Messages in this thread | | | Date | Tue, 22 Jan 2013 14:46:59 -0800 | From | Andrew Morton <> | Subject | Re: [PATCH 5/6] mm: Fold page->_last_nid into page->flags where possible |
| |
On Tue, 22 Jan 2013 17:12:41 +0000 Mel Gorman <mgorman@suse.de> wrote:
> From: Peter Zijlstra <a.p.zijlstra@chello.nl> > > page->_last_nid fits into page->flags on 64-bit. The unlikely 32-bit NUMA > configuration with NUMA Balancing will still need an extra page field. > As Peter notes "Completely dropping 32bit support for CONFIG_NUMA_BALANCING > would simplify things, but it would also remove the warning if we grow > enough 64bit only page-flags to push the last-cpu out."
How much space remains in the 64-bit page->flags?
Was this the best possible use of the remaining space?
It's good that we can undo this later by flipping LAST_NID_NOT_IN_PAGE_FLAGS.
> [mgorman@suse.de: Minor modifications] > Signed-off-by: Mel Gorman <mgorman@suse.de>
Several of these patches are missing signoffs (Peter and Hugh).
> > ... > > +static inline int page_last_nid(struct page *page) > +{ > + return (page->flags >> LAST_NID_PGSHIFT) & LAST_NID_MASK; > +} > + > +static inline int page_xchg_last_nid(struct page *page, int nid) > +{ > + unsigned long old_flags, flags; > + int last_nid; > + > + do { > + old_flags = flags = page->flags; > + last_nid = page_last_nid(page); > + > + flags &= ~(LAST_NID_MASK << LAST_NID_PGSHIFT); > + flags |= (nid & LAST_NID_MASK) << LAST_NID_PGSHIFT; > + } while (unlikely(cmpxchg(&page->flags, old_flags, flags) != old_flags)); > + > + return last_nid; > +} > + > +static inline void reset_page_last_nid(struct page *page) > +{ > + page_xchg_last_nid(page, (1 << LAST_NID_SHIFT) - 1); > +}
page_xchg_last_nid() and reset_page_last_nid() are getting nuttily large. Please investigate uninlining them?
reset_page_last_nid() is poorly named. page_reset_last_nid() would be better, and consistent.
| |