Communication problem in Subsurface

[John Van Ostrand]

Hi,

I'm working on the Cochran dive computer import and I've often had
communications issues where the transfer fails at the end with a short
read. It happens quite consistently when I switch applications (i.e.
Alt-Tab) and works consistently when I don't switch.

The transfer is a little odd in that commands are sent at a lower baud rate
but bulk data, like logs and profile data are sent at what looks like the
MCU's speed, 800,000 baud, or so. The profile can be as much as 16MB and
there isn't flow control. My test profile has 2.2MB. The FTDI FT8U232AM
chip seems to have a small buffer. My laptio is not recent but still
decent: A Lenovo ThinkPad T510 with Intel I7 M620, 2.67GHz, 8GB RAM and
Fedora 22 X86_64.

I've tried to renice subsurface to priority -20 without success.

There are reasons I want to download the profile in chunks larger than each
dive. One is that there needs to be an 800ms to 1,800ms delay when giving
commands to the dive computer which can hold close to 1,000 dives. That
would result in a large delay in downloading.

I could break it up into more reasonable chunks, and retry on a failure but
I don't want to work on that until I ask these questions.

1. Does anyone have a DC that does large downloads and does a window-switch
cause a failure?
2. Is there something that subsurface does when it loses or gains focus
that would cause this?
3. Any ideas to resolve this?

-- 
John Van Ostrand
At large on sabbatical
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.subsurface-divelog.org/pipermail/subsurface/attachments/20160111/5796f7d0/attachment.html>