Galileo TriMix / IrDA

[William Perry]

On Apr 30, 2013, at 8:49 AM, Henrik Brautaset Aronsen <subsurface at henrik.synth.no> wrote:

> William Perry wrote:
>> I am looking at switching to Subsurface so that I can just keep a linux VM floating around to download dives from my Galileo (Trimix) instead of a Windows VM and the 'fabulous' SmarTrak software from ScubaPro.  I am running this Ubuntu 13.04 VM in Parallels 8 at the moment and am having issues getting this to download anything at all with a generic mcs7780 USB adapter.  I expect it is an issue with the low-level IR stack, but the IRDA documentation leaves a bit to be desired.
>> 
>> The irda0 driver gets loaded successfully and I have modified /etc/default/irda-utils to set
>> DEVICE=irda0
>> AUTOMATIC=true
>> DISCOVERY=true
>> 
>> And irattach gets launched successfully when I plug in the USB device, but it just spits out this:
>> 
>> Apr 30 08:20:36 bplinux irattach: Starting device irda0
>> Apr 30 08:20:36 bplinux irattach: ioctl(SIOCSIFFLAGS): Broken pipe
>> Apr 30 08:20:36 bplinux irattach: Stopping device irda0
>> Apr 30 08:20:36 bplinux irattach: exiting …
>> 
>> And then dies.  'ifconfig -a' shows the irda0 adapter still, and lsusb is showing the correct chip manufacturer / identifying information:
>> Bus 002 Device 005: ID 9710:7780 MosChip Semiconductor MCS7780 4Mbps Fast IrDA Adapter
>> 
>> But I get no joy trying to import from Subsurface, or from a quick test program I wrote to talk directly to libdivecomputer.  This is actually the first time in 20 years of using and developing software on Linux that I have tried to get IrDA working.  Am I missing something very basic here?  I was able to download the dives into SmarTrak in windows so I don't think I have a hardware issue, but the adapter is ~6 years old and has had a few accidents, perhaps the windows drivers are a bit more forgiving?
> 
> You shouldn't need to write your own test program.  Did you try one of the test programs listed at http://libdivecomputer.org/download.html ?

Both the 'universal' and 'smart' test programs in the libdivecomputer distro had the same issues.  I was toying around with writing a UDDF exporter using libdivecomputer anyway so I tried my own test program as well.

> I can't comment on the IrDA problem, but I guess this is a libdivecomputer issue more than a Subsurface issue.  I'm Cc-ing the libdivecomputer list.

I am really hoping that Uwatec gets rid of irda in their replacement for the galileo, but who knows.  That is still at least a year away from what I hear from our ScubaPro rep.  This may get me to just ditch them and buy a Nitek Q -- asked Lamar Hires how open he would be to contributing a parser/communications layer to libdivecomputer to read from it, if I hear anything positive I will let the libdivecomputer-devel list know.

-bill