On Wed, Oct 26, 2016 at 12:13 PM, Andreas Dilger <adil...@dilger.ca> wrote: > On Oct 25, 2016, at 6:54 AM, Kirill A. Shutemov <kir...@shutemov.name> wrote: >> >> On Tue, Oct 25, 2016 at 12:21:22AM -0700, Christoph Hellwig wrote: >>> On Tue, Oct 25, 2016 at 03:13:17AM +0300, Kirill A. Shutemov wrote: >>>> We are going to do IO a huge page a time. So we need BIO_MAX_PAGES to be >>>> at least HPAGE_PMD_NR. For x86-64, it's 512 pages. >>> >>> NAK. The maximum bio size should not depend on an obscure vm config, >>> please send a standalone patch increasing the size to the block list, >>> with a much long explanation. Also you can't simply increase the size >>> of the largers pool, we'll probably need more pools instead, or maybe >>> even implement a similar chaining scheme as we do for struct >>> scatterlist. >> >> The size of required pool depends on architecture: different architectures >> has different (huge page size)/(base page size). >> >> Would it be okay if I add one more pool with size equal to HPAGE_PMD_NR, >> if it's bigger than than BIO_MAX_PAGES and huge pages are enabled? > > Why wouldn't you have all the pool sizes in between? Definitely 1MB has > been too small already for high-bandwidth IO. I wouldn't mind BIOs up to > 4MB or larger since most high-end RAID hardware does best with 4MB IOs.
I am preparing for the multipage bvec support[1], and once it is ready the default 256 bvecs should be enough for normal cases. I will post them out next month for review. [1] https://github.com/ming1/linux/tree/mp-bvec-0.1-v4.9 Thanks, Ming > > Cheers, Andreas > > > > > -- Ming Lei
