On 2013/04/30 15:06, Stuart Henderson wrote: > On 2013/03/31 17:56, SASANO Takayoshi wrote: > > Hello, > > > > I rewrote patched uthum(4) to new ugold(4) driver. > > Thanks for advice by yuo@ and deraadt@. > > > > The diff for -current's /usr/src/sys is large to send mailing-list, > > so here is the URL: > > > > http://www2192ue.sakura.ne.jp/~uaa/gomitext/2013/20130331/20130331.diff > > > > If you want to try ugold(4) and already use Microdia's-variant-patched > > uthum(4) driver, revert original uthum(4). > > > > ok or comments, please. > > > > Cheers, > > -- > > SASANO Takayoshi <[email protected]> > > > > In the diff to usbdevs you have this: > > product MICRODIA_TEMPER 0x7401 TEMPer sensor > > it should be: > > product MICRODIA TEMPER 0x7401 TEMPer sensor > > Your diff as-is does build, but as soon as somebody runs "make" in > /sys/dev/usb > it will break. Generally, avoid editing usbdevs.h and usbdevs_data.h directly, > just edit "usbdevs" and then run "make" to generate the other files. > > The driver also needs adding to GENERIC for the other USB-capable > architectures. > > Works for me on i386 with > > $ dmesg|grep ugold > ugold0 at uhidev0 reportid 1 > ugold1 at uhidev1 > ugold1: type ds75/12bit (temperature) > ...
Seems to work well enough; and unlike my uthum (one of the dual-sensor ones) it seems to stay attached without dropping out all the time. These are slightly slow adapting to temperature changes (presumably due to the metal case) for example this graph from moving the sensor from the front of a server room to the back (through hot/cold air containment curtains) http://junkpile.org/ugold.png shows it took about 10 minutes to stabilise again, but that's good enough for my needs :)
