diff options
| -rw-r--r-- | labs/lock.html | 18 |
1 files changed, 11 insertions, 7 deletions
diff --git a/labs/lock.html b/labs/lock.html index b43a51b..a93eb3a 100644 --- a/labs/lock.html +++ b/labs/lock.html @@ -82,7 +82,7 @@ workloads. <p>Run usertests to see if you don't break anything. -<h2>Lock-free bcache lookup</h2> +<h2>More scalabale bcache lookup</h2> <p>Several processes reading different files repeatedly will @@ -116,15 +116,19 @@ against, including: <p>A challenge is testing whether you code is still correct. One way to do is to artificially delay certain operations - using <tt>sleepticks</tt>. + using <tt>sleepticks</tt>. <tt>test1</tt> trashes the buffer cache + and exercises more code paths. <p>Here are some hints: <ul> - <li> Use an atomic increment instruction for incrementing and - decrementing <tt>b->ref</tt> (see <tt>__sync_fetch_and_add() and - related primitives</tt>) - <li>Don't walk the <tt>bcache.head</tt> list to find a buffer + <li>Use a simple design: i.e., don't design a lock-free implementation. + <li>Use a simple hash table with locks per bucket </ul> - + +<p>Check that your implementation has less contention + on <tt>test0</tt> + +<p>Make sure your implementation passes bcachetest and usertests. + </body> </html> |