I have an owfs network with a usb adapter connected to nine DS2409’s, with sensors on the main and aux limbs of the DS2409’s. I have been using this arrangement for many months under Fedora with owfs-2.7p2. I just upgraded to Ubuntu Server 8.04 and owfs-2.7p4.

However, whenever I go to one of my DS2409’s aux limbs and issue an “ls”, there is no output until I open another window and do an ls of another limb; that “releases” the first “ls”, which then shows not only the content of the actual aux limb, but multiple echos of all of the DS2409’s (which are not even on that limb, as they are all at the top level of the 1-wire network).

I tried several approaches to resolve this, including the module.d/blacklist suggestion above; nothing worked until I downgraded back to 2.7p2, which fixed the problem. I then tried 2.7p3, but the problem recurred. Once I was back to 2.7p2, I removed the blacklist entries, and I am not having any problems.

Thus, this appears to be a problem introduced in 2.7p3 . Note that I am using owserver and having owfs (and owhttpd) access the 1-wire network through owserver.

With owhttpd, showing that directory produces a never ending listing (continually growing until I close the browser window) showing first the contents of the actual aux limb, and then repeatedly showing all of the DS2409’s in sequence, over and over.

The problem occurs with either owfs or owhttpd accessing the 1-wire network, so the problem is not with fuse or the fuse mounting code in owfs. It does not occur on all limbs of all of the DS2409’s; the one limb that always produces it (1F.E1C004000000/aux) has a DS18S20, three DS2438’s, and a DS18B20.

If I do not use owserver, but rather use owfs to access the 1-wire network directly, the problem also occurs. Thus, the problem is not with owserver itself, but with the common code used to access the 1-wire network.

Note that I am starting owserver with /opt/owfs/bin/owserver -u -F -p 4001 to access the usb network.

also (as noted at http://owfs.org/index.php?page=ubuntu), need to:
sudo chmod 4755 /bin/fusermount
otherwise, the mounted directory has question marks for the owner, group, and permissions
Note that thre mare be security concerns regarding making fusermount suid. However, the only change from the chmod was to add user execute permission. Before: -rwsr-xr–, after -rwsr-xr-x .

http://www.nabble.com/Debian%2BUbuntu-package-availability-p12741322.html

http://www.nabble.com/Ubuntu…I‘ve-been-spoilt-with-the-NSLU2-packaging…help-!-td12696120.html

There also appear to be repos for Gutsy, though it’s not mentioned in that post:

deb http://ppa.launchpad.net/smurf/ubuntu gutsy main universe
deb-src http://ppa.launchpad.net/smurf/ubuntu gutsy main universe

to stop the kernel w1_master_driver from opening the device. w1_master_driver causes tons of stutters to all programs (I think because it tries to do way too much in a kernel call).

A few posters wondered about which usb device to use. I just did “-u” on commandlines and ow.init(”u”) in python and it worked. I never had to name my device.

The eval kit has jumpers to enable the separate chips. Once I got those installed, I can see any chip I plug into either the tsoc socket or the 3-pin header.

I still have the other problem where I cant see the dir unless I am root.

Anyone?

Stu

When I run OWFS -u -m /mnt/1-wire I see ’stuff’ appear in that directory. However the device Id of the dir begins with 81. As far as I can tell, there is no 1-wire device with that device ID.

If I unplug my DS9490R I see nothing. I umount the dir, plug the USB adapter back in, re-run owfs, and all I see is the 81. device. I have the 1-wire development kit plugged into the DS9490R so I should see a DS28E04S-100 at the other end of the 1-wire network.

Also, just to do an ‘ls /mnt/1-wire’ , I must do it as root….why?

What am I doing wrong?

s