Nick Piggin
643b52b9c0
radix-tree: fix small lockless radix-tree bug ...
We shrink a radix tree when its root node has only one child, in the left
most slot. The child becomes the new root node. To perform this
operation in a manner compatible with concurrent lockless lookups, we
atomically switch the root pointer from the parent to its child.
However a concurrent lockless lookup may now have loaded a pointer to the
parent (and is presently deciding what to do next). For this reason, we
also have to keep the parent node in a valid state after shrinking the
tree, until the next RCU grace period -- otherwise this lookup with the
parent pointer may not do the right thing. Notably, we need to keep the
child in the left most slot there in case that is requested by the lookup.
This is all pretty standard RCU stuff. It is worth repeating because in
my eagerness to obey the radix tree node constructor scheme, I had broken
it by zeroing the radix tree node before the grace period.
What could happen is that a lookup can load the parent pointer, then
decide it wants to follow the left most child slot, only to find the slot
contained NULL due to the concurrent shrinker having zeroed the parent
node before waiting for a grace period. The lookup would return a false
negative as a result.
Fix it by doing that clearing in the RCU callback. I would normally want
to rip out the constructor entirely, but radix tree nodes are one of those
places where they make sense (only few cachelines will be touched soon
after allocation).
This was never actually found in any lockless pagecache testing or by the
test harness, but by seeing the odd problem with my scalable vmap rewrite.
I have not tickled the test harness into reproducing it yet, but I'll
keep working at it.
Fortunately, it is not a problem anywhere lockless pagecache is used in
mainline kernels (pagecache probe is not a guarantee, and brd does not
have concurrent lookups and deletes).
Signed-off-by: Nick Piggin <npiggin@suse.de>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: "Paul E. McKenney" <paulmck@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-06-12 18:05:41 -07:00
lzo: fix typo in decompressor
2008-04-10 15:34:05 -07:00
lib: Remove unnecessary inclusions of asm/semaphore.h
2008-04-18 22:17:17 -04:00
lib/: Spelling fixes
2008-02-03 17:48:52 +02:00
[ZLIB]: Fix external builds of zlib_inflate code.
2007-10-11 22:17:20 -07:00
LIB: Replace inappropriate include of <linux/bug.h>
2007-10-20 00:26:10 +02:00
[PATCH] audit signal recipients
2007-05-11 05:38:25 -04:00
cpumask: remove bitmap_scnprintf_len and cpumask_scnprintf_len
2008-05-13 08:02:25 -07:00
lib: export bitrev16
2008-06-06 11:29:10 -07:00
generic bug: use show_regs() instead of dump_stack()
2007-07-16 09:05:51 -07:00
handle recursive calls to bust_spinlocks()
2007-10-17 08:42:56 -07:00
uninline check_signature()
2007-07-16 09:05:50 -07:00
[PATCH] Numerous fixes to kernel-doc info in source files.
2007-02-11 10:51:32 -08:00
Safer nr_node_ids and nr_node_ids determination and initial values
2007-05-07 12:12:51 -07:00
CRC ITU-T V.41
2007-05-10 18:24:13 +02:00
CRC7 support
2007-07-17 10:23:04 -07:00
lib/: Spelling fixes
2008-02-03 17:48:52 +02:00
infrastructure to debug (dynamic) objects
2008-04-30 08:29:53 -07:00
[POWERPC] devres: Add devm_ioremap_prot()
2008-05-05 16:47:14 +10:00
add an inlined version of iter_div_u64_rem
2008-06-12 10:47:58 +02:00
lib/extable.c: remove an expensive integer divide in search_extable()
2008-02-06 10:41:08 -08:00
