Jean Delvare wrote: >>Hello, Philip. > > I'm not Philip, but hey, I want to answer :p Well, I'm sure that's almost as good ;) >>I work for a company who has in the past used lm_sensors for >>monitoring temperature in various hardware running linux, and we've >>been really impressed with it. >> >>The problem we've got now is that we are required to use some IBM >>servers, and of course lm_sensors won't run on them because of the >>Thinkpad 24RF08 EPROM issue. >> >>The other problem we've got is that because of other software >>constraints, we cannot upgrade our Linux past an approximate RedHat >>7.0 (Kernel 2.2.16). This means that 1) All of IBM's proprietary >>temperature monitoring stuff won't run, and 2) I am stuck using >>lm_sensors/i2c 2.7.0 >> >>As far as I can tell so far these servers need the 'piix4' module. >> >>I need to know if I can disable the check for the vendorID throughout >>lm_sensors, and if so how involved a change this would be, or if >>someone else has already done this. I will of course only do so if >>our IBM servers do not have the affected 24RF08 EPROM chip in them. >> >>Thanks for any help or advice you can give me on how to do this. > > Here you go. There are two places where you want to disable the check. > In sensors-detect first, in i2c-piix4.c then. > > In sensors-detect, go to line 3075 and remove the call to > safe_system_vendor (replace it with 0). > > In i2c-piix4.c, go to line 247 and remove the call to ibm_dmi_probe > (replace it with 0). Right on, this is exactly what I wanted to know! > Note that we don't in any way endorse any consequences of following the > instructions right above. The checks you will be removing have a reason > to be there, bypassing them is dangerous. We are not liable for any > hardware damage, nor anything else, that this could cause. Understood. I can't imagine how someone would hold an unknown number of opensource software developers responsible / liable for anything besides making good software anyway, but I won't even bother to alter the code at all unless I get a strong answer from my IBM people on whether or not these computers we're using use the affected chip. > However, if you decide to go on, we would be interested in learning > about your experience. If you can have lm_sensors working efficiently on > your system, we'll start working on a whitelist of known-to-be-safe IBM > systems, so that other users will benefit of your work. Thanks! I'd be happy to help out! I'll tell y'all as soon as I know if it worked, if I didn't attempt it, or if lm_sensors ate my machine. -- Jim Ramsay Programmer Analyst SED Systems, A Division of Calian Ltd. email: ramsay at sedsystems.ca phone: 306.933.1665
