libusb-devel Mailing List for libusb

On Thu, Jan 29, 2004, Nagarajan, Haripriyax <har...@in...> wrote:
> Is there any thought on support for USB 2.0? It is clear from the
> documentation that libusb is geared towards USB 1.1. But will there be
> any modifications made to LibUSB to accommodate the higher bulk transfer
> rate supported to USB 2.0?

There are no changes required to support the higher bulk transfer rates
in USB 2.0

Try for yourself.

JE

However, the speeds will still be in the multi-megabyte range which is
substantially faster than USB 1.1.

The problem is mostly in the kernel. If you look at the code in devio.c,
function proc_bulk, you can see that the limit is one page (4k under
x86):

	len1 = bulk.len;
	if (len1 > PAGE_SIZE)
		return -EINVAL;

Looking at the URB interface, we can do up to 16k per call, which might
be worth the change.

JE

On Thu, Jan 29, 2004, Nagarajan, Haripriyax <har...@in...> wrote:
> Below is the cut,paste from the Linux kernel 2.4.18 documentation.
> 
> "In typical situations, a usb_bulk_msg() loop writing out 4 KB chunks is
> going to waste more than half the USB 2.0 bandwidth.  Delays between the
> I/O completion and the driver issuing the next request will take longer
> than the I/O.  If that same loop used 16 KB chunks, it'd be better; a
> sequence of 128 KB chunks would waste a lot less"
> 
> And since I saw the usb_bulk_read and usb_bulk_write functions in LIBUSB
> to be limiting the maximum bulk.len to 4KB, I assumed some changes might
> be required. I'll try it out.
> 
> Thanks!
> 
> -----Original Message-----
> From: Johannes Erdfelt [mailto:joh...@er...]
> Sent: Thursday, January 29, 2004 2:07 PM
> To: Nagarajan, Haripriyax
> Cc: lib...@li...
> Subject: Re: USB 2.0 support
> 
> 
> On Thu, Jan 29, 2004, Nagarajan, Haripriyax
> <har...@in...> wrote:
> > Is there any thought on support for USB 2.0? It is clear from the
> > documentation that libusb is geared towards USB 1.1. But will there be
> > any modifications made to LibUSB to accommodate the higher bulk
> transfer
> > rate supported to USB 2.0?
> 
> There are no changes required to support the higher bulk transfer rates
> in USB 2.0
> 
> Try for yourself.
> 
> JE
> 
> 
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> Libusb-devel mailing list
> Lib...@li...
> https://lists.sourceforge.net/lists/listinfo/libusb-devel
> 

Hi All,

Is there any thought on support for USB 2.0? It is clear from the
documentation that libusb is geared towards USB 1.1. But will there be
any modifications made to LibUSB to accommodate the higher bulk transfer
rate supported to USB 2.0?

Thanks,
Haripriya

Below is the cut,paste from the Linux kernel 2.4.18 documentation.

"In typical situations, a usb_bulk_msg() loop writing out 4 KB chunks is
going to waste more than half the USB 2.0 bandwidth.  Delays between the
I/O completion and the driver issuing the next request will take longer
than the I/O.  If that same loop used 16 KB chunks, it'd be better; a
sequence of 128 KB chunks would waste a lot less"

And since I saw the usb_bulk_read and usb_bulk_write functions in LIBUSB
to be limiting the maximum bulk.len to 4KB, I assumed some changes might
be required. I'll try it out.

Thanks!

-----Original Message-----
From: Johannes Erdfelt [mailto:joh...@er...]
Sent: Thursday, January 29, 2004 2:07 PM
To: Nagarajan, Haripriyax
Cc: lib...@li...
Subject: Re: USB 2.0 support


On Thu, Jan 29, 2004, Nagarajan, Haripriyax
<har...@in...> wrote:
> Is there any thought on support for USB 2.0? It is clear from the
> documentation that libusb is geared towards USB 1.1. But will there be
> any modifications made to LibUSB to accommodate the higher bulk
transfer
> rate supported to USB 2.0?

There are no changes required to support the higher bulk transfer rates
in USB 2.0

Try for yourself.

JE

On Wed, Jan 28, 2004, ???? ???? <abt...@ho...> wrote:
> I try to control one led on usb port to turn it on and off.
> is it possible?

What LED?

My USB ports have no LEDs.

JE

On Wed, Jan 14, 2004 at 12:38:08AM +0100, Oliver Neukum wrote:
> Hi,
> 
> I'll do it slowly this time.
> Here is the first one.

Applied, thanks.

greg k-h

Hi Johannes,

On Wed, 28 Jan 2004, Johannes Erdfelt wrote:

> > For crying out loud what's wrong with plain text?!
> 
> There is plain text too, the source code :)

Hehe, I'm on the USERS list.

