aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorPaul E. McKenney <paulmck@linux.vnet.ibm.com>2013-09-10 11:26:09 -0400
committerPaul E. McKenney <paulmck@linux.vnet.ibm.com>2013-09-25 09:49:46 -0400
commit64f26e5c86af9ce8615721340c8282f2b148c9aa (patch)
tree31bfe356b8a59dac492aa38f22b83fe20717bcb5
parent64d3b7a1d5289486df2d8bce36e23ed5ebc80a3d (diff)
kthread: Add pointer to vmstat-avoidance patch
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Reviewed-by: Josh Triplett <josh@joshtriplett.org>
-rw-r--r--Documentation/kernel-per-CPU-kthreads.txt17
1 files changed, 11 insertions, 6 deletions
diff --git a/Documentation/kernel-per-CPU-kthreads.txt b/Documentation/kernel-per-CPU-kthreads.txt
index 32351bfabf20..827104fb9364 100644
--- a/Documentation/kernel-per-CPU-kthreads.txt
+++ b/Documentation/kernel-per-CPU-kthreads.txt
@@ -181,12 +181,17 @@ To reduce its OS jitter, do any of the following:
181 make sure that this is safe on your particular system. 181 make sure that this is safe on your particular system.
182 d. It is not possible to entirely get rid of OS jitter 182 d. It is not possible to entirely get rid of OS jitter
183 from vmstat_update() on CONFIG_SMP=y systems, but you 183 from vmstat_update() on CONFIG_SMP=y systems, but you
184 can decrease its frequency by writing a large value to 184 can decrease its frequency by writing a large value
185 /proc/sys/vm/stat_interval. The default value is HZ, 185 to /proc/sys/vm/stat_interval. The default value is
186 for an interval of one second. Of course, larger values 186 HZ, for an interval of one second. Of course, larger
187 will make your virtual-memory statistics update more 187 values will make your virtual-memory statistics update
188 slowly. Of course, you can also run your workload at 188 more slowly. Of course, you can also run your workload
189 a real-time priority, thus preempting vmstat_update(). 189 at a real-time priority, thus preempting vmstat_update(),
190 but if your workload is CPU-bound, this is a bad idea.
191 However, there is an RFC patch from Christoph Lameter
192 (based on an earlier one from Gilad Ben-Yossef) that
193 reduces or even eliminates vmstat overhead for some
194 workloads at https://lkml.org/lkml/2013/9/4/379.
190 e. If running on high-end powerpc servers, build with 195 e. If running on high-end powerpc servers, build with
191 CONFIG_PPC_RTAS_DAEMON=n. This prevents the RTAS 196 CONFIG_PPC_RTAS_DAEMON=n. This prevents the RTAS
192 daemon from running on each CPU every second or so. 197 daemon from running on each CPU every second or so.