspring Visors, ConnectTech
Whiteheat converters, KTDI converters, and the begining of keyspan PDA's. 
This driver has issues with OHCI controlers, but works great with my Visor
on UHCI.
 * scanner: handles various and assundry USB scanners(HP, Agfa, Umax,
Seiko/Epon, Mustek, Microtek). Appears to require user space support as
well, might be posible to come completly to user space in the future.
 * uss720: USS720 based USB-Parallel cables.  Looks good, but too tied to
the 2.3 parport code to backport.
 * wacom: Wacom USB graphics tablets.  Functional for some tablets, but
there isn't the proper application support yet to use it properly, only as
an emulated mouse.
 * wmforce: Logitech Wingman Force: No force feedback yet, but supposed to
work as a joystick.

  Well, that's my take on the state of the USB code.  At the very least I
think we should have an experimental usb kernel image for legacy free
systems.  A normal kernel image, but with a driver for both host controlers,
and, keyboads, mice, and maybe ethernet's.  I can put together the patched
image, if it would be accepted. I'd need a sponsor.
  Oh, there have been some backport crash bugs reported recently, but they
appear not to be too difficult to remedy.

  - Nick Lopez

Reply to: