Merge branch 'master' into next

1 files changed, 32 insertions, 0 deletions

diff --git a/Documentation/sysctl/vm.txt b/Documentation/sysctl/vm.txt
index 3197fc83bc51..b716d33912d8 100644
--- a/Documentation/sysctl/vm.txt
+++ b/Documentation/sysctl/vm.txt
@@ -39,6 +39,8 @@ Currently, these files are in /proc/sys/vm:
 - nr_hugepages
 - nr_overcommit_hugepages
 - nr_pdflush_threads
+- nr_pdflush_threads_min
+- nr_pdflush_threads_max
 - nr_trim_pages         (only if CONFIG_MMU=n)
 - numa_zonelist_order
 - oom_dump_tasks
@@ -88,6 +90,10 @@ will itself start writeback.
 If dirty_bytes is written, dirty_ratio becomes a function of its value
 (dirty_bytes / the amount of dirtyable system memory).
 
+Note: the minimum value allowed for dirty_bytes is two pages (in bytes); any
+value lower than this limit will be ignored and the old configuration will be
+retained.
+
 ==============================================================
 
 dirty_expire_centisecs
@@ -463,6 +469,32 @@ The default value is 0.
 
 ==============================================================
 
+nr_pdflush_threads_min
+
+This value controls the minimum number of pdflush threads.
+
+At boot time, the kernel will create and maintain 'nr_pdflush_threads_min'
+threads for the kernel's lifetime.
+
+The default value is 2.  The minimum value you can specify is 1, and
+the maximum value is the current setting of 'nr_pdflush_threads_max'.
+
+See 'nr_pdflush_threads_max' below for more information.
+
+==============================================================
+
+nr_pdflush_threads_max
+
+This value controls the maximum number of pdflush threads that can be
+created.  The pdflush algorithm will create a new pdflush thread (up to
+this maximum) if no pdflush threads have been available for >= 1 second.
+
+The default value is 8.  The minimum value you can specify is the
+current value of 'nr_pdflush_threads_min' and the
+maximum is 1000.
+
+==============================================================
+
 overcommit_memory:
 
 This value contains a flag that enables memory overcommitment.