From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from sog-mx-3.v43.ch3.sourceforge.com ([172.29.43.193] helo=mx.sourceforge.net) by sfs-ml-2.v29.ch3.sourceforge.com with esmtp (Exim 4.76) (envelope-from ) id 1WpgzU-00039W-37 for ltp-list@lists.sourceforge.net; Wed, 28 May 2014 16:46:12 +0000 Received: from sestofw01.enea.se ([192.36.1.252] helo=mx-3.enea.com) by sog-mx-3.v43.ch3.sourceforge.com with smtp (Exim 4.76) id 1WpgzS-0002IU-6H for ltp-list@lists.sourceforge.net; Wed, 28 May 2014 16:46:12 +0000 Date: Wed, 28 May 2014 18:45:34 +0200 From: Mats Liljegren Message-ID: <20140528184534.56c465fc@mats-desktop> In-Reply-To: <1399990269-3383-1-git-send-email-mats.liljegren@enea.com> References: <1398697619-27372-1-git-send-email-mats.liljegren@enea.com> <1399990269-3383-1-git-send-email-mats.liljegren@enea.com> MIME-Version: 1.0 Subject: Re: [LTP] [PATCH v3] partrt_nohz_full: Introducing a new test case List-Id: Linux Test Project General Discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: ltp-list-bounces@lists.sourceforge.net To: Mats Liljegren Cc: ltp-list@lists.sourceforge.net, Kevin Hilman , Frederic Weisbecker Ping? Regards Mats Liljegren On Tue, 13 May 2014 16:11:08 +0200 Mats Liljegren wrote: > Changes from version 2: > - Added "git submodule init" to README > - Scrapped "verbose" command line parameter entirely, it wasn't > implemented anyway > - asprintf() replaced with fprintf() in launch_child() > - Introduced cleanup function parameter to assert_exit_status() and > shell(), making cleanup_entered variable in cleanup() obsolete > - Documented 32 CPU limitation in README. The test code itself has > been updated to handle all configured CPUs, but the partrt tool does > not. > - Simplified tool_available() by using tst_get_path() > - Removed unnecessary exit() and return statements > - Removed rand() to a simple "++" operation as work generator in the > RT partition. Seemed like rand() could cause ticks for some reason. > - Use SAFE_POPEN() instead of popen(). > - Use a 2 second sleep before counting ticks, as a primitive way of > emptying deferred work in the kernel. > > Changes from version 1: > - Style fixes proposed by checkpatch.pl script > - Have rt-tools as a git sub-module instead of requiring it to be > cloned manually, builds can still be done without checking out > rt-tools but you can't run partrt_nohz_full test successfully then > - Check if cpuset supported by kernel, TCONF error if not > - Check if rt-tools has been checked out before build, TCONF error if > not > - Simplified determined_nohz_mask() > - Use parse_opts() instead of getopt_long() when parsing command line > options > - Moved partrt in SUBDIRS so that it is alphabetically correct in > testcases/kernel/Makefile > - launch_child will now handle moving child thread to the right > CPUSET partition and CPU, which also eliminates the need for > synchronization between parent and child > - Children will no longer be able to cause any tst_* calls > - Removed shell_silent(), above changes made it obsolete > - Removed base parameter to str_to_ulong(), will always be 0 now > - setup_children() is now called from main() rather than test() > function. > - Macros defining file names > - Use C code for moving children to the right place instead of > external scripts. Using script invocation from RT partition using NFS > mounted file system will most likely invoke ticks which can make the > test case fail > - Added standard test loop in case reptitions are requested > - Removed atomic operation for cleanup recursion detection, since it > is only trying to protect recursive invocations from itself > - Using timerfd to wait for timeout instead of busy-waiting > - Removed pid_popen(), pid_wait() and shell_int() > - assert_exit_status() now also checks if status is -1, i.e. command > failed > - Moved _GNU_SOURCE macro from Makefile to C file > - Removed -Werror from Makefile > - Removed err(), fail(), pass() and info() macros > - Replaced usage of READ and WRITE macros with STDIN_FILENO and > STDOUT_FILENO, READ and WRITE macros removed > - Removed some unnecessary comments > - setup() is renamed to setup_children() > - Informational message are made mandatory, verbose only affects > output from sub-commands, partrt in particular > - Test case name is now partrt_nohz_full > > partrt_nohz_full is a test case which checks whether nohz_full mode > can achieve 0Hz ticks (if relevant extra patches has been applied) or > 1Hz ticks (mainline kernel). > > Appropriate kernel configurations and kernel boot parameters needed > are documented in testcases/kernel/partrt/README. > > Mats Liljegren (1): > partrt_nohz_full: Introducing a new test case > > .gitmodules | 3 + > README.kernel_config | 6 + > runtest/partrt_nohz_full | 1 + > testcases/kernel/Makefile | 1 + > testcases/kernel/partrt/Makefile | 17 + > testcases/kernel/partrt/README | 66 +++ > testcases/kernel/partrt/nohz_full/.gitignore | 1 + > testcases/kernel/partrt/nohz_full/Makefile | 23 + > .../partrt/nohz_full/test_partrt_nohz_full.c | 593 > ++++++++++++++++++++ > utils/.gitignore | 3 + > utils/Makefile | 7 + > utils/rt-tools | 1 + 12 files > changed, 722 insertions(+) create mode 100644 runtest/partrt_nohz_full > create mode 100644 testcases/kernel/partrt/Makefile > create mode 100644 testcases/kernel/partrt/README > create mode 100644 testcases/kernel/partrt/nohz_full/.gitignore > create mode 100644 testcases/kernel/partrt/nohz_full/Makefile > create mode 100644 > testcases/kernel/partrt/nohz_full/test_partrt_nohz_full.c create mode > 160000 utils/rt-tools > ------------------------------------------------------------------------------ Time is money. Stop wasting it! Get your web API in 5 minutes. www.restlet.com/download http://p.sf.net/sfu/restlet _______________________________________________ Ltp-list mailing list Ltp-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ltp-list