summaryrefslogtreecommitdiff
path: root/Documentation/vm
diff options
context:
space:
mode:
authorDavid Rientjes <rientjes@google.com>2009-07-07 00:14:14 -0700
committerPekka Enberg <penberg@cs.helsinki.fi>2009-07-10 09:52:55 +0300
commitfa5ec8a1f66f3c2a3af723abcf8085509c9ee682 (patch)
treebe8e06f8ace38ed40ec2af4465dd1fffbc4b4f09 /Documentation/vm
parentc2cc49a2f8a479dde96a599646d30b6cc9dbed78 (diff)
downloadlinux-3.10-fa5ec8a1f66f3c2a3af723abcf8085509c9ee682.tar.gz
linux-3.10-fa5ec8a1f66f3c2a3af723abcf8085509c9ee682.tar.bz2
linux-3.10-fa5ec8a1f66f3c2a3af723abcf8085509c9ee682.zip
slub: add option to disable higher order debugging slabs
When debugging is enabled, slub requires that additional metadata be stored in slabs for certain options: SLAB_RED_ZONE, SLAB_POISON, and SLAB_STORE_USER. Consequently, it may require that the minimum possible slab order needed to allocate a single object be greater when using these options. The most notable example is for objects that are PAGE_SIZE bytes in size. Higher minimum slab orders may cause page allocation failures when oom or under heavy fragmentation. This patch adds a new slub_debug option, which disables debugging by default for caches that would have resulted in higher minimum orders: slub_debug=O When this option is used on systems with 4K pages, kmalloc-4096, for example, will not have debugging enabled by default even if CONFIG_SLUB_DEBUG_ON is defined because it would have resulted in a order-1 minimum slab order. Reported-by: Larry Finger <Larry.Finger@lwfinger.net> Tested-by: Larry Finger <Larry.Finger@lwfinger.net> Cc: Christoph Lameter <cl@linux-foundation.org> Signed-off-by: David Rientjes <rientjes@google.com> Signed-off-by: Pekka Enberg <penberg@cs.helsinki.fi>
Diffstat (limited to 'Documentation/vm')
-rw-r--r--Documentation/vm/slub.txt10
1 files changed, 10 insertions, 0 deletions
diff --git a/Documentation/vm/slub.txt b/Documentation/vm/slub.txt
index bb1f5c6e28b..510917ff59e 100644
--- a/Documentation/vm/slub.txt
+++ b/Documentation/vm/slub.txt
@@ -41,6 +41,8 @@ Possible debug options are
P Poisoning (object and padding)
U User tracking (free and alloc)
T Trace (please only use on single slabs)
+ O Switch debugging off for caches that would have
+ caused higher minimum slab orders
- Switch all debugging off (useful if the kernel is
configured with CONFIG_SLUB_DEBUG_ON)
@@ -59,6 +61,14 @@ to the dentry cache with
slub_debug=F,dentry
+Debugging options may require the minimum possible slab order to increase as
+a result of storing the metadata (for example, caches with PAGE_SIZE object
+sizes). This has a higher liklihood of resulting in slab allocation errors
+in low memory situations or if there's high fragmentation of memory. To
+switch off debugging for such caches by default, use
+
+ slub_debug=O
+
In case you forgot to enable debugging on the kernel command line: It is
possible to enable debugging manually when the kernel is up. Look at the
contents of: