[sane-devel] about maxPacketSize

peter peter at mail.mustek-hz.com.cn
Thu Jun 14 08:56:26 UTC 2007

Many thanks! 
Yes, i think the usb2.0 is compatiable with usb 1.1 , but the code not working.(maybe it too old). 
The scanner is an usb2.0 device(sq113c),when is plugging to an usb1.1, it perform as usb1.1 device. when it plugging to usb2.0 port, it perform as usb2.0 device.

but the code seems only working when it perform as usb1.1 device.  the maxPacketSize  is only different between them, so does anyone have suggestion about what i should caution when i bulk read & write with the scanner?   


----- Original Message ----- 
From: "Gerald Murray" <gmurray at cloudnet.com>
To: "peter" <peter at mail.mustek-hz.com.cn>
Cc: <sane-devel at lists.alioth.debian.org>
Sent: Thursday, June 14, 2007 4:03 PM
Subject: Re: [sane-devel] about maxPacketSize

> Quoting peter <peter at mail.mustek-hz.com.cn>:
> > The working usb port is 64 ,and the bad usb port is 512 . In my code
> --snip--
> > so , my questing is why the usb port in OS10.2  maxPacketSize is   
> > 64,but  in OS 10.4 it change to 512 ?  and what does the   
> > maxPacketSize  mean??
> If the usb port is an older one (version 1.0 or version 1.10) then it
> is full speed device.  These are limited to maxpacket size of 64.
> Newer usb is version 2.0, high speed, which can have maxpacket size of 512,
> and are compatible with lower speed devices.  Computers also mix
> the versions, such as having 2 2.0 usb ports, and 4 1.1 ports.
> If the device is usb version2.0, it can use the bigger maxpacket size,  
> that every hub between the root port and the device can handle 2.0 speed.
> If any one hub cannot, then the smaller 64 size packets is the maximum
> that can be used.  There should be no problem in using smaller packet
> sizes on the higher speed capable hubs, but it will take longer.
> For greatest throughput of data, use the longest length data urbs
> that is available.  In a series of urbs, the data stream is considered
> terminated if it has less than the expected maximum, a short packet.
> There may be nothing wrong with the "bad usb" port.  It is likely only
> capable of usb version 1.1 specifications.
> regards,
> Gerald
> __________ NOD32 2328 (20070613) Information __________
> This message was checked by NOD32 antivirus system.
> http://www.nod32.com.hk

More information about the sane-devel mailing list