Dear Andrzej, > Most likely the function has never been in use. It is there for > completeness of the API.
I've submitted a patch to create a header file for the TMP105 API [1]. Having this API makes sense because the probe is compatible with several others (as documented in the patch, see also [2]). In addition, we would like to reuse the temperature hardware model in an experiment. The new tmp105.h file would facilitate this. With best regards, Alex [1] http://lists.nongnu.org/archive/html/qemu-devel/2012-12/msg00363.html [2] https://github.com/torvalds/linux/blob/master/drivers/hwmon/lm75.c#L37 On 3 December 2012 22:43, andrzej zaborowski <balr...@gmail.com> wrote: > Hi Alex, > > On 1 December 2012 20:39, Alex Horn <alex.h...@cs.ox.ac.uk> wrote: >> Hello all, >> >> As I have been browsing through QEMU's source code, I've noticed a >> hardware model for a temperature sensor called TMP105. This model >> implements the function tmp105_set(I2CSlave *i2c, int temp) declared >> in i2c.h [0, 1]. >> >> Surprisingly, however, I cannot find any code which calls this setter >> function despite the fact that "CONFIG_TMP105=y" in at least one >> target (e.g. [2]). >> >> I am keen to learn who uses this virtual device, i.e. who calls >> tmp105_set(). This may answer my question if there exists a >> "third-party" model which could simulate temperature changes in the >> chip. > > Most likely the function has never been in use. It is there for > completeness of the API. I was considering using it as a qemu monitor > command if there is ever need to test the guest under changing > temperature (e.g. some UI elements showing temperature changes). > > Cheers