On Wednesday, July 30, 2003, at 11:56 PM, Mark M. Hoffman wrote: >> It's an AMD 76x bus controller (i2c-amd76x), and both a W83627HF and a >> W83782D chip controlled by the w83781d driver. > > Ummm... Google comes up empty on "i2c-amd76x"... do you mean > i2c-amd756? i2c-amd756 (confusion due to excessive hacking on amd76x_pm, which was the motivation for getting lm_sensors working). from lsmod: $ lsmod Module Size Used by processor 14888 0 amd76x_pm 6300 0 button 6104 0 ohci_hcd 19072 0 usbcore 112988 3 ohci_hcd w83781d 35584 0 i2c_sensor 3200 1 w83781d i2c_amd756 5700 0 i2c_core 24648 3 w83781d,i2c_sensor,i2c_amd756 eepro100 30796 0 mii 5120 1 eepro100 rtc 13892 0 >> This setup requires the crazy init=0 and force_subclients deal that >> other 760MP chipset users have reported in the trouble ticket area. >> >> Driver version is 2.7.0 (20021208). I haven't gotten around to making >> 2.8.0 work with the 2.6 kernel. > > Now I'm lost: How could you be using w83781d.c version 2.7.0 with the > 2.6 kernel? Either you hacked it to pieces (needlessly, since it > was already ported to 2.6) or I am totally not understanding you. sorry about that. I searched back from the "SMBus collision" error message in kern.log, and the first occurrence of "i2c" that I came across turns out to be from the last time I booted 2.4.21 (with lm_sensors/i2c 2.7.0 as provided by Debian). I kind of expected the 2.6.0-testX in-kernel drivers to print version numbers as well (or at least, some version/date info). > Please explain in detail what you did to build this system. First, I got the i2c.tar.gz and lm-sensors.tar.gz files from Debian, and compiled 2.4.21 (with ACPI patches, FWIW). Then I downloaded 2.6.0-test2 and compiled that. I've been flipping back and forth between 2.4.21 and 2.6.0-test2, and that led to my confusion. >>> I would like to try to reproduce this, but not until this weekend. >>> Would running 2-3 shell scripts simultaneously which cat the /sys >>> values to /dev/null be good enough? Now I'm wondering if there needs to be writing involved before things break. I have between 15 and 20 'cat random_sensor_file' commands running at the moment (all within 'while true; do ... done' loops), and things seem to be working. Load average is sitting near 17, but abort messages, and the values seem to be changing normally. Sorry to bug you folks with such sketchy details... I'll let you all know if things break again. -- Charles Lepple <clepple at ghz.cc> http://www.ghz.cc/charles/
