Copyright (c) 1984 The Regents of the University of California.
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
3. Neither the name of the University nor the names of its contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
SUCH DAMAGE.
Conclusions
We have created a profiler that aids in the evaluation of the kernel. For each routine in the kernel, the profile shows the extent to which that routine helps support various abstractions, and how that routine uses other abstractions. The profile assesses the cost of routines at all levels of the kernel decomposition. The profiler is easily used, and can be compiled into the kernel. It adds only five to thirty percent execution overhead to the kernel being profiled, produces no additional output while the kernel is running and allows the kernel to be measured in its real environment. Kernel profiles can be used to identify bottlenecks in performance. We have shown how to improve performance by caching recently calculated name translations. The combined caches added to the name translation process reduce the average cost of translating a pathname to an inode by 35%. These changes reduce the percentage of time spent running in the system by nearly 9%. .nr H2 1
\s+2Acknowledgements\s0I would like to thank Robert Elz for sharing his ideas and his code for cacheing system wide names. Thanks also to all the users at Berkeley who provided all the input to generate the kernel profiles. This work was supported by the Defense Advance Research Projects Agency (DoD) under Arpa Order No. 4031 monitored by Naval Electronic System Command under Contract No. N00039-82-C-0235. .nr H2 1
\s+2References\s-2