[newlib-cygwin] Cygwin: testsuite: Busy-wait in cancel3 and cancel5
Jon Turney
jturney@sourceware.org
Fri Jul 14 15:28:28 GMT 2023
https://sourceware.org/git/gitweb.cgi?p=newlib-cygwin.git;h=e867f8c3e89f4f6a85de3a96ddfd2de8f4bb882f
commit e867f8c3e89f4f6a85de3a96ddfd2de8f4bb882f
Author: Jon Turney <jon.turney@dronecode.org.uk>
Date: Mon Jul 10 19:04:58 2023 +0100
Cygwin: testsuite: Busy-wait in cancel3 and cancel5
These tests async thread cancellation of a thread that doesn't have any
cancellation points.
Unfortunately, since 2b165a45 the async cancellation silently fails when
the thread is inside the kernel function Sleep(), so it just exits
normally after 10 seconds. (See the commentary in pthread::cancel() in
thread.cc, where it checks if the target thread is inside the kernel,
and silently converts the cancellation into a deferred one)
Work around this by busy-waiting rather than Sleep()ing for 10 seconds.
This is still somewhat fragile: the async cancel could still fail, if it
happens to occur while we're inside the kernel function that time()
calls.
v2:
Do nothing more efficiently
Signed-off-by: Jon Turney <jon.turney@dronecode.org.uk>
Diff:
---
winsup/testsuite/winsup.api/pthread/cancel3.c | 24 ++++++++++++++++++------
winsup/testsuite/winsup.api/pthread/cancel5.c | 24 ++++++++++++++++++------
2 files changed, 36 insertions(+), 12 deletions(-)
diff --git a/winsup/testsuite/winsup.api/pthread/cancel3.c b/winsup/testsuite/winsup.api/pthread/cancel3.c
index 832fe2e3f..07feb7c9b 100644
--- a/winsup/testsuite/winsup.api/pthread/cancel3.c
+++ b/winsup/testsuite/winsup.api/pthread/cancel3.c
@@ -75,11 +75,22 @@ mythread(void * arg)
assert(pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL) == 0);
/*
- * We wait up to 10 seconds, waking every 0.1 seconds,
- * for a cancelation to be applied to us.
+ * We wait up to 10 seconds for a cancelation to be applied to us.
*/
- for (bag->count = 0; bag->count < 100; bag->count++)
- Sleep(100);
+ for (bag->count = 0; bag->count < 10; bag->count++)
+ {
+ /* Busy wait to avoid Sleep(), since we can't asynchronous cancel inside a
+ kernel function. (This is still somewhat fragile as if the async cancel
+ can fail if it happens to occur while we're inside the kernel function
+ that time() calls...) */
+ time_t start = time(NULL);
+ while ((time(NULL) - start) < 1)
+ {
+ int i;
+ for (i = 0; i < 1E7; i++)
+ __asm__ volatile ("pause":::);
+ }
+ }
return result;
}
@@ -149,10 +160,11 @@ main()
if (fail)
{
- fprintf(stderr, "Thread %d: started %d: count %d\n",
+ fprintf(stderr, "Thread %d: started %d: count %d: result %d \n",
i,
threadbag[i].started,
- threadbag[i].count);
+ threadbag[i].count,
+ result);
}
failed = (failed || fail);
}
diff --git a/winsup/testsuite/winsup.api/pthread/cancel5.c b/winsup/testsuite/winsup.api/pthread/cancel5.c
index 8b7240615..999b3c95c 100644
--- a/winsup/testsuite/winsup.api/pthread/cancel5.c
+++ b/winsup/testsuite/winsup.api/pthread/cancel5.c
@@ -76,11 +76,22 @@ mythread(void * arg)
assert(pthread_setcanceltype(PTHREAD_CANCEL_ASYNCHRONOUS, NULL) == 0);
/*
- * We wait up to 10 seconds, waking every 0.1 seconds,
- * for a cancelation to be applied to us.
+ * We wait up to 10 seconds for a cancelation to be applied to us.
*/
- for (bag->count = 0; bag->count < 100; bag->count++)
- Sleep(100);
+ for (bag->count = 0; bag->count < 10; bag->count++)
+ {
+ /* Busy wait to avoid Sleep(), since we can't asynchronous cancel inside a
+ kernel function. (This is still somewhat fragile as if the async cancel
+ can fail if it happens to occur while we're inside the kernel function
+ that time() calls...) */
+ time_t start = time(NULL);
+ while ((time(NULL) - start) < 1)
+ {
+ int i;
+ for (i = 0; i < 1E7; i++)
+ __asm__ volatile ("pause":::);
+ }
+ }
return result;
}
@@ -148,10 +159,11 @@ main()
if (fail)
{
- fprintf(stderr, "Thread %d: started %d: count %d\n",
+ fprintf(stderr, "Thread %d: started %d: count %d: result %d\n",
i,
threadbag[i].started,
- threadbag[i].count);
+ threadbag[i].count,
+ result);
}
failed = (failed || fail);
}
More information about the Cygwin-cvs
mailing list