* change to cyclictest -h option?
@ 2011-09-29 14:28 Clark Williams
0 siblings, 0 replies; only message in thread
From: Clark Williams @ 2011-09-29 14:28 UTC (permalink / raw)
To: Carsten Emde; +Cc: Frank Rowand, RT
[-- Attachment #1.1: Type: text/plain, Size: 976 bytes --]
Carsten,
I'm considering applying the following patch to cyclictest, which will
change an implied behavior for the -h/--histogram option. Currently
when you specify -h without --smp or --numa, any realtime priority
specified is spread across the range of measurement threads. This patch
makes that spread an explicit option, rather than implicit in the
histogram option.
Previous command line:
$ sudo cyclictest --h -t -p90
New command line:
$ sudo cyclictest --h -t -p90 -Q
Where the -Q/--priospread option causes the priority assigned to
measurement threads to be decremented after each thread is created.
With this option in place, the first command line would create a
measurement thread for each core, all running at fifo:90.
Is this going to cause you pain and heartache in your automated tests?
Clark
P.S. Yes, I know that -Q makes no sense, but I couldn't come up with a
one-letter option that did, that wasn't already taken :)
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1.2: priospread.patch --]
[-- Type: text/x-patch, Size: 2638 bytes --]
diff --git a/src/cyclictest/cyclictest.c b/src/cyclictest/cyclictest.c
index 731b4bd..c0cde4a 100644
--- a/src/cyclictest/cyclictest.c
+++ b/src/cyclictest/cyclictest.c
@@ -170,6 +170,7 @@ static int duration = 0;
static int use_nsecs = 0;
static int refresh_on_max;
static int force_sched_other;
+static int priospread = 0;
static pthread_cond_t refresh_on_max_cond = PTHREAD_COND_INITIALIZER;
static pthread_mutex_t refresh_on_max_lock = PTHREAD_MUTEX_INITIALIZER;
@@ -845,6 +846,7 @@ static void display_help(int error)
"-p PRIO --prio=PRIO priority of highest prio thread\n"
"-P --preemptoff Preempt off tracing (used with -b)\n"
"-q --quiet print only a summary on exit\n"
+ "-Q --priospread spread priority levels starting at specified value\n"
"-r --relative use relative timer instead of absolute\n"
"-s --system use sys_nanosleep and sys_setitimer\n"
"-t --threads one thread per available processor\n"
@@ -984,9 +986,10 @@ static void process_options (int argc, char *argv[])
{"smp", no_argument, NULL, 'S'},
{"numa", no_argument, NULL, 'U'},
{"latency", required_argument, NULL, 'e'},
+ {"priospread", no_argument, NULL, 'Q'},
{NULL, 0, NULL, 0}
};
- int c = getopt_long(argc, argv, "a::b:Bc:Cd:Efh:H:i:Il:MnNo:O:p:PmqrsSt::uUvD:wWT:y:e:",
+ int c = getopt_long(argc, argv, "a::b:Bc:Cd:Efh:H:i:Il:MnNo:O:p:PmqQrsSt::uUvD:wWT:y:e:",
long_options, &option_index);
if (c == -1)
break;
@@ -1045,6 +1048,7 @@ static void process_options (int argc, char *argv[])
}
break;
case 'q': quiet = 1; break;
+ case 'Q': priospread = 1; break;
case 'r': timermode = TIMER_RELTIME; break;
case 's': use_system = MODE_SYS_OFFSET; break;
case 't':
@@ -1140,6 +1144,12 @@ static void process_options (int argc, char *argv[])
if (priority < 0 || priority > 99)
error = 1;
+ if (priospread && priority == 0) {
+ fprintf(stderr, "defaulting realtime priority to %d\n",
+ num_threads+1);
+ priority = num_threads+1;
+ }
+
if (priority && (policy != SCHED_FIFO && policy != SCHED_RR)) {
fprintf(stderr, "policy and priority don't match: setting policy to SCHED_FIFO\n");
policy = SCHED_FIFO;
@@ -1456,7 +1466,7 @@ int main(int argc, char **argv)
par->policy = SCHED_OTHER;
force_sched_other = 1;
}
- if (priority && !histogram && !smp && !numa)
+ if (priospread)
priority--;
par->clock = clocksources[clocksel];
par->mode = mode;
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 198 bytes --]
^ permalink raw reply related [flat|nested] only message in thread
only message in thread, other threads:[~2011-09-29 14:28 UTC | newest]
Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2011-09-29 14:28 change to cyclictest -h option? Clark Williams
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.