[bug report] perf lock: Print the number of lost entries for BPF

[Dan Carpenter <dan.carpenter@linaro.org>]

Hello Namhyung Kim,

Commit 6d499a6b3d90 ("perf lock: Print the number of lost entries for
BPF") from Aug 2, 2022 (linux-next), leads to the following Smatch
static checker warning:

	./tools/perf/util/bpf_skel/lock_contention.bpf.c:379 contention_begin()
	warn: unsigned 'pelem->stack_id' is never less than zero.

./tools/perf/util/bpf_skel/lock_contention.bpf.c
    360 SEC("tp_btf/contention_begin")
    361 int contention_begin(u64 *ctx)
    362 {
    363         struct tstamp_data *pelem;
    364 
    365         if (!enabled || !can_record(ctx))
    366                 return 0;
    367 
    368         pelem = get_tstamp_elem(ctx[1]);
    369         if (pelem == NULL)
    370                 return 0;
    371 
    372         pelem->timestamp = bpf_ktime_get_ns();
    373         pelem->lock = (__u64)ctx[0];
    374         pelem->flags = (__u32)ctx[1];
    375 
    376         if (needs_callstack) {
    377                 pelem->stack_id = bpf_get_stackid(ctx, &stacks,
    378                                                   BPF_F_FAST_STACK_CMP | stack_skip);
--> 379                 if (pelem->stack_id < 0)

->stack_id is a u32 so it can't be less than zero

    380                         __sync_fetch_and_add(&stack_fail, 1);
    381         } else if (aggr_mode == LOCK_AGGR_TASK) {
    382                 struct task_struct *task;
    383 
    384                 if (lock_owner) {
    385                         task = get_lock_owner(pelem->lock, pelem->flags);
    386 
    387                         /* The flags is not used anymore.  Pass the owner pid. */
    388                         if (task)
    389                                 pelem->flags = BPF_CORE_READ(task, pid);

regards,
dan carpenter