Hi,
you have a race in the testcase, attached patch removes the race and it no
longer fails for me afterwards on kernel-2.6.25-14.fc9.x86_64 .
The while-loop last command there is
ptrace(PTRACE_SYSCALL, child, 0, 0);
and then you try to do
ptrace(PTRACE_CONT, child, 0, SIGKILL);
You must do waitpid() in between as otherwise CHILD may still be running (and
not stopped). It is a correct behavior ptrace(2) returns ESRCH if CHILD is
still not stopped.
If you do
ptrace(PTRACE_KILL, child, 0,0);
there is no race as this command works in any state of CHILD - either already
stopped or still running.
Anyway thanks for testing ptrace/utrace,
Jan
--- ptrace-syscall-ret.c-orig 2008-05-15 18:44:29.000000000 +0200
+++ ptrace-syscall-ret.c 2008-05-15 18:46:18.000000000 +0200
@@ -67,20 +67,18 @@ int main(int argc, char **argv) {
assert(WIFSTOPPED(status));
assert(WSTOPSIG(status) == SIGSTOP);
- l = ptrace(PTRACE_SYSCALL,child, 0, 0);
- assert(l==0);
-
j = 0;
while (j < LOOP){
+ l = ptrace(PTRACE_SYSCALL, child, 0, 0);
+ assert(l==0);
+
pid = waitpid(child, &status, 0);
assert(child==pid);
assert(WIFSTOPPED(status));
assert(WSTOPSIG(status) == SIGTRAP);
- l = ptrace(PTRACE_SYSCALL, child, 0, 0);
- assert(l==0);
j++;
} //end loop
