linter: document benchmarking

Change-Id: I274de96698cf7f54295ec78cf409b62a453dae1c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/391681
Reviewed-by: Brian Wilkerson <[email protected]>
Reviewed-by: Phil Quitslund <[email protected]>
Commit-Queue: Samuel Rawlins <[email protected]>

Author: srawlins

pkg/linter/CONTRIBUTING.md

@@ -82,6 +82,39 @@ entry. Notably, any change that:
 should have a short entry in the `CHANGELOG`. Feel free to bring up any
 questions in your PR.
 
+### Benchmarking
+
+Lint rules can be benchmarked with real code on disk. (There is no
+micro-benchmarking suite.) Use the `tool/benchmark.dart` script to execute one
+or more lint rules against a corpus of code. Here are some examples:
+
+Execute all of the lint rules enabled by a specified analysis options file
+against all of the code in the specified directory.
+
+```none
+dart tool/benchmark.dart \
+    --config $HOME/my/example/analysis_options.yaml \
+    $HOME/my/example/package
+```
+
+Execute all of the lint rules enabled by a specified analysis options file
+against all of the code in the specified file.
+
+```none
+dart tool/benchmark.dart \
+    --config $HOME/my/example/analysis_options.yaml \
+    $HOME/my/example/package/lib/some_file.dart
+```
+
+Execute all of a set of specified lint rules against all of the code in the
+specified directory.
+
+```none
+dart tool/benchmark.dart \
+    --rules rule_1,rule_2 \
+    $HOME/my/example/package
+```
+
 **Thank you!**
 
 ### The small print