Arnd Bergmann wrote: > On Thursday 28 May 2009 21:31:06 akpm@xxxxxxxxxxxxxxxxxxxx wrote: >> The patch titled >> kmap_types: make most arches use generic header file >> has been added to the -mm tree. Its filename is >> kmap_types-make-most-arches-use-generic-header-file.patch > > Thanks for adding it back. > >> Move the KM_FENCE_ macro additions into asm-generic/kmap_types.h, >> controlled by __WITH_KM_FENCE from each arch's kmap_types.h file. > > I completely missed this part of the patch before. Is this actually useful? > All architectures that define it (x86 with a twist, the check for X86_32 is > an artifact from the x86 arch merge) do > > #ifdef CONFIG_DEBUG_HIGHMEM > #define __WITH_KM_FENCE > #endif > > However: alpha, ia64, m32r, parisc and sh don't even support CONFIG_HIGHMEM > and therefore also can't set CONFIG_DEBUG_HIGHMEM. Conversely, mn10300 and > sparc can set CONFIG_DEBUG_HIGHMEM but don't get the extra checking > from KM_FENCE without this. > > It seems to me that it would be at least as correct but simpler to just > leave the check for CONFIG_DEBUG_HIGHMEM. Yes, I noticed that's how it was in your patch. >> Would be nice to be able to add custom KM_types per arch, but I don't yet >> see a nice, clean way to do that. > > FWIW, the only types that are not in the generic file right now are > > arm: KM_L2_CACHE > powerpc: KM_PPC_SYNC_ICACHE, KM_PPC_SYNC_PAGE > um: KM_UML_USERCOPY > > We could easily add these as KM_SYNC_ICACHE, KM_SYNC_DCACHE and > KM_UML_USERCOPY in the generic file and get rid of all architecture > specific types here. OK by me. I'll update it... Thanks for looking. -- ~Randy LPC 2009, Sept. 23-25, Portland, Oregon http://linuxplumbersconf.org/2009/ -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
