[PATCH] test/triggers: ensure bogus-ioctl has a reasonable fd limit

[PATCH] test/triggers: ensure bogus-ioctl has a reasonable fd limit

[Kris Van Hees]

The bogus-ioctl trigger has a loop to close all possible fds, and it
was limit by rlim_max.  On some systems, that was such a high limit
that the loop was taking too long, causing tests to time out.  Set
the limit to 1024, so that the loop will end in reasonable time.

Signed-off-by: Kris Van Hees <kris.van.hees@oracle.com>
---
 test/triggers/bogus-ioctl.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/test/triggers/bogus-ioctl.c b/test/triggers/bogus-ioctl.c
index 25933229..f1546445 100644
--- a/test/triggers/bogus-ioctl.c
+++ b/test/triggers/bogus-ioctl.c
@@ -38,7 +38,10 @@ main(void)
 	sigaction(SIGUSR1, &act, NULL);
 
 	getrlimit(RLIMIT_NOFILE, &rl);
-	for (i = 0; i < rl.rlim_max; i++)
+	rl.rlim_cur = 1024;
+	setrlimit(RLIMIT_NOFILE, &rl);
+	getrlimit(RLIMIT_NOFILE, &rl);
+	for (i = 0; i < rl.rlim_cur; i++)
 		close(i);
 	n = 0;
 
-- 
2.45.2

Re: [DTrace-devel] [PATCH] test/triggers: ensure bogus-ioctl has a reasonable fd limit

[Nick Alcock]

On 31 Jul 2024, Kris Van Hees via DTrace-devel spake thusly:

> The bogus-ioctl trigger has a loop to close all possible fds, and it
> was limit by rlim_max.  On some systems, that was such a high limit
> that the loop was taking too long, causing tests to time out.  Set
> the limit to 1024, so that the loop will end in reasonable time.

Boggle. I guess this has grown a lot with time :)

> Signed-off-by: Kris Van Hees <kris.van.hees@oracle.com>

Signed-off-by: Nick Alcock <nick.alcock@oracle.com>

(in practice, probably closing the first three would work almost all the
time. I can't imagine any shell weird enough to leave a thousand fds
open on entry!)

-- 
NULL && (void)

Re: [DTrace-devel] [PATCH] test/triggers: ensure bogus-ioctl has a reasonable fd limit

[Nick Alcock]

On 31 Jul 2024, Nick Alcock uttered the following:

> On 31 Jul 2024, Kris Van Hees via DTrace-devel spake thusly:
>
>> The bogus-ioctl trigger has a loop to close all possible fds, and it
>> was limit by rlim_max.  On some systems, that was such a high limit
>> that the loop was taking too long, causing tests to time out.  Set
>> the limit to 1024, so that the loop will end in reasonable time.
>
> Boggle. I guess this has grown a lot with time :)
>
>> Signed-off-by: Kris Van Hees <kris.van.hees@oracle.com>
>
> Signed-off-by: Nick Alcock <nick.alcock@oracle.com>

Er, I mean

Reviewed-by: Nick Alcock <nick.alcock@oracle.com> of course.

The heat is clearly destroying my mind...

-- 
NULL && (void)