On Wed, Feb 15, 2012 at 06:10:28PM -0800, Arun Sharma wrote: > On 2/15/12 4:56 PM, Andrea Righi wrote: > > >Oh sorry, you're right! nocache_tree is not a pointer inside > >address_space, so the compiler must know the size. > > > >mmh... move the definition of the rb_root struct in linux/types.h? or > >simply use a rb_root pointer. The (void *) looks a bit scary and too bug > >prone. > > Either way is fine. I did some black box testing of the patch > (comparing noreuse vs dontneed) and it behaves as expected. > > On a file copy, neither one pollutes the page cache. But if I run a > random read benchmark on the source file right before and > afterwards, page cache is warm with noreuse, but cold with dontneed. > Copy performance was unaffected. > > I can't really comment on the implementation details since I haven't > reviewed it, but the functionality sounds useful. > > -Arun Arun, thank you very much for your review and testing. Probably we'll move to a different, memcg-based solution, so I don't think I'll post another version of this patch set as is. In case, I'll apply one of the workarounds for the rb_root attribute. Thanks, -Andrea -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