73,
Ged.

On Wed, Jan 28, 2004, Ged Haywood <ge...@ww...> wrote:
> On Wed, 28 Jan 2004, ebourqui wrote:
> > The html docs will be created from the sgml files once you do a
> > "./configure", then a "make"
> 
> Not if my sgml tools package is screwed.  Which it is.
> 
> For crying out loud what's wrong with plain text?!

There is plain text too, the source code :)

JE

Hi there,

On Wed, 28 Jan 2004, ebourqui wrote:

> The html docs will be created from the sgml files once you do a
> "./configure", then a "make"

Not if my sgml tools package is screwed.  Which it is.

For crying out loud what's wrong with plain text?!

73,
Ged.

Ged,
    The html docs will be created from the sgml files once you do a
"./configure", then a "make" , from within the directory where you unzipped
libusb. The html docs will then be in libusb-0.1.7/doc/html .

See the file INSTALL.libusb in your libusb directory for more details.

Emile

----- Original Message ----- 
From: "Ged Haywood" <ge...@ww...>
To: "Douglas S. Blank" <db...@br...>
Cc: <lib...@li...>; "linux-usb mlist"
<lin...@li...>
Sent: Saturday, January 24, 2004 12:25 PM
Subject: Re: [Libusb-devel] Re: [Linux-usb-users] Next step in reverse
engineering USB robot


> Hi there,
>
> On Fri, 23 Jan 2004, Douglas S. Blank wrote:
>
> > But we can't get the most basic pieces of the URB's figured out. If you
> > can help, please send me a hint.
>
> I'm not sure at what level your problem lies.  Is it using libusb
> itself or is it what to tell libusb to do?  I've tried to read the
> libusb docs but it's agony because they all seem to be in sgml.
>
> I tried to install a new sgml tools package and it's all screwed up now,
> bits missing, don't know how long it will take to fix.  :(
>
> 73,
> Ged.
>
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> Libusb-devel mailing list
> Lib...@li...
> https://lists.sourceforge.net/lists/listinfo/libusb-devel
>

According to Stavros Markou <sm...@pa...>:
> Hi,
> 
> Does libusb work on a linux pda ?

If the PDA can be USB Host (most of the time they only are slaves) and
that the PDA has the proper USB subsystem in the kernel, then why not.


Hub
-- 
AbiWord maintainer - Lille, France
http://www.figuiere.net/hub/ 
GPG fingerprint: 6C44 DB3E 0BF3 EAF5 B433  239A 5FEE 05E6 A56E 15A3

Hello, friends

I try to control one led on usb port to turn it on and off.
is it possible?

I compiled the example file testlibusb.c included in tests folder and I got 
this result

If it is possible, what is the modifation should i do to testlibusb.c ?


[root@localhost tests]# ./testlibusb
bus/device  idVendor/idProduct
001/001     0000/0000
- Product      : USB OHCI Root Hub
- Serial Number: d084c000
  wTotalLength:         25
  bNumInterfaces:       1
  bConfigurationValue:  1
  iConfiguration:       0
  bmAttributes:         40h
  MaxPower:             0
    bInterfaceNumber:   0
    bAlternateSetting:  0
    bNumEndpoints:      1
    bInterfaceClass:    9
    bInterfaceSubClass: 0
    bInterfaceProtocol: 0
    iInterface:         0
      bEndpointAddress: 81h
      bmAttributes:     03h
      wMaxPacketSize:   2
      bInterval:        255
      bRefresh:         0
      bSynchAddress:    0

_________________________________________________________________
Add photos to your messages with MSN 8. Get 2 months FREE*. 
http://join.msn.com/?page=features/featuredemail

The message could not be delivered to:

Addressee: bo...@mu...
: Remote system's reason for rejecting: 5.1.1 <bo...@mu...>... Address not found

The message could not be delivered to:

Addressee: ad...@pc...
: Remote system's reason for rejecting: RCPT TO:<ad...@pc...=
.edu> User unknown

Please be advised that I will be out of the office from January 19 until =
February 2nd and will not be checking email.

I will address all your creditsand RA requests upon my return.  For urgent =
issues, please contact Cal...@cl....

Thanks you for your patience and understanding.

Sincerely,

Kara Hale
From Cloud Ten Pictures

The message contains Unicode characters and has been sent as a binary attachment.

Hi there,

On Fri, 23 Jan 2004, Douglas S. Blank wrote:

> But we can't get the most basic pieces of the URB's figured out. If you 
> can help, please send me a hint.

