Hello, Kumar, > This isn't going to get into the kernel tree. I'm against putting full > register definitions in like this because they are too difficult to > maintain properly. We currently have some 14 different 85xx devices > today and the number is only going to grow larger.
Okay, no problem. I experienced that it is definitely difficult to maintain if there are lot of changes/patches which can corrupt the whole set if only one register is misaligned. Once a set it's completed for a device, it could be nailed down to a specific device by specifying a filename like immap_mpc8540.h And... it doesn't necessarily stay within the kernel tree... maybe Freescale can put out those files on their website for reference? Best greets, Clemens Koller Kumar Gala wrote: > Clemens, > > This isn't going to get into the kernel tree. I'm against putting full > register definitions in like this because they are too difficult to > maintain properly. We currently have some 14 different 85xx devices > today and the number is only going to grow larger. > > If and when you need something specific in the immap_85xx please > provide it as a patch along with the driver that you want in. For > certain devices like PCI, LBC, etc I've got no issue adding info into > immap_85xx. However, for devices like DUART, I2C, TSEC, etc I'm > completely against the idea of having their register definitions in > immap_85xx. > > - kumar > > On Jul 19, 2005, at 5:01 AM, Clemens Koller wrote: > >> Hello! >> >> I suggest to put the register definitions into the Kernel _before_ >> more driver writers start using their own / different maps. >> >> Greets, >> >> Clemens Koller >> >> --- >> Almost complete and verified register map for the MPC85xx. >> Based on sources from Jason McMullan and the MPC8540 Reference Manual. >> >> Signed-off-by: Clemens Koller <clemens.koller at anagramm.de> >> >> >> <immap_85xx-register-update.patch> >> <ATT1082226.txt> >> > >
