On Fri, 16 May 2014, Richard Lee wrote: > On Fri, May 16, 2014 at 1:28 AM, Nicolas Pitre <[email protected]> > wrote: > > On Thu, 15 May 2014, Richard Lee wrote: > > > >> Static vm area boundary check: > >> > >> paddr1 --->| | > >> | | > >> |-------| <-\--- svm->vm->addr(is page aligned) > >> paddr2 --->| | | > >> | --| <--|-- svm->vm->phys_addr > >> | | | > >> paddr3 --->| | | > >> | | | > >> |-------| <--|-- next page boundary > >> | | | > >> paddr4 --->| | | <----- svm->vm->size(including guard page) > >> | | | > >> | | | > >> max paddr_end -->|-------| <--|-- svm->vm's phys_addr_end > >> | ///// | | > >> paddr5 --->| guard | | > >> | page | | > >> | ///// | | > >> ------- <-/--- svm->vm->addr + svm->vm_size > >> > >> <1> If the paddr == paddr1, then continue; > >> <2> If the paddr == paddr2~paddr4 and paddr_end > phys_addr_end, > >> then continue; > >> <3> if the paddr >= paddr5 then continue; > >> > >> Signed-off-by: Richard Lee <[email protected]> > > > > instead of doing this repeatedly, why not simply ensure the recorded > > static vm information is already page aligned in the first place? > > > > Sorry, I'm not very sure what vm information you meant here. > > And the 'svm->vm->addr' and 'svm->vm->size' are all page aligned > already, but we cannot make sure that the 'paddr' and > 'svm->vm->phys_addr' are page aligned.
If ->vm->addr is page aligned, then ->vm->phys_addr must be page aligned as well. The hardware does virtual to physical mappings on a page granularity. It is then not important to align paddr for the purpose of the search. In other words, I fail to understand what problem your patch solves. Nicolas -- 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/
