| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
| |
It seems better to not use any random ASAN variable.
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
| |
Signed-off-by: Ruediger Meier <ruediger.meier@ga-group.nl>
|
|
|
|
|
|
| |
We have to wait for the killed helper.
Signed-off-by: Ruediger Meier <ruediger.meier@ga-group.nl>
|
|
|
|
|
|
|
| |
Sending signal indicating invalid memory reference makes AddressSanitizer
to report false positive test failure.
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Correction. The final failure to catch signal exit should not use
exiter() function.
--->8----
From: Sami Kerola <kerolasa@iki.fi>
Date: Mon, 14 Jul 2014 18:35:56 +0100
Subject: [PATCH] tests: add kill --queue check
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The sigqueue(3) takes two values, signal and sigval. Contents of the
signal can be altered with --signal option argument, so the --queue
argument should be reserved to affect sigval_int.
This is regression fix introduced by commit
9e8dffd5cd29f03029b1ac99eecb129532ca5c0f.
Reference: http://man7.org/linux/man-pages/man3/sigqueue.3.html
Reported-by: Benno Schulenberg <bensberg@justemail.net>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Unavailability of /proc is fatal for kill, and continuing with the test
in that case does not make sense as it will only mean false positive
errors.
Where /proc/<pid>/status file(s) does not exist the check will perform
opportunistic sleep with assumption the test_sigreceive will be ready to
be killed if it has some time to init.
[kzak@redhat.com: - remove dependence on gawk, just use shell
- fix typo in "test /proc"]
CC: Ruediger Meier <sweet_f_a@gmx.de>
Reference: https://travis-ci.org/rudimeier/util-linux/jobs/24561058
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
There were several failures (missing tmp files, missing pids) on travis
build farm which I couldn't track down completely. Here we fix some
possible issues:
- mktemp -u is unsafe, maybe even in practice if there is a bad random
generator
- make sure that mktemp does not give us pure integer filnames which
would not work for "kill by-name"
- avoid using trap, could be that it removed tmp files on wrong signals
Signed-off-by: Ruediger Meier <ruediger.meier@ga-group.nl>
|
|
|
|
| |
Signed-off-by: Ruediger Meier <ruediger.meier@ga-group.nl>
|
|
|
|
| |
Signed-off-by: Karel Zak <kzak@redhat.com>
|
|
|
|
|
| |
CC: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|
|
|
|
|
| |
CC: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|
|
|
|
|
| |
CC: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|
|
CC: Bernhard Voelker <mail@bernhard-voelker.de>
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
|