[sane-devel] Re: usb_control/bulk_msg: timeout

A E Lawrence A.E.Lawrence at lboro.ac.uk
Sun Jan 5 19:16:04 GMT 2003


Bill Moseley wrote:
> On Sun, 5 Jan 2003, A E Lawrence wrote:
> 
> 
>> > I made this change and ran
>>
>> >   $ make modules
>> >   # make modules_install
>>
>> > and rebooted.  Didn't seem to make any difference for my situation.  The
>> > timeout messages still happen -- and about the same rate.
>>
>>Sorry not to reply in the same thread: I had deleted everything before 
>>realizing that you had not mentioned make config or make xconfig.
>>
>>You are sure that you had set CONFIG_USB_LONG_TIMEOUT? Otherwise you would 
>>have recompiled exactly the same code as before.
> 
> 
> moseley at bumby:/usr/src/linux$ fgrep CONFIG_USB_LONG_TIMEOUT .config
> CONFIG_USB_LONG_TIMEOUT=y
> 
> But I'm not kernel hacker.  I edited usb.c, edited .config, and ran the
> two make commands as above.

Ok. That looks ok. There are some other timeouts in usb.c. It might be 
worth experimenting with those values. Just search on the strings `timeout' 
and `TIMEOUT'. But at a quick glance, there seems to be only one place that 
the "usb_control/bulk_msg: timeout" is generated, and it looked as if that 
was always controlled by GET_TIMEOUT. So I don't understand why you didn't 
see a change. Maybe that code gets called somewhere else with a different 
timeout value passed in.

It was fairly clear that I needed a longer timeout since usb.c had to get a 
reply in order to pass onto the hotplug script the information
to load the scanner module. I don't understand your original problem so 
can't comment on whether that might be so in your case as well.

ael







More information about the sane-devel mailing list