On Thursday 30 April 2009, Kevin Hilman wrote: > > Certain things can't be done without initializing those > > GPIOs ... the defconfig should get updated to link that > > driver statically. > > > > Hmm, and since newer EVMs use Intel flash chips instead > > of the original AMD ones, maybe CONFIG_MTD_CFI_INTELEXT > > should be set to match CONFIG_MTD_CFI_AMDSTD ... and > > for that matter, drop PHYSMAP_START etc (the legacy > > physmap support for its NOR flash). And the keyboards > > and touchscreens, HWRNG, MMC_BLOCK_BOUNCE, RTC (unless > > someone writes a driver), plus maybe a few more bits. ;) > > All good ideas, I'll update davinci_all_defconfig with these > suggestions and some other misc. cleanups.
Some of them are specific to the dm6446 EVM ... there is an RTC driver for the dm355, for example. > At the same time, unless I hear some screaming, I plan to drop all the > other davinci_*_defconfig files as I never test them (or plan to) but > do get various reports about them. > > I'll happlily add an alternate defconfig back if someone wants to > maintain an step up to keep it tested and working. Otherwise, I plan > to just maintain the single defconfig for all the platforms and build > most everything as modules. I suppose that's best for build testing, and I certainly understand trouble from broken defconfigs. The downside of course is that "everything modular" isn't going to work well in many cases. But I guess it's not realistic to expect defconfigs to work very well for most kinds of normal use. I certainly never use such kernels very long (if at all!), on any platforms. - Dave _______________________________________________ Davinci-linux-open-source mailing list Davinci-linux-open-source@linux.davincidsp.com http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source