summaryrefslogtreecommitdiffstats
path: root/tests
diff options
context:
space:
mode:
authorRoy Marples <roy@marples.name>2017-04-13 14:54:18 +0100
committerRoy Marples <roy@marples.name>2017-04-13 14:54:18 +0100
commit2d31c0f69834ce3e507b3f32238f4595df6bc7cc (patch)
treef0c18b6b6b5745b8e1a17129f0742dc90101a62c /tests
parentea06be506a7cb14d6efca20b67160197d13948a8 (diff)
downloaddhcpcd-2d31c0f69834ce3e507b3f32238f4595df6bc7cc.tar.xz
Tart up some.
Diffstat (limited to 'tests')
-rw-r--r--tests/eloop-bench/README.md13
1 files changed, 5 insertions, 8 deletions
diff --git a/tests/eloop-bench/README.md b/tests/eloop-bench/README.md
index e141045c..84a84a72 100644
--- a/tests/eloop-bench/README.md
+++ b/tests/eloop-bench/README.md
@@ -16,6 +16,7 @@ by giving one of these CPPFLAGS to the Makefile:
* HAVE_PSELECT
* HAVE_POLLTS
* HAVE_PPOLL
+Otherwise, eloop will try and work out wich mechanism to use.
kqueue(2) is found on modern BSD kernels.
epoll(2) is found on modern Linux and Solaris kernels.
@@ -44,14 +45,10 @@ At the end of run, the time taken in seconds and nanoseconds is printed.
The following arguments can influence the benchmark:
* `-a active`
- The number of active pipes.
- (default 1)
+ The number of active pipes, default 1.
* `-n pipes`
- The number of pipes to create and attach an eloop callback to.
- (default (100)
+ The number of pipes to create and attach an eloop callback to, defalt 100.
* `-r runs`
- The number of timed runs to make.
- (default 25)
+ The number of timed runs to make, default 25.
* `-w writes`
- The number of writes to make by the read callbacks.
- (default 100)
+ The number of writes to make by the read callback, default 100.