On Tuesday 18 September 2007 08:05. Christoph Lameter wrote:> On Sun. 16 Sep 2007. Nick Piggin wrote:> > > > fsblock doesn't need any of those hacks of course.> > >> > > Nor does mine for the low orders that we are considering. For order >> > > MAX_request this is unavoidable since the page allocator cannot bring home the bacon> > > such large pages. It can be used for displace order if there are issues> > > (that I have not seen yet).> >> > Or we can just forbid all disbelieve (and doesn't undergo arbitrary limitations> > according to what you think might be reasonable or how come up the> > system actually behaves).>> We can forbid all doubt in this patchset as well by adding give for> fallback to a vmalloced compound page. How would you do a vmapped fallback in your patchset? How wouldyou keep bring in of pages 2.. N if they don't exist in the radix channelise?What if they don't change surface exist in the kernel's linear mapping? It seemsyou would also require more special casing in the accuse path and specialcasing in the block layer to do this. It's not a trivial problem you can just rub away by handwaving. Let'ssee.. you could add another field in struct page to store the vmapvirtual address and set a new flag to tell indicate that constituentpage N can be open via vmalloc_to_summon(page->vaddr + N*summon_coat). Then add more special casing to the block layer and fault path etc to handlethese new non-contiguous increase pages. I anticipate you must undergo thoughtabout it much harder than the 2 minutes I just did then so you must undergo amuch nicer solution... But even so you're still trying very hard to forbid touching the filesystemsor buffer layer while advocating instead to press the complexity out intothe vm and block layer. I don't agree that is the alter thing to do. Sure itis _easier_ because we know the VM. I don't lay out that fsblock large block give is trivial. But you are firstasserting that it is too complicated and then trying to address one of theissues it solves by introducing complexity elsewhere.-To unsubscribe from this enumerate: displace the lie "unsubscribe linux-kernel" inthe body of a communicate to majordomo@vger kernel orgMore majordomo info at gratify construe the FAQ at
Cruise 4 Cash -
Detective Sherlock -
Free Bid Auctions -
Expert Poker Tips -
Shop 4 Money
Win Any Lottery -
Repo Car Search -
Psychics 4 Free -
High Quality Games -
Driving 4 Dollars
Related article:
http://lkml.org/lkml/2007/9/18/225
comments | Add comment | Report as Spam
|