Re: [Gumstix-users] Re: i2c-dev and EINPROGRESS error.

[Matthew S. <ma...@ec...>]

Well the problem turned out to be bad code leaving the bus locked up.

I am in the progress of making sure the pic times out the clock stretch.

Interestingly and the bus hang will hang the gumstix. I am sure this is 
not appropriate behaviour, I shall investigate and fix when I get some 
free time.

Just clock stretch for infinity and the gumstix will eventually stop 
responding.

Thanks for your help,

Matthew Swabey

Dave Hylands wrote:
> Hi Matthew,
> 
>> I have been successfully using the I2C bus for some time, I have had my
>> PICs hanging off the bus chatting perfectly.
>>
>> At the moment however:
>>
>> They use a protocol on the I2C bus of
>>
>> write(cmd)
>> read(result)
>>
>> Now when the read instruction is issued by the gumstix if the data for
>> result isn't ready the pic stretches the clock to accomplish this (as
>> specified in the I2C standard).
>>
>> Currently the read ioctl is returning with EINPROGRESS and then exiting.
>>
>> This seems to be a return due to the driver timing out, however it
>> leaves the bus in an unusable state, as my PIC is clock stretching for ever.
>>
>> Is there somewhere the bus timeout on the gumstix, and is it simple to
>> control (a fcntl perhaps?), or if not would you suggest a different
>> method of solving this problem?
> 
> Which version of the kernel are you using? In the 2.6.11 kernel, I
> found a bug where the timeout was way too short.
> 
> If your PIC is stretching the clock forever, then that's whats leaving
> the bus in an unusable state. There's no way for a master to recover
> from that. This is because with the i2c bus you can only drive the bus
> low, and not high, so whoever is driving it low is at fault.
> 
> There are some ioctls mentioned in i2c-dev.h which indicate that the
> number of retries and the timeout can be modifed, but I haven't tested
> these.
> 
> --
> Dave Hylands
> Vancouver, BC, Canada
> http://www.DaveHylands.com/
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting language
> that extends applications into web and mobile media. Attend the live webcast
> and join the prime developer group breaking into this new coding territory!
> http://sel.as-us.falkag.net/sel?cmd=k&kid0944&bid$1720&dat1642
> _______________________________________________
> gumstix-users mailing list
> gum...@li...
> https://lists.sourceforge.net/lists/listinfo/gumstix-users


-- 
Matthew Swabey
Rm7025, Bldg21 Faraday Tower
Department of ECS
University of Southampton
University Road
Southampton, SO17 1BJ
England

Tel: 02380 597381
Mob: 07966 440946