RE: Device Support for I2C and GPIO

[Mark Rivers <rivers@cars.uchicago.edu>]

Hi Florian,

If you set autoConnect=1 then after iocInit please send the output of

asynReport 1 I2C

Your driver printed this:
2016/01/14 13:11:54.967 I2C: wrote 1 bytes to /dev/i2c-1, return asynError

Why is it returning asynError if the write was successful?

> I have a question concerning the asynRecord.
> To read out the control register of my ADC I have to send the two bytes 0x42 0x02.
> Is this possible using the asyn record with OFMT="Binary" ? 
> How would the corresponding put to the BOUT field look like?

Yes, this is possible.  If you are doing the put from a programming language (C, Python, etc.) you would just send a byte array of length 2 containing those bytes.

>From the command line you can do this with caput.  I don't think you can use hex notation, but the following appears to work using decimal notation:

corvette:~>caput -a quadEMTest:TetrAMM:Asyn.BOUT 2 66 2
Old : quadEMTest:TetrAMM:Asyn.BOUT 80 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
New : quadEMTest:TetrAMM:Asyn.BOUT 80 66 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Mark

________________________________________
From: Florian Feldbauer [florian@ep1.ruhr-uni-bochum.de]
Sent: Thursday, January 14, 2016 7:33 AM
To: Mark Rivers; 'Pete Jemian'; tech-talk@aps.anl.gov
Subject: Re: Device Support for I2C and GPIO

Hey,

concerning the compiler error, I uploaded the unfinished version from my
desktop PC instead of
the one used on the Beagle Bone...I now uploaded the correct version.
Link is still the same.
Sorry for the mistake!

I also tested the asynRecord. This is my record:
record( asyn, "TEST" ) {
   field( DTYP,  "asynRecordDevice" )
   field( PORT,  "I2C" )
   field( ADDR,  "0" )
   field( IFACE, "asynOctet" )

   ## AsynOctet output
   field( OEOS,  "" )
   field( OFMT,  "Binary" )

   ## AsynOctet input
   field( IEOS,  "" )
   field( IFMT,  "Binary" )
}

in the IOC I checked the following:
iocRun: All initialization complete
dbgf TEST.PCNCT
DBR_STRING:          "Connect"
dbgf TEST.CNCT
DBR_STRING:          "Connect"
dbgf TEST.ENBL
DBR_STRING:          "Enable"
dbgf TEST.AUCT
DBR_STRING:          "autoConnect"
dbgf TEST.ERRS
DBR_CHAR[100]:       "" +      ""
dbpf TEST.NOWT 1
DBR_LONG:           1         0x1
dbpf TEST.BOUT B
DBR_CHAR[2]:         "B"
dbgf TEST.ERRS
DBR_CHAR[100]:       "error  nread 0 I2C: /dev/i2c-1 disconnected:" +  ""

"I2C: /dev/i2c-1 disconnected:" is the error message when
asynOctet->read is called,
but the file handle (drvAsynI2c::_fd) is negative, so either ::connect()
was not called,
or the connect failed without printing the error message.

I set autoConnect to false in my code, and tried it again:
dbgf TEST.PCNCT
DBR_STRING:          "Connect"
dbgf TEST.CNCT
DBR_STRING:          "Disconnect"
dbgf TEST.ENBL
DBR_STRING:          "Enable"
dbgf TEST.AUCT
DBR_STRING:          "noAutoConnect"
dbpf TEST.CNCT "Connect"
2016/01/14 13:11:54.957 I2C: Open connection to /dev/i2c-1
DBR_STRING:          "Connect"
dbgf TEST.ERRS
DBR_CHAR[100]:       "" +      ""
dbpf TEST.NOWT 1
DBR_LONG:           1         0x1
dbpf TEST.BOUT B
2016/01/14 13:11:54.964 I2C: read 0 bytes from /dev/i2c-1, return asynError
2016/01/14 13:11:54.965 drvAsynI2C::writeOctet: trying to send 2 bytes
drvAsynI2C::writeOctet: sending: 0x420x00
2016/01/14 13:11:54.967 I2C: wrote 1 bytes to /dev/i2c-1, return asynError
2016/01/14 13:11:54.969 I2C: read 0 bytes from /dev/i2c-1, return asynError
DBR_CHAR[2]:         "B"
dbgf TEST.ERRS
DBR_CHAR[100]:
  "error  nread 0 I2C: /dev/i2c-1 read error: Remote I/O error" +       ""

So with this second test, writing works, but reading not. The beaglebone
doesn't continue the clock for reading...
I will further study the problem...

I have a question concerning the asynRecord. To read out the control
register of my ADC I have to send
the two bytes 0x42 0x02.
Is this possible using the asyn record with OFMT="Binary" ? How would
the corresponding put to the BOUT field look like?

