diff options
| author | Harald Welte <laforge@gnumonks.org> | 2016-12-09 17:58:17 +0100 |
|---|---|---|
| committer | Harald Welte <laforge@gnumonks.org> | 2016-12-09 17:59:58 +0100 |
| commit | a24be85dc41885ee77072af9f079c69c575061e9 (patch) | |
| tree | 2c1591987d0e7f08e81412bd1ae9548bcb822445 | |
| parent | 665d48bf15602e8666b292dbe66881786af2ea95 (diff) | |
Fix msgb_test on 64 bit architectures
truncating the unsigned long pointer msg->data to 'int' and then passin
git into msgb_resize_area() is unsafe as depending on the 32rd address
bit it will be eiether negative or positive. That will in turn change
the expected "Sub area is not fully contained in the msg data\n" error
message into "Negative sizes are not allowed\n" which is not what the
autotest case expects.
Change-Id: I87ce13c265704d4ba8724e7dc7ed874c1128e0fa
| -rw-r--r-- | tests/msgb/msgb_test.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/tests/msgb/msgb_test.c b/tests/msgb/msgb_test.c index a726ef4d..ac103829 100644 --- a/tests/msgb/msgb_test.c +++ b/tests/msgb/msgb_test.c @@ -195,7 +195,7 @@ static void test_msgb_resize_area() OSMO_ASSERT(e != 0); if (OSMO_PANIC_TRY(&e)) - msgb_resize_area(msg, NULL, (int)msg->data, 0); + msgb_resize_area(msg, NULL, 0, 0); OSMO_ASSERT(e != 0); if (OSMO_PANIC_TRY(&e)) |