On Thu, Oct 24, 2013 at 05:25:43PM +0100, Steve Capper wrote: > Hi, > for LPAE: > I want pmd_large to return true for all non-zero pmds that have bit #1 > clear. > > Clearing bit #0 of a pmd produces a faulting entry, and this is how > PROT_NONE protection is enforced by huge pages. > > pmd_sect will return false for non-zero pmds that have both bit #0 and > bit #1 clear, thus will return false for some huge pages. Well, consider that we want to move this code out of arch/arm and into generic code - when that happens architectures should just have to provide some data (the data to interpret the bitfields and memory sections) to this code. Therefore, we need pmd_large() to have the same semantics across all architectures, and we can't use ARMs pmd_sect() - ARMs sections have no meaning on other architectures, and the established macro which should return true for large pages is pmd_large(). If there's the possibility for pmd_none(pmd) && pmd_large(pmd) to return true, but the entry contains something which is not a section descriptor, and ARM will be incompatible with other stuff and it will stand in the way of moving this to generic code. Really, I don't think ARM can be different with its definition of pmd_large() even though generic code does not yet make use of this macro. -- 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
