[sane-devel] Xerox 3220 USB problem

Alex Eftimie alex at rosedu.org
Mon Jul 22 20:33:58 UTC 2013

On Mon, Jul 8, 2013 at 10:39 AM, ABC <abc at telekom.ru> wrote:
> Alex Eftimie,
> On Sun, Jun 30, 2013 at 01:56:08PM +0300, Alex Eftimie wrote:
>> On Sun, Jun 30, 2013 at 1:31 PM, ABC <abc at telekom.ru> wrote:
>> > It maybe be becasue (it is possible) that smfp using different scanning
>> > mode, like different resolution. (And then doing some image conversion
>> > internally.) xerox_mfp driver's settings are 'physical' for the device.
>> Well, I'm setting the resolution myself, with --resolution=300 (the
>> same for xerox_mfp and smfp). Also, I'm using as color mode: "Lineart"
>> for xerox_mfp and "Black and White - Line Art" for smfp. But still 4
>> seconds vs 10 seconds. Can I get 'physical' with smfp :D?
> Regarding longer scans via smfp driver, I looked your usbmon logs. Scan
> parameters seems to be nearly equal. And I see that smfp (1.mon.out)
> driver send device INQUIRY reqest 4 times with interval from first to last
> request ~3 seconds. Total time of smfp scan is 8.7 seconds. Total time
> of xerox_mfp scan is 4.3 seconds. Image data reading time in both cases
> is nearly 1 second. Time between first image request and image data flow
> is ~3.5 seconds.  So I conclude that smfp driver spend difference in
> time internally and actual speed of working with device and device
> itself is the same.  Why is smfp driver spending time inernally?
> Unknown. It may be just some hardcoded intervals between commands.

Sorry for the late reply.

> It seems your usbmon logs don't contain case when scanimage fail with
> xerox_mfp driver. As I understand it would fail if `scanimage -L' will
> be run twice (it will fail second time). So it will be useful to have
> usbmon log of such case.

Here: https://www.dropbox.com/sh/u4p87ustkef1dw5/12nemzRUYO are the
usbmon logs for: default (xerox_mfp) and proprietari (smfp) drivers,
running two consecutive `scanimage -L`.

Hope this will give you some hints regarding the underlying issue.


More information about the sane-devel mailing list