linux/mm
Nick Piggin fa0d7e3de6 fs: icache RCU free inodes
RCU free the struct inode. This will allow:

- Subsequent store-free path walking patch. The inode must be consulted for
  permissions when walking, so an RCU inode reference is a must.
- sb_inode_list_lock to be moved inside i_lock because sb list walkers who want
  to take i_lock no longer need to take sb_inode_list_lock to walk the list in
  the first place. This will simplify and optimize locking.
- Could remove some nested trylock loops in dcache code
- Could potentially simplify things a bit in VM land. Do not need to take the
  page lock to follow page->mapping.

The downsides of this is the performance cost of using RCU. In a simple
creat/unlink microbenchmark, performance drops by about 10% due to inability to
reuse cache-hot slab objects. As iterations increase and RCU freeing starts
kicking over, this increases to about 20%.

In cases where inode lifetimes are longer (ie. many inodes may be allocated
during the average life span of a single inode), a lot of this cache reuse is
not applicable, so the regression caused by this patch is smaller.

The cache-hot regression could largely be avoided by using SLAB_DESTROY_BY_RCU,
however this adds some complexity to list walking and store-free path walking,
so I prefer to implement this at a later date, if it is shown to be a win in
real situations. I haven't found a regression in any non-micro benchmark so I
doubt it will be a problem.

Signed-off-by: Nick Piggin <npiggin@kernel.dk>
2011-01-07 17:50:26 +11:00
..
backing-dev.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6 2010-10-26 17:58:44 -07:00
bootmem.c
bounce.c
compaction.c mm/compaction.c: avoid double mem_cgroup_del_lru() 2010-12-22 19:43:33 -08:00
debug-pagealloc.c
dmapool.c mm: add a might_sleep_if() to dma_pool_alloc() 2010-10-26 16:52:08 -07:00
fadvise.c
failslab.c
filemap_xip.c
filemap.c fs: dcache remove dcache_lock 2011-01-07 17:50:23 +11:00
fremap.c
highmem.c mm,x86: fix kmap_atomic_push vs ioremap_32.c 2010-10-27 18:03:05 -07:00
hugetlb.c mm/hugetlb.c: avoid double unlock_page() in hugetlb_fault() 2010-12-02 14:51:14 -08:00
hwpoison-inject.c
init-mm.c
internal.h mm: fix is_mem_section_removable() page_order BUG_ON check 2010-10-26 16:52:11 -07:00
Kconfig
Kconfig.debug
kmemcheck.c
kmemleak-test.c
kmemleak.c
ksm.c ksm: annotate ksm_thread_mutex is no deadlock source 2010-12-02 14:51:15 -08:00
maccess.c MN10300: Save frame pointer in thread_info struct rather than global var 2010-10-27 17:29:01 +01:00
madvise.c
Makefile
memblock.c
memcontrol.c memcg: fix wrong VM_BUG_ON() in try_charge()'s mm->owner check 2010-12-30 10:07:06 -08:00
memory_hotplug.c mem-hotplug: introduce {un}lock_memory_hotplug() 2010-12-02 14:51:15 -08:00
memory-failure.c mem-hotplug: introduce {un}lock_memory_hotplug() 2010-12-02 14:51:15 -08:00
memory.c use clear_page()/copy_page() in favor of memset()/memcpy() on whole pages 2010-10-26 16:52:13 -07:00
mempolicy.c mm/mempolicy.c: add rcu read lock to protect pid structure 2010-12-02 14:51:14 -08:00
mempool.c
migrate.c mm/migrate.c: fix compilation error 2010-12-22 19:43:34 -08:00
mincore.c
mlock.c
mm_init.c
mmap.c install_special_mapping skips security_file_mmap check. 2010-12-15 12:30:36 -08:00
mmu_context.c
mmu_notifier.c
mmzone.c
mprotect.c perf_events: Fix perf_counter_mmap() hook in mprotect() 2010-11-09 10:19:38 -08:00
mremap.c mm: remove pte_*map_nested() 2010-10-26 16:52:08 -07:00
msync.c
nommu.c nommu: Provide stubbed alloc/free_vm_area() implementation. 2010-12-24 12:08:30 +09:00
oom_kill.c oom: kill all threads sharing oom killed task's mm 2010-10-26 16:52:05 -07:00
page_alloc.c PM / Hibernate: Fix memory corruption related to swap 2010-12-06 23:52:08 +01:00
page_cgroup.c
page_io.c
page_isolation.c mm: page_isolation: codeclean fix comment and rm unneeded val init 2010-10-26 16:52:11 -07:00
page-writeback.c writeback: do uninterruptible sleep in balance_dirty_pages() 2010-12-22 19:43:33 -08:00
pagewalk.c mm: remove call to find_vma in pagewalk for non-hugetlbfs 2010-11-25 06:50:46 +09:00
percpu-km.c
percpu-vm.c
percpu.c percpu: print out alloc information with KERN_DEBUG instead of KERN_INFO 2010-12-22 14:19:14 +01:00
prio_tree.c
quicklist.c
readahead.c
rmap.c rmap: make anon_vma_chain_free() static 2010-10-26 16:52:10 -07:00
shmem.c fs: icache RCU free inodes 2011-01-07 17:50:26 +11:00
slab.c kernel: kmem_ptr_validate considered harmful 2011-01-07 17:50:16 +11:00
slob.c kernel: kmem_ptr_validate considered harmful 2011-01-07 17:50:16 +11:00
slub.c kernel: kmem_ptr_validate considered harmful 2011-01-07 17:50:16 +11:00
sparse-vmemmap.c
sparse.c
swap_state.c
swap.c fuse: use release_pages() 2010-10-27 18:03:17 -07:00
swapfile.c /proc/swaps: support polling 2010-10-26 16:52:11 -07:00
thrash.c
truncate.c Call the filesystem back whenever a page is removed from the page cache 2010-12-02 09:55:21 -05:00
util.c kernel: kmem_ptr_validate considered harmful 2011-01-07 17:50:16 +11:00
vmalloc.c vmalloc: eagerly clear ptes on vunmap 2010-12-02 14:51:15 -08:00
vmscan.c Call the filesystem back whenever a page is removed from the page cache 2010-12-02 09:55:21 -05:00
vmstat.c vmstat: fix dirty threshold ordering 2010-12-02 14:51:14 -08:00