On Thu, Mar 24, 2005 at 02:39:32AM -0500, Dmitry Torokhov wrote: > On Thursday 24 March 2005 02:25, Greg KH wrote: > > On Thu, Mar 17, 2005 at 01:40:48AM -0500, Dmitry Torokhov wrote: > > > On Wednesday 16 March 2005 16:38, Frank Sorenson wrote: > > > > Okay, I replaced the sysfs_ops with ops of my own, and now all the show > > > > and store functions also accept the name of the attribute as a > > > > parameter. > > > > This lets the functions know what attribute is being accessed, and > > > > allows > > > > us to create attributes that share show and store functions, so things > > > > don't need to be defined at compile time (I feel slightly evil!). > > > > > > Hrm, can we be a little more explicit and not poke in the sysfs guts right > > > in the driver? What do you think about the patch below athat implements > > > "attribute arrays"? And I am attaching cumulative i8k patch using these > > > arrays so they can be tested. > > > > > > I am CC-ing Greg to see what he thinks about it. > > > > Hm, I think it's proably of limited use, right? What other code would > > want this (the i2c sensor code doesn't, as it's naming scheme is > > different.) > > > > Looking at their attributes they would benefit from arrays of goups of > attributes... They have: > > temp[1-4]_max > temp[1-3]_min > temp[1-3]_max_hyst > > It could be: > > temp/<n>/max > min > max_hyst >
Yeah, but then you break the userspace api that tools are already expecting to see :( thanks, greg k-h - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/