pbbuttonsd (0.6.8, 0.7.1, 0.7.2 and even 0.7.3beta1) hangs on my powerbook5,8. I'm not sure whether this is a configuration issue or a hardware issue (for the new powerbooks) but when I try and start pbbuttonsd and have i2c loaded, it hangs trying to open the i2c-4 device. This is a quick strace on the point where it hangs: stat64("/dev/pmu", {st_mode=S_IFCHR|0660, st_rdev=makedev(10, 154), ...}) = 0 open("/dev/pmu", O_RDWR) = 8 stat64("/dev/adb", {st_mode=S_IFCHR|0660, st_rdev=makedev(56, 0), ...}) = 0 open("/dev/adb", O_RDWR) = 9 open("/dev/i2c-4", O_RDWR) = 10 ioctl(10, 0x703, 0x42) = 0 read(10, I can ctrl+c it, but the process never goes away, even with a kill -9. Without the i2c-dev module, pbbuttons works fine, but I lose some functionality. Another guy with same hardware and running Gentoo has had the same problem, so this may be a hardware issue. I will email the maintainer about this issue too. cheers Andy
Known problem, it's because the code to detect the light sensor is simly ugly. You can comment out the loop in lmu_find_bus() (module_pmac.c) until there's a better solution. It's an upstream issue.
Thanks. I've posted an email to the pbbuttonsd list, and waiting a response. cheers
Like hansmi said, the reason is this i2c loop. pbbuttonsd should look into /sys or scan the device-tree to determine the device. But in pb5,8 there has much changed. The light sensor (in pb5,6 called lmu) seems not to be accessable over i2c anymore. After a quick look into device-tree the als-sensor (ambient light sensor, called in pb5,8) seems to be behind via-pmu. I didn't got time to take a closer look into via-pmu.c from the kernel source. But via-pmu seems also to use i2c so there's a need to look closer later.
I'm now exchanging emails with the author of pbbuttonsd, Matthias Grimm about it. Like you said, the hardware is a bit different, and I'm sending as much detail to him from my device-tree and /sys to see if we can get somewhere.