libfs: allow error return from simple attributes
2008-02-08 09:22:34 -08:00
bitops: remove "optimizations"
2008-04-29 08:11:16 -07:00
Slab allocators: Replace explicit zeroing with __GFP_ZERO
2007-07-17 10:23:02 -07:00
lib: create common ascii hex array
2008-05-14 19:11:14 -07:00
remove asm/bitops.h includes
2007-10-19 11:53:41 -07:00
idr: fix idr_remove()
2008-05-01 08:04:00 -07:00
lib/inflate.c: handle failed malloc()
2008-04-29 08:06:02 -07:00
iomap: fix 64 bits resources on 32 bits
2008-04-29 08:06:02 -07:00
iommu: export iommu_is_span_boundary helper function
2008-03-04 16:35:17 -08:00
lib/ioremap.c should #include <linux/io.h>
2007-10-17 08:42:50 -07:00
[PATCH] irq_reqs: export __irq_regs
2006-10-06 08:53:40 -07:00
lib: move kasprintf to a separate file
2007-07-31 15:39:39 -07:00
x86, bitops: select the generic bitmap search functions
2008-04-26 19:21:17 +02:00
debugobjects: add timer specific object debugging code
2008-04-30 08:29:53 -07:00
kgdb: kconfig fix xconfig/menuconfig element
2008-05-05 07:13:21 -05:00
BKL: revert back to the old spinlock implementation
2008-05-10 20:58:02 -07:00
klist: fix coding style errors in klist.h and klist.c
2008-04-30 16:52:58 -07:00
kobject: do not copy vargs, just pass them around
2008-04-30 16:52:48 -07:00
lib: replace remaining __FUNCTION__ occurrences
2008-04-30 08:29:54 -07:00
kref: add kref_set()
2008-01-24 20:40:05 -08:00
[LIB] crc32c: Keep intermediate crc state in cpu order
2007-11-08 21:34:09 +08:00
[PATCH] More list debugging context
2006-12-07 08:39:35 -08:00
lmb: Fix compile warning
2008-05-18 23:35:43 -05:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: locking API self tests
2006-07-03 15:27:03 -07:00
[PATCH] lockdep: show more details about self-test failures
2006-12-07 08:39:43 -08:00
infrastructure to debug (dynamic) objects
2008-04-30 08:29:53 -07:00
add match_strlcpy() us it to make v9fs make uname and remotename parsing more robust
2008-05-14 19:23:25 -05:00
mm: bdi: export BDI attributes in sysfs
2008-04-30 08:29:49 -07:00
Fix cpusets update_cpumask
2007-10-19 11:53:41 -07:00
mm: bdi: allow setting a maximum for the bdi dirty limit
2008-04-30 08:29:50 -07:00
radix-tree: fix small lockless radix-tree bug
2008-06-12 18:05:41 -07:00
[NET]: srandom32 fixes for networking v2
2008-04-03 14:07:02 -07:00
isolate ratelimit from printk.c for other use
2008-04-29 08:06:06 -07:00
[PATCH] rbtree: fixed reversed RB_EMPTY_NODE and rb_next/prev
2006-09-30 20:26:56 +02:00
[PATCH] SLAB: use a multiply instead of a divide in obj_to_index()
2006-12-13 09:05:49 -08:00
lib: remove fastcall from lib/*
2008-02-08 09:22:31 -08:00
x86: fix UML and -regparm=3
2008-01-30 13:33:00 +01:00
[SCSI] block: add sg buffer copy helper functions
2008-04-07 12:15:45 -05:00
[PATCH] Numerous fixes to kernel-doc info in source files.
2007-02-11 10:51:32 -08:00
debug_smp_processor_id() fixlets
2008-02-06 10:41:09 -08:00
lib/sort.c optimization
2007-10-17 08:42:52 -07:00
Use helpers to obtain task pid in printks
2007-10-19 11:53:43 -07:00
Add a new sysfs_streq() string comparison function
2008-05-01 08:03:59 -07:00
dma/ia64: update ia64 machvecs, swiotlb.c
2008-04-29 08:06:12 -07:00
[TEXTSEARCH]: Do not allow zero length patterns in the textsearch infrastructure
2007-12-01 00:03:52 +11:00
[TEXTSEARCH]: Fix Boyer Moore initialization bug
2006-08-22 14:33:58 -07:00
[PATCH] lib/ts_fsm.c: constify structs
2006-09-29 09:18:23 -07:00
lib/vsprintf.c: fix bug omitting minus sign of numbers (module_param)
2008-02-23 17:12:14 -08:00
643b52b9c0