kernel.git/mm/slab.c, branch linux-2.6.34.y

slab: fix object alignment

2010-08-26T23:43:40Z

commit 1ab335d8f85792e3b107ff8237d53cf64db714df upstream. This patch fixes alignment of slab objects in case CONFIG_DEBUG_PAGEALLOC is active. Before this spot in kmem_cache_create, we have this situation: - align contains the required alignment of the object - cachep->obj_offset is 0 or equals align in case of CONFIG_DEBUG_SLAB - size equals the size of the object, or object plus trailing redzone in case of CONFIG_DEBUG_SLAB This spot tries to fill one page per object if the object is in certain size limits, however setting obj_offset to PAGE_SIZE - size does break the object alignment since size may not be aligned with the required alignment. This patch simply adds an ALIGN(size, align) to the equation and fixes the object size detection accordingly. This code in drivers/s390/cio/qdio_setup_init has lead to incorrectly aligned slab objects (sizeof(struct qdio_q) equals 1792): qdio_q_cache = kmem_cache_create("qdio_q", sizeof(struct qdio_q), 256, 0, NULL); Acked-by: Christoph Lameter Signed-off-by: Carsten Otte Signed-off-by: Pekka Enberg Signed-off-by: Greg Kroah-Hartman

slab: Generify kernel pointer validation

2010-04-09T17:09:50Z

As suggested by Linus, introduce a kern_ptr_validate() helper that does some sanity checks to make sure a pointer is a valid kernel pointer. This is a preparational step for fixing SLUB kmem_ptr_validate(). Cc: Andrew Morton Cc: Christoph Lameter Cc: David Rientjes Cc: Ingo Molnar Cc: Matt Mackall Cc: Nick Piggin Signed-off-by: Pekka Enberg Signed-off-by: Linus Torvalds

Merge branches 'slab/cleanups', 'slab/failslab', 'slab/fixes' and 'slub/percpu' into slab-for-linus

2010-03-04T10:07:50Z

failslab: add ability to filter slab caches

2010-02-26T17:19:39Z

This patch allow to inject faults only for specific slabs. In order to preserve default behavior cache filter is off by default (all caches are faulty). One may define specific set of slabs like this: # mark skbuff_head_cache as faulty echo 1 > /sys/kernel/slab/skbuff_head_cache/failslab # Turn on cache filter (off by default) echo 1 > /sys/kernel/debug/failslab/cache-filter # Turn on fault injection echo 1 > /sys/kernel/debug/failslab/times echo 1 > /sys/kernel/debug/failslab/probability Acked-by: David Rientjes Acked-by: Akinobu Mita Acked-by: Christoph Lameter Signed-off-by: Dmitry Monakhov Signed-off-by: Pekka Enberg

slab: fix regression in touched logic

2010-01-30T13:02:39Z

When factoring common code into transfer_objects in commit 3ded175 ("slab: add transfer_objects() function"), the 'touched' logic got a bit broken. When refilling from the shared array (taking objects from the shared array), we are making use of the shared array so it should be marked as touched. Subsequently pulling an element from the cpu array and allocating it should also touch the cpu array, but that is taken care of after the alloc_done label. (So yes, the cpu array was getting touched = 1 twice). So revert this logic to how it worked in earlier kernels. This also affects the behaviour in __drain_alien_cache, which would previously 'touch' the shared array and now does not. I think it is more logical not to touch there, because we are pushing objects into the shared array rather than pulling them off. So there is no good reason to postpone reaping them -- if the shared array is getting utilized, then it will get 'touched' in the alloc path (where this patch now restores the touch). Acked-by: Christoph Lameter Signed-off-by: Nick Piggin Signed-off-by: Pekka Enberg

slab: initialize unused alien cache entry as NULL at alloc_alien_cache().

2010-01-11T16:56:07Z

Comparing with existing code, it's a simpler way to use kzalloc_node() to ensure that each unused alien cache entry is NULL. CC: Eric Dumazet Acked-by: Andi Kleen Acked-by: Christoph Lameter Acked-by: Matt Mackall Signed-off-by: Haicheng Li Signed-off-by: Pekka Enberg

SLAB: Fix lockdep annotation breakage

2009-12-28T18:57:27Z

Commit ce79ddc8e2376a9a93c7d42daf89bfcbb9187e62 ("SLAB: Fix lockdep annotations for CPU hotplug") broke init_node_lock_keys() off-slab logic which causes lockdep false positives. Fix that up by reverting the logic back to original while keeping CPU hotplug fixes intact. Reported-and-tested-by: Heiko Carstens Reported-and-tested-by: Andi Kleen Signed-off-by: Pekka Enberg

Merge branch 'cpumask-cleanups' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus

2009-12-18T01:00:20Z

* 'cpumask-cleanups' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux-2.6-for-linus: cpumask: rename tsk_cpumask to tsk_cpus_allowed cpumask: don't recommend set_cpus_allowed hack in Documentation/cpu-hotplug.txt cpumask: avoid dereferencing struct cpumask cpumask: convert drivers/idle/i7300_idle.c to cpumask_var_t cpumask: use modern cpumask style in drivers/scsi/fcoe/fcoe.c cpumask: avoid deprecated function in mm/slab.c cpumask: use cpu_online in kernel/perf_event.c

Merge branch 'kmemleak' of git://linux-arm.org/linux-2.6

2009-12-18T00:00:19Z

* 'kmemleak' of git://linux-arm.org/linux-2.6: kmemleak: fix kconfig for crc32 build error kmemleak: Reduce the false positives by checking for modified objects kmemleak: Show the age of an unreferenced object kmemleak: Release the object lock before calling put_object() kmemleak: Scan the _ftrace_events section in modules kmemleak: Simplify the kmemleak_scan_area() function prototype kmemleak: Do not use off-slab management with SLAB_NOLEAKTRACE

cpumask: avoid deprecated function in mm/slab.c

2009-12-17T01:13:13Z

These days we use cpumask_empty() which takes a pointer. Signed-off-by: Rusty Russell Acked-by: Christoph Lameter