https://bugs.kde.org/show_bug.cgi?id=405394
--- Comment #3 from Daniel Wagner <w...@monom.org> ---
With PRE_MEM_READ valgrind will complain for code like this (which should be
fine, right?):
void run_deadline() {
struct sched_attr attr;
ret = sched_getattr(0, &attr, sizeof(attr), 0);
}
==11537== Thread 2:
==11537== Syscall param sched_getattr(p) points to uninitialised byte(s)
==11537== at 0x49A1EFD: syscall (in /usr/lib64/libc-2.28.so)
==11537== by 0x403E68: run_deadline (cyclicdeadline.c:784)
==11537== by 0x489058D: start_thread (in /usr/lib64/libpthread-2.28.so)
==11537== by 0x49A76A2: clone (in /usr/lib64/libc-2.28.so)
==11537== Address 0x5671ea0 is on thread 2's stack
==11537== in frame #1, created by run_deadline (cyclicdeadline.c:772)
==11537==
No complains with PRE_MEM_WRITE.
--
You are receiving this mail because:
You are watching all bug changes.
