Fix the edge-case of MAXTRYLOCK=0

We didn't really have good semantics for what is meant by MAXTRYLOCK=0,
so when skipped.exp tried it, we ended up locking the variable and then
reporting a skip without ever unlocking it.

This is now cleaning up the semantics such that MAXTRYLOCK defines how
many times we should loop if the lock is busy.  Thus MAXTRYLOCK=0 means
we try only once and fail immediately.

The testcase was accidentally creating contention due to the broken
unlock behavior.  We now have to try a bit harder to create real
contention, so some lengthy delays are inserted to hoard the lock.

* runtime/probe_lock.h (stp_probe_lock): Fix the skip behavior.
* testsuite/systemtap.base/skipped.exp: Add a big udelay.

1 files changed, 7 insertions, 1 deletions

diff --git a/testsuite/systemtap.base/skipped.exp b/testsuite/systemtap.base/skipped.exp
index 8815c737..f12bc6f1 100644
--- a/testsuite/systemtap.base/skipped.exp
+++ b/testsuite/systemtap.base/skipped.exp
@@ -5,10 +5,16 @@ if {! [installtest_p]} { untested $test; return }
 set nr_cpus [exec sh -c "grep ^processor /proc/cpuinfo | wc -l"]
 if {$nr_cpus < 2} { unsupported $test; return }
 
+set script {
+global f
+probe timer.profile { f++; snooze() }
+function snooze() %{ udelay(10000); %}
+}
+
 set errs 0
 set warns 0
 set oks 0
-spawn stap -e "probe timer.s(5) {exit()} probe timer.profile,syscall.* {f++} global f" -DMAXTRYLOCK=0 -tu
+spawn stap -e $script -DMAXTRYLOCK=0 -DSTP_NO_OVERLOAD -tug
 expect {
     -timeout 60
     -re {^ERROR: Skipped too many probes, check MAXSKIPPED or try again with stap -t for more details.\r\n} { incr errs; exp_continue }