I'm not sure at what level your problem lies.  Is it using libusb
itself or is it what to tell libusb to do?  I've tried to read the
libusb docs but it's agony because they all seem to be in sgml.

I tried to install a new sgml tools package and it's all screwed up now,
bits missing, don't know how long it will take to fix.  :(

73,
Ged.

=2D----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Wed, 31 Dec 2003 17:24 pm, Brad Hards wrote:
> I've finally completed the C++ bindings. See attached patch for the gory
> details. There is also a tarball - I can't create a patch for the new
> apidocs directory without commit access to the repository.
I didn't get any feedback on this. If no-one has any suggestions, I'd like =
to=20
apply it.

JE: can I be added to the developer list?

Brad
=2D----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFAFjqGGwwszQ/PZzgRAkeEAJ4134mIMIOV/yisHe7AelnaRR+FDQCgo7Uv
bNOZob3RADouCB/cWTNWUTQ=3D
=3Dw4Vr
=2D----END PGP SIGNATURE-----

also sprach Brad Hards <bh...@bi...> [2004.01.24.0541 +0100]:
> I have used HID and usbfs (though libusb) simultaneously. Don't do
> the claim_interface(), and don't rely on consecutive transfers
> being to a consistent state, but it does work - that is how my
> Logitech mouse applet is implemented.

If I don't claim the interface, then there are various functions in
libusb that will exit with errors.

The patch made it to 0.1.8beta and will be released in the 0.1.8
version. And it works on Linux 2.6.1. It would be cool to hear from
other kernel versions and/or systems(*).

(*) it's a _np (not portable) function because of of this writing,
I don't know whether BSD or Darwin (or Win32) require a disconnect
=66rom the kernel driver to claim the interface with libusb.

--=20
martin;              (greetings from the heart of the sun.)
  \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck
=20
invalid/expired pgp subkeys? use subkeys.pgp.net as keyserver!
=20
"quite frankly, teachers are the only profession
 that teach our children."=20
                                                      - george w. bush=20

Ged Haywood wrote:
> Hi there,
> On Thu, 22 Jan 2004, Pedro M. wrote:
>>I love this idea of USB / Linux Robot ( well a WIFI robot is OK too ).
>>
>>Where there is a starting website about it ??..
> 
> http://www.google.com
> 
> The words "linux robot" gave me just under 300,000 hits and when I
> added "usb" that took it down to about 30,400.

Actually, you won't find anything on this particular project yet, as far 
as the USB is concerned. I'm one of the designers of Pyro, Python 
Robotics, which is an open source, mainly Linux-based, set of libraries 
and GUI for learning and doing research in robotics. You can find out 
more at http://PyroRobotics.org

Pyro works with some free simulators, and a few commercial robots. We 
are attempting to interface with a nice little commercial product, 
called Evolution Robotics 1, or er1 (http://www.evolution.com/).

But we can't get the most basic pieces of the URB's figured out. If you 
can help, please send me a hint. Details in my previous post. Thanks!

(BTW, I'm going to check out the usb-robot sourceforge site... thanks 
for that hint!)

-Doug

> 73,
> Ged.

=2D----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tue, 20 Jan 2004 07:00 am, martin f krafft wrote:
> I'd be interested to hear if there are other supported platforms
> that even need this. The reason on Linux is quite simply that there
> are devices (like HID) where the kernel takes control and thusly,
> usbfs won't work anymore. If this is not the case and not needed on
> other platforms, then it'll be a linux-specific thing. if it is
> needed, then we should write it for all.
I have used HID and usbfs (though libusb) simultaneously. Don't do the=20
claim_interface(), and don't rely on consecutive transfers being to a=20
consistent state, but it does work - that is how my Logitech mouse applet i=
s=20
implemented.

Brad
=2D----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)

iD8DBQFAEfeAGwwszQ/PZzgRArd0AJ0SLV++73Z68d28VOuNSiStoDG9ZACgiRGB
Y5BkFcJEbUzAyn779y8+Plg=3D
=3DUoqH
=2D----END PGP SIGNATURE-----

Hi,

Does libusb work on a linux pda ?

Stavros Markou.

Hi,

Does libusb work on a linux pda ?

Stavros Markou.

Is there a way to determine bus speed "USB_SPEED_{LOW,FULL,HIGH}"using
libusb interface?

Thanks!

Hi there,

On Thu, 22 Jan 2004, Pedro M. wrote:

> I love this idea of USB / Linux Robot ( well a WIFI robot is OK too ).
> 
> Where there is a starting website about it ??..

http://www.google.com

The words "linux robot" gave me just under 300,000 hits and when I
added "usb" that took it down to about 30,400.

73,
Ged.