Am 30.04.2015 um 14:28 schrieb Paul W Panish:
>
> I have some questions on your comments. The issue of the w1 drivers
> is new to me, and sounds like it may be an explanation of what I'm
> seeing, though I still don't understand why there should be a
> temperature dependence.
>
I don't see any connection either. That's the moment when I start to
check everything.
> I'm running Debian Linux on a Beaglebone Black. I've done nothing
> special for a kernel build, and I've built owfs using defaults, so
> the configure output indicates that W1 is enabled.
>
This means OWFS is able to use the w1 kernel driver. It doesn't mean it
uses it, the latter is configured through *each* owfs application
command line and owfs.conf.
> My application starts at boot and runs as root with full permissions.
> The OWFS information on the W1 linux kernel module states there is a
> conflict between W1 and OWFS "without OWFS 1-wire support". Given the
> owfs default config supports W1 this seems to imply there should be
> no conflict. Is that correct?
>
Not correct.
The meaning is: If you use --w1, there is no conflict, as OWFS uses the
kernel driver as it's underlying system and the kernel driver nicely
integrates OWFS' onewire commands and onewire commands from other
sources -- kernel's battery/temperature/backlight/gpio systems.
But if you use --usb or --i2c instead, you will have a conflict between
OWFS and kernel's battery/temperature/backlight/gpio systems, as the
latter still uses the w1 kernel subsystem and it accesses the onewire
host adaptor chips at the same time as OWFS, producing garbage on the
onewire.
So it's not about enabling --w1 but about avoiding --usb or --i2c. Or
use the latter and disable kernel's w1 subsystem.
>
> Also, if having W1 enabled in the configuration does not avoid
> conflicts there are two workarounds mentioned. The second is to run
> as root, which I'm doing.
>
> The first is to disable ds9490.ko using udev. I can find no
> information on how to do this. Checking dmesg I can find no instance
> of w1 (or W1) when the LinkUSB device is registered or at any other
> time.
>
Correct. The LinkUSB isn't DS2490-based, but a RS232 based onewire host
(the old "Link") connected a FTDI USB-RS232 converter.
The kernel w1 subsystem has no driver module for this chip, so there
cannot be a problem with this. It also means you cannot use --w1 with
this host adaptor.
>
> I've never had any problem with power-on-reset issues such as you
> mention. Why would I need to assume that any device on the bus might
> reset without cause at any time?
>
Power may fail in mysterious ways. Been there, seen that.
For the DS2408 powering through the 1W line helps it to keep its state.
Both power sources (1W and peripheral power) have to fail. But it's not
so unlikely for the USB to have a short power failure just the moment
where the mains has a short break. OWFS immediately reconnects a failed
USB host adaptor so there is no other error message but to check the POR
flags. Plus, hotplugging USB and onewire devices is only possible when
you check for POR.
For the DS182x, bus powering is only possible if you *never* do
simultaneous conversions. If you want to do that, you need to check for
POR as the peripheral power may have failed during the conversion. In
all other cases, same as DS2408.
Kind regards
Jan
------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Owfs-developers mailing list
Owfs-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/owfs-developers