Cheers
Florian


On 01/12/2016 09:02 PM, Mark Rivers wrote:
> You can also load an asynRecord in your IOC.  With that you can turn AutoConnect on and off, and you can manually force calls to ::connect() and ::disconnect() from medm.  If you do that then do you see your methods being called?
>
> Mark
>
> ________________________________________
> From: Florian Feldbauer [florian@ep1.ruhr-uni-bochum.de]
> Sent: Tuesday, January 12, 2016 8:58 AM
> To: Mark Rivers; 'Pete Jemian'; tech-talk@aps.anl.gov
> Subject: Re: Device Support for I2C and GPIO
>
> Hey,
>
> the current version of my code can be found here:
> http://www.ep1.rub.de/~florian/drvAsynI2C.tar.gz
>
> The value for the writeOctet and readOctet is expected to be
> a bitstream.
> Currently I tried to use StreamDevice as the interface between records
> and asynPortDriver...
>
>     Florian
>
>
> On 01/12/2016 01:22 PM, Mark Rivers wrote:
>> Your derived method's connect function should be called.  Can you send your code?  I'll take a look and see if I can spot anything wrong.
>>
>> Mark
>>
>> ________________________________________
>> From: Florian Feldbauer [florian@ep1.ruhr-uni-bochum.de]
>> Sent: Tuesday, January 12, 2016 4:38 AM
>> To: Mark Rivers; 'Pete Jemian'; tech-talk@aps.anl.gov
>> Subject: Re: Device Support for I2C and GPIO
>>
>> Hey Marc, hey all!
>>
>> unfortunately, I did not had the time to write and test the I2C device
>> support last year.
>> But now I have the time and I'm currently facing a problem:
>> My I2C device support is written as a class derived from the asynPortDriver,
>> reimplementing writeOctet, readOctet, connect and disconnect.
>>
>> When calling the ctor of asynPortDriver I set the "autoConnect" flag to 1,
>> but the connect function of my class is never called, thus the file
>> handle for the
>> I2C interface is never opened.
>>
>> Is there something I'm missing?
>> I thought autoConnect means that the connect function is automatically
>> called?
>>
>> Cheers
>> Florian
>>
>> On 11/25/2015 06:24 PM, Mark Rivers wrote:
>>> The I2C asyn driver could be considered for going into asyn itself, since it is conceptually similar to the serial, IP, GPIB, USBTMC, and VXI11 drivers.
>>>
>>> Note that "ip" is really not for software related to Industry Pack modules any more.   Most industry pack drivers now have their own modules (ip330, ipUnidig, dac128V, softGlue, ipac for ipOctal).  "ip" is really now software for asynOctet devices (serial, GPIB, TCP) that do not have another home like "vac" and "delayGen".  "ip" is no longer a good name for it.
>>>
>>> Similarly "std" was originally the home for a lot "standard" beamline control stuff.  Most of this has now been moved to other modules (calc, busy, optics, sscan), so "std" is more like "misc" now.
>>>
>>> Mark
>>>
>>>
>>> -----Original Message-----
>>> From: tech-talk-bounces@aps.anl.gov [mailto:tech-talk-bounces@aps.anl.gov] On Behalf Of Pete Jemian
>>> Sent: Wednesday, November 25, 2015 11:10 AM
>>> To: tech-talk@aps.anl.gov
>>> Subject: Re: Device Support for I2C and GPIO
>>>
>>> Michael describes a good procedure for completely new functionality.
>>> New modules should provide support distinct from other modules.
>>>
>>> This _might_ have a place in one of the existing synApps modules.  This
>>> is not "ip" functionality (related to Industry Pack modules), not
>>> "calc", not "optics", perhaps "std" is the catchall module for new
>>> synApps support?
>>>
>>> On 11/25/2015 9:08 AM, Michael Davidsaver wrote:
>>>> On 11/25/2015 03:08 AM, Florian Feldbauer wrote:
>>>>> I would like to upload the sources to github. Is it possible to upload
>>>>> it to the "epics-modules" group?
>>>> At this point there isn't any established criteria for when/if
>>>> repositories in this group are created.
>>>> It would suggest starting with a repository under your own account (what
>>>> I do), then think about moving to epics-modules at some later date,
>>>> which isn't difficult.
>>>>
>>>> That said, if you want to go straight to epics-modules I'm inclined to
>>>> just add it.  Let me know if you want to do this and I'll do this in a
>>>> day or so, assuming no one objects.  Of course, first you need to solve
>>>> the hard problem, the repository name ;)
>>>>
>>>>
>>>> Michael
>>>>
>> --
>> ----------------------------------------
>> | Dr. Florian Feldbauer                |
>> |                                      |
>> | Helmholtz-Institut Mainz /           |
>> | Johannes Gutenberg-Universität Mainz |
>> | Johann-Joachim-Becher-Weg 36         |
>> | D-55128 Mainz                        |
>> |                                      |
>> | Office: SB1 / 00-213                 |
>> | Phone:  (+49)6131 / 39-29605         |
>> ----------------------------------------
>>
> --
> ----------------------------------------
> | Dr. Florian Feldbauer                |
> |                                      |
> | Helmholtz-Institut Mainz /           |
> | Johannes Gutenberg-Universität Mainz |
> | Johann-Joachim-Becher-Weg 36         |
> | D-55128 Mainz                        |
> |                                      |
> | Office: SB1 / 00-213                 |
> | Phone:  (+49)6131 / 39-29605         |
> ----------------------------------------
>

--
----------------------------------------
| Dr. Florian Feldbauer                |
|                                      |
| Helmholtz-Institut Mainz /           |
| Johannes Gutenberg-Universität Mainz |
| Johann-Joachim-Becher-Weg 36         |
| D-55128 Mainz                        |
|                                      |
| Office: SB1 / 00-213                 |
| Phone:  (+49)6131 / 39-29605         |
----------------------------------------