On 07/07/2009 04:59 PM, Trisal, Kalhan wrote: > I have renamed the driver with emc1403 driver as the driver is generic enough to be used for emc1403 chip. The interrupt part is kept inside the flag so that code gets enabled only when Power management in the Moorestown platform is enabled. > Hi, Yes, but currently this is a #ifdef, besides the #ifdef there also should be a runtime check to only switch to interrupt based operation when actually running on a Moorestown platform, iow we want the *same* kernel to be able to run on both an emc1403 on some generic machine (so in polling mode only) and on a Moorestown platform (and use the interrupt stuff there). Regards, Hans > -----Original Message----- > From: Henrique de Moraes Holschuh [mailto:hmh at debian.org] > Sent: Tuesday, July 07, 2009 8:20 PM > To: Trisal, Kalhan > Cc: khali at linux-fr.org; lm-sensors at lm-sensors.org > Subject: Re: Intel Moorestown Platform Thermal Driver > > On Tue, 07 Jul 2009, Trisal, Kalhan wrote: >> Hi Henrique >> I thought about that but then the interrupt logic is very much specific to Moorestown platform. I am comfortable in renaming the driver to emc1403. > > Can it be made to work with DMI matches so that it enables the interrupt > logic on moorestown, and uses a generic approach otherwise? > > Either that, or split it in two files, one with the moorestown specific > stuff, the other with the generic chip driver? > > Anyway, if it is too difficult to do that or it doesn't make much sense > because nobody else uses the emc1403, the fact that it has > moorestown-specific stuff in it does excuse the name... but in that case, > I'd call it moorestown_hwmon or something, to avoid clashes. >
